Merge remote-tracking branch 'airlied/drm-next' into topic/drm-misc

Backmerge drm-next to be able to apply Chris' connector_unregister_all
cleanup (need latest i915 and sun4i state for that).

Also there's a trivial conflict in ttm_bo.c that git rerere fails to
remember.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

439 files changed, 19131 insertions, 10070 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index da3d02154fa6..eb09037a7161 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -85,8 +85,12 @@ extern int amdgpu_vm_debug;
 extern int amdgpu_sched_jobs;
 extern int amdgpu_sched_hw_submission;
 extern int amdgpu_powerplay;
+extern int amdgpu_powercontainment;
 extern unsigned amdgpu_pcie_gen_cap;
 extern unsigned amdgpu_pcie_lane_cap;
+extern unsigned amdgpu_cg_mask;
+extern unsigned amdgpu_pg_mask;
+extern char *amdgpu_disable_cu;
 
 #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS	        3000
 #define AMDGPU_MAX_USEC_TIMEOUT			100000	/* 100 ms */
@@ -183,6 +187,10 @@ int amdgpu_set_clockgating_state(struct amdgpu_device *adev,
 int amdgpu_set_powergating_state(struct amdgpu_device *adev,
 				  enum amd_ip_block_type block_type,
 				  enum amd_powergating_state state);
+int amdgpu_wait_for_idle(struct amdgpu_device *adev,
+			 enum amd_ip_block_type block_type);
+bool amdgpu_is_idle(struct amdgpu_device *adev,
+		    enum amd_ip_block_type block_type);
 
 struct amdgpu_ip_block_version {
 	enum amd_ip_block_type type;
@@ -594,11 +602,9 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
 		     struct amdgpu_sync *sync,
 		     struct reservation_object *resv,
 		     void *owner);
-bool amdgpu_sync_is_idle(struct amdgpu_sync *sync);
-int amdgpu_sync_cycle_fences(struct amdgpu_sync *dst, struct amdgpu_sync *src,
-			     struct fence *fence);
+struct fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
+				     struct amdgpu_ring *ring);
 struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
-int amdgpu_sync_wait(struct amdgpu_sync *sync);
 void amdgpu_sync_free(struct amdgpu_sync *sync);
 int amdgpu_sync_init(void);
 void amdgpu_sync_fini(void);
@@ -754,12 +760,11 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
 int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size,
 			     struct amdgpu_job **job);
 
+void amdgpu_job_free_resources(struct amdgpu_job *job);
 void amdgpu_job_free(struct amdgpu_job *job);
-void amdgpu_job_free_func(struct kref *refcount);
 int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
 		      struct amd_sched_entity *entity, void *owner,
 		      struct fence **f);
-void amdgpu_job_timeout_func(struct work_struct *work);
 
 struct amdgpu_ring {
 	struct amdgpu_device		*adev;
@@ -771,8 +776,6 @@ struct amdgpu_ring {
 	struct amdgpu_bo	*ring_obj;
 	volatile uint32_t	*ring;
 	unsigned		rptr_offs;
-	u64			next_rptr_gpu_addr;
-	volatile u32		*next_rptr_cpu_addr;
 	unsigned		wptr;
 	unsigned		wptr_old;
 	unsigned		ring_size;
@@ -791,7 +794,6 @@ struct amdgpu_ring {
 	u32			doorbell_index;
 	bool			use_doorbell;
 	unsigned		wptr_offs;
-	unsigned		next_rptr_offs;
 	unsigned		fence_offs;
 	uint64_t		current_ctx;
 	enum amdgpu_ring_type	type;
@@ -799,6 +801,9 @@ struct amdgpu_ring {
 	unsigned		cond_exe_offs;
 	u64				cond_exe_gpu_addr;
 	volatile u32	*cond_exe_cpu_addr;
+#if defined(CONFIG_DEBUG_FS)
+	struct dentry *ent;
+#endif
 };
 
 /*
@@ -861,6 +866,7 @@ struct amdgpu_vm {
 	struct amdgpu_bo	*page_directory;
 	unsigned		max_pde_used;
 	struct fence		*page_directory_fence;
+	uint64_t		last_eviction_counter;
 
 	/* array of page tables, one for each page directory entry */
 	struct amdgpu_vm_pt	*page_tables;
@@ -883,13 +889,14 @@ struct amdgpu_vm_id {
 	struct fence		*first;
 	struct amdgpu_sync	active;
 	struct fence		*last_flush;
-	struct amdgpu_ring      *last_user;
 	atomic64_t		owner;
 
 	uint64_t		pd_gpu_addr;
 	/* last flushed PD/PT update */
 	struct fence		*flushed_updates;
 
+	uint32_t                current_gpu_reset_count;
+
 	uint32_t		gds_base;
 	uint32_t		gds_size;
 	uint32_t		gws_base;
@@ -905,6 +912,10 @@ struct amdgpu_vm_manager {
 	struct list_head			ids_lru;
 	struct amdgpu_vm_id			ids[AMDGPU_NUM_VM];
 
+	/* Handling of VM fences */
+	u64					fence_context;
+	unsigned				seqno[AMDGPU_MAX_RINGS];
+
 	uint32_t				max_pfn;
 	/* vram base address for page table entry  */
 	u64					vram_base_offset;
@@ -926,17 +937,14 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 			 struct list_head *validated,
 			 struct amdgpu_bo_list_entry *entry);
-void amdgpu_vm_get_pt_bos(struct amdgpu_vm *vm, struct list_head *duplicates);
+void amdgpu_vm_get_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			  struct list_head *duplicates);
 void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev,
 				  struct amdgpu_vm *vm);
 int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 		      struct amdgpu_sync *sync, struct fence *fence,
-		      unsigned *vm_id, uint64_t *vm_pd_addr);
-int amdgpu_vm_flush(struct amdgpu_ring *ring,
-		    unsigned vm_id, uint64_t pd_addr,
-		    uint32_t gds_base, uint32_t gds_size,
-		    uint32_t gws_base, uint32_t gws_size,
-		    uint32_t oa_base, uint32_t oa_size);
+		      struct amdgpu_job *job);
+int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job);
 void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vm_id);
 uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr);
 int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
@@ -1142,6 +1150,12 @@ struct amdgpu_cu_info {
 	uint32_t bitmap[4][4];
 };
 
+struct amdgpu_gfx_funcs {
+	/* get the gpu clock counter */
+	uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev);
+	void (*select_se_sh)(struct amdgpu_device *adev, u32 se_num, u32 sh_num, u32 instance);
+};
+
 struct amdgpu_gfx {
 	struct mutex			gpu_clock_mutex;
 	struct amdgpu_gca_config	config;
@@ -1178,6 +1192,7 @@ struct amdgpu_gfx {
 	/* ce ram size*/
 	unsigned			ce_ram_size;
 	struct amdgpu_cu_info		cu_info;
+	const struct amdgpu_gfx_funcs	*funcs;
 };
 
 int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
@@ -1195,10 +1210,6 @@ void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count);
 void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib);
 void amdgpu_ring_commit(struct amdgpu_ring *ring);
 void amdgpu_ring_undo(struct amdgpu_ring *ring);
-unsigned amdgpu_ring_backup(struct amdgpu_ring *ring,
-			    uint32_t **data);
-int amdgpu_ring_restore(struct amdgpu_ring *ring,
-			unsigned size, uint32_t *data);
 int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		     unsigned ring_size, u32 nop, u32 align_mask,
 		     struct amdgpu_irq_src *irq_src, unsigned irq_type,
@@ -1250,6 +1261,7 @@ struct amdgpu_job {
 	uint32_t		num_ibs;
 	void			*owner;
 	uint64_t		ctx;
+	bool                    vm_needs_flush;
 	unsigned		vm_id;
 	uint64_t		vm_pd_addr;
 	uint32_t		gds_base, gds_size;
@@ -1257,8 +1269,7 @@ struct amdgpu_job {
 	uint32_t		oa_base, oa_size;
 
 	/* user fence handling */
-	struct amdgpu_bo	*uf_bo;
-	uint32_t		uf_offset;
+	uint64_t		uf_addr;
 	uint64_t		uf_sequence;
 
 };
@@ -1560,6 +1571,12 @@ struct amdgpu_dpm_funcs {
 	u32 (*get_fan_control_mode)(struct amdgpu_device *adev);
 	int (*set_fan_speed_percent)(struct amdgpu_device *adev, u32 speed);
 	int (*get_fan_speed_percent)(struct amdgpu_device *adev, u32 *speed);
+	int (*force_clock_level)(struct amdgpu_device *adev, enum pp_clock_type type, uint32_t mask);
+	int (*print_clock_levels)(struct amdgpu_device *adev, enum pp_clock_type type, char *buf);
+	int (*get_sclk_od)(struct amdgpu_device *adev);
+	int (*set_sclk_od)(struct amdgpu_device *adev, uint32_t value);
+	int (*get_mclk_od)(struct amdgpu_device *adev);
+	int (*set_mclk_od)(struct amdgpu_device *adev, uint32_t value);
 };
 
 struct amdgpu_dpm {
@@ -1767,6 +1784,8 @@ int amdgpu_debugfs_init(struct drm_minor *minor);
 void amdgpu_debugfs_cleanup(struct drm_minor *minor);
 #endif
 
+int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev);
+
 /*
  * amdgpu smumgr functions
  */
@@ -1811,15 +1830,13 @@ struct amdgpu_asic_funcs {
 			     u32 sh_num, u32 reg_offset, u32 *value);
 	void (*set_vga_state)(struct amdgpu_device *adev, bool state);
 	int (*reset)(struct amdgpu_device *adev);
-	/* wait for mc_idle */
-	int (*wait_for_mc_idle)(struct amdgpu_device *adev);
 	/* get the reference clock */
 	u32 (*get_xclk)(struct amdgpu_device *adev);
-	/* get the gpu clock counter */
-	uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev);
 	/* MM block clocks */
 	int (*set_uvd_clocks)(struct amdgpu_device *adev, u32 vclk, u32 dclk);
 	int (*set_vce_clocks)(struct amdgpu_device *adev, u32 evclk, u32 ecclk);
+	/* query virtual capabilities */
+	u32 (*get_virtual_caps)(struct amdgpu_device *adev);
 };
 
 /*
@@ -1914,8 +1931,12 @@ void amdgpu_cgs_destroy_device(struct cgs_device *cgs_device);
 
 
 /* GPU virtualization */
+#define AMDGPU_VIRT_CAPS_SRIOV_EN       (1 << 0)
+#define AMDGPU_VIRT_CAPS_IS_VF          (1 << 1)
 struct amdgpu_virtualization {
 	bool supports_sr_iov;
+	bool is_virtual;
+	u32 caps;
 };
 
 /*
@@ -1997,6 +2018,10 @@ struct amdgpu_device {
 	spinlock_t didt_idx_lock;
 	amdgpu_rreg_t			didt_rreg;
 	amdgpu_wreg_t			didt_wreg;
+	/* protects concurrent gc_cac register access */
+	spinlock_t gc_cac_idx_lock;
+	amdgpu_rreg_t			gc_cac_rreg;
+	amdgpu_wreg_t			gc_cac_wreg;
 	/* protects concurrent ENDPOINT (audio) register access */
 	spinlock_t audio_endpt_idx_lock;
 	amdgpu_block_rreg_t		audio_endpt_rreg;
@@ -2022,6 +2047,7 @@ struct amdgpu_device {
 	atomic64_t			vram_vis_usage;
 	atomic64_t			gtt_usage;
 	atomic64_t			num_bytes_moved;
+	atomic64_t			num_evictions;
 	atomic_t			gpu_reset_counter;
 
 	/* display */
@@ -2125,6 +2151,8 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
 #define WREG32_UVD_CTX(reg, v) adev->uvd_ctx_wreg(adev, (reg), (v))
 #define RREG32_DIDT(reg) adev->didt_rreg(adev, (reg))
 #define WREG32_DIDT(reg, v) adev->didt_wreg(adev, (reg), (v))
+#define RREG32_GC_CAC(reg) adev->gc_cac_rreg(adev, (reg))
+#define WREG32_GC_CAC(reg, v) adev->gc_cac_wreg(adev, (reg), (v))
 #define RREG32_AUDIO_ENDPT(block, reg) adev->audio_endpt_rreg(adev, (block), (reg))
 #define WREG32_AUDIO_ENDPT(block, reg, v) adev->audio_endpt_wreg(adev, (block), (reg), (v))
 #define WREG32_P(reg, val, mask)				\
@@ -2200,11 +2228,10 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
  */
 #define amdgpu_asic_set_vga_state(adev, state) (adev)->asic_funcs->set_vga_state((adev), (state))
 #define amdgpu_asic_reset(adev) (adev)->asic_funcs->reset((adev))
-#define amdgpu_asic_wait_for_mc_idle(adev) (adev)->asic_funcs->wait_for_mc_idle((adev))
 #define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev))
 #define amdgpu_asic_set_uvd_clocks(adev, v, d) (adev)->asic_funcs->set_uvd_clocks((adev), (v), (d))
 #define amdgpu_asic_set_vce_clocks(adev, ev, ec) (adev)->asic_funcs->set_vce_clocks((adev), (ev), (ec))
-#define amdgpu_asic_get_gpu_clock_counter(adev) (adev)->asic_funcs->get_gpu_clock_counter((adev))
+#define amdgpu_asic_get_virtual_caps(adev) ((adev)->asic_funcs->get_virtual_caps((adev)))
 #define amdgpu_asic_read_disabled_bios(adev) (adev)->asic_funcs->read_disabled_bios((adev))
 #define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l))
 #define amdgpu_asic_read_register(adev, se, sh, offset, v)((adev)->asic_funcs->read_register((adev), (se), (sh), (offset), (v)))
@@ -2257,6 +2284,8 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_dpm_print_power_state(adev, ps) (adev)->pm.funcs->print_power_state((adev), (ps))
 #define amdgpu_dpm_vblank_too_short(adev) (adev)->pm.funcs->vblank_too_short((adev))
 #define amdgpu_dpm_enable_bapm(adev, e) (adev)->pm.funcs->enable_bapm((adev), (e))
+#define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev))
+#define amdgpu_gfx_select_se_sh(adev, se, sh, instance) (adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance))
 
 #define amdgpu_dpm_get_temperature(adev) \
 	((adev)->pp_enabled ?						\
@@ -2335,6 +2364,18 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_dpm_force_clock_level(adev, type, level) \
 		(adev)->powerplay.pp_funcs->force_clock_level((adev)->powerplay.pp_handle, type, level)
 
+#define amdgpu_dpm_get_sclk_od(adev) \
+	(adev)->powerplay.pp_funcs->get_sclk_od((adev)->powerplay.pp_handle)
+
+#define amdgpu_dpm_set_sclk_od(adev, value) \
+	(adev)->powerplay.pp_funcs->set_sclk_od((adev)->powerplay.pp_handle, value)
+
+#define amdgpu_dpm_get_mclk_od(adev) \
+	((adev)->powerplay.pp_funcs->get_mclk_od((adev)->powerplay.pp_handle))
+
+#define amdgpu_dpm_set_mclk_od(adev, value) \
+	((adev)->powerplay.pp_funcs->set_mclk_od((adev)->powerplay.pp_handle, value))
+
 #define amdgpu_dpm_dispatch_task(adev, event_id, input, output)		\
 	(adev)->powerplay.pp_funcs->dispatch_tasks((adev)->powerplay.pp_handle, (event_id), (input), (output))
 
@@ -2376,9 +2417,13 @@ bool amdgpu_device_is_px(struct drm_device *dev);
 #if defined(CONFIG_VGA_SWITCHEROO)
 void amdgpu_register_atpx_handler(void);
 void amdgpu_unregister_atpx_handler(void);
+bool amdgpu_has_atpx_dgpu_power_cntl(void);
+bool amdgpu_is_atpx_hybrid(void);
 #else
 static inline void amdgpu_register_atpx_handler(void) {}
 static inline void amdgpu_unregister_atpx_handler(void) {}
+static inline bool amdgpu_has_atpx_dgpu_power_cntl(void) { return false; }
+static inline bool amdgpu_is_atpx_hybrid(void) { return false; }
 #endif
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 32809f749903..d080d0807a5b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -240,8 +240,8 @@ uint64_t get_gpu_clock_counter(struct kgd_dev *kgd)
 {
 	struct amdgpu_device *rdev = (struct amdgpu_device *)kgd;
 
-	if (rdev->asic_funcs->get_gpu_clock_counter)
-		return rdev->asic_funcs->get_gpu_clock_counter(rdev);
+	if (rdev->gfx.funcs->get_gpu_clock_counter)
+		return rdev->gfx.funcs->get_gpu_clock_counter(rdev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
index 35a1248aaa77..0494fe7b62c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c
@@ -10,6 +10,7 @@
 #include <linux/slab.h>
 #include <linux/acpi.h>
 #include <linux/pci.h>
+#include <linux/delay.h>
 
 #include "amd_acpi.h"
 
@@ -27,6 +28,7 @@ struct amdgpu_atpx_functions {
 struct amdgpu_atpx {
 	acpi_handle handle;
 	struct amdgpu_atpx_functions functions;
+	bool is_hybrid;
 };
 
 static struct amdgpu_atpx_priv {
@@ -63,6 +65,14 @@ bool amdgpu_has_atpx(void) {
 	return amdgpu_atpx_priv.atpx_detected;
 }
 
+bool amdgpu_has_atpx_dgpu_power_cntl(void) {
+	return amdgpu_atpx_priv.atpx.functions.power_cntl;
+}
+
+bool amdgpu_is_atpx_hybrid(void) {
+	return amdgpu_atpx_priv.atpx.is_hybrid;
+}
+
 /**
  * amdgpu_atpx_call - call an ATPX method
  *
@@ -142,18 +152,12 @@ static void amdgpu_atpx_parse_functions(struct amdgpu_atpx_functions *f, u32 mas
  */
 static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx)
 {
-	/* make sure required functions are enabled */
-	/* dGPU power control is required */
-	if (atpx->functions.power_cntl == false) {
-		printk("ATPX dGPU power cntl not present, forcing\n");
-		atpx->functions.power_cntl = true;
-	}
+	u32 valid_bits = 0;
 
 	if (atpx->functions.px_params) {
 		union acpi_object *info;
 		struct atpx_px_params output;
 		size_t size;
-		u32 valid_bits;
 
 		info = amdgpu_atpx_call(atpx->handle, ATPX_FUNCTION_GET_PX_PARAMETERS, NULL);
 		if (!info)
@@ -172,19 +176,43 @@ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx)
 		memcpy(&output, info->buffer.pointer, size);
 
 		valid_bits = output.flags & output.valid_flags;
-		/* if separate mux flag is set, mux controls are required */
-		if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) {
-			atpx->functions.i2c_mux_cntl = true;
-			atpx->functions.disp_mux_cntl = true;
-		}
-		/* if any outputs are muxed, mux controls are required */
-		if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED |
-				  ATPX_TV_SIGNAL_MUXED |
-				  ATPX_DFP_SIGNAL_MUXED))
-			atpx->functions.disp_mux_cntl = true;
 
 		kfree(info);
 	}
+
+	/* if separate mux flag is set, mux controls are required */
+	if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) {
+		atpx->functions.i2c_mux_cntl = true;
+		atpx->functions.disp_mux_cntl = true;
+	}
+	/* if any outputs are muxed, mux controls are required */
+	if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED |
+			  ATPX_TV_SIGNAL_MUXED |
+			  ATPX_DFP_SIGNAL_MUXED))
+		atpx->functions.disp_mux_cntl = true;
+
+
+	/* some bioses set these bits rather than flagging power_cntl as supported */
+	if (valid_bits & (ATPX_DYNAMIC_PX_SUPPORTED |
+			  ATPX_DYNAMIC_DGPU_POWER_OFF_SUPPORTED))
+		atpx->functions.power_cntl = true;
+
+	atpx->is_hybrid = false;
+	if (valid_bits & ATPX_MS_HYBRID_GFX_SUPPORTED) {
+		printk("ATPX Hybrid Graphics\n");
+#if 1
+		/* This is a temporary hack until the D3 cold support
+		 * makes it upstream.  The ATPX power_control method seems
+		 * to still work on even if the system should be using
+		 * the new standardized hybrid D3 cold ACPI interface.
+		 */
+		atpx->functions.power_cntl = true;
+#else
+		atpx->functions.power_cntl = false;
+#endif
+		atpx->is_hybrid = true;
+	}
+
 	return 0;
 }
 
@@ -259,6 +287,10 @@ static int amdgpu_atpx_set_discrete_state(struct amdgpu_atpx *atpx, u8 state)
 		if (!info)
 			return -EIO;
 		kfree(info);
+
+		/* 200ms delay is required after off */
+		if (state == 0)
+			msleep(200);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
index 823bf5e0b0c8..651115dcce12 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
@@ -94,6 +94,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev,
 	unsigned last_entry = 0, first_userptr = num_entries;
 	unsigned i;
 	int r;
+	unsigned long total_size = 0;
 
 	array = drm_malloc_ab(num_entries, sizeof(struct amdgpu_bo_list_entry));
 	if (!array)
@@ -140,6 +141,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev,
 		if (entry->robj->prefered_domains == AMDGPU_GEM_DOMAIN_OA)
 			oa_obj = entry->robj;
 
+		total_size += amdgpu_bo_size(entry->robj);
 		trace_amdgpu_bo_list_set(list, entry->robj);
 	}
 
@@ -155,6 +157,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev,
 	list->array = array;
 	list->num_entries = num_entries;
 
+	trace_amdgpu_cs_bo_status(list->num_entries, total_size);
 	return 0;
 
 error_free:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
index 199f76baf22c..5556ce979199 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
@@ -312,6 +312,8 @@ static uint32_t amdgpu_cgs_read_ind_register(struct cgs_device *cgs_device,
 		return RREG32_UVD_CTX(index);
 	case CGS_IND_REG__DIDT:
 		return RREG32_DIDT(index);
+	case CGS_IND_REG_GC_CAC:
+		return RREG32_GC_CAC(index);
 	case CGS_IND_REG__AUDIO_ENDPT:
 		DRM_ERROR("audio endpt register access not implemented.\n");
 		return 0;
@@ -336,6 +338,8 @@ static void amdgpu_cgs_write_ind_register(struct cgs_device *cgs_device,
 		return WREG32_UVD_CTX(index, value);
 	case CGS_IND_REG__DIDT:
 		return WREG32_DIDT(index, value);
+	case CGS_IND_REG_GC_CAC:
+		return WREG32_GC_CAC(index, value);
 	case CGS_IND_REG__AUDIO_ENDPT:
 		DRM_ERROR("audio endpt register access not implemented.\n");
 		return;
@@ -696,6 +700,17 @@ static uint32_t fw_type_convert(struct cgs_device *cgs_device, uint32_t fw_type)
 	return result;
 }
 
+static int amdgpu_cgs_rel_firmware(struct cgs_device *cgs_device, enum cgs_ucode_id type)
+{
+	CGS_FUNC_ADEV;
+	if ((CGS_UCODE_ID_SMU == type) || (CGS_UCODE_ID_SMU_SK == type)) {
+		release_firmware(adev->pm.fw);
+		return 0;
+	}
+	/* cannot release other firmware because they are not created by cgs */
+	return -EINVAL;
+}
+
 static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 					enum cgs_ucode_id type,
 					struct cgs_firmware_info *info)
@@ -776,6 +791,7 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 		}
 
 		hdr = (const struct smc_firmware_header_v1_0 *)	adev->pm.fw->data;
+		amdgpu_ucode_print_smc_hdr(&hdr->header);
 		adev->pm.fw_version = le32_to_cpu(hdr->header.ucode_version);
 		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
 		ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
@@ -819,6 +835,9 @@ static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device,
 	case CGS_SYSTEM_INFO_GFX_CU_INFO:
 		sys_info->value = adev->gfx.cu_info.number;
 		break;
+	case CGS_SYSTEM_INFO_GFX_SE_INFO:
+		sys_info->value = adev->gfx.config.max_shader_engines;
+		break;
 	default:
 		return -ENODEV;
 	}
@@ -898,7 +917,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device,
 	struct cgs_acpi_method_argument *argument = NULL;
 	uint32_t i, count;
 	acpi_status status;
-	int result;
+	int result = 0;
 	uint32_t func_no = 0xFFFFFFFF;
 
 	handle = ACPI_HANDLE(&adev->pdev->dev);
@@ -961,11 +980,11 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device,
 				params->integer.value = argument->value;
 				break;
 			case ACPI_TYPE_STRING:
-				params->string.length = argument->method_length;
+				params->string.length = argument->data_length;
 				params->string.pointer = argument->pointer;
 				break;
 			case ACPI_TYPE_BUFFER:
-				params->buffer.length = argument->method_length;
+				params->buffer.length = argument->data_length;
 				params->buffer.pointer = argument->pointer;
 				break;
 			default:
@@ -1068,17 +1087,14 @@ int amdgpu_cgs_call_acpi_method(struct cgs_device *cgs_device,
 	struct cgs_acpi_method_info info = {0};
 
 	acpi_input[0].type = CGS_ACPI_TYPE_INTEGER;
-	acpi_input[0].method_length = sizeof(uint32_t);
 	acpi_input[0].data_length = sizeof(uint32_t);
 	acpi_input[0].value = acpi_function;
 
 	acpi_input[1].type = CGS_ACPI_TYPE_BUFFER;
-	acpi_input[1].method_length = CGS_ACPI_MAX_BUFFER_SIZE;
 	acpi_input[1].data_length = input_size;
 	acpi_input[1].pointer = pinput;
 
 	acpi_output.type = CGS_ACPI_TYPE_BUFFER;
-	acpi_output.method_length = CGS_ACPI_MAX_BUFFER_SIZE;
 	acpi_output.data_length = output_size;
 	acpi_output.pointer = poutput;
 
@@ -1125,6 +1141,7 @@ static const struct cgs_ops amdgpu_cgs_ops = {
 	amdgpu_cgs_pm_query_clock_limits,
 	amdgpu_cgs_set_camera_voltages,
 	amdgpu_cgs_get_firmware_info,
+	amdgpu_cgs_rel_firmware,
 	amdgpu_cgs_set_powergating_state,
 	amdgpu_cgs_set_clockgating_state,
 	amdgpu_cgs_get_active_displays_info,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
index cb07da41152b..ff0b55a65ca3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
@@ -1690,7 +1690,6 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 						   DRM_MODE_SCALE_NONE);
 			/* no HPD on analog connectors */
 			amdgpu_connector->hpd.hpd = AMDGPU_HPD_NONE;
-			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 			connector->interlace_allowed = true;
 			connector->doublescan_allowed = true;
 			break;
@@ -1893,8 +1892,10 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 	}
 
 	if (amdgpu_connector->hpd.hpd == AMDGPU_HPD_NONE) {
-		if (i2c_bus->valid)
-			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+		if (i2c_bus->valid) {
+			connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+			                    DRM_CONNECTOR_POLL_DISCONNECT;
+		}
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 9bc8f1d99733..0307ff5887c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -216,11 +216,8 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	if (ret)
 		goto free_all_kdata;
 
-	if (p->uf_entry.robj) {
-		p->job->uf_bo = amdgpu_bo_ref(p->uf_entry.robj);
-		p->job->uf_offset = uf_offset;
-	}
-
+	if (p->uf_entry.robj)
+		p->job->uf_addr = uf_offset;
 	kfree(chunk_array);
 	return 0;
 
@@ -459,7 +456,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		list_splice(&need_pages, &p->validated);
 	}
 
-	amdgpu_vm_get_pt_bos(&fpriv->vm, &duplicates);
+	amdgpu_vm_get_pt_bos(p->adev, &fpriv->vm, &duplicates);
 
 	p->bytes_moved_threshold = amdgpu_cs_get_threshold_for_moves(p->adev);
 	p->bytes_moved = 0;
@@ -472,6 +469,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	if (r)
 		goto error_validate;
 
+	fpriv->vm.last_eviction_counter =
+		atomic64_read(&p->adev->num_evictions);
+
 	if (p->bo_list) {
 		struct amdgpu_bo *gds = p->bo_list->gds_obj;
 		struct amdgpu_bo *gws = p->bo_list->gws_obj;
@@ -499,6 +499,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		}
 	}
 
+	if (p->uf_entry.robj)
+		p->job->uf_addr += amdgpu_bo_gpu_offset(p->uf_entry.robj);
+
 error_validate:
 	if (r) {
 		amdgpu_vm_move_pt_bos_in_lru(p->adev, &fpriv->vm);
@@ -653,18 +656,21 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev,
 
 	/* Only for UVD/VCE VM emulation */
 	if (ring->funcs->parse_cs) {
+		p->job->vm = NULL;
 		for (i = 0; i < p->job->num_ibs; i++) {
 			r = amdgpu_ring_parse_cs(ring, p, i);
 			if (r)
 				return r;
 		}
-	}
+	} else {
+		p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->page_directory);
 
-	r = amdgpu_bo_vm_update_pte(p, vm);
-	if (!r)
-		amdgpu_cs_sync_rings(p);
+		r = amdgpu_bo_vm_update_pte(p, vm);
+		if (r)
+			return r;
+	}
 
-	return r;
+	return amdgpu_cs_sync_rings(p);
 }
 
 static int amdgpu_cs_handle_lockup(struct amdgpu_device *adev, int r)
@@ -761,7 +767,7 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 	}
 
 	/* UVD & VCE fw doesn't support user fences */
-	if (parser->job->uf_bo && (
+	if (parser->job->uf_addr && (
 	    parser->job->ring->type == AMDGPU_RING_TYPE_UVD ||
 	    parser->job->ring->type == AMDGPU_RING_TYPE_VCE))
 		return -EINVAL;
@@ -830,17 +836,13 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 {
 	struct amdgpu_ring *ring = p->job->ring;
 	struct amd_sched_entity *entity = &p->ctx->rings[ring->idx].entity;
-	struct fence *fence;
 	struct amdgpu_job *job;
 	int r;
 
 	job = p->job;
 	p->job = NULL;
 
-	r = amd_sched_job_init(&job->base, &ring->sched,
-			       entity, amdgpu_job_timeout_func,
-			       amdgpu_job_free_func,
-			       p->filp, &fence);
+	r = amd_sched_job_init(&job->base, &ring->sched, entity, p->filp);
 	if (r) {
 		amdgpu_job_free(job);
 		return r;
@@ -848,9 +850,10 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 
 	job->owner = p->filp;
 	job->ctx = entity->fence_context;
-	p->fence = fence_get(fence);
-	cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, fence);
+	p->fence = fence_get(&job->base.s_fence->finished);
+	cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, p->fence);
 	job->uf_sequence = cs->out.handle;
+	amdgpu_job_free_resources(job);
 
 	trace_amdgpu_cs_ioctl(job);
 	amd_sched_entity_push_job(&job->base);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 2ab5e0b184fb..614fb026436d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -25,6 +25,7 @@
  *          Alex Deucher
  *          Jerome Glisse
  */
+#include <linux/kthread.h>
 #include <linux/console.h>
 #include <linux/slab.h>
 #include <linux/debugfs.h>
@@ -35,6 +36,7 @@
 #include <linux/vga_switcheroo.h>
 #include <linux/efi.h>
 #include "amdgpu.h"
+#include "amdgpu_trace.h"
 #include "amdgpu_i2c.h"
 #include "atom.h"
 #include "amdgpu_atombios.h"
@@ -79,24 +81,27 @@ bool amdgpu_device_is_px(struct drm_device *dev)
 uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg,
 			bool always_indirect)
 {
+	uint32_t ret;
+
 	if ((reg * 4) < adev->rmmio_size && !always_indirect)
-		return readl(((void __iomem *)adev->rmmio) + (reg * 4));
+		ret = readl(((void __iomem *)adev->rmmio) + (reg * 4));
 	else {
 		unsigned long flags;
-		uint32_t ret;
 
 		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
 		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
 		ret = readl(((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
 		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
-
-		return ret;
 	}
+	trace_amdgpu_mm_rreg(adev->pdev->device, reg, ret);
+	return ret;
 }
 
 void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 		    bool always_indirect)
 {
+	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
+	
 	if ((reg * 4) < adev->rmmio_size && !always_indirect)
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	else {
@@ -827,8 +832,10 @@ static uint32_t cail_ioreg_read(struct card_info *info, uint32_t reg)
  */
 static void amdgpu_atombios_fini(struct amdgpu_device *adev)
 {
-	if (adev->mode_info.atom_context)
+	if (adev->mode_info.atom_context) {
 		kfree(adev->mode_info.atom_context->scratch);
+		kfree(adev->mode_info.atom_context->iio);
+	}
 	kfree(adev->mode_info.atom_context);
 	adev->mode_info.atom_context = NULL;
 	kfree(adev->mode_info.atom_card_info);
@@ -1068,11 +1075,14 @@ int amdgpu_set_clockgating_state(struct amdgpu_device *adev,
 	int i, r = 0;
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
 		if (adev->ip_blocks[i].type == block_type) {
 			r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
 									    state);
 			if (r)
 				return r;
+			break;
 		}
 	}
 	return r;
@@ -1085,16 +1095,53 @@ int amdgpu_set_powergating_state(struct amdgpu_device *adev,
 	int i, r = 0;
 
 	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
 		if (adev->ip_blocks[i].type == block_type) {
 			r = adev->ip_blocks[i].funcs->set_powergating_state((void *)adev,
 									    state);
 			if (r)
 				return r;
+			break;
 		}
 	}
 	return r;
 }
 
+int amdgpu_wait_for_idle(struct amdgpu_device *adev,
+			 enum amd_ip_block_type block_type)
+{
+	int i, r;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_blocks[i].type == block_type) {
+			r = adev->ip_blocks[i].funcs->wait_for_idle((void *)adev);
+			if (r)
+				return r;
+			break;
+		}
+	}
+	return 0;
+
+}
+
+bool amdgpu_is_idle(struct amdgpu_device *adev,
+		    enum amd_ip_block_type block_type)
+{
+	int i;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_blocks[i].type == block_type)
+			return adev->ip_blocks[i].funcs->is_idle((void *)adev);
+	}
+	return true;
+
+}
+
 const struct amdgpu_ip_block_version * amdgpu_get_ip_block(
 					struct amdgpu_device *adev,
 					enum amd_ip_block_type type)
@@ -1207,6 +1254,9 @@ static int amdgpu_early_init(struct amdgpu_device *adev)
 		}
 	}
 
+	adev->cg_flags &= amdgpu_cg_mask;
+	adev->pg_flags &= amdgpu_pg_mask;
+
 	return 0;
 }
 
@@ -1325,6 +1375,11 @@ static int amdgpu_fini(struct amdgpu_device *adev)
 		adev->ip_block_status[i].valid = false;
 	}
 
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (adev->ip_blocks[i].funcs->late_fini)
+			adev->ip_blocks[i].funcs->late_fini((void *)adev);
+	}
+
 	return 0;
 }
 
@@ -1378,6 +1433,15 @@ static int amdgpu_resume(struct amdgpu_device *adev)
 	return 0;
 }
 
+static bool amdgpu_device_is_virtual(void)
+{
+#ifdef CONFIG_X86
+	return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#else
+	return false;
+#endif
+}
+
 /**
  * amdgpu_device_init - initialize the driver
  *
@@ -1424,9 +1488,12 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->uvd_ctx_wreg = &amdgpu_invalid_wreg;
 	adev->didt_rreg = &amdgpu_invalid_rreg;
 	adev->didt_wreg = &amdgpu_invalid_wreg;
+	adev->gc_cac_rreg = &amdgpu_invalid_rreg;
+	adev->gc_cac_wreg = &amdgpu_invalid_wreg;
 	adev->audio_endpt_rreg = &amdgpu_block_invalid_rreg;
 	adev->audio_endpt_wreg = &amdgpu_block_invalid_wreg;
 
+
 	DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X 0x%02X).\n",
 		 amdgpu_asic_name[adev->asic_type], pdev->vendor, pdev->device,
 		 pdev->subsystem_vendor, pdev->subsystem_device, pdev->revision);
@@ -1451,6 +1518,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	spin_lock_init(&adev->pcie_idx_lock);
 	spin_lock_init(&adev->uvd_ctx_idx_lock);
 	spin_lock_init(&adev->didt_idx_lock);
+	spin_lock_init(&adev->gc_cac_idx_lock);
 	spin_lock_init(&adev->audio_endpt_idx_lock);
 
 	adev->rmmio_base = pci_resource_start(adev->pdev, 5);
@@ -1495,29 +1563,38 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain);
 
 	/* Read BIOS */
-	if (!amdgpu_get_bios(adev))
-		return -EINVAL;
+	if (!amdgpu_get_bios(adev)) {
+		r = -EINVAL;
+		goto failed;
+	}
 	/* Must be an ATOMBIOS */
 	if (!adev->is_atom_bios) {
 		dev_err(adev->dev, "Expecting atombios for GPU\n");
-		return -EINVAL;
+		r = -EINVAL;
+		goto failed;
 	}
 	r = amdgpu_atombios_init(adev);
 	if (r) {
 		dev_err(adev->dev, "amdgpu_atombios_init failed\n");
-		return r;
+		goto failed;
 	}
 
 	/* See if the asic supports SR-IOV */
 	adev->virtualization.supports_sr_iov =
 		amdgpu_atombios_has_gpu_virtualization_table(adev);
 
+	/* Check if we are executing in a virtualized environment */
+	adev->virtualization.is_virtual = amdgpu_device_is_virtual();
+	adev->virtualization.caps = amdgpu_asic_get_virtual_caps(adev);
+
 	/* Post card if necessary */
 	if (!amdgpu_card_posted(adev) ||
-	    adev->virtualization.supports_sr_iov) {
+	    (adev->virtualization.is_virtual &&
+	     !(adev->virtualization.caps & AMDGPU_VIRT_CAPS_SRIOV_EN))) {
 		if (!adev->bios) {
 			dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n");
-			return -EINVAL;
+			r = -EINVAL;
+			goto failed;
 		}
 		DRM_INFO("GPU not posted. posting now...\n");
 		amdgpu_atom_asic_init(adev->mode_info.atom_context);
@@ -1527,7 +1604,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	r = amdgpu_atombios_get_clock_info(adev);
 	if (r) {
 		dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n");
-		return r;
+		goto failed;
 	}
 	/* init i2c buses */
 	amdgpu_atombios_i2c_init(adev);
@@ -1536,7 +1613,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	r = amdgpu_fence_driver_init(adev);
 	if (r) {
 		dev_err(adev->dev, "amdgpu_fence_driver_init failed\n");
-		return r;
+		goto failed;
 	}
 
 	/* init the mode config */
@@ -1546,7 +1623,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (r) {
 		dev_err(adev->dev, "amdgpu_init failed\n");
 		amdgpu_fini(adev);
-		return r;
+		goto failed;
 	}
 
 	adev->accel_working = true;
@@ -1556,7 +1633,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	r = amdgpu_ib_pool_init(adev);
 	if (r) {
 		dev_err(adev->dev, "IB initialization failed (%d).\n", r);
-		return r;
+		goto failed;
 	}
 
 	r = amdgpu_ib_ring_tests(adev);
@@ -1573,6 +1650,12 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		DRM_ERROR("registering register debugfs failed (%d).\n", r);
 	}
 
+	r = amdgpu_debugfs_firmware_init(adev);
+	if (r) {
+		DRM_ERROR("registering firmware debugfs failed (%d).\n", r);
+		return r;
+	}
+
 	if ((amdgpu_testing & 1)) {
 		if (adev->accel_working)
 			amdgpu_test_moves(adev);
@@ -1598,10 +1681,15 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	r = amdgpu_late_init(adev);
 	if (r) {
 		dev_err(adev->dev, "amdgpu_late_init failed\n");
-		return r;
+		goto failed;
 	}
 
 	return 0;
+
+failed:
+	if (runtime)
+		vga_switcheroo_fini_domain_pm_ops(adev->dev);
+	return r;
 }
 
 static void amdgpu_debugfs_remove_files(struct amdgpu_device *adev);
@@ -1636,6 +1724,8 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
 	kfree(adev->bios);
 	adev->bios = NULL;
 	vga_switcheroo_unregister_client(adev->pdev);
+	if (adev->flags & AMD_IS_PX)
+		vga_switcheroo_fini_domain_pm_ops(adev->dev);
 	vga_client_register(adev->pdev, NULL, NULL, NULL);
 	if (adev->rio_mem)
 		pci_iounmap(adev->pdev, adev->rio_mem);
@@ -1841,11 +1931,6 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
  */
 int amdgpu_gpu_reset(struct amdgpu_device *adev)
 {
-	unsigned ring_sizes[AMDGPU_MAX_RINGS];
-	uint32_t *ring_data[AMDGPU_MAX_RINGS];
-
-	bool saved = false;
-
 	int i, r;
 	int resched;
 
@@ -1854,22 +1939,30 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev)
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
 
-	r = amdgpu_suspend(adev);
-
+	/* block scheduler */
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 		struct amdgpu_ring *ring = adev->rings[i];
+
 		if (!ring)
 			continue;
-
-		ring_sizes[i] = amdgpu_ring_backup(ring, &ring_data[i]);
-		if (ring_sizes[i]) {
-			saved = true;
-			dev_info(adev->dev, "Saved %d dwords of commands "
-				 "on ring %d.\n", ring_sizes[i], i);
-		}
+		kthread_park(ring->sched.thread);
+		amd_sched_hw_job_reset(&ring->sched);
 	}
+	/* after all hw jobs are reset, hw fence is meaningless, so force_completion */
+	amdgpu_fence_driver_force_completion(adev);
+
+	/* save scratch */
+	amdgpu_atombios_scratch_regs_save(adev);
+	r = amdgpu_suspend(adev);
 
 retry:
+	/* Disable fb access */
+	if (adev->mode_info.num_crtc) {
+		struct amdgpu_mode_mc_save save;
+		amdgpu_display_stop_mc_access(adev, &save);
+		amdgpu_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GMC);
+	}
+
 	r = amdgpu_asic_reset(adev);
 	/* post card */
 	amdgpu_atom_asic_init(adev->mode_info.atom_context);
@@ -1878,32 +1971,29 @@ retry:
 		dev_info(adev->dev, "GPU reset succeeded, trying to resume\n");
 		r = amdgpu_resume(adev);
 	}
-
+	/* restore scratch */
+	amdgpu_atombios_scratch_regs_restore(adev);
 	if (!r) {
+		r = amdgpu_ib_ring_tests(adev);
+		if (r) {
+			dev_err(adev->dev, "ib ring test failed (%d).\n", r);
+			r = amdgpu_suspend(adev);
+			goto retry;
+		}
+
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = adev->rings[i];
 			if (!ring)
 				continue;
-
-			amdgpu_ring_restore(ring, ring_sizes[i], ring_data[i]);
-			ring_sizes[i] = 0;
-			ring_data[i] = NULL;
-		}
-
-		r = amdgpu_ib_ring_tests(adev);
-		if (r) {
-			dev_err(adev->dev, "ib ring test failed (%d).\n", r);
-			if (saved) {
-				saved = false;
-				r = amdgpu_suspend(adev);
-				goto retry;
-			}
+			amd_sched_job_recovery(&ring->sched);
+			kthread_unpark(ring->sched.thread);
 		}
 	} else {
-		amdgpu_fence_driver_force_completion(adev);
+		dev_err(adev->dev, "asic resume failed (%d).\n", r);
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-			if (adev->rings[i])
-				kfree(ring_data[i]);
+			if (adev->rings[i]) {
+				kthread_unpark(adev->rings[i]->sched.thread);
+			}
 		}
 	}
 
@@ -1914,13 +2004,11 @@ retry:
 		/* bad news, how to tell it to userspace ? */
 		dev_info(adev->dev, "GPU reset failed\n");
 	}
+	amdgpu_irq_gpu_reset_resume_helper(adev);
 
 	return r;
 }
 
-#define AMDGPU_DEFAULT_PCIE_GEN_MASK 0x30007  /* gen: chipset 1/2, asic 1/2/3 */
-#define AMDGPU_DEFAULT_PCIE_MLW_MASK 0x2f0000 /* 1/2/4/8/16 lanes */
-
 void amdgpu_get_pcie_info(struct amdgpu_device *adev)
 {
 	u32 mask;
@@ -2074,20 +2162,43 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf,
 	struct amdgpu_device *adev = f->f_inode->i_private;
 	ssize_t result = 0;
 	int r;
+	bool use_bank;
+	unsigned instance_bank, sh_bank, se_bank;
 
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
+	if (*pos & (1ULL << 62)) {
+		se_bank = (*pos >> 24) & 0x3FF;
+		sh_bank = (*pos >> 34) & 0x3FF;
+		instance_bank = (*pos >> 44) & 0x3FF;
+		use_bank = 1;
+		*pos &= 0xFFFFFF;
+	} else {
+		use_bank = 0;
+	}
+
+	if (use_bank) {
+		if (sh_bank >= adev->gfx.config.max_sh_per_se ||
+		    se_bank >= adev->gfx.config.max_shader_engines)
+			return -EINVAL;
+		mutex_lock(&adev->grbm_idx_mutex);
+		amdgpu_gfx_select_se_sh(adev, se_bank,
+					sh_bank, instance_bank);
+	}
+
 	while (size) {
 		uint32_t value;
 
 		if (*pos > adev->rmmio_size)
-			return result;
+			goto end;
 
 		value = RREG32(*pos >> 2);
 		r = put_user(value, (uint32_t *)buf);
-		if (r)
-			return r;
+		if (r) {
+			result = r;
+			goto end;
+		}
 
 		result += 4;
 		buf += 4;
@@ -2095,6 +2206,12 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
+end:
+	if (use_bank) {
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		mutex_unlock(&adev->grbm_idx_mutex);
+	}
+
 	return result;
 }
 
@@ -2294,6 +2411,68 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 	return result;
 }
 
+static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
+					size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = f->f_inode->i_private;
+	ssize_t result = 0;
+	int r;
+	uint32_t *config, no_regs = 0;
+
+	if (size & 0x3 || *pos & 0x3)
+		return -EINVAL;
+
+	config = kmalloc(256 * sizeof(*config), GFP_KERNEL);
+	if (!config)
+		return -ENOMEM;
+
+	/* version, increment each time something is added */
+	config[no_regs++] = 0;
+	config[no_regs++] = adev->gfx.config.max_shader_engines;
+	config[no_regs++] = adev->gfx.config.max_tile_pipes;
+	config[no_regs++] = adev->gfx.config.max_cu_per_sh;
+	config[no_regs++] = adev->gfx.config.max_sh_per_se;
+	config[no_regs++] = adev->gfx.config.max_backends_per_se;
+	config[no_regs++] = adev->gfx.config.max_texture_channel_caches;
+	config[no_regs++] = adev->gfx.config.max_gprs;
+	config[no_regs++] = adev->gfx.config.max_gs_threads;
+	config[no_regs++] = adev->gfx.config.max_hw_contexts;
+	config[no_regs++] = adev->gfx.config.sc_prim_fifo_size_frontend;
+	config[no_regs++] = adev->gfx.config.sc_prim_fifo_size_backend;
+	config[no_regs++] = adev->gfx.config.sc_hiz_tile_fifo_size;
+	config[no_regs++] = adev->gfx.config.sc_earlyz_tile_fifo_size;
+	config[no_regs++] = adev->gfx.config.num_tile_pipes;
+	config[no_regs++] = adev->gfx.config.backend_enable_mask;
+	config[no_regs++] = adev->gfx.config.mem_max_burst_length_bytes;
+	config[no_regs++] = adev->gfx.config.mem_row_size_in_kb;
+	config[no_regs++] = adev->gfx.config.shader_engine_tile_size;
+	config[no_regs++] = adev->gfx.config.num_gpus;
+	config[no_regs++] = adev->gfx.config.multi_gpu_tile_size;
+	config[no_regs++] = adev->gfx.config.mc_arb_ramcfg;
+	config[no_regs++] = adev->gfx.config.gb_addr_config;
+	config[no_regs++] = adev->gfx.config.num_rbs;
+
+	while (size && (*pos < no_regs * 4)) {
+		uint32_t value;
+
+		value = config[*pos >> 2];
+		r = put_user(value, (uint32_t *)buf);
+		if (r) {
+			kfree(config);
+			return r;
+		}
+
+		result += 4;
+		buf += 4;
+		*pos += 4;
+		size -= 4;
+	}
+
+	kfree(config);
+	return result;
+}
+
+
 static const struct file_operations amdgpu_debugfs_regs_fops = {
 	.owner = THIS_MODULE,
 	.read = amdgpu_debugfs_regs_read,
@@ -2319,11 +2498,18 @@ static const struct file_operations amdgpu_debugfs_regs_smc_fops = {
 	.llseek = default_llseek
 };
 
+static const struct file_operations amdgpu_debugfs_gca_config_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_gca_config_read,
+	.llseek = default_llseek
+};
+
 static const struct file_operations *debugfs_regs[] = {
 	&amdgpu_debugfs_regs_fops,
 	&amdgpu_debugfs_regs_didt_fops,
 	&amdgpu_debugfs_regs_pcie_fops,
 	&amdgpu_debugfs_regs_smc_fops,
+	&amdgpu_debugfs_gca_config_fops,
 };
 
 static const char *debugfs_regs_names[] = {
@@ -2331,6 +2517,7 @@ static const char *debugfs_regs_names[] = {
 	"amdgpu_regs_didt",
 	"amdgpu_regs_pcie",
 	"amdgpu_regs_smc",
+	"amdgpu_gca_config",
 };
 
 static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index a6eecf6f9065..7dbe8d02c5a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -220,19 +220,17 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 
 	r = amdgpu_bo_pin_restricted(new_rbo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, &base);
 	if (unlikely(r != 0)) {
-		amdgpu_bo_unreserve(new_rbo);
 		r = -EINVAL;
 		DRM_ERROR("failed to pin new rbo buffer before flip\n");
-		goto cleanup;
+		goto unreserve;
 	}
 
 	r = reservation_object_get_fences_rcu(new_rbo->tbo.resv, &work->excl,
 					      &work->shared_count,
 					      &work->shared);
 	if (unlikely(r != 0)) {
-		amdgpu_bo_unreserve(new_rbo);
 		DRM_ERROR("failed to get fences for buffer\n");
-		goto cleanup;
+		goto unpin;
 	}
 
 	amdgpu_bo_get_tiling_flags(new_rbo, &tiling_flags);
@@ -275,9 +273,11 @@ pflip_cleanup:
 		DRM_ERROR("failed to reserve new rbo in error path\n");
 		goto cleanup;
 	}
+unpin:
 	if (unlikely(amdgpu_bo_unpin(new_rbo) != 0)) {
 		DRM_ERROR("failed to unpin new rbo in error path\n");
 	}
+unreserve:
 	amdgpu_bo_unreserve(new_rbo);
 
 cleanup:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index f888c015f76c..015f1f4aae53 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -82,8 +82,12 @@ int amdgpu_exp_hw_support = 0;
 int amdgpu_sched_jobs = 32;
 int amdgpu_sched_hw_submission = 2;
 int amdgpu_powerplay = -1;
+int amdgpu_powercontainment = 1;
 unsigned amdgpu_pcie_gen_cap = 0;
 unsigned amdgpu_pcie_lane_cap = 0;
+unsigned amdgpu_cg_mask = 0xffffffff;
+unsigned amdgpu_pg_mask = 0xffffffff;
+char *amdgpu_disable_cu = NULL;
 
 MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes");
 module_param_named(vramlimit, amdgpu_vram_limit, int, 0600);
@@ -160,6 +164,9 @@ module_param_named(sched_hw_submission, amdgpu_sched_hw_submission, int, 0444);
 #ifdef CONFIG_DRM_AMD_POWERPLAY
 MODULE_PARM_DESC(powerplay, "Powerplay component (1 = enable, 0 = disable, -1 = auto (default))");
 module_param_named(powerplay, amdgpu_powerplay, int, 0444);
+
+MODULE_PARM_DESC(powercontainment, "Power Containment (1 = enable (default), 0 = disable)");
+module_param_named(powercontainment, amdgpu_powercontainment, int, 0444);
 #endif
 
 MODULE_PARM_DESC(pcie_gen_cap, "PCIE Gen Caps (0: autodetect (default))");
@@ -168,6 +175,15 @@ module_param_named(pcie_gen_cap, amdgpu_pcie_gen_cap, uint, 0444);
 MODULE_PARM_DESC(pcie_lane_cap, "PCIE Lane Caps (0: autodetect (default))");
 module_param_named(pcie_lane_cap, amdgpu_pcie_lane_cap, uint, 0444);
 
+MODULE_PARM_DESC(cg_mask, "Clockgating flags mask (0 = disable clock gating)");
+module_param_named(cg_mask, amdgpu_cg_mask, uint, 0444);
+
+MODULE_PARM_DESC(pg_mask, "Powergating flags mask (0 = disable power gating)");
+module_param_named(pg_mask, amdgpu_pg_mask, uint, 0444);
+
+MODULE_PARM_DESC(disable_cu, "Disable CUs (se.sh.cu,...)");
+module_param_named(disable_cu, amdgpu_disable_cu, charp, 0444);
+
 static const struct pci_device_id pciidlist[] = {
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	/* Kaveri */
@@ -413,7 +429,10 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	pci_save_state(pdev);
 	pci_disable_device(pdev);
 	pci_ignore_hotplug(pdev);
-	pci_set_power_state(pdev, PCI_D3cold);
+	if (amdgpu_is_atpx_hybrid())
+		pci_set_power_state(pdev, PCI_D3cold);
+	else if (!amdgpu_has_atpx_dgpu_power_cntl())
+		pci_set_power_state(pdev, PCI_D3hot);
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
 
 	return 0;
@@ -430,7 +449,9 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
-	pci_set_power_state(pdev, PCI_D0);
+	if (amdgpu_is_atpx_hybrid() ||
+	    !amdgpu_has_atpx_dgpu_power_cntl())
+		pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 	ret = pci_enable_device(pdev);
 	if (ret)
@@ -515,7 +536,7 @@ static struct drm_driver kms_driver = {
 	.driver_features =
 	    DRIVER_USE_AGP |
 	    DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM |
-	    DRIVER_PRIME | DRIVER_RENDER,
+	    DRIVER_PRIME | DRIVER_RENDER | DRIVER_MODESET,
 	.dev_priv_size = 0,
 	.load = amdgpu_driver_load_kms,
 	.open = amdgpu_driver_open_kms,
@@ -590,7 +611,6 @@ static int __init amdgpu_init(void)
 	DRM_INFO("amdgpu kernel modesetting enabled.\n");
 	driver = &kms_driver;
 	pdriver = &amdgpu_kms_pci_driver;
-	driver->driver_features |= DRIVER_MODESET;
 	driver->num_ioctls = amdgpu_max_kms_ioctl;
 	amdgpu_register_atpx_handler();
 	/* let modprobe override vga console setting */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 8fab6486064f..88fbed2389c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -503,7 +503,7 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 	if (r)
 		goto error_print;
 
-	amdgpu_vm_get_pt_bos(bo_va->vm, &duplicates);
+	amdgpu_vm_get_pt_bos(adev, bo_va->vm, &duplicates);
 	list_for_each_entry(entry, &list, head) {
 		domain = amdgpu_mem_type_to_domain(entry->bo->mem.mem_type);
 		/* if anything is swapped out don't swap it in here,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 9f95da4f0536..a074edd95c70 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -70,3 +70,47 @@ void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg)
 		}
 	}
 }
+
+/**
+ * amdgpu_gfx_parse_disable_cu - Parse the disable_cu module parameter
+ *
+ * @mask: array in which the per-shader array disable masks will be stored
+ * @max_se: number of SEs
+ * @max_sh: number of SHs
+ *
+ * The bitmask of CUs to be disabled in the shader array determined by se and
+ * sh is stored in mask[se * max_sh + sh].
+ */
+void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_sh)
+{
+	unsigned se, sh, cu;
+	const char *p;
+
+	memset(mask, 0, sizeof(*mask) * max_se * max_sh);
+
+	if (!amdgpu_disable_cu || !*amdgpu_disable_cu)
+		return;
+
+	p = amdgpu_disable_cu;
+	for (;;) {
+		char *next;
+		int ret = sscanf(p, "%u.%u.%u", &se, &sh, &cu);
+		if (ret < 3) {
+			DRM_ERROR("amdgpu: could not parse disable_cu\n");
+			return;
+		}
+
+		if (se < max_se && sh < max_sh && cu < 16) {
+			DRM_INFO("amdgpu: disabling CU %u.%u.%u\n", se, sh, cu);
+			mask[se * max_sh + sh] |= 1u << cu;
+		} else {
+			DRM_ERROR("amdgpu: disable_cu %u.%u.%u is out of range\n",
+				  se, sh, cu);
+		}
+
+		next = strchr(p, ',');
+		if (!next)
+			break;
+		p = next + 1;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index dc06cbda7be6..51321e154c09 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -27,4 +27,6 @@
 int amdgpu_gfx_scratch_get(struct amdgpu_device *adev, uint32_t *reg);
 void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg);
 
+unsigned amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_sh);
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 34e35423b78e..46c3097c5224 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -160,10 +160,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		patch_offset = amdgpu_ring_init_cond_exec(ring);
 
 	if (vm) {
-		r = amdgpu_vm_flush(ring, job->vm_id, job->vm_pd_addr,
-				    job->gds_base, job->gds_size,
-				    job->gws_base, job->gws_size,
-				    job->oa_base, job->oa_size);
+		r = amdgpu_vm_flush(ring, job);
 		if (r) {
 			amdgpu_ring_undo(ring);
 			return r;
@@ -203,11 +200,8 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	}
 
 	/* wrap the last IB with fence */
-	if (job && job->uf_bo) {
-		uint64_t addr = amdgpu_bo_gpu_offset(job->uf_bo);
-
-		addr += job->uf_offset;
-		amdgpu_ring_emit_fence(ring, addr, job->uf_sequence,
+	if (job && job->uf_addr) {
+		amdgpu_ring_emit_fence(ring, job->uf_addr, job->uf_sequence,
 				       AMDGPU_FENCE_FLAG_64BIT);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index 835a3fa8d8df..278708f5a744 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -383,6 +383,18 @@ int amdgpu_irq_update(struct amdgpu_device *adev,
 	return r;
 }
 
+void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev)
+{
+	int i, j;
+	for (i = 0; i < AMDGPU_MAX_IRQ_SRC_ID; i++) {
+		struct amdgpu_irq_src *src = adev->irq.sources[i];
+		if (!src)
+			continue;
+		for (j = 0; j < src->num_types; j++)
+			amdgpu_irq_update(adev, src, j);
+	}
+}
+
 /**
  * amdgpu_irq_get - enable interrupt
  *
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
index e124b59f39c1..7ef09352e534 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
@@ -94,6 +94,7 @@ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
 		   unsigned type);
 bool amdgpu_irq_enabled(struct amdgpu_device *adev, struct amdgpu_irq_src *src,
 			unsigned type);
+void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev);
 
 int amdgpu_irq_add_domain(struct amdgpu_device *adev);
 void amdgpu_irq_remove_domain(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index f0dafa514fe4..aaee0c8f6731 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -28,21 +28,15 @@
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 
-static void amdgpu_job_free_handler(struct work_struct *ws)
+static void amdgpu_job_timedout(struct amd_sched_job *s_job)
 {
-	struct amdgpu_job *job = container_of(ws, struct amdgpu_job, base.work_free_job);
-	amd_sched_job_put(&job->base);
-}
+	struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base);
 
-void amdgpu_job_timeout_func(struct work_struct *work)
-{
-	struct amdgpu_job *job = container_of(work, struct amdgpu_job, base.work_tdr.work);
 	DRM_ERROR("ring %s timeout, last signaled seq=%u, last emitted seq=%u\n",
-				job->base.sched->name,
-				(uint32_t)atomic_read(&job->ring->fence_drv.last_seq),
-				job->ring->fence_drv.sync_seq);
-
-	amd_sched_job_put(&job->base);
+		  job->base.sched->name,
+		  atomic_read(&job->ring->fence_drv.last_seq),
+		  job->ring->fence_drv.sync_seq);
+	amdgpu_gpu_reset(job->adev);
 }
 
 int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
@@ -63,7 +57,6 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
 	(*job)->vm = vm;
 	(*job)->ibs = (void *)&(*job)[1];
 	(*job)->num_ibs = num_ibs;
-	INIT_WORK(&(*job)->base.work_free_job, amdgpu_job_free_handler);
 
 	amdgpu_sync_create(&(*job)->sync);
 
@@ -86,27 +79,33 @@ int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size,
 	return r;
 }
 
-void amdgpu_job_free(struct amdgpu_job *job)
+void amdgpu_job_free_resources(struct amdgpu_job *job)
 {
-	unsigned i;
 	struct fence *f;
+	unsigned i;
+
 	/* use sched fence if available */
-	f = (job->base.s_fence)? &job->base.s_fence->base : job->fence;
+	f = job->base.s_fence ? &job->base.s_fence->finished : job->fence;
 
 	for (i = 0; i < job->num_ibs; ++i)
-		amdgpu_sa_bo_free(job->adev, &job->ibs[i].sa_bo, f);
-	fence_put(job->fence);
+		amdgpu_ib_free(job->adev, &job->ibs[i], f);
+}
 
-	amdgpu_bo_unref(&job->uf_bo);
-	amdgpu_sync_free(&job->sync);
+void amdgpu_job_free_cb(struct amd_sched_job *s_job)
+{
+	struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base);
 
-	if (!job->base.use_sched)
-		kfree(job);
+	fence_put(job->fence);
+	amdgpu_sync_free(&job->sync);
+	kfree(job);
 }
 
-void amdgpu_job_free_func(struct kref *refcount)
+void amdgpu_job_free(struct amdgpu_job *job)
 {
-	struct amdgpu_job *job = container_of(refcount, struct amdgpu_job, base.refcount);
+	amdgpu_job_free_resources(job);
+
+	fence_put(job->fence);
+	amdgpu_sync_free(&job->sync);
 	kfree(job);
 }
 
@@ -114,22 +113,20 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
 		      struct amd_sched_entity *entity, void *owner,
 		      struct fence **f)
 {
-	struct fence *fence;
 	int r;
 	job->ring = ring;
 
 	if (!f)
 		return -EINVAL;
 
-	r = amd_sched_job_init(&job->base, &ring->sched,
-			       entity, amdgpu_job_timeout_func,
-			       amdgpu_job_free_func, owner, &fence);
+	r = amd_sched_job_init(&job->base, &ring->sched, entity, owner);
 	if (r)
 		return r;
 
 	job->owner = owner;
 	job->ctx = entity->fence_context;
-	*f = fence_get(fence);
+	*f = fence_get(&job->base.s_fence->finished);
+	amdgpu_job_free_resources(job);
 	amd_sched_entity_push_job(&job->base);
 
 	return 0;
@@ -147,8 +144,8 @@ static struct fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
 		int r;
 
 		r = amdgpu_vm_grab_id(vm, ring, &job->sync,
-				      &job->base.s_fence->base,
-				      &job->vm_id, &job->vm_pd_addr);
+				      &job->base.s_fence->finished,
+				      job);
 		if (r)
 			DRM_ERROR("Error getting VM ID (%d)\n", r);
 
@@ -170,11 +167,7 @@ static struct fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	}
 	job = to_amdgpu_job(sched_job);
 
-	r = amdgpu_sync_wait(&job->sync);
-	if (r) {
-		DRM_ERROR("failed to sync wait (%d)\n", r);
-		return NULL;
-	}
+	BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL));
 
 	trace_amdgpu_sched_run_job(job);
 	r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs,
@@ -185,14 +178,15 @@ static struct fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	}
 
 err:
+	/* if gpu reset, hw fence will be replaced here */
+	fence_put(job->fence);
 	job->fence = fence;
-	amdgpu_job_free(job);
 	return fence;
 }
 
 const struct amd_sched_backend_ops amdgpu_sched_ops = {
 	.dependency = amdgpu_job_dependency,
 	.run_job = amdgpu_job_run,
-	.begin_job = amd_sched_job_begin,
-	.finish_job = amd_sched_job_finish,
+	.timedout_job = amdgpu_job_timedout,
+	.free_job = amdgpu_job_free_cb
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 38a28d137f1d..a8efbb54423f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -148,6 +148,65 @@ out:
 	return r;
 }
 
+static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
+				struct drm_amdgpu_query_fw *query_fw,
+				struct amdgpu_device *adev)
+{
+	switch (query_fw->fw_type) {
+	case AMDGPU_INFO_FW_VCE:
+		fw_info->ver = adev->vce.fw_version;
+		fw_info->feature = adev->vce.fb_version;
+		break;
+	case AMDGPU_INFO_FW_UVD:
+		fw_info->ver = adev->uvd.fw_version;
+		fw_info->feature = 0;
+		break;
+	case AMDGPU_INFO_FW_GMC:
+		fw_info->ver = adev->mc.fw_version;
+		fw_info->feature = 0;
+		break;
+	case AMDGPU_INFO_FW_GFX_ME:
+		fw_info->ver = adev->gfx.me_fw_version;
+		fw_info->feature = adev->gfx.me_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_PFP:
+		fw_info->ver = adev->gfx.pfp_fw_version;
+		fw_info->feature = adev->gfx.pfp_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_CE:
+		fw_info->ver = adev->gfx.ce_fw_version;
+		fw_info->feature = adev->gfx.ce_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_RLC:
+		fw_info->ver = adev->gfx.rlc_fw_version;
+		fw_info->feature = adev->gfx.rlc_feature_version;
+		break;
+	case AMDGPU_INFO_FW_GFX_MEC:
+		if (query_fw->index == 0) {
+			fw_info->ver = adev->gfx.mec_fw_version;
+			fw_info->feature = adev->gfx.mec_feature_version;
+		} else if (query_fw->index == 1) {
+			fw_info->ver = adev->gfx.mec2_fw_version;
+			fw_info->feature = adev->gfx.mec2_feature_version;
+		} else
+			return -EINVAL;
+		break;
+	case AMDGPU_INFO_FW_SMC:
+		fw_info->ver = adev->pm.fw_version;
+		fw_info->feature = 0;
+		break;
+	case AMDGPU_INFO_FW_SDMA:
+		if (query_fw->index >= adev->sdma.num_instances)
+			return -EINVAL;
+		fw_info->ver = adev->sdma.instance[query_fw->index].fw_version;
+		fw_info->feature = adev->sdma.instance[query_fw->index].feature_version;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
 /*
  * Userspace get information ioctl
  */
@@ -294,67 +353,20 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		return copy_to_user(out, &count, min(size, 4u)) ? -EFAULT : 0;
 	}
 	case AMDGPU_INFO_TIMESTAMP:
-		ui64 = amdgpu_asic_get_gpu_clock_counter(adev);
+		ui64 = amdgpu_gfx_get_gpu_clock_counter(adev);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_FW_VERSION: {
 		struct drm_amdgpu_info_firmware fw_info;
+		int ret;
 
 		/* We only support one instance of each IP block right now. */
 		if (info->query_fw.ip_instance != 0)
 			return -EINVAL;
 
-		switch (info->query_fw.fw_type) {
-		case AMDGPU_INFO_FW_VCE:
-			fw_info.ver = adev->vce.fw_version;
-			fw_info.feature = adev->vce.fb_version;
-			break;
-		case AMDGPU_INFO_FW_UVD:
-			fw_info.ver = adev->uvd.fw_version;
-			fw_info.feature = 0;
-			break;
-		case AMDGPU_INFO_FW_GMC:
-			fw_info.ver = adev->mc.fw_version;
-			fw_info.feature = 0;
-			break;
-		case AMDGPU_INFO_FW_GFX_ME:
-			fw_info.ver = adev->gfx.me_fw_version;
-			fw_info.feature = adev->gfx.me_feature_version;
-			break;
-		case AMDGPU_INFO_FW_GFX_PFP:
-			fw_info.ver = adev->gfx.pfp_fw_version;
-			fw_info.feature = adev->gfx.pfp_feature_version;
-			break;
-		case AMDGPU_INFO_FW_GFX_CE:
-			fw_info.ver = adev->gfx.ce_fw_version;
-			fw_info.feature = adev->gfx.ce_feature_version;
-			break;
-		case AMDGPU_INFO_FW_GFX_RLC:
-			fw_info.ver = adev->gfx.rlc_fw_version;
-			fw_info.feature = adev->gfx.rlc_feature_version;
-			break;
-		case AMDGPU_INFO_FW_GFX_MEC:
-			if (info->query_fw.index == 0) {
-				fw_info.ver = adev->gfx.mec_fw_version;
-				fw_info.feature = adev->gfx.mec_feature_version;
-			} else if (info->query_fw.index == 1) {
-				fw_info.ver = adev->gfx.mec2_fw_version;
-				fw_info.feature = adev->gfx.mec2_feature_version;
-			} else
-				return -EINVAL;
-			break;
-		case AMDGPU_INFO_FW_SMC:
-			fw_info.ver = adev->pm.fw_version;
-			fw_info.feature = 0;
-			break;
-		case AMDGPU_INFO_FW_SDMA:
-			if (info->query_fw.index >= adev->sdma.num_instances)
-				return -EINVAL;
-			fw_info.ver = adev->sdma.instance[info->query_fw.index].fw_version;
-			fw_info.feature = adev->sdma.instance[info->query_fw.index].feature_version;
-			break;
-		default:
-			return -EINVAL;
-		}
+		ret = amdgpu_firmware_info(&fw_info, &info->query_fw, adev);
+		if (ret)
+			return ret;
+
 		return copy_to_user(out, &fw_info,
 				    min((size_t)size, sizeof(fw_info))) ? -EFAULT : 0;
 	}
@@ -453,7 +465,8 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 			dev_info.max_memory_clock = adev->pm.default_mclk * 10;
 		}
 		dev_info.enabled_rb_pipes_mask = adev->gfx.config.backend_enable_mask;
-		dev_info.num_rb_pipes = adev->gfx.config.num_rbs;
+		dev_info.num_rb_pipes = adev->gfx.config.max_backends_per_se *
+			adev->gfx.config.max_shader_engines;
 		dev_info.num_hw_gfx_contexts = adev->gfx.config.max_hw_contexts;
 		dev_info._pad = 0;
 		dev_info.ids_flags = 0;
@@ -761,3 +774,130 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_USERPTR, amdgpu_gem_userptr_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 };
 const int amdgpu_max_kms_ioctl = ARRAY_SIZE(amdgpu_ioctls_kms);
+
+/*
+ * Debugfs info
+ */
+#if defined(CONFIG_DEBUG_FS)
+
+static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct drm_amdgpu_info_firmware fw_info;
+	struct drm_amdgpu_query_fw query_fw;
+	int ret, i;
+
+	/* VCE */
+	query_fw.fw_type = AMDGPU_INFO_FW_VCE;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "VCE feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* UVD */
+	query_fw.fw_type = AMDGPU_INFO_FW_UVD;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "UVD feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* GMC */
+	query_fw.fw_type = AMDGPU_INFO_FW_GMC;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "MC feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* ME */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_ME;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "ME feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* PFP */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_PFP;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "PFP feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* CE */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_CE;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "CE feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* RLC */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLC;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "RLC feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* MEC */
+	query_fw.fw_type = AMDGPU_INFO_FW_GFX_MEC;
+	query_fw.index = 0;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "MEC feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* MEC2 */
+	if (adev->asic_type == CHIP_KAVERI ||
+	    (adev->asic_type > CHIP_TOPAZ && adev->asic_type != CHIP_STONEY)) {
+		query_fw.index = 1;
+		ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+		if (ret)
+			return ret;
+		seq_printf(m, "MEC2 feature version: %u, firmware version: 0x%08x\n",
+			   fw_info.feature, fw_info.ver);
+	}
+
+	/* SMC */
+	query_fw.fw_type = AMDGPU_INFO_FW_SMC;
+	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+	if (ret)
+		return ret;
+	seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x\n",
+		   fw_info.feature, fw_info.ver);
+
+	/* SDMA */
+	query_fw.fw_type = AMDGPU_INFO_FW_SDMA;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		query_fw.index = i;
+		ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
+		if (ret)
+			return ret;
+		seq_printf(m, "SDMA%d feature version: %u, firmware version: 0x%08x\n",
+			   i, fw_info.feature, fw_info.ver);
+	}
+
+	return 0;
+}
+
+static const struct drm_info_list amdgpu_firmware_info_list[] = {
+	{"amdgpu_firmware_info", amdgpu_debugfs_firmware_info, 0, NULL},
+};
+#endif
+
+int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	return amdgpu_debugfs_add_files(adev, amdgpu_firmware_info_list,
+					ARRAY_SIZE(amdgpu_firmware_info_list));
+#else
+	return 0;
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 7ecea83ce453..6f0873c75a25 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -589,6 +589,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 			   struct ttm_mem_reg *new_mem)
 {
 	struct amdgpu_bo *rbo;
+	struct ttm_mem_reg *old_mem = &bo->mem;
 
 	if (!amdgpu_ttm_bo_is_amdgpu_bo(bo))
 		return;
@@ -602,6 +603,8 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 
 	/* move_notify is called before move happens */
 	amdgpu_update_memory_usage(rbo->adev, &bo->mem, new_mem);
+
+	trace_amdgpu_ttm_bo_move(rbo, new_mem->mem_type, old_mem->mem_type);
 }
 
 int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
index 589b36e8c5cf..ff63b88b0ffa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
@@ -270,30 +270,28 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
 	enum amd_pm_state_type state = 0;
-	long idx;
+	unsigned long idx;
 	int ret;
 
 	if (strlen(buf) == 1)
 		adev->pp_force_state_enabled = false;
-	else {
-		ret = kstrtol(buf, 0, &idx);
+	else if (adev->pp_enabled) {
+		struct pp_states_info data;
 
-		if (ret) {
+		ret = kstrtoul(buf, 0, &idx);
+		if (ret || idx >= ARRAY_SIZE(data.states)) {
 			count = -EINVAL;
 			goto fail;
 		}
 
-		if (adev->pp_enabled) {
-			struct pp_states_info data;
-			amdgpu_dpm_get_pp_num_states(adev, &data);
-			state = data.states[idx];
-			/* only set user selected power states */
-			if (state != POWER_STATE_TYPE_INTERNAL_BOOT &&
-				state != POWER_STATE_TYPE_DEFAULT) {
-				amdgpu_dpm_dispatch_task(adev,
-						AMD_PP_EVENT_ENABLE_USER_STATE, &state, NULL);
-				adev->pp_force_state_enabled = true;
-			}
+		amdgpu_dpm_get_pp_num_states(adev, &data);
+		state = data.states[idx];
+		/* only set user selected power states */
+		if (state != POWER_STATE_TYPE_INTERNAL_BOOT &&
+		    state != POWER_STATE_TYPE_DEFAULT) {
+			amdgpu_dpm_dispatch_task(adev,
+					AMD_PP_EVENT_ENABLE_USER_STATE, &state, NULL);
+			adev->pp_force_state_enabled = true;
 		}
 	}
 fail:
@@ -349,6 +347,8 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev,
 
 	if (adev->pp_enabled)
 		size = amdgpu_dpm_print_clock_levels(adev, PP_SCLK, buf);
+	else if (adev->pm.funcs->print_clock_levels)
+		size = adev->pm.funcs->print_clock_levels(adev, PP_SCLK, buf);
 
 	return size;
 }
@@ -365,7 +365,9 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev,
 	uint32_t i, mask = 0;
 	char sub_str[2];
 
-	for (i = 0; i < strlen(buf) - 1; i++) {
+	for (i = 0; i < strlen(buf); i++) {
+		if (*(buf + i) == '\n')
+			continue;
 		sub_str[0] = *(buf + i);
 		sub_str[1] = '\0';
 		ret = kstrtol(sub_str, 0, &level);
@@ -379,6 +381,8 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev,
 
 	if (adev->pp_enabled)
 		amdgpu_dpm_force_clock_level(adev, PP_SCLK, mask);
+	else if (adev->pm.funcs->force_clock_level)
+		adev->pm.funcs->force_clock_level(adev, PP_SCLK, mask);
 fail:
 	return count;
 }
@@ -393,6 +397,8 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev,
 
 	if (adev->pp_enabled)
 		size = amdgpu_dpm_print_clock_levels(adev, PP_MCLK, buf);
+	else if (adev->pm.funcs->print_clock_levels)
+		size = adev->pm.funcs->print_clock_levels(adev, PP_MCLK, buf);
 
 	return size;
 }
@@ -409,7 +415,9 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev,
 	uint32_t i, mask = 0;
 	char sub_str[2];
 
-	for (i = 0; i < strlen(buf) - 1; i++) {
+	for (i = 0; i < strlen(buf); i++) {
+		if (*(buf + i) == '\n')
+			continue;
 		sub_str[0] = *(buf + i);
 		sub_str[1] = '\0';
 		ret = kstrtol(sub_str, 0, &level);
@@ -423,6 +431,8 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev,
 
 	if (adev->pp_enabled)
 		amdgpu_dpm_force_clock_level(adev, PP_MCLK, mask);
+	else if (adev->pm.funcs->force_clock_level)
+		adev->pm.funcs->force_clock_level(adev, PP_MCLK, mask);
 fail:
 	return count;
 }
@@ -437,6 +447,8 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev,
 
 	if (adev->pp_enabled)
 		size = amdgpu_dpm_print_clock_levels(adev, PP_PCIE, buf);
+	else if (adev->pm.funcs->print_clock_levels)
+		size = adev->pm.funcs->print_clock_levels(adev, PP_PCIE, buf);
 
 	return size;
 }
@@ -453,7 +465,9 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
 	uint32_t i, mask = 0;
 	char sub_str[2];
 
-	for (i = 0; i < strlen(buf) - 1; i++) {
+	for (i = 0; i < strlen(buf); i++) {
+		if (*(buf + i) == '\n')
+			continue;
 		sub_str[0] = *(buf + i);
 		sub_str[1] = '\0';
 		ret = kstrtol(sub_str, 0, &level);
@@ -467,6 +481,100 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev,
 
 	if (adev->pp_enabled)
 		amdgpu_dpm_force_clock_level(adev, PP_PCIE, mask);
+	else if (adev->pm.funcs->force_clock_level)
+		adev->pm.funcs->force_clock_level(adev, PP_PCIE, mask);
+fail:
+	return count;
+}
+
+static ssize_t amdgpu_get_pp_sclk_od(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+	uint32_t value = 0;
+
+	if (adev->pp_enabled)
+		value = amdgpu_dpm_get_sclk_od(adev);
+	else if (adev->pm.funcs->get_sclk_od)
+		value = adev->pm.funcs->get_sclk_od(adev);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", value);
+}
+
+static ssize_t amdgpu_set_pp_sclk_od(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf,
+		size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+	int ret;
+	long int value;
+
+	ret = kstrtol(buf, 0, &value);
+
+	if (ret) {
+		count = -EINVAL;
+		goto fail;
+	}
+
+	if (adev->pp_enabled) {
+		amdgpu_dpm_set_sclk_od(adev, (uint32_t)value);
+		amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_READJUST_POWER_STATE, NULL, NULL);
+	} else if (adev->pm.funcs->set_sclk_od) {
+		adev->pm.funcs->set_sclk_od(adev, (uint32_t)value);
+		adev->pm.dpm.current_ps = adev->pm.dpm.boot_ps;
+		amdgpu_pm_compute_clocks(adev);
+	}
+
+fail:
+	return count;
+}
+
+static ssize_t amdgpu_get_pp_mclk_od(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+	uint32_t value = 0;
+
+	if (adev->pp_enabled)
+		value = amdgpu_dpm_get_mclk_od(adev);
+	else if (adev->pm.funcs->get_mclk_od)
+		value = adev->pm.funcs->get_mclk_od(adev);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", value);
+}
+
+static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf,
+		size_t count)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+	int ret;
+	long int value;
+
+	ret = kstrtol(buf, 0, &value);
+
+	if (ret) {
+		count = -EINVAL;
+		goto fail;
+	}
+
+	if (adev->pp_enabled) {
+		amdgpu_dpm_set_mclk_od(adev, (uint32_t)value);
+		amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_READJUST_POWER_STATE, NULL, NULL);
+	} else if (adev->pm.funcs->set_mclk_od) {
+		adev->pm.funcs->set_mclk_od(adev, (uint32_t)value);
+		adev->pm.dpm.current_ps = adev->pm.dpm.boot_ps;
+		amdgpu_pm_compute_clocks(adev);
+	}
+
 fail:
 	return count;
 }
@@ -492,6 +600,12 @@ static DEVICE_ATTR(pp_dpm_mclk, S_IRUGO | S_IWUSR,
 static DEVICE_ATTR(pp_dpm_pcie, S_IRUGO | S_IWUSR,
 		amdgpu_get_pp_dpm_pcie,
 		amdgpu_set_pp_dpm_pcie);
+static DEVICE_ATTR(pp_sclk_od, S_IRUGO | S_IWUSR,
+		amdgpu_get_pp_sclk_od,
+		amdgpu_set_pp_sclk_od);
+static DEVICE_ATTR(pp_mclk_od, S_IRUGO | S_IWUSR,
+		amdgpu_get_pp_mclk_od,
+		amdgpu_set_pp_mclk_od);
 
 static ssize_t amdgpu_hwmon_show_temp(struct device *dev,
 				      struct device_attribute *attr,
@@ -1110,22 +1224,34 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev)
 			DRM_ERROR("failed to create device file pp_table\n");
 			return ret;
 		}
-		ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk);
-		if (ret) {
-			DRM_ERROR("failed to create device file pp_dpm_sclk\n");
-			return ret;
-		}
-		ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk);
-		if (ret) {
-			DRM_ERROR("failed to create device file pp_dpm_mclk\n");
-			return ret;
-		}
-		ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie);
-		if (ret) {
-			DRM_ERROR("failed to create device file pp_dpm_pcie\n");
-			return ret;
-		}
 	}
+
+	ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk);
+	if (ret) {
+		DRM_ERROR("failed to create device file pp_dpm_sclk\n");
+		return ret;
+	}
+	ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk);
+	if (ret) {
+		DRM_ERROR("failed to create device file pp_dpm_mclk\n");
+		return ret;
+	}
+	ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie);
+	if (ret) {
+		DRM_ERROR("failed to create device file pp_dpm_pcie\n");
+		return ret;
+	}
+	ret = device_create_file(adev->dev, &dev_attr_pp_sclk_od);
+	if (ret) {
+		DRM_ERROR("failed to create device file pp_sclk_od\n");
+		return ret;
+	}
+	ret = device_create_file(adev->dev, &dev_attr_pp_mclk_od);
+	if (ret) {
+		DRM_ERROR("failed to create device file pp_mclk_od\n");
+		return ret;
+	}
+
 	ret = amdgpu_debugfs_pm_init(adev);
 	if (ret) {
 		DRM_ERROR("Failed to register debugfs file for dpm!\n");
@@ -1148,10 +1274,12 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev)
 		device_remove_file(adev->dev, &dev_attr_pp_cur_state);
 		device_remove_file(adev->dev, &dev_attr_pp_force_state);
 		device_remove_file(adev->dev, &dev_attr_pp_table);
-		device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk);
-		device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk);
-		device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie);
 	}
+	device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk);
+	device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk);
+	device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie);
+	device_remove_file(adev->dev, &dev_attr_pp_sclk_od);
+	device_remove_file(adev->dev, &dev_attr_pp_mclk_od);
 }
 
 void amdgpu_pm_compute_clocks(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c
index 6bd961fb43dc..c5738a22b690 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c
@@ -52,6 +52,7 @@ static int amdgpu_powerplay_init(struct amdgpu_device *adev)
 		pp_init->chip_family = adev->family;
 		pp_init->chip_id = adev->asic_type;
 		pp_init->device = amdgpu_cgs_create_device(adev);
+		pp_init->powercontainment_enabled = amdgpu_powercontainment;
 
 		ret = amd_powerplay_init(pp_init, amd_pp);
 		kfree(pp_init);
@@ -183,13 +184,6 @@ static int amdgpu_pp_sw_fini(void *handle)
 	if (ret)
 		return ret;
 
-#ifdef CONFIG_DRM_AMD_POWERPLAY
-	if (adev->pp_enabled) {
-		amdgpu_pm_sysfs_fini(adev);
-		amd_powerplay_fini(adev->powerplay.pp_handle);
-	}
-#endif
-
 	return ret;
 }
 
@@ -223,6 +217,22 @@ static int amdgpu_pp_hw_fini(void *handle)
 	return ret;
 }
 
+static void amdgpu_pp_late_fini(void *handle)
+{
+#ifdef CONFIG_DRM_AMD_POWERPLAY
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (adev->pp_enabled) {
+		amdgpu_pm_sysfs_fini(adev);
+		amd_powerplay_fini(adev->powerplay.pp_handle);
+	}
+
+	if (adev->powerplay.ip_funcs->late_fini)
+		adev->powerplay.ip_funcs->late_fini(
+			  adev->powerplay.pp_handle);
+#endif
+}
+
 static int amdgpu_pp_suspend(void *handle)
 {
 	int ret = 0;
@@ -311,6 +321,7 @@ const struct amd_ip_funcs amdgpu_pp_ip_funcs = {
 	.sw_fini = amdgpu_pp_sw_fini,
 	.hw_init = amdgpu_pp_hw_init,
 	.hw_fini = amdgpu_pp_hw_fini,
+	.late_fini = amdgpu_pp_late_fini,
 	.suspend = amdgpu_pp_suspend,
 	.resume = amdgpu_pp_resume,
 	.is_idle = amdgpu_pp_is_idle,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index 3b02272db678..3b885e3e9b56 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -28,6 +28,7 @@
  */
 #include <linux/seq_file.h>
 #include <linux/slab.h>
+#include <linux/debugfs.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
@@ -48,6 +49,7 @@
  */
 static int amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
 				    struct amdgpu_ring *ring);
+static void amdgpu_debugfs_ring_fini(struct amdgpu_ring *ring);
 
 /**
  * amdgpu_ring_alloc - allocate space on the ring buffer
@@ -140,78 +142,6 @@ void amdgpu_ring_undo(struct amdgpu_ring *ring)
 }
 
 /**
- * amdgpu_ring_backup - Back up the content of a ring
- *
- * @ring: the ring we want to back up
- *
- * Saves all unprocessed commits from a ring, returns the number of dwords saved.
- */
-unsigned amdgpu_ring_backup(struct amdgpu_ring *ring,
-			    uint32_t **data)
-{
-	unsigned size, ptr, i;
-
-	*data = NULL;
-
-	if (ring->ring_obj == NULL)
-		return 0;
-
-	/* it doesn't make sense to save anything if all fences are signaled */
-	if (!amdgpu_fence_count_emitted(ring))
-		return 0;
-
-	ptr = le32_to_cpu(*ring->next_rptr_cpu_addr);
-
-	size = ring->wptr + (ring->ring_size / 4);
-	size -= ptr;
-	size &= ring->ptr_mask;
-	if (size == 0)
-		return 0;
-
-	/* and then save the content of the ring */
-	*data = kmalloc_array(size, sizeof(uint32_t), GFP_KERNEL);
-	if (!*data)
-		return 0;
-	for (i = 0; i < size; ++i) {
-		(*data)[i] = ring->ring[ptr++];
-		ptr &= ring->ptr_mask;
-	}
-
-	return size;
-}
-
-/**
- * amdgpu_ring_restore - append saved commands to the ring again
- *
- * @ring: ring to append commands to
- * @size: number of dwords we want to write
- * @data: saved commands
- *
- * Allocates space on the ring and restore the previously saved commands.
- */
-int amdgpu_ring_restore(struct amdgpu_ring *ring,
-			unsigned size, uint32_t *data)
-{
-	int i, r;
-
-	if (!size || !data)
-		return 0;
-
-	/* restore the saved ring content */
-	r = amdgpu_ring_alloc(ring, size);
-	if (r)
-		return r;
-
-	for (i = 0; i < size; ++i) {
-		amdgpu_ring_write(ring, data[i]);
-	}
-
-	amdgpu_ring_commit(ring);
-	kfree(data);
-	return 0;
-}
-
-/**
  * amdgpu_ring_init - init driver ring struct.
  *
  * @adev: amdgpu_device pointer
@@ -260,14 +190,6 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		return r;
 	}
 
-	r = amdgpu_wb_get(adev, &ring->next_rptr_offs);
-	if (r) {
-		dev_err(adev->dev, "(%d) ring next_rptr wb alloc failed\n", r);
-		return r;
-	}
-	ring->next_rptr_gpu_addr = adev->wb.gpu_addr + ring->next_rptr_offs * 4;
-	ring->next_rptr_cpu_addr = &adev->wb.wb[ring->next_rptr_offs];
-
 	r = amdgpu_wb_get(adev, &ring->cond_exe_offs);
 	if (r) {
 		dev_err(adev->dev, "(%d) ring cond_exec_polling wb alloc failed\n", r);
@@ -310,6 +232,9 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		}
 		r = amdgpu_bo_kmap(ring->ring_obj,
 				       (void **)&ring->ring);
+
+		memset((void *)ring->ring, 0, ring->ring_size);
+
 		amdgpu_bo_unreserve(ring->ring_obj);
 		if (r) {
 			dev_err(adev->dev, "(%d) ring map failed\n", r);
@@ -343,10 +268,10 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
 	ring->ring = NULL;
 	ring->ring_obj = NULL;
 
+	amdgpu_wb_free(ring->adev, ring->cond_exe_offs);
 	amdgpu_wb_free(ring->adev, ring->fence_offs);
 	amdgpu_wb_free(ring->adev, ring->rptr_offs);
 	amdgpu_wb_free(ring->adev, ring->wptr_offs);
-	amdgpu_wb_free(ring->adev, ring->next_rptr_offs);
 
 	if (ring_obj) {
 		r = amdgpu_bo_reserve(ring_obj, false);
@@ -357,6 +282,7 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
 		}
 		amdgpu_bo_unref(&ring_obj);
 	}
+	amdgpu_debugfs_ring_fini(ring);
 }
 
 /*
@@ -364,57 +290,62 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
  */
 #if defined(CONFIG_DEBUG_FS)
 
-static int amdgpu_debugfs_ring_info(struct seq_file *m, void *data)
+/* Layout of file is 12 bytes consisting of
+ * - rptr
+ * - wptr
+ * - driver's copy of wptr
+ *
+ * followed by n-words of ring data
+ */
+static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf,
+					size_t size, loff_t *pos)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = dev->dev_private;
-	int roffset = (unsigned long)node->info_ent->data;
-	struct amdgpu_ring *ring = (void *)(((uint8_t*)adev) + roffset);
-	uint32_t rptr, wptr, rptr_next;
-	unsigned i;
-
-	wptr = amdgpu_ring_get_wptr(ring);
-	seq_printf(m, "wptr: 0x%08x [%5d]\n", wptr, wptr);
-
-	rptr = amdgpu_ring_get_rptr(ring);
-	rptr_next = le32_to_cpu(*ring->next_rptr_cpu_addr);
-
-	seq_printf(m, "rptr: 0x%08x [%5d]\n", rptr, rptr);
-
-	seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n",
-		   ring->wptr, ring->wptr);
-
-	if (!ring->ready)
-		return 0;
-
-	/* print 8 dw before current rptr as often it's the last executed
-	 * packet that is the root issue
-	 */
-	i = (rptr + ring->ptr_mask + 1 - 32) & ring->ptr_mask;
-	while (i != rptr) {
-		seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]);
-		if (i == rptr)
-			seq_puts(m, " *");
-		if (i == rptr_next)
-			seq_puts(m, " #");
-		seq_puts(m, "\n");
-		i = (i + 1) & ring->ptr_mask;
+	struct amdgpu_ring *ring = (struct amdgpu_ring*)f->f_inode->i_private;
+	int r, i;
+	uint32_t value, result, early[3];
+
+	if (*pos & 3 || size & 3)
+		return -EINVAL;
+
+	result = 0;
+
+	if (*pos < 12) {
+		early[0] = amdgpu_ring_get_rptr(ring);
+		early[1] = amdgpu_ring_get_wptr(ring);
+		early[2] = ring->wptr;
+		for (i = *pos / 4; i < 3 && size; i++) {
+			r = put_user(early[i], (uint32_t *)buf);
+			if (r)
+				return r;
+			buf += 4;
+			result += 4;
+			size -= 4;
+			*pos += 4;
+		}
 	}
-	while (i != wptr) {
-		seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]);
-		if (i == rptr)
-			seq_puts(m, " *");
-		if (i == rptr_next)
-			seq_puts(m, " #");
-		seq_puts(m, "\n");
-		i = (i + 1) & ring->ptr_mask;
+
+	while (size) {
+		if (*pos >= (ring->ring_size + 12))
+			return result;
+			
+		value = ring->ring[(*pos - 12)/4];
+		r = put_user(value, (uint32_t*)buf);
+		if (r)
+			return r;
+		buf += 4;
+		result += 4;
+		size -= 4;
+		*pos += 4;
 	}
-	return 0;
+
+	return result;
 }
 
-static struct drm_info_list amdgpu_debugfs_ring_info_list[AMDGPU_MAX_RINGS];
-static char amdgpu_debugfs_ring_names[AMDGPU_MAX_RINGS][32];
+static const struct file_operations amdgpu_debugfs_ring_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_ring_read,
+	.llseek = default_llseek
+};
 
 #endif
 
@@ -422,28 +353,27 @@ static int amdgpu_debugfs_ring_init(struct amdgpu_device *adev,
 				    struct amdgpu_ring *ring)
 {
 #if defined(CONFIG_DEBUG_FS)
-	unsigned offset = (uint8_t*)ring - (uint8_t*)adev;
-	unsigned i;
-	struct drm_info_list *info;
-	char *name;
-
-	for (i = 0; i < ARRAY_SIZE(amdgpu_debugfs_ring_info_list); ++i) {
-		info = &amdgpu_debugfs_ring_info_list[i];
-		if (!info->data)
-			break;
-	}
+	struct drm_minor *minor = adev->ddev->primary;
+	struct dentry *ent, *root = minor->debugfs_root;
+	char name[32];
 
-	if (i == ARRAY_SIZE(amdgpu_debugfs_ring_info_list))
-		return -ENOSPC;
-
-	name = &amdgpu_debugfs_ring_names[i][0];
 	sprintf(name, "amdgpu_ring_%s", ring->name);
-	info->name = name;
-	info->show = amdgpu_debugfs_ring_info;
-	info->driver_features = 0;
-	info->data = (void*)(uintptr_t)offset;
 
-	return amdgpu_debugfs_add_files(adev, info, 1);
+	ent = debugfs_create_file(name,
+				  S_IFREG | S_IRUGO, root,
+				  ring, &amdgpu_debugfs_ring_fops);
+	if (IS_ERR(ent))
+		return PTR_ERR(ent);
+
+	i_size_write(ent->d_inode, ring->ring_size + 12);
+	ring->ent = ent;
 #endif
 	return 0;
 }
+
+static void amdgpu_debugfs_ring_fini(struct amdgpu_ring *ring)
+{
+#if defined(CONFIG_DEBUG_FS)
+	debugfs_remove(ring->ent);
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
index b16366c2b4a0..d8af37a845f4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
@@ -115,6 +115,7 @@ int amdgpu_sa_bo_manager_start(struct amdgpu_device *adev,
 		return r;
 	}
 	r = amdgpu_bo_kmap(sa_manager->bo, &sa_manager->cpu_ptr);
+	memset(sa_manager->cpu_ptr, 0, sa_manager->size);
 	amdgpu_bo_unreserve(sa_manager->bo);
 	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index 34a92808bbd4..5c8d3022fb87 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -223,13 +223,16 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
 }
 
 /**
- * amdgpu_sync_is_idle - test if all fences are signaled
+ * amdgpu_sync_peek_fence - get the next fence not signaled yet
  *
  * @sync: the sync object
+ * @ring: optional ring to use for test
  *
- * Returns true if all fences in the sync object are signaled.
+ * Returns the next fence not signaled yet without removing it from the sync
+ * object.
  */
-bool amdgpu_sync_is_idle(struct amdgpu_sync *sync)
+struct fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
+				     struct amdgpu_ring *ring)
 {
 	struct amdgpu_sync_entry *e;
 	struct hlist_node *tmp;
@@ -237,6 +240,19 @@ bool amdgpu_sync_is_idle(struct amdgpu_sync *sync)
 
 	hash_for_each_safe(sync->fences, i, tmp, e, node) {
 		struct fence *f = e->fence;
+		struct amd_sched_fence *s_fence = to_amd_sched_fence(f);
+
+		if (ring && s_fence) {
+			/* For fences from the same ring it is sufficient
+			 * when they are scheduled.
+			 */
+			if (s_fence->sched == &ring->sched) {
+				if (fence_is_signaled(&s_fence->scheduled))
+					continue;
+
+				return &s_fence->scheduled;
+			}
+		}
 
 		if (fence_is_signaled(f)) {
 			hash_del(&e->node);
@@ -245,58 +261,19 @@ bool amdgpu_sync_is_idle(struct amdgpu_sync *sync)
 			continue;
 		}
 
-		return false;
+		return f;
 	}
 
-	return true;
+	return NULL;
 }
 
 /**
- * amdgpu_sync_cycle_fences - move fences from one sync object into another
+ * amdgpu_sync_get_fence - get the next fence from the sync object
  *
- * @dst: the destination sync object
- * @src: the source sync object
- * @fence: fence to add to source
+ * @sync: sync object to use
  *
- * Remove all fences from source and put them into destination and add
- * fence as new one into source.
+ * Get and removes the next fence from the sync object not signaled yet.
  */
-int amdgpu_sync_cycle_fences(struct amdgpu_sync *dst, struct amdgpu_sync *src,
-			     struct fence *fence)
-{
-	struct amdgpu_sync_entry *e, *newone;
-	struct hlist_node *tmp;
-	int i;
-
-	/* Allocate the new entry before moving the old ones */
-	newone = kmem_cache_alloc(amdgpu_sync_slab, GFP_KERNEL);
-	if (!newone)
-		return -ENOMEM;
-
-	hash_for_each_safe(src->fences, i, tmp, e, node) {
-		struct fence *f = e->fence;
-
-		hash_del(&e->node);
-		if (fence_is_signaled(f)) {
-			fence_put(f);
-			kmem_cache_free(amdgpu_sync_slab, e);
-			continue;
-		}
-
-		if (amdgpu_sync_add_later(dst, f)) {
-			kmem_cache_free(amdgpu_sync_slab, e);
-			continue;
-		}
-
-		hash_add(dst->fences, &e->node, f->context);
-	}
-
-	hash_add(src->fences, &newone->node, fence->context);
-	newone->fence = fence_get(fence);
-
-	return 0;
-}
-
 struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync)
 {
 	struct amdgpu_sync_entry *e;
@@ -319,25 +296,6 @@ struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync)
 	return NULL;
 }
 
-int amdgpu_sync_wait(struct amdgpu_sync *sync)
-{
-	struct amdgpu_sync_entry *e;
-	struct hlist_node *tmp;
-	int i, r;
-
-	hash_for_each_safe(sync->fences, i, tmp, e, node) {
-		r = fence_wait(e->fence, false);
-		if (r)
-			return r;
-
-		hash_del(&e->node);
-		fence_put(e->fence);
-		kmem_cache_free(amdgpu_sync_slab, e);
-	}
-
-	return 0;
-}
-
 /**
  * amdgpu_sync_free - free the sync object
  *
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
index 26a5f4acf584..499803f3ce3a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
@@ -11,19 +11,68 @@
 #define TRACE_SYSTEM amdgpu
 #define TRACE_INCLUDE_FILE amdgpu_trace
 
+TRACE_EVENT(amdgpu_mm_rreg,
+	    TP_PROTO(unsigned did, uint32_t reg, uint32_t value),
+	    TP_ARGS(did, reg, value),
+	    TP_STRUCT__entry(
+				__field(unsigned, did)
+				__field(uint32_t, reg)
+				__field(uint32_t, value)
+			    ),
+	    TP_fast_assign(
+			   __entry->did = did;
+			   __entry->reg = reg;
+			   __entry->value = value;
+			   ),
+	    TP_printk("0x%04lx, 0x%04lx, 0x%08lx",
+		      (unsigned long)__entry->did,
+		      (unsigned long)__entry->reg,
+		      (unsigned long)__entry->value)
+);
+
+TRACE_EVENT(amdgpu_mm_wreg,
+	    TP_PROTO(unsigned did, uint32_t reg, uint32_t value),
+	    TP_ARGS(did, reg, value),
+	    TP_STRUCT__entry(
+				__field(unsigned, did)
+				__field(uint32_t, reg)
+				__field(uint32_t, value)
+			    ),
+	    TP_fast_assign(
+			   __entry->did = did;
+			   __entry->reg = reg;
+			   __entry->value = value;
+			   ),
+	    TP_printk("0x%04lx, 0x%04lx, 0x%08lx",
+		      (unsigned long)__entry->did,
+		      (unsigned long)__entry->reg,
+		      (unsigned long)__entry->value)
+);
+
 TRACE_EVENT(amdgpu_bo_create,
 	    TP_PROTO(struct amdgpu_bo *bo),
 	    TP_ARGS(bo),
 	    TP_STRUCT__entry(
 			     __field(struct amdgpu_bo *, bo)
 			     __field(u32, pages)
+			     __field(u32, type)
+			     __field(u32, prefer)
+			     __field(u32, allow)
+			     __field(u32, visible)
 			     ),
 
 	    TP_fast_assign(
 			   __entry->bo = bo;
 			   __entry->pages = bo->tbo.num_pages;
+			   __entry->type = bo->tbo.mem.mem_type;
+			   __entry->prefer = bo->prefered_domains;
+			   __entry->allow = bo->allowed_domains;
+			   __entry->visible = bo->flags;
 			   ),
-	    TP_printk("bo=%p, pages=%u", __entry->bo, __entry->pages)
+
+	    TP_printk("bo=%p,pages=%u,type=%d,prefered=%d,allowed=%d,visible=%d",
+		       __entry->bo, __entry->pages, __entry->type,
+		       __entry->prefer, __entry->allow, __entry->visible)
 );
 
 TRACE_EVENT(amdgpu_cs,
@@ -64,7 +113,7 @@ TRACE_EVENT(amdgpu_cs_ioctl,
 			   __entry->adev = job->adev;
 			   __entry->sched_job = &job->base;
 			   __entry->ib = job->ibs;
-			   __entry->fence = &job->base.s_fence->base;
+			   __entry->fence = &job->base.s_fence->finished;
 			   __entry->ring_name = job->ring->name;
 			   __entry->num_ibs = job->num_ibs;
 			   ),
@@ -89,7 +138,7 @@ TRACE_EVENT(amdgpu_sched_run_job,
 			   __entry->adev = job->adev;
 			   __entry->sched_job = &job->base;
 			   __entry->ib = job->ibs;
-			   __entry->fence = &job->base.s_fence->base;
+			   __entry->fence = &job->base.s_fence->finished;
 			   __entry->ring_name = job->ring->name;
 			   __entry->num_ibs = job->num_ibs;
 			   ),
@@ -244,13 +293,55 @@ TRACE_EVENT(amdgpu_bo_list_set,
 	    TP_STRUCT__entry(
 			     __field(struct amdgpu_bo_list *, list)
 			     __field(struct amdgpu_bo *, bo)
+			     __field(u64, bo_size)
 			     ),
 
 	    TP_fast_assign(
 			   __entry->list = list;
 			   __entry->bo = bo;
+			   __entry->bo_size = amdgpu_bo_size(bo);
 			   ),
-	    TP_printk("list=%p, bo=%p", __entry->list, __entry->bo)
+	    TP_printk("list=%p, bo=%p, bo_size = %Ld",
+		      __entry->list,
+		      __entry->bo,
+		      __entry->bo_size)
+);
+
+TRACE_EVENT(amdgpu_cs_bo_status,
+	    TP_PROTO(uint64_t total_bo, uint64_t total_size),
+	    TP_ARGS(total_bo, total_size),
+	    TP_STRUCT__entry(
+			__field(u64, total_bo)
+			__field(u64, total_size)
+			),
+
+	    TP_fast_assign(
+			__entry->total_bo = total_bo;
+			__entry->total_size = total_size;
+			),
+	    TP_printk("total bo size = %Ld, total bo count = %Ld",
+			__entry->total_bo, __entry->total_size)
+);
+
+TRACE_EVENT(amdgpu_ttm_bo_move,
+	    TP_PROTO(struct amdgpu_bo* bo, uint32_t new_placement, uint32_t old_placement),
+	    TP_ARGS(bo, new_placement, old_placement),
+	    TP_STRUCT__entry(
+			__field(struct amdgpu_bo *, bo)
+			__field(u64, bo_size)
+			__field(u32, new_placement)
+			__field(u32, old_placement)
+			),
+
+	    TP_fast_assign(
+			__entry->bo      = bo;
+			__entry->bo_size = amdgpu_bo_size(bo);
+			__entry->new_placement = new_placement;
+			__entry->old_placement = old_placement;
+			),
+	    TP_printk("bo=%p from:%d to %d with size = %Ld",
+			__entry->bo, __entry->old_placement,
+			__entry->new_placement, __entry->bo_size)
 );
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 3b9053af4762..b7742e62972a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -286,9 +286,10 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 	r = amdgpu_copy_buffer(ring, old_start, new_start,
 			       new_mem->num_pages * PAGE_SIZE, /* bytes */
 			       bo->resv, &fence);
-	/* FIXME: handle copy error */
-	r = ttm_bo_move_accel_cleanup(bo, fence,
-				      evict, no_wait_gpu, new_mem);
+	if (r)
+		return r;
+
+	r = ttm_bo_pipeline_move(bo, fence, evict, new_mem);
 	fence_put(fence);
 	return r;
 }
@@ -396,6 +397,11 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo,
 		return -EINVAL;
 
 	adev = amdgpu_get_adev(bo->bdev);
+
+	/* remember the eviction */
+	if (evict)
+		atomic64_inc(&adev->num_evictions);
+
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
 		amdgpu_move_null(bo, new_mem);
 		return 0;
@@ -429,7 +435,8 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo,
 
 	if (r) {
 memcpy:
-		r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+		r = ttm_bo_move_memcpy(bo, evict, interruptible,
+				       no_wait_gpu, new_mem);
 		if (r) {
 			return r;
 		}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 01abfc21b4a2..e19520c4b4b6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -253,19 +253,20 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 {
 	int r;
 
-	if (adev->uvd.vcpu_bo == NULL)
-		return 0;
+	kfree(adev->uvd.saved_bo);
 
 	amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity);
 
-	r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false);
-	if (!r) {
-		amdgpu_bo_kunmap(adev->uvd.vcpu_bo);
-		amdgpu_bo_unpin(adev->uvd.vcpu_bo);
-		amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
-	}
+	if (adev->uvd.vcpu_bo) {
+		r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false);
+		if (!r) {
+			amdgpu_bo_kunmap(adev->uvd.vcpu_bo);
+			amdgpu_bo_unpin(adev->uvd.vcpu_bo);
+			amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
+		}
 
-	amdgpu_bo_unref(&adev->uvd.vcpu_bo);
+		amdgpu_bo_unref(&adev->uvd.vcpu_bo);
+	}
 
 	amdgpu_ring_fini(&adev->uvd.ring);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 9f36ed30ba11..2f8496d48c94 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -25,6 +25,7 @@
  *          Alex Deucher
  *          Jerome Glisse
  */
+#include <linux/fence-array.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
@@ -114,16 +115,26 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 /**
  * amdgpu_vm_get_bos - add the vm BOs to a duplicates list
  *
+ * @adev: amdgpu device pointer
  * @vm: vm providing the BOs
  * @duplicates: head of duplicates list
  *
  * Add the page directory to the BO duplicates list
  * for command submission.
  */
-void amdgpu_vm_get_pt_bos(struct amdgpu_vm *vm, struct list_head *duplicates)
+void amdgpu_vm_get_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+			  struct list_head *duplicates)
 {
+	uint64_t num_evictions;
 	unsigned i;
 
+	/* We only need to validate the page tables
+	 * if they aren't already valid.
+	 */
+	num_evictions = atomic64_read(&adev->num_evictions);
+	if (num_evictions == vm->last_eviction_counter)
+		return;
+
 	/* add the vm page table to the list */
 	for (i = 0; i <= vm->max_pde_used; ++i) {
 		struct amdgpu_bo_list_entry *entry = &vm->page_tables[i].entry;
@@ -162,6 +173,13 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev,
 	spin_unlock(&glob->lru_lock);
 }
 
+static bool amdgpu_vm_is_gpu_reset(struct amdgpu_device *adev,
+			      struct amdgpu_vm_id *id)
+{
+	return id->current_gpu_reset_count !=
+		atomic_read(&adev->gpu_reset_counter) ? true : false;
+}
+
 /**
  * amdgpu_vm_grab_id - allocate the next free VMID
  *
@@ -174,20 +192,69 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev,
  */
 int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 		      struct amdgpu_sync *sync, struct fence *fence,
-		      unsigned *vm_id, uint64_t *vm_pd_addr)
+		      struct amdgpu_job *job)
 {
-	uint64_t pd_addr = amdgpu_bo_gpu_offset(vm->page_directory);
 	struct amdgpu_device *adev = ring->adev;
 	struct fence *updates = sync->last_vm_update;
-	struct amdgpu_vm_id *id;
-	unsigned i = ring->idx;
-	int r;
+	struct amdgpu_vm_id *id, *idle;
+	struct fence **fences;
+	unsigned i;
+	int r = 0;
+
+	fences = kmalloc_array(sizeof(void *), adev->vm_manager.num_ids,
+			       GFP_KERNEL);
+	if (!fences)
+		return -ENOMEM;
 
 	mutex_lock(&adev->vm_manager.lock);
 
+	/* Check if we have an idle VMID */
+	i = 0;
+	list_for_each_entry(idle, &adev->vm_manager.ids_lru, list) {
+		fences[i] = amdgpu_sync_peek_fence(&idle->active, ring);
+		if (!fences[i])
+			break;
+		++i;
+	}
+
+	/* If we can't find a idle VMID to use, wait till one becomes available */
+	if (&idle->list == &adev->vm_manager.ids_lru) {
+		u64 fence_context = adev->vm_manager.fence_context + ring->idx;
+		unsigned seqno = ++adev->vm_manager.seqno[ring->idx];
+		struct fence_array *array;
+		unsigned j;
+
+		for (j = 0; j < i; ++j)
+			fence_get(fences[j]);
+
+		array = fence_array_create(i, fences, fence_context,
+					   seqno, true);
+		if (!array) {
+			for (j = 0; j < i; ++j)
+				fence_put(fences[j]);
+			kfree(fences);
+			r = -ENOMEM;
+			goto error;
+		}
+
+
+		r = amdgpu_sync_fence(ring->adev, sync, &array->base);
+		fence_put(&array->base);
+		if (r)
+			goto error;
+
+		mutex_unlock(&adev->vm_manager.lock);
+		return 0;
+
+	}
+	kfree(fences);
+
+	job->vm_needs_flush = true;
 	/* Check if we can use a VMID already assigned to this VM */
+	i = ring->idx;
 	do {
 		struct fence *flushed;
+		bool same_ring = ring->idx == i;
 
 		id = vm->ids[i++];
 		if (i == AMDGPU_MAX_RINGS)
@@ -196,67 +263,49 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 		/* Check all the prerequisites to using this VMID */
 		if (!id)
 			continue;
+		if (amdgpu_vm_is_gpu_reset(adev, id))
+			continue;
 
 		if (atomic64_read(&id->owner) != vm->client_id)
 			continue;
 
-		if (pd_addr != id->pd_gpu_addr)
+		if (job->vm_pd_addr != id->pd_gpu_addr)
 			continue;
 
-		if (id->last_user != ring &&
+		if (!same_ring &&
 		    (!id->last_flush || !fence_is_signaled(id->last_flush)))
 			continue;
 
 		flushed  = id->flushed_updates;
-		if (updates && (!flushed || fence_is_later(updates, flushed)))
+		if (updates &&
+		    (!flushed || fence_is_later(updates, flushed)))
 			continue;
 
-		/* Good we can use this VMID */
-		if (id->last_user == ring) {
-			r = amdgpu_sync_fence(ring->adev, sync,
-					      id->first);
-			if (r)
-				goto error;
-		}
-
-		/* And remember this submission as user of the VMID */
+		/* Good we can use this VMID. Remember this submission as
+		 * user of the VMID.
+		 */
 		r = amdgpu_sync_fence(ring->adev, &id->active, fence);
 		if (r)
 			goto error;
 
+		id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter);
 		list_move_tail(&id->list, &adev->vm_manager.ids_lru);
 		vm->ids[ring->idx] = id;
 
-		*vm_id = id - adev->vm_manager.ids;
-		*vm_pd_addr = AMDGPU_VM_NO_FLUSH;
-		trace_amdgpu_vm_grab_id(vm, ring->idx, *vm_id, *vm_pd_addr);
+		job->vm_id = id - adev->vm_manager.ids;
+		job->vm_needs_flush = false;
+		trace_amdgpu_vm_grab_id(vm, ring->idx, job->vm_id, job->vm_pd_addr);
 
 		mutex_unlock(&adev->vm_manager.lock);
 		return 0;
 
 	} while (i != ring->idx);
 
-	id = list_first_entry(&adev->vm_manager.ids_lru,
-			      struct amdgpu_vm_id,
-			      list);
+	/* Still no ID to use? Then use the idle one found earlier */
+	id = idle;
 
-	if (!amdgpu_sync_is_idle(&id->active)) {
-		struct list_head *head = &adev->vm_manager.ids_lru;
-		struct amdgpu_vm_id *tmp;
-
-		list_for_each_entry_safe(id, tmp, &adev->vm_manager.ids_lru,
-					 list) {
-			if (amdgpu_sync_is_idle(&id->active)) {
-				list_move(&id->list, head);
-				head = &id->list;
-			}
-		}
-		id = list_first_entry(&adev->vm_manager.ids_lru,
-				      struct amdgpu_vm_id,
-				      list);
-	}
-
-	r = amdgpu_sync_cycle_fences(sync, &id->active, fence);
+	/* Remember this submission as user of the VMID */
+	r = amdgpu_sync_fence(ring->adev, &id->active, fence);
 	if (r)
 		goto error;
 
@@ -269,22 +318,46 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 	fence_put(id->flushed_updates);
 	id->flushed_updates = fence_get(updates);
 
-	id->pd_gpu_addr = pd_addr;
-
+	id->pd_gpu_addr = job->vm_pd_addr;
+	id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter);
 	list_move_tail(&id->list, &adev->vm_manager.ids_lru);
-	id->last_user = ring;
 	atomic64_set(&id->owner, vm->client_id);
 	vm->ids[ring->idx] = id;
 
-	*vm_id = id - adev->vm_manager.ids;
-	*vm_pd_addr = pd_addr;
-	trace_amdgpu_vm_grab_id(vm, ring->idx, *vm_id, *vm_pd_addr);
+	job->vm_id = id - adev->vm_manager.ids;
+	trace_amdgpu_vm_grab_id(vm, ring->idx, job->vm_id, job->vm_pd_addr);
 
 error:
 	mutex_unlock(&adev->vm_manager.lock);
 	return r;
 }
 
+static bool amdgpu_vm_ring_has_compute_vm_bug(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	const struct amdgpu_ip_block_version *ip_block;
+
+	if (ring->type != AMDGPU_RING_TYPE_COMPUTE)
+		/* only compute rings */
+		return false;
+
+	ip_block = amdgpu_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX);
+	if (!ip_block)
+		return false;
+
+	if (ip_block->major <= 7) {
+		/* gfx7 has no workaround */
+		return true;
+	} else if (ip_block->major == 8) {
+		if (adev->gfx.mec_fw_version >= 673)
+			/* gfx8 is fixed in MEC firmware 673 */
+			return false;
+		else
+			return true;
+	}
+	return false;
+}
+
 /**
  * amdgpu_vm_flush - hardware flush the vm
  *
@@ -294,59 +367,52 @@ error:
  *
  * Emit a VM flush when it is necessary.
  */
-int amdgpu_vm_flush(struct amdgpu_ring *ring,
-		    unsigned vm_id, uint64_t pd_addr,
-		    uint32_t gds_base, uint32_t gds_size,
-		    uint32_t gws_base, uint32_t gws_size,
-		    uint32_t oa_base, uint32_t oa_size)
+int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job)
 {
 	struct amdgpu_device *adev = ring->adev;
-	struct amdgpu_vm_id *id = &adev->vm_manager.ids[vm_id];
+	struct amdgpu_vm_id *id = &adev->vm_manager.ids[job->vm_id];
 	bool gds_switch_needed = ring->funcs->emit_gds_switch && (
-		id->gds_base != gds_base ||
-		id->gds_size != gds_size ||
-		id->gws_base != gws_base ||
-		id->gws_size != gws_size ||
-		id->oa_base != oa_base ||
-		id->oa_size != oa_size);
+		id->gds_base != job->gds_base ||
+		id->gds_size != job->gds_size ||
+		id->gws_base != job->gws_base ||
+		id->gws_size != job->gws_size ||
+		id->oa_base != job->oa_base ||
+		id->oa_size != job->oa_size);
 	int r;
 
 	if (ring->funcs->emit_pipeline_sync && (
-	    pd_addr != AMDGPU_VM_NO_FLUSH || gds_switch_needed ||
-		    ring->type == AMDGPU_RING_TYPE_COMPUTE))
+	    job->vm_needs_flush || gds_switch_needed ||
+	    amdgpu_vm_ring_has_compute_vm_bug(ring)))
 		amdgpu_ring_emit_pipeline_sync(ring);
 
-	if (ring->funcs->emit_vm_flush &&
-	    pd_addr != AMDGPU_VM_NO_FLUSH) {
+	if (ring->funcs->emit_vm_flush && (job->vm_needs_flush ||
+	    amdgpu_vm_is_gpu_reset(adev, id))) {
 		struct fence *fence;
 
-		trace_amdgpu_vm_flush(pd_addr, ring->idx, vm_id);
-		amdgpu_ring_emit_vm_flush(ring, vm_id, pd_addr);
+		trace_amdgpu_vm_flush(job->vm_pd_addr, ring->idx, job->vm_id);
+		amdgpu_ring_emit_vm_flush(ring, job->vm_id, job->vm_pd_addr);
+
+		r = amdgpu_fence_emit(ring, &fence);
+		if (r)
+			return r;
 
 		mutex_lock(&adev->vm_manager.lock);
-		if ((id->pd_gpu_addr == pd_addr) && (id->last_user == ring)) {
-			r = amdgpu_fence_emit(ring, &fence);
-			if (r) {
-				mutex_unlock(&adev->vm_manager.lock);
-				return r;
-			}
-			fence_put(id->last_flush);
-			id->last_flush = fence;
-		}
+		fence_put(id->last_flush);
+		id->last_flush = fence;
 		mutex_unlock(&adev->vm_manager.lock);
 	}
 
 	if (gds_switch_needed) {
-		id->gds_base = gds_base;
-		id->gds_size = gds_size;
-		id->gws_base = gws_base;
-		id->gws_size = gws_size;
-		id->oa_base = oa_base;
-		id->oa_size = oa_size;
-		amdgpu_ring_emit_gds_switch(ring, vm_id,
-					    gds_base, gds_size,
-					    gws_base, gws_size,
-					    oa_base, oa_size);
+		id->gds_base = job->gds_base;
+		id->gds_size = job->gds_size;
+		id->gws_base = job->gws_base;
+		id->gws_size = job->gws_size;
+		id->oa_base = job->oa_base;
+		id->oa_size = job->oa_size;
+		amdgpu_ring_emit_gds_switch(ring, job->vm_id,
+					    job->gds_base, job->gds_size,
+					    job->gws_base, job->gws_size,
+					    job->oa_base, job->oa_size);
 	}
 
 	return 0;
@@ -723,7 +789,7 @@ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev,
  * @vm: requested vm
  * @start: start of GPU address range
  * @end: end of GPU address range
- * @dst: destination address to map to
+ * @dst: destination address to map to, the next dst inside the function
  * @flags: mapping flags
  *
  * Update the page tables in the range @start - @end.
@@ -737,49 +803,75 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 {
 	const uint64_t mask = AMDGPU_VM_PTE_COUNT - 1;
 
-	uint64_t last_pe_start = ~0, last_pe_end = ~0, last_dst = ~0;
-	uint64_t addr;
+	uint64_t cur_pe_start, cur_pe_end, cur_dst;
+	uint64_t addr; /* next GPU address to be updated */
+	uint64_t pt_idx;
+	struct amdgpu_bo *pt;
+	unsigned nptes; /* next number of ptes to be updated */
+	uint64_t next_pe_start;
+
+	/* initialize the variables */
+	addr = start;
+	pt_idx = addr >> amdgpu_vm_block_size;
+	pt = vm->page_tables[pt_idx].entry.robj;
+
+	if ((addr & ~mask) == (end & ~mask))
+		nptes = end - addr;
+	else
+		nptes = AMDGPU_VM_PTE_COUNT - (addr & mask);
+
+	cur_pe_start = amdgpu_bo_gpu_offset(pt);
+	cur_pe_start += (addr & mask) * 8;
+	cur_pe_end = cur_pe_start + 8 * nptes;
+	cur_dst = dst;
+
+	/* for next ptb*/
+	addr += nptes;
+	dst += nptes * AMDGPU_GPU_PAGE_SIZE;
 
 	/* walk over the address space and update the page tables */
-	for (addr = start; addr < end; ) {
-		uint64_t pt_idx = addr >> amdgpu_vm_block_size;
-		struct amdgpu_bo *pt = vm->page_tables[pt_idx].entry.robj;
-		unsigned nptes;
-		uint64_t pe_start;
+	while (addr < end) {
+		pt_idx = addr >> amdgpu_vm_block_size;
+		pt = vm->page_tables[pt_idx].entry.robj;
 
 		if ((addr & ~mask) == (end & ~mask))
 			nptes = end - addr;
 		else
 			nptes = AMDGPU_VM_PTE_COUNT - (addr & mask);
 
-		pe_start = amdgpu_bo_gpu_offset(pt);
-		pe_start += (addr & mask) * 8;
-
-		if (last_pe_end != pe_start) {
+		next_pe_start = amdgpu_bo_gpu_offset(pt);
+		next_pe_start += (addr & mask) * 8;
 
+		if (cur_pe_end == next_pe_start) {
+			/* The next ptb is consecutive to current ptb.
+			 * Don't call amdgpu_vm_frag_ptes now.
+			 * Will update two ptbs together in future.
+			*/
+			cur_pe_end += 8 * nptes;
+		} else {
 			amdgpu_vm_frag_ptes(adev, vm_update_params,
-					    last_pe_start, last_pe_end,
-					    last_dst, flags);
+					    cur_pe_start, cur_pe_end,
+					    cur_dst, flags);
 
-			last_pe_start = pe_start;
-			last_pe_end = pe_start + 8 * nptes;
-			last_dst = dst;
-		} else {
-			last_pe_end += 8 * nptes;
+			cur_pe_start = next_pe_start;
+			cur_pe_end = next_pe_start + 8 * nptes;
+			cur_dst = dst;
 		}
 
+		/* for next ptb*/
 		addr += nptes;
 		dst += nptes * AMDGPU_GPU_PAGE_SIZE;
 	}
 
-	amdgpu_vm_frag_ptes(adev, vm_update_params, last_pe_start,
-			    last_pe_end, last_dst, flags);
+	amdgpu_vm_frag_ptes(adev, vm_update_params, cur_pe_start,
+			    cur_pe_end, cur_dst, flags);
 }
 
 /**
  * amdgpu_vm_bo_update_mapping - update a mapping in the vm page table
  *
  * @adev: amdgpu_device pointer
+ * @exclusive: fence we need to sync to
  * @src: address where to copy page table entries from
  * @pages_addr: DMA addresses to use for mapping
  * @vm: requested vm
@@ -793,6 +885,7 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
  * Returns 0 for success, -EINVAL for failure.
  */
 static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
+				       struct fence *exclusive,
 				       uint64_t src,
 				       dma_addr_t *pages_addr,
 				       struct amdgpu_vm *vm,
@@ -853,6 +946,10 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 
 	vm_update_params.ib = &job->ibs[0];
 
+	r = amdgpu_sync_fence(adev, &job->sync, exclusive);
+	if (r)
+		goto error_free;
+
 	r = amdgpu_sync_resv(adev, &job->sync, vm->page_directory->tbo.resv,
 			     owner);
 	if (r)
@@ -889,6 +986,7 @@ error_free:
  * amdgpu_vm_bo_split_mapping - split a mapping into smaller chunks
  *
  * @adev: amdgpu_device pointer
+ * @exclusive: fence we need to sync to
  * @gtt_flags: flags as they are used for GTT
  * @pages_addr: DMA addresses to use for mapping
  * @vm: requested vm
@@ -902,6 +1000,7 @@ error_free:
  * Returns 0 for success, -EINVAL for failure.
  */
 static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev,
+				      struct fence *exclusive,
 				      uint32_t gtt_flags,
 				      dma_addr_t *pages_addr,
 				      struct amdgpu_vm *vm,
@@ -932,7 +1031,8 @@ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev,
 	addr += mapping->offset;
 
 	if (!pages_addr || src)
-		return amdgpu_vm_bo_update_mapping(adev, src, pages_addr, vm,
+		return amdgpu_vm_bo_update_mapping(adev, exclusive,
+						   src, pages_addr, vm,
 						   start, mapping->it.last,
 						   flags, addr, fence);
 
@@ -940,7 +1040,8 @@ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev,
 		uint64_t last;
 
 		last = min((uint64_t)mapping->it.last, start + max_size - 1);
-		r = amdgpu_vm_bo_update_mapping(adev, src, pages_addr, vm,
+		r = amdgpu_vm_bo_update_mapping(adev, exclusive,
+						src, pages_addr, vm,
 						start, last, flags, addr,
 						fence);
 		if (r)
@@ -973,6 +1074,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 	struct amdgpu_bo_va_mapping *mapping;
 	dma_addr_t *pages_addr = NULL;
 	uint32_t gtt_flags, flags;
+	struct fence *exclusive;
 	uint64_t addr;
 	int r;
 
@@ -994,8 +1096,11 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 		default:
 			break;
 		}
+
+		exclusive = reservation_object_get_excl(bo_va->bo->tbo.resv);
 	} else {
 		addr = 0;
+		exclusive = NULL;
 	}
 
 	flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem);
@@ -1007,7 +1112,8 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 	spin_unlock(&vm->status_lock);
 
 	list_for_each_entry(mapping, &bo_va->invalids, list) {
-		r = amdgpu_vm_bo_split_mapping(adev, gtt_flags, pages_addr, vm,
+		r = amdgpu_vm_bo_split_mapping(adev, exclusive,
+					       gtt_flags, pages_addr, vm,
 					       mapping, flags, addr,
 					       &bo_va->last_pt_update);
 		if (r)
@@ -1054,7 +1160,7 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
 			struct amdgpu_bo_va_mapping, list);
 		list_del(&mapping->list);
 
-		r = amdgpu_vm_bo_split_mapping(adev, 0, NULL, vm, mapping,
+		r = amdgpu_vm_bo_split_mapping(adev, NULL, 0, NULL, vm, mapping,
 					       0, 0, NULL);
 		kfree(mapping);
 		if (r)
@@ -1445,6 +1551,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	amdgpu_bo_unreserve(vm->page_directory);
 	if (r)
 		goto error_free_page_directory;
+	vm->last_eviction_counter = atomic64_read(&adev->num_evictions);
 
 	return 0;
 
@@ -1516,6 +1623,10 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev)
 			      &adev->vm_manager.ids_lru);
 	}
 
+	adev->vm_manager.fence_context = fence_context_alloc(AMDGPU_MAX_RINGS);
+	for (i = 0; i < AMDGPU_MAX_RINGS; ++i)
+		adev->vm_manager.seqno[i] = 0;
+
 	atomic_set(&adev->vm_manager.vm_pte_next_ring, 0);
 	atomic64_set(&adev->vm_manager.client_counter, 0);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
index ea407db1fbcf..5c33ed862695 100644
--- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
@@ -50,7 +50,9 @@
 #include "gmc/gmc_7_1_sh_mask.h"
 
 MODULE_FIRMWARE("radeon/bonaire_smc.bin");
+MODULE_FIRMWARE("radeon/bonaire_k_smc.bin");
 MODULE_FIRMWARE("radeon/hawaii_smc.bin");
+MODULE_FIRMWARE("radeon/hawaii_k_smc.bin");
 
 #define MC_CG_ARB_FREQ_F0           0x0a
 #define MC_CG_ARB_FREQ_F1           0x0b
@@ -736,19 +738,19 @@ static int ci_enable_didt(struct amdgpu_device *adev, bool enable)
 
 	if (pi->caps_sq_ramping || pi->caps_db_ramping ||
 	    pi->caps_td_ramping || pi->caps_tcp_ramping) {
-		gfx_v7_0_enter_rlc_safe_mode(adev);
+		adev->gfx.rlc.funcs->enter_safe_mode(adev);
 
 		if (enable) {
 			ret = ci_program_pt_config_registers(adev, didt_config_ci);
 			if (ret) {
-				gfx_v7_0_exit_rlc_safe_mode(adev);
+				adev->gfx.rlc.funcs->exit_safe_mode(adev);
 				return ret;
 			}
 		}
 
 		ci_do_enable_didt(adev, enable);
 
-		gfx_v7_0_exit_rlc_safe_mode(adev);
+		adev->gfx.rlc.funcs->exit_safe_mode(adev);
 	}
 
 	return 0;
@@ -3636,6 +3638,10 @@ static int ci_setup_default_dpm_tables(struct amdgpu_device *adev)
 
 	ci_setup_default_pcie_tables(adev);
 
+	/* save a copy of the default DPM table */
+	memcpy(&(pi->golden_dpm_table), &(pi->dpm_table),
+			sizeof(struct ci_dpm_table));
+
 	return 0;
 }
 
@@ -5754,10 +5760,18 @@ static int ci_dpm_init_microcode(struct amdgpu_device *adev)
 
 	switch (adev->asic_type) {
 	case CHIP_BONAIRE:
-		chip_name = "bonaire";
+		if ((adev->pdev->revision == 0x80) ||
+		    (adev->pdev->revision == 0x81) ||
+		    (adev->pdev->device == 0x665f))
+			chip_name = "bonaire_k";
+		else
+			chip_name = "bonaire";
 		break;
 	case CHIP_HAWAII:
-		chip_name = "hawaii";
+		if (adev->pdev->revision == 0x80)
+			chip_name = "hawaii_k";
+		else
+			chip_name = "hawaii";
 		break;
 	case CHIP_KAVERI:
 	case CHIP_KABINI:
@@ -6221,6 +6235,9 @@ static int ci_dpm_sw_fini(void *handle)
 	ci_dpm_fini(adev);
 	mutex_unlock(&adev->pm.mutex);
 
+	release_firmware(adev->pm.fw);
+	adev->pm.fw = NULL;
+
 	return 0;
 }
 
@@ -6401,6 +6418,186 @@ static int ci_dpm_set_powergating_state(void *handle,
 	return 0;
 }
 
+static int ci_dpm_print_clock_levels(struct amdgpu_device *adev,
+		enum pp_clock_type type, char *buf)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+	struct ci_single_dpm_table *sclk_table = &pi->dpm_table.sclk_table;
+	struct ci_single_dpm_table *mclk_table = &pi->dpm_table.mclk_table;
+	struct ci_single_dpm_table *pcie_table = &pi->dpm_table.pcie_speed_table;
+
+	int i, now, size = 0;
+	uint32_t clock, pcie_speed;
+
+	switch (type) {
+	case PP_SCLK:
+		amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetSclkFrequency);
+		clock = RREG32(mmSMC_MSG_ARG_0);
+
+		for (i = 0; i < sclk_table->count; i++) {
+			if (clock > sclk_table->dpm_levels[i].value)
+				continue;
+			break;
+		}
+		now = i;
+
+		for (i = 0; i < sclk_table->count; i++)
+			size += sprintf(buf + size, "%d: %uMhz %s\n",
+					i, sclk_table->dpm_levels[i].value / 100,
+					(i == now) ? "*" : "");
+		break;
+	case PP_MCLK:
+		amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetMclkFrequency);
+		clock = RREG32(mmSMC_MSG_ARG_0);
+
+		for (i = 0; i < mclk_table->count; i++) {
+			if (clock > mclk_table->dpm_levels[i].value)
+				continue;
+			break;
+		}
+		now = i;
+
+		for (i = 0; i < mclk_table->count; i++)
+			size += sprintf(buf + size, "%d: %uMhz %s\n",
+					i, mclk_table->dpm_levels[i].value / 100,
+					(i == now) ? "*" : "");
+		break;
+	case PP_PCIE:
+		pcie_speed = ci_get_current_pcie_speed(adev);
+		for (i = 0; i < pcie_table->count; i++) {
+			if (pcie_speed != pcie_table->dpm_levels[i].value)
+				continue;
+			break;
+		}
+		now = i;
+
+		for (i = 0; i < pcie_table->count; i++)
+			size += sprintf(buf + size, "%d: %s %s\n", i,
+					(pcie_table->dpm_levels[i].value == 0) ? "2.5GB, x1" :
+					(pcie_table->dpm_levels[i].value == 1) ? "5.0GB, x16" :
+					(pcie_table->dpm_levels[i].value == 2) ? "8.0GB, x16" : "",
+					(i == now) ? "*" : "");
+		break;
+	default:
+		break;
+	}
+
+	return size;
+}
+
+static int ci_dpm_force_clock_level(struct amdgpu_device *adev,
+		enum pp_clock_type type, uint32_t mask)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+
+	if (adev->pm.dpm.forced_level
+			!= AMDGPU_DPM_FORCED_LEVEL_MANUAL)
+		return -EINVAL;
+
+	switch (type) {
+	case PP_SCLK:
+		if (!pi->sclk_dpm_key_disabled)
+			amdgpu_ci_send_msg_to_smc_with_parameter(adev,
+					PPSMC_MSG_SCLKDPM_SetEnabledMask,
+					pi->dpm_level_enable_mask.sclk_dpm_enable_mask & mask);
+		break;
+
+	case PP_MCLK:
+		if (!pi->mclk_dpm_key_disabled)
+			amdgpu_ci_send_msg_to_smc_with_parameter(adev,
+					PPSMC_MSG_MCLKDPM_SetEnabledMask,
+					pi->dpm_level_enable_mask.mclk_dpm_enable_mask & mask);
+		break;
+
+	case PP_PCIE:
+	{
+		uint32_t tmp = mask & pi->dpm_level_enable_mask.pcie_dpm_enable_mask;
+		uint32_t level = 0;
+
+		while (tmp >>= 1)
+			level++;
+
+		if (!pi->pcie_dpm_key_disabled)
+			amdgpu_ci_send_msg_to_smc_with_parameter(adev,
+					PPSMC_MSG_PCIeDPM_ForceLevel,
+					level);
+		break;
+	}
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int ci_dpm_get_sclk_od(struct amdgpu_device *adev)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+	struct ci_single_dpm_table *sclk_table = &(pi->dpm_table.sclk_table);
+	struct ci_single_dpm_table *golden_sclk_table =
+			&(pi->golden_dpm_table.sclk_table);
+	int value;
+
+	value = (sclk_table->dpm_levels[sclk_table->count - 1].value -
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) *
+			100 /
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return value;
+}
+
+static int ci_dpm_set_sclk_od(struct amdgpu_device *adev, uint32_t value)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+	struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps);
+	struct ci_single_dpm_table *golden_sclk_table =
+			&(pi->golden_dpm_table.sclk_table);
+
+	if (value > 20)
+		value = 20;
+
+	ps->performance_levels[ps->performance_level_count - 1].sclk =
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value *
+			value / 100 +
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return 0;
+}
+
+static int ci_dpm_get_mclk_od(struct amdgpu_device *adev)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+	struct ci_single_dpm_table *mclk_table = &(pi->dpm_table.mclk_table);
+	struct ci_single_dpm_table *golden_mclk_table =
+			&(pi->golden_dpm_table.mclk_table);
+	int value;
+
+	value = (mclk_table->dpm_levels[mclk_table->count - 1].value -
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) *
+			100 /
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return value;
+}
+
+static int ci_dpm_set_mclk_od(struct amdgpu_device *adev, uint32_t value)
+{
+	struct ci_power_info *pi = ci_get_pi(adev);
+	struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps);
+	struct ci_single_dpm_table *golden_mclk_table =
+			&(pi->golden_dpm_table.mclk_table);
+
+	if (value > 20)
+		value = 20;
+
+	ps->performance_levels[ps->performance_level_count - 1].mclk =
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value *
+			value / 100 +
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return 0;
+}
+
 const struct amd_ip_funcs ci_dpm_ip_funcs = {
 	.name = "ci_dpm",
 	.early_init = ci_dpm_early_init,
@@ -6435,6 +6632,12 @@ static const struct amdgpu_dpm_funcs ci_dpm_funcs = {
 	.get_fan_control_mode = &ci_dpm_get_fan_control_mode,
 	.set_fan_speed_percent = &ci_dpm_set_fan_speed_percent,
 	.get_fan_speed_percent = &ci_dpm_get_fan_speed_percent,
+	.print_clock_levels = ci_dpm_print_clock_levels,
+	.force_clock_level = ci_dpm_force_clock_level,
+	.get_sclk_od = ci_dpm_get_sclk_od,
+	.set_sclk_od = ci_dpm_set_sclk_od,
+	.get_mclk_od = ci_dpm_get_mclk_od,
+	.set_mclk_od = ci_dpm_set_mclk_od,
 };
 
 static void ci_dpm_set_dpm_funcs(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h b/drivers/gpu/drm/amd/amdgpu/ci_dpm.h
index faccc30c93bf..91be2996ae7c 100644
--- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h
+++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.h
@@ -193,6 +193,7 @@ struct ci_pt_defaults {
 
 struct ci_power_info {
 	struct ci_dpm_table dpm_table;
+	struct ci_dpm_table golden_dpm_table;
 	u32 voltage_control;
 	u32 mvdd_control;
 	u32 vddci_control;
diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c
index 07bc795a4ca9..a7de4d18ac94 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik.c
@@ -962,6 +962,12 @@ static bool cik_read_bios_from_rom(struct amdgpu_device *adev,
 	return true;
 }
 
+static u32 cik_get_virtual_caps(struct amdgpu_device *adev)
+{
+	/* CIK does not support SR-IOV */
+	return 0;
+}
+
 static const struct amdgpu_allowed_register_entry cik_allowed_read_registers[] = {
 	{mmGRBM_STATUS, false},
 	{mmGB_ADDR_CONFIG, false},
@@ -1029,12 +1035,12 @@ static uint32_t cik_read_indexed_register(struct amdgpu_device *adev,
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		gfx_v7_0_select_se_sh(adev, se_num, sh_num);
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
 
 	val = RREG32(reg_offset);
 
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 	return val;
 }
@@ -1152,10 +1158,11 @@ static void kv_restore_regs_for_reset(struct amdgpu_device *adev,
 	WREG32(mmGMCON_RENG_EXECUTE, save->gmcon_reng_execute);
 }
 
-static void cik_gpu_pci_config_reset(struct amdgpu_device *adev)
+static int cik_gpu_pci_config_reset(struct amdgpu_device *adev)
 {
 	struct kv_reset_save_regs kv_save = { 0 };
 	u32 i;
+	int r = -EINVAL;
 
 	dev_info(adev->dev, "GPU pci config reset\n");
 
@@ -1171,14 +1178,20 @@ static void cik_gpu_pci_config_reset(struct amdgpu_device *adev)
 
 	/* wait for asic to come out of reset */
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff)
+		if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) {
+			/* enable BM */
+			pci_set_master(adev->pdev);
+			r = 0;
 			break;
+		}
 		udelay(1);
 	}
 
 	/* does asic init need to be run first??? */
 	if (adev->flags & AMD_IS_APU)
 		kv_restore_regs_for_reset(adev, &kv_save);
+
+	return r;
 }
 
 static void cik_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hung)
@@ -1204,13 +1217,14 @@ static void cik_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hu
  */
 static int cik_asic_reset(struct amdgpu_device *adev)
 {
+	int r;
 	cik_set_bios_scratch_engine_hung(adev, true);
 
-	cik_gpu_pci_config_reset(adev);
+	r = cik_gpu_pci_config_reset(adev);
 
 	cik_set_bios_scratch_engine_hung(adev, false);
 
-	return 0;
+	return r;
 }
 
 static int cik_set_uvd_clock(struct amdgpu_device *adev, u32 clock,
@@ -2007,9 +2021,7 @@ static const struct amdgpu_asic_funcs cik_asic_funcs =
 	.get_xclk = &cik_get_xclk,
 	.set_uvd_clocks = &cik_set_uvd_clocks,
 	.set_vce_clocks = &cik_set_vce_clocks,
-	/* these should be moved to their own ip modules */
-	.get_gpu_clock_counter = &gfx_v7_0_get_gpu_clock_counter,
-	.wait_for_mc_idle = &gmc_v7_0_mc_wait_for_idle,
+	.get_virtual_caps = &cik_get_virtual_caps,
 };
 
 static int cik_common_early_init(void *handle)
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
index 518dca43b133..46aca16a40aa 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
@@ -66,6 +66,16 @@ MODULE_FIRMWARE("radeon/mullins_sdma1.bin");
 
 u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev);
 
+
+static void cik_sdma_free_microcode(struct amdgpu_device *adev)
+{
+	int i;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+			release_firmware(adev->sdma.instance[i].fw);
+			adev->sdma.instance[i].fw = NULL;
+	}
+}
+
 /*
  * sDMA - System DMA
  * Starting with CIK, the GPU has new asynchronous
@@ -214,17 +224,6 @@ static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring,
 				  unsigned vm_id, bool ctx_switch)
 {
 	u32 extra_bits = vm_id & 0xf;
-	u32 next_rptr = ring->wptr + 5;
-
-	while ((next_rptr & 7) != 4)
-		next_rptr++;
-
-	next_rptr += 4;
-	amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
-	amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
-	amdgpu_ring_write(ring, 1); /* number of DWs to follow */
-	amdgpu_ring_write(ring, next_rptr);
 
 	/* IB packet must end on a 8 DW boundary */
 	cik_sdma_ring_insert_nop(ring, (12 - (ring->wptr & 7)) % 8);
@@ -419,6 +418,8 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 		/* Initialize the ring buffer's read and write pointers */
 		WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0);
 		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0);
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
@@ -446,7 +447,12 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
 
 		ring->ready = true;
+	}
+
+	cik_sdma_enable(adev, true);
 
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
 		r = amdgpu_ring_test_ring(ring);
 		if (r) {
 			ring->ready = false;
@@ -529,8 +535,8 @@ static int cik_sdma_start(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	/* unhalt the MEs */
-	cik_sdma_enable(adev, true);
+	/* halt the engine before programing */
+	cik_sdma_enable(adev, false);
 
 	/* start the gfx rings and rlc compute queues */
 	r = cik_sdma_gfx_resume(adev);
@@ -998,6 +1004,7 @@ static int cik_sdma_sw_fini(void *handle)
 	for (i = 0; i < adev->sdma.num_instances; i++)
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
 
+	cik_sdma_free_microcode(adev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c
index 933e425a8154..8ba07e79d4cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c
@@ -2219,6 +2219,7 @@ static void cz_dpm_powergate_vce(struct amdgpu_device *adev, bool gate)
 			}
 		}
 	} else { /*pi->caps_vce_pg*/
+		pi->vce_power_gated = gate;
 		cz_update_vce_dpm(adev);
 		cz_enable_vce_dpm(adev, !gate);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
index c90408bc0fde..d4bf133908b1 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
@@ -307,11 +307,10 @@ static void dce_v11_0_page_flip(struct amdgpu_device *adev,
 	struct amdgpu_crtc *amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
 	u32 tmp;
 
-	/* flip at hsync for async, default is vsync */
-	/* use UPDATE_IMMEDIATE_EN instead for async? */
+	/* flip immediate for async, default is vsync */
 	tmp = RREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset);
 	tmp = REG_SET_FIELD(tmp, GRPH_FLIP_CONTROL,
-			    GRPH_SURFACE_UPDATE_H_RETRACE_EN, async ? 1 : 0);
+			    GRPH_SURFACE_UPDATE_IMMEDIATE_EN, async ? 1 : 0);
 	WREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, tmp);
 	/* update the scanout addresses */
 	WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index 300ff4aab0fd..4fdfab1e9200 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -526,36 +526,16 @@ static void dce_v8_0_stop_mc_access(struct amdgpu_device *adev,
 		crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]),
 					     CRTC_CONTROL, CRTC_MASTER_EN);
 		if (crtc_enabled) {
-#if 0
-			u32 frame_count;
-			int j;
-
+#if 1
 			save->crtc_enabled[i] = true;
 			tmp = RREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i]);
 			if (REG_GET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN) == 0) {
-				amdgpu_display_vblank_wait(adev, i);
-				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+				/*it is correct only for RGB ; black is 0*/
+				WREG32(mmCRTC_BLANK_DATA_COLOR + crtc_offsets[i], 0);
 				tmp = REG_SET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN, 1);
 				WREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
-				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
-			}
-			/* wait for the next frame */
-			frame_count = amdgpu_display_vblank_get_counter(adev, i);
-			for (j = 0; j < adev->usec_timeout; j++) {
-				if (amdgpu_display_vblank_get_counter(adev, i) != frame_count)
-					break;
-				udelay(1);
-			}
-			tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK) == 0) {
-				tmp = REG_SET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK, 1);
-				WREG32(mmGRPH_UPDATE + crtc_offsets[i], tmp);
-			}
-			tmp = RREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK) == 0) {
-				tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK, 1);
-				WREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
 			}
+			mdelay(20);
 #else
 			/* XXX this is a hack to avoid strange behavior with EFI on certain systems */
 			WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
@@ -575,55 +555,22 @@ static void dce_v8_0_stop_mc_access(struct amdgpu_device *adev,
 static void dce_v8_0_resume_mc_access(struct amdgpu_device *adev,
 				      struct amdgpu_mode_mc_save *save)
 {
-	u32 tmp, frame_count;
-	int i, j;
+	u32 tmp;
+	int i;
 
 	/* update crtc base addresses */
 	for (i = 0; i < adev->mode_info.num_crtc; i++) {
 		WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i],
 		       upper_32_bits(adev->mc.vram_start));
-		WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i],
-		       upper_32_bits(adev->mc.vram_start));
 		WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + crtc_offsets[i],
 		       (u32)adev->mc.vram_start);
-		WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS + crtc_offsets[i],
-		       (u32)adev->mc.vram_start);
 
 		if (save->crtc_enabled[i]) {
-			tmp = RREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE) != 3) {
-				tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE, 3);
-				WREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i], tmp);
-			}
-			tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK)) {
-				tmp = REG_SET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK, 0);
-				WREG32(mmGRPH_UPDATE + crtc_offsets[i], tmp);
-			}
-			tmp = RREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK)) {
-				tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK, 0);
-				WREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
-			}
-			for (j = 0; j < adev->usec_timeout; j++) {
-				tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]);
-				if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_SURFACE_UPDATE_PENDING) == 0)
-					break;
-				udelay(1);
-			}
 			tmp = RREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i]);
 			tmp = REG_SET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN, 0);
-			WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 			WREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
-			WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
-			/* wait for the next frame */
-			frame_count = amdgpu_display_vblank_get_counter(adev, i);
-			for (j = 0; j < adev->usec_timeout; j++) {
-				if (amdgpu_display_vblank_get_counter(adev, i) != frame_count)
-					break;
-				udelay(1);
-			}
 		}
+		mdelay(20);
 	}
 
 	WREG32(mmVGA_MEMORY_BASE_ADDRESS_HIGH, upper_32_bits(adev->mc.vram_start));
diff --git a/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c b/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c
index 245cabf06575..ed03b75175d4 100644
--- a/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c
@@ -72,6 +72,11 @@ static int fiji_dpm_sw_init(void *handle)
 
 static int fiji_dpm_sw_fini(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	release_firmware(adev->pm.fw);
+	adev->pm.fw = NULL;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index 7f18a53ab53a..f6bd9465dbdc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -991,6 +991,22 @@ out:
 	return err;
 }
 
+static void gfx_v7_0_free_microcode(struct amdgpu_device *adev)
+{
+	release_firmware(adev->gfx.pfp_fw);
+	adev->gfx.pfp_fw = NULL;
+	release_firmware(adev->gfx.me_fw);
+	adev->gfx.me_fw = NULL;
+	release_firmware(adev->gfx.ce_fw);
+	adev->gfx.ce_fw = NULL;
+	release_firmware(adev->gfx.mec_fw);
+	adev->gfx.mec_fw = NULL;
+	release_firmware(adev->gfx.mec2_fw);
+	adev->gfx.mec2_fw = NULL;
+	release_firmware(adev->gfx.rlc_fw);
+	adev->gfx.rlc_fw = NULL;
+}
+
 /**
  * gfx_v7_0_tiling_mode_table_init - init the hw tiling table
  *
@@ -1567,9 +1583,15 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev)
  * registers are instanced per SE or SH.  0xffffffff means
  * broadcast to all SEs or SHs (CIK).
  */
-void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num)
+static void gfx_v7_0_select_se_sh(struct amdgpu_device *adev,
+				  u32 se_num, u32 sh_num, u32 instance)
 {
-	u32 data = GRBM_GFX_INDEX__INSTANCE_BROADCAST_WRITES_MASK;
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX, instance);
 
 	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff))
 		data |= GRBM_GFX_INDEX__SH_BROADCAST_WRITES_MASK |
@@ -1643,13 +1665,13 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v7_0_select_se_sh(adev, i, j);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
 			data = gfx_v7_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
@@ -1730,7 +1752,7 @@ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev)
 	 * making sure that the following register writes will be broadcasted
 	 * to all the shaders
 	 */
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
@@ -2034,17 +2056,6 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 				      unsigned vm_id, bool ctx_switch)
 {
 	u32 header, control = 0;
-	u32 next_rptr = ring->wptr + 5;
-
-	if (ctx_switch)
-		next_rptr += 2;
-
-	next_rptr += 4;
-	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM);
-	amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
-	amdgpu_ring_write(ring, next_rptr);
 
 	/* insert SWITCH_BUFFER packet before first IB in the ring frame */
 	if (ctx_switch) {
@@ -2073,22 +2084,9 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					  struct amdgpu_ib *ib,
 					  unsigned vm_id, bool ctx_switch)
 {
-	u32 header, control = 0;
-	u32 next_rptr = ring->wptr + 5;
-
-	control |= INDIRECT_BUFFER_VALID;
-	next_rptr += 4;
-	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM);
-	amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
-	amdgpu_ring_write(ring, next_rptr);
-
-	header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
-
-	control |= ib->length_dw | (vm_id << 24);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24);
 
-	amdgpu_ring_write(ring, header);
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	amdgpu_ring_write(ring,
 #ifdef __BIG_ENDIAN
 					  (2 << 0) |
@@ -3205,7 +3203,8 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev)
 		}
 	}
 	adev->gfx.rlc.cs_data = ci_cs_data;
-	adev->gfx.rlc.cp_table_size = CP_ME_TABLE_SIZE * 5 * 4;
+	adev->gfx.rlc.cp_table_size = ALIGN(CP_ME_TABLE_SIZE * 5 * 4, 2048); /* CP JT */
+	adev->gfx.rlc.cp_table_size += 64 * 1024; /* GDS */
 
 	src_ptr = adev->gfx.rlc.reg_list;
 	dws = adev->gfx.rlc.reg_list_size;
@@ -3363,7 +3362,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v7_0_select_se_sh(adev, i, j);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
 			for (k = 0; k < adev->usec_timeout; k++) {
 				if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0)
 					break;
@@ -3371,7 +3370,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -3418,7 +3417,7 @@ static u32 gfx_v7_0_halt_rlc(struct amdgpu_device *adev)
 	return orig;
 }
 
-void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev)
+static void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev)
 {
 	u32 tmp, i, mask;
 
@@ -3440,7 +3439,7 @@ void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev)
 	}
 }
 
-void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev)
+static void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev)
 {
 	u32 tmp;
 
@@ -3455,7 +3454,7 @@ void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev)
  *
  * Halt the RLC ME (MicroEngine) (CIK).
  */
-void gfx_v7_0_rlc_stop(struct amdgpu_device *adev)
+static void gfx_v7_0_rlc_stop(struct amdgpu_device *adev)
 {
 	WREG32(mmRLC_CNTL, 0);
 
@@ -3531,7 +3530,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev)
 	WREG32(mmRLC_LB_CNTR_MAX, 0x00008000);
 
 	mutex_lock(&adev->grbm_idx_mutex);
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	WREG32(mmRLC_LB_INIT_CU_MASK, 0xffffffff);
 	WREG32(mmRLC_LB_PARAMS, 0x00600408);
 	WREG32(mmRLC_LB_CNTL, 0x80000004);
@@ -3571,7 +3570,7 @@ static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		tmp2 = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK |
@@ -3622,7 +3621,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK |
@@ -3673,7 +3672,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
 		tmp = gfx_v7_0_halt_rlc(adev);
 
 		mutex_lock(&adev->grbm_idx_mutex);
-		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+		gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | RLC_SERDES_WR_CTRL__MGCG_OVERRIDE_1_MASK;
@@ -3851,6 +3850,20 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev,
 	}
 }
 
+static void gfx_v7_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev,
+						 u32 bitmap)
+{
+	u32 data;
+
+	if (!bitmap)
+		return;
+
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
+
+	WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data);
+}
+
 static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev)
 {
 	u32 data, mask;
@@ -4107,7 +4120,7 @@ static void gfx_v7_0_fini_pg(struct amdgpu_device *adev)
  * Fetches a GPU clock counter snapshot (SI).
  * Returns the 64 bit clock counter snapshot.
  */
-uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev)
+static uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
 	uint64_t clock;
 
@@ -4167,12 +4180,24 @@ static void gfx_v7_0_ring_emit_gds_switch(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, (1 << (oa_size + oa_base)) - (1 << oa_base));
 }
 
+static const struct amdgpu_gfx_funcs gfx_v7_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v7_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v7_0_select_se_sh,
+};
+
+static const struct amdgpu_rlc_funcs gfx_v7_0_rlc_funcs = {
+	.enter_safe_mode = gfx_v7_0_enter_rlc_safe_mode,
+	.exit_safe_mode = gfx_v7_0_exit_rlc_safe_mode
+};
+
 static int gfx_v7_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	adev->gfx.num_gfx_rings = GFX7_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = GFX7_NUM_COMPUTE_RINGS;
+	adev->gfx.funcs = &gfx_v7_0_gfx_funcs;
+	adev->gfx.rlc.funcs = &gfx_v7_0_rlc_funcs;
 	gfx_v7_0_set_ring_funcs(adev);
 	gfx_v7_0_set_irq_funcs(adev);
 	gfx_v7_0_set_gds_init(adev);
@@ -4489,6 +4514,7 @@ static int gfx_v7_0_sw_fini(void *handle)
 	gfx_v7_0_cp_compute_fini(adev);
 	gfx_v7_0_rlc_fini(adev);
 	gfx_v7_0_mec_fini(adev);
+	gfx_v7_0_free_microcode(adev);
 
 	return 0;
 }
@@ -4816,7 +4842,7 @@ static int gfx_v7_0_eop_irq(struct amdgpu_device *adev,
 	case 2:
 		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 			ring = &adev->gfx.compute_ring[i];
-			if ((ring->me == me_id) & (ring->pipe == pipe_id))
+			if ((ring->me == me_id) && (ring->pipe == pipe_id))
 				amdgpu_fence_process(ring);
 		}
 		break;
@@ -5015,16 +5041,22 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
 	int i, j, k, counter, active_cu_number = 0;
 	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
 	struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
+	unsigned disable_masks[4 * 2];
 
 	memset(cu_info, 0, sizeof(*cu_info));
 
+	amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2);
+
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v7_0_select_se_sh(adev, i, j);
+			gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff);
+			if (i < 4 && j < 2)
+				gfx_v7_0_set_user_cu_inactive_bitmap(
+					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v7_0_get_cu_active_bitmap(adev);
 			cu_info->bitmap[i][j] = bitmap;
 
@@ -5040,7 +5072,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev)
 			ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8));
 		}
 	}
-	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h
index e747aa935c88..94e3ea147c26 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h
@@ -26,11 +26,4 @@
 
 extern const struct amd_ip_funcs gfx_v7_0_ip_funcs;
 
-/* XXX these shouldn't be exported */
-void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev);
-void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev);
-void gfx_v7_0_rlc_stop(struct amdgpu_device *adev);
-uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev);
-void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num);
-
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index f19bab68fd83..c30b6ac25d89 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -297,7 +297,8 @@ static const u32 polaris11_golden_common_all[] =
 static const u32 golden_settings_polaris10_a11[] =
 {
 	mmATC_MISC_CG, 0x000c0fc0, 0x000c0200,
-	mmCB_HW_CONTROL, 0xfffdf3cf, 0x00006208,
+	mmCB_HW_CONTROL, 0xfffdf3cf, 0x00007208,
+	mmCB_HW_CONTROL_2, 0, 0x0f000000,
 	mmCB_HW_CONTROL_3, 0x000001ff, 0x00000040,
 	mmDB_DEBUG2, 0xf00fffff, 0x00000400,
 	mmPA_SC_ENHANCE, 0xffffffff, 0x20000001,
@@ -836,6 +837,26 @@ err1:
 	return r;
 }
 
+
+static void gfx_v8_0_free_microcode(struct amdgpu_device *adev) {
+	release_firmware(adev->gfx.pfp_fw);
+	adev->gfx.pfp_fw = NULL;
+	release_firmware(adev->gfx.me_fw);
+	adev->gfx.me_fw = NULL;
+	release_firmware(adev->gfx.ce_fw);
+	adev->gfx.ce_fw = NULL;
+	release_firmware(adev->gfx.rlc_fw);
+	adev->gfx.rlc_fw = NULL;
+	release_firmware(adev->gfx.mec_fw);
+	adev->gfx.mec_fw = NULL;
+	if ((adev->asic_type != CHIP_STONEY) &&
+	    (adev->asic_type != CHIP_TOPAZ))
+		release_firmware(adev->gfx.mec2_fw);
+	adev->gfx.mec2_fw = NULL;
+
+	kfree(adev->gfx.rlc.register_list_format);
+}
+
 static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
 {
 	const char *chip_name;
@@ -1129,6 +1150,71 @@ static void gfx_v8_0_get_csb_buffer(struct amdgpu_device *adev,
 	buffer[count++] = cpu_to_le32(0);
 }
 
+static void cz_init_cp_jump_table(struct amdgpu_device *adev)
+{
+	const __le32 *fw_data;
+	volatile u32 *dst_ptr;
+	int me, i, max_me = 4;
+	u32 bo_offset = 0;
+	u32 table_offset, table_size;
+
+	if (adev->asic_type == CHIP_CARRIZO)
+		max_me = 5;
+
+	/* write the cp table buffer */
+	dst_ptr = adev->gfx.rlc.cp_table_ptr;
+	for (me = 0; me < max_me; me++) {
+		if (me == 0) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.ce_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 1) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.pfp_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 2) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.me_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 3) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.mec_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else  if (me == 4) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.mec2_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.mec2_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		}
+
+		for (i = 0; i < table_size; i ++) {
+			dst_ptr[bo_offset + i] =
+				cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
+		}
+
+		bo_offset += table_size;
+	}
+}
+
 static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev)
 {
 	int r;
@@ -1144,6 +1230,18 @@ static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev)
 		amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj);
 		adev->gfx.rlc.clear_state_obj = NULL;
 	}
+
+	/* jump table block */
+	if (adev->gfx.rlc.cp_table_obj) {
+		r = amdgpu_bo_reserve(adev->gfx.rlc.cp_table_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r);
+		amdgpu_bo_unpin(adev->gfx.rlc.cp_table_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
+
+		amdgpu_bo_unref(&adev->gfx.rlc.cp_table_obj);
+		adev->gfx.rlc.cp_table_obj = NULL;
+	}
 }
 
 static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
@@ -1200,6 +1298,46 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
 	}
 
+	if ((adev->asic_type == CHIP_CARRIZO) ||
+	    (adev->asic_type == CHIP_STONEY)) {
+		adev->gfx.rlc.cp_table_size = ALIGN(96 * 5 * 4, 2048) + (64 * 1024); /* JT + GDS */
+		if (adev->gfx.rlc.cp_table_obj == NULL) {
+			r = amdgpu_bo_create(adev, adev->gfx.rlc.cp_table_size, PAGE_SIZE, true,
+					     AMDGPU_GEM_DOMAIN_VRAM,
+					     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
+					     NULL, NULL,
+					     &adev->gfx.rlc.cp_table_obj);
+			if (r) {
+				dev_warn(adev->dev, "(%d) create RLC cp table bo failed\n", r);
+				return r;
+			}
+		}
+
+		r = amdgpu_bo_reserve(adev->gfx.rlc.cp_table_obj, false);
+		if (unlikely(r != 0)) {
+			dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r);
+			return r;
+		}
+		r = amdgpu_bo_pin(adev->gfx.rlc.cp_table_obj, AMDGPU_GEM_DOMAIN_VRAM,
+				  &adev->gfx.rlc.cp_table_gpu_addr);
+		if (r) {
+			amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
+			dev_warn(adev->dev, "(%d) pin RLC cp_table bo failed\n", r);
+			return r;
+		}
+		r = amdgpu_bo_kmap(adev->gfx.rlc.cp_table_obj, (void **)&adev->gfx.rlc.cp_table_ptr);
+		if (r) {
+			dev_warn(adev->dev, "(%d) map RLC cp table bo failed\n", r);
+			return r;
+		}
+
+		cz_init_cp_jump_table(adev);
+
+		amdgpu_bo_kunmap(adev->gfx.rlc.cp_table_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
+
+	}
+
 	return 0;
 }
 
@@ -1983,7 +2121,7 @@ static int gfx_v8_0_sw_fini(void *handle)
 
 	gfx_v8_0_rlc_fini(adev);
 
-	kfree(adev->gfx.rlc.register_list_format);
+	gfx_v8_0_free_microcode(adev);
 
 	return 0;
 }
@@ -3308,9 +3446,15 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev)
 	}
 }
 
-void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num)
+static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev,
+				  u32 se_num, u32 sh_num, u32 instance)
 {
-	u32 data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1);
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1);
+	else
+		data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX, instance);
 
 	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff)) {
 		data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_BROADCAST_WRITES, 1);
@@ -3360,13 +3504,13 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v8_0_select_se_sh(adev, i, j);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
 			data = gfx_v8_0_get_rb_active_bitmap(adev);
 			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
 					       rb_bitmap_width_per_sh);
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	adev->gfx.config.backend_enable_mask = active_rbs;
@@ -3470,7 +3614,7 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
 	 * making sure that the following register writes will be broadcasted
 	 * to all the shaders
 	 */
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 
 	WREG32(mmPA_SC_FIFO_SIZE,
 		   (adev->gfx.config.sc_prim_fifo_size_frontend <<
@@ -3493,7 +3637,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			gfx_v8_0_select_se_sh(adev, i, j);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
 			for (k = 0; k < adev->usec_timeout; k++) {
 				if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0)
 					break;
@@ -3501,7 +3645,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
 			}
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK |
@@ -3662,13 +3806,13 @@ static void gfx_v8_0_enable_save_restore_machine(struct amdgpu_device *adev)
 	WREG32(mmRLC_SRM_CNTL, data);
 }
 
-static void polaris11_init_power_gating(struct amdgpu_device *adev)
+static void gfx_v8_0_init_power_gating(struct amdgpu_device *adev)
 {
 	uint32_t data;
 
 	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
-			AMD_PG_SUPPORT_GFX_SMG |
-			AMD_PG_SUPPORT_GFX_DMG)) {
+			      AMD_PG_SUPPORT_GFX_SMG |
+			      AMD_PG_SUPPORT_GFX_DMG)) {
 		data = RREG32(mmCP_RB_WPTR_POLL_CNTL);
 		data &= ~CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK;
 		data |= (0x60 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT);
@@ -3693,6 +3837,53 @@ static void polaris11_init_power_gating(struct amdgpu_device *adev)
 	}
 }
 
+static void cz_enable_sck_slow_down_on_power_up(struct amdgpu_device *adev,
+						bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(mmRLC_PG_CNTL);
+
+	if (enable)
+		data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK;
+
+	if (orig != data)
+		WREG32(mmRLC_PG_CNTL, data);
+}
+
+static void cz_enable_sck_slow_down_on_power_down(struct amdgpu_device *adev,
+						  bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(mmRLC_PG_CNTL);
+
+	if (enable)
+		data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK;
+
+	if (orig != data)
+		WREG32(mmRLC_PG_CNTL, data);
+}
+
+static void cz_enable_cp_power_gating(struct amdgpu_device *adev, bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(mmRLC_PG_CNTL);
+
+	if (enable)
+		data &= ~RLC_PG_CNTL__CP_PG_DISABLE_MASK;
+	else
+		data |= RLC_PG_CNTL__CP_PG_DISABLE_MASK;
+
+	if (orig != data)
+		WREG32(mmRLC_PG_CNTL, data);
+}
+
 static void gfx_v8_0_init_pg(struct amdgpu_device *adev)
 {
 	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
@@ -3705,8 +3896,25 @@ static void gfx_v8_0_init_pg(struct amdgpu_device *adev)
 		gfx_v8_0_init_save_restore_list(adev);
 		gfx_v8_0_enable_save_restore_machine(adev);
 
-		if (adev->asic_type == CHIP_POLARIS11)
-			polaris11_init_power_gating(adev);
+		if ((adev->asic_type == CHIP_CARRIZO) ||
+		    (adev->asic_type == CHIP_STONEY)) {
+			WREG32(mmRLC_JUMP_TABLE_RESTORE, adev->gfx.rlc.cp_table_gpu_addr >> 8);
+			gfx_v8_0_init_power_gating(adev);
+			WREG32(mmRLC_PG_ALWAYS_ON_CU_MASK, adev->gfx.cu_info.ao_cu_mask);
+			if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) {
+				cz_enable_sck_slow_down_on_power_up(adev, true);
+				cz_enable_sck_slow_down_on_power_down(adev, true);
+			} else {
+				cz_enable_sck_slow_down_on_power_up(adev, false);
+				cz_enable_sck_slow_down_on_power_down(adev, false);
+			}
+			if (adev->pg_flags & AMD_PG_SUPPORT_CP)
+				cz_enable_cp_power_gating(adev, true);
+			else
+				cz_enable_cp_power_gating(adev, false);
+		} else if (adev->asic_type == CHIP_POLARIS11) {
+			gfx_v8_0_init_power_gating(adev);
+		}
 	}
 }
 
@@ -3974,11 +4182,15 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev)
 		amdgpu_ring_write(ring, 0x3a00161a);
 		amdgpu_ring_write(ring, 0x0000002e);
 		break;
-	case CHIP_TOPAZ:
 	case CHIP_CARRIZO:
 		amdgpu_ring_write(ring, 0x00000002);
 		amdgpu_ring_write(ring, 0x00000000);
 		break;
+	case CHIP_TOPAZ:
+		amdgpu_ring_write(ring, adev->gfx.config.num_rbs == 1 ?
+				0x00000000 : 0x00000002);
+		amdgpu_ring_write(ring, 0x00000000);
+		break;
 	case CHIP_STONEY:
 		amdgpu_ring_write(ring, 0x00000000);
 		amdgpu_ring_write(ring, 0x00000000);
@@ -4941,7 +5153,7 @@ static int gfx_v8_0_soft_reset(void *handle)
  * Fetches a GPU clock counter snapshot.
  * Returns the 64 bit clock counter snapshot.
  */
-uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev)
+static uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
 	uint64_t clock;
 
@@ -5001,12 +5213,18 @@ static void gfx_v8_0_ring_emit_gds_switch(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, (1 << (oa_size + oa_base)) - (1 << oa_base));
 }
 
+static const struct amdgpu_gfx_funcs gfx_v8_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v8_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v8_0_select_se_sh,
+};
+
 static int gfx_v8_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	adev->gfx.num_gfx_rings = GFX8_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = GFX8_NUM_COMPUTE_RINGS;
+	adev->gfx.funcs = &gfx_v8_0_gfx_funcs;
 	gfx_v8_0_set_ring_funcs(adev);
 	gfx_v8_0_set_irq_funcs(adev);
 	gfx_v8_0_set_gds_init(adev);
@@ -5039,51 +5257,43 @@ static int gfx_v8_0_late_init(void *handle)
 	return 0;
 }
 
-static void polaris11_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev,
-		bool enable)
+static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev,
+						       bool enable)
 {
 	uint32_t data, temp;
 
-	/* Send msg to SMU via Powerplay */
-	amdgpu_set_powergating_state(adev,
-			AMD_IP_BLOCK_TYPE_SMC,
-			enable ? AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE);
+	if (adev->asic_type == CHIP_POLARIS11)
+		/* Send msg to SMU via Powerplay */
+		amdgpu_set_powergating_state(adev,
+					     AMD_IP_BLOCK_TYPE_SMC,
+					     enable ?
+					     AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE);
 
-	if (enable) {
-		/* Enable static MGPG */
-		temp = data = RREG32(mmRLC_PG_CNTL);
+	temp = data = RREG32(mmRLC_PG_CNTL);
+	/* Enable static MGPG */
+	if (enable)
 		data |= RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK;
-
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
-	} else {
-		temp = data = RREG32(mmRLC_PG_CNTL);
+	else
 		data &= ~RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK;
 
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
-	}
+	if (temp != data)
+		WREG32(mmRLC_PG_CNTL, data);
 }
 
-static void polaris11_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *adev,
-		bool enable)
+static void gfx_v8_0_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *adev,
+							bool enable)
 {
 	uint32_t data, temp;
 
-	if (enable) {
-		/* Enable dynamic MGPG */
-		temp = data = RREG32(mmRLC_PG_CNTL);
+	temp = data = RREG32(mmRLC_PG_CNTL);
+	/* Enable dynamic MGPG */
+	if (enable)
 		data |= RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK;
-
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
-	} else {
-		temp = data = RREG32(mmRLC_PG_CNTL);
+	else
 		data &= ~RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK;
 
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
-	}
+	if (temp != data)
+		WREG32(mmRLC_PG_CNTL, data);
 }
 
 static void polaris11_enable_gfx_quick_mg_power_gating(struct amdgpu_device *adev,
@@ -5091,19 +5301,63 @@ static void polaris11_enable_gfx_quick_mg_power_gating(struct amdgpu_device *ade
 {
 	uint32_t data, temp;
 
-	if (enable) {
-		/* Enable quick PG */
-		temp = data = RREG32(mmRLC_PG_CNTL);
-		data |= 0x100000;
+	temp = data = RREG32(mmRLC_PG_CNTL);
+	/* Enable quick PG */
+	if (enable)
+		data |= RLC_PG_CNTL__QUICK_PG_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__QUICK_PG_ENABLE_MASK;
 
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
-	} else {
-		temp = data = RREG32(mmRLC_PG_CNTL);
-		data &= ~0x100000;
+	if (temp != data)
+		WREG32(mmRLC_PG_CNTL, data);
+}
 
-		if (temp != data)
-			WREG32(mmRLC_PG_CNTL, data);
+static void cz_enable_gfx_cg_power_gating(struct amdgpu_device *adev,
+					  bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(mmRLC_PG_CNTL);
+
+	if (enable)
+		data |= RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK;
+
+	if (orig != data)
+		WREG32(mmRLC_PG_CNTL, data);
+}
+
+static void cz_enable_gfx_pipeline_power_gating(struct amdgpu_device *adev,
+						bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(mmRLC_PG_CNTL);
+
+	if (enable)
+		data |= RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK;
+	else
+		data &= ~RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK;
+
+	if (orig != data)
+		WREG32(mmRLC_PG_CNTL, data);
+
+	/* Read any GFX register to wake up GFX. */
+	if (!enable)
+		data = RREG32(mmDB_RENDER_CONTROL);
+}
+
+static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev,
+					  bool enable)
+{
+	if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) {
+		cz_enable_gfx_cg_power_gating(adev, true);
+		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PIPELINE)
+			cz_enable_gfx_pipeline_power_gating(adev, true);
+	} else {
+		cz_enable_gfx_cg_power_gating(adev, false);
+		cz_enable_gfx_pipeline_power_gating(adev, false);
 	}
 }
 
@@ -5111,21 +5365,42 @@ static int gfx_v8_0_set_powergating_state(void *handle,
 					  enum amd_powergating_state state)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	bool enable = (state == AMD_PG_STATE_GATE) ? true : false;
 
 	if (!(adev->pg_flags & AMD_PG_SUPPORT_GFX_PG))
 		return 0;
 
 	switch (adev->asic_type) {
+	case CHIP_CARRIZO:
+	case CHIP_STONEY:
+		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)
+			cz_update_gfx_cg_power_gating(adev, enable);
+
+		if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) && enable)
+			gfx_v8_0_enable_gfx_static_mg_power_gating(adev, true);
+		else
+			gfx_v8_0_enable_gfx_static_mg_power_gating(adev, false);
+
+		if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG) && enable)
+			gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, true);
+		else
+			gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, false);
+		break;
 	case CHIP_POLARIS11:
-		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG)
-			polaris11_enable_gfx_static_mg_power_gating(adev,
-					state == AMD_PG_STATE_GATE ? true : false);
-		else if (adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG)
-			polaris11_enable_gfx_dynamic_mg_power_gating(adev,
-					state == AMD_PG_STATE_GATE ? true : false);
+		if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) && enable)
+			gfx_v8_0_enable_gfx_static_mg_power_gating(adev, true);
+		else
+			gfx_v8_0_enable_gfx_static_mg_power_gating(adev, false);
+
+		if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG) && enable)
+			gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, true);
+		else
+			gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, false);
+
+		if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_QUICK_MG) && enable)
+			polaris11_enable_gfx_quick_mg_power_gating(adev, true);
 		else
-			polaris11_enable_gfx_quick_mg_power_gating(adev,
-					state == AMD_PG_STATE_GATE ? true : false);
+			polaris11_enable_gfx_quick_mg_power_gating(adev, false);
 		break;
 	default:
 		break;
@@ -5139,7 +5414,7 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev,
 {
 	uint32_t data;
 
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 
 	WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 	WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
@@ -5527,6 +5802,8 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
 			WREG32(mmRLC_CGCG_CGLS_CTRL, data);
 	}
 
+	gfx_v8_0_wait_for_rlc_serdes(adev);
+
 	adev->gfx.rlc.funcs->exit_safe_mode(adev);
 }
 static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
@@ -5652,17 +5929,6 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring,
 				      unsigned vm_id, bool ctx_switch)
 {
 	u32 header, control = 0;
-	u32 next_rptr = ring->wptr + 5;
-
-	if (ctx_switch)
-		next_rptr += 2;
-
-	next_rptr += 4;
-	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM);
-	amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
-	amdgpu_ring_write(ring, next_rptr);
 
 	/* insert SWITCH_BUFFER packet before first IB in the ring frame */
 	if (ctx_switch) {
@@ -5691,23 +5957,9 @@ static void gfx_v8_0_ring_emit_ib_compute(struct amdgpu_ring *ring,
 					  struct amdgpu_ib *ib,
 					  unsigned vm_id, bool ctx_switch)
 {
-	u32 header, control = 0;
-	u32 next_rptr = ring->wptr + 5;
-
-	control |= INDIRECT_BUFFER_VALID;
-
-	next_rptr += 4;
-	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM);
-	amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
-	amdgpu_ring_write(ring, next_rptr);
-
-	header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+	u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24);
 
-	control |= ib->length_dw | (vm_id << 24);
-
-	amdgpu_ring_write(ring, header);
+	amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
 	amdgpu_ring_write(ring,
 #ifdef __BIG_ENDIAN
 					  (2 << 0) |
@@ -6160,9 +6412,9 @@ static void gfx_v8_0_set_rlc_funcs(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
 	case CHIP_TOPAZ:
-	case CHIP_STONEY:
 		adev->gfx.rlc.funcs = &iceland_rlc_funcs;
 		break;
+	case CHIP_STONEY:
 	case CHIP_CARRIZO:
 		adev->gfx.rlc.funcs = &cz_rlc_funcs;
 		break;
@@ -6200,6 +6452,20 @@ static void gfx_v8_0_set_gds_init(struct amdgpu_device *adev)
 	}
 }
 
+static void gfx_v8_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev,
+						 u32 bitmap)
+{
+	u32 data;
+
+	if (!bitmap)
+		return;
+
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
+
+	WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data);
+}
+
 static u32 gfx_v8_0_get_cu_active_bitmap(struct amdgpu_device *adev)
 {
 	u32 data, mask;
@@ -6220,16 +6486,22 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev)
 	int i, j, k, counter, active_cu_number = 0;
 	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
 	struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
+	unsigned disable_masks[4 * 2];
 
 	memset(cu_info, 0, sizeof(*cu_info));
 
+	amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2);
+
 	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			gfx_v8_0_select_se_sh(adev, i, j);
+			gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff);
+			if (i < 4 && j < 2)
+				gfx_v8_0_set_user_cu_inactive_bitmap(
+					adev, disable_masks[i * 2 + j]);
 			bitmap = gfx_v8_0_get_cu_active_bitmap(adev);
 			cu_info->bitmap[i][j] = bitmap;
 
@@ -6245,7 +6517,7 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev)
 			ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8));
 		}
 	}
-	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+	gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 
 	cu_info->number = active_cu_number;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h
index 16a49f53a2fa..bc82c794312c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h
@@ -26,7 +26,6 @@
 
 extern const struct amd_ip_funcs gfx_v8_0_ip_funcs;
 
-uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev);
 void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index 1feb6439cb0b..d24a82bd0c7a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -39,6 +39,7 @@
 
 static void gmc_v7_0_set_gart_funcs(struct amdgpu_device *adev);
 static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev);
+static int gmc_v7_0_wait_for_idle(void *handle);
 
 MODULE_FIRMWARE("radeon/bonaire_mc.bin");
 MODULE_FIRMWARE("radeon/hawaii_mc.bin");
@@ -73,39 +74,15 @@ static void gmc_v7_0_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
-/**
- * gmc7_mc_wait_for_idle - wait for MC idle callback.
- *
- * @adev: amdgpu_device pointer
- *
- * Wait for the MC (memory controller) to be idle.
- * (evergreen+).
- * Returns 0 if the MC is idle, -1 if not.
- */
-int gmc_v7_0_mc_wait_for_idle(struct amdgpu_device *adev)
-{
-	unsigned i;
-	u32 tmp;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		/* read MC_STATUS */
-		tmp = RREG32(mmSRBM_STATUS) & 0x1F00;
-		if (!tmp)
-			return 0;
-		udelay(1);
-	}
-	return -1;
-}
-
-void gmc_v7_0_mc_stop(struct amdgpu_device *adev,
-		      struct amdgpu_mode_mc_save *save)
+static void gmc_v7_0_mc_stop(struct amdgpu_device *adev,
+			     struct amdgpu_mode_mc_save *save)
 {
 	u32 blackout;
 
 	if (adev->mode_info.num_crtc)
 		amdgpu_display_stop_mc_access(adev, save);
 
-	amdgpu_asic_wait_for_mc_idle(adev);
+	gmc_v7_0_wait_for_idle((void *)adev);
 
 	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
 	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
@@ -120,8 +97,8 @@ void gmc_v7_0_mc_stop(struct amdgpu_device *adev,
 	udelay(100);
 }
 
-void gmc_v7_0_mc_resume(struct amdgpu_device *adev,
-			struct amdgpu_mode_mc_save *save)
+static void gmc_v7_0_mc_resume(struct amdgpu_device *adev,
+			       struct amdgpu_mode_mc_save *save)
 {
 	u32 tmp;
 
@@ -311,7 +288,7 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
 		amdgpu_display_set_vga_render_state(adev, false);
 
 	gmc_v7_0_mc_stop(adev, &save);
-	if (amdgpu_asic_wait_for_mc_idle(adev)) {
+	if (gmc_v7_0_wait_for_idle((void *)adev)) {
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
 	}
 	/* Update configuration */
@@ -331,7 +308,7 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev)
 	WREG32(mmMC_VM_AGP_BASE, 0);
 	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
 	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
-	if (amdgpu_asic_wait_for_mc_idle(adev)) {
+	if (gmc_v7_0_wait_for_idle((void *)adev)) {
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
 	}
 	gmc_v7_0_mc_resume(adev, &save);
@@ -1137,7 +1114,7 @@ static int gmc_v7_0_soft_reset(void *handle)
 
 	if (srbm_soft_reset) {
 		gmc_v7_0_mc_stop(adev, &save);
-		if (gmc_v7_0_wait_for_idle(adev)) {
+		if (gmc_v7_0_wait_for_idle((void *)adev)) {
 			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
 		}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h
index 36fcbbc46ada..0b386b5d2f7a 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h
@@ -26,11 +26,4 @@
 
 extern const struct amd_ip_funcs gmc_v7_0_ip_funcs;
 
-/* XXX these shouldn't be exported */
-void gmc_v7_0_mc_stop(struct amdgpu_device *adev,
-		      struct amdgpu_mode_mc_save *save);
-void gmc_v7_0_mc_resume(struct amdgpu_device *adev,
-			struct amdgpu_mode_mc_save *save);
-int gmc_v7_0_mc_wait_for_idle(struct amdgpu_device *adev);
-
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 9945d5bbf1fe..717359d3ba8c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -41,6 +41,7 @@
 
 static void gmc_v8_0_set_gart_funcs(struct amdgpu_device *adev);
 static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev);
+static int gmc_v8_0_wait_for_idle(void *handle);
 
 MODULE_FIRMWARE("amdgpu/tonga_mc.bin");
 MODULE_FIRMWARE("amdgpu/polaris11_mc.bin");
@@ -147,44 +148,15 @@ static void gmc_v8_0_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
-/**
- * gmc8_mc_wait_for_idle - wait for MC idle callback.
- *
- * @adev: amdgpu_device pointer
- *
- * Wait for the MC (memory controller) to be idle.
- * (evergreen+).
- * Returns 0 if the MC is idle, -1 if not.
- */
-int gmc_v8_0_mc_wait_for_idle(struct amdgpu_device *adev)
-{
-	unsigned i;
-	u32 tmp;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		/* read MC_STATUS */
-		tmp = RREG32(mmSRBM_STATUS) & (SRBM_STATUS__VMC_BUSY_MASK |
-					       SRBM_STATUS__MCB_BUSY_MASK |
-					       SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK |
-					       SRBM_STATUS__MCC_BUSY_MASK |
-					       SRBM_STATUS__MCD_BUSY_MASK |
-					       SRBM_STATUS__VMC1_BUSY_MASK);
-		if (!tmp)
-			return 0;
-		udelay(1);
-	}
-	return -1;
-}
-
-void gmc_v8_0_mc_stop(struct amdgpu_device *adev,
-		      struct amdgpu_mode_mc_save *save)
+static void gmc_v8_0_mc_stop(struct amdgpu_device *adev,
+			     struct amdgpu_mode_mc_save *save)
 {
 	u32 blackout;
 
 	if (adev->mode_info.num_crtc)
 		amdgpu_display_stop_mc_access(adev, save);
 
-	amdgpu_asic_wait_for_mc_idle(adev);
+	gmc_v8_0_wait_for_idle(adev);
 
 	blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL);
 	if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) {
@@ -199,8 +171,8 @@ void gmc_v8_0_mc_stop(struct amdgpu_device *adev,
 	udelay(100);
 }
 
-void gmc_v8_0_mc_resume(struct amdgpu_device *adev,
-			struct amdgpu_mode_mc_save *save)
+static void gmc_v8_0_mc_resume(struct amdgpu_device *adev,
+			       struct amdgpu_mode_mc_save *save)
 {
 	u32 tmp;
 
@@ -393,7 +365,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 		amdgpu_display_set_vga_render_state(adev, false);
 
 	gmc_v8_0_mc_stop(adev, &save);
-	if (amdgpu_asic_wait_for_mc_idle(adev)) {
+	if (gmc_v8_0_wait_for_idle((void *)adev)) {
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
 	}
 	/* Update configuration */
@@ -413,7 +385,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev)
 	WREG32(mmMC_VM_AGP_BASE, 0);
 	WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF);
 	WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF);
-	if (amdgpu_asic_wait_for_mc_idle(adev)) {
+	if (gmc_v8_0_wait_for_idle((void *)adev)) {
 		dev_warn(adev->dev, "Wait for MC idle timedout !\n");
 	}
 	gmc_v8_0_mc_resume(adev, &save);
@@ -1140,7 +1112,7 @@ static int gmc_v8_0_soft_reset(void *handle)
 
 	if (srbm_soft_reset) {
 		gmc_v8_0_mc_stop(adev, &save);
-		if (gmc_v8_0_wait_for_idle(adev)) {
+		if (gmc_v8_0_wait_for_idle((void *)adev)) {
 			dev_warn(adev->dev, "Wait for GMC idle timed out !\n");
 		}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h
index 973436086b38..fc5001a8119d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h
@@ -26,11 +26,4 @@
 
 extern const struct amd_ip_funcs gmc_v8_0_ip_funcs;
 
-/* XXX these shouldn't be exported */
-void gmc_v8_0_mc_stop(struct amdgpu_device *adev,
-		      struct amdgpu_mode_mc_save *save);
-void gmc_v8_0_mc_resume(struct amdgpu_device *adev,
-			struct amdgpu_mode_mc_save *save);
-int gmc_v8_0_mc_wait_for_idle(struct amdgpu_device *adev);
-
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c b/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c
index 460bc8ad37e6..825ccd63f2dc 100644
--- a/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c
@@ -72,6 +72,11 @@ static int iceland_dpm_sw_init(void *handle)
 
 static int iceland_dpm_sw_fini(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	release_firmware(adev->pm.fw);
+	adev->pm.fw = NULL;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
index a789a863d677..5a0e245771ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
@@ -507,19 +507,19 @@ static int kv_enable_didt(struct amdgpu_device *adev, bool enable)
 	    pi->caps_db_ramping ||
 	    pi->caps_td_ramping ||
 	    pi->caps_tcp_ramping) {
-		gfx_v7_0_enter_rlc_safe_mode(adev);
+		adev->gfx.rlc.funcs->enter_safe_mode(adev);
 
 		if (enable) {
 			ret = kv_program_pt_config_registers(adev, didt_config_kv);
 			if (ret) {
-				gfx_v7_0_exit_rlc_safe_mode(adev);
+				adev->gfx.rlc.funcs->exit_safe_mode(adev);
 				return ret;
 			}
 		}
 
 		kv_do_enable_didt(adev, enable);
 
-		gfx_v7_0_exit_rlc_safe_mode(adev);
+		adev->gfx.rlc.funcs->exit_safe_mode(adev);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
index f4c3130d3fdb..ac3730a6e49f 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
@@ -105,6 +105,15 @@ static void sdma_v2_4_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
+static void sdma_v2_4_free_microcode(struct amdgpu_device *adev)
+{
+	int i;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		release_firmware(adev->sdma.instance[i].fw);
+		adev->sdma.instance[i].fw = NULL;
+	}
+}
+
 /**
  * sdma_v2_4_init_microcode - load ucode images from disk
  *
@@ -246,19 +255,6 @@ static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring,
 				   unsigned vm_id, bool ctx_switch)
 {
 	u32 vmid = vm_id & 0xf;
-	u32 next_rptr = ring->wptr + 5;
-
-	while ((next_rptr & 7) != 2)
-		next_rptr++;
-
-	next_rptr += 6;
-
-	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) |
-			  SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR));
-	amdgpu_ring_write(ring, lower_32_bits(ring->next_rptr_gpu_addr) & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr));
-	amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1));
-	amdgpu_ring_write(ring, next_rptr);
 
 	/* IB packet must end on a 8 DW boundary */
 	sdma_v2_4_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8);
@@ -461,6 +457,8 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 		/* Initialize the ring buffer's read and write pointers */
 		WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0);
 		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0);
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
@@ -489,7 +487,11 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
 
 		ring->ready = true;
+	}
 
+	sdma_v2_4_enable(adev, true);
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
 		r = amdgpu_ring_test_ring(ring);
 		if (r) {
 			ring->ready = false;
@@ -580,8 +582,8 @@ static int sdma_v2_4_start(struct amdgpu_device *adev)
 			return -EINVAL;
 	}
 
-	/* unhalt the MEs */
-	sdma_v2_4_enable(adev, true);
+	/* halt the engine before programing */
+	sdma_v2_4_enable(adev, false);
 
 	/* start the gfx rings and rlc compute queues */
 	r = sdma_v2_4_gfx_resume(adev);
@@ -1012,6 +1014,7 @@ static int sdma_v2_4_sw_fini(void *handle)
 	for (i = 0; i < adev->sdma.num_instances; i++)
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
 
+	sdma_v2_4_free_microcode(adev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
index 31d99b0010f7..f00db6f4c04c 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
@@ -236,6 +236,15 @@ static void sdma_v3_0_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
+static void sdma_v3_0_free_microcode(struct amdgpu_device *adev)
+{
+	int i;
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		release_firmware(adev->sdma.instance[i].fw);
+		adev->sdma.instance[i].fw = NULL;
+	}
+}
+
 /**
  * sdma_v3_0_init_microcode - load ucode images from disk
  *
@@ -406,18 +415,6 @@ static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring,
 				   unsigned vm_id, bool ctx_switch)
 {
 	u32 vmid = vm_id & 0xf;
-	u32 next_rptr = ring->wptr + 5;
-
-	while ((next_rptr & 7) != 2)
-		next_rptr++;
-	next_rptr += 6;
-
-	amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) |
-			  SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR));
-	amdgpu_ring_write(ring, lower_32_bits(ring->next_rptr_gpu_addr) & 0xfffffffc);
-	amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr));
-	amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1));
-	amdgpu_ring_write(ring, next_rptr);
 
 	/* IB packet must end on a 8 DW boundary */
 	sdma_v3_0_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8);
@@ -672,6 +669,8 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		/* Initialize the ring buffer's read and write pointers */
 		WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0);
 		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0);
+		WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0);
 
 		/* set the wb address whether it's enabled or not */
 		WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i],
@@ -711,7 +710,15 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl);
 
 		ring->ready = true;
+	}
 
+	/* unhalt the MEs */
+	sdma_v3_0_enable(adev, true);
+	/* enable sdma ring preemption */
+	sdma_v3_0_ctx_switch_enable(adev, true);
+
+	for (i = 0; i < adev->sdma.num_instances; i++) {
+		ring = &adev->sdma.instance[i].ring;
 		r = amdgpu_ring_test_ring(ring);
 		if (r) {
 			ring->ready = false;
@@ -804,10 +811,9 @@ static int sdma_v3_0_start(struct amdgpu_device *adev)
 		}
 	}
 
-	/* unhalt the MEs */
-	sdma_v3_0_enable(adev, true);
-	/* enable sdma ring preemption */
-	sdma_v3_0_ctx_switch_enable(adev, true);
+	/* disble sdma engine before programing it */
+	sdma_v3_0_ctx_switch_enable(adev, false);
+	sdma_v3_0_enable(adev, false);
 
 	/* start the gfx rings and rlc compute queues */
 	r = sdma_v3_0_gfx_resume(adev);
@@ -1247,6 +1253,7 @@ static int sdma_v3_0_sw_fini(void *handle)
 	for (i = 0; i < adev->sdma.num_instances; i++)
 		amdgpu_ring_fini(&adev->sdma.instance[i].ring);
 
+	sdma_v3_0_free_microcode(adev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c b/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c
index b7615cefcac4..f06f6f4dc3a8 100644
--- a/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c
@@ -71,6 +71,11 @@ static int tonga_dpm_sw_init(void *handle)
 
 static int tonga_dpm_sw_fini(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	release_firmware(adev->pm.fw);
+	adev->pm.fw = NULL;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
index f07551476a70..416c8567d3ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
@@ -34,6 +34,8 @@
 #include "oss/oss_2_0_d.h"
 #include "oss/oss_2_0_sh_mask.h"
 
+#include "bif/bif_4_1_d.h"
+
 static void uvd_v4_2_mc_resume(struct amdgpu_device *adev);
 static void uvd_v4_2_init_cg(struct amdgpu_device *adev);
 static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev);
@@ -439,6 +441,32 @@ static void uvd_v4_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq
 }
 
 /**
+ * uvd_v4_2_ring_emit_hdp_flush - emit an hdp flush
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp flush.
+ */
+static void uvd_v4_2_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0));
+	amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * uvd_v4_2_ring_hdp_invalidate - emit an hdp invalidate
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp invalidate.
+ */
+static void uvd_v4_2_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0));
+	amdgpu_ring_write(ring, 1);
+}
+
+/**
  * uvd_v4_2_ring_test_ring - register write test
  *
  * @ring: amdgpu_ring pointer
@@ -763,6 +791,8 @@ static const struct amdgpu_ring_funcs uvd_v4_2_ring_funcs = {
 	.parse_cs = amdgpu_uvd_ring_parse_cs,
 	.emit_ib = uvd_v4_2_ring_emit_ib,
 	.emit_fence = uvd_v4_2_ring_emit_fence,
+	.emit_hdp_flush = uvd_v4_2_ring_emit_hdp_flush,
+	.emit_hdp_invalidate = uvd_v4_2_ring_emit_hdp_invalidate,
 	.test_ring = uvd_v4_2_ring_test_ring,
 	.test_ib = uvd_v4_2_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
index e0a76a883d46..dd636c4c4b08 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
@@ -31,6 +31,7 @@
 #include "uvd/uvd_5_0_sh_mask.h"
 #include "oss/oss_2_0_d.h"
 #include "oss/oss_2_0_sh_mask.h"
+#include "bif/bif_5_0_d.h"
 #include "vi.h"
 
 static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev);
@@ -489,6 +490,32 @@ static void uvd_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq
 }
 
 /**
+ * uvd_v5_0_ring_emit_hdp_flush - emit an hdp flush
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp flush.
+ */
+static void uvd_v5_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0));
+	amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * uvd_v5_0_ring_hdp_invalidate - emit an hdp invalidate
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp invalidate.
+ */
+static void uvd_v5_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0));
+	amdgpu_ring_write(ring, 1);
+}
+
+/**
  * uvd_v5_0_ring_test_ring - register write test
  *
  * @ring: amdgpu_ring pointer
@@ -815,6 +842,8 @@ static const struct amdgpu_ring_funcs uvd_v5_0_ring_funcs = {
 	.parse_cs = amdgpu_uvd_ring_parse_cs,
 	.emit_ib = uvd_v5_0_ring_emit_ib,
 	.emit_fence = uvd_v5_0_ring_emit_fence,
+	.emit_hdp_flush = uvd_v5_0_ring_emit_hdp_flush,
+	.emit_hdp_invalidate = uvd_v5_0_ring_emit_hdp_invalidate,
 	.test_ring = uvd_v5_0_ring_test_ring,
 	.test_ib = uvd_v5_0_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index c9929d665c01..07e9a987fbee 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -33,6 +33,7 @@
 #include "oss/oss_2_0_sh_mask.h"
 #include "smu/smu_7_1_3_d.h"
 #include "smu/smu_7_1_3_sh_mask.h"
+#include "bif/bif_5_1_d.h"
 #include "vi.h"
 
 static void uvd_v6_0_set_ring_funcs(struct amdgpu_device *adev);
@@ -385,8 +386,8 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 	uint32_t mp_swap_cntl;
 	int i, j, r;
 
-	/*disable DPG */
-	WREG32_P(mmUVD_POWER_STATUS, 0, ~(1 << 2));
+	/* disable DPG */
+	WREG32_P(mmUVD_POWER_STATUS, 0, ~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
 
 	/* disable byte swapping */
 	lmi_swap_cntl = 0;
@@ -405,17 +406,21 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 	}
 
 	/* disable interupt */
-	WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1));
+	WREG32_P(mmUVD_MASTINT_EN, 0, ~UVD_MASTINT_EN__VCPU_EN_MASK);
 
 	/* stall UMC and register bus before resetting VCPU */
-	WREG32_P(mmUVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
+	WREG32_P(mmUVD_LMI_CTRL2, UVD_LMI_CTRL2__STALL_ARB_UMC_MASK, ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
 	mdelay(1);
 
 	/* put LMI, VCPU, RBC etc... into reset */
-	WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK | UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__RBC_SOFT_RESET_MASK | UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__CXW_SOFT_RESET_MASK | UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
+	WREG32(mmUVD_SOFT_RESET,
+		UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__RBC_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__CXW_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
 		UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
 	mdelay(5);
 
@@ -424,8 +429,13 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 	mdelay(5);
 
 	/* initialize UVD memory controller */
-	WREG32(mmUVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) |
-			     (1 << 21) | (1 << 9) | (1 << 20));
+	WREG32(mmUVD_LMI_CTRL,
+		(0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+		UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__REQ_MODE_MASK |
+		UVD_LMI_CTRL__DISABLE_ON_FWV_FAIL_MASK);
 
 #ifdef __BIG_ENDIAN
 	/* swap (8 in 32) RB and IB */
@@ -447,10 +457,10 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 	mdelay(5);
 
 	/* enable VCPU clock */
-	WREG32(mmUVD_VCPU_CNTL,  1 << 9);
+	WREG32(mmUVD_VCPU_CNTL, UVD_VCPU_CNTL__CLK_EN_MASK);
 
 	/* enable UMC */
-	WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8));
+	WREG32_P(mmUVD_LMI_CTRL2, 0, ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
 
 	/* boot up the VCPU */
 	WREG32(mmUVD_SOFT_RESET, 0);
@@ -484,10 +494,12 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 		return r;
 	}
 	/* enable master interrupt */
-	WREG32_P(mmUVD_MASTINT_EN, 3 << 1, ~(3 << 1));
+	WREG32_P(mmUVD_MASTINT_EN,
+		(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK),
+		~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK));
 
 	/* clear the bit 4 of UVD_STATUS */
-	WREG32_P(mmUVD_STATUS, 0, ~(2 << 1));
+	WREG32_P(mmUVD_STATUS, 0, ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
 
 	rb_bufsz = order_base_2(ring->ring_size);
 	tmp = 0;
@@ -581,6 +593,32 @@ static void uvd_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq
 }
 
 /**
+ * uvd_v6_0_ring_emit_hdp_flush - emit an hdp flush
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp flush.
+ */
+static void uvd_v6_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0));
+	amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * uvd_v6_0_ring_hdp_invalidate - emit an hdp invalidate
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp invalidate.
+ */
+static void uvd_v6_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0));
+	amdgpu_ring_write(ring, 1);
+}
+
+/**
  * uvd_v6_0_ring_test_ring - register write test
  *
  * @ring: amdgpu_ring pointer
@@ -847,7 +885,8 @@ static int uvd_v6_0_set_clockgating_state(void *handle,
 	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
 	static int curstate = -1;
 
-	if (adev->asic_type == CHIP_FIJI)
+	if (adev->asic_type == CHIP_FIJI ||
+			adev->asic_type == CHIP_POLARIS10)
 		uvd_v6_set_bypass_mode(adev, enable);
 
 	if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG))
@@ -919,6 +958,8 @@ static const struct amdgpu_ring_funcs uvd_v6_0_ring_funcs = {
 	.parse_cs = amdgpu_uvd_ring_parse_cs,
 	.emit_ib = uvd_v6_0_ring_emit_ib,
 	.emit_fence = uvd_v6_0_ring_emit_fence,
+	.emit_hdp_flush = uvd_v6_0_ring_emit_hdp_flush,
+	.emit_hdp_invalidate = uvd_v6_0_ring_emit_hdp_invalidate,
 	.test_ring = uvd_v6_0_ring_test_ring,
 	.test_ib = uvd_v6_0_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index 2c88d0b66cf3..cda7def9dc2c 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -203,6 +203,29 @@ static void vi_didt_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
 	spin_unlock_irqrestore(&adev->didt_idx_lock, flags);
 }
 
+static u32 vi_gc_cac_rreg(struct amdgpu_device *adev, u32 reg)
+{
+	unsigned long flags;
+	u32 r;
+
+	spin_lock_irqsave(&adev->gc_cac_idx_lock, flags);
+	WREG32(mmGC_CAC_IND_INDEX, (reg));
+	r = RREG32(mmGC_CAC_IND_DATA);
+	spin_unlock_irqrestore(&adev->gc_cac_idx_lock, flags);
+	return r;
+}
+
+static void vi_gc_cac_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&adev->gc_cac_idx_lock, flags);
+	WREG32(mmGC_CAC_IND_INDEX, (reg));
+	WREG32(mmGC_CAC_IND_DATA, (v));
+	spin_unlock_irqrestore(&adev->gc_cac_idx_lock, flags);
+}
+
+
 static const u32 tonga_mgcg_cgcg_init[] =
 {
 	mmCGTT_DRM_CLK_CTRL0, 0xffffffff, 0x00600100,
@@ -421,6 +444,20 @@ static bool vi_read_bios_from_rom(struct amdgpu_device *adev,
 	return true;
 }
 
+static u32 vi_get_virtual_caps(struct amdgpu_device *adev)
+{
+	u32 caps = 0;
+	u32 reg = RREG32(mmBIF_IOV_FUNC_IDENTIFIER);
+
+	if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, IOV_ENABLE))
+		caps |= AMDGPU_VIRT_CAPS_SRIOV_EN;
+
+	if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, FUNC_IDENTIFIER))
+		caps |= AMDGPU_VIRT_CAPS_IS_VF;
+
+	return caps;
+}
+
 static const struct amdgpu_allowed_register_entry tonga_allowed_read_registers[] = {
 	{mmGB_MACROTILE_MODE7, true},
 };
@@ -519,12 +556,12 @@ static uint32_t vi_read_indexed_register(struct amdgpu_device *adev, u32 se_num,
 
 	mutex_lock(&adev->grbm_idx_mutex);
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		gfx_v8_0_select_se_sh(adev, se_num, sh_num);
+		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
 
 	val = RREG32(reg_offset);
 
 	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff);
+		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
 	return val;
 }
@@ -583,7 +620,7 @@ static int vi_read_register(struct amdgpu_device *adev, u32 se_num,
 	return -EINVAL;
 }
 
-static void vi_gpu_pci_config_reset(struct amdgpu_device *adev)
+static int vi_gpu_pci_config_reset(struct amdgpu_device *adev)
 {
 	u32 i;
 
@@ -598,11 +635,14 @@ static void vi_gpu_pci_config_reset(struct amdgpu_device *adev)
 
 	/* wait for asic to come out of reset */
 	for (i = 0; i < adev->usec_timeout; i++) {
-		if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff)
-			break;
+		if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) {
+			/* enable BM */
+			pci_set_master(adev->pdev);
+			return 0;
+		}
 		udelay(1);
 	}
-
+	return -EINVAL;
 }
 
 static void vi_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hung)
@@ -628,13 +668,15 @@ static void vi_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hun
  */
 static int vi_asic_reset(struct amdgpu_device *adev)
 {
+	int r;
+
 	vi_set_bios_scratch_engine_hung(adev, true);
 
-	vi_gpu_pci_config_reset(adev);
+	r = vi_gpu_pci_config_reset(adev);
 
 	vi_set_bios_scratch_engine_hung(adev, false);
 
-	return 0;
+	return r;
 }
 
 static int vi_set_uvd_clock(struct amdgpu_device *adev, u32 clock,
@@ -1118,9 +1160,7 @@ static const struct amdgpu_asic_funcs vi_asic_funcs =
 	.get_xclk = &vi_get_xclk,
 	.set_uvd_clocks = &vi_set_uvd_clocks,
 	.set_vce_clocks = &vi_set_vce_clocks,
-	/* these should be moved to their own ip modules */
-	.get_gpu_clock_counter = &gfx_v8_0_get_gpu_clock_counter,
-	.wait_for_mc_idle = &gmc_v8_0_mc_wait_for_idle,
+	.get_virtual_caps = &vi_get_virtual_caps,
 };
 
 static int vi_common_early_init(void *handle)
@@ -1141,6 +1181,8 @@ static int vi_common_early_init(void *handle)
 	adev->uvd_ctx_wreg = &vi_uvd_ctx_wreg;
 	adev->didt_rreg = &vi_didt_rreg;
 	adev->didt_wreg = &vi_didt_wreg;
+	adev->gc_cac_rreg = &vi_gc_cac_rreg;
+	adev->gc_cac_wreg = &vi_gc_cac_wreg;
 
 	adev->asic_funcs = &vi_asic_funcs;
 
@@ -1207,19 +1249,39 @@ static int vi_common_early_init(void *handle)
 			AMD_CG_SUPPORT_HDP_LS |
 			AMD_CG_SUPPORT_SDMA_MGCG |
 			AMD_CG_SUPPORT_SDMA_LS;
+		/* rev0 hardware doesn't support PG */
 		adev->pg_flags = 0;
+		if (adev->rev_id != 0x00)
+			adev->pg_flags |= AMD_PG_SUPPORT_GFX_PG |
+				AMD_PG_SUPPORT_GFX_SMG |
+				AMD_PG_SUPPORT_GFX_DMG |
+				AMD_PG_SUPPORT_CP |
+				AMD_PG_SUPPORT_RLC_SMU_HS |
+				AMD_PG_SUPPORT_GFX_PIPELINE;
 		adev->external_rev_id = adev->rev_id + 0x1;
 		break;
 	case CHIP_STONEY:
 		adev->cg_flags = AMD_CG_SUPPORT_UVD_MGCG |
 			AMD_CG_SUPPORT_GFX_MGCG |
 			AMD_CG_SUPPORT_GFX_MGLS |
+			AMD_CG_SUPPORT_GFX_RLC_LS |
+			AMD_CG_SUPPORT_GFX_CP_LS |
+			AMD_CG_SUPPORT_GFX_CGTS |
+			AMD_CG_SUPPORT_GFX_MGLS |
+			AMD_CG_SUPPORT_GFX_CGTS_LS |
+			AMD_CG_SUPPORT_GFX_CGCG |
+			AMD_CG_SUPPORT_GFX_CGLS |
 			AMD_CG_SUPPORT_BIF_LS |
 			AMD_CG_SUPPORT_HDP_MGCG |
 			AMD_CG_SUPPORT_HDP_LS |
 			AMD_CG_SUPPORT_SDMA_MGCG |
 			AMD_CG_SUPPORT_SDMA_LS;
-		adev->pg_flags = 0;
+		adev->pg_flags = AMD_PG_SUPPORT_GFX_PG |
+			AMD_PG_SUPPORT_GFX_SMG |
+			AMD_PG_SUPPORT_GFX_DMG |
+			AMD_PG_SUPPORT_GFX_PIPELINE |
+			AMD_PG_SUPPORT_CP |
+			AMD_PG_SUPPORT_RLC_SMU_HS;
 		adev->external_rev_id = adev->rev_id + 0x1;
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index ac005796b71c..4f3849ac8c07 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -63,13 +63,12 @@ static struct kfd_process *create_process(const struct task_struct *thread);
 void kfd_process_create_wq(void)
 {
 	if (!kfd_process_wq)
-		kfd_process_wq = create_workqueue("kfd_process_wq");
+		kfd_process_wq = alloc_workqueue("kfd_process_wq", 0, 0);
 }
 
 void kfd_process_destroy_wq(void)
 {
 	if (kfd_process_wq) {
-		flush_workqueue(kfd_process_wq);
 		destroy_workqueue(kfd_process_wq);
 		kfd_process_wq = NULL;
 	}
@@ -242,13 +241,19 @@ static void kfd_process_notifier_release(struct mmu_notifier *mn,
 	pqm_uninit(&p->pqm);
 
 	/* Iterate over all process device data structure and check
-	 * if we should reset all wavefronts */
-	list_for_each_entry(pdd, &p->per_device_data, per_device_list)
+	 * if we should delete debug managers and reset all wavefronts
+	 */
+	list_for_each_entry(pdd, &p->per_device_data, per_device_list) {
+		if ((pdd->dev->dbgmgr) &&
+				(pdd->dev->dbgmgr->pasid == p->pasid))
+			kfd_dbgmgr_destroy(pdd->dev->dbgmgr);
+
 		if (pdd->reset_wavefronts) {
 			pr_warn("amdkfd: Resetting all wave fronts\n");
 			dbgdev_wave_reset_wavefronts(pdd->dev, p);
 			pdd->reset_wavefronts = false;
 		}
+	}
 
 	mutex_unlock(&p->mutex);
 
@@ -324,6 +329,7 @@ err_process_pqm_init:
 	synchronize_rcu();
 	mmu_notifier_unregister_no_release(&process->mmu_notifier, process->mm);
 err_mmu_notifier:
+	mutex_destroy(&process->mutex);
 	kfd_pasid_free(process->pasid);
 err_alloc_pasid:
 	kfree(process->queues);
@@ -404,42 +410,52 @@ void kfd_unbind_process_from_device(struct kfd_dev *dev, unsigned int pasid)
 
 	idx = srcu_read_lock(&kfd_processes_srcu);
 
+	/*
+	 * Look for the process that matches the pasid. If there is no such
+	 * process, we either released it in amdkfd's own notifier, or there
+	 * is a bug. Unfortunately, there is no way to tell...
+	 */
 	hash_for_each_rcu(kfd_processes_table, i, p, kfd_processes)
-		if (p->pasid == pasid)
-			break;
+		if (p->pasid == pasid) {
 
-	srcu_read_unlock(&kfd_processes_srcu, idx);
+			srcu_read_unlock(&kfd_processes_srcu, idx);
 
-	BUG_ON(p->pasid != pasid);
+			pr_debug("Unbinding process %d from IOMMU\n", pasid);
 
-	mutex_lock(&p->mutex);
+			mutex_lock(&p->mutex);
 
-	if ((dev->dbgmgr) && (dev->dbgmgr->pasid == p->pasid))
-		kfd_dbgmgr_destroy(dev->dbgmgr);
+			if ((dev->dbgmgr) && (dev->dbgmgr->pasid == p->pasid))
+				kfd_dbgmgr_destroy(dev->dbgmgr);
 
-	pqm_uninit(&p->pqm);
+			pqm_uninit(&p->pqm);
 
-	pdd = kfd_get_process_device_data(dev, p);
+			pdd = kfd_get_process_device_data(dev, p);
 
-	if (!pdd) {
-		mutex_unlock(&p->mutex);
-		return;
-	}
+			if (!pdd) {
+				mutex_unlock(&p->mutex);
+				return;
+			}
 
-	if (pdd->reset_wavefronts) {
-		dbgdev_wave_reset_wavefronts(pdd->dev, p);
-		pdd->reset_wavefronts = false;
-	}
+			if (pdd->reset_wavefronts) {
+				dbgdev_wave_reset_wavefronts(pdd->dev, p);
+				pdd->reset_wavefronts = false;
+			}
 
-	/*
-	 * Just mark pdd as unbound, because we still need it to call
-	 * amd_iommu_unbind_pasid() in when the process exits.
-	 * We don't call amd_iommu_unbind_pasid() here
-	 * because the IOMMU called us.
-	 */
-	pdd->bound = false;
+			/*
+			 * Just mark pdd as unbound, because we still need it
+			 * to call amd_iommu_unbind_pasid() in when the
+			 * process exits.
+			 * We don't call amd_iommu_unbind_pasid() here
+			 * because the IOMMU called us.
+			 */
+			pdd->bound = false;
 
-	mutex_unlock(&p->mutex);
+			mutex_unlock(&p->mutex);
+
+			return;
+		}
+
+	srcu_read_unlock(&kfd_processes_srcu, idx);
 }
 
 struct kfd_process_device *kfd_get_first_process_device_data(struct kfd_process *p)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index 74909e72a009..884c96f50c3d 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -666,7 +666,7 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 			dev->node_props.simd_count);
 
 	if (dev->mem_bank_count < dev->node_props.mem_banks_count) {
-		pr_warn("kfd: mem_banks_count truncated from %d to %d\n",
+		pr_info_once("kfd: mem_banks_count truncated from %d to %d\n",
 				dev->node_props.mem_banks_count,
 				dev->mem_bank_count);
 		sysfs_show_32bit_prop(buffer, "mem_banks_count",
diff --git a/drivers/gpu/drm/amd/include/amd_pcie.h b/drivers/gpu/drm/amd/include/amd_pcie.h
index 7c2a916c1e63..5eb895fd98bf 100644
--- a/drivers/gpu/drm/amd/include/amd_pcie.h
+++ b/drivers/gpu/drm/amd/include/amd_pcie.h
@@ -37,6 +37,13 @@
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_MASK   0x0000FFFF
 #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_SHIFT  0
 
+/* gen: chipset 1/2, asic 1/2/3 */
+#define AMDGPU_DEFAULT_PCIE_GEN_MASK (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 \
+				      | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 \
+				      | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 \
+				      | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 \
+				      | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3)
+
 /* Following flags shows PCIe lane width switch supported in driver which are decided by chipset and ASIC */
 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X1          0x00010000
 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X2          0x00020000
@@ -47,4 +54,11 @@
 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X32         0x00400000
 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_SHIFT       16
 
+/* 1/2/4/8/16 lanes */
+#define AMDGPU_DEFAULT_PCIE_MLW_MASK (CAIL_PCIE_LINK_WIDTH_SUPPORT_X1 \
+				      | CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 \
+				      | CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 \
+				      | CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 \
+				      | CAIL_PCIE_LINK_WIDTH_SUPPORT_X16)
+
 #endif
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index 6080951d539d..a74a0d2ff1ca 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -26,15 +26,6 @@
 #define AMD_MAX_USEC_TIMEOUT		100000  /* 100 ms */
 
 /*
-* Supported GPU families (aligned with amdgpu_drm.h)
-*/
-#define AMD_FAMILY_UNKNOWN              0
-#define AMD_FAMILY_CI                   120 /* Bonaire, Hawaii */
-#define AMD_FAMILY_KV                   125 /* Kaveri, Kabini, Mullins */
-#define AMD_FAMILY_VI                   130 /* Iceland, Tonga */
-#define AMD_FAMILY_CZ                   135 /* Carrizo */
-
-/*
  * Supported ASIC types
  */
 enum amd_asic_type {
@@ -120,6 +111,8 @@ enum amd_powergating_state {
 #define AMD_PG_SUPPORT_SDMA			(1 << 8)
 #define AMD_PG_SUPPORT_ACP			(1 << 9)
 #define AMD_PG_SUPPORT_SAMU			(1 << 10)
+#define AMD_PG_SUPPORT_GFX_QUICK_MG		(1 << 11)
+#define AMD_PG_SUPPORT_GFX_PIPELINE		(1 << 12)
 
 enum amd_pm_state_type {
 	/* not used for dpm */
@@ -157,6 +150,7 @@ struct amd_ip_funcs {
 	int (*hw_init)(void *handle);
 	/* tears down the hw state */
 	int (*hw_fini)(void *handle);
+	void (*late_fini)(void *handle);
 	/* handles IP specific hw/sw changes for suspend */
 	int (*suspend)(void *handle);
 	/* handles IP specific hw/sw changes for resume */
diff --git a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h
index ebaf67bb1589..90ff7c8a6011 100644
--- a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h
+++ b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h
@@ -2823,4 +2823,7 @@
 #define mmDC_EDC_CSINVOC_CNT                                                    0x3192
 #define mmDC_EDC_RESTORE_CNT                                                    0x3193
 
+#define mmGC_CAC_IND_INDEX                                                      0x129a
+#define mmGC_CAC_IND_DATA                                                       0x129b
+
 #endif /* GFX_8_0_D_H */
diff --git a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h
index 7d722458d9f5..4070ca3a68eb 100644
--- a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h
+++ b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h
@@ -8730,8 +8730,6 @@
 #define RLC_GPM_STAT__DYN_CU_POWERING_DOWN__SHIFT 0x10
 #define RLC_GPM_STAT__ABORTED_PD_SEQUENCE_MASK 0x20000
 #define RLC_GPM_STAT__ABORTED_PD_SEQUENCE__SHIFT 0x11
-#define RLC_GPM_STAT__RESERVED_MASK 0xfc0000
-#define RLC_GPM_STAT__RESERVED__SHIFT 0x12
 #define RLC_GPM_STAT__PG_ERROR_STATUS_MASK 0xff000000
 #define RLC_GPM_STAT__PG_ERROR_STATUS__SHIFT 0x18
 #define RLC_GPU_CLOCK_32_RES_SEL__RES_SEL_MASK 0x3f
@@ -8764,8 +8762,10 @@
 #define RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE__SHIFT 0x12
 #define RLC_PG_CNTL__SMU_HANDSHAKE_ENABLE_MASK 0x80000
 #define RLC_PG_CNTL__SMU_HANDSHAKE_ENABLE__SHIFT 0x13
-#define RLC_PG_CNTL__RESERVED1_MASK 0xf00000
-#define RLC_PG_CNTL__RESERVED1__SHIFT 0x14
+#define RLC_PG_CNTL__QUICK_PG_ENABLE_MASK 0x100000
+#define RLC_PG_CNTL__QUICK_PG_ENABLE__SHIFT 0x14
+#define RLC_PG_CNTL__RESERVED1_MASK 0xe00000
+#define RLC_PG_CNTL__RESERVED1__SHIFT 0x15
 #define RLC_GPM_THREAD_PRIORITY__THREAD0_PRIORITY_MASK 0xff
 #define RLC_GPM_THREAD_PRIORITY__THREAD0_PRIORITY__SHIFT 0x0
 #define RLC_GPM_THREAD_PRIORITY__THREAD1_PRIORITY_MASK 0xff00
@@ -9102,8 +9102,6 @@
 #define RLC_GPM_LOG_CONT__CONT__SHIFT 0x0
 #define RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK 0xff
 #define RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG__SHIFT 0x0
-#define RLC_PG_DELAY_3__RESERVED_MASK 0xffffff00
-#define RLC_PG_DELAY_3__RESERVED__SHIFT 0x8
 #define RLC_GPM_INT_DISABLE_TH0__DISABLE_MASK 0xffffffff
 #define RLC_GPM_INT_DISABLE_TH0__DISABLE__SHIFT 0x0
 #define RLC_GPM_INT_DISABLE_TH1__DISABLE_MASK 0xffffffff
@@ -9124,14 +9122,8 @@
 #define RLC_SRM_DEBUG_SELECT__RESERVED__SHIFT 0x8
 #define RLC_SRM_DEBUG__DATA_MASK 0xffffffff
 #define RLC_SRM_DEBUG__DATA__SHIFT 0x0
-#define RLC_SRM_ARAM_ADDR__ADDR_MASK 0x3ff
-#define RLC_SRM_ARAM_ADDR__ADDR__SHIFT 0x0
-#define RLC_SRM_ARAM_ADDR__RESERVED_MASK 0xfffffc00
-#define RLC_SRM_ARAM_ADDR__RESERVED__SHIFT 0xa
 #define RLC_SRM_ARAM_DATA__DATA_MASK 0xffffffff
 #define RLC_SRM_ARAM_DATA__DATA__SHIFT 0x0
-#define RLC_SRM_DRAM_ADDR__ADDR_MASK 0x3ff
-#define RLC_SRM_DRAM_ADDR__ADDR__SHIFT 0x0
 #define RLC_SRM_DRAM_ADDR__RESERVED_MASK 0xfffffc00
 #define RLC_SRM_DRAM_ADDR__RESERVED__SHIFT 0xa
 #define RLC_SRM_DRAM_DATA__DATA_MASK 0xffffffff
@@ -17946,8 +17938,6 @@
 #define VGT_TESS_DISTRIBUTION__ACCUM_TRI__SHIFT 0x8
 #define VGT_TESS_DISTRIBUTION__ACCUM_QUAD_MASK 0xff0000
 #define VGT_TESS_DISTRIBUTION__ACCUM_QUAD__SHIFT 0x10
-#define VGT_TESS_DISTRIBUTION__DONUT_SPLIT_MASK 0xff000000
-#define VGT_TESS_DISTRIBUTION__DONUT_SPLIT__SHIFT 0x18
 #define VGT_TF_RING_SIZE__SIZE_MASK 0xffff
 #define VGT_TF_RING_SIZE__SIZE__SHIFT 0x0
 #define VGT_SYS_CONFIG__DUAL_CORE_EN_MASK 0x1
@@ -20502,8 +20492,6 @@
 #define DIDT_SQ_CTRL0__DIDT_CTRL_RST__SHIFT 0x4
 #define DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20
 #define DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5
-#define DIDT_SQ_CTRL0__UNUSED_0_MASK 0xffffffc0
-#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT 0x6
 #define DIDT_SQ_CTRL1__MIN_POWER_MASK 0xffff
 #define DIDT_SQ_CTRL1__MIN_POWER__SHIFT 0x0
 #define DIDT_SQ_CTRL1__MAX_POWER_MASK 0xffff0000
@@ -20558,8 +20546,6 @@
 #define DIDT_DB_CTRL0__DIDT_CTRL_RST__SHIFT 0x4
 #define DIDT_DB_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20
 #define DIDT_DB_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5
-#define DIDT_DB_CTRL0__UNUSED_0_MASK 0xffffffc0
-#define DIDT_DB_CTRL0__UNUSED_0__SHIFT 0x6
 #define DIDT_DB_CTRL1__MIN_POWER_MASK 0xffff
 #define DIDT_DB_CTRL1__MIN_POWER__SHIFT 0x0
 #define DIDT_DB_CTRL1__MAX_POWER_MASK 0xffff0000
@@ -20614,8 +20600,6 @@
 #define DIDT_TD_CTRL0__DIDT_CTRL_RST__SHIFT 0x4
 #define DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20
 #define DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5
-#define DIDT_TD_CTRL0__UNUSED_0_MASK 0xffffffc0
-#define DIDT_TD_CTRL0__UNUSED_0__SHIFT 0x6
 #define DIDT_TD_CTRL1__MIN_POWER_MASK 0xffff
 #define DIDT_TD_CTRL1__MIN_POWER__SHIFT 0x0
 #define DIDT_TD_CTRL1__MAX_POWER_MASK 0xffff0000
@@ -20670,8 +20654,6 @@
 #define DIDT_TCP_CTRL0__DIDT_CTRL_RST__SHIFT 0x4
 #define DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20
 #define DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5
-#define DIDT_TCP_CTRL0__UNUSED_0_MASK 0xffffffc0
-#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT 0x6
 #define DIDT_TCP_CTRL1__MIN_POWER_MASK 0xffff
 #define DIDT_TCP_CTRL1__MIN_POWER__SHIFT 0x0
 #define DIDT_TCP_CTRL1__MAX_POWER_MASK 0xffff0000
@@ -20726,8 +20708,6 @@
 #define DIDT_DBR_CTRL0__DIDT_CTRL_RST__SHIFT 0x4
 #define DIDT_DBR_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20
 #define DIDT_DBR_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5
-#define DIDT_DBR_CTRL0__UNUSED_0_MASK 0xffffffc0
-#define DIDT_DBR_CTRL0__UNUSED_0__SHIFT 0x6
 #define DIDT_DBR_CTRL1__MIN_POWER_MASK 0xffff
 #define DIDT_DBR_CTRL1__MIN_POWER__SHIFT 0x0
 #define DIDT_DBR_CTRL1__MAX_POWER_MASK 0xffff0000
@@ -20773,4 +20753,84 @@
 #define DIDT_DBR_WEIGHT8_11__WEIGHT11_MASK 0xff000000
 #define DIDT_DBR_WEIGHT8_11__WEIGHT11__SHIFT 0x18
 
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK    0x00000001
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT  0x00000000
+
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK       0x0000007e
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK       0x00001f80L
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT     0x00000001
+#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT     0x00000007
+
+#define DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK   0x1fffe000L
+#define DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d
+
+#define DIDT_SQ_STALL_CTRL__UNUSED_0_MASK                  0xe0000000L
+#define DIDT_SQ_STALL_CTRL__UNUSED_0__SHIFT                0x0000001d
+
+#define DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK       0x00000001L
+#define DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT     0x00000000
+
+#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK       0x00007ffeL
+#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT     0x00000001
+#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK       0x1fff8000L
+#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT     0x0000000f
+
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK    0x00000001L
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT  0x00000000
+
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK       0x0000007eL
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK       0x00001f80L
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT     0x00000001
+#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT     0x00000007
+
+#define DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK   0x1fffe000L
+#define DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d
+
+#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK     0x00000fc0L
+#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK     0x0003f000L
+#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT   0x00000006
+#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT   0x0000000c
+
+#define DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK       0x00000001L
+#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK       0x00007ffeL
+#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK       0x1fff8000L
+
+#define DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT     0x00000000
+#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT     0x00000001
+#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT     0x0000000f
+
+#define DIDT_TD_STALL_CTRL__UNUSED_0_MASK                  0xe0000000L
+#define DIDT_TD_STALL_CTRL__UNUSED_0__SHIFT                0x0000001d
+
+#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK     0x00000fc0L
+#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK     0x0003f000L
+#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT   0x00000006
+#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT   0x0000000c
+
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK   0x00000001L
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT 0x00000000
+
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK      0x0000007eL
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK      0x00001f80L
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT    0x00000001
+#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT    0x00000007
+
+#define DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK  0x1fffe000L
+#define DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d
+
+#define DIDT_TCP_STALL_CTRL__UNUSED_0_MASK                 0xe0000000L
+#define DIDT_TCP_STALL_CTRL__UNUSED_0__SHIFT               0x0000001d
+
+#define DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK      0x00000001L
+#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK      0x00007ffeL
+#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK      0x1fff8000L
+#define DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT    0x00000000
+#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT    0x00000001
+#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT    0x0000000f
+
+#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK    0x00000fc0L
+#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK    0x0003f000L
+#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT  0x00000006
+#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT  0x0000000c
+
 #endif /* GFX_8_0_SH_MASK_H */
diff --git a/drivers/gpu/drm/amd/include/atombios.h b/drivers/gpu/drm/amd/include/atombios.h
index 32f3e345de08..3493da5c8f0e 100644
--- a/drivers/gpu/drm/amd/include/atombios.h
+++ b/drivers/gpu/drm/amd/include/atombios.h
@@ -5538,6 +5538,78 @@ typedef struct  _ATOM_ASIC_PROFILING_INFO_V3_5
   ULONG  ulReserved[12];
 }ATOM_ASIC_PROFILING_INFO_V3_5;
 
+/* for Polars10/11 AVFS parameters */
+typedef struct  _ATOM_ASIC_PROFILING_INFO_V3_6
+{
+  ATOM_COMMON_TABLE_HEADER         asHeader;
+  ULONG  ulMaxVddc;
+  ULONG  ulMinVddc;
+  USHORT usLkgEuseIndex;
+  UCHAR  ucLkgEfuseBitLSB;
+  UCHAR  ucLkgEfuseLength;
+  ULONG  ulLkgEncodeLn_MaxDivMin;
+  ULONG  ulLkgEncodeMax;
+  ULONG  ulLkgEncodeMin;
+  EFUSE_LINEAR_FUNC_PARAM sRoFuse;
+  ULONG  ulEvvDefaultVddc;
+  ULONG  ulEvvNoCalcVddc;
+  ULONG  ulSpeed_Model;
+  ULONG  ulSM_A0;
+  ULONG  ulSM_A1;
+  ULONG  ulSM_A2;
+  ULONG  ulSM_A3;
+  ULONG  ulSM_A4;
+  ULONG  ulSM_A5;
+  ULONG  ulSM_A6;
+  ULONG  ulSM_A7;
+  UCHAR  ucSM_A0_sign;
+  UCHAR  ucSM_A1_sign;
+  UCHAR  ucSM_A2_sign;
+  UCHAR  ucSM_A3_sign;
+  UCHAR  ucSM_A4_sign;
+  UCHAR  ucSM_A5_sign;
+  UCHAR  ucSM_A6_sign;
+  UCHAR  ucSM_A7_sign;
+  ULONG  ulMargin_RO_a;
+  ULONG  ulMargin_RO_b;
+  ULONG  ulMargin_RO_c;
+  ULONG  ulMargin_fixed;
+  ULONG  ulMargin_Fmax_mean;
+  ULONG  ulMargin_plat_mean;
+  ULONG  ulMargin_Fmax_sigma;
+  ULONG  ulMargin_plat_sigma;
+  ULONG  ulMargin_DC_sigma;
+  ULONG  ulLoadLineSlop;
+  ULONG  ulaTDClimitPerDPM[8];
+  ULONG  ulaNoCalcVddcPerDPM[8];
+  ULONG  ulAVFS_meanNsigma_Acontant0;
+  ULONG  ulAVFS_meanNsigma_Acontant1;
+  ULONG  ulAVFS_meanNsigma_Acontant2;
+  USHORT usAVFS_meanNsigma_DC_tol_sigma;
+  USHORT usAVFS_meanNsigma_Platform_mean;
+  USHORT usAVFS_meanNsigma_Platform_sigma;
+  ULONG  ulGB_VDROOP_TABLE_CKSOFF_a0;
+  ULONG  ulGB_VDROOP_TABLE_CKSOFF_a1;
+  ULONG  ulGB_VDROOP_TABLE_CKSOFF_a2;
+  ULONG  ulGB_VDROOP_TABLE_CKSON_a0;
+  ULONG  ulGB_VDROOP_TABLE_CKSON_a1;
+  ULONG  ulGB_VDROOP_TABLE_CKSON_a2;
+  ULONG  ulAVFSGB_FUSE_TABLE_CKSOFF_m1;
+  USHORT usAVFSGB_FUSE_TABLE_CKSOFF_m2;
+  ULONG  ulAVFSGB_FUSE_TABLE_CKSOFF_b;
+  ULONG  ulAVFSGB_FUSE_TABLE_CKSON_m1;
+  USHORT usAVFSGB_FUSE_TABLE_CKSON_m2;
+  ULONG  ulAVFSGB_FUSE_TABLE_CKSON_b;
+  USHORT usMaxVoltage_0_25mv;
+  UCHAR  ucEnableGB_VDROOP_TABLE_CKSOFF;
+  UCHAR  ucEnableGB_VDROOP_TABLE_CKSON;
+  UCHAR  ucEnableGB_FUSE_TABLE_CKSOFF;
+  UCHAR  ucEnableGB_FUSE_TABLE_CKSON;
+  USHORT usPSM_Age_ComFactor;
+  UCHAR  ucEnableApplyAVFS_CKS_OFF_Voltage;
+  UCHAR  ucReserved;
+}ATOM_ASIC_PROFILING_INFO_V3_6;
+
 
 typedef struct _ATOM_SCLK_FCW_RANGE_ENTRY_V1{
   ULONG  ulMaxSclkFreq;
diff --git a/drivers/gpu/drm/amd/include/cgs_common.h b/drivers/gpu/drm/amd/include/cgs_common.h
index a461e155a160..0c8c85d2a2a5 100644
--- a/drivers/gpu/drm/amd/include/cgs_common.h
+++ b/drivers/gpu/drm/amd/include/cgs_common.h
@@ -49,6 +49,7 @@ enum cgs_ind_reg {
 	CGS_IND_REG__SMC,
 	CGS_IND_REG__UVD_CTX,
 	CGS_IND_REG__DIDT,
+	CGS_IND_REG_GC_CAC,
 	CGS_IND_REG__AUDIO_ENDPT
 };
 
@@ -115,6 +116,7 @@ enum cgs_system_info_id {
 	CGS_SYSTEM_INFO_CG_FLAGS,
 	CGS_SYSTEM_INFO_PG_FLAGS,
 	CGS_SYSTEM_INFO_GFX_CU_INFO,
+	CGS_SYSTEM_INFO_GFX_SE_INFO,
 	CGS_SYSTEM_INFO_ID_MAXIMUM,
 };
 
@@ -189,7 +191,6 @@ typedef unsigned long cgs_handle_t;
 
 struct cgs_acpi_method_argument {
 	uint32_t type;
-	uint32_t method_length;
 	uint32_t data_length;
 	union{
 		uint32_t value;
@@ -581,6 +582,9 @@ typedef int (*cgs_get_firmware_info)(struct cgs_device *cgs_device,
 				     enum cgs_ucode_id type,
 				     struct cgs_firmware_info *info);
 
+typedef int (*cgs_rel_firmware)(struct cgs_device *cgs_device,
+					 enum cgs_ucode_id type);
+
 typedef int(*cgs_set_powergating_state)(struct cgs_device *cgs_device,
 				  enum amd_ip_block_type block_type,
 				  enum amd_powergating_state state);
@@ -645,6 +649,7 @@ struct cgs_ops {
 	cgs_set_camera_voltages_t set_camera_voltages;
 	/* Firmware Info */
 	cgs_get_firmware_info get_firmware_info;
+	cgs_rel_firmware rel_firmware;
 	/* cg pg interface*/
 	cgs_set_powergating_state set_powergating_state;
 	cgs_set_clockgating_state set_clockgating_state;
@@ -738,6 +743,8 @@ struct cgs_device
 	CGS_CALL(set_camera_voltages,dev,mask,voltages)
 #define cgs_get_firmware_info(dev, type, info)	\
 	CGS_CALL(get_firmware_info, dev, type, info)
+#define cgs_rel_firmware(dev, type)	\
+	CGS_CALL(rel_firmware, dev, type)
 #define cgs_set_powergating_state(dev, block_type, state)	\
 	CGS_CALL(set_powergating_state, dev, block_type, state)
 #define cgs_set_clockgating_state(dev, block_type, state)	\
diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
index 8e345bfddb69..f9e03ad0baa2 100644
--- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
+++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
@@ -73,11 +73,14 @@ static int pp_sw_init(void *handle)
 
 	ret = hwmgr->hwmgr_func->backend_init(hwmgr);
 	if (ret)
-		goto err;
+		goto err1;
 
 	pr_info("amdgpu: powerplay initialized\n");
 
 	return 0;
+err1:
+	if (hwmgr->pptable_func->pptable_fini)
+		hwmgr->pptable_func->pptable_fini(hwmgr);
 err:
 	pr_err("amdgpu: powerplay initialization failed\n");
 	return ret;
@@ -100,6 +103,9 @@ static int pp_sw_fini(void *handle)
 	if (hwmgr->hwmgr_func->backend_fini != NULL)
 		ret = hwmgr->hwmgr_func->backend_fini(hwmgr);
 
+	if (hwmgr->pptable_func->pptable_fini)
+		hwmgr->pptable_func->pptable_fini(hwmgr);
+
 	return ret;
 }
 
@@ -530,6 +536,10 @@ int pp_dpm_dispatch_tasks(void *handle, enum amd_pp_event event_id, void *input,
 	case AMD_PP_EVENT_COMPLETE_INIT:
 		ret = pem_handle_event(pp_handle->eventmgr, event_id, &data);
 		break;
+	case AMD_PP_EVENT_READJUST_POWER_STATE:
+		pp_handle->hwmgr->current_ps = pp_handle->hwmgr->boot_ps;
+		ret = pem_handle_event(pp_handle->eventmgr, event_id, &data);
+		break;
 	default:
 		break;
 	}
@@ -734,12 +744,12 @@ static int pp_dpm_get_pp_table(void *handle, char **table)
 
 	PP_CHECK_HW(hwmgr);
 
-	if (hwmgr->hwmgr_func->get_pp_table == NULL) {
-		printk(KERN_INFO "%s was not implemented.\n", __func__);
-		return 0;
-	}
+	if (!hwmgr->soft_pp_table)
+		return -EINVAL;
+
+	*table = (char *)hwmgr->soft_pp_table;
 
-	return hwmgr->hwmgr_func->get_pp_table(hwmgr, table);
+	return hwmgr->soft_pp_table_size;
 }
 
 static int pp_dpm_set_pp_table(void *handle, const char *buf, size_t size)
@@ -753,12 +763,23 @@ static int pp_dpm_set_pp_table(void *handle, const char *buf, size_t size)
 
 	PP_CHECK_HW(hwmgr);
 
-	if (hwmgr->hwmgr_func->set_pp_table == NULL) {
-		printk(KERN_INFO "%s was not implemented.\n", __func__);
-		return 0;
+	if (!hwmgr->hardcode_pp_table) {
+		hwmgr->hardcode_pp_table =
+				kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL);
+
+		if (!hwmgr->hardcode_pp_table)
+			return -ENOMEM;
+
+		/* to avoid powerplay crash when hardcode pptable is empty */
+		memcpy(hwmgr->hardcode_pp_table, hwmgr->soft_pp_table,
+				hwmgr->soft_pp_table_size);
 	}
 
-	return hwmgr->hwmgr_func->set_pp_table(hwmgr, buf, size);
+	memcpy(hwmgr->hardcode_pp_table, buf, size);
+
+	hwmgr->soft_pp_table = hwmgr->hardcode_pp_table;
+
+	return amd_powerplay_reset(handle);
 }
 
 static int pp_dpm_force_clock_level(void *handle,
@@ -800,6 +821,82 @@ static int pp_dpm_print_clock_levels(void *handle,
 	return hwmgr->hwmgr_func->print_clock_levels(hwmgr, type, buf);
 }
 
+static int pp_dpm_get_sclk_od(void *handle)
+{
+	struct pp_hwmgr *hwmgr;
+
+	if (!handle)
+		return -EINVAL;
+
+	hwmgr = ((struct pp_instance *)handle)->hwmgr;
+
+	PP_CHECK_HW(hwmgr);
+
+	if (hwmgr->hwmgr_func->get_sclk_od == NULL) {
+		printk(KERN_INFO "%s was not implemented.\n", __func__);
+		return 0;
+	}
+
+	return hwmgr->hwmgr_func->get_sclk_od(hwmgr);
+}
+
+static int pp_dpm_set_sclk_od(void *handle, uint32_t value)
+{
+	struct pp_hwmgr *hwmgr;
+
+	if (!handle)
+		return -EINVAL;
+
+	hwmgr = ((struct pp_instance *)handle)->hwmgr;
+
+	PP_CHECK_HW(hwmgr);
+
+	if (hwmgr->hwmgr_func->set_sclk_od == NULL) {
+		printk(KERN_INFO "%s was not implemented.\n", __func__);
+		return 0;
+	}
+
+	return hwmgr->hwmgr_func->set_sclk_od(hwmgr, value);
+}
+
+static int pp_dpm_get_mclk_od(void *handle)
+{
+	struct pp_hwmgr *hwmgr;
+
+	if (!handle)
+		return -EINVAL;
+
+	hwmgr = ((struct pp_instance *)handle)->hwmgr;
+
+	PP_CHECK_HW(hwmgr);
+
+	if (hwmgr->hwmgr_func->get_mclk_od == NULL) {
+		printk(KERN_INFO "%s was not implemented.\n", __func__);
+		return 0;
+	}
+
+	return hwmgr->hwmgr_func->get_mclk_od(hwmgr);
+}
+
+static int pp_dpm_set_mclk_od(void *handle, uint32_t value)
+{
+	struct pp_hwmgr *hwmgr;
+
+	if (!handle)
+		return -EINVAL;
+
+	hwmgr = ((struct pp_instance *)handle)->hwmgr;
+
+	PP_CHECK_HW(hwmgr);
+
+	if (hwmgr->hwmgr_func->set_mclk_od == NULL) {
+		printk(KERN_INFO "%s was not implemented.\n", __func__);
+		return 0;
+	}
+
+	return hwmgr->hwmgr_func->set_mclk_od(hwmgr, value);
+}
+
 const struct amd_powerplay_funcs pp_dpm_funcs = {
 	.get_temperature = pp_dpm_get_temperature,
 	.load_firmware = pp_dpm_load_fw,
@@ -822,6 +919,10 @@ const struct amd_powerplay_funcs pp_dpm_funcs = {
 	.set_pp_table = pp_dpm_set_pp_table,
 	.force_clock_level = pp_dpm_force_clock_level,
 	.print_clock_levels = pp_dpm_print_clock_levels,
+	.get_sclk_od = pp_dpm_get_sclk_od,
+	.set_sclk_od = pp_dpm_set_sclk_od,
+	.get_mclk_od = pp_dpm_get_mclk_od,
+	.set_mclk_od = pp_dpm_set_mclk_od,
 };
 
 static int amd_pp_instance_init(struct amd_pp_init *pp_init,
@@ -903,6 +1004,44 @@ int amd_powerplay_fini(void *handle)
 	return 0;
 }
 
+int amd_powerplay_reset(void *handle)
+{
+	struct pp_instance *instance = (struct pp_instance *)handle;
+	struct pp_eventmgr *eventmgr;
+	struct pem_event_data event_data = { {0} };
+	int ret;
+
+	if (instance == NULL)
+		return -EINVAL;
+
+	eventmgr = instance->eventmgr;
+	if (!eventmgr || !eventmgr->pp_eventmgr_fini)
+		return -EINVAL;
+
+	eventmgr->pp_eventmgr_fini(eventmgr);
+
+	ret = pp_sw_fini(handle);
+	if (ret)
+		return ret;
+
+	kfree(instance->hwmgr->ps);
+
+	ret = pp_sw_init(handle);
+	if (ret)
+		return ret;
+
+	hw_init_power_state_table(instance->hwmgr);
+
+	if (eventmgr == NULL || eventmgr->pp_eventmgr_init == NULL)
+		return -EINVAL;
+
+	ret = eventmgr->pp_eventmgr_init(eventmgr);
+	if (ret)
+		return ret;
+
+	return pem_handle_event(eventmgr, AMD_PP_EVENT_COMPLETE_INIT, &event_data);
+}
+
 /* export this function to DAL */
 
 int amd_powerplay_display_configuration_change(void *handle,
diff --git a/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c b/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c
index 46410e3c7349..fb88e4e5d625 100644
--- a/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c
@@ -58,9 +58,6 @@ static void pem_fini(struct pp_eventmgr *eventmgr)
 	pem_unregister_interrupts(eventmgr);
 
 	pem_handle_event(eventmgr, AMD_PP_EVENT_UNINITIALIZE, &event_data);
-
-	if (eventmgr != NULL)
-		kfree(eventmgr);
 }
 
 int eventmgr_init(struct pp_instance *handle)
diff --git a/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c b/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c
index 5cd123472db4..b6f45fd01fa6 100644
--- a/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c
+++ b/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c
@@ -132,8 +132,7 @@ int pem_task_enable_dynamic_state_management(struct pp_eventmgr *eventmgr, struc
 
 int pem_task_disable_dynamic_state_management(struct pp_eventmgr *eventmgr, struct pem_event_data *event_data)
 {
-	/* TODO */
-	return 0;
+	return phm_disable_dynamic_state_management(eventmgr->hwmgr);
 }
 
 int pem_task_enable_clock_power_gatings_tasks(struct pp_eventmgr *eventmgr, struct pem_event_data *event_data)
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
index 436fc16dabb6..2da548f6337e 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
@@ -206,7 +206,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate)
 							AMD_IP_BLOCK_TYPE_VCE,
 							AMD_PG_STATE_GATE);
 				cz_enable_disable_vce_dpm(hwmgr, false);
-			/* TODO: to figure out why vce can't be poweroff*/
+				cz_dpm_powerdown_vce(hwmgr);
 				cz_hwmgr->vce_power_gated = true;
 			} else {
 				cz_dpm_powerup_vce(hwmgr);
@@ -225,6 +225,7 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate)
 			}
 		}
 	} else {
+		cz_hwmgr->vce_power_gated = bgate;
 		cz_dpm_update_vce_dpm(hwmgr);
 		cz_enable_disable_vce_dpm(hwmgr, !bgate);
 		return 0;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
index 1f14c477d15d..9bf622e123b6 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
@@ -1180,6 +1180,13 @@ static int cz_apply_state_adjust_rules(struct pp_hwmgr *hwmgr,
 static int cz_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 {
 	int result = 0;
+	struct cz_hwmgr *data;
+
+	data = kzalloc(sizeof(struct cz_hwmgr), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+
+	hwmgr->backend = data;
 
 	result = cz_initialize_dpm_defaults(hwmgr);
 	if (result != 0) {
@@ -1909,15 +1916,7 @@ static const struct pp_hwmgr_func cz_hwmgr_funcs = {
 
 int cz_hwmgr_init(struct pp_hwmgr *hwmgr)
 {
-	struct cz_hwmgr *cz_hwmgr;
-	int ret = 0;
-
-	cz_hwmgr = kzalloc(sizeof(struct cz_hwmgr), GFP_KERNEL);
-	if (cz_hwmgr == NULL)
-		return -ENOMEM;
-
-	hwmgr->backend = cz_hwmgr;
 	hwmgr->hwmgr_func = &cz_hwmgr_funcs;
 	hwmgr->pptable_func = &pptable_funcs;
-	return ret;
+	return 0;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c
index 24a16e49b571..744aa886a2be 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c
@@ -581,25 +581,24 @@ static int fiji_patch_boot_state(struct pp_hwmgr *hwmgr,
 
 static int fiji_hwmgr_backend_fini(struct pp_hwmgr *hwmgr)
 {
-	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
-
-	if (data->soft_pp_table) {
-		kfree(data->soft_pp_table);
-		data->soft_pp_table = NULL;
-	}
-
 	return phm_hwmgr_backend_fini(hwmgr);
 }
 
 static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 {
-	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_hwmgr *data;
 	uint32_t i;
 	struct phm_ppt_v1_information *table_info =
 			(struct phm_ppt_v1_information *)(hwmgr->pptable);
 	bool stay_in_boot;
 	int result;
 
+	data = kzalloc(sizeof(struct fiji_hwmgr), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+
+	hwmgr->backend = data;
+
 	data->dll_default_on = false;
 	data->sram_end = SMC_RAM_END;
 
@@ -633,6 +632,8 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 	data->vddci_control = FIJI_VOLTAGE_CONTROL_NONE;
 	data->mvdd_control = FIJI_VOLTAGE_CONTROL_NONE;
 
+	data->force_pcie_gen = PP_PCIEGenInvalid;
+
 	if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr,
 			VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2))
 		data->voltage_control = FIJI_VOLTAGE_CONTROL_BY_SVID2;
@@ -732,7 +733,7 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_gen_cap = 0x30007;
+			data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK;
 		else
 			data->pcie_gen_cap = (uint32_t)sys_info.value;
 		if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
@@ -741,7 +742,7 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_lane_cap = 0x2f0000;
+			data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK;
 		else
 			data->pcie_lane_cap = (uint32_t)sys_info.value;
 	} else {
@@ -1234,6 +1235,34 @@ static int fiji_program_voting_clients(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int fiji_clear_voting_clients(struct pp_hwmgr *hwmgr)
+{
+	/* Reset voting clients before disabling DPM */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			SCLK_PWRMGT_CNTL, RESET_SCLK_CNT, 1);
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			SCLK_PWRMGT_CNTL, RESET_BUSY_CNT, 1);
+
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_0, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_1, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_2, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_3, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_4, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_5, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_6, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_7, 0);
+
+	return 0;
+}
+
 /**
 * Get the location of various tables inside the FW image.
 *
@@ -1361,6 +1390,17 @@ static int fiji_copy_and_switch_arb_sets(struct pp_hwmgr *hwmgr,
 }
 
 /**
+* Call SMC to reset S0/S1 to S1 and Reset SMIO to initial value
+*
+* @param    hwmgr  the address of the powerplay hardware manager.
+* @return   if success then 0;
+*/
+static int fiji_reset_to_default(struct pp_hwmgr *hwmgr)
+{
+	return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_ResetToDefaults);
+}
+
+/**
 * Initial switch from ARB F0->F1
 *
 * @param    hwmgr  the address of the powerplay hardware manager.
@@ -1373,6 +1413,21 @@ static int fiji_initial_switch_from_arbf0_to_f1(struct pp_hwmgr *hwmgr)
 			MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
 }
 
+static int fiji_force_switch_to_arbf0(struct pp_hwmgr *hwmgr)
+{
+	uint32_t tmp;
+
+	tmp = (cgs_read_ind_register(hwmgr->device,
+			CGS_IND_REG__SMC, ixSMC_SCRATCH9) &
+			0x0000ff00) >> 8;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return fiji_copy_and_switch_arb_sets(hwmgr,
+			tmp, MC_CG_ARB_FREQ_F0);
+}
+
 static int fiji_reset_single_dpm_table(struct pp_hwmgr *hwmgr,
 		struct fiji_single_dpm_table *dpm_table, uint32_t count)
 {
@@ -1830,7 +1885,7 @@ static uint16_t fiji_find_closest_vddci(struct pp_hwmgr *hwmgr, uint16_t vddci)
 
 	PP_ASSERT_WITH_CODE(false,
 			"VDDCI is larger than max VDDCI in VDDCI Voltage Table!",
-			return vddci_table->entries[i].value);
+			return vddci_table->entries[i-1].value);
 }
 
 static int fiji_get_dependency_volt_by_clk(struct pp_hwmgr *hwmgr,
@@ -3175,6 +3230,17 @@ static int fiji_enable_ulv(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int fiji_disable_ulv(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_ulv_parm *ulv = &(data->ulv);
+
+	if (ulv->ulv_supported)
+		return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_DisableULV);
+
+	return 0;
+}
+
 static int fiji_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
 {
 	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
@@ -3195,6 +3261,21 @@ static int fiji_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int fiji_disable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
+{
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_SclkDeepSleep)) {
+		if (smum_send_msg_to_smc(hwmgr->smumgr,
+				PPSMC_MSG_MASTER_DeepSleep_OFF)) {
+			PP_ASSERT_WITH_CODE(false,
+					"Attempt to disable Master Deep Sleep switch failed!",
+					return -1);
+		}
+	}
+
+	return 0;
+}
+
 static int fiji_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
 {
 	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
@@ -3355,6 +3436,70 @@ static int fiji_start_dpm(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int fiji_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+
+	/* disable SCLK dpm */
+	if (!data->sclk_dpm_key_disabled)
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_DPM_Disable) == 0),
+				"Failed to disable SCLK DPM!",
+				return -1);
+
+	/* disable MCLK dpm */
+	if (!data->mclk_dpm_key_disabled) {
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc_with_parameter(hwmgr->smumgr,
+				PPSMC_MSG_MCLKDPM_SetEnabledMask, 1) == 0),
+				"Failed to force MCLK DPM0!",
+				return -1);
+
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_MCLKDPM_Disable) == 0),
+				"Failed to disable MCLK DPM!",
+				return -1);
+	}
+
+	return 0;
+}
+
+static int fiji_stop_dpm(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+
+	/* disable general power management */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, GENERAL_PWRMGT,
+			GLOBAL_PWRMGT_EN, 0);
+	/* disable sclk deep sleep */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, SCLK_PWRMGT_CNTL,
+			DYNAMIC_PM_EN, 0);
+
+	/* disable PCIE dpm */
+	if (!data->pcie_dpm_key_disabled) {
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_PCIeDPM_Disable) == 0),
+				"Failed to disable pcie DPM during DPM Stop Function!",
+				return -1);
+	}
+
+	if (fiji_disable_sclk_mclk_dpm(hwmgr)) {
+		printk(KERN_ERR "Failed to disable Sclk DPM and Mclk DPM!");
+		return -1;
+	}
+
+	PP_ASSERT_WITH_CODE(
+			(smum_send_msg_to_smc(hwmgr->smumgr,
+					PPSMC_MSG_Voltage_Cntl_Disable) == 0),
+			"Failed to disable voltage DPM during DPM Stop Function!",
+			return -1);
+
+	return 0;
+}
+
 static void fiji_set_dpm_event_sources(struct pp_hwmgr *hwmgr,
 		uint32_t sources)
 {
@@ -3413,6 +3558,23 @@ static int fiji_enable_thermal_auto_throttle(struct pp_hwmgr *hwmgr)
 	return fiji_enable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal);
 }
 
+static int fiji_disable_auto_throttle_source(struct pp_hwmgr *hwmgr,
+		PHM_AutoThrottleSource source)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+
+	if (data->active_auto_throttle_sources & (1 << source)) {
+		data->active_auto_throttle_sources &= ~(1 << source);
+		fiji_set_dpm_event_sources(hwmgr, data->active_auto_throttle_sources);
+	}
+	return 0;
+}
+
+static int fiji_disable_thermal_auto_throttle(struct pp_hwmgr *hwmgr)
+{
+	return fiji_disable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal);
+}
+
 static int fiji_enable_dpm_tasks(struct pp_hwmgr *hwmgr)
 {
 	int tmp_result, result = 0;
@@ -3527,6 +3689,64 @@ static int fiji_enable_dpm_tasks(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
+static int fiji_disable_dpm_tasks(struct pp_hwmgr *hwmgr)
+{
+	int tmp_result, result = 0;
+
+	tmp_result = (fiji_is_dpm_running(hwmgr)) ? 0 : -1;
+	PP_ASSERT_WITH_CODE(tmp_result == 0,
+			"DPM is not running right now, no need to disable DPM!",
+			return 0);
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_ThermalController))
+		PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+				GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, 1);
+
+	tmp_result = fiji_disable_power_containment(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable power containment!", result = tmp_result);
+
+	tmp_result = fiji_disable_smc_cac(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable SMC CAC!", result = tmp_result);
+
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			CG_SPLL_SPREAD_SPECTRUM, SSEN, 0);
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, 0);
+
+	tmp_result = fiji_disable_thermal_auto_throttle(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable thermal auto throttle!", result = tmp_result);
+
+	tmp_result = fiji_stop_dpm(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to stop DPM!", result = tmp_result);
+
+	tmp_result = fiji_disable_deep_sleep_master_switch(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable deep sleep master switch!", result = tmp_result);
+
+	tmp_result = fiji_disable_ulv(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable ULV!", result = tmp_result);
+
+	tmp_result = fiji_clear_voting_clients(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to clear voting clients!", result = tmp_result);
+
+	tmp_result = fiji_reset_to_default(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to reset to default!", result = tmp_result);
+
+	tmp_result = fiji_force_switch_to_arbf0(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to force to switch arbf0!", result = tmp_result);
+
+	return result;
+}
+
 static int fiji_force_dpm_highest(struct pp_hwmgr *hwmgr)
 {
 	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
@@ -5069,42 +5289,6 @@ static int fiji_get_fan_control_mode(struct pp_hwmgr *hwmgr)
 				CG_FDO_CTRL2, FDO_PWM_MODE);
 }
 
-static int fiji_get_pp_table(struct pp_hwmgr *hwmgr, char **table)
-{
-	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kmemdup(hwmgr->soft_pp_table,
-					      hwmgr->soft_pp_table_size,
-					      GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	*table = (char *)&data->soft_pp_table;
-
-	return hwmgr->soft_pp_table_size;
-}
-
-static int fiji_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size)
-{
-	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	memcpy(data->soft_pp_table, buf, size);
-
-	hwmgr->soft_pp_table = data->soft_pp_table;
-
-	/* TODO: re-init powerplay to implement modified pptable */
-
-	return 0;
-}
-
 static int fiji_force_clock_level(struct pp_hwmgr *hwmgr,
 		enum pp_clock_type type, uint32_t mask)
 {
@@ -5274,12 +5458,96 @@ bool fiji_check_smc_update_required_for_display_configuration(struct pp_hwmgr *h
 	return is_update_required;
 }
 
+static int fiji_get_sclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table);
+	struct fiji_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	int value;
+
+	value = (sclk_table->dpm_levels[sclk_table->count - 1].value -
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) *
+			100 /
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return value;
+}
+
+static int fiji_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	struct pp_power_state  *ps;
+	struct fiji_power_state  *fiji_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	fiji_ps = cast_phw_fiji_power_state(&ps->hardware);
+
+	fiji_ps->performance_levels[fiji_ps->performance_level_count - 1].engine_clock =
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value *
+			value / 100 +
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return 0;
+}
+
+static int fiji_get_mclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table);
+	struct fiji_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	int value;
+
+	value = (mclk_table->dpm_levels[mclk_table->count - 1].value -
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) *
+			100 /
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return value;
+}
+
+static int fiji_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	struct fiji_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	struct pp_power_state  *ps;
+	struct fiji_power_state  *fiji_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	fiji_ps = cast_phw_fiji_power_state(&ps->hardware);
+
+	fiji_ps->performance_levels[fiji_ps->performance_level_count - 1].memory_clock =
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value *
+			value / 100 +
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return 0;
+}
 
 static const struct pp_hwmgr_func fiji_hwmgr_funcs = {
 	.backend_init = &fiji_hwmgr_backend_init,
 	.backend_fini = &fiji_hwmgr_backend_fini,
 	.asic_setup = &fiji_setup_asic_task,
 	.dynamic_state_management_enable = &fiji_enable_dpm_tasks,
+	.dynamic_state_management_disable = &fiji_disable_dpm_tasks,
 	.force_dpm_level = &fiji_dpm_force_dpm_level,
 	.get_num_of_pp_table_entries = &tonga_get_number_of_powerplay_table_entries,
 	.get_power_state_size = &fiji_get_power_state_size,
@@ -5312,24 +5580,18 @@ static const struct pp_hwmgr_func fiji_hwmgr_funcs = {
 	.get_fan_control_mode = fiji_get_fan_control_mode,
 	.check_states_equal = fiji_check_states_equal,
 	.check_smc_update_required_for_display_configuration = fiji_check_smc_update_required_for_display_configuration,
-	.get_pp_table = fiji_get_pp_table,
-	.set_pp_table = fiji_set_pp_table,
 	.force_clock_level = fiji_force_clock_level,
 	.print_clock_levels = fiji_print_clock_levels,
+	.get_sclk_od = fiji_get_sclk_od,
+	.set_sclk_od = fiji_set_sclk_od,
+	.get_mclk_od = fiji_get_mclk_od,
+	.set_mclk_od = fiji_set_mclk_od,
 };
 
 int fiji_hwmgr_init(struct pp_hwmgr *hwmgr)
 {
-	struct fiji_hwmgr  *data;
-	int ret = 0;
-
-	data = kzalloc(sizeof(struct fiji_hwmgr), GFP_KERNEL);
-	if (data == NULL)
-		return -ENOMEM;
-
-	hwmgr->backend = data;
 	hwmgr->hwmgr_func = &fiji_hwmgr_funcs;
 	hwmgr->pptable_func = &tonga_pptable_funcs;
 	pp_fiji_thermal_initialize(hwmgr);
-	return ret;
+	return 0;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h
index 170edf5a772d..bf67c2a92c68 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h
@@ -302,9 +302,6 @@ struct fiji_hwmgr {
 	bool                           pg_acp_init;
 	bool                           frtc_enabled;
 	bool                           frtc_status_changed;
-
-	/* soft pptable for re-uploading into smu */
-	void *soft_pp_table;
 };
 
 /* To convert to Q8.8 format for firmware */
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c
index db23a4068baf..44658451a8d2 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c
@@ -73,17 +73,18 @@ void fiji_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr)
 
 	if (!tmp) {
 		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
-				PHM_PlatformCaps_PowerContainment);
-
-		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 				PHM_PlatformCaps_CAC);
 
 		fiji_hwmgr->fast_watermark_threshold = 100;
 
-		tmp = 1;
-		fiji_hwmgr->enable_dte_feature = tmp ? false : true;
-		fiji_hwmgr->enable_tdc_limit_feature = tmp ? true : false;
-		fiji_hwmgr->enable_pkg_pwr_tracking_feature = tmp ? true : false;
+		if (hwmgr->powercontainment_enabled) {
+			phm_cap_set(hwmgr->platform_descriptor.platformCaps,
+				    PHM_PlatformCaps_PowerContainment);
+			tmp = 1;
+			fiji_hwmgr->enable_dte_feature = tmp ? false : true;
+			fiji_hwmgr->enable_tdc_limit_feature = tmp ? true : false;
+			fiji_hwmgr->enable_pkg_pwr_tracking_feature = tmp ? true : false;
+		}
 	}
 }
 
@@ -459,6 +460,23 @@ int fiji_enable_smc_cac(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
+int fiji_disable_smc_cac(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	int result = 0;
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_CAC) && data->cac_enabled) {
+		int smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+				(uint16_t)(PPSMC_MSG_DisableCac));
+		PP_ASSERT_WITH_CODE((smc_result == 0),
+				"Failed to disable CAC in SMC.", result = -1);
+
+		data->cac_enabled = false;
+	}
+	return result;
+}
+
 int fiji_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n)
 {
 	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
@@ -528,6 +546,48 @@ int fiji_enable_power_containment(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
+int fiji_disable_power_containment(struct pp_hwmgr *hwmgr)
+{
+	struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend);
+	int result = 0;
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_PowerContainment) &&
+			data->power_containment_features) {
+		int smc_result;
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_TDCLimit) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_TDCLimitDisable));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable TDCLimit in SMC.",
+					result = smc_result);
+		}
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_DTE) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_DisableDTE));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable DTE in SMC.",
+					result = smc_result);
+		}
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_PkgPwrLimit) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_PkgPwrLimitDisable));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable PkgPwrTracking in SMC.",
+					result = smc_result);
+		}
+		data->power_containment_features = 0;
+	}
+
+	return result;
+}
+
 int fiji_power_control_set_level(struct pp_hwmgr *hwmgr)
 {
 	struct phm_ppt_v1_information *table_info =
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h
index 55e58200f33a..fec772421733 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h
@@ -36,6 +36,19 @@ enum fiji_pt_config_reg_type {
 #define POWERCONTAINMENT_FEATURE_TDCLimit        0x00000002
 #define POWERCONTAINMENT_FEATURE_PkgPwrLimit     0x00000004
 
+#define DIDT_SQ_CTRL0__UNUSED_0_MASK             0xffffffc0
+#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT           0x6
+#define DIDT_TD_CTRL0__UNUSED_0_MASK             0xffffffc0
+#define DIDT_TD_CTRL0__UNUSED_0__SHIFT           0x6
+#define DIDT_TCP_CTRL0__UNUSED_0_MASK            0xffffffc0
+#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT          0x6
+#define DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK                 0xe0000000
+#define DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT               0x0000001d
+#define DIDT_TD_TUNING_CTRL__UNUSED_0_MASK                 0xe0000000
+#define DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT               0x0000001d
+#define DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK                0xe0000000
+#define DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT              0x0000001d
+
 struct fiji_pt_config_reg {
 	uint32_t                           offset;
 	uint32_t                           mask;
@@ -58,7 +71,9 @@ void fiji_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr);
 int fiji_populate_bapm_parameters_in_dpm_table(struct pp_hwmgr *hwmgr);
 int fiji_populate_pm_fuses(struct pp_hwmgr *hwmgr);
 int fiji_enable_smc_cac(struct pp_hwmgr *hwmgr);
+int fiji_disable_smc_cac(struct pp_hwmgr *hwmgr);
 int fiji_enable_power_containment(struct pp_hwmgr *hwmgr);
+int fiji_disable_power_containment(struct pp_hwmgr *hwmgr);
 int fiji_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n);
 int fiji_power_control_set_level(struct pp_hwmgr *hwmgr);
 
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c
index fa208ada6892..789f98ad2615 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c
@@ -154,6 +154,30 @@ int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr)
 	return ret;
 }
 
+int phm_disable_dynamic_state_management(struct pp_hwmgr *hwmgr)
+{
+	int ret = -1;
+	bool enabled;
+
+	PHM_FUNC_CHECK(hwmgr);
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+		PHM_PlatformCaps_TablelessHardwareInterface)) {
+		if (hwmgr->hwmgr_func->dynamic_state_management_disable)
+			ret = hwmgr->hwmgr_func->dynamic_state_management_disable(hwmgr);
+	} else {
+		ret = phm_dispatch_table(hwmgr,
+				&(hwmgr->disable_dynamic_state_management),
+				NULL, NULL);
+	}
+
+	enabled = ret == 0 ? false : true;
+
+	cgs_notify_dpm_enabled(hwmgr->device, enabled);
+
+	return ret;
+}
+
 int phm_force_dpm_levels(struct pp_hwmgr *hwmgr, enum amd_dpm_forced_level level)
 {
 	PHM_FUNC_CHECK(hwmgr);
@@ -306,11 +330,15 @@ int phm_store_dal_configuration_data(struct pp_hwmgr *hwmgr,
 {
 	PHM_FUNC_CHECK(hwmgr);
 
-	if (hwmgr->hwmgr_func->store_cc6_data == NULL)
+	if (display_config == NULL)
 		return -EINVAL;
 
 	hwmgr->display_config = *display_config;
-	/* to do pass other display configuration in furture */
+
+	if (hwmgr->hwmgr_func->store_cc6_data == NULL)
+		return -EINVAL;
+
+	/* TODO: pass other display configuration in the future */
 
 	if (hwmgr->hwmgr_func->store_cc6_data)
 		hwmgr->hwmgr_func->store_cc6_data(hwmgr,
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c
index 1c48917da3cf..03b6128ebc20 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c
@@ -24,6 +24,7 @@
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/slab.h>
+#include <drm/amdgpu_drm.h>
 #include "cgs_common.h"
 #include "power_state.h"
 #include "hwmgr.h"
@@ -58,12 +59,13 @@ int hwmgr_init(struct amd_pp_init *pp_init, struct pp_instance *handle)
 	hwmgr->hw_revision = pp_init->rev_id;
 	hwmgr->usec_timeout = AMD_MAX_USEC_TIMEOUT;
 	hwmgr->power_source = PP_PowerSource_AC;
+	hwmgr->powercontainment_enabled = pp_init->powercontainment_enabled;
 
 	switch (hwmgr->chip_family) {
-	case AMD_FAMILY_CZ:
+	case AMDGPU_FAMILY_CZ:
 		cz_hwmgr_init(hwmgr);
 		break;
-	case AMD_FAMILY_VI:
+	case AMDGPU_FAMILY_VI:
 		switch (hwmgr->chip_id) {
 		case CHIP_TONGA:
 			tonga_hwmgr_init(hwmgr);
@@ -93,6 +95,15 @@ int hwmgr_fini(struct pp_hwmgr *hwmgr)
 	if (hwmgr == NULL || hwmgr->ps == NULL)
 		return -EINVAL;
 
+	/* do hwmgr finish*/
+	kfree(hwmgr->hardcode_pp_table);
+
+	kfree(hwmgr->backend);
+
+	kfree(hwmgr->start_thermal_controller.function_list);
+
+	kfree(hwmgr->set_temperature_range.function_list);
+
 	kfree(hwmgr->ps);
 	kfree(hwmgr);
 	return 0;
@@ -462,7 +473,7 @@ uint16_t phm_find_closest_vddci(struct pp_atomctrl_voltage_table *vddci_table, u
 
 	PP_ASSERT_WITH_CODE(false,
 			"VDDCI is larger than max VDDCI in VDDCI Voltage Table!",
-			return vddci_table->entries[i].value);
+			return vddci_table->entries[i-1].value);
 }
 
 int phm_find_boot_level(void *table,
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h
index 347fef127ce9..2930a3355948 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h
@@ -39,6 +39,7 @@ struct phm_ppt_v1_clock_voltage_dependency_record {
 	uint8_t phases;
 	uint8_t cks_enable;
 	uint8_t cks_voffset;
+	uint32_t sclk_offset;
 };
 
 typedef struct phm_ppt_v1_clock_voltage_dependency_record phm_ppt_v1_clock_voltage_dependency_record;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c
index 8f142a74ad08..aeec25c66aa8 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c
@@ -106,11 +106,17 @@ int polaris10_phm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate)
 	data->uvd_power_gated = bgate;
 
 	if (bgate) {
+		cgs_set_clockgating_state(hwmgr->device,
+				AMD_IP_BLOCK_TYPE_UVD,
+				AMD_CG_STATE_GATE);
 		polaris10_update_uvd_dpm(hwmgr, true);
 		polaris10_phm_powerdown_uvd(hwmgr);
 	} else {
 		polaris10_phm_powerup_uvd(hwmgr);
 		polaris10_update_uvd_dpm(hwmgr, false);
+		cgs_set_clockgating_state(hwmgr->device,
+				AMD_IP_BLOCK_TYPE_UVD,
+				AMD_PG_STATE_UNGATE);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c
index aa6be033f21b..9d764c4d253e 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c
@@ -389,6 +389,34 @@ static int polaris10_program_voting_clients(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int polaris10_clear_voting_clients(struct pp_hwmgr *hwmgr)
+{
+	/* Reset voting clients before disabling DPM */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			SCLK_PWRMGT_CNTL, RESET_SCLK_CNT, 1);
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			SCLK_PWRMGT_CNTL, RESET_BUSY_CNT, 1);
+
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_0, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_1, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_2, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_3, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_4, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_5, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_6, 0);
+	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+			ixCG_FREQ_TRAN_VOTING_7, 0);
+
+	return 0;
+}
+
 /**
 * Get the location of various tables inside the FW image.
 *
@@ -515,6 +543,11 @@ static int polaris10_copy_and_switch_arb_sets(struct pp_hwmgr *hwmgr,
 	return 0;
 }
 
+static int polaris10_reset_to_default(struct pp_hwmgr *hwmgr)
+{
+	return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_ResetToDefaults);
+}
+
 /**
 * Initial switch from ARB F0->F1
 *
@@ -528,6 +561,21 @@ static int polaris10_initial_switch_from_arbf0_to_f1(struct pp_hwmgr *hwmgr)
 			MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
 }
 
+static int polaris10_force_switch_to_arbf0(struct pp_hwmgr *hwmgr)
+{
+	uint32_t tmp;
+
+	tmp = (cgs_read_ind_register(hwmgr->device,
+			CGS_IND_REG__SMC, ixSMC_SCRATCH9) &
+			0x0000ff00) >> 8;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return polaris10_copy_and_switch_arb_sets(hwmgr,
+			tmp, MC_CG_ARB_FREQ_F0);
+}
+
 static int polaris10_setup_default_pcie_table(struct pp_hwmgr *hwmgr)
 {
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
@@ -999,7 +1047,7 @@ static int polaris10_get_dependency_volt_by_clk(struct pp_hwmgr *hwmgr,
 				vddci = phm_find_closest_vddci(&(data->vddci_voltage_table),
 						(dep_table->entries[i].vddc -
 								(uint16_t)data->vddc_vddci_delta));
-				*voltage |= (vddci * VOLTAGE_SCALE) <<	VDDCI_SHIFT;
+				*voltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT;
 			}
 
 			if (POLARIS10_VOLTAGE_CONTROL_NONE == data->mvdd_control)
@@ -1296,7 +1344,6 @@ static int polaris10_populate_single_memory_level(struct pp_hwmgr *hwmgr,
 	}
 
 	mem_level->MclkFrequency = clock;
-	mem_level->StutterEnable = 0;
 	mem_level->EnabledForThrottle = 1;
 	mem_level->EnabledForActivity = 0;
 	mem_level->UpHyst = 0;
@@ -1304,7 +1351,6 @@ static int polaris10_populate_single_memory_level(struct pp_hwmgr *hwmgr,
 	mem_level->VoltageDownHyst = 0;
 	mem_level->ActivityLevel = (uint16_t)data->mclk_activity_target;
 	mem_level->StutterEnable = false;
-
 	mem_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW;
 
 	data->display_timing.num_existing_displays = info.display_count;
@@ -1358,12 +1404,12 @@ static int polaris10_populate_all_memory_levels(struct pp_hwmgr *hwmgr)
 			return result;
 	}
 
-	/* in order to prevent MC activity from stutter mode to push DPM up.
+	/* In order to prevent MC activity from stutter mode to push DPM up,
 	 * the UVD change complements this by putting the MCLK in
-	 * a higher state by default such that we are not effected by
+	 * a higher state by default such that we are not affected by
 	 * up threshold or and MCLK DPM latency.
 	 */
-	levels[0].ActivityLevel = (uint16_t)data->mclk_dpm0_activity_target;
+	levels[0].ActivityLevel = 0x1f;
 	CONVERT_FROM_HOST_TO_SMC_US(levels[0].ActivityLevel);
 
 	data->smc_state_table.MemoryDpmLevelCount =
@@ -1761,12 +1807,9 @@ static int polaris10_populate_smc_initailial_state(struct pp_hwmgr *hwmgr)
 
 static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr)
 {
-	uint32_t ro, efuse, efuse2, clock_freq, volt_without_cks,
-			volt_with_cks, value;
-	uint16_t clock_freq_u16;
+	uint32_t ro, efuse, volt_without_cks, volt_with_cks, value, max, min;
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
-	uint8_t type, i, j, cks_setting, stretch_amount, stretch_amount2,
-			volt_offset = 0;
+	uint8_t i, stretch_amount, stretch_amount2, volt_offset = 0;
 	struct phm_ppt_v1_information *table_info =
 			(struct phm_ppt_v1_information *)(hwmgr->pptable);
 	struct phm_ppt_v1_clock_voltage_dependency_table *sclk_table =
@@ -1778,50 +1821,38 @@ static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr)
 	 * if the part is SS or FF. if RO >= 1660MHz, part is FF.
 	 */
 	efuse = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC,
-			ixSMU_EFUSE_0 + (146 * 4));
-	efuse2 = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC,
-			ixSMU_EFUSE_0 + (148 * 4));
+			ixSMU_EFUSE_0 + (67 * 4));
 	efuse &= 0xFF000000;
 	efuse = efuse >> 24;
-	efuse2 &= 0xF;
-
-	if (efuse2 == 1)
-		ro = (2300 - 1350) * efuse / 255 + 1350;
-	else
-		ro = (2500 - 1000) * efuse / 255 + 1000;
 
-	if (ro >= 1660)
-		type = 0;
-	else
-		type = 1;
+	if (hwmgr->chip_id == CHIP_POLARIS10) {
+		min = 1000;
+		max = 2300;
+	} else {
+		min = 1100;
+		max = 2100;
+	}
 
-	/* Populate Stretch amount */
-	data->smc_state_table.ClockStretcherAmount = stretch_amount;
+	ro = efuse * (max -min)/255 + min;
 
 	/* Populate Sclk_CKS_masterEn0_7 and Sclk_voltageOffset */
 	for (i = 0; i < sclk_table->count; i++) {
 		data->smc_state_table.Sclk_CKS_masterEn0_7 |=
 				sclk_table->entries[i].cks_enable << i;
-		volt_without_cks = (uint32_t)((14041 *
-			(sclk_table->entries[i].clk/100) / 10000 + 3571 + 75 - ro) * 1000 /
-			(4026 - (13924 * (sclk_table->entries[i].clk/100) / 10000)));
-		volt_with_cks = (uint32_t)((13946 *
-			(sclk_table->entries[i].clk/100) / 10000 + 3320 + 45 - ro) * 1000 /
-			(3664 - (11454 * (sclk_table->entries[i].clk/100) / 10000)));
+
+		volt_without_cks =  (uint32_t)(((ro - 40) * 1000 - 2753594 - sclk_table->entries[i].clk/100 * 136418 /1000) / \
+					(sclk_table->entries[i].clk/100 * 1132925 /10000 - 242418)/100);
+
+		volt_with_cks = (uint32_t)((ro * 1000 -2396351 - sclk_table->entries[i].clk/100 * 329021/1000) / \
+				(sclk_table->entries[i].clk/10000 * 649434 /1000  - 18005)/10);
+
 		if (volt_without_cks >= volt_with_cks)
 			volt_offset = (uint8_t)(((volt_without_cks - volt_with_cks +
 					sclk_table->entries[i].cks_voffset) * 100 / 625) + 1);
+
 		data->smc_state_table.Sclk_voltageOffset[i] = volt_offset;
 	}
 
-	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE,
-			STRETCH_ENABLE, 0x0);
-	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE,
-			masterReset, 0x1);
-	/* PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE, staticEnable, 0x1); */
-	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE,
-			masterReset, 0x0);
-
 	/* Populate CKS Lookup Table */
 	if (stretch_amount == 1 || stretch_amount == 2 || stretch_amount == 5)
 		stretch_amount2 = 0;
@@ -1835,69 +1866,6 @@ static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr)
 				return -EINVAL);
 	}
 
-	value = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC,
-			ixPWR_CKS_CNTL);
-	value &= 0xFFC2FF87;
-	data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].minFreq =
-			polaris10_clock_stretcher_lookup_table[stretch_amount2][0];
-	data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].maxFreq =
-			polaris10_clock_stretcher_lookup_table[stretch_amount2][1];
-	clock_freq_u16 = (uint16_t)(PP_SMC_TO_HOST_UL(data->smc_state_table.
-			GraphicsLevel[data->smc_state_table.GraphicsDpmLevelCount - 1].SclkSetting.SclkFrequency) / 100);
-	if (polaris10_clock_stretcher_lookup_table[stretch_amount2][0] < clock_freq_u16
-	&& polaris10_clock_stretcher_lookup_table[stretch_amount2][1] > clock_freq_u16) {
-		/* Program PWR_CKS_CNTL. CKS_USE_FOR_LOW_FREQ */
-		value |= (polaris10_clock_stretcher_lookup_table[stretch_amount2][3]) << 16;
-		/* Program PWR_CKS_CNTL. CKS_LDO_REFSEL */
-		value |= (polaris10_clock_stretcher_lookup_table[stretch_amount2][2]) << 18;
-		/* Program PWR_CKS_CNTL. CKS_STRETCH_AMOUNT */
-		value |= (polaris10_clock_stretch_amount_conversion
-				[polaris10_clock_stretcher_lookup_table[stretch_amount2][3]]
-				 [stretch_amount]) << 3;
-	}
-	CONVERT_FROM_HOST_TO_SMC_US(data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].minFreq);
-	CONVERT_FROM_HOST_TO_SMC_US(data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].maxFreq);
-	data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].setting =
-			polaris10_clock_stretcher_lookup_table[stretch_amount2][2] & 0x7F;
-	data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].setting |=
-			(polaris10_clock_stretcher_lookup_table[stretch_amount2][3]) << 7;
-
-	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
-			ixPWR_CKS_CNTL, value);
-
-	/* Populate DDT Lookup Table */
-	for (i = 0; i < 4; i++) {
-		/* Assign the minimum and maximum VID stored
-		 * in the last row of Clock Stretcher Voltage Table.
-		 */
-		data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].minVID =
-				(uint8_t) polaris10_clock_stretcher_ddt_table[type][i][2];
-		data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].maxVID =
-				(uint8_t) polaris10_clock_stretcher_ddt_table[type][i][3];
-		/* Loop through each SCLK and check the frequency
-		 * to see if it lies within the frequency for clock stretcher.
-		 */
-		for (j = 0; j < data->smc_state_table.GraphicsDpmLevelCount; j++) {
-			cks_setting = 0;
-			clock_freq = PP_SMC_TO_HOST_UL(
-					data->smc_state_table.GraphicsLevel[j].SclkSetting.SclkFrequency);
-			/* Check the allowed frequency against the sclk level[j].
-			 *  Sclk's endianness has already been converted,
-			 *  and it's in 10Khz unit,
-			 *  as opposed to Data table, which is in Mhz unit.
-			 */
-			if (clock_freq >= (polaris10_clock_stretcher_ddt_table[type][i][0]) * 100) {
-				cks_setting |= 0x2;
-				if (clock_freq < (polaris10_clock_stretcher_ddt_table[type][i][1]) * 100)
-					cks_setting |= 0x1;
-			}
-			data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].setting
-							|= cks_setting << (j * 2);
-		}
-		CONVERT_FROM_HOST_TO_SMC_US(
-			data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].setting);
-	}
-
 	value = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixPWR_CKS_CNTL);
 	value &= 0xFFFFFFFE;
 	cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixPWR_CKS_CNTL, value);
@@ -1945,9 +1913,8 @@ static int polaris10_populate_vr_config(struct pp_hwmgr *hwmgr,
 	if (POLARIS10_VOLTAGE_CONTROL_BY_SVID2 == data->mvdd_control) {
 		config = VR_SVI2_PLANE_2;
 		table->VRConfig |= (config << VRCONF_MVDD_SHIFT);
-	} else if (POLARIS10_VOLTAGE_CONTROL_BY_GPIO == data->mvdd_control) {
-		config = VR_SMIO_PATTERN_2;
-		table->VRConfig |= (config << VRCONF_MVDD_SHIFT);
+		cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, data->soft_regs_start +
+			offsetof(SMU74_SoftRegisters, AllowMvddSwitch), 0x1);
 	} else {
 		config = VR_STATIC_VOLTAGE;
 		table->VRConfig |= (config << VRCONF_MVDD_SHIFT);
@@ -1956,6 +1923,90 @@ static int polaris10_populate_vr_config(struct pp_hwmgr *hwmgr,
 	return 0;
 }
 
+
+int polaris10_populate_avfs_parameters(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	SMU74_Discrete_DpmTable  *table = &(data->smc_state_table);
+	int result = 0;
+	struct pp_atom_ctrl__avfs_parameters avfs_params = {0};
+	AVFS_meanNsigma_t AVFS_meanNsigma = { {0} };
+	AVFS_Sclk_Offset_t AVFS_SclkOffset = { {0} };
+	uint32_t tmp, i;
+	struct pp_smumgr *smumgr = hwmgr->smumgr;
+	struct polaris10_smumgr *smu_data = (struct polaris10_smumgr *)(smumgr->backend);
+
+	struct phm_ppt_v1_information *table_info =
+			(struct phm_ppt_v1_information *)hwmgr->pptable;
+	struct phm_ppt_v1_clock_voltage_dependency_table *sclk_table =
+			table_info->vdd_dep_on_sclk;
+
+
+	if (smu_data->avfs.avfs_btc_status == AVFS_BTC_NOTSUPPORTED)
+		return result;
+
+	result = atomctrl_get_avfs_information(hwmgr, &avfs_params);
+
+	if (0 == result) {
+		table->BTCGB_VDROOP_TABLE[0].a0  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a0);
+		table->BTCGB_VDROOP_TABLE[0].a1  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a1);
+		table->BTCGB_VDROOP_TABLE[0].a2  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a2);
+		table->BTCGB_VDROOP_TABLE[1].a0  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0);
+		table->BTCGB_VDROOP_TABLE[1].a1  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1);
+		table->BTCGB_VDROOP_TABLE[1].a2  = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2);
+		table->AVFSGB_VDROOP_TABLE[0].m1 = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSON_m1);
+		table->AVFSGB_VDROOP_TABLE[0].m2 = PP_HOST_TO_SMC_US(avfs_params.usAVFSGB_FUSE_TABLE_CKSON_m2);
+		table->AVFSGB_VDROOP_TABLE[0].b  = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSON_b);
+		table->AVFSGB_VDROOP_TABLE[0].m1_shift = 24;
+		table->AVFSGB_VDROOP_TABLE[0].m2_shift  = 12;
+		table->AVFSGB_VDROOP_TABLE[1].m1 = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1);
+		table->AVFSGB_VDROOP_TABLE[1].m2 = PP_HOST_TO_SMC_US(avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2);
+		table->AVFSGB_VDROOP_TABLE[1].b  = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b);
+		table->AVFSGB_VDROOP_TABLE[1].m1_shift = 24;
+		table->AVFSGB_VDROOP_TABLE[1].m2_shift  = 12;
+		table->MaxVoltage                = PP_HOST_TO_SMC_US(avfs_params.usMaxVoltage_0_25mv);
+		AVFS_meanNsigma.Aconstant[0]      = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant0);
+		AVFS_meanNsigma.Aconstant[1]      = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant1);
+		AVFS_meanNsigma.Aconstant[2]      = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant2);
+		AVFS_meanNsigma.DC_tol_sigma      = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_DC_tol_sigma);
+		AVFS_meanNsigma.Platform_mean     = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_Platform_mean);
+		AVFS_meanNsigma.PSM_Age_CompFactor = PP_HOST_TO_SMC_US(avfs_params.usPSM_Age_ComFactor);
+		AVFS_meanNsigma.Platform_sigma     = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_Platform_sigma);
+
+		for (i = 0; i < NUM_VFT_COLUMNS; i++) {
+			AVFS_meanNsigma.Static_Voltage_Offset[i] = (uint8_t)(sclk_table->entries[i].cks_voffset * 100 / 625);
+			AVFS_SclkOffset.Sclk_Offset[i] = PP_HOST_TO_SMC_US((uint16_t)(sclk_table->entries[i].sclk_offset) / 100);
+		}
+
+		result = polaris10_read_smc_sram_dword(smumgr,
+				SMU7_FIRMWARE_HEADER_LOCATION + offsetof(SMU74_Firmware_Header, AvfsMeanNSigma),
+				&tmp, data->sram_end);
+
+		polaris10_copy_bytes_to_smc(smumgr,
+					tmp,
+					(uint8_t *)&AVFS_meanNsigma,
+					sizeof(AVFS_meanNsigma_t),
+					data->sram_end);
+
+		result = polaris10_read_smc_sram_dword(smumgr,
+				SMU7_FIRMWARE_HEADER_LOCATION + offsetof(SMU74_Firmware_Header, AvfsSclkOffsetTable),
+				&tmp, data->sram_end);
+		polaris10_copy_bytes_to_smc(smumgr,
+					tmp,
+					(uint8_t *)&AVFS_SclkOffset,
+					sizeof(AVFS_Sclk_Offset_t),
+					data->sram_end);
+
+		data->avfs_vdroop_override_setting = (avfs_params.ucEnableGB_VDROOP_TABLE_CKSON << BTCGB0_Vdroop_Enable_SHIFT) |
+						(avfs_params.ucEnableGB_VDROOP_TABLE_CKSOFF << BTCGB1_Vdroop_Enable_SHIFT) |
+						(avfs_params.ucEnableGB_FUSE_TABLE_CKSON << AVFSGB0_Vdroop_Enable_SHIFT) |
+						(avfs_params.ucEnableGB_FUSE_TABLE_CKSOFF << AVFSGB1_Vdroop_Enable_SHIFT);
+		data->apply_avfs_cks_off_voltage = (avfs_params.ucEnableApplyAVFS_CKS_OFF_Voltage == 1) ? true : false;
+	}
+	return result;
+}
+
+
 /**
 * Initializes the SMC table and uploads it
 *
@@ -2056,6 +2107,10 @@ static int polaris10_init_smc_table(struct pp_hwmgr *hwmgr)
 				"Failed to populate Clock Stretcher Data Table!",
 				return result);
 	}
+
+	result = polaris10_populate_avfs_parameters(hwmgr);
+	PP_ASSERT_WITH_CODE(0 == result, "Failed to populate AVFS Parameters!", return result;);
+
 	table->CurrSclkPllRange = 0xff;
 	table->GraphicsVoltageChangeEnable  = 1;
 	table->GraphicsThermThrottleEnable  = 1;
@@ -2229,6 +2284,17 @@ static int polaris10_enable_ulv(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int polaris10_disable_ulv(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_ulv_parm *ulv = &(data->ulv);
+
+	if (ulv->ulv_supported)
+		return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_DisableULV);
+
+	return 0;
+}
+
 static int polaris10_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
 {
 	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
@@ -2249,9 +2315,27 @@ static int polaris10_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int polaris10_disable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr)
+{
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_SclkDeepSleep)) {
+		if (smum_send_msg_to_smc(hwmgr->smumgr,
+				PPSMC_MSG_MASTER_DeepSleep_OFF)) {
+			PP_ASSERT_WITH_CODE(false,
+					"Attempt to disable Master Deep Sleep switch failed!",
+					return -1);
+		}
+	}
+
+	return 0;
+}
+
 static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
 {
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	uint32_t soft_register_value = 0;
+	uint32_t handshake_disables_offset = data->soft_regs_start
+				+ offsetof(SMU74_SoftRegisters, HandshakeDisables);
 
 	/* enable SCLK dpm */
 	if (!data->sclk_dpm_key_disabled)
@@ -2262,6 +2346,12 @@ static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
 
 	/* enable MCLK dpm */
 	if (0 == data->mclk_dpm_key_disabled) {
+/* Disable UVD - SMU handshake for MCLK. */
+		soft_register_value = cgs_read_ind_register(hwmgr->device,
+					CGS_IND_REG__SMC, handshake_disables_offset);
+		soft_register_value |= SMU7_UVD_MCLK_HANDSHAKE_DISABLE;
+		cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC,
+				handshake_disables_offset, soft_register_value);
 
 		PP_ASSERT_WITH_CODE(
 				(0 == smum_send_msg_to_smc(hwmgr->smumgr,
@@ -2269,7 +2359,6 @@ static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
 				"Failed to enable MCLK DPM during DPM Start Function!",
 				return -1);
 
-
 		PHM_WRITE_FIELD(hwmgr->device, MC_SEQ_CNTL_3, CAC_EN, 0x1);
 
 		cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixLCAC_MC0_CNTL, 0x5);
@@ -2338,6 +2427,58 @@ static int polaris10_start_dpm(struct pp_hwmgr *hwmgr)
 	return 0;
 }
 
+static int polaris10_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+
+	/* disable SCLK dpm */
+	if (!data->sclk_dpm_key_disabled)
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_DPM_Disable) == 0),
+				"Failed to disable SCLK DPM!",
+				return -1);
+
+	/* disable MCLK dpm */
+	if (!data->mclk_dpm_key_disabled) {
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_MCLKDPM_Disable) == 0),
+				"Failed to disable MCLK DPM!",
+				return -1);
+	}
+
+	return 0;
+}
+
+static int polaris10_stop_dpm(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+
+	/* disable general power management */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, GENERAL_PWRMGT,
+			GLOBAL_PWRMGT_EN, 0);
+	/* disable sclk deep sleep */
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, SCLK_PWRMGT_CNTL,
+			DYNAMIC_PM_EN, 0);
+
+	/* disable PCIE dpm */
+	if (!data->pcie_dpm_key_disabled) {
+		PP_ASSERT_WITH_CODE(
+				(smum_send_msg_to_smc(hwmgr->smumgr,
+						PPSMC_MSG_PCIeDPM_Disable) == 0),
+				"Failed to disable pcie DPM during DPM Stop Function!",
+				return -1);
+	}
+
+	if (polaris10_disable_sclk_mclk_dpm(hwmgr)) {
+		printk(KERN_ERR "Failed to disable Sclk DPM and Mclk DPM!");
+		return -1;
+	}
+
+	return 0;
+}
+
 static void polaris10_set_dpm_event_sources(struct pp_hwmgr *hwmgr, uint32_t sources)
 {
 	bool protection;
@@ -2395,6 +2536,23 @@ static int polaris10_enable_thermal_auto_throttle(struct pp_hwmgr *hwmgr)
 	return polaris10_enable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal);
 }
 
+static int polaris10_disable_auto_throttle_source(struct pp_hwmgr *hwmgr,
+		PHM_AutoThrottleSource source)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+
+	if (data->active_auto_throttle_sources & (1 << source)) {
+		data->active_auto_throttle_sources &= ~(1 << source);
+		polaris10_set_dpm_event_sources(hwmgr, data->active_auto_throttle_sources);
+	}
+	return 0;
+}
+
+static int polaris10_disable_thermal_auto_throttle(struct pp_hwmgr *hwmgr)
+{
+	return polaris10_disable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal);
+}
+
 int polaris10_pcie_performance_request(struct pp_hwmgr *hwmgr)
 {
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
@@ -2516,8 +2674,60 @@ int polaris10_enable_dpm_tasks(struct pp_hwmgr *hwmgr)
 
 int polaris10_disable_dpm_tasks(struct pp_hwmgr *hwmgr)
 {
+	int tmp_result, result = 0;
 
-	return 0;
+	tmp_result = (polaris10_is_dpm_running(hwmgr)) ? 0 : -1;
+	PP_ASSERT_WITH_CODE(tmp_result == 0,
+			"DPM is not running right now, no need to disable DPM!",
+			return 0);
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_ThermalController))
+		PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+				GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, 1);
+
+	tmp_result = polaris10_disable_power_containment(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable power containment!", result = tmp_result);
+
+	tmp_result = polaris10_disable_smc_cac(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable SMC CAC!", result = tmp_result);
+
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			CG_SPLL_SPREAD_SPECTRUM, SSEN, 0);
+	PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC,
+			GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, 0);
+
+	tmp_result = polaris10_disable_thermal_auto_throttle(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable thermal auto throttle!", result = tmp_result);
+
+	tmp_result = polaris10_stop_dpm(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to stop DPM!", result = tmp_result);
+
+	tmp_result = polaris10_disable_deep_sleep_master_switch(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable deep sleep master switch!", result = tmp_result);
+
+	tmp_result = polaris10_disable_ulv(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to disable ULV!", result = tmp_result);
+
+	tmp_result = polaris10_clear_voting_clients(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to clear voting clients!", result = tmp_result);
+
+	tmp_result = polaris10_reset_to_default(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to reset to default!", result = tmp_result);
+
+	tmp_result = polaris10_force_switch_to_arbf0(hwmgr);
+	PP_ASSERT_WITH_CODE((tmp_result == 0),
+			"Failed to force to switch arbf0!", result = tmp_result);
+
+	return result;
 }
 
 int polaris10_reset_asic_tasks(struct pp_hwmgr *hwmgr)
@@ -2528,13 +2738,6 @@ int polaris10_reset_asic_tasks(struct pp_hwmgr *hwmgr)
 
 int polaris10_hwmgr_backend_fini(struct pp_hwmgr *hwmgr)
 {
-	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
-
-	if (data->soft_pp_table) {
-		kfree(data->soft_pp_table);
-		data->soft_pp_table = NULL;
-	}
-
 	return phm_hwmgr_backend_fini(hwmgr);
 }
 
@@ -2590,8 +2793,13 @@ int polaris10_set_features_platform_caps(struct pp_hwmgr *hwmgr)
 	phm_cap_unset(hwmgr->platform_descriptor.platformCaps,
 						PHM_PlatformCaps_TCPRamping);
 
-	phm_cap_set(hwmgr->platform_descriptor.platformCaps,
-					PHM_PlatformCaps_PowerContainment);
+	if (hwmgr->powercontainment_enabled)
+		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
+			    PHM_PlatformCaps_PowerContainment);
+	else
+		phm_cap_unset(hwmgr->platform_descriptor.platformCaps,
+			    PHM_PlatformCaps_PowerContainment);
+
 	phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 							PHM_PlatformCaps_CAC);
 
@@ -2606,6 +2814,7 @@ int polaris10_set_features_platform_caps(struct pp_hwmgr *hwmgr)
 
 	phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 						PHM_PlatformCaps_FanSpeedInTableIsRPM);
+
 	if (hwmgr->chip_id == CHIP_POLARIS11)
 		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 					PHM_PlatformCaps_SPLLShutdownSupport);
@@ -2662,12 +2871,12 @@ static int polaris10_get_evv_voltages(struct pp_hwmgr *hwmgr)
 				}
 			}
 
-
-			PP_ASSERT_WITH_CODE(0 == atomctrl_get_voltage_evv_on_sclk_ai(hwmgr,
-							VOLTAGE_TYPE_VDDC, sclk, vv_id, &vddc),
-						"Error retrieving EVV voltage value!",
-						continue);
-
+			if (atomctrl_get_voltage_evv_on_sclk_ai(hwmgr,
+						VOLTAGE_TYPE_VDDC,
+						sclk, vv_id, &vddc) != 0) {
+				printk(KERN_WARNING "failed to retrieving EVV voltage!\n");
+				continue;
+			}
 
 			/* need to make sure vddc is less than 2v or else, it could burn the ASIC. */
 			PP_ASSERT_WITH_CODE((vddc < 2000 && vddc != 0),
@@ -2898,13 +3107,19 @@ static int polaris10_set_private_data_based_on_pptable(struct pp_hwmgr *hwmgr)
 
 int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 {
-	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_hwmgr *data;
 	struct pp_atomctrl_gpio_pin_assignment gpio_pin_assignment;
 	uint32_t temp_reg;
 	int result;
 	struct phm_ppt_v1_information *table_info =
 			(struct phm_ppt_v1_information *)(hwmgr->pptable);
 
+	data = kzalloc(sizeof(struct polaris10_hwmgr), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+
+	hwmgr->backend = data;
+
 	data->dll_default_on = false;
 	data->sram_end = SMC_RAM_END;
 	data->mclk_dpm0_activity_target = 0xa;
@@ -2938,6 +3153,11 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 	data->vddci_control = POLARIS10_VOLTAGE_CONTROL_NONE;
 	data->mvdd_control = POLARIS10_VOLTAGE_CONTROL_NONE;
 
+	data->enable_tdc_limit_feature = true;
+	data->enable_pkg_pwr_tracking_feature = true;
+	data->force_pcie_gen = PP_PCIEGenInvalid;
+	data->mclk_stutter_mode_threshold = 40000;
+
 	if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr,
 			VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2))
 		data->voltage_control = POLARIS10_VOLTAGE_CONTROL_BY_SVID2;
@@ -2962,6 +3182,10 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 			data->vddci_control = POLARIS10_VOLTAGE_CONTROL_BY_SVID2;
 	}
 
+	if (table_info->cac_dtp_table->usClockStretchAmount != 0)
+		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
+					PHM_PlatformCaps_ClockStretcher);
+
 	polaris10_set_features_platform_caps(hwmgr);
 
 	polaris10_init_dpm_defaults(hwmgr);
@@ -3068,7 +3292,7 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_gen_cap = 0x30007;
+			data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK;
 		else
 			data->pcie_gen_cap = (uint32_t)sys_info.value;
 		if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
@@ -3077,7 +3301,7 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_lane_cap = 0x2f0000;
+			data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK;
 		else
 			data->pcie_lane_cap = (uint32_t)sys_info.value;
 
@@ -3520,10 +3744,11 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr,
 	ATOM_Tonga_State *state_entry = (ATOM_Tonga_State *)state;
 	ATOM_Tonga_POWERPLAYTABLE *powerplay_table =
 			(ATOM_Tonga_POWERPLAYTABLE *)pp_table;
-	ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table =
-			(ATOM_Tonga_SCLK_Dependency_Table *)
+	PPTable_Generic_SubTable_Header *sclk_dep_table =
+			(PPTable_Generic_SubTable_Header *)
 			(((unsigned long)powerplay_table) +
 				le16_to_cpu(powerplay_table->usSclkDependencyTableOffset));
+
 	ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table =
 			(ATOM_Tonga_MCLK_Dependency_Table *)
 			(((unsigned long)powerplay_table) +
@@ -3575,7 +3800,11 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr,
 	/* Performance levels are arranged from low to high. */
 	performance_level->memory_clock = mclk_dep_table->entries
 			[state_entry->ucMemoryClockIndexLow].ulMclk;
-	performance_level->engine_clock = sclk_dep_table->entries
+	if (sclk_dep_table->ucRevId == 0)
+		performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries
+			[state_entry->ucEngineClockIndexLow].ulSclk;
+	else if (sclk_dep_table->ucRevId == 1)
+		performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries
 			[state_entry->ucEngineClockIndexLow].ulSclk;
 	performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap,
 			state_entry->ucPCIEGenLow);
@@ -3586,8 +3815,14 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr,
 			[polaris10_power_state->performance_level_count++]);
 	performance_level->memory_clock = mclk_dep_table->entries
 			[state_entry->ucMemoryClockIndexHigh].ulMclk;
-	performance_level->engine_clock = sclk_dep_table->entries
+
+	if (sclk_dep_table->ucRevId == 0)
+		performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries
+			[state_entry->ucEngineClockIndexHigh].ulSclk;
+	else if (sclk_dep_table->ucRevId == 1)
+		performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries
 			[state_entry->ucEngineClockIndexHigh].ulSclk;
+
 	performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap,
 			state_entry->ucPCIEGenHigh);
 	performance_level->pcie_lane = get_pcie_lane_support(data->pcie_lane_cap,
@@ -3645,7 +3880,6 @@ static int polaris10_get_pp_table_entry(struct pp_hwmgr *hwmgr,
 		switch (state->classification.ui_label) {
 		case PP_StateUILabel_Performance:
 			data->use_pcie_performance_levels = true;
-
 			for (i = 0; i < ps->performance_level_count; i++) {
 				if (data->pcie_gen_performance.max <
 						ps->performance_levels[i].pcie_gen)
@@ -3661,7 +3895,6 @@ static int polaris10_get_pp_table_entry(struct pp_hwmgr *hwmgr,
 						ps->performance_levels[i].pcie_lane)
 					data->pcie_lane_performance.max =
 							ps->performance_levels[i].pcie_lane;
-
 				if (data->pcie_lane_performance.min >
 						ps->performance_levels[i].pcie_lane)
 					data->pcie_lane_performance.min =
@@ -4187,12 +4420,9 @@ int polaris10_update_samu_dpm(struct pp_hwmgr *hwmgr, bool bgate)
 {
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
 	uint32_t mm_boot_level_offset, mm_boot_level_value;
-	struct phm_ppt_v1_information *table_info =
-			(struct phm_ppt_v1_information *)(hwmgr->pptable);
 
 	if (!bgate) {
-		data->smc_state_table.SamuBootLevel =
-				(uint8_t) (table_info->mm_dep_table->count - 1);
+		data->smc_state_table.SamuBootLevel = 0;
 		mm_boot_level_offset = data->dpm_table_start +
 				offsetof(SMU74_Discrete_DpmTable, SamuBootLevel);
 		mm_boot_level_offset /= 4;
@@ -4721,42 +4951,6 @@ int polaris10_setup_asic_task(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
-static int polaris10_get_pp_table(struct pp_hwmgr *hwmgr, char **table)
-{
-	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kmemdup(hwmgr->soft_pp_table,
-					      hwmgr->soft_pp_table_size,
-					      GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	*table = (char *)&data->soft_pp_table;
-
-	return hwmgr->soft_pp_table_size;
-}
-
-static int polaris10_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size)
-{
-	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	memcpy(data->soft_pp_table, buf, size);
-
-	hwmgr->soft_pp_table = data->soft_pp_table;
-
-	/* TODO: re-init powerplay to implement modified pptable */
-
-	return 0;
-}
-
 static int polaris10_force_clock_level(struct pp_hwmgr *hwmgr,
 		enum pp_clock_type type, uint32_t mask)
 {
@@ -4899,6 +5093,89 @@ static int polaris10_get_fan_control_mode(struct pp_hwmgr *hwmgr)
 				CG_FDO_CTRL2, FDO_PWM_MODE);
 }
 
+static int polaris10_get_sclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table);
+	struct polaris10_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	int value;
+
+	value = (sclk_table->dpm_levels[sclk_table->count - 1].value -
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) *
+			100 /
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return value;
+}
+
+static int polaris10_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	struct pp_power_state  *ps;
+	struct polaris10_power_state  *polaris10_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	polaris10_ps = cast_phw_polaris10_power_state(&ps->hardware);
+
+	polaris10_ps->performance_levels[polaris10_ps->performance_level_count - 1].engine_clock =
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value *
+			value / 100 +
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return 0;
+}
+
+static int polaris10_get_mclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table);
+	struct polaris10_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	int value;
+
+	value = (mclk_table->dpm_levels[mclk_table->count - 1].value -
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) *
+			100 /
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return value;
+}
+
+static int polaris10_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	struct polaris10_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	struct pp_power_state  *ps;
+	struct polaris10_power_state  *polaris10_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	polaris10_ps = cast_phw_polaris10_power_state(&ps->hardware);
+
+	polaris10_ps->performance_levels[polaris10_ps->performance_level_count - 1].memory_clock =
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value *
+			value / 100 +
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return 0;
+}
 static const struct pp_hwmgr_func polaris10_hwmgr_funcs = {
 	.backend_init = &polaris10_hwmgr_backend_init,
 	.backend_fini = &polaris10_hwmgr_backend_fini,
@@ -4937,22 +5214,17 @@ static const struct pp_hwmgr_func polaris10_hwmgr_funcs = {
 	.check_states_equal = polaris10_check_states_equal,
 	.set_fan_control_mode = polaris10_set_fan_control_mode,
 	.get_fan_control_mode = polaris10_get_fan_control_mode,
-	.get_pp_table = polaris10_get_pp_table,
-	.set_pp_table = polaris10_set_pp_table,
 	.force_clock_level = polaris10_force_clock_level,
 	.print_clock_levels = polaris10_print_clock_levels,
 	.enable_per_cu_power_gating = polaris10_phm_enable_per_cu_power_gating,
+	.get_sclk_od = polaris10_get_sclk_od,
+	.set_sclk_od = polaris10_set_sclk_od,
+	.get_mclk_od = polaris10_get_mclk_od,
+	.set_mclk_od = polaris10_set_mclk_od,
 };
 
 int polaris10_hwmgr_init(struct pp_hwmgr *hwmgr)
 {
-	struct polaris10_hwmgr  *data;
-
-	data = kzalloc (sizeof(struct polaris10_hwmgr), GFP_KERNEL);
-	if (data == NULL)
-		return -ENOMEM;
-
-	hwmgr->backend = data;
 	hwmgr->hwmgr_func = &polaris10_hwmgr_funcs;
 	hwmgr->pptable_func = &tonga_pptable_funcs;
 	pp_polaris10_thermal_initialize(hwmgr);
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h
index beedf35cbfa6..fd38b0d7a3c2 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h
@@ -309,9 +309,8 @@ struct polaris10_hwmgr {
 	uint32_t                           up_hyst;
 	uint32_t disable_dpm_mask;
 	bool apply_optimized_settings;
-
-	/* soft pptable for re-uploading into smu */
-	void *soft_pp_table;
+	uint32_t                              avfs_vdroop_override_setting;
+	bool                                  apply_avfs_cks_off_voltage;
 };
 
 /* To convert to Q8.8 format for firmware */
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c
index 0b99ab3ba0c5..5620e268b553 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c
@@ -286,7 +286,7 @@ int polaris10_populate_pm_fuses(struct pp_hwmgr *hwmgr)
 
 		if (polaris10_copy_bytes_to_smc(hwmgr->smumgr, pm_fuse_table_offset,
 				(uint8_t *)&data->power_tune_table,
-				sizeof(struct SMU74_Discrete_PmFuses), data->sram_end))
+				(sizeof(struct SMU74_Discrete_PmFuses) - 92), data->sram_end))
 			PP_ASSERT_WITH_CODE(false,
 					"Attempt to download PmFuseTable Failed!",
 					return -EINVAL);
@@ -312,6 +312,23 @@ int polaris10_enable_smc_cac(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
+int polaris10_disable_smc_cac(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	int result = 0;
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_CAC) && data->cac_enabled) {
+		int smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+				(uint16_t)(PPSMC_MSG_DisableCac));
+		PP_ASSERT_WITH_CODE((smc_result == 0),
+				"Failed to disable CAC in SMC.", result = -1);
+
+		data->cac_enabled = false;
+	}
+	return result;
+}
+
 int polaris10_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n)
 {
 	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
@@ -373,6 +390,48 @@ int polaris10_enable_power_containment(struct pp_hwmgr *hwmgr)
 	return result;
 }
 
+int polaris10_disable_power_containment(struct pp_hwmgr *hwmgr)
+{
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
+	int result = 0;
+
+	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
+			PHM_PlatformCaps_PowerContainment) &&
+			data->power_containment_features) {
+		int smc_result;
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_TDCLimit) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_TDCLimitDisable));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable TDCLimit in SMC.",
+					result = smc_result);
+		}
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_DTE) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_DisableDTE));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable DTE in SMC.",
+					result = smc_result);
+		}
+
+		if (data->power_containment_features &
+				POWERCONTAINMENT_FEATURE_PkgPwrLimit) {
+			smc_result = smum_send_msg_to_smc(hwmgr->smumgr,
+					(uint16_t)(PPSMC_MSG_PkgPwrLimitDisable));
+			PP_ASSERT_WITH_CODE((smc_result == 0),
+					"Failed to disable PkgPwrTracking in SMC.",
+					result = smc_result);
+		}
+		data->power_containment_features = 0;
+	}
+
+	return result;
+}
+
 int polaris10_power_control_set_level(struct pp_hwmgr *hwmgr)
 {
 	struct phm_ppt_v1_information *table_info =
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h
index 68bc1cb6d40c..d492d6d28867 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h
@@ -31,6 +31,19 @@ enum polaris10_pt_config_reg_type {
 	POLARIS10_CONFIGREG_MAX
 };
 
+#define DIDT_SQ_CTRL0__UNUSED_0_MASK    0xfffc0000
+#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT  0x12
+#define DIDT_TD_CTRL0__UNUSED_0_MASK    0xfffc0000
+#define DIDT_TD_CTRL0__UNUSED_0__SHIFT  0x12
+#define DIDT_TCP_CTRL0__UNUSED_0_MASK   0xfffc0000
+#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT 0x12
+#define DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK                 0xc0000000
+#define DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT               0x0000001e
+#define DIDT_TD_TUNING_CTRL__UNUSED_0_MASK                 0xc0000000
+#define DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT               0x0000001e
+#define DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK                0xc0000000
+#define DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT              0x0000001e
+
 /* PowerContainment Features */
 #define POWERCONTAINMENT_FEATURE_DTE             0x00000001
 #define POWERCONTAINMENT_FEATURE_TDCLimit        0x00000002
@@ -62,7 +75,9 @@ void polaris10_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr);
 int polaris10_populate_bapm_parameters_in_dpm_table(struct pp_hwmgr *hwmgr);
 int polaris10_populate_pm_fuses(struct pp_hwmgr *hwmgr);
 int polaris10_enable_smc_cac(struct pp_hwmgr *hwmgr);
+int polaris10_disable_smc_cac(struct pp_hwmgr *hwmgr);
 int polaris10_enable_power_containment(struct pp_hwmgr *hwmgr);
+int polaris10_disable_power_containment(struct pp_hwmgr *hwmgr);
 int polaris10_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n);
 int polaris10_power_control_set_level(struct pp_hwmgr *hwmgr);
 
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c
index aba167f7d167..b206632d4650 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c
@@ -625,10 +625,14 @@ static int tf_polaris10_thermal_avfs_enable(struct pp_hwmgr *hwmgr,
 	int ret;
 	struct pp_smumgr *smumgr = (struct pp_smumgr *)(hwmgr->smumgr);
 	struct polaris10_smumgr *smu_data = (struct polaris10_smumgr *)(smumgr->backend);
+	struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
 
-	if (smu_data->avfs.avfs_btc_status != AVFS_BTC_ENABLEAVFS)
+	if (smu_data->avfs.avfs_btc_status == AVFS_BTC_NOTSUPPORTED)
 		return 0;
 
+	ret = smum_send_msg_to_smc_with_parameter(hwmgr->smumgr,
+			PPSMC_MSG_SetGBDroopSettings, data->avfs_vdroop_override_setting);
+
 	ret = (smum_send_msg_to_smc(smumgr, PPSMC_MSG_EnableAvfs) == 0) ?
 			0 : -1;
 
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c
index 58742e0d1492..a3c38bbd1e94 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c
@@ -44,6 +44,20 @@ bool acpi_atcs_functions_supported(void *device, uint32_t index)
 	return result == 0 ? (output_buf.function_bits & (1 << (index - 1))) != 0 : false;
 }
 
+bool acpi_atcs_notify_pcie_device_ready(void *device)
+{
+	int32_t temp_buffer = 1;
+
+	return cgs_call_acpi_method(device, CGS_ACPI_METHOD_ATCS,
+				ATCS_FUNCTION_PCIE_DEVICE_READY_NOTIFICATION,
+						&temp_buffer,
+						NULL,
+						0,
+						sizeof(temp_buffer),
+						0);
+}
+
+
 int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise)
 {
 	struct atcs_pref_req_input atcs_input;
@@ -52,7 +66,7 @@ int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise)
 	int result;
 	struct cgs_system_info info = {0};
 
-	if (!acpi_atcs_functions_supported(device, ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST))
+	if( 0 != acpi_atcs_notify_pcie_device_ready(device))
 		return -EINVAL;
 
 	info.size = sizeof(struct cgs_system_info);
@@ -77,7 +91,7 @@ int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise)
 						ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST,
 						&atcs_input,
 						&atcs_output,
-						0,
+						1,
 						sizeof(atcs_input),
 						sizeof(atcs_output));
 		if (result != 0)
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c
index da9f5f1b6dc2..5d70e2c47faf 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c
@@ -552,13 +552,13 @@ static bool atomctrl_lookup_gpio_pin(
 				pin_assignment->ucGpioPinBitShift;
 			gpio_pin_assignment->us_gpio_pin_aindex =
 				le16_to_cpu(pin_assignment->usGpioPin_AIndex);
-			return false;
+			return true;
 		}
 
 		offset += offsetof(ATOM_GPIO_PIN_ASSIGNMENT, ucGPIO_ID) + 1;
 	}
 
-	return true;
+	return false;
 }
 
 /**
@@ -1302,3 +1302,46 @@ int atomctrl_get_smc_sclk_range_table(struct pp_hwmgr *hwmgr, struct pp_atom_ctr
 
 	return 0;
 }
+
+int atomctrl_get_avfs_information(struct pp_hwmgr *hwmgr, struct pp_atom_ctrl__avfs_parameters *param)
+{
+	ATOM_ASIC_PROFILING_INFO_V3_6 *profile = NULL;
+
+	if (param == NULL)
+		return -EINVAL;
+
+	profile = (ATOM_ASIC_PROFILING_INFO_V3_6 *)
+			cgs_atom_get_data_table(hwmgr->device,
+					GetIndexIntoMasterTable(DATA, ASIC_ProfilingInfo),
+					NULL, NULL, NULL);
+	if (!profile)
+		return -1;
+
+	param->ulAVFS_meanNsigma_Acontant0 = profile->ulAVFS_meanNsigma_Acontant0;
+	param->ulAVFS_meanNsigma_Acontant1 = profile->ulAVFS_meanNsigma_Acontant1;
+	param->ulAVFS_meanNsigma_Acontant2 = profile->ulAVFS_meanNsigma_Acontant2;
+	param->usAVFS_meanNsigma_DC_tol_sigma = profile->usAVFS_meanNsigma_DC_tol_sigma;
+	param->usAVFS_meanNsigma_Platform_mean = profile->usAVFS_meanNsigma_Platform_mean;
+	param->usAVFS_meanNsigma_Platform_sigma = profile->usAVFS_meanNsigma_Platform_sigma;
+	param->ulGB_VDROOP_TABLE_CKSOFF_a0 = profile->ulGB_VDROOP_TABLE_CKSOFF_a0;
+	param->ulGB_VDROOP_TABLE_CKSOFF_a1 = profile->ulGB_VDROOP_TABLE_CKSOFF_a1;
+	param->ulGB_VDROOP_TABLE_CKSOFF_a2 = profile->ulGB_VDROOP_TABLE_CKSOFF_a2;
+	param->ulGB_VDROOP_TABLE_CKSON_a0 = profile->ulGB_VDROOP_TABLE_CKSON_a0;
+	param->ulGB_VDROOP_TABLE_CKSON_a1 = profile->ulGB_VDROOP_TABLE_CKSON_a1;
+	param->ulGB_VDROOP_TABLE_CKSON_a2 = profile->ulGB_VDROOP_TABLE_CKSON_a2;
+	param->ulAVFSGB_FUSE_TABLE_CKSOFF_m1 = profile->ulAVFSGB_FUSE_TABLE_CKSOFF_m1;
+	param->usAVFSGB_FUSE_TABLE_CKSOFF_m2 = profile->usAVFSGB_FUSE_TABLE_CKSOFF_m2;
+	param->ulAVFSGB_FUSE_TABLE_CKSOFF_b = profile->ulAVFSGB_FUSE_TABLE_CKSOFF_b;
+	param->ulAVFSGB_FUSE_TABLE_CKSON_m1 = profile->ulAVFSGB_FUSE_TABLE_CKSON_m1;
+	param->usAVFSGB_FUSE_TABLE_CKSON_m2 = profile->usAVFSGB_FUSE_TABLE_CKSON_m2;
+	param->ulAVFSGB_FUSE_TABLE_CKSON_b = profile->ulAVFSGB_FUSE_TABLE_CKSON_b;
+	param->usMaxVoltage_0_25mv = profile->usMaxVoltage_0_25mv;
+	param->ucEnableGB_VDROOP_TABLE_CKSOFF = profile->ucEnableGB_VDROOP_TABLE_CKSOFF;
+	param->ucEnableGB_VDROOP_TABLE_CKSON = profile->ucEnableGB_VDROOP_TABLE_CKSON;
+	param->ucEnableGB_FUSE_TABLE_CKSOFF = profile->ucEnableGB_FUSE_TABLE_CKSOFF;
+	param->ucEnableGB_FUSE_TABLE_CKSON = profile->ucEnableGB_FUSE_TABLE_CKSON;
+	param->usPSM_Age_ComFactor = profile->usPSM_Age_ComFactor;
+	param->ucEnableApplyAVFS_CKS_OFF_Voltage = profile->ucEnableApplyAVFS_CKS_OFF_Voltage;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h
index d24ebb566905..248c5db5f380 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h
@@ -250,6 +250,35 @@ struct pp_atomctrl_gpio_pin_assignment {
 };
 typedef struct pp_atomctrl_gpio_pin_assignment pp_atomctrl_gpio_pin_assignment;
 
+struct pp_atom_ctrl__avfs_parameters {
+	uint32_t  ulAVFS_meanNsigma_Acontant0;
+	uint32_t  ulAVFS_meanNsigma_Acontant1;
+	uint32_t  ulAVFS_meanNsigma_Acontant2;
+	uint16_t usAVFS_meanNsigma_DC_tol_sigma;
+	uint16_t usAVFS_meanNsigma_Platform_mean;
+	uint16_t usAVFS_meanNsigma_Platform_sigma;
+	uint32_t  ulGB_VDROOP_TABLE_CKSOFF_a0;
+	uint32_t  ulGB_VDROOP_TABLE_CKSOFF_a1;
+	uint32_t  ulGB_VDROOP_TABLE_CKSOFF_a2;
+	uint32_t  ulGB_VDROOP_TABLE_CKSON_a0;
+	uint32_t  ulGB_VDROOP_TABLE_CKSON_a1;
+	uint32_t  ulGB_VDROOP_TABLE_CKSON_a2;
+	uint32_t  ulAVFSGB_FUSE_TABLE_CKSOFF_m1;
+	uint16_t  usAVFSGB_FUSE_TABLE_CKSOFF_m2;
+	uint32_t  ulAVFSGB_FUSE_TABLE_CKSOFF_b;
+	uint32_t  ulAVFSGB_FUSE_TABLE_CKSON_m1;
+	uint16_t  usAVFSGB_FUSE_TABLE_CKSON_m2;
+	uint32_t  ulAVFSGB_FUSE_TABLE_CKSON_b;
+	uint16_t  usMaxVoltage_0_25mv;
+	uint8_t  ucEnableGB_VDROOP_TABLE_CKSOFF;
+	uint8_t  ucEnableGB_VDROOP_TABLE_CKSON;
+	uint8_t  ucEnableGB_FUSE_TABLE_CKSOFF;
+	uint8_t  ucEnableGB_FUSE_TABLE_CKSON;
+	uint16_t usPSM_Age_ComFactor;
+	uint8_t  ucEnableApplyAVFS_CKS_OFF_Voltage;
+	uint8_t  ucReserved;
+};
+
 extern bool atomctrl_get_pp_assign_pin(struct pp_hwmgr *hwmgr, const uint32_t pinId, pp_atomctrl_gpio_pin_assignment *gpio_pin_assignment);
 extern int atomctrl_get_voltage_evv_on_sclk(struct pp_hwmgr *hwmgr, uint8_t voltage_type, uint32_t sclk, uint16_t virtual_voltage_Id, uint16_t *voltage);
 extern uint32_t atomctrl_get_mpll_reference_clock(struct pp_hwmgr *hwmgr);
@@ -278,5 +307,8 @@ extern int atomctrl_set_ac_timing_ai(struct pp_hwmgr *hwmgr, uint32_t memory_clo
 extern int atomctrl_get_voltage_evv_on_sclk_ai(struct pp_hwmgr *hwmgr, uint8_t voltage_type,
 				uint32_t sclk, uint16_t virtual_voltage_Id, uint16_t *voltage);
 extern int atomctrl_get_smc_sclk_range_table(struct pp_hwmgr *hwmgr, struct pp_atom_ctrl_sclk_range_table *table);
+
+extern int atomctrl_get_avfs_information(struct pp_hwmgr *hwmgr, struct pp_atom_ctrl__avfs_parameters *param);
+
 #endif
 
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c
index 2f1a14fe05b1..35bc8a29b773 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c
@@ -794,15 +794,18 @@ static const ATOM_PPLIB_STATE_V2 *get_state_entry_v2(
 static const ATOM_PPLIB_POWERPLAYTABLE *get_powerplay_table(
 				     struct pp_hwmgr *hwmgr)
 {
-	const void *table_addr = NULL;
+	const void *table_addr = hwmgr->soft_pp_table;
 	uint8_t frev, crev;
 	uint16_t size;
 
-	table_addr = cgs_atom_get_data_table(hwmgr->device,
-			GetIndexIntoMasterTable(DATA, PowerPlayInfo),
-			&size, &frev, &crev);
+	if (!table_addr) {
+		table_addr = cgs_atom_get_data_table(hwmgr->device,
+				GetIndexIntoMasterTable(DATA, PowerPlayInfo),
+				&size, &frev, &crev);
 
-	hwmgr->soft_pp_table = table_addr;
+		hwmgr->soft_pp_table = table_addr;
+		hwmgr->soft_pp_table_size = size;
+	}
 
 	return (const ATOM_PPLIB_POWERPLAYTABLE *)table_addr;
 }
@@ -1499,7 +1502,7 @@ int get_number_of_vce_state_table_entries(
 	const ATOM_PPLIB_VCE_State_Table *vce_table =
 				    get_vce_state_table(hwmgr, table);
 
-	if (vce_table > 0)
+	if (vce_table)
 		return vce_table->numEntries;
 
 	return 0;
@@ -1589,11 +1592,6 @@ static int pp_tables_initialize(struct pp_hwmgr *hwmgr)
 
 static int pp_tables_uninitialize(struct pp_hwmgr *hwmgr)
 {
-	if (NULL != hwmgr->soft_pp_table) {
-		kfree(hwmgr->soft_pp_table);
-		hwmgr->soft_pp_table = NULL;
-	}
-
 	if (NULL != hwmgr->dyn_state.vddc_dependency_on_sclk) {
 		kfree(hwmgr->dyn_state.vddc_dependency_on_sclk);
 		hwmgr->dyn_state.vddc_dependency_on_sclk = NULL;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c
index 16fed487973b..6c4553cf8023 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c
@@ -2847,27 +2847,6 @@ static int tonga_setup_default_dpm_tables(struct pp_hwmgr *hwmgr)
 		}
 	}
 
-	/* Initialize Vddc DPM table based on allow Vddc values.  And populate corresponding std values. */
-	for (i = 0; i < allowed_vdd_sclk_table->count; i++) {
-		data->dpm_table.vddc_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].vddc;
-		/* tonga_hwmgr->dpm_table.VddcTable.dpm_levels[i].param1 = stdVoltageTable->entries[i].Leakage; */
-		/* param1 is for corresponding std voltage */
-		data->dpm_table.vddc_table.dpm_levels[i].enabled = 1;
-	}
-	data->dpm_table.vddc_table.count = allowed_vdd_sclk_table->count;
-
-	if (NULL != allowed_vdd_mclk_table) {
-		/* Initialize Vddci DPM table based on allow Mclk values */
-		for (i = 0; i < allowed_vdd_mclk_table->count; i++) {
-			data->dpm_table.vdd_ci_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].vddci;
-			data->dpm_table.vdd_ci_table.dpm_levels[i].enabled = 1;
-			data->dpm_table.mvdd_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].mvdd;
-			data->dpm_table.mvdd_table.dpm_levels[i].enabled = 1;
-		}
-		data->dpm_table.vdd_ci_table.count = allowed_vdd_mclk_table->count;
-		data->dpm_table.mvdd_table.count = allowed_vdd_mclk_table->count;
-	}
-
 	/* setup PCIE gen speed levels*/
 	tonga_setup_default_pcie_tables(hwmgr);
 
@@ -3047,8 +3026,8 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr)
 
 	reg_value = 0;
 	if ((0 == reg_value) &&
-		(0 == atomctrl_get_pp_assign_pin(hwmgr,
-			VDDC_VRHOT_GPIO_PINID, &gpio_pin_assignment))) {
+		(atomctrl_get_pp_assign_pin(hwmgr, VDDC_VRHOT_GPIO_PINID,
+						&gpio_pin_assignment))) {
 		table->VRHotGpio = gpio_pin_assignment.uc_gpio_pin_bit_shift;
 		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 			PHM_PlatformCaps_RegulatorHot);
@@ -3061,8 +3040,8 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr)
 	/* ACDC Switch GPIO */
 	reg_value = 0;
 	if ((0 == reg_value) &&
-		(0 == atomctrl_get_pp_assign_pin(hwmgr,
-			PP_AC_DC_SWITCH_GPIO_PINID, &gpio_pin_assignment))) {
+		(atomctrl_get_pp_assign_pin(hwmgr, PP_AC_DC_SWITCH_GPIO_PINID,
+						&gpio_pin_assignment))) {
 		table->AcDcGpio = gpio_pin_assignment.uc_gpio_pin_bit_shift;
 		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 			PHM_PlatformCaps_AutomaticDCTransition);
@@ -3084,8 +3063,7 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr)
 	}
 
 	reg_value = 0;
-	if ((0 == reg_value) &&
-		(0 == atomctrl_get_pp_assign_pin(hwmgr,
+	if ((0 == reg_value) && (atomctrl_get_pp_assign_pin(hwmgr,
 			THERMAL_INT_OUTPUT_GPIO_PINID, &gpio_pin_assignment))) {
 		phm_cap_set(hwmgr->platform_descriptor.platformCaps,
 			PHM_PlatformCaps_ThermalOutGPIO);
@@ -4443,13 +4421,6 @@ int tonga_reset_asic_tasks(struct pp_hwmgr *hwmgr)
 
 int tonga_hwmgr_backend_fini(struct pp_hwmgr *hwmgr)
 {
-	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
-
-	if (data->soft_pp_table) {
-		kfree(data->soft_pp_table);
-		data->soft_pp_table = NULL;
-	}
-
 	return phm_hwmgr_backend_fini(hwmgr);
 }
 
@@ -4463,7 +4434,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 {
 	int result = 0;
 	SMU72_Discrete_DpmTable  *table = NULL;
-	tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
+	tonga_hwmgr *data;
 	pp_atomctrl_gpio_pin_assignment gpio_pin_assignment;
 	struct phm_ppt_v1_information *pptable_info = (struct phm_ppt_v1_information *)(hwmgr->pptable);
 	phw_tonga_ulv_parm *ulv;
@@ -4472,6 +4443,12 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 	PP_ASSERT_WITH_CODE((NULL != hwmgr),
 		"Invalid Parameter!", return -1;);
 
+	data = kzalloc(sizeof(struct tonga_hwmgr), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+
+	hwmgr->backend = data;
+
 	data->dll_defaule_on = 0;
 	data->sram_end = SMC_RAM_END;
 
@@ -4510,6 +4487,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 	data->vdd_ci_control = TONGA_VOLTAGE_CONTROL_NONE;
 	data->vdd_gfx_control = TONGA_VOLTAGE_CONTROL_NONE;
 	data->mvdd_control = TONGA_VOLTAGE_CONTROL_NONE;
+	data->force_pcie_gen = PP_PCIEGenInvalid;
 
 	if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr,
 				VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) {
@@ -4591,7 +4569,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 	* if ucGPIO_ID=VDDC_PCC_GPIO_PINID in GPIO_LUTable,
 	* Peak Current Control feature is enabled and we should program PCC HW register
 	*/
-	if (0 == atomctrl_get_pp_assign_pin(hwmgr, VDDC_PCC_GPIO_PINID, &gpio_pin_assignment)) {
+	if (atomctrl_get_pp_assign_pin(hwmgr, VDDC_PCC_GPIO_PINID, &gpio_pin_assignment)) {
 		uint32_t temp_reg = cgs_read_ind_register(hwmgr->device,
 										CGS_IND_REG__SMC, ixCNB_PWRMGT_CNTL);
 
@@ -4659,7 +4637,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_gen_cap = 0x30007;
+			data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK;
 		else
 			data->pcie_gen_cap = (uint32_t)sys_info.value;
 		if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3)
@@ -4668,7 +4646,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr)
 		sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW;
 		result = cgs_query_system_info(hwmgr->device, &sys_info);
 		if (result)
-			data->pcie_lane_cap = 0x2f0000;
+			data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK;
 		else
 			data->pcie_lane_cap = (uint32_t)sys_info.value;
 	} else {
@@ -6051,42 +6029,6 @@ static int tonga_get_fan_control_mode(struct pp_hwmgr *hwmgr)
 				CG_FDO_CTRL2, FDO_PWM_MODE);
 }
 
-static int tonga_get_pp_table(struct pp_hwmgr *hwmgr, char **table)
-{
-	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kmemdup(hwmgr->soft_pp_table,
-					      hwmgr->soft_pp_table_size,
-					      GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	*table = (char *)&data->soft_pp_table;
-
-	return hwmgr->soft_pp_table_size;
-}
-
-static int tonga_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size)
-{
-	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
-
-	if (!data->soft_pp_table) {
-		data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL);
-		if (!data->soft_pp_table)
-			return -ENOMEM;
-	}
-
-	memcpy(data->soft_pp_table, buf, size);
-
-	hwmgr->soft_pp_table = data->soft_pp_table;
-
-	/* TODO: re-init powerplay to implement modified pptable */
-
-	return 0;
-}
-
 static int tonga_force_clock_level(struct pp_hwmgr *hwmgr,
 		enum pp_clock_type type, uint32_t mask)
 {
@@ -6194,11 +6136,96 @@ static int tonga_print_clock_levels(struct pp_hwmgr *hwmgr,
 	return size;
 }
 
+static int tonga_get_sclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
+	struct tonga_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table);
+	struct tonga_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	int value;
+
+	value = (sclk_table->dpm_levels[sclk_table->count - 1].value -
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) *
+			100 /
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return value;
+}
+
+static int tonga_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
+	struct tonga_single_dpm_table *golden_sclk_table =
+			&(data->golden_dpm_table.sclk_table);
+	struct pp_power_state  *ps;
+	struct tonga_power_state  *tonga_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	tonga_ps = cast_phw_tonga_power_state(&ps->hardware);
+
+	tonga_ps->performance_levels[tonga_ps->performance_level_count - 1].engine_clock =
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value *
+			value / 100 +
+			golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value;
+
+	return 0;
+}
+
+static int tonga_get_mclk_od(struct pp_hwmgr *hwmgr)
+{
+	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
+	struct tonga_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table);
+	struct tonga_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	int value;
+
+	value = (mclk_table->dpm_levels[mclk_table->count - 1].value -
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) *
+			100 /
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return value;
+}
+
+static int tonga_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value)
+{
+	struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend);
+	struct tonga_single_dpm_table *golden_mclk_table =
+			&(data->golden_dpm_table.mclk_table);
+	struct pp_power_state  *ps;
+	struct tonga_power_state  *tonga_ps;
+
+	if (value > 20)
+		value = 20;
+
+	ps = hwmgr->request_ps;
+
+	if (ps == NULL)
+		return -EINVAL;
+
+	tonga_ps = cast_phw_tonga_power_state(&ps->hardware);
+
+	tonga_ps->performance_levels[tonga_ps->performance_level_count - 1].memory_clock =
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value *
+			value / 100 +
+			golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value;
+
+	return 0;
+}
+
 static const struct pp_hwmgr_func tonga_hwmgr_funcs = {
 	.backend_init = &tonga_hwmgr_backend_init,
 	.backend_fini = &tonga_hwmgr_backend_fini,
 	.asic_setup = &tonga_setup_asic_task,
 	.dynamic_state_management_enable = &tonga_enable_dpm_tasks,
+	.dynamic_state_management_disable = &tonga_disable_dpm_tasks,
 	.apply_state_adjust_rules = tonga_apply_state_adjust_rules,
 	.force_dpm_level = &tonga_force_dpm_level,
 	.power_state_set = tonga_set_power_state_tasks,
@@ -6232,22 +6259,16 @@ static const struct pp_hwmgr_func tonga_hwmgr_funcs = {
 	.check_states_equal = tonga_check_states_equal,
 	.set_fan_control_mode = tonga_set_fan_control_mode,
 	.get_fan_control_mode = tonga_get_fan_control_mode,
-	.get_pp_table = tonga_get_pp_table,
-	.set_pp_table = tonga_set_pp_table,
 	.force_clock_level = tonga_force_clock_level,
 	.print_clock_levels = tonga_print_clock_levels,
+	.get_sclk_od = tonga_get_sclk_od,
+	.set_sclk_od = tonga_set_sclk_od,
+	.get_mclk_od = tonga_get_mclk_od,
+	.set_mclk_od = tonga_set_mclk_od,
 };
 
 int tonga_hwmgr_init(struct pp_hwmgr *hwmgr)
 {
-	tonga_hwmgr  *data;
-
-	data = kzalloc (sizeof(tonga_hwmgr), GFP_KERNEL);
-	if (data == NULL)
-		return -ENOMEM;
-	memset(data, 0x00, sizeof(tonga_hwmgr));
-
-	hwmgr->backend = data;
 	hwmgr->hwmgr_func = &tonga_hwmgr_funcs;
 	hwmgr->pptable_func = &tonga_pptable_funcs;
 	pp_tonga_thermal_initialize(hwmgr);
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h
index 573cd39fe78d..3961884bfa9b 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h
@@ -352,9 +352,6 @@ struct tonga_hwmgr {
 	bool                           samu_power_gated; /* 1: gated, 0:not gated */
 	bool                           acp_power_gated;  /* 1: gated, 0:not gated */
 	bool                           pg_acp_init;
-
-	/* soft pptable for re-uploading into smu */
-	void *soft_pp_table;
 };
 
 typedef struct tonga_hwmgr tonga_hwmgr;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h
index 1b44f4e9b8f5..f127198aafc4 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h
@@ -197,6 +197,22 @@ typedef struct _ATOM_Tonga_SCLK_Dependency_Table {
 	ATOM_Tonga_SCLK_Dependency_Record entries[1];				 /* Dynamically allocate entries. */
 } ATOM_Tonga_SCLK_Dependency_Table;
 
+typedef struct _ATOM_Polaris_SCLK_Dependency_Record {
+	UCHAR  ucVddInd;											/* Base voltage */
+	USHORT usVddcOffset;										/* Offset relative to base voltage */
+	ULONG ulSclk;
+	USHORT usEdcCurrent;
+	UCHAR  ucReliabilityTemperature;
+	UCHAR  ucCKSVOffsetandDisable;			/* Bits 0~6: Voltage offset for CKS, Bit 7: Disable/enable for the SCLK level. */
+	ULONG  ulSclkOffset;
+} ATOM_Polaris_SCLK_Dependency_Record;
+
+typedef struct _ATOM_Polaris_SCLK_Dependency_Table {
+	UCHAR ucRevId;
+	UCHAR ucNumEntries;							/* Number of entries. */
+	ATOM_Polaris_SCLK_Dependency_Record entries[1];				 /* Dynamically allocate entries. */
+} ATOM_Polaris_SCLK_Dependency_Table;
+
 typedef struct _ATOM_Tonga_PCIE_Record {
 	UCHAR ucPCIEGenSpeed;
 	UCHAR usPCIELaneWidth;
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c
index 10e3630ee39d..94d6b472e1fe 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c
@@ -408,41 +408,78 @@ static int get_mclk_voltage_dependency_table(
 static int get_sclk_voltage_dependency_table(
 		struct pp_hwmgr *hwmgr,
 		phm_ppt_v1_clock_voltage_dependency_table **pp_tonga_sclk_dep_table,
-		const ATOM_Tonga_SCLK_Dependency_Table * sclk_dep_table
+		const PPTable_Generic_SubTable_Header *sclk_dep_table
 		)
 {
 	uint32_t table_size, i;
 	phm_ppt_v1_clock_voltage_dependency_table *sclk_table;
 
-	PP_ASSERT_WITH_CODE((0 != sclk_dep_table->ucNumEntries),
-		"Invalid PowerPlay Table!", return -1);
+	if (sclk_dep_table->ucRevId < 1) {
+		const ATOM_Tonga_SCLK_Dependency_Table *tonga_table =
+			    (ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table;
 
-	table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
-		* sclk_dep_table->ucNumEntries;
+		PP_ASSERT_WITH_CODE((0 != tonga_table->ucNumEntries),
+			"Invalid PowerPlay Table!", return -1);
 
-	sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
-		kzalloc(table_size, GFP_KERNEL);
+		table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
+			* tonga_table->ucNumEntries;
 
-	if (NULL == sclk_table)
-		return -ENOMEM;
+		sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
+			kzalloc(table_size, GFP_KERNEL);
 
-	memset(sclk_table, 0x00, table_size);
-
-	sclk_table->count = (uint32_t)sclk_dep_table->ucNumEntries;
-
-	for (i = 0; i < sclk_dep_table->ucNumEntries; i++) {
-		sclk_table->entries[i].vddInd =
-			sclk_dep_table->entries[i].ucVddInd;
-		sclk_table->entries[i].vdd_offset =
-			sclk_dep_table->entries[i].usVddcOffset;
-		sclk_table->entries[i].clk =
-			sclk_dep_table->entries[i].ulSclk;
-		sclk_table->entries[i].cks_enable =
-			(((sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
-		sclk_table->entries[i].cks_voffset =
-			(sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
-	}
+		if (NULL == sclk_table)
+			return -ENOMEM;
 
+		memset(sclk_table, 0x00, table_size);
+
+		sclk_table->count = (uint32_t)tonga_table->ucNumEntries;
+
+		for (i = 0; i < tonga_table->ucNumEntries; i++) {
+			sclk_table->entries[i].vddInd =
+				tonga_table->entries[i].ucVddInd;
+			sclk_table->entries[i].vdd_offset =
+				tonga_table->entries[i].usVddcOffset;
+			sclk_table->entries[i].clk =
+				tonga_table->entries[i].ulSclk;
+			sclk_table->entries[i].cks_enable =
+				(((tonga_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
+			sclk_table->entries[i].cks_voffset =
+				(tonga_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
+		}
+	} else {
+		const ATOM_Polaris_SCLK_Dependency_Table *polaris_table =
+			    (ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table;
+
+		PP_ASSERT_WITH_CODE((0 != polaris_table->ucNumEntries),
+			"Invalid PowerPlay Table!", return -1);
+
+		table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
+			* polaris_table->ucNumEntries;
+
+		sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
+			kzalloc(table_size, GFP_KERNEL);
+
+		if (NULL == sclk_table)
+			return -ENOMEM;
+
+		memset(sclk_table, 0x00, table_size);
+
+		sclk_table->count = (uint32_t)polaris_table->ucNumEntries;
+
+		for (i = 0; i < polaris_table->ucNumEntries; i++) {
+			sclk_table->entries[i].vddInd =
+				polaris_table->entries[i].ucVddInd;
+			sclk_table->entries[i].vdd_offset =
+				polaris_table->entries[i].usVddcOffset;
+			sclk_table->entries[i].clk =
+				polaris_table->entries[i].ulSclk;
+			sclk_table->entries[i].cks_enable =
+				(((polaris_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
+			sclk_table->entries[i].cks_voffset =
+				(polaris_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
+			sclk_table->entries[i].sclk_offset = polaris_table->entries[i].ulSclkOffset;
+		}
+	}
 	*pp_tonga_sclk_dep_table = sclk_table;
 
 	return 0;
@@ -708,8 +745,8 @@ static int init_clock_voltage_dependency(
 	const ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table =
 		(const ATOM_Tonga_MCLK_Dependency_Table *)(((unsigned long) powerplay_table) +
 		le16_to_cpu(powerplay_table->usMclkDependencyTableOffset));
-	const ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table =
-		(const ATOM_Tonga_SCLK_Dependency_Table *)(((unsigned long) powerplay_table) +
+	const PPTable_Generic_SubTable_Header *sclk_dep_table =
+		(const PPTable_Generic_SubTable_Header *)(((unsigned long) powerplay_table) +
 		le16_to_cpu(powerplay_table->usSclkDependencyTableOffset));
 	const ATOM_Tonga_Hard_Limit_Table *pHardLimits =
 		(const ATOM_Tonga_Hard_Limit_Table *)(((unsigned long) powerplay_table) +
@@ -1040,48 +1077,41 @@ int tonga_pp_tables_uninitialize(struct pp_hwmgr *hwmgr)
 	struct phm_ppt_v1_information *pp_table_information =
 		(struct phm_ppt_v1_information *)(hwmgr->pptable);
 
-	if (NULL != hwmgr->soft_pp_table) {
-		kfree(hwmgr->soft_pp_table);
-		hwmgr->soft_pp_table = NULL;
-	}
-
-	if (NULL != pp_table_information->vdd_dep_on_sclk)
-		pp_table_information->vdd_dep_on_sclk = NULL;
+	kfree(pp_table_information->vdd_dep_on_sclk);
+	pp_table_information->vdd_dep_on_sclk = NULL;
 
-	if (NULL != pp_table_information->vdd_dep_on_mclk)
-		pp_table_information->vdd_dep_on_mclk = NULL;
+	kfree(pp_table_information->vdd_dep_on_mclk);
+	pp_table_information->vdd_dep_on_mclk = NULL;
 
-	if (NULL != pp_table_information->valid_mclk_values)
-		pp_table_information->valid_mclk_values = NULL;
+	kfree(pp_table_information->valid_mclk_values);
+	pp_table_information->valid_mclk_values = NULL;
 
-	if (NULL != pp_table_information->valid_sclk_values)
-		pp_table_information->valid_sclk_values = NULL;
+	kfree(pp_table_information->valid_sclk_values);
+	pp_table_information->valid_sclk_values = NULL;
 
-	if (NULL != pp_table_information->vddc_lookup_table)
-		pp_table_information->vddc_lookup_table = NULL;
+	kfree(pp_table_information->vddc_lookup_table);
+	pp_table_information->vddc_lookup_table = NULL;
 
-	if (NULL != pp_table_information->vddgfx_lookup_table)
-		pp_table_information->vddgfx_lookup_table = NULL;
+	kfree(pp_table_information->vddgfx_lookup_table);
+	pp_table_information->vddgfx_lookup_table = NULL;
 
-	if (NULL != pp_table_information->mm_dep_table)
-		pp_table_information->mm_dep_table = NULL;
+	kfree(pp_table_information->mm_dep_table);
+	pp_table_information->mm_dep_table = NULL;
 
-	if (NULL != pp_table_information->cac_dtp_table)
-		pp_table_information->cac_dtp_table = NULL;
+	kfree(pp_table_information->cac_dtp_table);
+	pp_table_information->cac_dtp_table = NULL;
 
-	if (NULL != hwmgr->dyn_state.cac_dtp_table)
-		hwmgr->dyn_state.cac_dtp_table = NULL;
+	kfree(hwmgr->dyn_state.cac_dtp_table);
+	hwmgr->dyn_state.cac_dtp_table = NULL;
 
-	if (NULL != pp_table_information->ppm_parameter_table)
-		pp_table_information->ppm_parameter_table = NULL;
+	kfree(pp_table_information->ppm_parameter_table);
+	pp_table_information->ppm_parameter_table = NULL;
 
-	if (NULL != pp_table_information->pcie_table)
-		pp_table_information->pcie_table = NULL;
+	kfree(pp_table_information->pcie_table);
+	pp_table_information->pcie_table = NULL;
 
-	if (NULL != hwmgr->pptable) {
-		kfree(hwmgr->pptable);
-		hwmgr->pptable = NULL;
-	}
+	kfree(hwmgr->pptable);
+	hwmgr->pptable = NULL;
 
 	return result;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h
index 50b367d44307..b764c8c05ec8 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h
@@ -132,6 +132,7 @@ struct amd_pp_init {
 	uint32_t chip_family;
 	uint32_t chip_id;
 	uint32_t rev_id;
+	bool powercontainment_enabled;
 };
 enum amd_pp_display_config_type{
 	AMD_PP_DisplayConfigType_None = 0,
@@ -342,6 +343,10 @@ struct amd_powerplay_funcs {
 	int (*set_pp_table)(void *handle, const char *buf, size_t size);
 	int (*force_clock_level)(void *handle, enum pp_clock_type type, uint32_t mask);
 	int (*print_clock_levels)(void *handle, enum pp_clock_type type, char *buf);
+	int (*get_sclk_od)(void *handle);
+	int (*set_sclk_od)(void *handle, uint32_t value);
+	int (*get_mclk_od)(void *handle);
+	int (*set_mclk_od)(void *handle, uint32_t value);
 };
 
 struct amd_powerplay {
@@ -355,6 +360,8 @@ int amd_powerplay_init(struct amd_pp_init *pp_init,
 
 int amd_powerplay_fini(void *handle);
 
+int amd_powerplay_reset(void *handle);
+
 int amd_powerplay_display_configuration_change(void *handle,
 		const struct amd_pp_display_configuration *input);
 
diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h
index 56f712c7d07a..962cb5385951 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h
@@ -340,6 +340,7 @@ extern int phm_powergate_vce(struct pp_hwmgr *hwmgr, bool gate);
 extern int phm_powerdown_uvd(struct pp_hwmgr *hwmgr);
 extern int phm_setup_asic(struct pp_hwmgr *hwmgr);
 extern int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr);
+extern int phm_disable_dynamic_state_management(struct pp_hwmgr *hwmgr);
 extern void phm_init_dynamic_caps(struct pp_hwmgr *hwmgr);
 extern bool phm_is_hw_access_blocked(struct pp_hwmgr *hwmgr);
 extern int phm_block_hw_access(struct pp_hwmgr *hwmgr, bool block);
diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h
index 28f571449495..0bbc42a224e5 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h
@@ -278,6 +278,8 @@ struct pp_hwmgr_func {
 
 	int (*dynamic_state_management_enable)(
 						struct pp_hwmgr *hw_mgr);
+	int (*dynamic_state_management_disable)(
+						struct pp_hwmgr *hw_mgr);
 
 	int (*patch_boot_state)(struct pp_hwmgr *hwmgr,
 				     struct pp_hw_power_state *hw_ps);
@@ -333,11 +335,13 @@ struct pp_hwmgr_func {
 	int (*get_clock_by_type)(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks);
 	int (*get_max_high_clocks)(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks);
 	int (*power_off_asic)(struct pp_hwmgr *hwmgr);
-	int (*get_pp_table)(struct pp_hwmgr *hwmgr, char **table);
-	int (*set_pp_table)(struct pp_hwmgr *hwmgr, const char *buf, size_t size);
 	int (*force_clock_level)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask);
 	int (*print_clock_levels)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf);
 	int (*enable_per_cu_power_gating)(struct pp_hwmgr *hwmgr, bool enable);
+	int (*get_sclk_od)(struct pp_hwmgr *hwmgr);
+	int (*set_sclk_od)(struct pp_hwmgr *hwmgr, uint32_t value);
+	int (*get_mclk_od)(struct pp_hwmgr *hwmgr);
+	int (*set_mclk_od)(struct pp_hwmgr *hwmgr, uint32_t value);
 };
 
 struct pp_table_func {
@@ -578,6 +582,7 @@ struct pp_hwmgr {
 	struct pp_smumgr *smumgr;
 	const void *soft_pp_table;
 	uint32_t soft_pp_table_size;
+	void *hardcode_pp_table;
 	bool need_pp_table_upload;
 	enum amd_dpm_forced_level dpm_level;
 	bool block_hw_access;
@@ -607,6 +612,7 @@ struct pp_hwmgr {
 	uint32_t num_ps;
 	struct pp_thermal_controller_info thermal_controller;
 	bool fan_ctrl_is_in_default_mode;
+	bool powercontainment_enabled;
 	uint32_t fan_ctrl_default_mode;
 	uint32_t tmin;
 	struct phm_microcode_version_info microcode_version_info;
diff --git a/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h b/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h
index 0c6a413eaa5b..d41d37ab5b7c 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h
@@ -27,6 +27,7 @@
 
 #pragma pack(push, 1)
 
+#define PPSMC_MSG_SetGBDroopSettings          ((uint16_t) 0x305)
 
 #define PPSMC_SWSTATE_FLAG_DC                           0x01
 #define PPSMC_SWSTATE_FLAG_UVD                          0x02
diff --git a/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h b/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h
index 3bd5e69b9045..3df5de2cdab0 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h
@@ -26,3 +26,4 @@ extern bool acpi_atcs_functions_supported(void *device,
 extern int acpi_pcie_perf_request(void *device,
 						uint8_t perf_req,
 						bool advertise);
+extern bool acpi_atcs_notify_pcie_device_ready(void *device);
diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu74.h b/drivers/gpu/drm/amd/powerplay/inc/smu74.h
index 1a12d85b8e97..fd10a9fa843d 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/smu74.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/smu74.h
@@ -34,6 +34,30 @@
 #define SMU__NUM_LCLK_DPM_LEVELS 8
 #define SMU__NUM_PCIE_DPM_LEVELS 8
 
+#define EXP_M1  35
+#define EXP_M2  92821
+#define EXP_B   66629747
+
+#define EXP_M1_1  365
+#define EXP_M2_1  658700
+#define EXP_B_1   305506134
+
+#define EXP_M1_2  189
+#define EXP_M2_2  379692
+#define EXP_B_2   194609469
+
+#define EXP_M1_3  99
+#define EXP_M2_3  217915
+#define EXP_B_3   122255994
+
+#define EXP_M1_4  51
+#define EXP_M2_4  122643
+#define EXP_B_4   74893384
+
+#define EXP_M1_5  423
+#define EXP_M2_5  1103326
+#define EXP_B_5   728122621
+
 enum SID_OPTION {
 	SID_OPTION_HI,
 	SID_OPTION_LO,
@@ -548,20 +572,20 @@ struct SMU74_Firmware_Header {
 	uint32_t CacConfigTable;
 	uint32_t CacStatusTable;
 
-
 	uint32_t mcRegisterTable;
 
-
 	uint32_t mcArbDramTimingTable;
 
-
-
-
 	uint32_t PmFuseTable;
 	uint32_t Globals;
 	uint32_t ClockStretcherTable;
 	uint32_t VftTable;
-	uint32_t Reserved[21];
+	uint32_t Reserved1;
+	uint32_t AvfsTable;
+	uint32_t AvfsCksOffGbvTable;
+	uint32_t AvfsMeanNSigma;
+	uint32_t AvfsSclkOffsetTable;
+	uint32_t Reserved[16];
 	uint32_t Signature;
 };
 
@@ -701,8 +725,6 @@ VR Config info is contained in dpmTable.VRConfig */
 struct SMU_ClockStretcherDataTableEntry {
 	uint8_t minVID;
 	uint8_t maxVID;
-
-
 	uint16_t setting;
 };
 typedef struct SMU_ClockStretcherDataTableEntry SMU_ClockStretcherDataTableEntry;
@@ -769,6 +791,43 @@ struct VFT_TABLE_t {
 typedef struct VFT_TABLE_t VFT_TABLE_t;
 
 
+/* Total margin, root mean square of Fmax + DC + Platform */
+struct AVFS_Margin_t {
+	VFT_CELL_t Cell[NUM_VFT_COLUMNS];
+};
+typedef struct AVFS_Margin_t AVFS_Margin_t;
+
+#define BTCGB_VDROOP_TABLE_MAX_ENTRIES 2
+#define AVFSGB_VDROOP_TABLE_MAX_ENTRIES 2
+
+struct GB_VDROOP_TABLE_t {
+	int32_t a0;
+	int32_t a1;
+	int32_t a2;
+	uint32_t spare;
+};
+typedef struct GB_VDROOP_TABLE_t GB_VDROOP_TABLE_t;
+
+struct AVFS_CksOff_Gbv_t {
+	VFT_CELL_t Cell[NUM_VFT_COLUMNS];
+};
+typedef struct AVFS_CksOff_Gbv_t AVFS_CksOff_Gbv_t;
+
+struct AVFS_meanNsigma_t {
+	uint32_t Aconstant[3];
+	uint16_t DC_tol_sigma;
+	uint16_t Platform_mean;
+	uint16_t Platform_sigma;
+	uint16_t PSM_Age_CompFactor;
+	uint8_t  Static_Voltage_Offset[NUM_VFT_COLUMNS];
+};
+typedef struct AVFS_meanNsigma_t AVFS_meanNsigma_t;
+
+struct AVFS_Sclk_Offset_t {
+	uint16_t Sclk_Offset[8];
+};
+typedef struct AVFS_Sclk_Offset_t AVFS_Sclk_Offset_t;
+
 #endif
 
 
diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h b/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h
index 0dfe82336dc7..b85ff5400e57 100644
--- a/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h
+++ b/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h
@@ -223,6 +223,16 @@ struct SMU74_Discrete_StateInfo {
 
 typedef struct SMU74_Discrete_StateInfo SMU74_Discrete_StateInfo;
 
+struct SMU_QuadraticCoeffs {
+	int32_t m1;
+	uint32_t b;
+
+	int16_t m2;
+	uint8_t m1_shift;
+	uint8_t m2_shift;
+};
+typedef struct SMU_QuadraticCoeffs SMU_QuadraticCoeffs;
+
 struct SMU74_Discrete_DpmTable {
 
 	SMU74_PIDController                  GraphicsPIDController;
@@ -258,7 +268,14 @@ struct SMU74_Discrete_DpmTable {
 	uint8_t                             ThermOutPolarity;
 	uint8_t                             ThermOutMode;
 	uint8_t                             BootPhases;
-	uint32_t                            Reserved[4];
+
+	uint8_t                             VRHotLevel;
+	uint8_t                             Reserved1[3];
+	uint16_t                            FanStartTemperature;
+	uint16_t                            FanStopTemperature;
+	uint16_t                            MaxVoltage;
+	uint16_t                            Reserved2;
+	uint32_t                            Reserved[1];
 
 	SMU74_Discrete_GraphicsLevel        GraphicsLevel[SMU74_MAX_LEVELS_GRAPHICS];
 	SMU74_Discrete_MemoryLevel          MemoryACPILevel;
@@ -347,6 +364,8 @@ struct SMU74_Discrete_DpmTable {
 
 	uint32_t                            CurrSclkPllRange;
 	sclkFcwRange_t                      SclkFcwRangeTable[NUM_SCLK_RANGE];
+	GB_VDROOP_TABLE_t                   BTCGB_VDROOP_TABLE[BTCGB_VDROOP_TABLE_MAX_ENTRIES];
+	SMU_QuadraticCoeffs                 AVFSGB_VDROOP_TABLE[AVFSGB_VDROOP_TABLE_MAX_ENTRIES];
 };
 
 typedef struct SMU74_Discrete_DpmTable SMU74_Discrete_DpmTable;
@@ -550,16 +569,6 @@ struct SMU7_AcpiScoreboard {
 
 typedef struct SMU7_AcpiScoreboard SMU7_AcpiScoreboard;
 
-struct SMU_QuadraticCoeffs {
-	int32_t m1;
-	uint32_t b;
-
-	int16_t m2;
-	uint8_t m1_shift;
-	uint8_t m2_shift;
-};
-typedef struct SMU_QuadraticCoeffs SMU_QuadraticCoeffs;
-
 struct SMU74_Discrete_PmFuses {
 	uint8_t BapmVddCVidHiSidd[8];
 	uint8_t BapmVddCVidLoSidd[8];
@@ -821,6 +830,17 @@ typedef struct SMU7_GfxCuPgScoreboard SMU7_GfxCuPgScoreboard;
 #define DB_PCC_SHIFT 26 
 #define DB_EDC_SHIFT 27
 
+#define BTCGB0_Vdroop_Enable_MASK  0x1
+#define BTCGB1_Vdroop_Enable_MASK  0x2
+#define AVFSGB0_Vdroop_Enable_MASK 0x4
+#define AVFSGB1_Vdroop_Enable_MASK 0x8
+
+#define BTCGB0_Vdroop_Enable_SHIFT  0
+#define BTCGB1_Vdroop_Enable_SHIFT  1
+#define AVFSGB0_Vdroop_Enable_SHIFT 2
+#define AVFSGB1_Vdroop_Enable_SHIFT 3
+
+
 #pragma pack(pop)
 
 
diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c
index 673a75c74e18..8e52a2e82db5 100644
--- a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c
+++ b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c
@@ -1006,10 +1006,16 @@ static int fiji_smu_init(struct pp_smumgr *smumgr)
 
 static int fiji_smu_fini(struct pp_smumgr *smumgr)
 {
+	struct fiji_smumgr *priv = (struct fiji_smumgr *)(smumgr->backend);
+
+	smu_free_memory(smumgr->device, (void *)priv->header_buffer.handle);
+
 	if (smumgr->backend) {
 		kfree(smumgr->backend);
 		smumgr->backend = NULL;
 	}
+
+	cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c
index de618ead9db8..5dba7c509710 100644
--- a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c
+++ b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c
@@ -52,19 +52,18 @@
 static const SMU74_Discrete_GraphicsLevel avfs_graphics_level_polaris10[8] = {
 	/*  Min      pcie   DeepSleep Activity  CgSpll      CgSpll    CcPwr  CcPwr  Sclk         Enabled      Enabled                       Voltage    Power */
 	/* Voltage, DpmLevel, DivId,  Level,  FuncCntl3,  FuncCntl4,  DynRm, DynRm1 Did, Padding,ForActivity, ForThrottle, UpHyst, DownHyst, DownHyst, Throttle */
-	{ 0x3c0fd047, 0x00, 0x03, 0x1e00, 0x00200410, 0x87020000, 0, 0, 0x16, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x30750000, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0xa00fd047, 0x01, 0x04, 0x1e00, 0x00800510, 0x87020000, 0, 0, 0x16, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x409c0000, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0x0410d047, 0x01, 0x00, 0x1e00, 0x00600410, 0x87020000, 0, 0, 0x0e, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x50c30000, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0x6810d047, 0x01, 0x00, 0x1e00, 0x00800410, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x60ea0000, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0xcc10d047, 0x01, 0x00, 0x1e00, 0x00e00410, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0xe8fd0000, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0x3011d047, 0x01, 0x00, 0x1e00, 0x00400510, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x70110100, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0x9411d047, 0x01, 0x00, 0x1e00, 0x00a00510, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0xf8240100, 0, 0, 0, 0, 0, 0, 0 } },
-	{ 0xf811d047, 0x01, 0x00, 0x1e00, 0x00000610, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x80380100, 0, 0, 0, 0, 0, 0, 0 } }
+	{ 0x100ea446, 0x00, 0x03, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x30750000, 0x3000, 0, 0x2600, 0, 0, 0x0004, 0x8f02, 0xffff, 0x2f00, 0x300e, 0x2700 } },
+	{ 0x400ea446, 0x01, 0x04, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x409c0000, 0x2000, 0, 0x1e00, 1, 1, 0x0004, 0x8300, 0xffff, 0x1f00, 0xcb5e, 0x1a00 } },
+	{ 0x740ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x50c30000, 0x2800, 0, 0x2000, 1, 1, 0x0004, 0x0c02, 0xffff, 0x2700, 0x6433, 0x2100 } },
+	{ 0xa40ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x60ea0000, 0x3000, 0, 0x2600, 1, 1, 0x0004, 0x8f02, 0xffff, 0x2f00, 0x300e, 0x2700 } },
+	{ 0xd80ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x70110100, 0x3800, 0, 0x2c00, 1, 1, 0x0004, 0x1203, 0xffff, 0x3600, 0xc9e2, 0x2e00 } },
+	{ 0x3c0fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x80380100, 0x2000, 0, 0x1e00, 2, 1, 0x0004, 0x8300, 0xffff, 0x1f00, 0xcb5e, 0x1a00 } },
+	{ 0x6c0fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x905f0100, 0x2400, 0, 0x1e00, 2, 1, 0x0004, 0x8901, 0xffff, 0x2300, 0x314c, 0x1d00 } },
+	{ 0xa00fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0xa0860100, 0x2800, 0, 0x2000, 2, 1, 0x0004, 0x0c02, 0xffff, 0x2700, 0x6433, 0x2100 } }
 };
 
 static const SMU74_Discrete_MemoryLevel avfs_memory_level_polaris10 =
-	{0x50140000, 0x50140000, 0x00320000, 0x00, 0x00,
-	 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x0000, 0x00, 0x00};
+	{0x100ea446, 0, 0x30750000, 0x01, 0x01, 0x01, 0x00, 0x00, 0x64, 0x00, 0x00, 0x1f00, 0x00, 0x00};
 
 /**
 * Set the address for reading/writing the SMC SRAM space.
@@ -219,6 +218,18 @@ bool polaris10_is_smc_ram_running(struct pp_smumgr *smumgr)
 	&& (0x20100 <= cgs_read_ind_register(smumgr->device, CGS_IND_REG__SMC, ixSMC_PC_C)));
 }
 
+static bool polaris10_is_hw_avfs_present(struct pp_smumgr *smumgr)
+{
+	uint32_t efuse;
+
+	efuse = cgs_read_ind_register(smumgr->device, CGS_IND_REG__SMC, ixSMU_EFUSE_0 + (49*4));
+	efuse &= 0x00000001;
+	if (efuse)
+		return true;
+
+	return false;
+}
+
 /**
 * Send a message to the SMC, and wait for its response.
 *
@@ -228,21 +239,27 @@ bool polaris10_is_smc_ram_running(struct pp_smumgr *smumgr)
 */
 int polaris10_send_msg_to_smc(struct pp_smumgr *smumgr, uint16_t msg)
 {
+	int ret;
+
 	if (!polaris10_is_smc_ram_running(smumgr))
 		return -1;
 
+
 	SMUM_WAIT_FIELD_UNEQUAL(smumgr, SMC_RESP_0, SMC_RESP, 0);
 
-	if (1 != SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP))
-		printk("Failed to send Previous Message.\n");
+	ret = SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP);
 
+	if (ret != 1)
+		printk("\n failed to send pre message %x ret is %d \n",  msg, ret);
 
 	cgs_write_register(smumgr->device, mmSMC_MESSAGE_0, msg);
 
 	SMUM_WAIT_FIELD_UNEQUAL(smumgr, SMC_RESP_0, SMC_RESP, 0);
 
-	if (1 != SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP))
-		printk("Failed to send Message.\n");
+	ret = SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP);
+
+	if (ret != 1)
+		printk("\n failed to send message %x ret is %d \n",  msg, ret);
 
 	return 0;
 }
@@ -469,6 +486,7 @@ int polaris10_smu_fini(struct pp_smumgr *smumgr)
 		kfree(smumgr->backend);
 		smumgr->backend = NULL;
 	}
+	cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU);
 	return 0;
 }
 
@@ -952,6 +970,11 @@ static int polaris10_smu_init(struct pp_smumgr *smumgr)
 		(cgs_handle_t)smu_data->smu_buffer.handle);
 		return -1;);
 
+	if (polaris10_is_hw_avfs_present(smumgr))
+		smu_data->avfs.avfs_btc_status = AVFS_BTC_BOOT;
+	else
+		smu_data->avfs.avfs_btc_status = AVFS_BTC_NOTSUPPORTED;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
index c483baf6b4fb..7723473e51a0 100644
--- a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
+++ b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
@@ -23,6 +23,7 @@
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/slab.h>
+#include <drm/amdgpu_drm.h>
 #include "pp_instance.h"
 #include "smumgr.h"
 #include "cgs_common.h"
@@ -52,10 +53,10 @@ int smum_init(struct amd_pp_init *pp_init, struct pp_instance *handle)
 	handle->smu_mgr = smumgr;
 
 	switch (smumgr->chip_family) {
-	case AMD_FAMILY_CZ:
+	case AMDGPU_FAMILY_CZ:
 		cz_smum_init(smumgr);
 		break;
-	case AMD_FAMILY_VI:
+	case AMDGPU_FAMILY_VI:
 		switch (smumgr->chip_id) {
 		case CHIP_TONGA:
 			tonga_smum_init(smumgr);
@@ -81,6 +82,7 @@ int smum_init(struct amd_pp_init *pp_init, struct pp_instance *handle)
 
 int smum_fini(struct pp_smumgr *smumgr)
 {
+	kfree(smumgr->device);
 	kfree(smumgr);
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c
index 32820b680d88..b22722eabafc 100644
--- a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c
+++ b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c
@@ -328,10 +328,17 @@ int tonga_write_smc_sram_dword(struct pp_smumgr *smumgr,
 
 static int tonga_smu_fini(struct pp_smumgr *smumgr)
 {
+	struct tonga_smumgr *priv = (struct tonga_smumgr *)(smumgr->backend);
+
+	smu_free_memory(smumgr->device, (void *)priv->smu_buffer.handle);
+	smu_free_memory(smumgr->device, (void *)priv->header_buffer.handle);
+
 	if (smumgr->backend != NULL) {
 		kfree(smumgr->backend);
 		smumgr->backend = NULL;
 	}
+
+	cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h b/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h
index c89dc777768f..b961a1c6caf3 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h
+++ b/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h
@@ -26,7 +26,7 @@ TRACE_EVENT(amd_sched_job,
 	    TP_fast_assign(
 			   __entry->entity = sched_job->s_entity;
 			   __entry->sched_job = sched_job;
-			   __entry->fence = &sched_job->s_fence->base;
+			   __entry->fence = &sched_job->s_fence->finished;
 			   __entry->name = sched_job->sched->name;
 			   __entry->job_count = kfifo_len(
 				   &sched_job->s_entity->job_queue) / sizeof(sched_job);
@@ -46,7 +46,7 @@ TRACE_EVENT(amd_sched_process_job,
 		    ),
 
 	    TP_fast_assign(
-		    __entry->fence = &fence->base;
+		    __entry->fence = &fence->finished;
 		    ),
 	    TP_printk("fence=%p signaled", __entry->fence)
 );
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
index c16248cee779..70ff09d10885 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
+++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
@@ -32,6 +32,7 @@
 
 static bool amd_sched_entity_is_ready(struct amd_sched_entity *entity);
 static void amd_sched_wakeup(struct amd_gpu_scheduler *sched);
+static void amd_sched_process_job(struct fence *f, struct fence_cb *cb);
 
 struct kmem_cache *sched_fence_slab;
 atomic_t sched_fence_slab_ref = ATOMIC_INIT(0);
@@ -140,7 +141,7 @@ int amd_sched_entity_init(struct amd_gpu_scheduler *sched,
 		return r;
 
 	atomic_set(&entity->fence_seq, 0);
-	entity->fence_context = fence_context_alloc(1);
+	entity->fence_context = fence_context_alloc(2);
 
 	return 0;
 }
@@ -251,17 +252,21 @@ static bool amd_sched_entity_add_dependency_cb(struct amd_sched_entity *entity)
 
 	s_fence = to_amd_sched_fence(fence);
 	if (s_fence && s_fence->sched == sched) {
-		/* Fence is from the same scheduler */
-		if (test_bit(AMD_SCHED_FENCE_SCHEDULED_BIT, &fence->flags)) {
-			/* Ignore it when it is already scheduled */
-			fence_put(entity->dependency);
-			return false;
-		}
 
-		/* Wait for fence to be scheduled */
-		entity->cb.func = amd_sched_entity_clear_dep;
-		list_add_tail(&entity->cb.node, &s_fence->scheduled_cb);
-		return true;
+		/*
+		 * Fence is from the same scheduler, only need to wait for
+		 * it to be scheduled
+		 */
+		fence = fence_get(&s_fence->scheduled);
+		fence_put(entity->dependency);
+		entity->dependency = fence;
+		if (!fence_add_callback(fence, &entity->cb,
+					amd_sched_entity_clear_dep))
+			return true;
+
+		/* Ignore it when it is already scheduled */
+		fence_put(fence);
+		return false;
 	}
 
 	if (!fence_add_callback(entity->dependency, &entity->cb,
@@ -319,46 +324,108 @@ static bool amd_sched_entity_in(struct amd_sched_job *sched_job)
 	return added;
 }
 
-static void amd_sched_free_job(struct fence *f, struct fence_cb *cb) {
-	struct amd_sched_job *job = container_of(cb, struct amd_sched_job, cb_free_job);
-	schedule_work(&job->work_free_job);
-}
-
 /* job_finish is called after hw fence signaled, and
  * the job had already been deleted from ring_mirror_list
  */
-void amd_sched_job_finish(struct amd_sched_job *s_job)
+static void amd_sched_job_finish(struct work_struct *work)
 {
-	struct amd_sched_job *next;
+	struct amd_sched_job *s_job = container_of(work, struct amd_sched_job,
+						   finish_work);
 	struct amd_gpu_scheduler *sched = s_job->sched;
 
+	/* remove job from ring_mirror_list */
+	spin_lock(&sched->job_list_lock);
+	list_del_init(&s_job->node);
 	if (sched->timeout != MAX_SCHEDULE_TIMEOUT) {
-		if (cancel_delayed_work(&s_job->work_tdr))
-			amd_sched_job_put(s_job);
+		struct amd_sched_job *next;
+
+		spin_unlock(&sched->job_list_lock);
+		cancel_delayed_work_sync(&s_job->work_tdr);
+		spin_lock(&sched->job_list_lock);
 
 		/* queue TDR for next job */
 		next = list_first_entry_or_null(&sched->ring_mirror_list,
 						struct amd_sched_job, node);
 
-		if (next) {
-			INIT_DELAYED_WORK(&next->work_tdr, s_job->timeout_callback);
-			amd_sched_job_get(next);
+		if (next)
 			schedule_delayed_work(&next->work_tdr, sched->timeout);
-		}
 	}
+	spin_unlock(&sched->job_list_lock);
+	sched->ops->free_job(s_job);
 }
 
-void amd_sched_job_begin(struct amd_sched_job *s_job)
+static void amd_sched_job_finish_cb(struct fence *f, struct fence_cb *cb)
+{
+	struct amd_sched_job *job = container_of(cb, struct amd_sched_job,
+						 finish_cb);
+	schedule_work(&job->finish_work);
+}
+
+static void amd_sched_job_begin(struct amd_sched_job *s_job)
 {
 	struct amd_gpu_scheduler *sched = s_job->sched;
 
+	spin_lock(&sched->job_list_lock);
+	list_add_tail(&s_job->node, &sched->ring_mirror_list);
 	if (sched->timeout != MAX_SCHEDULE_TIMEOUT &&
-		list_first_entry_or_null(&sched->ring_mirror_list, struct amd_sched_job, node) == s_job)
-	{
-		INIT_DELAYED_WORK(&s_job->work_tdr, s_job->timeout_callback);
-		amd_sched_job_get(s_job);
+	    list_first_entry_or_null(&sched->ring_mirror_list,
+				     struct amd_sched_job, node) == s_job)
+		schedule_delayed_work(&s_job->work_tdr, sched->timeout);
+	spin_unlock(&sched->job_list_lock);
+}
+
+static void amd_sched_job_timedout(struct work_struct *work)
+{
+	struct amd_sched_job *job = container_of(work, struct amd_sched_job,
+						 work_tdr.work);
+
+	job->sched->ops->timedout_job(job);
+}
+
+void amd_sched_hw_job_reset(struct amd_gpu_scheduler *sched)
+{
+	struct amd_sched_job *s_job;
+
+	spin_lock(&sched->job_list_lock);
+	list_for_each_entry_reverse(s_job, &sched->ring_mirror_list, node) {
+		if (fence_remove_callback(s_job->s_fence->parent, &s_job->s_fence->cb)) {
+			fence_put(s_job->s_fence->parent);
+			s_job->s_fence->parent = NULL;
+		}
+	}
+	spin_unlock(&sched->job_list_lock);
+}
+
+void amd_sched_job_recovery(struct amd_gpu_scheduler *sched)
+{
+	struct amd_sched_job *s_job;
+	int r;
+
+	spin_lock(&sched->job_list_lock);
+	s_job = list_first_entry_or_null(&sched->ring_mirror_list,
+					 struct amd_sched_job, node);
+	if (s_job)
 		schedule_delayed_work(&s_job->work_tdr, sched->timeout);
+
+	list_for_each_entry(s_job, &sched->ring_mirror_list, node) {
+		struct amd_sched_fence *s_fence = s_job->s_fence;
+		struct fence *fence = sched->ops->run_job(s_job);
+		if (fence) {
+			s_fence->parent = fence_get(fence);
+			r = fence_add_callback(fence, &s_fence->cb,
+					       amd_sched_process_job);
+			if (r == -ENOENT)
+				amd_sched_process_job(fence, &s_fence->cb);
+			else if (r)
+				DRM_ERROR("fence add callback failed (%d)\n",
+					  r);
+			fence_put(fence);
+		} else {
+			DRM_ERROR("Failed to run job!\n");
+			amd_sched_process_job(NULL, &s_fence->cb);
+		}
 	}
+	spin_unlock(&sched->job_list_lock);
 }
 
 /**
@@ -372,36 +439,29 @@ void amd_sched_entity_push_job(struct amd_sched_job *sched_job)
 {
 	struct amd_sched_entity *entity = sched_job->s_entity;
 
-	sched_job->use_sched = 1;
-	fence_add_callback(&sched_job->s_fence->base,
-					&sched_job->cb_free_job, amd_sched_free_job);
 	trace_amd_sched_job(sched_job);
+	fence_add_callback(&sched_job->s_fence->finished, &sched_job->finish_cb,
+			   amd_sched_job_finish_cb);
 	wait_event(entity->sched->job_scheduled,
 		   amd_sched_entity_in(sched_job));
 }
 
 /* init a sched_job with basic field */
 int amd_sched_job_init(struct amd_sched_job *job,
-						struct amd_gpu_scheduler *sched,
-						struct amd_sched_entity *entity,
-						void (*timeout_cb)(struct work_struct *work),
-						void (*free_cb)(struct kref *refcount),
-						void *owner, struct fence **fence)
+		       struct amd_gpu_scheduler *sched,
+		       struct amd_sched_entity *entity,
+		       void *owner)
 {
-	INIT_LIST_HEAD(&job->node);
-	kref_init(&job->refcount);
 	job->sched = sched;
 	job->s_entity = entity;
 	job->s_fence = amd_sched_fence_create(entity, owner);
 	if (!job->s_fence)
 		return -ENOMEM;
 
-	job->s_fence->s_job = job;
-	job->timeout_callback = timeout_cb;
-	job->free_callback = free_cb;
+	INIT_WORK(&job->finish_work, amd_sched_job_finish);
+	INIT_LIST_HEAD(&job->node);
+	INIT_DELAYED_WORK(&job->work_tdr, amd_sched_job_timedout);
 
-	if (fence)
-		*fence = &job->s_fence->base;
 	return 0;
 }
 
@@ -450,23 +510,25 @@ static void amd_sched_process_job(struct fence *f, struct fence_cb *cb)
 	struct amd_sched_fence *s_fence =
 		container_of(cb, struct amd_sched_fence, cb);
 	struct amd_gpu_scheduler *sched = s_fence->sched;
-	unsigned long flags;
 
 	atomic_dec(&sched->hw_rq_count);
-
-	/* remove job from ring_mirror_list */
-	spin_lock_irqsave(&sched->job_list_lock, flags);
-	list_del_init(&s_fence->s_job->node);
-	sched->ops->finish_job(s_fence->s_job);
-	spin_unlock_irqrestore(&sched->job_list_lock, flags);
-
-	amd_sched_fence_signal(s_fence);
+	amd_sched_fence_finished(s_fence);
 
 	trace_amd_sched_process_job(s_fence);
-	fence_put(&s_fence->base);
+	fence_put(&s_fence->finished);
 	wake_up_interruptible(&sched->wake_up_worker);
 }
 
+static bool amd_sched_blocked(struct amd_gpu_scheduler *sched)
+{
+	if (kthread_should_park()) {
+		kthread_parkme();
+		return true;
+	}
+
+	return false;
+}
+
 static int amd_sched_main(void *param)
 {
 	struct sched_param sparam = {.sched_priority = 1};
@@ -476,14 +538,15 @@ static int amd_sched_main(void *param)
 	sched_setscheduler(current, SCHED_FIFO, &sparam);
 
 	while (!kthread_should_stop()) {
-		struct amd_sched_entity *entity;
+		struct amd_sched_entity *entity = NULL;
 		struct amd_sched_fence *s_fence;
 		struct amd_sched_job *sched_job;
 		struct fence *fence;
 
 		wait_event_interruptible(sched->wake_up_worker,
-			(entity = amd_sched_select_entity(sched)) ||
-			kthread_should_stop());
+					 (!amd_sched_blocked(sched) &&
+					  (entity = amd_sched_select_entity(sched))) ||
+					 kthread_should_stop());
 
 		if (!entity)
 			continue;
@@ -495,16 +558,19 @@ static int amd_sched_main(void *param)
 		s_fence = sched_job->s_fence;
 
 		atomic_inc(&sched->hw_rq_count);
-		amd_sched_job_pre_schedule(sched, sched_job);
+		amd_sched_job_begin(sched_job);
+
 		fence = sched->ops->run_job(sched_job);
 		amd_sched_fence_scheduled(s_fence);
 		if (fence) {
+			s_fence->parent = fence_get(fence);
 			r = fence_add_callback(fence, &s_fence->cb,
 					       amd_sched_process_job);
 			if (r == -ENOENT)
 				amd_sched_process_job(fence, &s_fence->cb);
 			else if (r)
-				DRM_ERROR("fence add callback failed (%d)\n", r);
+				DRM_ERROR("fence add callback failed (%d)\n",
+					  r);
 			fence_put(fence);
 		} else {
 			DRM_ERROR("Failed to run job!\n");
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
index 070095a9433c..7cbbbfb502ef 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
+++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
@@ -27,8 +27,6 @@
 #include <linux/kfifo.h>
 #include <linux/fence.h>
 
-#define AMD_SCHED_FENCE_SCHEDULED_BIT	FENCE_FLAG_USER_BITS
-
 struct amd_gpu_scheduler;
 struct amd_sched_rq;
 
@@ -68,36 +66,34 @@ struct amd_sched_rq {
 };
 
 struct amd_sched_fence {
-	struct fence                    base;
+	struct fence                    scheduled;
+	struct fence                    finished;
 	struct fence_cb                 cb;
-	struct list_head		scheduled_cb;
+	struct fence                    *parent;
 	struct amd_gpu_scheduler	*sched;
 	spinlock_t			lock;
 	void                            *owner;
-	struct amd_sched_job	*s_job;
 };
 
 struct amd_sched_job {
-	struct kref refcount;
 	struct amd_gpu_scheduler        *sched;
 	struct amd_sched_entity         *s_entity;
 	struct amd_sched_fence          *s_fence;
-	bool	use_sched;	/* true if the job goes to scheduler */
-	struct fence_cb                cb_free_job;
-	struct work_struct             work_free_job;
-	struct list_head			   node;
-	struct delayed_work work_tdr;
-	void (*timeout_callback) (struct work_struct *work);
-	void (*free_callback)(struct kref *refcount);
+	struct fence_cb			finish_cb;
+	struct work_struct		finish_work;
+	struct list_head		node;
+	struct delayed_work		work_tdr;
 };
 
-extern const struct fence_ops amd_sched_fence_ops;
+extern const struct fence_ops amd_sched_fence_ops_scheduled;
+extern const struct fence_ops amd_sched_fence_ops_finished;
 static inline struct amd_sched_fence *to_amd_sched_fence(struct fence *f)
 {
-	struct amd_sched_fence *__f = container_of(f, struct amd_sched_fence, base);
+	if (f->ops == &amd_sched_fence_ops_scheduled)
+		return container_of(f, struct amd_sched_fence, scheduled);
 
-	if (__f->base.ops == &amd_sched_fence_ops)
-		return __f;
+	if (f->ops == &amd_sched_fence_ops_finished)
+		return container_of(f, struct amd_sched_fence, finished);
 
 	return NULL;
 }
@@ -109,8 +105,8 @@ static inline struct amd_sched_fence *to_amd_sched_fence(struct fence *f)
 struct amd_sched_backend_ops {
 	struct fence *(*dependency)(struct amd_sched_job *sched_job);
 	struct fence *(*run_job)(struct amd_sched_job *sched_job);
-	void (*begin_job)(struct amd_sched_job *sched_job);
-	void (*finish_job)(struct amd_sched_job *sched_job);
+	void (*timedout_job)(struct amd_sched_job *sched_job);
+	void (*free_job)(struct amd_sched_job *sched_job);
 };
 
 enum amd_sched_priority {
@@ -152,25 +148,11 @@ void amd_sched_entity_push_job(struct amd_sched_job *sched_job);
 struct amd_sched_fence *amd_sched_fence_create(
 	struct amd_sched_entity *s_entity, void *owner);
 void amd_sched_fence_scheduled(struct amd_sched_fence *fence);
-void amd_sched_fence_signal(struct amd_sched_fence *fence);
+void amd_sched_fence_finished(struct amd_sched_fence *fence);
 int amd_sched_job_init(struct amd_sched_job *job,
-					struct amd_gpu_scheduler *sched,
-					struct amd_sched_entity *entity,
-					void (*timeout_cb)(struct work_struct *work),
-					void (*free_cb)(struct kref* refcount),
-					void *owner, struct fence **fence);
-void amd_sched_job_pre_schedule(struct amd_gpu_scheduler *sched ,
-								struct amd_sched_job *s_job);
-void amd_sched_job_finish(struct amd_sched_job *s_job);
-void amd_sched_job_begin(struct amd_sched_job *s_job);
-static inline void amd_sched_job_get(struct amd_sched_job *job) {
-	if (job)
-		kref_get(&job->refcount);
-}
-
-static inline void amd_sched_job_put(struct amd_sched_job *job) {
-	if (job)
-		kref_put(&job->refcount, job->free_callback);
-}
-
+		       struct amd_gpu_scheduler *sched,
+		       struct amd_sched_entity *entity,
+		       void *owner);
+void amd_sched_hw_job_reset(struct amd_gpu_scheduler *sched);
+void amd_sched_job_recovery(struct amd_gpu_scheduler *sched);
 #endif
diff --git a/drivers/gpu/drm/amd/scheduler/sched_fence.c b/drivers/gpu/drm/amd/scheduler/sched_fence.c
index 2a732c490375..6b63beaf7574 100644
--- a/drivers/gpu/drm/amd/scheduler/sched_fence.c
+++ b/drivers/gpu/drm/amd/scheduler/sched_fence.c
@@ -27,7 +27,8 @@
 #include <drm/drmP.h>
 #include "gpu_scheduler.h"
 
-struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *s_entity, void *owner)
+struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *entity,
+					       void *owner)
 {
 	struct amd_sched_fence *fence = NULL;
 	unsigned seq;
@@ -36,46 +37,37 @@ struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *s_entity
 	if (fence == NULL)
 		return NULL;
 
-	INIT_LIST_HEAD(&fence->scheduled_cb);
 	fence->owner = owner;
-	fence->sched = s_entity->sched;
+	fence->sched = entity->sched;
 	spin_lock_init(&fence->lock);
 
-	seq = atomic_inc_return(&s_entity->fence_seq);
-	fence_init(&fence->base, &amd_sched_fence_ops, &fence->lock,
-		   s_entity->fence_context, seq);
+	seq = atomic_inc_return(&entity->fence_seq);
+	fence_init(&fence->scheduled, &amd_sched_fence_ops_scheduled,
+		   &fence->lock, entity->fence_context, seq);
+	fence_init(&fence->finished, &amd_sched_fence_ops_finished,
+		   &fence->lock, entity->fence_context + 1, seq);
 
 	return fence;
 }
 
-void amd_sched_fence_signal(struct amd_sched_fence *fence)
+void amd_sched_fence_scheduled(struct amd_sched_fence *fence)
 {
-	int ret = fence_signal(&fence->base);
+	int ret = fence_signal(&fence->scheduled);
+
 	if (!ret)
-		FENCE_TRACE(&fence->base, "signaled from irq context\n");
+		FENCE_TRACE(&fence->scheduled, "signaled from irq context\n");
 	else
-		FENCE_TRACE(&fence->base, "was already signaled\n");
-}
-
-void amd_sched_job_pre_schedule(struct amd_gpu_scheduler *sched ,
-				struct amd_sched_job *s_job)
-{
-	unsigned long flags;
-	spin_lock_irqsave(&sched->job_list_lock, flags);
-	list_add_tail(&s_job->node, &sched->ring_mirror_list);
-	sched->ops->begin_job(s_job);
-	spin_unlock_irqrestore(&sched->job_list_lock, flags);
+		FENCE_TRACE(&fence->scheduled, "was already signaled\n");
 }
 
-void amd_sched_fence_scheduled(struct amd_sched_fence *s_fence)
+void amd_sched_fence_finished(struct amd_sched_fence *fence)
 {
-	struct fence_cb *cur, *tmp;
+	int ret = fence_signal(&fence->finished);
 
-	set_bit(AMD_SCHED_FENCE_SCHEDULED_BIT, &s_fence->base.flags);
-	list_for_each_entry_safe(cur, tmp, &s_fence->scheduled_cb, node) {
-		list_del_init(&cur->node);
-		cur->func(&s_fence->base, cur);
-	}
+	if (!ret)
+		FENCE_TRACE(&fence->finished, "signaled from irq context\n");
+	else
+		FENCE_TRACE(&fence->finished, "was already signaled\n");
 }
 
 static const char *amd_sched_fence_get_driver_name(struct fence *fence)
@@ -105,6 +97,8 @@ static void amd_sched_fence_free(struct rcu_head *rcu)
 {
 	struct fence *f = container_of(rcu, struct fence, rcu);
 	struct amd_sched_fence *fence = to_amd_sched_fence(f);
+
+	fence_put(fence->parent);
 	kmem_cache_free(sched_fence_slab, fence);
 }
 
@@ -116,16 +110,41 @@ static void amd_sched_fence_free(struct rcu_head *rcu)
  * This function is called when the reference count becomes zero.
  * It just RCU schedules freeing up the fence.
  */
-static void amd_sched_fence_release(struct fence *f)
+static void amd_sched_fence_release_scheduled(struct fence *f)
 {
-	call_rcu(&f->rcu, amd_sched_fence_free);
+	struct amd_sched_fence *fence = to_amd_sched_fence(f);
+
+	call_rcu(&fence->finished.rcu, amd_sched_fence_free);
 }
 
-const struct fence_ops amd_sched_fence_ops = {
+/**
+ * amd_sched_fence_release_scheduled - drop extra reference
+ *
+ * @f: fence
+ *
+ * Drop the extra reference from the scheduled fence to the base fence.
+ */
+static void amd_sched_fence_release_finished(struct fence *f)
+{
+	struct amd_sched_fence *fence = to_amd_sched_fence(f);
+
+	fence_put(&fence->scheduled);
+}
+
+const struct fence_ops amd_sched_fence_ops_scheduled = {
+	.get_driver_name = amd_sched_fence_get_driver_name,
+	.get_timeline_name = amd_sched_fence_get_timeline_name,
+	.enable_signaling = amd_sched_fence_enable_signaling,
+	.signaled = NULL,
+	.wait = fence_default_wait,
+	.release = amd_sched_fence_release_scheduled,
+};
+
+const struct fence_ops amd_sched_fence_ops_finished = {
 	.get_driver_name = amd_sched_fence_get_driver_name,
 	.get_timeline_name = amd_sched_fence_get_timeline_name,
 	.enable_signaling = amd_sched_fence_enable_signaling,
 	.signaled = NULL,
 	.wait = fence_default_wait,
-	.release = amd_sched_fence_release,
+	.release = amd_sched_fence_release_finished,
 };
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 59f2f93b6f84..b29a41218fc9 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -186,17 +186,6 @@ static void ast_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *
 {
 }
 
-static int ast_bo_move(struct ttm_buffer_object *bo,
-		       bool evict, bool interruptible,
-		       bool no_wait_gpu,
-		       struct ttm_mem_reg *new_mem)
-{
-	int r;
-	r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
-	return r;
-}
-
-
 static void ast_ttm_backend_destroy(struct ttm_tt *tt)
 {
 	ttm_tt_fini(tt);
@@ -241,7 +230,7 @@ struct ttm_bo_driver ast_bo_driver = {
 	.ttm_tt_unpopulate = ast_ttm_tt_unpopulate,
 	.init_mem_type = ast_bo_init_mem_type,
 	.evict_flags = ast_bo_evict_flags,
-	.move = ast_bo_move,
+	.move = NULL,
 	.verify_access = ast_bo_verify_access,
 	.io_mem_reserve = &ast_ttm_io_mem_reserve,
 	.io_mem_free = &ast_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c
index 473a475f27b1..6119b5085501 100644
--- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c
+++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c
@@ -254,9 +254,10 @@ int atmel_hlcdc_create_outputs(struct drm_device *dev)
 		if (!ret)
 			ret = atmel_hlcdc_check_endpoint(dev, &ep);
 
-		of_node_put(ep_np);
-		if (ret)
+		if (ret) {
+			of_node_put(ep_np);
 			return ret;
+		}
 	}
 
 	for_each_endpoint_of_node(dev->dev->of_node, ep_np) {
@@ -264,9 +265,10 @@ int atmel_hlcdc_create_outputs(struct drm_device *dev)
 		if (!ret)
 			ret = atmel_hlcdc_attach_endpoint(dev, &ep);
 
-		of_node_put(ep_np);
-		if (ret)
+		if (ret) {
+			of_node_put(ep_np);
 			return ret;
+		}
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c
index aef3ca8a81fa..016c191221f3 100644
--- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c
+++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c
@@ -339,6 +339,8 @@ atmel_hlcdc_plane_update_pos_and_size(struct atmel_hlcdc_plane *plane,
 
 		atmel_hlcdc_layer_update_cfg(&plane->layer, 13, 0xffffffff,
 					     factor_reg);
+	} else {
+		atmel_hlcdc_layer_update_cfg(&plane->layer, 13, 0xffffffff, 0);
 	}
 }
 
diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c
index 28a60a7dc2f8..5c5638a777a1 100644
--- a/drivers/gpu/drm/bochs/bochs_mm.c
+++ b/drivers/gpu/drm/bochs/bochs_mm.c
@@ -165,15 +165,6 @@ static void bochs_ttm_io_mem_free(struct ttm_bo_device *bdev,
 {
 }
 
-static int bochs_bo_move(struct ttm_buffer_object *bo,
-			 bool evict, bool interruptible,
-			 bool no_wait_gpu,
-			 struct ttm_mem_reg *new_mem)
-{
-	return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
-}
-
-
 static void bochs_ttm_backend_destroy(struct ttm_tt *tt)
 {
 	ttm_tt_fini(tt);
@@ -208,7 +199,7 @@ struct ttm_bo_driver bochs_bo_driver = {
 	.ttm_tt_unpopulate = ttm_pool_unpopulate,
 	.init_mem_type = bochs_bo_init_mem_type,
 	.evict_flags = bochs_bo_evict_flags,
-	.move = bochs_bo_move,
+	.move = NULL,
 	.verify_access = bochs_bo_verify_access,
 	.io_mem_reserve = &bochs_ttm_io_mem_reserve,
 	.io_mem_free = &bochs_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
index 7699597070a1..32715daf73cb 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
@@ -938,7 +938,7 @@ int analogix_dp_get_modes(struct drm_connector *connector)
 		num_modes += drm_panel_get_modes(dp->plat_data->panel);
 
 	if (dp->plat_data->get_modes)
-		num_modes += dp->plat_data->get_modes(dp->plat_data);
+		num_modes += dp->plat_data->get_modes(dp->plat_data, connector);
 
 	return num_modes;
 }
@@ -1208,6 +1208,7 @@ static int analogix_dp_dt_parse_pdata(struct analogix_dp_device *dp)
 
 	switch (dp->plat_data->dev_type) {
 	case RK3288_DP:
+	case RK3399_EDP:
 		/*
 		 * Like Rk3288 DisplayPort TRM indicate that "Main link
 		 * containing 4 physical lanes of 2.7/1.62 Gbps/lane".
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h
index f09275d40f70..b45638043ec4 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h
@@ -127,10 +127,10 @@ enum analog_power_block {
 };
 
 enum dp_irq_type {
-	DP_IRQ_TYPE_HP_CABLE_IN,
-	DP_IRQ_TYPE_HP_CABLE_OUT,
-	DP_IRQ_TYPE_HP_CHANGE,
-	DP_IRQ_TYPE_UNKNOWN,
+	DP_IRQ_TYPE_HP_CABLE_IN  = BIT(0),
+	DP_IRQ_TYPE_HP_CABLE_OUT = BIT(1),
+	DP_IRQ_TYPE_HP_CHANGE    = BIT(2),
+	DP_IRQ_TYPE_UNKNOWN      = BIT(3),
 };
 
 struct video_info {
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
index 49205ef02be3..48030f0cf497 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
@@ -74,8 +74,12 @@ void analogix_dp_init_analog_param(struct analogix_dp_device *dp)
 	reg = SEL_24M | TX_DVDD_BIT_1_0625V;
 	writel(reg, dp->reg_base + ANALOGIX_DP_ANALOG_CTL_2);
 
-	if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP)) {
-		writel(REF_CLK_24M, dp->reg_base + ANALOGIX_DP_PLL_REG_1);
+	if (dp->plat_data && is_rockchip(dp->plat_data->dev_type)) {
+		reg = REF_CLK_24M;
+		if (dp->plat_data->dev_type == RK3288_DP)
+			reg ^= REF_CLK_MASK;
+
+		writel(reg, dp->reg_base + ANALOGIX_DP_PLL_REG_1);
 		writel(0x95, dp->reg_base + ANALOGIX_DP_PLL_REG_2);
 		writel(0x40, dp->reg_base + ANALOGIX_DP_PLL_REG_3);
 		writel(0x58, dp->reg_base + ANALOGIX_DP_PLL_REG_4);
@@ -244,7 +248,7 @@ void analogix_dp_set_analog_power_down(struct analogix_dp_device *dp,
 	u32 reg;
 	u32 phy_pd_addr = ANALOGIX_DP_PHY_PD;
 
-	if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP))
+	if (dp->plat_data && is_rockchip(dp->plat_data->dev_type))
 		phy_pd_addr = ANALOGIX_DP_PD;
 
 	switch (block) {
@@ -448,7 +452,7 @@ void analogix_dp_init_aux(struct analogix_dp_device *dp)
 	analogix_dp_reset_aux(dp);
 
 	/* Disable AUX transaction H/W retry */
-	if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP))
+	if (dp->plat_data && is_rockchip(dp->plat_data->dev_type))
 		reg = AUX_BIT_PERIOD_EXPECTED_DELAY(0) |
 		      AUX_HW_RETRY_COUNT_SEL(3) |
 		      AUX_HW_RETRY_INTERVAL_600_MICROSECONDS;
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h
index 337912b0aeab..cdcc6c5add5e 100644
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h
+++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h
@@ -163,8 +163,9 @@
 #define HSYNC_POLARITY_CFG			(0x1 << 0)
 
 /* ANALOGIX_DP_PLL_REG_1 */
-#define REF_CLK_24M				(0x1 << 1)
-#define REF_CLK_27M				(0x0 << 1)
+#define REF_CLK_24M				(0x1 << 0)
+#define REF_CLK_27M				(0x0 << 0)
+#define REF_CLK_MASK				(0x1 << 0)
 
 /* ANALOGIX_DP_LANE_MAP */
 #define LANE3_MAP_LOGIC_LANE_0			(0x0 << 6)
diff --git a/drivers/gpu/drm/bridge/dw-hdmi.c b/drivers/gpu/drm/bridge/dw-hdmi.c
index 70b1f7d4270b..77ab47341658 100644
--- a/drivers/gpu/drm/bridge/dw-hdmi.c
+++ b/drivers/gpu/drm/bridge/dw-hdmi.c
@@ -1495,14 +1495,6 @@ static void dw_hdmi_connector_force(struct drm_connector *connector)
 }
 
 static const struct drm_connector_funcs dw_hdmi_connector_funcs = {
-	.dpms = drm_helper_connector_dpms,
-	.fill_modes = drm_helper_probe_single_connector_modes,
-	.detect = dw_hdmi_connector_detect,
-	.destroy = dw_hdmi_connector_destroy,
-	.force = dw_hdmi_connector_force,
-};
-
-static const struct drm_connector_funcs dw_hdmi_atomic_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.detect = dw_hdmi_connector_detect,
@@ -1634,14 +1626,9 @@ static int dw_hdmi_register(struct drm_device *drm, struct dw_hdmi *hdmi)
 	drm_connector_helper_add(&hdmi->connector,
 				 &dw_hdmi_connector_helper_funcs);
 
-	if (drm_core_check_feature(drm, DRIVER_ATOMIC))
-		drm_connector_init(drm, &hdmi->connector,
-				   &dw_hdmi_atomic_connector_funcs,
-				   DRM_MODE_CONNECTOR_HDMIA);
-	else
-		drm_connector_init(drm, &hdmi->connector,
-				   &dw_hdmi_connector_funcs,
-				   DRM_MODE_CONNECTOR_HDMIA);
+	drm_connector_init(drm, &hdmi->connector,
+			   &dw_hdmi_connector_funcs,
+			   DRM_MODE_CONNECTOR_HDMIA);
 
 	drm_mode_connector_attach_encoder(&hdmi->connector, encoder);
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 6768b7b1af32..1cc9ee607128 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -186,17 +186,6 @@ static void cirrus_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_re
 {
 }
 
-static int cirrus_bo_move(struct ttm_buffer_object *bo,
-		       bool evict, bool interruptible,
-		       bool no_wait_gpu,
-		       struct ttm_mem_reg *new_mem)
-{
-	int r;
-	r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
-	return r;
-}
-
-
 static void cirrus_ttm_backend_destroy(struct ttm_tt *tt)
 {
 	ttm_tt_fini(tt);
@@ -241,7 +230,7 @@ struct ttm_bo_driver cirrus_bo_driver = {
 	.ttm_tt_unpopulate = cirrus_ttm_tt_unpopulate,
 	.init_mem_type = cirrus_bo_init_mem_type,
 	.evict_flags = cirrus_bo_evict_flags,
-	.move = cirrus_bo_move,
+	.move = NULL,
 	.verify_access = cirrus_bo_verify_access,
 	.io_mem_reserve = &cirrus_ttm_io_mem_reserve,
 	.io_mem_free = &cirrus_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index d99ab2f6663f..3cee084e9d28 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -1299,14 +1299,39 @@ EXPORT_SYMBOL(drm_atomic_add_affected_planes);
  */
 void drm_atomic_legacy_backoff(struct drm_atomic_state *state)
 {
+	struct drm_device *dev = state->dev;
+	unsigned crtc_mask = 0;
+	struct drm_crtc *crtc;
 	int ret;
+	bool global = false;
+
+	drm_for_each_crtc(crtc, dev) {
+		if (crtc->acquire_ctx != state->acquire_ctx)
+			continue;
+
+		crtc_mask |= drm_crtc_mask(crtc);
+		crtc->acquire_ctx = NULL;
+	}
+
+	if (WARN_ON(dev->mode_config.acquire_ctx == state->acquire_ctx)) {
+		global = true;
+
+		dev->mode_config.acquire_ctx = NULL;
+	}
 
 retry:
 	drm_modeset_backoff(state->acquire_ctx);
 
-	ret = drm_modeset_lock_all_ctx(state->dev, state->acquire_ctx);
+	ret = drm_modeset_lock_all_ctx(dev, state->acquire_ctx);
 	if (ret)
 		goto retry;
+
+	drm_for_each_crtc(crtc, dev)
+		if (drm_crtc_mask(crtc) & crtc_mask)
+			crtc->acquire_ctx = state->acquire_ctx;
+
+	if (global)
+		dev->mode_config.acquire_ctx = state->acquire_ctx;
 }
 EXPORT_SYMBOL(drm_atomic_legacy_backoff);
 
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index d61591274ff6..604d3ef72ffa 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -531,11 +531,11 @@ drm_crtc_helper_disable(struct drm_crtc *crtc)
 int drm_crtc_helper_set_config(struct drm_mode_set *set)
 {
 	struct drm_device *dev;
-	struct drm_crtc *new_crtc;
-	struct drm_encoder *save_encoders, *new_encoder, *encoder;
+	struct drm_crtc **save_encoder_crtcs, *new_crtc;
+	struct drm_encoder **save_connector_encoders, *new_encoder, *encoder;
 	bool mode_changed = false; /* if true do a full mode set */
 	bool fb_changed = false; /* if true and !mode_changed just do a flip */
-	struct drm_connector *save_connectors, *connector;
+	struct drm_connector *connector;
 	int count = 0, ro, fail = 0;
 	const struct drm_crtc_helper_funcs *crtc_funcs;
 	struct drm_mode_set save_set;
@@ -577,15 +577,15 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 	 * Allocate space for the backup of all (non-pointer) encoder and
 	 * connector data.
 	 */
-	save_encoders = kzalloc(dev->mode_config.num_encoder *
-				sizeof(struct drm_encoder), GFP_KERNEL);
-	if (!save_encoders)
+	save_encoder_crtcs = kzalloc(dev->mode_config.num_encoder *
+				sizeof(struct drm_crtc *), GFP_KERNEL);
+	if (!save_encoder_crtcs)
 		return -ENOMEM;
 
-	save_connectors = kzalloc(dev->mode_config.num_connector *
-				sizeof(struct drm_connector), GFP_KERNEL);
-	if (!save_connectors) {
-		kfree(save_encoders);
+	save_connector_encoders = kzalloc(dev->mode_config.num_connector *
+				sizeof(struct drm_encoder *), GFP_KERNEL);
+	if (!save_connector_encoders) {
+		kfree(save_encoder_crtcs);
 		return -ENOMEM;
 	}
 
@@ -596,12 +596,12 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 	 */
 	count = 0;
 	drm_for_each_encoder(encoder, dev) {
-		save_encoders[count++] = *encoder;
+		save_encoder_crtcs[count++] = encoder->crtc;
 	}
 
 	count = 0;
 	drm_for_each_connector(connector, dev) {
-		save_connectors[count++] = *connector;
+		save_connector_encoders[count++] = connector->encoder;
 	}
 
 	save_set.crtc = set->crtc;
@@ -634,8 +634,12 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 		mode_changed = true;
 	}
 
-	/* take a reference on all connectors in set */
+	/* take a reference on all unbound connectors in set, reuse the
+	 * already taken reference for bound connectors
+	 */
 	for (ro = 0; ro < set->num_connectors; ro++) {
+		if (set->connectors[ro]->encoder)
+			continue;
 		drm_connector_reference(set->connectors[ro]);
 	}
 
@@ -757,30 +761,28 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 		}
 	}
 
-	/* after fail drop reference on all connectors in save set */
-	count = 0;
-	drm_for_each_connector(connector, dev) {
-		drm_connector_unreference(&save_connectors[count++]);
-	}
-
-	kfree(save_connectors);
-	kfree(save_encoders);
+	kfree(save_connector_encoders);
+	kfree(save_encoder_crtcs);
 	return 0;
 
 fail:
 	/* Restore all previous data. */
 	count = 0;
 	drm_for_each_encoder(encoder, dev) {
-		*encoder = save_encoders[count++];
+		encoder->crtc = save_encoder_crtcs[count++];
 	}
 
 	count = 0;
 	drm_for_each_connector(connector, dev) {
-		*connector = save_connectors[count++];
+		connector->encoder = save_connector_encoders[count++];
 	}
 
-	/* after fail drop reference on all connectors in set */
+	/* after fail drop reference on all unbound connectors in set, let
+	 * bound connectors keep their reference
+	 */
 	for (ro = 0; ro < set->num_connectors; ro++) {
+		if (set->connectors[ro]->encoder)
+			continue;
 		drm_connector_unreference(set->connectors[ro]);
 	}
 
@@ -790,8 +792,8 @@ fail:
 				      save_set.y, save_set.fb))
 		DRM_ERROR("failed to restore config after modeset failure\n");
 
-	kfree(save_connectors);
-	kfree(save_encoders);
+	kfree(save_connector_encoders);
+	kfree(save_encoder_crtcs);
 	return ret;
 }
 EXPORT_SYMBOL(drm_crtc_helper_set_config);
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index a13edf5de2d6..6537908050d7 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -2927,11 +2927,9 @@ static void drm_dp_destroy_connector_work(struct work_struct *work)
 		drm_dp_port_teardown_pdt(port, port->pdt);
 
 		if (!port->input && port->vcpi.vcpi > 0) {
-			if (mgr->mst_state) {
-				drm_dp_mst_reset_vcpi_slots(mgr, port);
-				drm_dp_update_payload_part1(mgr);
-				drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi);
-			}
+			drm_dp_mst_reset_vcpi_slots(mgr, port);
+			drm_dp_update_payload_part1(mgr);
+			drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi);
 		}
 
 		kref_put(&port->kref, drm_dp_free_mst_port);
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index c0b0c718994a..1fd6eac1400c 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -596,3 +596,18 @@ void drm_fbdev_cma_hotplug_event(struct drm_fbdev_cma *fbdev_cma)
 		drm_fb_helper_hotplug_event(&fbdev_cma->fb_helper);
 }
 EXPORT_SYMBOL_GPL(drm_fbdev_cma_hotplug_event);
+
+/**
+ * drm_fbdev_cma_set_suspend - wrapper around drm_fb_helper_set_suspend
+ * @fbdev_cma: The drm_fbdev_cma struct, may be NULL
+ * @state: desired state, zero to resume, non-zero to suspend
+ *
+ * Calls drm_fb_helper_set_suspend, which is a wrapper around
+ * fb_set_suspend implemented by fbdev core.
+ */
+void drm_fbdev_cma_set_suspend(struct drm_fbdev_cma *fbdev_cma, int state)
+{
+	if (fbdev_cma)
+		drm_fb_helper_set_suspend(&fbdev_cma->fb_helper, state);
+}
+EXPORT_SYMBOL(drm_fbdev_cma_set_suspend);
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.c b/drivers/gpu/drm/etnaviv/etnaviv_drv.c
index 340d390306d8..ffd1b32caa8d 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_drv.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.c
@@ -91,10 +91,8 @@ static void load_gpu(struct drm_device *dev)
 			int ret;
 
 			ret = etnaviv_gpu_init(g);
-			if (ret) {
-				dev_err(g->dev, "hw init failed: %d\n", ret);
+			if (ret)
 				priv->gpu[i] = NULL;
-			}
 		}
 	}
 }
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
index ff6aa5dfb2d7..87ef34150d46 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
@@ -487,6 +487,47 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 	return 0;
 }
 
+static void etnaviv_gpu_enable_mlcg(struct etnaviv_gpu *gpu)
+{
+	u32 pmc, ppc;
+
+	/* enable clock gating */
+	ppc = gpu_read(gpu, VIVS_PM_POWER_CONTROLS);
+	ppc |= VIVS_PM_POWER_CONTROLS_ENABLE_MODULE_CLOCK_GATING;
+
+	/* Disable stall module clock gating for 4.3.0.1 and 4.3.0.2 revs */
+	if (gpu->identity.revision == 0x4301 ||
+	    gpu->identity.revision == 0x4302)
+		ppc |= VIVS_PM_POWER_CONTROLS_DISABLE_STALL_MODULE_CLOCK_GATING;
+
+	gpu_write(gpu, VIVS_PM_POWER_CONTROLS, ppc);
+
+	pmc = gpu_read(gpu, VIVS_PM_MODULE_CONTROLS);
+
+	/* Disable PA clock gating for GC400+ except for GC420 */
+	if (gpu->identity.model >= chipModel_GC400 &&
+	    gpu->identity.model != chipModel_GC420)
+		pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PA;
+
+	/*
+	 * Disable PE clock gating on revs < 5.0.0.0 when HZ is
+	 * present without a bug fix.
+	 */
+	if (gpu->identity.revision < 0x5000 &&
+	    gpu->identity.minor_features0 & chipMinorFeatures0_HZ &&
+	    !(gpu->identity.minor_features1 &
+	      chipMinorFeatures1_DISABLE_PE_GATING))
+		pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PE;
+
+	if (gpu->identity.revision < 0x5422)
+		pmc |= BIT(15); /* Unknown bit */
+
+	pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_HZ;
+	pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_EZ;
+
+	gpu_write(gpu, VIVS_PM_MODULE_CONTROLS, pmc);
+}
+
 static void etnaviv_gpu_hw_init(struct etnaviv_gpu *gpu)
 {
 	u16 prefetch;
@@ -506,6 +547,9 @@ static void etnaviv_gpu_hw_init(struct etnaviv_gpu *gpu)
 		gpu_write(gpu, VIVS_MC_DEBUG_MEMORY, mc_memory_debug);
 	}
 
+	/* enable module-level clock gating */
+	etnaviv_gpu_enable_mlcg(gpu);
+
 	/*
 	 * Update GPU AXI cache atttribute to "cacheable, no allocate".
 	 * This is necessary to prevent the iMX6 SoC locking up.
@@ -553,8 +597,10 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 	bool mmuv2;
 
 	ret = pm_runtime_get_sync(gpu->dev);
-	if (ret < 0)
+	if (ret < 0) {
+		dev_err(gpu->dev, "Failed to enable GPU power domain\n");
 		return ret;
+	}
 
 	etnaviv_hw_identify(gpu);
 
@@ -591,8 +637,10 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 	}
 
 	ret = etnaviv_hw_reset(gpu);
-	if (ret)
+	if (ret) {
+		dev_err(gpu->dev, "GPU reset failed\n");
 		goto fail;
+	}
 
 	/* Setup IOMMU.. eventually we will (I think) do this once per context
 	 * and have separate page tables per context.  For now, to keep things
@@ -610,12 +658,14 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 	}
 
 	if (!iommu) {
+		dev_err(gpu->dev, "Failed to allocate GPU IOMMU domain\n");
 		ret = -ENOMEM;
 		goto fail;
 	}
 
 	gpu->mmu = etnaviv_iommu_new(gpu, iommu, version);
 	if (!gpu->mmu) {
+		dev_err(gpu->dev, "Failed to instantiate GPU IOMMU\n");
 		iommu_domain_free(iommu);
 		ret = -ENOMEM;
 		goto fail;
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
index 522cfd447892..16353ee81651 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
@@ -225,6 +225,7 @@ struct iommu_domain *etnaviv_iommu_domain_alloc(struct etnaviv_gpu *gpu)
 
 	etnaviv_domain->domain.type = __IOMMU_DOMAIN_PAGING;
 	etnaviv_domain->domain.ops = &etnaviv_iommu_ops.ops;
+	etnaviv_domain->domain.pgsize_bitmap = SZ_4K;
 	etnaviv_domain->domain.geometry.aperture_start = GPU_MEM_START;
 	etnaviv_domain->domain.geometry.aperture_end = GPU_MEM_START + PT_ENTRIES * SZ_4K - 1;
 
diff --git a/drivers/gpu/drm/etnaviv/state_hi.xml.h b/drivers/gpu/drm/etnaviv/state_hi.xml.h
index 6a7de5f1454a..807a3d9e0dd5 100644
--- a/drivers/gpu/drm/etnaviv/state_hi.xml.h
+++ b/drivers/gpu/drm/etnaviv/state_hi.xml.h
@@ -218,6 +218,13 @@ Copyright (C) 2015
 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_FE	0x00000001
 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_DE	0x00000002
 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PE	0x00000004
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_SH	0x00000008
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PA	0x00000010
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_SE	0x00000020
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA	0x00000040
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_TX	0x00000080
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_EZ	0x00010000
+#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_HZ	0x00020000
 
 #define VIVS_PM_MODULE_STATUS					0x00000108
 #define VIVS_PM_MODULE_STATUS_MODULE_CLOCK_GATED_FE		0x00000001
diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig
index 1c7e14cf2781..83f61c513b7e 100644
--- a/drivers/gpu/drm/exynos/Kconfig
+++ b/drivers/gpu/drm/exynos/Kconfig
@@ -11,7 +11,7 @@ if DRM_EXYNOS
 
 config DRM_EXYNOS_IOMMU
 	bool
-	depends on EXYNOS_IOMMU && ARM_DMA_USE_IOMMU
+	depends on EXYNOS_IOMMU
 	default y
 
 comment "CRTCs"
diff --git a/drivers/gpu/drm/exynos/exynos7_drm_decon.c b/drivers/gpu/drm/exynos/exynos7_drm_decon.c
index f6223f907c15..7f9901b7777b 100644
--- a/drivers/gpu/drm/exynos/exynos7_drm_decon.c
+++ b/drivers/gpu/drm/exynos/exynos7_drm_decon.c
@@ -31,7 +31,6 @@
 #include "exynos_drm_plane.h"
 #include "exynos_drm_drv.h"
 #include "exynos_drm_fb.h"
-#include "exynos_drm_fbdev.h"
 #include "exynos_drm_iommu.h"
 
 /*
diff --git a/drivers/gpu/drm/exynos/exynos_dp.c b/drivers/gpu/drm/exynos/exynos_dp.c
index 468498e3fec1..4f0850585b8e 100644
--- a/drivers/gpu/drm/exynos/exynos_dp.c
+++ b/drivers/gpu/drm/exynos/exynos_dp.c
@@ -34,7 +34,7 @@
 
 struct exynos_dp_device {
 	struct drm_encoder         encoder;
-	struct drm_connector       connector;
+	struct drm_connector       *connector;
 	struct drm_bridge          *ptn_bridge;
 	struct drm_device          *drm_dev;
 	struct device              *dev;
@@ -67,10 +67,10 @@ static int exynos_dp_poweroff(struct analogix_dp_plat_data *plat_data)
 	return exynos_dp_crtc_clock_enable(plat_data, false);
 }
 
-static int exynos_dp_get_modes(struct analogix_dp_plat_data *plat_data)
+static int exynos_dp_get_modes(struct analogix_dp_plat_data *plat_data,
+			       struct drm_connector *connector)
 {
 	struct exynos_dp_device *dp = to_dp(plat_data);
-	struct drm_connector *connector = &dp->connector;
 	struct drm_display_mode *mode;
 	int num_modes = 0;
 
@@ -103,6 +103,7 @@ static int exynos_dp_bridge_attach(struct analogix_dp_plat_data *plat_data,
 	int ret;
 
 	drm_connector_register(connector);
+	dp->connector = connector;
 
 	/* Pre-empt DP connector creation if there's a bridge */
 	if (dp->ptn_bridge) {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c
index 011211e4167d..edbd98ff293e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_core.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_core.c
@@ -15,7 +15,6 @@
 #include <drm/drmP.h>
 #include "exynos_drm_drv.h"
 #include "exynos_drm_crtc.h"
-#include "exynos_drm_fbdev.h"
 
 static LIST_HEAD(exynos_drm_subdrv_list);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index 13d28d4229e2..877d2efa28e2 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -159,12 +159,7 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 	DRM_INFO("Exynos DRM: using %s device for DMA mapping operations\n",
 		 dev_name(private->dma_dev));
 
-	/*
-	 * create mapping to manage iommu table and set a pointer to iommu
-	 * mapping structure to iommu_mapping of private data.
-	 * also this iommu_mapping can be used to check if iommu is supported
-	 * or not.
-	 */
+	/* create common IOMMU mapping for all devices attached to Exynos DRM */
 	ret = drm_create_iommu_mapping(dev);
 	if (ret < 0) {
 		DRM_ERROR("failed to create iommu mapping.\n");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index cc33ec9296e7..b39d521f093d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -224,8 +224,6 @@ struct exynos_drm_private {
 	struct drm_property *plane_zpos_property;
 
 	struct device *dma_dev;
-	unsigned long da_start;
-	unsigned long da_space_size;
 	void *mapping;
 
 	unsigned int pipe;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 3efe1aa89416..d47216488985 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -30,7 +30,6 @@
 
 #include "exynos_drm_drv.h"
 #include "exynos_drm_fb.h"
-#include "exynos_drm_fbdev.h"
 #include "exynos_drm_crtc.h"
 #include "exynos_drm_plane.h"
 #include "exynos_drm_iommu.h"
@@ -120,7 +119,6 @@ static struct fimd_driver_data s3c64xx_fimd_driver_data = {
 	.timing_base = 0x0,
 	.has_clksel = 1,
 	.has_limited_fmt = 1,
-	.has_hw_trigger = 1,
 };
 
 static struct fimd_driver_data exynos3_fimd_driver_data = {
@@ -171,14 +169,11 @@ static struct fimd_driver_data exynos5420_fimd_driver_data = {
 	.lcdblk_vt_shift = 24,
 	.lcdblk_bypass_shift = 15,
 	.lcdblk_mic_bypass_shift = 11,
-	.trg_type = I80_HW_TRG,
 	.has_shadowcon = 1,
 	.has_vidoutcon = 1,
 	.has_vtsel = 1,
 	.has_mic_bypass = 1,
 	.has_dp_clk = 1,
-	.has_hw_trigger = 1,
-	.has_trigger_per_te = 1,
 };
 
 struct fimd_context {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index 493552368295..8564c3da0d22 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -48,13 +48,13 @@
 
 /* registers for base address */
 #define G2D_SRC_BASE_ADDR		0x0304
-#define G2D_SRC_STRIDE_REG		0x0308
+#define G2D_SRC_STRIDE			0x0308
 #define G2D_SRC_COLOR_MODE		0x030C
 #define G2D_SRC_LEFT_TOP		0x0310
 #define G2D_SRC_RIGHT_BOTTOM		0x0314
 #define G2D_SRC_PLANE2_BASE_ADDR	0x0318
 #define G2D_DST_BASE_ADDR		0x0404
-#define G2D_DST_STRIDE_REG		0x0408
+#define G2D_DST_STRIDE			0x0408
 #define G2D_DST_COLOR_MODE		0x040C
 #define G2D_DST_LEFT_TOP		0x0410
 #define G2D_DST_RIGHT_BOTTOM		0x0414
@@ -563,7 +563,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset)
 
 	switch (reg_offset) {
 	case G2D_SRC_BASE_ADDR:
-	case G2D_SRC_STRIDE_REG:
+	case G2D_SRC_STRIDE:
 	case G2D_SRC_COLOR_MODE:
 	case G2D_SRC_LEFT_TOP:
 	case G2D_SRC_RIGHT_BOTTOM:
@@ -573,7 +573,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset)
 		reg_type = REG_TYPE_SRC_PLANE2;
 		break;
 	case G2D_DST_BASE_ADDR:
-	case G2D_DST_STRIDE_REG:
+	case G2D_DST_STRIDE:
 	case G2D_DST_COLOR_MODE:
 	case G2D_DST_LEFT_TOP:
 	case G2D_DST_RIGHT_BOTTOM:
@@ -968,8 +968,8 @@ static int g2d_check_reg_offset(struct device *dev,
 			} else
 				buf_info->types[reg_type] = BUF_TYPE_GEM;
 			break;
-		case G2D_SRC_STRIDE_REG:
-		case G2D_DST_STRIDE_REG:
+		case G2D_SRC_STRIDE:
+		case G2D_DST_STRIDE:
 			if (for_addr)
 				goto err;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.c b/drivers/gpu/drm/exynos/exynos_drm_iommu.c
index 7ca09ee19656..0f373702414e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_iommu.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.c
@@ -14,13 +14,27 @@
 
 #include <linux/dma-mapping.h>
 #include <linux/iommu.h>
-#include <linux/kref.h>
-
-#include <asm/dma-iommu.h>
 
 #include "exynos_drm_drv.h"
 #include "exynos_drm_iommu.h"
 
+static inline int configure_dma_max_seg_size(struct device *dev)
+{
+	if (!dev->dma_parms)
+		dev->dma_parms = kzalloc(sizeof(*dev->dma_parms), GFP_KERNEL);
+	if (!dev->dma_parms)
+		return -ENOMEM;
+
+	dma_set_max_seg_size(dev, DMA_BIT_MASK(32));
+	return 0;
+}
+
+static inline void clear_dma_max_seg_size(struct device *dev)
+{
+	kfree(dev->dma_parms);
+	dev->dma_parms = NULL;
+}
+
 /*
  * drm_create_iommu_mapping - create a mapping structure
  *
@@ -28,38 +42,22 @@
  */
 int drm_create_iommu_mapping(struct drm_device *drm_dev)
 {
-	struct dma_iommu_mapping *mapping = NULL;
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 
-	if (!priv->da_start)
-		priv->da_start = EXYNOS_DEV_ADDR_START;
-	if (!priv->da_space_size)
-		priv->da_space_size = EXYNOS_DEV_ADDR_SIZE;
-
-	mapping = arm_iommu_create_mapping(&platform_bus_type, priv->da_start,
-						priv->da_space_size);
-
-	if (IS_ERR(mapping))
-		return PTR_ERR(mapping);
-
-	priv->mapping = mapping;
-
-	return 0;
+	return __exynos_iommu_create_mapping(priv, EXYNOS_DEV_ADDR_START,
+					     EXYNOS_DEV_ADDR_SIZE);
 }
 
 /*
  * drm_release_iommu_mapping - release iommu mapping structure
  *
  * @drm_dev: DRM device
- *
- * if mapping->kref becomes 0 then all things related to iommu mapping
- * will be released
  */
 void drm_release_iommu_mapping(struct drm_device *drm_dev)
 {
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 
-	arm_iommu_release_mapping(priv->mapping);
+	__exynos_iommu_release_mapping(priv);
 }
 
 /*
@@ -77,25 +75,19 @@ int drm_iommu_attach_device(struct drm_device *drm_dev,
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 	int ret;
 
-	if (!priv->mapping)
-		return 0;
-
-	subdrv_dev->dma_parms = devm_kzalloc(subdrv_dev,
-					sizeof(*subdrv_dev->dma_parms),
-					GFP_KERNEL);
-	if (!subdrv_dev->dma_parms)
-		return -ENOMEM;
-
-	dma_set_max_seg_size(subdrv_dev, 0xffffffffu);
-
-	if (subdrv_dev->archdata.mapping)
-		arm_iommu_detach_device(subdrv_dev);
+	if (get_dma_ops(priv->dma_dev) != get_dma_ops(subdrv_dev)) {
+		DRM_ERROR("Device %s lacks support for IOMMU\n",
+			  dev_name(subdrv_dev));
+		return -EINVAL;
+	}
 
-	ret = arm_iommu_attach_device(subdrv_dev, priv->mapping);
-	if (ret < 0) {
-		DRM_DEBUG_KMS("failed iommu attach.\n");
+	ret = configure_dma_max_seg_size(subdrv_dev);
+	if (ret)
 		return ret;
-	}
+
+	ret = __exynos_iommu_attach(priv, subdrv_dev);
+	if (ret)
+		clear_dma_max_seg_size(subdrv_dev);
 
 	return 0;
 }
@@ -113,10 +105,7 @@ void drm_iommu_detach_device(struct drm_device *drm_dev,
 				struct device *subdrv_dev)
 {
 	struct exynos_drm_private *priv = drm_dev->dev_private;
-	struct dma_iommu_mapping *mapping = priv->mapping;
-
-	if (!mapping || !mapping->domain)
-		return;
 
-	arm_iommu_detach_device(subdrv_dev);
+	__exynos_iommu_detach(priv, subdrv_dev);
+	clear_dma_max_seg_size(subdrv_dev);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.h b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
index 5ffebe02ee4d..c8de4913fdbe 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_iommu.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
@@ -17,6 +17,97 @@
 
 #ifdef CONFIG_DRM_EXYNOS_IOMMU
 
+#if defined(CONFIG_ARM_DMA_USE_IOMMU)
+#include <asm/dma-iommu.h>
+
+static inline int __exynos_iommu_create_mapping(struct exynos_drm_private *priv,
+					unsigned long start, unsigned long size)
+{
+	priv->mapping = arm_iommu_create_mapping(&platform_bus_type, start,
+						 size);
+	return IS_ERR(priv->mapping);
+}
+
+static inline void
+__exynos_iommu_release_mapping(struct exynos_drm_private *priv)
+{
+	arm_iommu_release_mapping(priv->mapping);
+}
+
+static inline int __exynos_iommu_attach(struct exynos_drm_private *priv,
+					struct device *dev)
+{
+	if (dev->archdata.mapping)
+		arm_iommu_detach_device(dev);
+
+	return arm_iommu_attach_device(dev, priv->mapping);
+}
+
+static inline void __exynos_iommu_detach(struct exynos_drm_private *priv,
+					 struct device *dev)
+{
+	arm_iommu_detach_device(dev);
+}
+
+#elif defined(CONFIG_IOMMU_DMA)
+#include <linux/dma-iommu.h>
+
+static inline int __exynos_iommu_create_mapping(struct exynos_drm_private *priv,
+					unsigned long start, unsigned long size)
+{
+	struct iommu_domain *domain;
+	int ret;
+
+	domain = iommu_domain_alloc(priv->dma_dev->bus);
+	if (!domain)
+		return -ENOMEM;
+
+	ret = iommu_get_dma_cookie(domain);
+	if (ret)
+		goto free_domain;
+
+	ret = iommu_dma_init_domain(domain, start, size);
+	if (ret)
+		goto put_cookie;
+
+	priv->mapping = domain;
+	return 0;
+
+put_cookie:
+	iommu_put_dma_cookie(domain);
+free_domain:
+	iommu_domain_free(domain);
+	return ret;
+}
+
+static inline void __exynos_iommu_release_mapping(struct exynos_drm_private *priv)
+{
+	struct iommu_domain *domain = priv->mapping;
+
+	iommu_put_dma_cookie(domain);
+	iommu_domain_free(domain);
+	priv->mapping = NULL;
+}
+
+static inline int __exynos_iommu_attach(struct exynos_drm_private *priv,
+					struct device *dev)
+{
+	struct iommu_domain *domain = priv->mapping;
+
+	return iommu_attach_device(domain, dev);
+}
+
+static inline void __exynos_iommu_detach(struct exynos_drm_private *priv,
+					 struct device *dev)
+{
+	struct iommu_domain *domain = priv->mapping;
+
+	iommu_detach_device(domain, dev);
+}
+#else
+#error Unsupported architecture and IOMMU/DMA-mapping glue code
+#endif
+
 int drm_create_iommu_mapping(struct drm_device *drm_dev);
 
 void drm_release_iommu_mapping(struct drm_device *drm_dev);
diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c
index 706de3278f1c..3371635cd4d7 100644
--- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c
+++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c
@@ -44,6 +44,8 @@ static void fsl_dcu_drm_disable_crtc(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct fsl_dcu_drm_device *fsl_dev = dev->dev_private;
 
+	drm_crtc_vblank_off(crtc);
+
 	regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE,
 			   DCU_MODE_DCU_MODE_MASK,
 			   DCU_MODE_DCU_MODE(DCU_MODE_OFF));
@@ -61,6 +63,8 @@ static void fsl_dcu_drm_crtc_enable(struct drm_crtc *crtc)
 			   DCU_MODE_DCU_MODE(DCU_MODE_NORMAL));
 	regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE,
 		     DCU_UPDATE_MODE_READREG);
+
+	drm_crtc_vblank_on(crtc);
 }
 
 static void fsl_dcu_drm_crtc_mode_set_nofb(struct drm_crtc *crtc)
@@ -137,9 +141,10 @@ int fsl_dcu_drm_crtc_create(struct fsl_dcu_drm_device *fsl_dev)
 {
 	struct drm_plane *primary;
 	struct drm_crtc *crtc = &fsl_dev->crtc;
-	unsigned int i, j, reg_num;
 	int ret;
 
+	fsl_dcu_drm_init_planes(fsl_dev->drm);
+
 	primary = fsl_dcu_drm_primary_create_plane(fsl_dev->drm);
 	if (!primary)
 		return -ENOMEM;
@@ -153,19 +158,5 @@ int fsl_dcu_drm_crtc_create(struct fsl_dcu_drm_device *fsl_dev)
 
 	drm_crtc_helper_add(crtc, &fsl_dcu_drm_crtc_helper_funcs);
 
-	if (!strcmp(fsl_dev->soc->name, "ls1021a"))
-		reg_num = LS1021A_LAYER_REG_NUM;
-	else
-		reg_num = VF610_LAYER_REG_NUM;
-	for (i = 0; i < fsl_dev->soc->total_layer; i++) {
-		for (j = 1; j <= reg_num; j++)
-			regmap_write(fsl_dev->regmap, DCU_CTRLDESCLN(i, j), 0);
-	}
-	regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE,
-			   DCU_MODE_DCU_MODE_MASK,
-			   DCU_MODE_DCU_MODE(DCU_MODE_OFF));
-	regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE,
-		     DCU_UPDATE_MODE_READREG);
-
 	return 0;
 }
diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c
index 33727d5d826a..7882387f9bff 100644
--- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c
+++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c
@@ -11,6 +11,7 @@
 
 #include <linux/clk.h>
 #include <linux/clk-provider.h>
+#include <linux/console.h>
 #include <linux/io.h>
 #include <linux/mfd/syscon.h>
 #include <linux/mm.h>
@@ -22,6 +23,7 @@
 #include <linux/regmap.h>
 
 #include <drm/drmP.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
@@ -42,7 +44,6 @@ static const struct regmap_config fsl_dcu_regmap_config = {
 	.reg_bits = 32,
 	.reg_stride = 4,
 	.val_bits = 32,
-	.cache_type = REGCACHE_RBTREE,
 
 	.volatile_reg = fsl_dcu_drm_is_volatile_reg,
 };
@@ -228,11 +229,26 @@ static int fsl_dcu_drm_pm_suspend(struct device *dev)
 	if (!fsl_dev)
 		return 0;
 
+	disable_irq(fsl_dev->irq);
 	drm_kms_helper_poll_disable(fsl_dev->drm);
-	regcache_cache_only(fsl_dev->regmap, true);
-	regcache_mark_dirty(fsl_dev->regmap);
-	clk_disable(fsl_dev->clk);
-	clk_unprepare(fsl_dev->clk);
+
+	console_lock();
+	drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 1);
+	console_unlock();
+
+	fsl_dev->state = drm_atomic_helper_suspend(fsl_dev->drm);
+	if (IS_ERR(fsl_dev->state)) {
+		console_lock();
+		drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 0);
+		console_unlock();
+
+		drm_kms_helper_poll_enable(fsl_dev->drm);
+		enable_irq(fsl_dev->irq);
+		return PTR_ERR(fsl_dev->state);
+	}
+
+	clk_disable_unprepare(fsl_dev->pix_clk);
+	clk_disable_unprepare(fsl_dev->clk);
 
 	return 0;
 }
@@ -245,21 +261,27 @@ static int fsl_dcu_drm_pm_resume(struct device *dev)
 	if (!fsl_dev)
 		return 0;
 
-	ret = clk_enable(fsl_dev->clk);
+	ret = clk_prepare_enable(fsl_dev->clk);
 	if (ret < 0) {
 		dev_err(dev, "failed to enable dcu clk\n");
-		clk_unprepare(fsl_dev->clk);
 		return ret;
 	}
-	ret = clk_prepare(fsl_dev->clk);
+
+	ret = clk_prepare_enable(fsl_dev->pix_clk);
 	if (ret < 0) {
-		dev_err(dev, "failed to prepare dcu clk\n");
+		dev_err(dev, "failed to enable pix clk\n");
 		return ret;
 	}
 
+	fsl_dcu_drm_init_planes(fsl_dev->drm);
+	drm_atomic_helper_resume(fsl_dev->drm, fsl_dev->state);
+
+	console_lock();
+	drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 0);
+	console_unlock();
+
 	drm_kms_helper_poll_enable(fsl_dev->drm);
-	regcache_cache_only(fsl_dev->regmap, false);
-	regcache_sync(fsl_dev->regmap);
+	enable_irq(fsl_dev->irq);
 
 	return 0;
 }
@@ -273,12 +295,14 @@ static const struct fsl_dcu_soc_data fsl_dcu_ls1021a_data = {
 	.name = "ls1021a",
 	.total_layer = 16,
 	.max_layer = 4,
+	.layer_regs = LS1021A_LAYER_REG_NUM,
 };
 
 static const struct fsl_dcu_soc_data fsl_dcu_vf610_data = {
 	.name = "vf610",
 	.total_layer = 64,
 	.max_layer = 6,
+	.layer_regs = VF610_LAYER_REG_NUM,
 };
 
 static const struct of_device_id fsl_dcu_of_match[] = {
diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h
index c275f900ff23..3b371fe7491e 100644
--- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h
+++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h
@@ -175,6 +175,7 @@ struct fsl_dcu_soc_data {
 	unsigned int total_layer;
 	/*max layer number DCU supported*/
 	unsigned int max_layer;
+	unsigned int layer_regs;
 };
 
 struct fsl_dcu_drm_device {
@@ -193,6 +194,7 @@ struct fsl_dcu_drm_device {
 	struct drm_encoder encoder;
 	struct fsl_dcu_drm_connector connector;
 	const struct fsl_dcu_soc_data *soc;
+	struct drm_atomic_state *state;
 };
 
 void fsl_dcu_fbdev_init(struct drm_device *dev);
diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c
index 274558b3b32b..e50467a0deb0 100644
--- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c
+++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c
@@ -217,6 +217,22 @@ static const u32 fsl_dcu_drm_plane_formats[] = {
 	DRM_FORMAT_YUV422,
 };
 
+void fsl_dcu_drm_init_planes(struct drm_device *dev)
+{
+	struct fsl_dcu_drm_device *fsl_dev = dev->dev_private;
+	int i, j;
+
+	for (i = 0; i < fsl_dev->soc->total_layer; i++) {
+		for (j = 1; j <= fsl_dev->soc->layer_regs; j++)
+			regmap_write(fsl_dev->regmap, DCU_CTRLDESCLN(i, j), 0);
+	}
+	regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE,
+			   DCU_MODE_DCU_MODE_MASK,
+			   DCU_MODE_DCU_MODE(DCU_MODE_OFF));
+	regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE,
+		     DCU_UPDATE_MODE_READREG);
+}
+
 struct drm_plane *fsl_dcu_drm_primary_create_plane(struct drm_device *dev)
 {
 	struct drm_plane *primary;
diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h
index d657f088d859..8ee45f813ee8 100644
--- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h
+++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h
@@ -12,6 +12,7 @@
 #ifndef __FSL_DCU_DRM_PLANE_H__
 #define __FSL_DCU_DRM_PLANE_H__
 
+void fsl_dcu_drm_init_planes(struct drm_device *dev);
 struct drm_plane *fsl_dcu_drm_primary_create_plane(struct drm_device *dev);
 
 #endif /* __FSL_DCU_DRM_PLANE_H__ */
diff --git a/drivers/gpu/drm/hisilicon/kirin/Kconfig b/drivers/gpu/drm/hisilicon/kirin/Kconfig
index ea0df6115f7e..499f64405dac 100644
--- a/drivers/gpu/drm/hisilicon/kirin/Kconfig
+++ b/drivers/gpu/drm/hisilicon/kirin/Kconfig
@@ -4,6 +4,7 @@ config DRM_HISI_KIRIN
 	select DRM_KMS_HELPER
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_CMA_HELPER
+	select HISI_KIRIN_DW_DSI
 	help
 	  Choose this option if you have a hisilicon Kirin chipsets(hi6220).
 	  If M is selected the module will be called kirin-drm.
diff --git a/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c b/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c
index 805f4326fafe..c3707d47cd89 100644
--- a/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c
+++ b/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c
@@ -967,21 +967,21 @@ static int ade_dts_parse(struct platform_device *pdev, struct ade_hw_ctx *ctx)
 	}
 
 	ctx->ade_core_clk = devm_clk_get(dev, "clk_ade_core");
-	if (!ctx->ade_core_clk) {
+	if (IS_ERR(ctx->ade_core_clk)) {
 		DRM_ERROR("failed to parse clk ADE_CORE\n");
-		return -ENODEV;
+		return PTR_ERR(ctx->ade_core_clk);
 	}
 
 	ctx->media_noc_clk = devm_clk_get(dev, "clk_codec_jpeg");
-	if (!ctx->media_noc_clk) {
+	if (IS_ERR(ctx->media_noc_clk)) {
 		DRM_ERROR("failed to parse clk CODEC_JPEG\n");
-	    return -ENODEV;
+		return PTR_ERR(ctx->media_noc_clk);
 	}
 
 	ctx->ade_pix_clk = devm_clk_get(dev, "clk_ade_pix");
-	if (!ctx->ade_pix_clk) {
+	if (IS_ERR(ctx->ade_pix_clk)) {
 		DRM_ERROR("failed to parse clk ADE_PIX\n");
-	    return -ENODEV;
+		return PTR_ERR(ctx->ade_pix_clk);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
index 8f404103341d..cee87bfd10c4 100644
--- a/drivers/gpu/drm/i915/Kconfig.debug
+++ b/drivers/gpu/drm/i915/Kconfig.debug
@@ -18,6 +18,9 @@ config DRM_I915_WERROR
 config DRM_I915_DEBUG
         bool "Enable additional driver debugging"
         depends on DRM_I915
+        select PREEMPT_COUNT
+        select X86_MSR # used by igt/pm_rpm
+        select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks)
         default n
         help
           Choose this option to turn on extra driver debugging that may affect
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 276abf1cac2b..684fc1cd08fa 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -10,9 +10,11 @@ subdir-ccflags-$(CONFIG_DRM_I915_WERROR) := -Werror
 i915-y := i915_drv.o \
 	  i915_irq.o \
 	  i915_params.o \
+	  i915_pci.o \
           i915_suspend.o \
 	  i915_sysfs.o \
 	  intel_csr.o \
+	  intel_device_info.o \
 	  intel_pm.o \
 	  intel_runtime_pm.o
 
@@ -37,6 +39,7 @@ i915-y += i915_cmd_parser.o \
 	  i915_gem_userptr.o \
 	  i915_gpu_error.o \
 	  i915_trace_points.o \
+	  intel_breadcrumbs.o \
 	  intel_lrc.o \
 	  intel_mocs.o \
 	  intel_ringbuffer.o \
@@ -101,9 +104,6 @@ i915-y += dvo_ch7017.o \
 # virtual gpu code
 i915-y += i915_vgpu.o
 
-# legacy horrors
-i915-y += i915_dma.o
-
 ifeq ($(CONFIG_DRM_I915_GVT),y)
 i915-y += intel_gvt.o
 include $(src)/gvt/Makefile
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 5b7526697838..844fea795bae 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -265,7 +265,7 @@ static int i915_gem_stolen_list_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	u64 total_obj_size, total_gtt_size;
 	LIST_HEAD(stolen);
@@ -440,15 +440,15 @@ static void print_context_stats(struct seq_file *m,
 
 	memset(&stats, 0, sizeof(stats));
 
-	mutex_lock(&dev_priv->dev->struct_mutex);
+	mutex_lock(&dev_priv->drm.struct_mutex);
 	if (dev_priv->kernel_context)
 		per_file_ctx_stats(0, dev_priv->kernel_context, &stats);
 
-	list_for_each_entry(file, &dev_priv->dev->filelist, lhead) {
+	list_for_each_entry(file, &dev_priv->drm.filelist, lhead) {
 		struct drm_i915_file_private *fpriv = file->driver_priv;
 		idr_for_each(&fpriv->context_idr, per_file_ctx_stats, &stats);
 	}
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 
 	print_file_stats(m, "[k]contexts", stats);
 }
@@ -591,7 +591,7 @@ static int i915_gem_gtt_info(struct seq_file *m, void *data)
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
 	uintptr_t list = (uintptr_t) node->info_ent->data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	u64 total_obj_size, total_gtt_size;
 	int count, ret;
@@ -625,7 +625,7 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc;
 	int ret;
 
@@ -662,8 +662,8 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 					   engine->name,
 					   i915_gem_request_get_seqno(work->flip_queued_req),
 					   dev_priv->next_seqno,
-					   engine->get_seqno(engine),
-					   i915_gem_request_completed(work->flip_queued_req, true));
+					   intel_engine_get_seqno(engine),
+					   i915_gem_request_completed(work->flip_queued_req));
 			} else
 				seq_printf(m, "Flip not associated with any ring\n");
 			seq_printf(m, "Flip queued on frame %d, (was ready on frame %d), now %d\n",
@@ -695,7 +695,7 @@ static int i915_gem_batch_pool_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	struct intel_engine_cs *engine;
 	int total = 0;
@@ -740,7 +740,7 @@ static int i915_gem_request_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	struct drm_i915_gem_request *req;
 	int ret, any;
@@ -788,17 +788,29 @@ static int i915_gem_request_info(struct seq_file *m, void *data)
 static void i915_ring_seqno_info(struct seq_file *m,
 				 struct intel_engine_cs *engine)
 {
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct rb_node *rb;
+
 	seq_printf(m, "Current sequence (%s): %x\n",
-		   engine->name, engine->get_seqno(engine));
-	seq_printf(m, "Current user interrupts (%s): %x\n",
-		   engine->name, READ_ONCE(engine->user_interrupts));
+		   engine->name, intel_engine_get_seqno(engine));
+	seq_printf(m, "Current user interrupts (%s): %lx\n",
+		   engine->name, READ_ONCE(engine->breadcrumbs.irq_wakeups));
+
+	spin_lock(&b->lock);
+	for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
+		struct intel_wait *w = container_of(rb, typeof(*w), node);
+
+		seq_printf(m, "Waiting (%s): %s [%d] on %x\n",
+			   engine->name, w->tsk->comm, w->tsk->pid, w->seqno);
+	}
+	spin_unlock(&b->lock);
 }
 
 static int i915_gem_seqno_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	int ret;
 
@@ -821,7 +833,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	int ret, i, pipe;
 
@@ -1012,7 +1024,7 @@ static int i915_gem_fence_regs_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i, ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -1040,7 +1052,7 @@ static int i915_hws_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	const u32 *hws;
 	int i;
@@ -1151,7 +1163,7 @@ static int
 i915_next_seqno_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -1188,7 +1200,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
 	intel_runtime_pm_get(dev_priv);
@@ -1391,7 +1403,7 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	u64 acthd[I915_NUM_ENGINES];
 	u32 seqno[I915_NUM_ENGINES];
@@ -1408,7 +1420,7 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused)
 
 	for_each_engine_id(engine, dev_priv, id) {
 		acthd[id] = intel_ring_get_active_head(engine);
-		seqno[id] = engine->get_seqno(engine);
+		seqno[id] = intel_engine_get_seqno(engine);
 	}
 
 	i915_get_extra_instdone(dev_priv, instdone);
@@ -1428,9 +1440,11 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused)
 			   engine->hangcheck.seqno,
 			   seqno[id],
 			   engine->last_submitted_seqno);
-		seq_printf(m, "\tuser interrupts = %x [current %x]\n",
+		seq_printf(m, "\twaiters? %d\n",
+			   intel_engine_has_waiter(engine));
+		seq_printf(m, "\tuser interrupts = %lx [current %lx]\n",
 			   engine->hangcheck.user_interrupts,
-			   READ_ONCE(engine->user_interrupts));
+			   READ_ONCE(engine->breadcrumbs.irq_wakeups));
 		seq_printf(m, "\tACTHD = 0x%08llx [current 0x%08llx]\n",
 			   (long long)engine->hangcheck.acthd,
 			   (long long)acthd[id]);
@@ -1460,7 +1474,7 @@ static int ironlake_drpc_info(struct seq_file *m)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 rgvmodectl, rstdbyctl;
 	u16 crstandvid;
 	int ret;
@@ -1528,7 +1542,7 @@ static int i915_forcewake_domains(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_uncore_forcewake_domain *fw_domain;
 
 	spin_lock_irq(&dev_priv->uncore.lock);
@@ -1546,7 +1560,7 @@ static int vlv_drpc_info(struct seq_file *m)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 rpmodectl1, rcctl1, pw_status;
 
 	intel_runtime_pm_get(dev_priv);
@@ -1586,7 +1600,7 @@ static int gen6_drpc_info(struct seq_file *m)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 rpmodectl1, gt_core_status, rcctl1, rc6vids = 0;
 	unsigned forcewake_count;
 	int count = 0, ret;
@@ -1698,7 +1712,7 @@ static int i915_frontbuffer_tracking(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	seq_printf(m, "FB tracking busy bits: 0x%08x\n",
 		   dev_priv->fb_tracking.busy_bits);
@@ -1713,7 +1727,7 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!HAS_FBC(dev)) {
 		seq_puts(m, "FBC unsupported on this chipset\n");
@@ -1743,7 +1757,7 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 static int i915_fbc_fc_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
 		return -ENODEV;
@@ -1756,7 +1770,7 @@ static int i915_fbc_fc_get(void *data, u64 *val)
 static int i915_fbc_fc_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 reg;
 
 	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
@@ -1783,7 +1797,7 @@ static int i915_ips_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!HAS_IPS(dev)) {
 		seq_puts(m, "not supported\n");
@@ -1813,7 +1827,7 @@ static int i915_sr_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool sr_enabled = false;
 
 	intel_runtime_pm_get(dev_priv);
@@ -1842,7 +1856,7 @@ static int i915_emon_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long temp, chipset, gfx;
 	int ret;
 
@@ -1870,7 +1884,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 	int gpu_freq, ia_freq;
 	unsigned int max_gpu_freq, min_gpu_freq;
@@ -1925,7 +1939,7 @@ static int i915_opregion(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_opregion *opregion = &dev_priv->opregion;
 	int ret;
 
@@ -1946,7 +1960,7 @@ static int i915_vbt(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_opregion *opregion = &dev_priv->opregion;
 
 	if (opregion->vbt)
@@ -1968,19 +1982,19 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 		return ret;
 
 #ifdef CONFIG_DRM_FBDEV_EMULATION
-       if (to_i915(dev)->fbdev) {
-               fbdev_fb = to_intel_framebuffer(to_i915(dev)->fbdev->helper.fb);
-
-               seq_printf(m, "fbcon size: %d x %d, depth %d, %d bpp, modifier 0x%llx, refcount %d, obj ",
-                         fbdev_fb->base.width,
-                         fbdev_fb->base.height,
-                         fbdev_fb->base.depth,
-                         fbdev_fb->base.bits_per_pixel,
-                         fbdev_fb->base.modifier[0],
-                         drm_framebuffer_read_refcount(&fbdev_fb->base));
-               describe_obj(m, fbdev_fb->obj);
-               seq_putc(m, '\n');
-       }
+	if (to_i915(dev)->fbdev) {
+		fbdev_fb = to_intel_framebuffer(to_i915(dev)->fbdev->helper.fb);
+
+		seq_printf(m, "fbcon size: %d x %d, depth %d, %d bpp, modifier 0x%llx, refcount %d, obj ",
+			   fbdev_fb->base.width,
+			   fbdev_fb->base.height,
+			   fbdev_fb->base.depth,
+			   fbdev_fb->base.bits_per_pixel,
+			   fbdev_fb->base.modifier[0],
+			   drm_framebuffer_read_refcount(&fbdev_fb->base));
+		describe_obj(m, fbdev_fb->obj);
+		seq_putc(m, '\n');
+	}
 #endif
 
 	mutex_lock(&dev->mode_config.fb_lock);
@@ -2017,7 +2031,7 @@ static int i915_context_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	struct i915_gem_context *ctx;
 	int ret;
@@ -2114,7 +2128,7 @@ static int i915_dump_lrc(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	struct i915_gem_context *ctx;
 	int ret;
@@ -2141,7 +2155,7 @@ static int i915_execlists(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *)m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	u32 status_pointer;
 	u8 read_pointer;
@@ -2244,7 +2258,7 @@ static int i915_swizzle_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -2317,7 +2331,7 @@ static int per_file_ctx(int id, void *ptr, void *data)
 
 static void gen8_ppgtt_info(struct seq_file *m, struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
 	int i;
@@ -2338,7 +2352,7 @@ static void gen8_ppgtt_info(struct seq_file *m, struct drm_device *dev)
 
 static void gen6_ppgtt_info(struct seq_file *m, struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 
 	if (IS_GEN6(dev_priv))
@@ -2372,7 +2386,7 @@ static int i915_ppgtt_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_file *file;
 
 	int ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -2415,7 +2429,7 @@ static int count_irq_waiters(struct drm_i915_private *i915)
 	int count = 0;
 
 	for_each_engine(engine, i915)
-		count += engine->irq_refcount;
+		count += intel_engine_has_waiter(engine);
 
 	return count;
 }
@@ -2424,11 +2438,12 @@ static int i915_rps_boost_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_file *file;
 
 	seq_printf(m, "RPS enabled? %d\n", dev_priv->rps.enabled);
-	seq_printf(m, "GPU busy? %d\n", dev_priv->mm.busy);
+	seq_printf(m, "GPU busy? %s [%x]\n",
+		   yesno(dev_priv->gt.awake), dev_priv->gt.active_engines);
 	seq_printf(m, "CPU waiting? %d\n", count_irq_waiters(dev_priv));
 	seq_printf(m, "Frequency requested %d; min hard:%d, soft:%d; max soft:%d, hard:%d\n",
 		   intel_gpu_freq(dev_priv, dev_priv->rps.cur_freq),
@@ -2469,7 +2484,7 @@ static int i915_llc(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const bool edram = INTEL_GEN(dev_priv) > 8;
 
 	seq_printf(m, "LLC: %s\n", yesno(HAS_LLC(dev)));
@@ -2482,7 +2497,7 @@ static int i915_llc(struct seq_file *m, void *data)
 static int i915_guc_load_status_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
-	struct drm_i915_private *dev_priv = node->minor->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(node->minor->dev);
 	struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw;
 	u32 tmp, i;
 
@@ -2544,9 +2559,9 @@ static void i915_guc_client_info(struct seq_file *m,
 
 	for_each_engine(engine, dev_priv) {
 		seq_printf(m, "\tSubmissions: %llu %s\n",
-				client->submissions[engine->guc_id],
+				client->submissions[engine->id],
 				engine->name);
-		tot += client->submissions[engine->guc_id];
+		tot += client->submissions[engine->id];
 	}
 	seq_printf(m, "\tTotal: %llu\n", tot);
 }
@@ -2555,7 +2570,7 @@ static int i915_guc_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc guc;
 	struct i915_guc_client client = {};
 	struct intel_engine_cs *engine;
@@ -2587,9 +2602,9 @@ static int i915_guc_info(struct seq_file *m, void *data)
 	seq_printf(m, "\nGuC submissions:\n");
 	for_each_engine(engine, dev_priv) {
 		seq_printf(m, "\t%-24s: %10llu, last seqno 0x%08x\n",
-			engine->name, guc.submissions[engine->guc_id],
-			guc.last_seqno[engine->guc_id]);
-		total += guc.submissions[engine->guc_id];
+			engine->name, guc.submissions[engine->id],
+			guc.last_seqno[engine->id]);
+		total += guc.submissions[engine->id];
 	}
 	seq_printf(m, "\t%s: %llu\n", "Total", total);
 
@@ -2605,7 +2620,7 @@ static int i915_guc_log_dump(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *log_obj = dev_priv->guc.log_obj;
 	u32 *log;
 	int i = 0, pg;
@@ -2633,7 +2648,7 @@ static int i915_edp_psr_status(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 psrperf = 0;
 	u32 stat[3];
 	enum pipe pipe;
@@ -2701,7 +2716,6 @@ static int i915_sink_crc(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct intel_encoder *encoder;
 	struct intel_connector *connector;
 	struct intel_dp *intel_dp = NULL;
 	int ret;
@@ -2709,18 +2723,19 @@ static int i915_sink_crc(struct seq_file *m, void *data)
 
 	drm_modeset_lock_all(dev);
 	for_each_intel_connector(dev, connector) {
+		struct drm_crtc *crtc;
 
-		if (connector->base.dpms != DRM_MODE_DPMS_ON)
+		if (!connector->base.state->best_encoder)
 			continue;
 
-		if (!connector->base.encoder)
+		crtc = connector->base.state->crtc;
+		if (!crtc->state->active)
 			continue;
 
-		encoder = to_intel_encoder(connector->base.encoder);
-		if (encoder->type != INTEL_OUTPUT_EDP)
+		if (connector->base.connector_type != DRM_MODE_CONNECTOR_eDP)
 			continue;
 
-		intel_dp = enc_to_intel_dp(&encoder->base);
+		intel_dp = enc_to_intel_dp(connector->base.state->best_encoder);
 
 		ret = intel_dp_sink_crc(intel_dp, crc);
 		if (ret)
@@ -2741,7 +2756,7 @@ static int i915_energy_uJ(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u64 power;
 	u32 units;
 
@@ -2767,12 +2782,12 @@ static int i915_runtime_pm_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!HAS_RUNTIME_PM(dev_priv))
 		seq_puts(m, "Runtime power management not supported\n");
 
-	seq_printf(m, "GPU idle: %s\n", yesno(!dev_priv->mm.busy));
+	seq_printf(m, "GPU idle: %s\n", yesno(!dev_priv->gt.awake));
 	seq_printf(m, "IRQs disabled: %s\n",
 		   yesno(!intel_irqs_enabled(dev_priv)));
 #ifdef CONFIG_PM
@@ -2782,8 +2797,8 @@ static int i915_runtime_pm_status(struct seq_file *m, void *unused)
 	seq_printf(m, "Device Power Management (CONFIG_PM) disabled\n");
 #endif
 	seq_printf(m, "PCI device power state: %s [%d]\n",
-		   pci_power_name(dev_priv->dev->pdev->current_state),
-		   dev_priv->dev->pdev->current_state);
+		   pci_power_name(dev_priv->drm.pdev->current_state),
+		   dev_priv->drm.pdev->current_state);
 
 	return 0;
 }
@@ -2792,7 +2807,7 @@ static int i915_power_domain_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_power_domains *power_domains = &dev_priv->power_domains;
 	int i;
 
@@ -2827,7 +2842,7 @@ static int i915_dmc_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_csr *csr;
 
 	if (!HAS_CSR(dev)) {
@@ -2950,7 +2965,7 @@ static void intel_dp_info(struct seq_file *m,
 
 	seq_printf(m, "\tDPCD rev: %x\n", intel_dp->dpcd[DP_DPCD_REV]);
 	seq_printf(m, "\taudio support: %s\n", yesno(intel_dp->has_audio));
-	if (intel_encoder->type == INTEL_OUTPUT_EDP)
+	if (intel_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)
 		intel_panel_info(m, &intel_connector->panel);
 }
 
@@ -2989,14 +3004,26 @@ static void intel_connector_info(struct seq_file *m,
 		seq_printf(m, "\tCEA rev: %d\n",
 			   connector->display_info.cea_rev);
 	}
-	if (intel_encoder) {
-		if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
-		    intel_encoder->type == INTEL_OUTPUT_EDP)
-			intel_dp_info(m, intel_connector);
-		else if (intel_encoder->type == INTEL_OUTPUT_HDMI)
-			intel_hdmi_info(m, intel_connector);
-		else if (intel_encoder->type == INTEL_OUTPUT_LVDS)
+
+	if (!intel_encoder || intel_encoder->type == INTEL_OUTPUT_DP_MST)
+		return;
+
+	switch (connector->connector_type) {
+	case DRM_MODE_CONNECTOR_DisplayPort:
+	case DRM_MODE_CONNECTOR_eDP:
+		intel_dp_info(m, intel_connector);
+		break;
+	case DRM_MODE_CONNECTOR_LVDS:
+		if (intel_encoder->type == INTEL_OUTPUT_LVDS)
 			intel_lvds_info(m, intel_connector);
+		break;
+	case DRM_MODE_CONNECTOR_HDMIA:
+		if (intel_encoder->type == INTEL_OUTPUT_HDMI ||
+		    intel_encoder->type == INTEL_OUTPUT_UNKNOWN)
+			intel_hdmi_info(m, intel_connector);
+		break;
+	default:
+		break;
 	}
 
 	seq_printf(m, "\tmodes:\n");
@@ -3006,7 +3033,7 @@ static void intel_connector_info(struct seq_file *m,
 
 static bool cursor_active(struct drm_device *dev, int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 state;
 
 	if (IS_845G(dev) || IS_I865G(dev))
@@ -3019,7 +3046,7 @@ static bool cursor_active(struct drm_device *dev, int pipe)
 
 static bool cursor_position(struct drm_device *dev, int pipe, int *x, int *y)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pos;
 
 	pos = I915_READ(CURPOS(pipe));
@@ -3140,7 +3167,7 @@ static int i915_display_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc;
 	struct drm_connector *connector;
 
@@ -3195,7 +3222,7 @@ static int i915_semaphore_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	int num_rings = hweight32(INTEL_INFO(dev)->ring_mask);
 	enum intel_engine_id id;
@@ -3268,7 +3295,7 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	drm_modeset_lock_all(dev);
@@ -3298,7 +3325,7 @@ static int i915_wa_registers(struct seq_file *m, void *unused)
 	struct intel_engine_cs *engine;
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_workarounds *workarounds = &dev_priv->workarounds;
 	enum intel_engine_id id;
 
@@ -3336,7 +3363,7 @@ static int i915_ddb_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct skl_ddb_allocation *ddb;
 	struct skl_ddb_entry *entry;
 	enum pipe pipe;
@@ -3374,31 +3401,16 @@ static int i915_ddb_info(struct seq_file *m, void *unused)
 static void drrs_status_per_crtc(struct seq_file *m,
 		struct drm_device *dev, struct intel_crtc *intel_crtc)
 {
-	struct intel_encoder *intel_encoder;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_drrs *drrs = &dev_priv->drrs;
 	int vrefresh = 0;
+	struct drm_connector *connector;
 
-	for_each_encoder_on_crtc(dev, &intel_crtc->base, intel_encoder) {
-		/* Encoder connected on this CRTC */
-		switch (intel_encoder->type) {
-		case INTEL_OUTPUT_EDP:
-			seq_puts(m, "eDP:\n");
-			break;
-		case INTEL_OUTPUT_DSI:
-			seq_puts(m, "DSI:\n");
-			break;
-		case INTEL_OUTPUT_HDMI:
-			seq_puts(m, "HDMI:\n");
-			break;
-		case INTEL_OUTPUT_DISPLAYPORT:
-			seq_puts(m, "DP:\n");
-			break;
-		default:
-			seq_printf(m, "Other encoder (id=%d).\n",
-						intel_encoder->type);
-			return;
-		}
+	drm_for_each_connector(connector, dev) {
+		if (connector->state->crtc != &intel_crtc->base)
+			continue;
+
+		seq_printf(m, "%s:\n", connector->name);
 	}
 
 	if (dev_priv->vbt.drrs_type == STATIC_DRRS_SUPPORT)
@@ -3461,18 +3473,16 @@ static int i915_drrs_status(struct seq_file *m, void *unused)
 	struct intel_crtc *intel_crtc;
 	int active_crtc_cnt = 0;
 
+	drm_modeset_lock_all(dev);
 	for_each_intel_crtc(dev, intel_crtc) {
-		drm_modeset_lock(&intel_crtc->base.mutex, NULL);
-
 		if (intel_crtc->base.state->active) {
 			active_crtc_cnt++;
 			seq_printf(m, "\nCRTC %d:  ", active_crtc_cnt);
 
 			drrs_status_per_crtc(m, dev, intel_crtc);
 		}
-
-		drm_modeset_unlock(&intel_crtc->base.mutex);
 	}
+	drm_modeset_unlock_all(dev);
 
 	if (!active_crtc_cnt)
 		seq_puts(m, "No active crtc found\n");
@@ -3490,17 +3500,23 @@ static int i915_dp_mst_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct drm_encoder *encoder;
 	struct intel_encoder *intel_encoder;
 	struct intel_digital_port *intel_dig_port;
+	struct drm_connector *connector;
+
 	drm_modeset_lock_all(dev);
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		intel_encoder = to_intel_encoder(encoder);
-		if (intel_encoder->type != INTEL_OUTPUT_DISPLAYPORT)
+	drm_for_each_connector(connector, dev) {
+		if (connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		intel_encoder = intel_attached_encoder(connector);
+		if (!intel_encoder || intel_encoder->type == INTEL_OUTPUT_DP_MST)
 			continue;
-		intel_dig_port = enc_to_dig_port(encoder);
+
+		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
 		if (!intel_dig_port->dp.can_mst)
 			continue;
+
 		seq_printf(m, "MST Source Port %c\n",
 			   port_name(intel_dig_port->port));
 		drm_dp_mst_dump_topology(m, &intel_dig_port->dp.mst_mgr);
@@ -3512,7 +3528,7 @@ static int i915_dp_mst_info(struct seq_file *m, void *unused)
 static int i915_pipe_crc_open(struct inode *inode, struct file *filep)
 {
 	struct pipe_crc_info *info = inode->i_private;
-	struct drm_i915_private *dev_priv = info->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(info->dev);
 	struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[info->pipe];
 
 	if (info->pipe >= INTEL_INFO(info->dev)->num_pipes)
@@ -3536,7 +3552,7 @@ static int i915_pipe_crc_open(struct inode *inode, struct file *filep)
 static int i915_pipe_crc_release(struct inode *inode, struct file *filep)
 {
 	struct pipe_crc_info *info = inode->i_private;
-	struct drm_i915_private *dev_priv = info->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(info->dev);
 	struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[info->pipe];
 
 	spin_lock_irq(&pipe_crc->lock);
@@ -3564,7 +3580,7 @@ i915_pipe_crc_read(struct file *filep, char __user *user_buf, size_t count,
 {
 	struct pipe_crc_info *info = filep->private_data;
 	struct drm_device *dev = info->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[info->pipe];
 	char buf[PIPE_CRC_BUFFER_LEN];
 	int n_entries;
@@ -3697,7 +3713,7 @@ static const char *pipe_crc_source_name(enum intel_pipe_crc_source source)
 static int display_crc_ctl_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	for (i = 0; i < I915_MAX_PIPES; i++)
@@ -3758,7 +3774,7 @@ static int i9xx_pipe_crc_auto_source(struct drm_device *dev, enum pipe pipe,
 		case INTEL_OUTPUT_TVOUT:
 			*source = INTEL_PIPE_CRC_SOURCE_TV;
 			break;
-		case INTEL_OUTPUT_DISPLAYPORT:
+		case INTEL_OUTPUT_DP:
 		case INTEL_OUTPUT_EDP:
 			dig_port = enc_to_dig_port(&encoder->base);
 			switch (dig_port->port) {
@@ -3791,7 +3807,7 @@ static int vlv_pipe_crc_ctl_reg(struct drm_device *dev,
 				enum intel_pipe_crc_source *source,
 				uint32_t *val)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool need_stable_symbols = false;
 
 	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) {
@@ -3862,7 +3878,7 @@ static int i9xx_pipe_crc_ctl_reg(struct drm_device *dev,
 				 enum intel_pipe_crc_source *source,
 				 uint32_t *val)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool need_stable_symbols = false;
 
 	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) {
@@ -3936,7 +3952,7 @@ static int i9xx_pipe_crc_ctl_reg(struct drm_device *dev,
 static void vlv_undo_pipe_scramble_reset(struct drm_device *dev,
 					 enum pipe pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t tmp = I915_READ(PORT_DFT2_G4X);
 
 	switch (pipe) {
@@ -3961,7 +3977,7 @@ static void vlv_undo_pipe_scramble_reset(struct drm_device *dev,
 static void g4x_undo_pipe_scramble_reset(struct drm_device *dev,
 					 enum pipe pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t tmp = I915_READ(PORT_DFT2_G4X);
 
 	if (pipe == PIPE_A)
@@ -4004,7 +4020,7 @@ static int ilk_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
 
 static void hsw_trans_edp_pipe_A_crc_wa(struct drm_device *dev, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc =
 		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_A]);
 	struct intel_crtc_state *pipe_config;
@@ -4072,7 +4088,7 @@ static int ivb_pipe_crc_ctl_reg(struct drm_device *dev,
 static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 			       enum intel_pipe_crc_source source)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe];
 	struct intel_crtc *crtc = to_intel_crtc(intel_get_crtc_for_pipe(dev,
 									pipe));
@@ -4579,7 +4595,7 @@ static void wm_latency_show(struct seq_file *m, const uint16_t wm[8])
 static int pri_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4595,7 +4611,7 @@ static int pri_wm_latency_show(struct seq_file *m, void *data)
 static int spr_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4611,7 +4627,7 @@ static int spr_wm_latency_show(struct seq_file *m, void *data)
 static int cur_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4702,7 +4718,7 @@ static ssize_t pri_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4718,7 +4734,7 @@ static ssize_t spr_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4734,7 +4750,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint16_t *latencies;
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4776,7 +4792,7 @@ static int
 i915_wedged_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	*val = i915_terminally_wedged(&dev_priv->gpu_error);
 
@@ -4787,7 +4803,7 @@ static int
 i915_wedged_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/*
 	 * There is no safeguard against this debugfs entry colliding
@@ -4815,44 +4831,10 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_wedged_fops,
 			"%llu\n");
 
 static int
-i915_ring_stop_get(void *data, u64 *val)
-{
-	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	*val = dev_priv->gpu_error.stop_rings;
-
-	return 0;
-}
-
-static int
-i915_ring_stop_set(void *data, u64 val)
-{
-	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	DRM_DEBUG_DRIVER("Stopping rings 0x%08llx\n", val);
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	dev_priv->gpu_error.stop_rings = val;
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
-DEFINE_SIMPLE_ATTRIBUTE(i915_ring_stop_fops,
-			i915_ring_stop_get, i915_ring_stop_set,
-			"0x%08llx\n");
-
-static int
 i915_ring_missed_irq_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	*val = dev_priv->gpu_error.missed_irq_rings;
 	return 0;
@@ -4862,7 +4844,7 @@ static int
 i915_ring_missed_irq_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	/* Lock against concurrent debugfs callers */
@@ -4883,7 +4865,7 @@ static int
 i915_ring_test_irq_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	*val = dev_priv->gpu_error.test_irq_rings;
 
@@ -4894,18 +4876,11 @@ static int
 i915_ring_test_irq_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
+	val &= INTEL_INFO(dev_priv)->ring_mask;
 	DRM_DEBUG_DRIVER("Masking interrupts on rings 0x%08llx\n", val);
-
-	/* Lock against concurrent debugfs callers */
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
 	dev_priv->gpu_error.test_irq_rings = val;
-	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
 }
@@ -4934,7 +4909,7 @@ static int
 i915_drop_caches_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	DRM_DEBUG("Dropping caches: 0x%08llx\n", val);
@@ -4946,7 +4921,7 @@ i915_drop_caches_set(void *data, u64 val)
 		return ret;
 
 	if (val & DROP_ACTIVE) {
-		ret = i915_gpu_idle(dev);
+		ret = i915_gem_wait_for_idle(dev_priv);
 		if (ret)
 			goto unlock;
 	}
@@ -4974,7 +4949,7 @@ static int
 i915_max_freq_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	if (INTEL_INFO(dev)->gen < 6)
@@ -4996,7 +4971,7 @@ static int
 i915_max_freq_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 hw_max, hw_min;
 	int ret;
 
@@ -5041,7 +5016,7 @@ static int
 i915_min_freq_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	if (INTEL_INFO(dev)->gen < 6)
@@ -5063,7 +5038,7 @@ static int
 i915_min_freq_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 hw_max, hw_min;
 	int ret;
 
@@ -5108,7 +5083,7 @@ static int
 i915_cache_sharing_get(void *data, u64 *val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 snpcr;
 	int ret;
 
@@ -5123,7 +5098,7 @@ i915_cache_sharing_get(void *data, u64 *val)
 	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
 
 	intel_runtime_pm_put(dev_priv);
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 
 	*val = (snpcr & GEN6_MBC_SNPCR_MASK) >> GEN6_MBC_SNPCR_SHIFT;
 
@@ -5134,7 +5109,7 @@ static int
 i915_cache_sharing_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 snpcr;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
@@ -5171,7 +5146,7 @@ struct sseu_dev_status {
 static void cherryview_sseu_device_status(struct drm_device *dev,
 					  struct sseu_dev_status *stat)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ss_max = 2;
 	int ss;
 	u32 sig1[ss_max], sig2[ss_max];
@@ -5203,7 +5178,7 @@ static void cherryview_sseu_device_status(struct drm_device *dev,
 static void gen9_sseu_device_status(struct drm_device *dev,
 				    struct sseu_dev_status *stat)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int s_max = 3, ss_max = 4;
 	int s, ss;
 	u32 s_reg[s_max], eu_reg[2*s_max], eu_mask[2];
@@ -5268,7 +5243,7 @@ static void gen9_sseu_device_status(struct drm_device *dev,
 static void broadwell_sseu_device_status(struct drm_device *dev,
 					 struct sseu_dev_status *stat)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int s;
 	u32 slice_info = I915_READ(GEN8_GT_SLICE_INFO);
 
@@ -5347,7 +5322,7 @@ static int i915_sseu_status(struct seq_file *m, void *unused)
 static int i915_forcewake_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_INFO(dev)->gen < 6)
 		return 0;
@@ -5361,7 +5336,7 @@ static int i915_forcewake_open(struct inode *inode, struct file *file)
 static int i915_forcewake_release(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_INFO(dev)->gen < 6)
 		return 0;
@@ -5477,7 +5452,6 @@ static const struct i915_debugfs_files {
 	{"i915_max_freq", &i915_max_freq_fops},
 	{"i915_min_freq", &i915_min_freq_fops},
 	{"i915_cache_sharing", &i915_cache_sharing_fops},
-	{"i915_ring_stop", &i915_ring_stop_fops},
 	{"i915_ring_missed_irq", &i915_ring_missed_irq_fops},
 	{"i915_ring_test_irq", &i915_ring_test_irq_fops},
 	{"i915_gem_drop_caches", &i915_drop_caches_fops},
@@ -5495,7 +5469,7 @@ static const struct i915_debugfs_files {
 
 void intel_display_crc_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 
 	for_each_pipe(dev_priv, pipe) {
@@ -5507,8 +5481,9 @@ void intel_display_crc_init(struct drm_device *dev)
 	}
 }
 
-int i915_debugfs_init(struct drm_minor *minor)
+int i915_debugfs_register(struct drm_i915_private *dev_priv)
 {
+	struct drm_minor *minor = dev_priv->drm.primary;
 	int ret, i;
 
 	ret = i915_forcewake_create(minor->debugfs_root, minor);
@@ -5534,8 +5509,9 @@ int i915_debugfs_init(struct drm_minor *minor)
 					minor->debugfs_root, minor);
 }
 
-void i915_debugfs_cleanup(struct drm_minor *minor)
+void i915_debugfs_unregister(struct drm_i915_private *dev_priv)
 {
+	struct drm_minor *minor = dev_priv->drm.primary;
 	int i;
 
 	drm_debugfs_remove_files(i915_debugfs_list,
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
deleted file mode 100644
index d15a461fa84a..000000000000
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ /dev/null
@@ -1,1692 +0,0 @@
-/* i915_dma.c -- DMA support for the I915 -*- linux-c -*-
- */
-/*
- * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
- * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
- * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
-#include <drm/drmP.h>
-#include <drm/drm_crtc_helper.h>
-#include <drm/drm_fb_helper.h>
-#include <drm/drm_legacy.h>
-#include "intel_drv.h"
-#include <drm/i915_drm.h>
-#include "i915_drv.h"
-#include "i915_vgpu.h"
-#include "i915_trace.h"
-#include <linux/pci.h>
-#include <linux/console.h>
-#include <linux/vt.h>
-#include <linux/vgaarb.h>
-#include <linux/acpi.h>
-#include <linux/pnp.h>
-#include <linux/vga_switcheroo.h>
-#include <linux/slab.h>
-#include <acpi/video.h>
-#include <linux/pm.h>
-#include <linux/pm_runtime.h>
-#include <linux/oom.h>
-
-static unsigned int i915_load_fail_count;
-
-bool __i915_inject_load_failure(const char *func, int line)
-{
-	if (i915_load_fail_count >= i915.inject_load_failure)
-		return false;
-
-	if (++i915_load_fail_count == i915.inject_load_failure) {
-		DRM_INFO("Injecting failure at checkpoint %u [%s:%d]\n",
-			 i915.inject_load_failure, func, line);
-		return true;
-	}
-
-	return false;
-}
-
-#define FDO_BUG_URL "https://bugs.freedesktop.org/enter_bug.cgi?product=DRI"
-#define FDO_BUG_MSG "Please file a bug at " FDO_BUG_URL " against DRM/Intel " \
-		    "providing the dmesg log by booting with drm.debug=0xf"
-
-void
-__i915_printk(struct drm_i915_private *dev_priv, const char *level,
-	      const char *fmt, ...)
-{
-	static bool shown_bug_once;
-	struct device *dev = dev_priv->dev->dev;
-	bool is_error = level[1] <= KERN_ERR[1];
-	bool is_debug = level[1] == KERN_DEBUG[1];
-	struct va_format vaf;
-	va_list args;
-
-	if (is_debug && !(drm_debug & DRM_UT_DRIVER))
-		return;
-
-	va_start(args, fmt);
-
-	vaf.fmt = fmt;
-	vaf.va = &args;
-
-	dev_printk(level, dev, "[" DRM_NAME ":%ps] %pV",
-		   __builtin_return_address(0), &vaf);
-
-	if (is_error && !shown_bug_once) {
-		dev_notice(dev, "%s", FDO_BUG_MSG);
-		shown_bug_once = true;
-	}
-
-	va_end(args);
-}
-
-static bool i915_error_injected(struct drm_i915_private *dev_priv)
-{
-	return i915.inject_load_failure &&
-	       i915_load_fail_count == i915.inject_load_failure;
-}
-
-#define i915_load_error(dev_priv, fmt, ...)				     \
-	__i915_printk(dev_priv,						     \
-		      i915_error_injected(dev_priv) ? KERN_DEBUG : KERN_ERR, \
-		      fmt, ##__VA_ARGS__)
-
-static int i915_getparam(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	drm_i915_getparam_t *param = data;
-	int value;
-
-	switch (param->param) {
-	case I915_PARAM_IRQ_ACTIVE:
-	case I915_PARAM_ALLOW_BATCHBUFFER:
-	case I915_PARAM_LAST_DISPATCH:
-		/* Reject all old ums/dri params. */
-		return -ENODEV;
-	case I915_PARAM_CHIPSET_ID:
-		value = dev->pdev->device;
-		break;
-	case I915_PARAM_REVISION:
-		value = dev->pdev->revision;
-		break;
-	case I915_PARAM_HAS_GEM:
-		value = 1;
-		break;
-	case I915_PARAM_NUM_FENCES_AVAIL:
-		value = dev_priv->num_fence_regs;
-		break;
-	case I915_PARAM_HAS_OVERLAY:
-		value = dev_priv->overlay ? 1 : 0;
-		break;
-	case I915_PARAM_HAS_PAGEFLIPPING:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_EXECBUF2:
-		/* depends on GEM */
-		value = 1;
-		break;
-	case I915_PARAM_HAS_BSD:
-		value = intel_engine_initialized(&dev_priv->engine[VCS]);
-		break;
-	case I915_PARAM_HAS_BLT:
-		value = intel_engine_initialized(&dev_priv->engine[BCS]);
-		break;
-	case I915_PARAM_HAS_VEBOX:
-		value = intel_engine_initialized(&dev_priv->engine[VECS]);
-		break;
-	case I915_PARAM_HAS_BSD2:
-		value = intel_engine_initialized(&dev_priv->engine[VCS2]);
-		break;
-	case I915_PARAM_HAS_RELAXED_FENCING:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_COHERENT_RINGS:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_EXEC_CONSTANTS:
-		value = INTEL_INFO(dev)->gen >= 4;
-		break;
-	case I915_PARAM_HAS_RELAXED_DELTA:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_GEN7_SOL_RESET:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_LLC:
-		value = HAS_LLC(dev);
-		break;
-	case I915_PARAM_HAS_WT:
-		value = HAS_WT(dev);
-		break;
-	case I915_PARAM_HAS_ALIASING_PPGTT:
-		value = USES_PPGTT(dev);
-		break;
-	case I915_PARAM_HAS_WAIT_TIMEOUT:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_SEMAPHORES:
-		value = i915_semaphore_is_enabled(dev_priv);
-		break;
-	case I915_PARAM_HAS_PRIME_VMAP_FLUSH:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_SECURE_BATCHES:
-		value = capable(CAP_SYS_ADMIN);
-		break;
-	case I915_PARAM_HAS_PINNED_BATCHES:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_EXEC_NO_RELOC:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_EXEC_HANDLE_LUT:
-		value = 1;
-		break;
-	case I915_PARAM_CMD_PARSER_VERSION:
-		value = i915_cmd_parser_get_version(dev_priv);
-		break;
-	case I915_PARAM_HAS_COHERENT_PHYS_GTT:
-		value = 1;
-		break;
-	case I915_PARAM_MMAP_VERSION:
-		value = 1;
-		break;
-	case I915_PARAM_SUBSLICE_TOTAL:
-		value = INTEL_INFO(dev)->subslice_total;
-		if (!value)
-			return -ENODEV;
-		break;
-	case I915_PARAM_EU_TOTAL:
-		value = INTEL_INFO(dev)->eu_total;
-		if (!value)
-			return -ENODEV;
-		break;
-	case I915_PARAM_HAS_GPU_RESET:
-		value = i915.enable_hangcheck && intel_has_gpu_reset(dev_priv);
-		break;
-	case I915_PARAM_HAS_RESOURCE_STREAMER:
-		value = HAS_RESOURCE_STREAMER(dev);
-		break;
-	case I915_PARAM_HAS_EXEC_SOFTPIN:
-		value = 1;
-		break;
-	default:
-		DRM_DEBUG("Unknown parameter %d\n", param->param);
-		return -EINVAL;
-	}
-
-	if (copy_to_user(param->value, &value, sizeof(int))) {
-		DRM_ERROR("copy_to_user failed\n");
-		return -EFAULT;
-	}
-
-	return 0;
-}
-
-static int i915_get_bridge_dev(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	dev_priv->bridge_dev = pci_get_bus_and_slot(0, PCI_DEVFN(0, 0));
-	if (!dev_priv->bridge_dev) {
-		DRM_ERROR("bridge device not found\n");
-		return -1;
-	}
-	return 0;
-}
-
-/* Allocate space for the MCH regs if needed, return nonzero on error */
-static int
-intel_alloc_mchbar_resource(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
-	u32 temp_lo, temp_hi = 0;
-	u64 mchbar_addr;
-	int ret;
-
-	if (INTEL_INFO(dev)->gen >= 4)
-		pci_read_config_dword(dev_priv->bridge_dev, reg + 4, &temp_hi);
-	pci_read_config_dword(dev_priv->bridge_dev, reg, &temp_lo);
-	mchbar_addr = ((u64)temp_hi << 32) | temp_lo;
-
-	/* If ACPI doesn't have it, assume we need to allocate it ourselves */
-#ifdef CONFIG_PNP
-	if (mchbar_addr &&
-	    pnp_range_reserved(mchbar_addr, mchbar_addr + MCHBAR_SIZE))
-		return 0;
-#endif
-
-	/* Get some space for it */
-	dev_priv->mch_res.name = "i915 MCHBAR";
-	dev_priv->mch_res.flags = IORESOURCE_MEM;
-	ret = pci_bus_alloc_resource(dev_priv->bridge_dev->bus,
-				     &dev_priv->mch_res,
-				     MCHBAR_SIZE, MCHBAR_SIZE,
-				     PCIBIOS_MIN_MEM,
-				     0, pcibios_align_resource,
-				     dev_priv->bridge_dev);
-	if (ret) {
-		DRM_DEBUG_DRIVER("failed bus alloc: %d\n", ret);
-		dev_priv->mch_res.start = 0;
-		return ret;
-	}
-
-	if (INTEL_INFO(dev)->gen >= 4)
-		pci_write_config_dword(dev_priv->bridge_dev, reg + 4,
-				       upper_32_bits(dev_priv->mch_res.start));
-
-	pci_write_config_dword(dev_priv->bridge_dev, reg,
-			       lower_32_bits(dev_priv->mch_res.start));
-	return 0;
-}
-
-/* Setup MCHBAR if possible, return true if we should disable it again */
-static void
-intel_setup_mchbar(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int mchbar_reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
-	u32 temp;
-	bool enabled;
-
-	if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev))
-		return;
-
-	dev_priv->mchbar_need_disable = false;
-
-	if (IS_I915G(dev) || IS_I915GM(dev)) {
-		pci_read_config_dword(dev_priv->bridge_dev, DEVEN, &temp);
-		enabled = !!(temp & DEVEN_MCHBAR_EN);
-	} else {
-		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
-		enabled = temp & 1;
-	}
-
-	/* If it's already enabled, don't have to do anything */
-	if (enabled)
-		return;
-
-	if (intel_alloc_mchbar_resource(dev))
-		return;
-
-	dev_priv->mchbar_need_disable = true;
-
-	/* Space is allocated or reserved, so enable it. */
-	if (IS_I915G(dev) || IS_I915GM(dev)) {
-		pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
-				       temp | DEVEN_MCHBAR_EN);
-	} else {
-		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
-		pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp | 1);
-	}
-}
-
-static void
-intel_teardown_mchbar(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int mchbar_reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
-
-	if (dev_priv->mchbar_need_disable) {
-		if (IS_I915G(dev) || IS_I915GM(dev)) {
-			u32 deven_val;
-
-			pci_read_config_dword(dev_priv->bridge_dev, DEVEN,
-					      &deven_val);
-			deven_val &= ~DEVEN_MCHBAR_EN;
-			pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
-					       deven_val);
-		} else {
-			u32 mchbar_val;
-
-			pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg,
-					      &mchbar_val);
-			mchbar_val &= ~1;
-			pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg,
-					       mchbar_val);
-		}
-	}
-
-	if (dev_priv->mch_res.start)
-		release_resource(&dev_priv->mch_res);
-}
-
-/* true = enable decode, false = disable decoder */
-static unsigned int i915_vga_set_decode(void *cookie, bool state)
-{
-	struct drm_device *dev = cookie;
-
-	intel_modeset_vga_set_state(dev, state);
-	if (state)
-		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
-		       VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
-	else
-		return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
-}
-
-static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state)
-{
-	struct drm_device *dev = pci_get_drvdata(pdev);
-	pm_message_t pmm = { .event = PM_EVENT_SUSPEND };
-
-	if (state == VGA_SWITCHEROO_ON) {
-		pr_info("switched on\n");
-		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
-		/* i915 resume handler doesn't set to D0 */
-		pci_set_power_state(dev->pdev, PCI_D0);
-		i915_resume_switcheroo(dev);
-		dev->switch_power_state = DRM_SWITCH_POWER_ON;
-	} else {
-		pr_info("switched off\n");
-		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
-		i915_suspend_switcheroo(dev, pmm);
-		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
-	}
-}
-
-static bool i915_switcheroo_can_switch(struct pci_dev *pdev)
-{
-	struct drm_device *dev = pci_get_drvdata(pdev);
-
-	/*
-	 * FIXME: open_count is protected by drm_global_mutex but that would lead to
-	 * locking inversion with the driver load path. And the access here is
-	 * completely racy anyway. So don't bother with locking for now.
-	 */
-	return dev->open_count == 0;
-}
-
-static const struct vga_switcheroo_client_ops i915_switcheroo_ops = {
-	.set_gpu_state = i915_switcheroo_set_state,
-	.reprobe = NULL,
-	.can_switch = i915_switcheroo_can_switch,
-};
-
-static void i915_gem_fini(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-
-	/*
-	 * Neither the BIOS, ourselves or any other kernel
-	 * expects the system to be in execlists mode on startup,
-	 * so we need to reset the GPU back to legacy mode. And the only
-	 * known way to disable logical contexts is through a GPU reset.
-	 *
-	 * So in order to leave the system in a known default configuration,
-	 * always reset the GPU upon unload. Afterwards we then clean up the
-	 * GEM state tracking, flushing off the requests and leaving the
-	 * system in a known idle state.
-	 *
-	 * Note that is of the upmost importance that the GPU is idle and
-	 * all stray writes are flushed *before* we dismantle the backing
-	 * storage for the pinned objects.
-	 *
-	 * However, since we are uncertain that reseting the GPU on older
-	 * machines is a good idea, we don't - just in case it leaves the
-	 * machine in an unusable condition.
-	 */
-	if (HAS_HW_CONTEXTS(dev)) {
-		int reset = intel_gpu_reset(dev_priv, ALL_ENGINES);
-		WARN_ON(reset && reset != -ENODEV);
-	}
-
-	mutex_lock(&dev->struct_mutex);
-	i915_gem_reset(dev);
-	i915_gem_cleanup_engines(dev);
-	i915_gem_context_fini(dev);
-	mutex_unlock(&dev->struct_mutex);
-
-	WARN_ON(!list_empty(&to_i915(dev)->context_list));
-}
-
-static int i915_load_modeset_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	if (i915_inject_load_failure())
-		return -ENODEV;
-
-	ret = intel_bios_init(dev_priv);
-	if (ret)
-		DRM_INFO("failed to find VBIOS tables\n");
-
-	/* If we have > 1 VGA cards, then we need to arbitrate access
-	 * to the common VGA resources.
-	 *
-	 * If we are a secondary display controller (!PCI_DISPLAY_CLASS_VGA),
-	 * then we do not take part in VGA arbitration and the
-	 * vga_client_register() fails with -ENODEV.
-	 */
-	ret = vga_client_register(dev->pdev, dev, NULL, i915_vga_set_decode);
-	if (ret && ret != -ENODEV)
-		goto out;
-
-	intel_register_dsm_handler();
-
-	ret = vga_switcheroo_register_client(dev->pdev, &i915_switcheroo_ops, false);
-	if (ret)
-		goto cleanup_vga_client;
-
-	/* must happen before intel_power_domains_init_hw() on VLV/CHV */
-	intel_update_rawclk(dev_priv);
-
-	intel_power_domains_init_hw(dev_priv, false);
-
-	intel_csr_ucode_init(dev_priv);
-
-	ret = intel_irq_install(dev_priv);
-	if (ret)
-		goto cleanup_csr;
-
-	intel_setup_gmbus(dev);
-
-	/* Important: The output setup functions called by modeset_init need
-	 * working irqs for e.g. gmbus and dp aux transfers. */
-	intel_modeset_init(dev);
-
-	intel_guc_init(dev);
-
-	ret = i915_gem_init(dev);
-	if (ret)
-		goto cleanup_irq;
-
-	intel_modeset_gem_init(dev);
-
-	if (INTEL_INFO(dev)->num_pipes == 0)
-		return 0;
-
-	ret = intel_fbdev_init(dev);
-	if (ret)
-		goto cleanup_gem;
-
-	/* Only enable hotplug handling once the fbdev is fully set up. */
-	intel_hpd_init(dev_priv);
-
-	/*
-	 * Some ports require correctly set-up hpd registers for detection to
-	 * work properly (leading to ghost connected connector status), e.g. VGA
-	 * on gm45.  Hence we can only set up the initial fbdev config after hpd
-	 * irqs are fully enabled. Now we should scan for the initial config
-	 * only once hotplug handling is enabled, but due to screwed-up locking
-	 * around kms/fbdev init we can't protect the fdbev initial config
-	 * scanning against hotplug events. Hence do this first and ignore the
-	 * tiny window where we will loose hotplug notifactions.
-	 */
-	intel_fbdev_initial_config_async(dev);
-
-	drm_kms_helper_poll_init(dev);
-
-	return 0;
-
-cleanup_gem:
-	i915_gem_fini(dev);
-cleanup_irq:
-	intel_guc_fini(dev);
-	drm_irq_uninstall(dev);
-	intel_teardown_gmbus(dev);
-cleanup_csr:
-	intel_csr_ucode_fini(dev_priv);
-	intel_power_domains_fini(dev_priv);
-	vga_switcheroo_unregister_client(dev->pdev);
-cleanup_vga_client:
-	vga_client_register(dev->pdev, NULL, NULL, NULL);
-out:
-	return ret;
-}
-
-#if IS_ENABLED(CONFIG_FB)
-static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
-{
-	struct apertures_struct *ap;
-	struct pci_dev *pdev = dev_priv->dev->pdev;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-	bool primary;
-	int ret;
-
-	ap = alloc_apertures(1);
-	if (!ap)
-		return -ENOMEM;
-
-	ap->ranges[0].base = ggtt->mappable_base;
-	ap->ranges[0].size = ggtt->mappable_end;
-
-	primary =
-		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
-
-	ret = remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
-
-	kfree(ap);
-
-	return ret;
-}
-#else
-static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
-{
-	return 0;
-}
-#endif
-
-#if !defined(CONFIG_VGA_CONSOLE)
-static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
-{
-	return 0;
-}
-#elif !defined(CONFIG_DUMMY_CONSOLE)
-static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
-{
-	return -ENODEV;
-}
-#else
-static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
-{
-	int ret = 0;
-
-	DRM_INFO("Replacing VGA console driver\n");
-
-	console_lock();
-	if (con_is_bound(&vga_con))
-		ret = do_take_over_console(&dummy_con, 0, MAX_NR_CONSOLES - 1, 1);
-	if (ret == 0) {
-		ret = do_unregister_con_driver(&vga_con);
-
-		/* Ignore "already unregistered". */
-		if (ret == -ENODEV)
-			ret = 0;
-	}
-	console_unlock();
-
-	return ret;
-}
-#endif
-
-static void i915_dump_device_info(struct drm_i915_private *dev_priv)
-{
-	const struct intel_device_info *info = &dev_priv->info;
-
-#define PRINT_S(name) "%s"
-#define SEP_EMPTY
-#define PRINT_FLAG(name) info->name ? #name "," : ""
-#define SEP_COMMA ,
-	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x rev=0x%02x flags="
-			 DEV_INFO_FOR_EACH_FLAG(PRINT_S, SEP_EMPTY),
-			 info->gen,
-			 dev_priv->dev->pdev->device,
-			 dev_priv->dev->pdev->revision,
-			 DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_COMMA));
-#undef PRINT_S
-#undef SEP_EMPTY
-#undef PRINT_FLAG
-#undef SEP_COMMA
-}
-
-static void cherryview_sseu_info_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_device_info *info;
-	u32 fuse, eu_dis;
-
-	info = (struct intel_device_info *)&dev_priv->info;
-	fuse = I915_READ(CHV_FUSE_GT);
-
-	info->slice_total = 1;
-
-	if (!(fuse & CHV_FGT_DISABLE_SS0)) {
-		info->subslice_per_slice++;
-		eu_dis = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK |
-				 CHV_FGT_EU_DIS_SS0_R1_MASK);
-		info->eu_total += 8 - hweight32(eu_dis);
-	}
-
-	if (!(fuse & CHV_FGT_DISABLE_SS1)) {
-		info->subslice_per_slice++;
-		eu_dis = fuse & (CHV_FGT_EU_DIS_SS1_R0_MASK |
-				 CHV_FGT_EU_DIS_SS1_R1_MASK);
-		info->eu_total += 8 - hweight32(eu_dis);
-	}
-
-	info->subslice_total = info->subslice_per_slice;
-	/*
-	 * CHV expected to always have a uniform distribution of EU
-	 * across subslices.
-	*/
-	info->eu_per_subslice = info->subslice_total ?
-				info->eu_total / info->subslice_total :
-				0;
-	/*
-	 * CHV supports subslice power gating on devices with more than
-	 * one subslice, and supports EU power gating on devices with
-	 * more than one EU pair per subslice.
-	*/
-	info->has_slice_pg = 0;
-	info->has_subslice_pg = (info->subslice_total > 1);
-	info->has_eu_pg = (info->eu_per_subslice > 2);
-}
-
-static void gen9_sseu_info_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_device_info *info;
-	int s_max = 3, ss_max = 4, eu_max = 8;
-	int s, ss;
-	u32 fuse2, s_enable, ss_disable, eu_disable;
-	u8 eu_mask = 0xff;
-
-	info = (struct intel_device_info *)&dev_priv->info;
-	fuse2 = I915_READ(GEN8_FUSE2);
-	s_enable = (fuse2 & GEN8_F2_S_ENA_MASK) >>
-		   GEN8_F2_S_ENA_SHIFT;
-	ss_disable = (fuse2 & GEN9_F2_SS_DIS_MASK) >>
-		     GEN9_F2_SS_DIS_SHIFT;
-
-	info->slice_total = hweight32(s_enable);
-	/*
-	 * The subslice disable field is global, i.e. it applies
-	 * to each of the enabled slices.
-	*/
-	info->subslice_per_slice = ss_max - hweight32(ss_disable);
-	info->subslice_total = info->slice_total *
-			       info->subslice_per_slice;
-
-	/*
-	 * Iterate through enabled slices and subslices to
-	 * count the total enabled EU.
-	*/
-	for (s = 0; s < s_max; s++) {
-		if (!(s_enable & (0x1 << s)))
-			/* skip disabled slice */
-			continue;
-
-		eu_disable = I915_READ(GEN9_EU_DISABLE(s));
-		for (ss = 0; ss < ss_max; ss++) {
-			int eu_per_ss;
-
-			if (ss_disable & (0x1 << ss))
-				/* skip disabled subslice */
-				continue;
-
-			eu_per_ss = eu_max - hweight8((eu_disable >> (ss*8)) &
-						      eu_mask);
-
-			/*
-			 * Record which subslice(s) has(have) 7 EUs. we
-			 * can tune the hash used to spread work among
-			 * subslices if they are unbalanced.
-			 */
-			if (eu_per_ss == 7)
-				info->subslice_7eu[s] |= 1 << ss;
-
-			info->eu_total += eu_per_ss;
-		}
-	}
-
-	/*
-	 * SKL is expected to always have a uniform distribution
-	 * of EU across subslices with the exception that any one
-	 * EU in any one subslice may be fused off for die
-	 * recovery. BXT is expected to be perfectly uniform in EU
-	 * distribution.
-	*/
-	info->eu_per_subslice = info->subslice_total ?
-				DIV_ROUND_UP(info->eu_total,
-					     info->subslice_total) : 0;
-	/*
-	 * SKL supports slice power gating on devices with more than
-	 * one slice, and supports EU power gating on devices with
-	 * more than one EU pair per subslice. BXT supports subslice
-	 * power gating on devices with more than one subslice, and
-	 * supports EU power gating on devices with more than one EU
-	 * pair per subslice.
-	*/
-	info->has_slice_pg = ((IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) &&
-			       (info->slice_total > 1));
-	info->has_subslice_pg = (IS_BROXTON(dev) && (info->subslice_total > 1));
-	info->has_eu_pg = (info->eu_per_subslice > 2);
-
-	if (IS_BROXTON(dev)) {
-#define IS_SS_DISABLED(_ss_disable, ss)    (_ss_disable & (0x1 << ss))
-		/*
-		 * There is a HW issue in 2x6 fused down parts that requires
-		 * Pooled EU to be enabled as a WA. The pool configuration
-		 * changes depending upon which subslice is fused down. This
-		 * doesn't affect if the device has all 3 subslices enabled.
-		 */
-		/* WaEnablePooledEuFor2x6:bxt */
-		info->has_pooled_eu = ((info->subslice_per_slice == 3) ||
-				       (info->subslice_per_slice == 2 &&
-					INTEL_REVID(dev) < BXT_REVID_C0));
-
-		info->min_eu_in_pool = 0;
-		if (info->has_pooled_eu) {
-			if (IS_SS_DISABLED(ss_disable, 0) ||
-			    IS_SS_DISABLED(ss_disable, 2))
-				info->min_eu_in_pool = 3;
-			else if (IS_SS_DISABLED(ss_disable, 1))
-				info->min_eu_in_pool = 6;
-			else
-				info->min_eu_in_pool = 9;
-		}
-#undef IS_SS_DISABLED
-	}
-}
-
-static void broadwell_sseu_info_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_device_info *info;
-	const int s_max = 3, ss_max = 3, eu_max = 8;
-	int s, ss;
-	u32 fuse2, eu_disable[s_max], s_enable, ss_disable;
-
-	fuse2 = I915_READ(GEN8_FUSE2);
-	s_enable = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
-	ss_disable = (fuse2 & GEN8_F2_SS_DIS_MASK) >> GEN8_F2_SS_DIS_SHIFT;
-
-	eu_disable[0] = I915_READ(GEN8_EU_DISABLE0) & GEN8_EU_DIS0_S0_MASK;
-	eu_disable[1] = (I915_READ(GEN8_EU_DISABLE0) >> GEN8_EU_DIS0_S1_SHIFT) |
-			((I915_READ(GEN8_EU_DISABLE1) & GEN8_EU_DIS1_S1_MASK) <<
-			 (32 - GEN8_EU_DIS0_S1_SHIFT));
-	eu_disable[2] = (I915_READ(GEN8_EU_DISABLE1) >> GEN8_EU_DIS1_S2_SHIFT) |
-			((I915_READ(GEN8_EU_DISABLE2) & GEN8_EU_DIS2_S2_MASK) <<
-			 (32 - GEN8_EU_DIS1_S2_SHIFT));
-
-
-	info = (struct intel_device_info *)&dev_priv->info;
-	info->slice_total = hweight32(s_enable);
-
-	/*
-	 * The subslice disable field is global, i.e. it applies
-	 * to each of the enabled slices.
-	 */
-	info->subslice_per_slice = ss_max - hweight32(ss_disable);
-	info->subslice_total = info->slice_total * info->subslice_per_slice;
-
-	/*
-	 * Iterate through enabled slices and subslices to
-	 * count the total enabled EU.
-	 */
-	for (s = 0; s < s_max; s++) {
-		if (!(s_enable & (0x1 << s)))
-			/* skip disabled slice */
-			continue;
-
-		for (ss = 0; ss < ss_max; ss++) {
-			u32 n_disabled;
-
-			if (ss_disable & (0x1 << ss))
-				/* skip disabled subslice */
-				continue;
-
-			n_disabled = hweight8(eu_disable[s] >> (ss * eu_max));
-
-			/*
-			 * Record which subslices have 7 EUs.
-			 */
-			if (eu_max - n_disabled == 7)
-				info->subslice_7eu[s] |= 1 << ss;
-
-			info->eu_total += eu_max - n_disabled;
-		}
-	}
-
-	/*
-	 * BDW is expected to always have a uniform distribution of EU across
-	 * subslices with the exception that any one EU in any one subslice may
-	 * be fused off for die recovery.
-	 */
-	info->eu_per_subslice = info->subslice_total ?
-		DIV_ROUND_UP(info->eu_total, info->subslice_total) : 0;
-
-	/*
-	 * BDW supports slice power gating on devices with more than
-	 * one slice.
-	 */
-	info->has_slice_pg = (info->slice_total > 1);
-	info->has_subslice_pg = 0;
-	info->has_eu_pg = 0;
-}
-
-/*
- * Determine various intel_device_info fields at runtime.
- *
- * Use it when either:
- *   - it's judged too laborious to fill n static structures with the limit
- *     when a simple if statement does the job,
- *   - run-time checks (eg read fuse/strap registers) are needed.
- *
- * This function needs to be called:
- *   - after the MMIO has been setup as we are reading registers,
- *   - after the PCH has been detected,
- *   - before the first usage of the fields it can tweak.
- */
-static void intel_device_info_runtime_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_device_info *info;
-	enum pipe pipe;
-
-	info = (struct intel_device_info *)&dev_priv->info;
-
-	/*
-	 * Skylake and Broxton currently don't expose the topmost plane as its
-	 * use is exclusive with the legacy cursor and we only want to expose
-	 * one of those, not both. Until we can safely expose the topmost plane
-	 * as a DRM_PLANE_TYPE_CURSOR with all the features exposed/supported,
-	 * we don't expose the topmost plane at all to prevent ABI breakage
-	 * down the line.
-	 */
-	if (IS_BROXTON(dev)) {
-		info->num_sprites[PIPE_A] = 2;
-		info->num_sprites[PIPE_B] = 2;
-		info->num_sprites[PIPE_C] = 1;
-	} else if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev))
-		for_each_pipe(dev_priv, pipe)
-			info->num_sprites[pipe] = 2;
-	else
-		for_each_pipe(dev_priv, pipe)
-			info->num_sprites[pipe] = 1;
-
-	if (i915.disable_display) {
-		DRM_INFO("Display disabled (module parameter)\n");
-		info->num_pipes = 0;
-	} else if (info->num_pipes > 0 &&
-		   (IS_GEN7(dev_priv) || IS_GEN8(dev_priv)) &&
-		   HAS_PCH_SPLIT(dev)) {
-		u32 fuse_strap = I915_READ(FUSE_STRAP);
-		u32 sfuse_strap = I915_READ(SFUSE_STRAP);
-
-		/*
-		 * SFUSE_STRAP is supposed to have a bit signalling the display
-		 * is fused off. Unfortunately it seems that, at least in
-		 * certain cases, fused off display means that PCH display
-		 * reads don't land anywhere. In that case, we read 0s.
-		 *
-		 * On CPT/PPT, we can detect this case as SFUSE_STRAP_FUSE_LOCK
-		 * should be set when taking over after the firmware.
-		 */
-		if (fuse_strap & ILK_INTERNAL_DISPLAY_DISABLE ||
-		    sfuse_strap & SFUSE_STRAP_DISPLAY_DISABLED ||
-		    (dev_priv->pch_type == PCH_CPT &&
-		     !(sfuse_strap & SFUSE_STRAP_FUSE_LOCK))) {
-			DRM_INFO("Display fused off, disabling\n");
-			info->num_pipes = 0;
-		} else if (fuse_strap & IVB_PIPE_C_DISABLE) {
-			DRM_INFO("PipeC fused off\n");
-			info->num_pipes -= 1;
-		}
-	} else if (info->num_pipes > 0 && IS_GEN9(dev_priv)) {
-		u32 dfsm = I915_READ(SKL_DFSM);
-		u8 disabled_mask = 0;
-		bool invalid;
-		int num_bits;
-
-		if (dfsm & SKL_DFSM_PIPE_A_DISABLE)
-			disabled_mask |= BIT(PIPE_A);
-		if (dfsm & SKL_DFSM_PIPE_B_DISABLE)
-			disabled_mask |= BIT(PIPE_B);
-		if (dfsm & SKL_DFSM_PIPE_C_DISABLE)
-			disabled_mask |= BIT(PIPE_C);
-
-		num_bits = hweight8(disabled_mask);
-
-		switch (disabled_mask) {
-		case BIT(PIPE_A):
-		case BIT(PIPE_B):
-		case BIT(PIPE_A) | BIT(PIPE_B):
-		case BIT(PIPE_A) | BIT(PIPE_C):
-			invalid = true;
-			break;
-		default:
-			invalid = false;
-		}
-
-		if (num_bits > info->num_pipes || invalid)
-			DRM_ERROR("invalid pipe fuse configuration: 0x%x\n",
-				  disabled_mask);
-		else
-			info->num_pipes -= num_bits;
-	}
-
-	/* Initialize slice/subslice/EU info */
-	if (IS_CHERRYVIEW(dev))
-		cherryview_sseu_info_init(dev);
-	else if (IS_BROADWELL(dev))
-		broadwell_sseu_info_init(dev);
-	else if (INTEL_INFO(dev)->gen >= 9)
-		gen9_sseu_info_init(dev);
-
-	info->has_snoop = !info->has_llc;
-
-	/* Snooping is broken on BXT A stepping. */
-	if (IS_BXT_REVID(dev, 0, BXT_REVID_A1))
-		info->has_snoop = false;
-
-	DRM_DEBUG_DRIVER("slice total: %u\n", info->slice_total);
-	DRM_DEBUG_DRIVER("subslice total: %u\n", info->subslice_total);
-	DRM_DEBUG_DRIVER("subslice per slice: %u\n", info->subslice_per_slice);
-	DRM_DEBUG_DRIVER("EU total: %u\n", info->eu_total);
-	DRM_DEBUG_DRIVER("EU per subslice: %u\n", info->eu_per_subslice);
-	DRM_DEBUG_DRIVER("Has Pooled EU: %s\n", HAS_POOLED_EU(dev) ? "y" : "n");
-	if (HAS_POOLED_EU(dev))
-		DRM_DEBUG_DRIVER("Min EU in pool: %u\n", info->min_eu_in_pool);
-	DRM_DEBUG_DRIVER("has slice power gating: %s\n",
-			 info->has_slice_pg ? "y" : "n");
-	DRM_DEBUG_DRIVER("has subslice power gating: %s\n",
-			 info->has_subslice_pg ? "y" : "n");
-	DRM_DEBUG_DRIVER("has EU power gating: %s\n",
-			 info->has_eu_pg ? "y" : "n");
-
-	i915.enable_execlists =
-		intel_sanitize_enable_execlists(dev_priv,
-					       	i915.enable_execlists);
-
-	/*
-	 * i915.enable_ppgtt is read-only, so do an early pass to validate the
-	 * user's requested state against the hardware/driver capabilities.  We
-	 * do this now so that we can print out any log messages once rather
-	 * than every time we check intel_enable_ppgtt().
-	 */
-	i915.enable_ppgtt =
-		intel_sanitize_enable_ppgtt(dev_priv, i915.enable_ppgtt);
-	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt);
-}
-
-static void intel_init_dpio(struct drm_i915_private *dev_priv)
-{
-	/*
-	 * IOSF_PORT_DPIO is used for VLV x2 PHY (DP/HDMI B and C),
-	 * CHV x1 PHY (DP/HDMI D)
-	 * IOSF_PORT_DPIO_2 is used for CHV x2 PHY (DP/HDMI B and C)
-	 */
-	if (IS_CHERRYVIEW(dev_priv)) {
-		DPIO_PHY_IOSF_PORT(DPIO_PHY0) = IOSF_PORT_DPIO_2;
-		DPIO_PHY_IOSF_PORT(DPIO_PHY1) = IOSF_PORT_DPIO;
-	} else if (IS_VALLEYVIEW(dev_priv)) {
-		DPIO_PHY_IOSF_PORT(DPIO_PHY0) = IOSF_PORT_DPIO;
-	}
-}
-
-static int i915_workqueues_init(struct drm_i915_private *dev_priv)
-{
-	/*
-	 * The i915 workqueue is primarily used for batched retirement of
-	 * requests (and thus managing bo) once the task has been completed
-	 * by the GPU. i915_gem_retire_requests() is called directly when we
-	 * need high-priority retirement, such as waiting for an explicit
-	 * bo.
-	 *
-	 * It is also used for periodic low-priority events, such as
-	 * idle-timers and recording error state.
-	 *
-	 * All tasks on the workqueue are expected to acquire the dev mutex
-	 * so there is no point in running more than one instance of the
-	 * workqueue at any time.  Use an ordered one.
-	 */
-	dev_priv->wq = alloc_ordered_workqueue("i915", 0);
-	if (dev_priv->wq == NULL)
-		goto out_err;
-
-	dev_priv->hotplug.dp_wq = alloc_ordered_workqueue("i915-dp", 0);
-	if (dev_priv->hotplug.dp_wq == NULL)
-		goto out_free_wq;
-
-	dev_priv->gpu_error.hangcheck_wq =
-		alloc_ordered_workqueue("i915-hangcheck", 0);
-	if (dev_priv->gpu_error.hangcheck_wq == NULL)
-		goto out_free_dp_wq;
-
-	return 0;
-
-out_free_dp_wq:
-	destroy_workqueue(dev_priv->hotplug.dp_wq);
-out_free_wq:
-	destroy_workqueue(dev_priv->wq);
-out_err:
-	DRM_ERROR("Failed to allocate workqueues.\n");
-
-	return -ENOMEM;
-}
-
-static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
-{
-	destroy_workqueue(dev_priv->gpu_error.hangcheck_wq);
-	destroy_workqueue(dev_priv->hotplug.dp_wq);
-	destroy_workqueue(dev_priv->wq);
-}
-
-/**
- * i915_driver_init_early - setup state not requiring device access
- * @dev_priv: device private
- *
- * Initialize everything that is a "SW-only" state, that is state not
- * requiring accessing the device or exposing the driver via kernel internal
- * or userspace interfaces. Example steps belonging here: lock initialization,
- * system memory allocation, setting up device specific attributes and
- * function hooks not requiring accessing the device.
- */
-static int i915_driver_init_early(struct drm_i915_private *dev_priv,
-				  struct drm_device *dev,
-				  struct intel_device_info *info)
-{
-	struct intel_device_info *device_info;
-	int ret = 0;
-
-	if (i915_inject_load_failure())
-		return -ENODEV;
-
-	/* Setup the write-once "constant" device info */
-	device_info = (struct intel_device_info *)&dev_priv->info;
-	memcpy(device_info, info, sizeof(dev_priv->info));
-	device_info->device_id = dev->pdev->device;
-
-	BUG_ON(device_info->gen > sizeof(device_info->gen_mask) * BITS_PER_BYTE);
-	device_info->gen_mask = BIT(device_info->gen - 1);
-
-	spin_lock_init(&dev_priv->irq_lock);
-	spin_lock_init(&dev_priv->gpu_error.lock);
-	mutex_init(&dev_priv->backlight_lock);
-	spin_lock_init(&dev_priv->uncore.lock);
-	spin_lock_init(&dev_priv->mm.object_stat_lock);
-	spin_lock_init(&dev_priv->mmio_flip_lock);
-	mutex_init(&dev_priv->sb_lock);
-	mutex_init(&dev_priv->modeset_restore_lock);
-	mutex_init(&dev_priv->av_mutex);
-	mutex_init(&dev_priv->wm.wm_mutex);
-	mutex_init(&dev_priv->pps_mutex);
-
-	ret = i915_workqueues_init(dev_priv);
-	if (ret < 0)
-		return ret;
-
-	ret = intel_gvt_init(dev_priv);
-	if (ret < 0)
-		goto err_workqueues;
-
-	/* This must be called before any calls to HAS_PCH_* */
-	intel_detect_pch(dev);
-
-	intel_pm_setup(dev);
-	intel_init_dpio(dev_priv);
-	intel_power_domains_init(dev_priv);
-	intel_irq_init(dev_priv);
-	intel_init_display_hooks(dev_priv);
-	intel_init_clock_gating_hooks(dev_priv);
-	intel_init_audio_hooks(dev_priv);
-	i915_gem_load_init(dev);
-
-	intel_display_crc_init(dev);
-
-	i915_dump_device_info(dev_priv);
-
-	/* Not all pre-production machines fall into this category, only the
-	 * very first ones. Almost everything should work, except for maybe
-	 * suspend/resume. And we don't implement workarounds that affect only
-	 * pre-production machines. */
-	if (IS_HSW_EARLY_SDV(dev))
-		DRM_INFO("This is an early pre-production Haswell machine. "
-			 "It may not be fully functional.\n");
-
-	return 0;
-
-err_workqueues:
-	i915_workqueues_cleanup(dev_priv);
-	return ret;
-}
-
-/**
- * i915_driver_cleanup_early - cleanup the setup done in i915_driver_init_early()
- * @dev_priv: device private
- */
-static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv)
-{
-	i915_gem_load_cleanup(dev_priv->dev);
-	i915_workqueues_cleanup(dev_priv);
-}
-
-static int i915_mmio_setup(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	int mmio_bar;
-	int mmio_size;
-
-	mmio_bar = IS_GEN2(dev) ? 1 : 0;
-	/*
-	 * Before gen4, the registers and the GTT are behind different BARs.
-	 * However, from gen4 onwards, the registers and the GTT are shared
-	 * in the same BAR, so we want to restrict this ioremap from
-	 * clobbering the GTT which we want ioremap_wc instead. Fortunately,
-	 * the register BAR remains the same size for all the earlier
-	 * generations up to Ironlake.
-	 */
-	if (INTEL_INFO(dev)->gen < 5)
-		mmio_size = 512 * 1024;
-	else
-		mmio_size = 2 * 1024 * 1024;
-	dev_priv->regs = pci_iomap(dev->pdev, mmio_bar, mmio_size);
-	if (dev_priv->regs == NULL) {
-		DRM_ERROR("failed to map registers\n");
-
-		return -EIO;
-	}
-
-	/* Try to make sure MCHBAR is enabled before poking at it */
-	intel_setup_mchbar(dev);
-
-	return 0;
-}
-
-static void i915_mmio_cleanup(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-
-	intel_teardown_mchbar(dev);
-	pci_iounmap(dev->pdev, dev_priv->regs);
-}
-
-/**
- * i915_driver_init_mmio - setup device MMIO
- * @dev_priv: device private
- *
- * Setup minimal device state necessary for MMIO accesses later in the
- * initialization sequence. The setup here should avoid any other device-wide
- * side effects or exposing the driver via kernel internal or user space
- * interfaces.
- */
-static int i915_driver_init_mmio(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	int ret;
-
-	if (i915_inject_load_failure())
-		return -ENODEV;
-
-	if (i915_get_bridge_dev(dev))
-		return -EIO;
-
-	ret = i915_mmio_setup(dev);
-	if (ret < 0)
-		goto put_bridge;
-
-	intel_uncore_init(dev_priv);
-
-	return 0;
-
-put_bridge:
-	pci_dev_put(dev_priv->bridge_dev);
-
-	return ret;
-}
-
-/**
- * i915_driver_cleanup_mmio - cleanup the setup done in i915_driver_init_mmio()
- * @dev_priv: device private
- */
-static void i915_driver_cleanup_mmio(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	intel_uncore_fini(dev_priv);
-	i915_mmio_cleanup(dev);
-	pci_dev_put(dev_priv->bridge_dev);
-}
-
-/**
- * i915_driver_init_hw - setup state requiring device access
- * @dev_priv: device private
- *
- * Setup state that requires accessing the device, but doesn't require
- * exposing the driver via kernel internal or userspace interfaces.
- */
-static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-	uint32_t aperture_size;
-	int ret;
-
-	if (i915_inject_load_failure())
-		return -ENODEV;
-
-	intel_device_info_runtime_init(dev);
-
-	ret = i915_ggtt_init_hw(dev);
-	if (ret)
-		return ret;
-
-	ret = i915_ggtt_enable_hw(dev);
-	if (ret) {
-		DRM_ERROR("failed to enable GGTT\n");
-		goto out_ggtt;
-	}
-
-	/* WARNING: Apparently we must kick fbdev drivers before vgacon,
-	 * otherwise the vga fbdev driver falls over. */
-	ret = i915_kick_out_firmware_fb(dev_priv);
-	if (ret) {
-		DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
-		goto out_ggtt;
-	}
-
-	ret = i915_kick_out_vgacon(dev_priv);
-	if (ret) {
-		DRM_ERROR("failed to remove conflicting VGA console\n");
-		goto out_ggtt;
-	}
-
-	pci_set_master(dev->pdev);
-
-	/* overlay on gen2 is broken and can't address above 1G */
-	if (IS_GEN2(dev)) {
-		ret = dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(30));
-		if (ret) {
-			DRM_ERROR("failed to set DMA mask\n");
-
-			goto out_ggtt;
-		}
-	}
-
-
-	/* 965GM sometimes incorrectly writes to hardware status page (HWS)
-	 * using 32bit addressing, overwriting memory if HWS is located
-	 * above 4GB.
-	 *
-	 * The documentation also mentions an issue with undefined
-	 * behaviour if any general state is accessed within a page above 4GB,
-	 * which also needs to be handled carefully.
-	 */
-	if (IS_BROADWATER(dev) || IS_CRESTLINE(dev)) {
-		ret = dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(32));
-
-		if (ret) {
-			DRM_ERROR("failed to set DMA mask\n");
-
-			goto out_ggtt;
-		}
-	}
-
-	aperture_size = ggtt->mappable_end;
-
-	ggtt->mappable =
-		io_mapping_create_wc(ggtt->mappable_base,
-				     aperture_size);
-	if (!ggtt->mappable) {
-		ret = -EIO;
-		goto out_ggtt;
-	}
-
-	ggtt->mtrr = arch_phys_wc_add(ggtt->mappable_base,
-					      aperture_size);
-
-	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY,
-			   PM_QOS_DEFAULT_VALUE);
-
-	intel_uncore_sanitize(dev_priv);
-
-	intel_opregion_setup(dev_priv);
-
-	i915_gem_load_init_fences(dev_priv);
-
-	/* On the 945G/GM, the chipset reports the MSI capability on the
-	 * integrated graphics even though the support isn't actually there
-	 * according to the published specs.  It doesn't appear to function
-	 * correctly in testing on 945G.
-	 * This may be a side effect of MSI having been made available for PEG
-	 * and the registers being closely associated.
-	 *
-	 * According to chipset errata, on the 965GM, MSI interrupts may
-	 * be lost or delayed, but we use them anyways to avoid
-	 * stuck interrupts on some machines.
-	 */
-	if (!IS_I945G(dev) && !IS_I945GM(dev)) {
-		if (pci_enable_msi(dev->pdev) < 0)
-			DRM_DEBUG_DRIVER("can't enable MSI");
-	}
-
-	return 0;
-
-out_ggtt:
-	i915_ggtt_cleanup_hw(dev);
-
-	return ret;
-}
-
-/**
- * i915_driver_cleanup_hw - cleanup the setup done in i915_driver_init_hw()
- * @dev_priv: device private
- */
-static void i915_driver_cleanup_hw(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-
-	if (dev->pdev->msi_enabled)
-		pci_disable_msi(dev->pdev);
-
-	pm_qos_remove_request(&dev_priv->pm_qos);
-	arch_phys_wc_del(ggtt->mtrr);
-	io_mapping_free(ggtt->mappable);
-	i915_ggtt_cleanup_hw(dev);
-}
-
-/**
- * i915_driver_register - register the driver with the rest of the system
- * @dev_priv: device private
- *
- * Perform any steps necessary to make the driver available via kernel
- * internal or userspace interfaces.
- */
-static void i915_driver_register(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	i915_gem_shrinker_init(dev_priv);
-	/*
-	 * Notify a valid surface after modesetting,
-	 * when running inside a VM.
-	 */
-	if (intel_vgpu_active(dev_priv))
-		I915_WRITE(vgtif_reg(display_ready), VGT_DRV_DISPLAY_READY);
-
-	i915_setup_sysfs(dev);
-
-	if (INTEL_INFO(dev_priv)->num_pipes) {
-		/* Must be done after probing outputs */
-		intel_opregion_register(dev_priv);
-		acpi_video_register();
-	}
-
-	if (IS_GEN5(dev_priv))
-		intel_gpu_ips_init(dev_priv);
-
-	i915_audio_component_init(dev_priv);
-}
-
-/**
- * i915_driver_unregister - cleanup the registration done in i915_driver_regiser()
- * @dev_priv: device private
- */
-static void i915_driver_unregister(struct drm_i915_private *dev_priv)
-{
-	i915_audio_component_cleanup(dev_priv);
-	intel_gpu_ips_teardown();
-	acpi_video_unregister();
-	intel_opregion_unregister(dev_priv);
-	i915_teardown_sysfs(dev_priv->dev);
-	i915_gem_shrinker_cleanup(dev_priv);
-}
-
-/**
- * i915_driver_load - setup chip and create an initial config
- * @dev: DRM device
- * @flags: startup flags
- *
- * The driver load routine has to do several things:
- *   - drive output discovery via intel_modeset_init()
- *   - initialize the memory manager
- *   - allocate initial config memory
- *   - setup the DRM framebuffer with the allocated memory
- */
-int i915_driver_load(struct drm_device *dev, unsigned long flags)
-{
-	struct drm_i915_private *dev_priv;
-	int ret = 0;
-
-	dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL);
-	if (dev_priv == NULL)
-		return -ENOMEM;
-
-	dev->dev_private = dev_priv;
-	/* Must be set before calling __i915_printk */
-	dev_priv->dev = dev;
-
-	ret = i915_driver_init_early(dev_priv, dev,
-				     (struct intel_device_info *)flags);
-
-	if (ret < 0)
-		goto out_free_priv;
-
-	intel_runtime_pm_get(dev_priv);
-
-	ret = i915_driver_init_mmio(dev_priv);
-	if (ret < 0)
-		goto out_runtime_pm_put;
-
-	ret = i915_driver_init_hw(dev_priv);
-	if (ret < 0)
-		goto out_cleanup_mmio;
-
-	/*
-	 * TODO: move the vblank init and parts of modeset init steps into one
-	 * of the i915_driver_init_/i915_driver_register functions according
-	 * to the role/effect of the given init step.
-	 */
-	if (INTEL_INFO(dev)->num_pipes) {
-		ret = drm_vblank_init(dev, INTEL_INFO(dev)->num_pipes);
-		if (ret)
-			goto out_cleanup_hw;
-	}
-
-	ret = i915_load_modeset_init(dev);
-	if (ret < 0)
-		goto out_cleanup_vblank;
-
-	i915_driver_register(dev_priv);
-
-	intel_runtime_pm_enable(dev_priv);
-
-	intel_runtime_pm_put(dev_priv);
-
-	return 0;
-
-out_cleanup_vblank:
-	drm_vblank_cleanup(dev);
-out_cleanup_hw:
-	i915_driver_cleanup_hw(dev_priv);
-out_cleanup_mmio:
-	i915_driver_cleanup_mmio(dev_priv);
-out_runtime_pm_put:
-	intel_runtime_pm_put(dev_priv);
-	i915_driver_cleanup_early(dev_priv);
-out_free_priv:
-	i915_load_error(dev_priv, "Device initialization failed (%d)\n", ret);
-
-	kfree(dev_priv);
-
-	return ret;
-}
-
-int i915_driver_unload(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	intel_fbdev_fini(dev);
-
-	intel_gvt_cleanup(dev_priv);
-
-	ret = i915_gem_suspend(dev);
-	if (ret) {
-		DRM_ERROR("failed to idle hardware: %d\n", ret);
-		return ret;
-	}
-
-	intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
-
-	i915_driver_unregister(dev_priv);
-
-	drm_vblank_cleanup(dev);
-
-	intel_modeset_cleanup(dev);
-
-	/*
-	 * free the memory space allocated for the child device
-	 * config parsed from VBT
-	 */
-	if (dev_priv->vbt.child_dev && dev_priv->vbt.child_dev_num) {
-		kfree(dev_priv->vbt.child_dev);
-		dev_priv->vbt.child_dev = NULL;
-		dev_priv->vbt.child_dev_num = 0;
-	}
-	kfree(dev_priv->vbt.sdvo_lvds_vbt_mode);
-	dev_priv->vbt.sdvo_lvds_vbt_mode = NULL;
-	kfree(dev_priv->vbt.lfp_lvds_vbt_mode);
-	dev_priv->vbt.lfp_lvds_vbt_mode = NULL;
-
-	vga_switcheroo_unregister_client(dev->pdev);
-	vga_client_register(dev->pdev, NULL, NULL, NULL);
-
-	intel_csr_ucode_fini(dev_priv);
-
-	/* Free error state after interrupts are fully disabled. */
-	cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
-	i915_destroy_error_state(dev);
-
-	/* Flush any outstanding unpin_work. */
-	flush_workqueue(dev_priv->wq);
-
-	intel_guc_fini(dev);
-	i915_gem_fini(dev);
-	intel_fbc_cleanup_cfb(dev_priv);
-
-	intel_power_domains_fini(dev_priv);
-
-	i915_driver_cleanup_hw(dev_priv);
-	i915_driver_cleanup_mmio(dev_priv);
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
-
-	i915_driver_cleanup_early(dev_priv);
-	kfree(dev_priv);
-
-	return 0;
-}
-
-int i915_driver_open(struct drm_device *dev, struct drm_file *file)
-{
-	int ret;
-
-	ret = i915_gem_open(dev, file);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-/**
- * i915_driver_lastclose - clean up after all DRM clients have exited
- * @dev: DRM device
- *
- * Take care of cleaning up after all DRM clients have exited.  In the
- * mode setting case, we want to restore the kernel's initial mode (just
- * in case the last client left us in a bad state).
- *
- * Additionally, in the non-mode setting case, we'll tear down the GTT
- * and DMA structures, since the kernel won't be using them, and clea
- * up any GEM state.
- */
-void i915_driver_lastclose(struct drm_device *dev)
-{
-	intel_fbdev_restore_mode(dev);
-	vga_switcheroo_process_delayed_switch();
-}
-
-void i915_driver_preclose(struct drm_device *dev, struct drm_file *file)
-{
-	mutex_lock(&dev->struct_mutex);
-	i915_gem_context_close(dev, file);
-	i915_gem_release(dev, file);
-	mutex_unlock(&dev->struct_mutex);
-}
-
-void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
-{
-	struct drm_i915_file_private *file_priv = file->driver_priv;
-
-	kfree(file_priv);
-}
-
-static int
-i915_gem_reject_pin_ioctl(struct drm_device *dev, void *data,
-			  struct drm_file *file)
-{
-	return -ENODEV;
-}
-
-const struct drm_ioctl_desc i915_ioctls[] = {
-	DRM_IOCTL_DEF_DRV(I915_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_FLUSH, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_FLIP, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_BATCHBUFFER, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_IRQ_EMIT, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_SETPARAM, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER, i915_gem_execbuffer, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2, i915_gem_execbuffer2, DRM_AUTH|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_PIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_UNPIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_BUSY, i915_gem_busy_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_SET_CACHING, i915_gem_set_caching_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_GET_CACHING, i915_gem_get_caching_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_THROTTLE, i915_gem_throttle_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_ENTERVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_LEAVEVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_CREATE, i915_gem_create_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_PREAD, i915_gem_pread_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_PWRITE, i915_gem_pwrite_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP, i915_gem_mmap_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_GTT, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_GET_APERTURE, i915_gem_get_aperture_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GET_PIPE_FROM_CRTC_ID, intel_get_pipe_from_crtc_id, 0),
-	DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_WAIT, i915_gem_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_CREATE, i915_gem_context_create_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_DESTROY, i915_gem_context_destroy_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_REG_READ, i915_reg_read_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GET_RESET_STATS, i915_gem_context_reset_stats_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_USERPTR, i915_gem_userptr_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_GETPARAM, i915_gem_context_getparam_ioctl, DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_SETPARAM, i915_gem_context_setparam_ioctl, DRM_RENDER_ALLOW),
-};
-
-int i915_max_ioctl = ARRAY_SIZE(i915_ioctls);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 3eb47fbcea73..b9a811750ca8 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -27,403 +27,92 @@
  *
  */
 
-#include <linux/device.h>
 #include <linux/acpi.h>
+#include <linux/device.h>
+#include <linux/oom.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/pnp.h>
+#include <linux/slab.h>
+#include <linux/vgaarb.h>
+#include <linux/vga_switcheroo.h>
+#include <linux/vt.h>
+#include <acpi/video.h>
+
 #include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
 #include <drm/i915_drm.h>
+
 #include "i915_drv.h"
 #include "i915_trace.h"
+#include "i915_vgpu.h"
 #include "intel_drv.h"
 
-#include <linux/console.h>
-#include <linux/module.h>
-#include <linux/pm_runtime.h>
-#include <linux/vga_switcheroo.h>
-#include <drm/drm_crtc_helper.h>
-
 static struct drm_driver driver;
 
-#define GEN_DEFAULT_PIPEOFFSETS \
-	.pipe_offsets = { PIPE_A_OFFSET, PIPE_B_OFFSET, \
-			  PIPE_C_OFFSET, PIPE_EDP_OFFSET }, \
-	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
-			   TRANSCODER_C_OFFSET, TRANSCODER_EDP_OFFSET }, \
-	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET }
-
-#define GEN_CHV_PIPEOFFSETS \
-	.pipe_offsets = { PIPE_A_OFFSET, PIPE_B_OFFSET, \
-			  CHV_PIPE_C_OFFSET }, \
-	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
-			   CHV_TRANSCODER_C_OFFSET, }, \
-	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET, \
-			     CHV_PALETTE_C_OFFSET }
-
-#define CURSOR_OFFSETS \
-	.cursor_offsets = { CURSOR_A_OFFSET, CURSOR_B_OFFSET, CHV_CURSOR_C_OFFSET }
-
-#define IVB_CURSOR_OFFSETS \
-	.cursor_offsets = { CURSOR_A_OFFSET, IVB_CURSOR_B_OFFSET, IVB_CURSOR_C_OFFSET }
-
-#define BDW_COLORS \
-	.color = { .degamma_lut_size = 512, .gamma_lut_size = 512 }
-#define CHV_COLORS \
-	.color = { .degamma_lut_size = 65, .gamma_lut_size = 257 }
-
-static const struct intel_device_info intel_i830_info = {
-	.gen = 2, .is_mobile = 1, .cursor_needs_physical = 1, .num_pipes = 2,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_845g_info = {
-	.gen = 2, .num_pipes = 1,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_i85x_info = {
-	.gen = 2, .is_i85x = 1, .is_mobile = 1, .num_pipes = 2,
-	.cursor_needs_physical = 1,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_i865g_info = {
-	.gen = 2, .num_pipes = 1,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_i915g_info = {
-	.gen = 3, .is_i915g = 1, .cursor_needs_physical = 1, .num_pipes = 2,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-static const struct intel_device_info intel_i915gm_info = {
-	.gen = 3, .is_mobile = 1, .num_pipes = 2,
-	.cursor_needs_physical = 1,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.supports_tv = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-static const struct intel_device_info intel_i945g_info = {
-	.gen = 3, .has_hotplug = 1, .cursor_needs_physical = 1, .num_pipes = 2,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-static const struct intel_device_info intel_i945gm_info = {
-	.gen = 3, .is_i945gm = 1, .is_mobile = 1, .num_pipes = 2,
-	.has_hotplug = 1, .cursor_needs_physical = 1,
-	.has_overlay = 1, .overlay_needs_physical = 1,
-	.supports_tv = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_i965g_info = {
-	.gen = 4, .is_broadwater = 1, .num_pipes = 2,
-	.has_hotplug = 1,
-	.has_overlay = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_i965gm_info = {
-	.gen = 4, .is_crestline = 1, .num_pipes = 2,
-	.is_mobile = 1, .has_fbc = 1, .has_hotplug = 1,
-	.has_overlay = 1,
-	.supports_tv = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_g33_info = {
-	.gen = 3, .is_g33 = 1, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_overlay = 1,
-	.ring_mask = RENDER_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_g45_info = {
-	.gen = 4, .is_g4x = 1, .need_gfx_hws = 1, .num_pipes = 2,
-	.has_pipe_cxsr = 1, .has_hotplug = 1,
-	.ring_mask = RENDER_RING | BSD_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_gm45_info = {
-	.gen = 4, .is_g4x = 1, .num_pipes = 2,
-	.is_mobile = 1, .need_gfx_hws = 1, .has_fbc = 1,
-	.has_pipe_cxsr = 1, .has_hotplug = 1,
-	.supports_tv = 1,
-	.ring_mask = RENDER_RING | BSD_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_pineview_info = {
-	.gen = 3, .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_overlay = 1,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_ironlake_d_info = {
-	.gen = 5, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.ring_mask = RENDER_RING | BSD_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_ironlake_m_info = {
-	.gen = 5, .is_mobile = 1, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING | BSD_RING,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_sandybridge_d_info = {
-	.gen = 6, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING,
-	.has_llc = 1,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-static const struct intel_device_info intel_sandybridge_m_info = {
-	.gen = 6, .is_mobile = 1, .num_pipes = 2,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING,
-	.has_llc = 1,
-	GEN_DEFAULT_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-};
-
-#define GEN7_FEATURES  \
-	.gen = 7, .num_pipes = 3, \
-	.need_gfx_hws = 1, .has_hotplug = 1, \
-	.has_fbc = 1, \
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING, \
-	.has_llc = 1, \
-	GEN_DEFAULT_PIPEOFFSETS, \
-	IVB_CURSOR_OFFSETS
-
-static const struct intel_device_info intel_ivybridge_d_info = {
-	GEN7_FEATURES,
-	.is_ivybridge = 1,
-};
-
-static const struct intel_device_info intel_ivybridge_m_info = {
-	GEN7_FEATURES,
-	.is_ivybridge = 1,
-	.is_mobile = 1,
-};
-
-static const struct intel_device_info intel_ivybridge_q_info = {
-	GEN7_FEATURES,
-	.is_ivybridge = 1,
-	.num_pipes = 0, /* legal, last one wins */
-};
-
-#define VLV_FEATURES  \
-	.gen = 7, .num_pipes = 2, \
-	.need_gfx_hws = 1, .has_hotplug = 1, \
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING, \
-	.display_mmio_offset = VLV_DISPLAY_BASE, \
-	GEN_DEFAULT_PIPEOFFSETS, \
-	CURSOR_OFFSETS
-
-static const struct intel_device_info intel_valleyview_m_info = {
-	VLV_FEATURES,
-	.is_valleyview = 1,
-	.is_mobile = 1,
-};
-
-static const struct intel_device_info intel_valleyview_d_info = {
-	VLV_FEATURES,
-	.is_valleyview = 1,
-};
-
-#define HSW_FEATURES  \
-	GEN7_FEATURES, \
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING, \
-	.has_ddi = 1, \
-	.has_fpga_dbg = 1
+static unsigned int i915_load_fail_count;
 
-static const struct intel_device_info intel_haswell_d_info = {
-	HSW_FEATURES,
-	.is_haswell = 1,
-};
+bool __i915_inject_load_failure(const char *func, int line)
+{
+	if (i915_load_fail_count >= i915.inject_load_failure)
+		return false;
 
-static const struct intel_device_info intel_haswell_m_info = {
-	HSW_FEATURES,
-	.is_haswell = 1,
-	.is_mobile = 1,
-};
+	if (++i915_load_fail_count == i915.inject_load_failure) {
+		DRM_INFO("Injecting failure at checkpoint %u [%s:%d]\n",
+			 i915.inject_load_failure, func, line);
+		return true;
+	}
 
-#define BDW_FEATURES \
-	HSW_FEATURES, \
-	BDW_COLORS
+	return false;
+}
 
-static const struct intel_device_info intel_broadwell_d_info = {
-	BDW_FEATURES,
-	.gen = 8,
-	.is_broadwell = 1,
-};
+#define FDO_BUG_URL "https://bugs.freedesktop.org/enter_bug.cgi?product=DRI"
+#define FDO_BUG_MSG "Please file a bug at " FDO_BUG_URL " against DRM/Intel " \
+		    "providing the dmesg log by booting with drm.debug=0xf"
 
-static const struct intel_device_info intel_broadwell_m_info = {
-	BDW_FEATURES,
-	.gen = 8, .is_mobile = 1,
-	.is_broadwell = 1,
-};
+void
+__i915_printk(struct drm_i915_private *dev_priv, const char *level,
+	      const char *fmt, ...)
+{
+	static bool shown_bug_once;
+	struct device *dev = dev_priv->drm.dev;
+	bool is_error = level[1] <= KERN_ERR[1];
+	bool is_debug = level[1] == KERN_DEBUG[1];
+	struct va_format vaf;
+	va_list args;
+
+	if (is_debug && !(drm_debug & DRM_UT_DRIVER))
+		return;
 
-static const struct intel_device_info intel_broadwell_gt3d_info = {
-	BDW_FEATURES,
-	.gen = 8,
-	.is_broadwell = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
-};
+	va_start(args, fmt);
 
-static const struct intel_device_info intel_broadwell_gt3m_info = {
-	BDW_FEATURES,
-	.gen = 8, .is_mobile = 1,
-	.is_broadwell = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
-};
+	vaf.fmt = fmt;
+	vaf.va = &args;
 
-static const struct intel_device_info intel_cherryview_info = {
-	.gen = 8, .num_pipes = 3,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
-	.is_cherryview = 1,
-	.display_mmio_offset = VLV_DISPLAY_BASE,
-	GEN_CHV_PIPEOFFSETS,
-	CURSOR_OFFSETS,
-	CHV_COLORS,
-};
+	dev_printk(level, dev, "[" DRM_NAME ":%ps] %pV",
+		   __builtin_return_address(0), &vaf);
 
-static const struct intel_device_info intel_skylake_info = {
-	BDW_FEATURES,
-	.is_skylake = 1,
-	.gen = 9,
-};
-
-static const struct intel_device_info intel_skylake_gt3_info = {
-	BDW_FEATURES,
-	.is_skylake = 1,
-	.gen = 9,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
-};
-
-static const struct intel_device_info intel_broxton_info = {
-	.is_preliminary = 1,
-	.is_broxton = 1,
-	.gen = 9,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
-	.num_pipes = 3,
-	.has_ddi = 1,
-	.has_fpga_dbg = 1,
-	.has_fbc = 1,
-	.has_pooled_eu = 0,
-	GEN_DEFAULT_PIPEOFFSETS,
-	IVB_CURSOR_OFFSETS,
-	BDW_COLORS,
-};
+	if (is_error && !shown_bug_once) {
+		dev_notice(dev, "%s", FDO_BUG_MSG);
+		shown_bug_once = true;
+	}
 
-static const struct intel_device_info intel_kabylake_info = {
-	BDW_FEATURES,
-	.is_kabylake = 1,
-	.gen = 9,
-};
+	va_end(args);
+}
 
-static const struct intel_device_info intel_kabylake_gt3_info = {
-	BDW_FEATURES,
-	.is_kabylake = 1,
-	.gen = 9,
-	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
-};
+static bool i915_error_injected(struct drm_i915_private *dev_priv)
+{
+	return i915.inject_load_failure &&
+	       i915_load_fail_count == i915.inject_load_failure;
+}
 
-/*
- * Make sure any device matches here are from most specific to most
- * general.  For example, since the Quanta match is based on the subsystem
- * and subvendor IDs, we need it to come before the more general IVB
- * PCI ID matches, otherwise we'll use the wrong info struct above.
- */
-static const struct pci_device_id pciidlist[] = {
-	INTEL_I830_IDS(&intel_i830_info),
-	INTEL_I845G_IDS(&intel_845g_info),
-	INTEL_I85X_IDS(&intel_i85x_info),
-	INTEL_I865G_IDS(&intel_i865g_info),
-	INTEL_I915G_IDS(&intel_i915g_info),
-	INTEL_I915GM_IDS(&intel_i915gm_info),
-	INTEL_I945G_IDS(&intel_i945g_info),
-	INTEL_I945GM_IDS(&intel_i945gm_info),
-	INTEL_I965G_IDS(&intel_i965g_info),
-	INTEL_G33_IDS(&intel_g33_info),
-	INTEL_I965GM_IDS(&intel_i965gm_info),
-	INTEL_GM45_IDS(&intel_gm45_info),
-	INTEL_G45_IDS(&intel_g45_info),
-	INTEL_PINEVIEW_IDS(&intel_pineview_info),
-	INTEL_IRONLAKE_D_IDS(&intel_ironlake_d_info),
-	INTEL_IRONLAKE_M_IDS(&intel_ironlake_m_info),
-	INTEL_SNB_D_IDS(&intel_sandybridge_d_info),
-	INTEL_SNB_M_IDS(&intel_sandybridge_m_info),
-	INTEL_IVB_Q_IDS(&intel_ivybridge_q_info), /* must be first IVB */
-	INTEL_IVB_M_IDS(&intel_ivybridge_m_info),
-	INTEL_IVB_D_IDS(&intel_ivybridge_d_info),
-	INTEL_HSW_D_IDS(&intel_haswell_d_info),
-	INTEL_HSW_M_IDS(&intel_haswell_m_info),
-	INTEL_VLV_M_IDS(&intel_valleyview_m_info),
-	INTEL_VLV_D_IDS(&intel_valleyview_d_info),
-	INTEL_BDW_GT12M_IDS(&intel_broadwell_m_info),
-	INTEL_BDW_GT12D_IDS(&intel_broadwell_d_info),
-	INTEL_BDW_GT3M_IDS(&intel_broadwell_gt3m_info),
-	INTEL_BDW_GT3D_IDS(&intel_broadwell_gt3d_info),
-	INTEL_CHV_IDS(&intel_cherryview_info),
-	INTEL_SKL_GT1_IDS(&intel_skylake_info),
-	INTEL_SKL_GT2_IDS(&intel_skylake_info),
-	INTEL_SKL_GT3_IDS(&intel_skylake_gt3_info),
-	INTEL_SKL_GT4_IDS(&intel_skylake_gt3_info),
-	INTEL_BXT_IDS(&intel_broxton_info),
-	INTEL_KBL_GT1_IDS(&intel_kabylake_info),
-	INTEL_KBL_GT2_IDS(&intel_kabylake_info),
-	INTEL_KBL_GT3_IDS(&intel_kabylake_gt3_info),
-	INTEL_KBL_GT4_IDS(&intel_kabylake_gt3_info),
-	{0, 0, 0}
-};
+#define i915_load_error(dev_priv, fmt, ...)				     \
+	__i915_printk(dev_priv,						     \
+		      i915_error_injected(dev_priv) ? KERN_DEBUG : KERN_ERR, \
+		      fmt, ##__VA_ARGS__)
 
-MODULE_DEVICE_TABLE(pci, pciidlist);
 
 static enum intel_pch intel_virt_detect_pch(struct drm_device *dev)
 {
@@ -453,9 +142,9 @@ static enum intel_pch intel_virt_detect_pch(struct drm_device *dev)
 	return ret;
 }
 
-void intel_detect_pch(struct drm_device *dev)
+static void intel_detect_pch(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct pci_dev *pch = NULL;
 
 	/* In all current cases, num_pipes is equivalent to the PCH_NOP setting
@@ -515,6 +204,10 @@ void intel_detect_pch(struct drm_device *dev)
 				DRM_DEBUG_KMS("Found SunrisePoint LP PCH\n");
 				WARN_ON(!IS_SKYLAKE(dev) &&
 					!IS_KABYLAKE(dev));
+			} else if (id == INTEL_PCH_KBP_DEVICE_ID_TYPE) {
+				dev_priv->pch_type = PCH_KBP;
+				DRM_DEBUG_KMS("Found KabyPoint PCH\n");
+				WARN_ON(!IS_KABYLAKE(dev));
 			} else if ((id == INTEL_PCH_P2X_DEVICE_ID_TYPE) ||
 				   (id == INTEL_PCH_P3X_DEVICE_ID_TYPE) ||
 				   ((id == INTEL_PCH_QEMU_DEVICE_ID_TYPE) &&
@@ -556,9 +249,1163 @@ bool i915_semaphore_is_enabled(struct drm_i915_private *dev_priv)
 	return true;
 }
 
+static int i915_getparam(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	drm_i915_getparam_t *param = data;
+	int value;
+
+	switch (param->param) {
+	case I915_PARAM_IRQ_ACTIVE:
+	case I915_PARAM_ALLOW_BATCHBUFFER:
+	case I915_PARAM_LAST_DISPATCH:
+		/* Reject all old ums/dri params. */
+		return -ENODEV;
+	case I915_PARAM_CHIPSET_ID:
+		value = dev->pdev->device;
+		break;
+	case I915_PARAM_REVISION:
+		value = dev->pdev->revision;
+		break;
+	case I915_PARAM_HAS_GEM:
+		value = 1;
+		break;
+	case I915_PARAM_NUM_FENCES_AVAIL:
+		value = dev_priv->num_fence_regs;
+		break;
+	case I915_PARAM_HAS_OVERLAY:
+		value = dev_priv->overlay ? 1 : 0;
+		break;
+	case I915_PARAM_HAS_PAGEFLIPPING:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_EXECBUF2:
+		/* depends on GEM */
+		value = 1;
+		break;
+	case I915_PARAM_HAS_BSD:
+		value = intel_engine_initialized(&dev_priv->engine[VCS]);
+		break;
+	case I915_PARAM_HAS_BLT:
+		value = intel_engine_initialized(&dev_priv->engine[BCS]);
+		break;
+	case I915_PARAM_HAS_VEBOX:
+		value = intel_engine_initialized(&dev_priv->engine[VECS]);
+		break;
+	case I915_PARAM_HAS_BSD2:
+		value = intel_engine_initialized(&dev_priv->engine[VCS2]);
+		break;
+	case I915_PARAM_HAS_RELAXED_FENCING:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_COHERENT_RINGS:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_EXEC_CONSTANTS:
+		value = INTEL_INFO(dev)->gen >= 4;
+		break;
+	case I915_PARAM_HAS_RELAXED_DELTA:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_GEN7_SOL_RESET:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_LLC:
+		value = HAS_LLC(dev);
+		break;
+	case I915_PARAM_HAS_WT:
+		value = HAS_WT(dev);
+		break;
+	case I915_PARAM_HAS_ALIASING_PPGTT:
+		value = USES_PPGTT(dev);
+		break;
+	case I915_PARAM_HAS_WAIT_TIMEOUT:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_SEMAPHORES:
+		value = i915_semaphore_is_enabled(dev_priv);
+		break;
+	case I915_PARAM_HAS_PRIME_VMAP_FLUSH:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_SECURE_BATCHES:
+		value = capable(CAP_SYS_ADMIN);
+		break;
+	case I915_PARAM_HAS_PINNED_BATCHES:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_EXEC_NO_RELOC:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_EXEC_HANDLE_LUT:
+		value = 1;
+		break;
+	case I915_PARAM_CMD_PARSER_VERSION:
+		value = i915_cmd_parser_get_version(dev_priv);
+		break;
+	case I915_PARAM_HAS_COHERENT_PHYS_GTT:
+		value = 1;
+		break;
+	case I915_PARAM_MMAP_VERSION:
+		value = 1;
+		break;
+	case I915_PARAM_SUBSLICE_TOTAL:
+		value = INTEL_INFO(dev)->subslice_total;
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_EU_TOTAL:
+		value = INTEL_INFO(dev)->eu_total;
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_HAS_GPU_RESET:
+		value = i915.enable_hangcheck && intel_has_gpu_reset(dev_priv);
+		break;
+	case I915_PARAM_HAS_RESOURCE_STREAMER:
+		value = HAS_RESOURCE_STREAMER(dev);
+		break;
+	case I915_PARAM_HAS_EXEC_SOFTPIN:
+		value = 1;
+		break;
+	case I915_PARAM_HAS_POOLED_EU:
+		value = HAS_POOLED_EU(dev);
+		break;
+	case I915_PARAM_MIN_EU_IN_POOL:
+		value = INTEL_INFO(dev)->min_eu_in_pool;
+		break;
+	default:
+		DRM_DEBUG("Unknown parameter %d\n", param->param);
+		return -EINVAL;
+	}
+
+	if (put_user(value, param->value))
+		return -EFAULT;
+
+	return 0;
+}
+
+static int i915_get_bridge_dev(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	dev_priv->bridge_dev = pci_get_bus_and_slot(0, PCI_DEVFN(0, 0));
+	if (!dev_priv->bridge_dev) {
+		DRM_ERROR("bridge device not found\n");
+		return -1;
+	}
+	return 0;
+}
+
+/* Allocate space for the MCH regs if needed, return nonzero on error */
+static int
+intel_alloc_mchbar_resource(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+	u32 temp_lo, temp_hi = 0;
+	u64 mchbar_addr;
+	int ret;
+
+	if (INTEL_INFO(dev)->gen >= 4)
+		pci_read_config_dword(dev_priv->bridge_dev, reg + 4, &temp_hi);
+	pci_read_config_dword(dev_priv->bridge_dev, reg, &temp_lo);
+	mchbar_addr = ((u64)temp_hi << 32) | temp_lo;
+
+	/* If ACPI doesn't have it, assume we need to allocate it ourselves */
+#ifdef CONFIG_PNP
+	if (mchbar_addr &&
+	    pnp_range_reserved(mchbar_addr, mchbar_addr + MCHBAR_SIZE))
+		return 0;
+#endif
+
+	/* Get some space for it */
+	dev_priv->mch_res.name = "i915 MCHBAR";
+	dev_priv->mch_res.flags = IORESOURCE_MEM;
+	ret = pci_bus_alloc_resource(dev_priv->bridge_dev->bus,
+				     &dev_priv->mch_res,
+				     MCHBAR_SIZE, MCHBAR_SIZE,
+				     PCIBIOS_MIN_MEM,
+				     0, pcibios_align_resource,
+				     dev_priv->bridge_dev);
+	if (ret) {
+		DRM_DEBUG_DRIVER("failed bus alloc: %d\n", ret);
+		dev_priv->mch_res.start = 0;
+		return ret;
+	}
+
+	if (INTEL_INFO(dev)->gen >= 4)
+		pci_write_config_dword(dev_priv->bridge_dev, reg + 4,
+				       upper_32_bits(dev_priv->mch_res.start));
+
+	pci_write_config_dword(dev_priv->bridge_dev, reg,
+			       lower_32_bits(dev_priv->mch_res.start));
+	return 0;
+}
+
+/* Setup MCHBAR if possible, return true if we should disable it again */
+static void
+intel_setup_mchbar(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int mchbar_reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+	u32 temp;
+	bool enabled;
+
+	if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev))
+		return;
+
+	dev_priv->mchbar_need_disable = false;
+
+	if (IS_I915G(dev) || IS_I915GM(dev)) {
+		pci_read_config_dword(dev_priv->bridge_dev, DEVEN, &temp);
+		enabled = !!(temp & DEVEN_MCHBAR_EN);
+	} else {
+		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
+		enabled = temp & 1;
+	}
+
+	/* If it's already enabled, don't have to do anything */
+	if (enabled)
+		return;
+
+	if (intel_alloc_mchbar_resource(dev))
+		return;
+
+	dev_priv->mchbar_need_disable = true;
+
+	/* Space is allocated or reserved, so enable it. */
+	if (IS_I915G(dev) || IS_I915GM(dev)) {
+		pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
+				       temp | DEVEN_MCHBAR_EN);
+	} else {
+		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
+		pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp | 1);
+	}
+}
+
+static void
+intel_teardown_mchbar(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int mchbar_reg = INTEL_INFO(dev)->gen >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+
+	if (dev_priv->mchbar_need_disable) {
+		if (IS_I915G(dev) || IS_I915GM(dev)) {
+			u32 deven_val;
+
+			pci_read_config_dword(dev_priv->bridge_dev, DEVEN,
+					      &deven_val);
+			deven_val &= ~DEVEN_MCHBAR_EN;
+			pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
+					       deven_val);
+		} else {
+			u32 mchbar_val;
+
+			pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg,
+					      &mchbar_val);
+			mchbar_val &= ~1;
+			pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg,
+					       mchbar_val);
+		}
+	}
+
+	if (dev_priv->mch_res.start)
+		release_resource(&dev_priv->mch_res);
+}
+
+/* true = enable decode, false = disable decoder */
+static unsigned int i915_vga_set_decode(void *cookie, bool state)
+{
+	struct drm_device *dev = cookie;
+
+	intel_modeset_vga_set_state(dev, state);
+	if (state)
+		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
+		       VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
+	else
+		return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
+}
+
+static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+	pm_message_t pmm = { .event = PM_EVENT_SUSPEND };
+
+	if (state == VGA_SWITCHEROO_ON) {
+		pr_info("switched on\n");
+		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
+		/* i915 resume handler doesn't set to D0 */
+		pci_set_power_state(dev->pdev, PCI_D0);
+		i915_resume_switcheroo(dev);
+		dev->switch_power_state = DRM_SWITCH_POWER_ON;
+	} else {
+		pr_info("switched off\n");
+		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
+		i915_suspend_switcheroo(dev, pmm);
+		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
+	}
+}
+
+static bool i915_switcheroo_can_switch(struct pci_dev *pdev)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+
+	/*
+	 * FIXME: open_count is protected by drm_global_mutex but that would lead to
+	 * locking inversion with the driver load path. And the access here is
+	 * completely racy anyway. So don't bother with locking for now.
+	 */
+	return dev->open_count == 0;
+}
+
+static const struct vga_switcheroo_client_ops i915_switcheroo_ops = {
+	.set_gpu_state = i915_switcheroo_set_state,
+	.reprobe = NULL,
+	.can_switch = i915_switcheroo_can_switch,
+};
+
+static void i915_gem_fini(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	/*
+	 * Neither the BIOS, ourselves or any other kernel
+	 * expects the system to be in execlists mode on startup,
+	 * so we need to reset the GPU back to legacy mode. And the only
+	 * known way to disable logical contexts is through a GPU reset.
+	 *
+	 * So in order to leave the system in a known default configuration,
+	 * always reset the GPU upon unload. Afterwards we then clean up the
+	 * GEM state tracking, flushing off the requests and leaving the
+	 * system in a known idle state.
+	 *
+	 * Note that is of the upmost importance that the GPU is idle and
+	 * all stray writes are flushed *before* we dismantle the backing
+	 * storage for the pinned objects.
+	 *
+	 * However, since we are uncertain that reseting the GPU on older
+	 * machines is a good idea, we don't - just in case it leaves the
+	 * machine in an unusable condition.
+	 */
+	if (HAS_HW_CONTEXTS(dev)) {
+		int reset = intel_gpu_reset(dev_priv, ALL_ENGINES);
+		WARN_ON(reset && reset != -ENODEV);
+	}
+
+	mutex_lock(&dev->struct_mutex);
+	i915_gem_reset(dev);
+	i915_gem_cleanup_engines(dev);
+	i915_gem_context_fini(dev);
+	mutex_unlock(&dev->struct_mutex);
+
+	WARN_ON(!list_empty(&to_i915(dev)->context_list));
+}
+
+static int i915_load_modeset_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int ret;
+
+	if (i915_inject_load_failure())
+		return -ENODEV;
+
+	ret = intel_bios_init(dev_priv);
+	if (ret)
+		DRM_INFO("failed to find VBIOS tables\n");
+
+	/* If we have > 1 VGA cards, then we need to arbitrate access
+	 * to the common VGA resources.
+	 *
+	 * If we are a secondary display controller (!PCI_DISPLAY_CLASS_VGA),
+	 * then we do not take part in VGA arbitration and the
+	 * vga_client_register() fails with -ENODEV.
+	 */
+	ret = vga_client_register(dev->pdev, dev, NULL, i915_vga_set_decode);
+	if (ret && ret != -ENODEV)
+		goto out;
+
+	intel_register_dsm_handler();
+
+	ret = vga_switcheroo_register_client(dev->pdev, &i915_switcheroo_ops, false);
+	if (ret)
+		goto cleanup_vga_client;
+
+	/* must happen before intel_power_domains_init_hw() on VLV/CHV */
+	intel_update_rawclk(dev_priv);
+
+	intel_power_domains_init_hw(dev_priv, false);
+
+	intel_csr_ucode_init(dev_priv);
+
+	ret = intel_irq_install(dev_priv);
+	if (ret)
+		goto cleanup_csr;
+
+	intel_setup_gmbus(dev);
+
+	/* Important: The output setup functions called by modeset_init need
+	 * working irqs for e.g. gmbus and dp aux transfers. */
+	intel_modeset_init(dev);
+
+	intel_guc_init(dev);
+
+	ret = i915_gem_init(dev);
+	if (ret)
+		goto cleanup_irq;
+
+	intel_modeset_gem_init(dev);
+
+	if (INTEL_INFO(dev)->num_pipes == 0)
+		return 0;
+
+	ret = intel_fbdev_init(dev);
+	if (ret)
+		goto cleanup_gem;
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	intel_hpd_init(dev_priv);
+
+	drm_kms_helper_poll_init(dev);
+
+	return 0;
+
+cleanup_gem:
+	i915_gem_fini(dev);
+cleanup_irq:
+	intel_guc_fini(dev);
+	drm_irq_uninstall(dev);
+	intel_teardown_gmbus(dev);
+cleanup_csr:
+	intel_csr_ucode_fini(dev_priv);
+	intel_power_domains_fini(dev_priv);
+	vga_switcheroo_unregister_client(dev->pdev);
+cleanup_vga_client:
+	vga_client_register(dev->pdev, NULL, NULL, NULL);
+out:
+	return ret;
+}
+
+#if IS_ENABLED(CONFIG_FB)
+static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
+{
+	struct apertures_struct *ap;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	struct i915_ggtt *ggtt = &dev_priv->ggtt;
+	bool primary;
+	int ret;
+
+	ap = alloc_apertures(1);
+	if (!ap)
+		return -ENOMEM;
+
+	ap->ranges[0].base = ggtt->mappable_base;
+	ap->ranges[0].size = ggtt->mappable_end;
+
+	primary =
+		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
+
+	ret = remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
+
+	kfree(ap);
+
+	return ret;
+}
+#else
+static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
+{
+	return 0;
+}
+#endif
+
+#if !defined(CONFIG_VGA_CONSOLE)
+static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
+{
+	return 0;
+}
+#elif !defined(CONFIG_DUMMY_CONSOLE)
+static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
+{
+	return -ENODEV;
+}
+#else
+static int i915_kick_out_vgacon(struct drm_i915_private *dev_priv)
+{
+	int ret = 0;
+
+	DRM_INFO("Replacing VGA console driver\n");
+
+	console_lock();
+	if (con_is_bound(&vga_con))
+		ret = do_take_over_console(&dummy_con, 0, MAX_NR_CONSOLES - 1, 1);
+	if (ret == 0) {
+		ret = do_unregister_con_driver(&vga_con);
+
+		/* Ignore "already unregistered". */
+		if (ret == -ENODEV)
+			ret = 0;
+	}
+	console_unlock();
+
+	return ret;
+}
+#endif
+
+static void intel_init_dpio(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * IOSF_PORT_DPIO is used for VLV x2 PHY (DP/HDMI B and C),
+	 * CHV x1 PHY (DP/HDMI D)
+	 * IOSF_PORT_DPIO_2 is used for CHV x2 PHY (DP/HDMI B and C)
+	 */
+	if (IS_CHERRYVIEW(dev_priv)) {
+		DPIO_PHY_IOSF_PORT(DPIO_PHY0) = IOSF_PORT_DPIO_2;
+		DPIO_PHY_IOSF_PORT(DPIO_PHY1) = IOSF_PORT_DPIO;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		DPIO_PHY_IOSF_PORT(DPIO_PHY0) = IOSF_PORT_DPIO;
+	}
+}
+
+static int i915_workqueues_init(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * The i915 workqueue is primarily used for batched retirement of
+	 * requests (and thus managing bo) once the task has been completed
+	 * by the GPU. i915_gem_retire_requests() is called directly when we
+	 * need high-priority retirement, such as waiting for an explicit
+	 * bo.
+	 *
+	 * It is also used for periodic low-priority events, such as
+	 * idle-timers and recording error state.
+	 *
+	 * All tasks on the workqueue are expected to acquire the dev mutex
+	 * so there is no point in running more than one instance of the
+	 * workqueue at any time.  Use an ordered one.
+	 */
+	dev_priv->wq = alloc_ordered_workqueue("i915", 0);
+	if (dev_priv->wq == NULL)
+		goto out_err;
+
+	dev_priv->hotplug.dp_wq = alloc_ordered_workqueue("i915-dp", 0);
+	if (dev_priv->hotplug.dp_wq == NULL)
+		goto out_free_wq;
+
+	return 0;
+
+out_free_wq:
+	destroy_workqueue(dev_priv->wq);
+out_err:
+	DRM_ERROR("Failed to allocate workqueues.\n");
+
+	return -ENOMEM;
+}
+
+static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
+{
+	destroy_workqueue(dev_priv->hotplug.dp_wq);
+	destroy_workqueue(dev_priv->wq);
+}
+
+/**
+ * i915_driver_init_early - setup state not requiring device access
+ * @dev_priv: device private
+ *
+ * Initialize everything that is a "SW-only" state, that is state not
+ * requiring accessing the device or exposing the driver via kernel internal
+ * or userspace interfaces. Example steps belonging here: lock initialization,
+ * system memory allocation, setting up device specific attributes and
+ * function hooks not requiring accessing the device.
+ */
+static int i915_driver_init_early(struct drm_i915_private *dev_priv,
+				  const struct pci_device_id *ent)
+{
+	const struct intel_device_info *match_info =
+		(struct intel_device_info *)ent->driver_data;
+	struct intel_device_info *device_info;
+	int ret = 0;
+
+	if (i915_inject_load_failure())
+		return -ENODEV;
+
+	/* Setup the write-once "constant" device info */
+	device_info = mkwrite_device_info(dev_priv);
+	memcpy(device_info, match_info, sizeof(*device_info));
+	device_info->device_id = dev_priv->drm.pdev->device;
+
+	BUG_ON(device_info->gen > sizeof(device_info->gen_mask) * BITS_PER_BYTE);
+	device_info->gen_mask = BIT(device_info->gen - 1);
+
+	spin_lock_init(&dev_priv->irq_lock);
+	spin_lock_init(&dev_priv->gpu_error.lock);
+	mutex_init(&dev_priv->backlight_lock);
+	spin_lock_init(&dev_priv->uncore.lock);
+	spin_lock_init(&dev_priv->mm.object_stat_lock);
+	spin_lock_init(&dev_priv->mmio_flip_lock);
+	mutex_init(&dev_priv->sb_lock);
+	mutex_init(&dev_priv->modeset_restore_lock);
+	mutex_init(&dev_priv->av_mutex);
+	mutex_init(&dev_priv->wm.wm_mutex);
+	mutex_init(&dev_priv->pps_mutex);
+
+	ret = i915_workqueues_init(dev_priv);
+	if (ret < 0)
+		return ret;
+
+	ret = intel_gvt_init(dev_priv);
+	if (ret < 0)
+		goto err_workqueues;
+
+	/* This must be called before any calls to HAS_PCH_* */
+	intel_detect_pch(&dev_priv->drm);
+
+	intel_pm_setup(&dev_priv->drm);
+	intel_init_dpio(dev_priv);
+	intel_power_domains_init(dev_priv);
+	intel_irq_init(dev_priv);
+	intel_init_display_hooks(dev_priv);
+	intel_init_clock_gating_hooks(dev_priv);
+	intel_init_audio_hooks(dev_priv);
+	i915_gem_load_init(&dev_priv->drm);
+
+	intel_display_crc_init(&dev_priv->drm);
+
+	intel_device_info_dump(dev_priv);
+
+	/* Not all pre-production machines fall into this category, only the
+	 * very first ones. Almost everything should work, except for maybe
+	 * suspend/resume. And we don't implement workarounds that affect only
+	 * pre-production machines. */
+	if (IS_HSW_EARLY_SDV(dev_priv))
+		DRM_INFO("This is an early pre-production Haswell machine. "
+			 "It may not be fully functional.\n");
+
+	return 0;
+
+err_workqueues:
+	i915_workqueues_cleanup(dev_priv);
+	return ret;
+}
+
+/**
+ * i915_driver_cleanup_early - cleanup the setup done in i915_driver_init_early()
+ * @dev_priv: device private
+ */
+static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv)
+{
+	i915_gem_load_cleanup(&dev_priv->drm);
+	i915_workqueues_cleanup(dev_priv);
+}
+
+static int i915_mmio_setup(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int mmio_bar;
+	int mmio_size;
+
+	mmio_bar = IS_GEN2(dev) ? 1 : 0;
+	/*
+	 * Before gen4, the registers and the GTT are behind different BARs.
+	 * However, from gen4 onwards, the registers and the GTT are shared
+	 * in the same BAR, so we want to restrict this ioremap from
+	 * clobbering the GTT which we want ioremap_wc instead. Fortunately,
+	 * the register BAR remains the same size for all the earlier
+	 * generations up to Ironlake.
+	 */
+	if (INTEL_INFO(dev)->gen < 5)
+		mmio_size = 512 * 1024;
+	else
+		mmio_size = 2 * 1024 * 1024;
+	dev_priv->regs = pci_iomap(dev->pdev, mmio_bar, mmio_size);
+	if (dev_priv->regs == NULL) {
+		DRM_ERROR("failed to map registers\n");
+
+		return -EIO;
+	}
+
+	/* Try to make sure MCHBAR is enabled before poking at it */
+	intel_setup_mchbar(dev);
+
+	return 0;
+}
+
+static void i915_mmio_cleanup(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	intel_teardown_mchbar(dev);
+	pci_iounmap(dev->pdev, dev_priv->regs);
+}
+
+/**
+ * i915_driver_init_mmio - setup device MMIO
+ * @dev_priv: device private
+ *
+ * Setup minimal device state necessary for MMIO accesses later in the
+ * initialization sequence. The setup here should avoid any other device-wide
+ * side effects or exposing the driver via kernel internal or user space
+ * interfaces.
+ */
+static int i915_driver_init_mmio(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	int ret;
+
+	if (i915_inject_load_failure())
+		return -ENODEV;
+
+	if (i915_get_bridge_dev(dev))
+		return -EIO;
+
+	ret = i915_mmio_setup(dev);
+	if (ret < 0)
+		goto put_bridge;
+
+	intel_uncore_init(dev_priv);
+
+	return 0;
+
+put_bridge:
+	pci_dev_put(dev_priv->bridge_dev);
+
+	return ret;
+}
+
+/**
+ * i915_driver_cleanup_mmio - cleanup the setup done in i915_driver_init_mmio()
+ * @dev_priv: device private
+ */
+static void i915_driver_cleanup_mmio(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+
+	intel_uncore_fini(dev_priv);
+	i915_mmio_cleanup(dev);
+	pci_dev_put(dev_priv->bridge_dev);
+}
+
+static void intel_sanitize_options(struct drm_i915_private *dev_priv)
+{
+	i915.enable_execlists =
+		intel_sanitize_enable_execlists(dev_priv,
+						i915.enable_execlists);
+
+	/*
+	 * i915.enable_ppgtt is read-only, so do an early pass to validate the
+	 * user's requested state against the hardware/driver capabilities.  We
+	 * do this now so that we can print out any log messages once rather
+	 * than every time we check intel_enable_ppgtt().
+	 */
+	i915.enable_ppgtt =
+		intel_sanitize_enable_ppgtt(dev_priv, i915.enable_ppgtt);
+	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt);
+}
+
+/**
+ * i915_driver_init_hw - setup state requiring device access
+ * @dev_priv: device private
+ *
+ * Setup state that requires accessing the device, but doesn't require
+ * exposing the driver via kernel internal or userspace interfaces.
+ */
+static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct i915_ggtt *ggtt = &dev_priv->ggtt;
+	uint32_t aperture_size;
+	int ret;
+
+	if (i915_inject_load_failure())
+		return -ENODEV;
+
+	intel_device_info_runtime_init(dev_priv);
+
+	intel_sanitize_options(dev_priv);
+
+	ret = i915_ggtt_init_hw(dev);
+	if (ret)
+		return ret;
+
+	ret = i915_ggtt_enable_hw(dev);
+	if (ret) {
+		DRM_ERROR("failed to enable GGTT\n");
+		goto out_ggtt;
+	}
+
+	/* WARNING: Apparently we must kick fbdev drivers before vgacon,
+	 * otherwise the vga fbdev driver falls over. */
+	ret = i915_kick_out_firmware_fb(dev_priv);
+	if (ret) {
+		DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
+		goto out_ggtt;
+	}
+
+	ret = i915_kick_out_vgacon(dev_priv);
+	if (ret) {
+		DRM_ERROR("failed to remove conflicting VGA console\n");
+		goto out_ggtt;
+	}
+
+	pci_set_master(dev->pdev);
+
+	/* overlay on gen2 is broken and can't address above 1G */
+	if (IS_GEN2(dev)) {
+		ret = dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(30));
+		if (ret) {
+			DRM_ERROR("failed to set DMA mask\n");
+
+			goto out_ggtt;
+		}
+	}
+
+
+	/* 965GM sometimes incorrectly writes to hardware status page (HWS)
+	 * using 32bit addressing, overwriting memory if HWS is located
+	 * above 4GB.
+	 *
+	 * The documentation also mentions an issue with undefined
+	 * behaviour if any general state is accessed within a page above 4GB,
+	 * which also needs to be handled carefully.
+	 */
+	if (IS_BROADWATER(dev) || IS_CRESTLINE(dev)) {
+		ret = dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(32));
+
+		if (ret) {
+			DRM_ERROR("failed to set DMA mask\n");
+
+			goto out_ggtt;
+		}
+	}
+
+	aperture_size = ggtt->mappable_end;
+
+	ggtt->mappable =
+		io_mapping_create_wc(ggtt->mappable_base,
+				     aperture_size);
+	if (!ggtt->mappable) {
+		ret = -EIO;
+		goto out_ggtt;
+	}
+
+	ggtt->mtrr = arch_phys_wc_add(ggtt->mappable_base,
+					      aperture_size);
+
+	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY,
+			   PM_QOS_DEFAULT_VALUE);
+
+	intel_uncore_sanitize(dev_priv);
+
+	intel_opregion_setup(dev_priv);
+
+	i915_gem_load_init_fences(dev_priv);
+
+	/* On the 945G/GM, the chipset reports the MSI capability on the
+	 * integrated graphics even though the support isn't actually there
+	 * according to the published specs.  It doesn't appear to function
+	 * correctly in testing on 945G.
+	 * This may be a side effect of MSI having been made available for PEG
+	 * and the registers being closely associated.
+	 *
+	 * According to chipset errata, on the 965GM, MSI interrupts may
+	 * be lost or delayed, but we use them anyways to avoid
+	 * stuck interrupts on some machines.
+	 */
+	if (!IS_I945G(dev) && !IS_I945GM(dev)) {
+		if (pci_enable_msi(dev->pdev) < 0)
+			DRM_DEBUG_DRIVER("can't enable MSI");
+	}
+
+	return 0;
+
+out_ggtt:
+	i915_ggtt_cleanup_hw(dev);
+
+	return ret;
+}
+
+/**
+ * i915_driver_cleanup_hw - cleanup the setup done in i915_driver_init_hw()
+ * @dev_priv: device private
+ */
+static void i915_driver_cleanup_hw(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct i915_ggtt *ggtt = &dev_priv->ggtt;
+
+	if (dev->pdev->msi_enabled)
+		pci_disable_msi(dev->pdev);
+
+	pm_qos_remove_request(&dev_priv->pm_qos);
+	arch_phys_wc_del(ggtt->mtrr);
+	io_mapping_free(ggtt->mappable);
+	i915_ggtt_cleanup_hw(dev);
+}
+
+/**
+ * i915_driver_register - register the driver with the rest of the system
+ * @dev_priv: device private
+ *
+ * Perform any steps necessary to make the driver available via kernel
+ * internal or userspace interfaces.
+ */
+static void i915_driver_register(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+
+	i915_gem_shrinker_init(dev_priv);
+
+	/*
+	 * Notify a valid surface after modesetting,
+	 * when running inside a VM.
+	 */
+	if (intel_vgpu_active(dev_priv))
+		I915_WRITE(vgtif_reg(display_ready), VGT_DRV_DISPLAY_READY);
+
+	/* Reveal our presence to userspace */
+	if (drm_dev_register(dev, 0) == 0) {
+		i915_debugfs_register(dev_priv);
+		i915_setup_sysfs(dev);
+	} else
+		DRM_ERROR("Failed to register driver for userspace access!\n");
+
+	if (INTEL_INFO(dev_priv)->num_pipes) {
+		/* Must be done after probing outputs */
+		intel_opregion_register(dev_priv);
+		acpi_video_register();
+	}
+
+	if (IS_GEN5(dev_priv))
+		intel_gpu_ips_init(dev_priv);
+
+	i915_audio_component_init(dev_priv);
+
+	/*
+	 * Some ports require correctly set-up hpd registers for detection to
+	 * work properly (leading to ghost connected connector status), e.g. VGA
+	 * on gm45.  Hence we can only set up the initial fbdev config after hpd
+	 * irqs are fully enabled. We do it last so that the async config
+	 * cannot run before the connectors are registered.
+	 */
+	intel_fbdev_initial_config_async(dev);
+}
+
+/**
+ * i915_driver_unregister - cleanup the registration done in i915_driver_regiser()
+ * @dev_priv: device private
+ */
+static void i915_driver_unregister(struct drm_i915_private *dev_priv)
+{
+	i915_audio_component_cleanup(dev_priv);
+
+	intel_gpu_ips_teardown();
+	acpi_video_unregister();
+	intel_opregion_unregister(dev_priv);
+
+	i915_teardown_sysfs(&dev_priv->drm);
+	i915_debugfs_unregister(dev_priv);
+	drm_dev_unregister(&dev_priv->drm);
+
+	i915_gem_shrinker_cleanup(dev_priv);
+}
+
+/**
+ * i915_driver_load - setup chip and create an initial config
+ * @dev: DRM device
+ * @flags: startup flags
+ *
+ * The driver load routine has to do several things:
+ *   - drive output discovery via intel_modeset_init()
+ *   - initialize the memory manager
+ *   - allocate initial config memory
+ *   - setup the DRM framebuffer with the allocated memory
+ */
+int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	struct drm_i915_private *dev_priv;
+	int ret;
+
+	if (i915.nuclear_pageflip)
+		driver.driver_features |= DRIVER_ATOMIC;
+
+	ret = -ENOMEM;
+	dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL);
+	if (dev_priv)
+		ret = drm_dev_init(&dev_priv->drm, &driver, &pdev->dev);
+	if (ret) {
+		dev_printk(KERN_ERR, &pdev->dev,
+			   "[" DRM_NAME ":%s] allocation failed\n", __func__);
+		kfree(dev_priv);
+		return ret;
+	}
+
+	dev_priv->drm.pdev = pdev;
+	dev_priv->drm.dev_private = dev_priv;
+
+	ret = pci_enable_device(pdev);
+	if (ret)
+		goto out_free_priv;
+
+	pci_set_drvdata(pdev, &dev_priv->drm);
+
+	ret = i915_driver_init_early(dev_priv, ent);
+	if (ret < 0)
+		goto out_pci_disable;
+
+	intel_runtime_pm_get(dev_priv);
+
+	ret = i915_driver_init_mmio(dev_priv);
+	if (ret < 0)
+		goto out_runtime_pm_put;
+
+	ret = i915_driver_init_hw(dev_priv);
+	if (ret < 0)
+		goto out_cleanup_mmio;
+
+	/*
+	 * TODO: move the vblank init and parts of modeset init steps into one
+	 * of the i915_driver_init_/i915_driver_register functions according
+	 * to the role/effect of the given init step.
+	 */
+	if (INTEL_INFO(dev_priv)->num_pipes) {
+		ret = drm_vblank_init(&dev_priv->drm,
+				      INTEL_INFO(dev_priv)->num_pipes);
+		if (ret)
+			goto out_cleanup_hw;
+	}
+
+	ret = i915_load_modeset_init(&dev_priv->drm);
+	if (ret < 0)
+		goto out_cleanup_vblank;
+
+	i915_driver_register(dev_priv);
+
+	intel_runtime_pm_enable(dev_priv);
+
+	intel_runtime_pm_put(dev_priv);
+
+	return 0;
+
+out_cleanup_vblank:
+	drm_vblank_cleanup(&dev_priv->drm);
+out_cleanup_hw:
+	i915_driver_cleanup_hw(dev_priv);
+out_cleanup_mmio:
+	i915_driver_cleanup_mmio(dev_priv);
+out_runtime_pm_put:
+	intel_runtime_pm_put(dev_priv);
+	i915_driver_cleanup_early(dev_priv);
+out_pci_disable:
+	pci_disable_device(pdev);
+out_free_priv:
+	i915_load_error(dev_priv, "Device initialization failed (%d)\n", ret);
+	drm_dev_unref(&dev_priv->drm);
+	return ret;
+}
+
+void i915_driver_unload(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	intel_fbdev_fini(dev);
+
+	if (i915_gem_suspend(dev))
+		DRM_ERROR("failed to idle hardware; continuing to unload!\n");
+
+	intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
+
+	i915_driver_unregister(dev_priv);
+
+	drm_vblank_cleanup(dev);
+
+	intel_modeset_cleanup(dev);
+
+	/*
+	 * free the memory space allocated for the child device
+	 * config parsed from VBT
+	 */
+	if (dev_priv->vbt.child_dev && dev_priv->vbt.child_dev_num) {
+		kfree(dev_priv->vbt.child_dev);
+		dev_priv->vbt.child_dev = NULL;
+		dev_priv->vbt.child_dev_num = 0;
+	}
+	kfree(dev_priv->vbt.sdvo_lvds_vbt_mode);
+	dev_priv->vbt.sdvo_lvds_vbt_mode = NULL;
+	kfree(dev_priv->vbt.lfp_lvds_vbt_mode);
+	dev_priv->vbt.lfp_lvds_vbt_mode = NULL;
+
+	vga_switcheroo_unregister_client(dev->pdev);
+	vga_client_register(dev->pdev, NULL, NULL, NULL);
+
+	intel_csr_ucode_fini(dev_priv);
+
+	/* Free error state after interrupts are fully disabled. */
+	cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
+	i915_destroy_error_state(dev);
+
+	/* Flush any outstanding unpin_work. */
+	flush_workqueue(dev_priv->wq);
+
+	intel_guc_fini(dev);
+	i915_gem_fini(dev);
+	intel_fbc_cleanup_cfb(dev_priv);
+
+	intel_power_domains_fini(dev_priv);
+
+	i915_driver_cleanup_hw(dev_priv);
+	i915_driver_cleanup_mmio(dev_priv);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
+
+	i915_driver_cleanup_early(dev_priv);
+}
+
+static int i915_driver_open(struct drm_device *dev, struct drm_file *file)
+{
+	int ret;
+
+	ret = i915_gem_open(dev, file);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/**
+ * i915_driver_lastclose - clean up after all DRM clients have exited
+ * @dev: DRM device
+ *
+ * Take care of cleaning up after all DRM clients have exited.  In the
+ * mode setting case, we want to restore the kernel's initial mode (just
+ * in case the last client left us in a bad state).
+ *
+ * Additionally, in the non-mode setting case, we'll tear down the GTT
+ * and DMA structures, since the kernel won't be using them, and clea
+ * up any GEM state.
+ */
+static void i915_driver_lastclose(struct drm_device *dev)
+{
+	intel_fbdev_restore_mode(dev);
+	vga_switcheroo_process_delayed_switch();
+}
+
+static void i915_driver_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	mutex_lock(&dev->struct_mutex);
+	i915_gem_context_close(dev, file);
+	i915_gem_release(dev, file);
+	mutex_unlock(&dev->struct_mutex);
+}
+
+static void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+
+	kfree(file_priv);
+}
+
 static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_encoder *encoder;
 
 	drm_modeset_lock_all(dev);
@@ -583,7 +1430,7 @@ static bool suspend_to_idle(struct drm_i915_private *dev_priv)
 
 static int i915_drm_suspend(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	pci_power_t opregion_target_state;
 	int error;
 
@@ -650,7 +1497,7 @@ out:
 
 static int i915_drm_suspend_late(struct drm_device *drm_dev, bool hibernation)
 {
-	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(drm_dev);
 	bool fw_csr;
 	int ret;
 
@@ -712,7 +1559,7 @@ int i915_suspend_switcheroo(struct drm_device *dev, pm_message_t state)
 {
 	int error;
 
-	if (!dev || !dev->dev_private) {
+	if (!dev) {
 		DRM_ERROR("dev: %p\n", dev);
 		DRM_ERROR("DRM not initialized, aborting suspend.\n");
 		return -ENODEV;
@@ -734,7 +1581,7 @@ int i915_suspend_switcheroo(struct drm_device *dev, pm_message_t state)
 
 static int i915_drm_resume(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	disable_rpm_wakeref_asserts(dev_priv);
@@ -768,7 +1615,7 @@ static int i915_drm_resume(struct drm_device *dev)
 	mutex_lock(&dev->struct_mutex);
 	if (i915_gem_init_hw(dev)) {
 		DRM_ERROR("failed to re-initialize GPU, declaring wedged!\n");
-			atomic_or(I915_WEDGED, &dev_priv->gpu_error.reset_counter);
+		atomic_or(I915_WEDGED, &dev_priv->gpu_error.reset_counter);
 	}
 	mutex_unlock(&dev->struct_mutex);
 
@@ -814,7 +1661,7 @@ static int i915_drm_resume(struct drm_device *dev)
 
 static int i915_drm_resume_early(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	/*
@@ -926,7 +1773,7 @@ int i915_resume_switcheroo(struct drm_device *dev)
  */
 int i915_reset(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct i915_gpu_error *error = &dev_priv->gpu_error;
 	unsigned reset_counter;
 	int ret;
@@ -945,24 +1792,11 @@ int i915_reset(struct drm_i915_private *dev_priv)
 		goto error;
 	}
 
+	pr_notice("drm/i915: Resetting chip after gpu hang\n");
+
 	i915_gem_reset(dev);
 
 	ret = intel_gpu_reset(dev_priv, ALL_ENGINES);
-
-	/* Also reset the gpu hangman. */
-	if (error->stop_rings != 0) {
-		DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
-		error->stop_rings = 0;
-		if (ret == -ENODEV) {
-			DRM_INFO("Reset not implemented, but ignoring "
-				 "error for simulated gpu hangs\n");
-			ret = 0;
-		}
-	}
-
-	if (i915_stop_ring_allow_warn(dev_priv))
-		pr_notice("drm/i915: Resetting chip after gpu hang\n");
-
 	if (ret) {
 		if (ret != -ENODEV)
 			DRM_ERROR("Failed to reset chip: %i\n", ret);
@@ -1012,45 +1846,12 @@ error:
 	return ret;
 }
 
-static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
-{
-	struct intel_device_info *intel_info =
-		(struct intel_device_info *) ent->driver_data;
-
-	if (IS_PRELIMINARY_HW(intel_info) && !i915.preliminary_hw_support) {
-		DRM_INFO("This hardware requires preliminary hardware support.\n"
-			 "See CONFIG_DRM_I915_PRELIMINARY_HW_SUPPORT, and/or modparam preliminary_hw_support\n");
-		return -ENODEV;
-	}
-
-	/* Only bind to function 0 of the device. Early generations
-	 * used function 1 as a placeholder for multi-head. This causes
-	 * us confusion instead, especially on the systems where both
-	 * functions have the same PCI-ID!
-	 */
-	if (PCI_FUNC(pdev->devfn))
-		return -ENODEV;
-
-	if (vga_switcheroo_client_probe_defer(pdev))
-		return -EPROBE_DEFER;
-
-	return drm_get_pci_dev(pdev, ent, &driver);
-}
-
-static void
-i915_pci_remove(struct pci_dev *pdev)
-{
-	struct drm_device *dev = pci_get_drvdata(pdev);
-
-	drm_put_dev(dev);
-}
-
 static int i915_pm_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 
-	if (!drm_dev || !drm_dev->dev_private) {
+	if (!drm_dev) {
 		dev_err(dev, "DRM not initialized, aborting suspend.\n");
 		return -ENODEV;
 	}
@@ -1063,7 +1864,7 @@ static int i915_pm_suspend(struct device *dev)
 
 static int i915_pm_suspend_late(struct device *dev)
 {
-	struct drm_device *drm_dev = dev_to_i915(dev)->dev;
+	struct drm_device *drm_dev = &dev_to_i915(dev)->drm;
 
 	/*
 	 * We have a suspend ordering issue with the snd-hda driver also
@@ -1082,7 +1883,7 @@ static int i915_pm_suspend_late(struct device *dev)
 
 static int i915_pm_poweroff_late(struct device *dev)
 {
-	struct drm_device *drm_dev = dev_to_i915(dev)->dev;
+	struct drm_device *drm_dev = &dev_to_i915(dev)->drm;
 
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
@@ -1092,7 +1893,7 @@ static int i915_pm_poweroff_late(struct device *dev)
 
 static int i915_pm_resume_early(struct device *dev)
 {
-	struct drm_device *drm_dev = dev_to_i915(dev)->dev;
+	struct drm_device *drm_dev = &dev_to_i915(dev)->drm;
 
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
@@ -1102,7 +1903,7 @@ static int i915_pm_resume_early(struct device *dev)
 
 static int i915_pm_resume(struct device *dev)
 {
-	struct drm_device *drm_dev = dev_to_i915(dev)->dev;
+	struct drm_device *drm_dev = &dev_to_i915(dev)->drm;
 
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
@@ -1352,8 +2153,6 @@ int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool force_on)
 	u32 val;
 	int err;
 
-#define COND (I915_READ(VLV_GTLC_SURVIVABILITY_REG) & VLV_GFX_CLK_STATUS_BIT)
-
 	val = I915_READ(VLV_GTLC_SURVIVABILITY_REG);
 	val &= ~VLV_GFX_CLK_FORCE_ON_BIT;
 	if (force_on)
@@ -1363,13 +2162,16 @@ int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool force_on)
 	if (!force_on)
 		return 0;
 
-	err = wait_for(COND, 20);
+	err = intel_wait_for_register(dev_priv,
+				      VLV_GTLC_SURVIVABILITY_REG,
+				      VLV_GFX_CLK_STATUS_BIT,
+				      VLV_GFX_CLK_STATUS_BIT,
+				      20);
 	if (err)
 		DRM_ERROR("timeout waiting for GFX clock force-on (%08x)\n",
 			  I915_READ(VLV_GTLC_SURVIVABILITY_REG));
 
 	return err;
-#undef COND
 }
 
 static int vlv_allow_gt_wake(struct drm_i915_private *dev_priv, bool allow)
@@ -1384,13 +2186,15 @@ static int vlv_allow_gt_wake(struct drm_i915_private *dev_priv, bool allow)
 	I915_WRITE(VLV_GTLC_WAKE_CTRL, val);
 	POSTING_READ(VLV_GTLC_WAKE_CTRL);
 
-#define COND (!!(I915_READ(VLV_GTLC_PW_STATUS) & VLV_GTLC_ALLOWWAKEACK) == \
-	      allow)
-	err = wait_for(COND, 1);
+	err = intel_wait_for_register(dev_priv,
+				      VLV_GTLC_PW_STATUS,
+				      VLV_GTLC_ALLOWWAKEACK,
+				      allow,
+				      1);
 	if (err)
 		DRM_ERROR("timeout disabling GT waking\n");
+
 	return err;
-#undef COND
 }
 
 static int vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv,
@@ -1402,8 +2206,7 @@ static int vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv,
 
 	mask = VLV_GTLC_PW_MEDIA_STATUS_MASK | VLV_GTLC_PW_RENDER_STATUS_MASK;
 	val = wait_for_on ? mask : 0;
-#define COND ((I915_READ(VLV_GTLC_PW_STATUS) & mask) == val)
-	if (COND)
+	if ((I915_READ(VLV_GTLC_PW_STATUS) & mask) == val)
 		return 0;
 
 	DRM_DEBUG_KMS("waiting for GT wells to go %s (%08x)\n",
@@ -1414,13 +2217,14 @@ static int vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv,
 	 * RC6 transitioning can be delayed up to 2 msec (see
 	 * valleyview_enable_rps), use 3 msec for safety.
 	 */
-	err = wait_for(COND, 3);
+	err = intel_wait_for_register(dev_priv,
+				      VLV_GTLC_PW_STATUS, mask, val,
+				      3);
 	if (err)
 		DRM_ERROR("timeout waiting for GT wells to go %s\n",
 			  onoff(wait_for_on));
 
 	return err;
-#undef COND
 }
 
 static void vlv_check_no_gt_access(struct drm_i915_private *dev_priv)
@@ -1477,7 +2281,7 @@ err1:
 static int vlv_resume_prepare(struct drm_i915_private *dev_priv,
 				bool rpm_resume)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	int err;
 	int ret;
 
@@ -1513,7 +2317,7 @@ static int intel_runtime_suspend(struct device *device)
 {
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct drm_device *dev = pci_get_drvdata(pdev);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	if (WARN_ON_ONCE(!(dev_priv->rps.enabled && intel_enable_rc6())))
@@ -1551,11 +2355,8 @@ static int intel_runtime_suspend(struct device *device)
 	i915_gem_release_all_mmaps(dev_priv);
 	mutex_unlock(&dev->struct_mutex);
 
-	cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
-
 	intel_guc_suspend(dev);
 
-	intel_suspend_gt_powersave(dev_priv);
 	intel_runtime_pm_disable_interrupts(dev_priv);
 
 	ret = 0;
@@ -1620,7 +2421,7 @@ static int intel_runtime_resume(struct device *device)
 {
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct drm_device *dev = pci_get_drvdata(pdev);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
 	if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev)))
@@ -1670,8 +2471,6 @@ static int intel_runtime_resume(struct device *device)
 	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
 		intel_hpd_init(dev_priv);
 
-	intel_enable_gt_powersave(dev_priv);
-
 	enable_rpm_wakeref_asserts(dev_priv);
 
 	if (ret)
@@ -1682,7 +2481,7 @@ static int intel_runtime_resume(struct device *device)
 	return ret;
 }
 
-static const struct dev_pm_ops i915_pm_ops = {
+const struct dev_pm_ops i915_pm_ops = {
 	/*
 	 * S0ix (via system suspend) and S3 event handlers [PMSG_SUSPEND,
 	 * PMSG_RESUME]
@@ -1741,6 +2540,68 @@ static const struct file_operations i915_driver_fops = {
 	.llseek = noop_llseek,
 };
 
+static int
+i915_gem_reject_pin_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	return -ENODEV;
+}
+
+static const struct drm_ioctl_desc i915_ioctls[] = {
+	DRM_IOCTL_DEF_DRV(I915_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_FLUSH, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FLIP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_BATCHBUFFER, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_EMIT, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_SETPARAM, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER, i915_gem_execbuffer, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2, i915_gem_execbuffer2, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_UNPIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_BUSY, i915_gem_busy_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_CACHING, i915_gem_set_caching_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_CACHING, i915_gem_get_caching_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_THROTTLE, i915_gem_throttle_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_ENTERVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_LEAVEVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CREATE, i915_gem_create_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PREAD, i915_gem_pread_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PWRITE, i915_gem_pwrite_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP, i915_gem_mmap_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_GTT, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_APERTURE, i915_gem_get_aperture_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GET_PIPE_FROM_CRTC_ID, intel_get_pipe_from_crtc_id, 0),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_WAIT, i915_gem_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_CREATE, i915_gem_context_create_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_DESTROY, i915_gem_context_destroy_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_REG_READ, i915_reg_read_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GET_RESET_STATS, i915_gem_context_reset_stats_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_USERPTR, i915_gem_userptr_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_GETPARAM, i915_gem_context_getparam_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_SETPARAM, i915_gem_context_setparam_ioctl, DRM_RENDER_ALLOW),
+};
+
 static struct drm_driver driver = {
 	/* Don't use MTRRs here; the Xserver or userspace app should
 	 * deal with them for Intel hardware.
@@ -1748,18 +2609,12 @@ static struct drm_driver driver = {
 	.driver_features =
 	    DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM | DRIVER_PRIME |
 	    DRIVER_RENDER | DRIVER_MODESET,
-	.load = i915_driver_load,
-	.unload = i915_driver_unload,
 	.open = i915_driver_open,
 	.lastclose = i915_driver_lastclose,
 	.preclose = i915_driver_preclose,
 	.postclose = i915_driver_postclose,
 	.set_busid = drm_pci_set_busid,
 
-#if defined(CONFIG_DEBUG_FS)
-	.debugfs_init = i915_debugfs_init,
-	.debugfs_cleanup = i915_debugfs_cleanup,
-#endif
 	.gem_free_object = i915_gem_free_object,
 	.gem_vm_ops = &i915_gem_vm_ops,
 
@@ -1772,6 +2627,7 @@ static struct drm_driver driver = {
 	.dumb_map_offset = i915_gem_mmap_gtt,
 	.dumb_destroy = drm_gem_dumb_destroy,
 	.ioctls = i915_ioctls,
+	.num_ioctls = ARRAY_SIZE(i915_ioctls),
 	.fops = &i915_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
@@ -1780,56 +2636,3 @@ static struct drm_driver driver = {
 	.minor = DRIVER_MINOR,
 	.patchlevel = DRIVER_PATCHLEVEL,
 };
-
-static struct pci_driver i915_pci_driver = {
-	.name = DRIVER_NAME,
-	.id_table = pciidlist,
-	.probe = i915_pci_probe,
-	.remove = i915_pci_remove,
-	.driver.pm = &i915_pm_ops,
-};
-
-static int __init i915_init(void)
-{
-	driver.num_ioctls = i915_max_ioctl;
-
-	/*
-	 * Enable KMS by default, unless explicitly overriden by
-	 * either the i915.modeset prarameter or by the
-	 * vga_text_mode_force boot option.
-	 */
-
-	if (i915.modeset == 0)
-		driver.driver_features &= ~DRIVER_MODESET;
-
-	if (vgacon_text_force() && i915.modeset == -1)
-		driver.driver_features &= ~DRIVER_MODESET;
-
-	if (!(driver.driver_features & DRIVER_MODESET)) {
-		/* Silently fail loading to not upset userspace. */
-		DRM_DEBUG_DRIVER("KMS and UMS disabled.\n");
-		return 0;
-	}
-
-	if (i915.nuclear_pageflip)
-		driver.driver_features |= DRIVER_ATOMIC;
-
-	return drm_pci_init(&driver, &i915_pci_driver);
-}
-
-static void __exit i915_exit(void)
-{
-	if (!(driver.driver_features & DRIVER_MODESET))
-		return; /* Never loaded a driver. */
-
-	drm_pci_exit(&driver, &i915_pci_driver);
-}
-
-module_init(i915_init);
-module_exit(i915_exit);
-
-MODULE_AUTHOR("Tungsten Graphics, Inc.");
-MODULE_AUTHOR("Intel Corporation");
-
-MODULE_DESCRIPTION(DRIVER_DESC);
-MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 24a86c64d22e..03e1bfaa5a41 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -69,7 +69,7 @@
 
 #define DRIVER_NAME		"i915"
 #define DRIVER_DESC		"Intel Graphics"
-#define DRIVER_DATE		"20160620"
+#define DRIVER_DATE		"20160711"
 
 #undef WARN_ON
 /* Many gcc seem to no see through this and fall over :( */
@@ -320,15 +320,16 @@ struct i915_hotplug {
 		for_each_if ((__ports_mask) & (1 << (__port)))
 
 #define for_each_crtc(dev, crtc) \
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+	list_for_each_entry(crtc, &(dev)->mode_config.crtc_list, head)
 
 #define for_each_intel_plane(dev, intel_plane) \
 	list_for_each_entry(intel_plane,			\
-			    &dev->mode_config.plane_list,	\
+			    &(dev)->mode_config.plane_list,	\
 			    base.head)
 
 #define for_each_intel_plane_mask(dev, intel_plane, plane_mask)		\
-	list_for_each_entry(intel_plane, &dev->mode_config.plane_list,	\
+	list_for_each_entry(intel_plane,				\
+			    &(dev)->mode_config.plane_list,		\
 			    base.head)					\
 		for_each_if ((plane_mask) &				\
 			     (1 << drm_plane_index(&intel_plane->base)))
@@ -339,11 +340,15 @@ struct i915_hotplug {
 			    base.head)					\
 		for_each_if ((intel_plane)->pipe == (intel_crtc)->pipe)
 
-#define for_each_intel_crtc(dev, intel_crtc) \
-	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list, base.head)
+#define for_each_intel_crtc(dev, intel_crtc)				\
+	list_for_each_entry(intel_crtc,					\
+			    &(dev)->mode_config.crtc_list,		\
+			    base.head)
 
-#define for_each_intel_crtc_mask(dev, intel_crtc, crtc_mask) \
-	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list, base.head) \
+#define for_each_intel_crtc_mask(dev, intel_crtc, crtc_mask)		\
+	list_for_each_entry(intel_crtc,					\
+			    &(dev)->mode_config.crtc_list,		\
+			    base.head)					\
 		for_each_if ((crtc_mask) & (1 << drm_crtc_index(&intel_crtc->base)))
 
 #define for_each_intel_encoder(dev, intel_encoder)		\
@@ -353,7 +358,7 @@ struct i915_hotplug {
 
 #define for_each_intel_connector(dev, intel_connector)		\
 	list_for_each_entry(intel_connector,			\
-			    &dev->mode_config.connector_list,	\
+			    &(dev)->mode_config.connector_list,	\
 			    base.head)
 
 #define for_each_encoder_on_crtc(dev, __crtc, intel_encoder) \
@@ -475,6 +480,7 @@ struct drm_i915_error_state {
 	struct timeval time;
 
 	char error_msg[128];
+	bool simulated;
 	int iommu;
 	u32 reset_count;
 	u32 suspend_count;
@@ -506,6 +512,7 @@ struct drm_i915_error_state {
 		bool valid;
 		/* Software tracked state */
 		bool waiting;
+		int num_waiters;
 		int hangcheck_score;
 		enum intel_ring_hangcheck_action hangcheck_action;
 		int num_requests;
@@ -551,6 +558,12 @@ struct drm_i915_error_state {
 			u32 tail;
 		} *requests;
 
+		struct drm_i915_error_waiter {
+			char comm[TASK_COMM_LEN];
+			pid_t pid;
+			u32 seqno;
+		} *waiters;
+
 		struct {
 			u32 gfx_mode;
 			union {
@@ -868,9 +881,12 @@ struct i915_gem_context {
 
 	/* Unique identifier for this context, used by the hw for tracking */
 	unsigned long flags;
+#define CONTEXT_NO_ZEROMAP		BIT(0)
+#define CONTEXT_NO_ERROR_CAPTURE	BIT(1)
 	unsigned hw_id;
 	u32 user_handle;
-#define CONTEXT_NO_ZEROMAP		(1<<0)
+
+	u32 ggtt_alignment;
 
 	struct intel_context {
 		struct drm_i915_gem_object *state;
@@ -1011,6 +1027,7 @@ enum intel_pch {
 	PCH_CPT,	/* Cougarpoint PCH */
 	PCH_LPT,	/* Lynxpoint PCH */
 	PCH_SPT,        /* Sunrisepoint PCH */
+	PCH_KBP,        /* Kabypoint PCH */
 	PCH_NOP,
 };
 
@@ -1305,37 +1322,11 @@ struct i915_gem_mm {
 	struct list_head fence_list;
 
 	/**
-	 * We leave the user IRQ off as much as possible,
-	 * but this means that requests will finish and never
-	 * be retired once the system goes idle. Set a timer to
-	 * fire periodically while the ring is running. When it
-	 * fires, go retire requests.
-	 */
-	struct delayed_work retire_work;
-
-	/**
-	 * When we detect an idle GPU, we want to turn on
-	 * powersaving features. So once we see that there
-	 * are no more requests outstanding and no more
-	 * arrive within a small period of time, we fire
-	 * off the idle_work.
-	 */
-	struct delayed_work idle_work;
-
-	/**
 	 * Are we in a non-interruptible section of code like
 	 * modesetting?
 	 */
 	bool interruptible;
 
-	/**
-	 * Is the GPU currently considered idle, or busy executing userspace
-	 * requests?  Whilst idle, we attempt to power down the hardware and
-	 * display clocks. In order to reduce the effect on performance, there
-	 * is a slight delay before we do so.
-	 */
-	bool busy;
-
 	/* the indicator for dispatch video commands on two BSD rings */
 	unsigned int bsd_ring_dispatch_index;
 
@@ -1372,7 +1363,6 @@ struct i915_gpu_error {
 	/* Hang gpu twice in this window and your context gets banned */
 #define DRM_I915_CTX_BAN_PERIOD DIV_ROUND_UP(8*DRM_I915_HANGCHECK_PERIOD, 1000)
 
-	struct workqueue_struct *hangcheck_wq;
 	struct delayed_work hangcheck_work;
 
 	/* For reset and error_state handling. */
@@ -1409,20 +1399,19 @@ struct i915_gpu_error {
 #define I915_WEDGED			(1 << 31)
 
 	/**
+	 * Waitqueue to signal when a hang is detected. Used to for waiters
+	 * to release the struct_mutex for the reset to procede.
+	 */
+	wait_queue_head_t wait_queue;
+
+	/**
 	 * Waitqueue to signal when the reset has completed. Used by clients
 	 * that wait for dev_priv->mm.wedged to settle.
 	 */
 	wait_queue_head_t reset_queue;
 
-	/* Userspace knobs for gpu hang simulation;
-	 * combines both a ring mask, and extra flags
-	 */
-	u32 stop_rings;
-#define I915_STOP_RING_ALLOW_BAN       (1 << 31)
-#define I915_STOP_RING_ALLOW_WARN      (1 << 30)
-
 	/* For missed irq/seqno simulation. */
-	unsigned int test_irq_rings;
+	unsigned long test_irq_rings;
 };
 
 enum modeset_restore {
@@ -1733,7 +1722,8 @@ struct intel_wm_config {
 };
 
 struct drm_i915_private {
-	struct drm_device *dev;
+	struct drm_device drm;
+
 	struct kmem_cache *objects;
 	struct kmem_cache *vmas;
 	struct kmem_cache *requests;
@@ -2029,6 +2019,34 @@ struct drm_i915_private {
 		int (*init_engines)(struct drm_device *dev);
 		void (*cleanup_engine)(struct intel_engine_cs *engine);
 		void (*stop_engine)(struct intel_engine_cs *engine);
+
+		/**
+		 * Is the GPU currently considered idle, or busy executing
+		 * userspace requests? Whilst idle, we allow runtime power
+		 * management to power down the hardware and display clocks.
+		 * In order to reduce the effect on performance, there
+		 * is a slight delay before we do so.
+		 */
+		unsigned int active_engines;
+		bool awake;
+
+		/**
+		 * We leave the user IRQ off as much as possible,
+		 * but this means that requests will finish and never
+		 * be retired once the system goes idle. Set a timer to
+		 * fire periodically while the ring is running. When it
+		 * fires, go retire requests.
+		 */
+		struct delayed_work retire_work;
+
+		/**
+		 * When we detect an idle GPU, we want to turn on
+		 * powersaving features. So once we see that there
+		 * are no more requests outstanding and no more
+		 * arrive within a small period of time, we fire
+		 * off the idle_work.
+		 */
+		struct delayed_work idle_work;
 	} gt;
 
 	/* perform PHY state sanity checks? */
@@ -2044,7 +2062,7 @@ struct drm_i915_private {
 
 static inline struct drm_i915_private *to_i915(const struct drm_device *dev)
 {
-	return dev->dev_private;
+	return container_of(dev, struct drm_i915_private, drm);
 }
 
 static inline struct drm_i915_private *dev_to_i915(struct device *dev)
@@ -2215,6 +2233,7 @@ struct drm_i915_gem_object {
 
 	unsigned int frontbuffer_bits:INTEL_FRONTBUFFER_BITS;
 
+	unsigned int has_wc_mmap;
 	unsigned int pin_display;
 
 	struct sg_table *pages;
@@ -2267,6 +2286,12 @@ struct drm_i915_gem_object {
 };
 #define to_intel_bo(x) container_of(x, struct drm_i915_gem_object, base)
 
+static inline bool
+i915_gem_object_has_struct_page(const struct drm_i915_gem_object *obj)
+{
+	return obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE;
+}
+
 /*
  * Optimised SGL iterator for GEM objects
  */
@@ -2357,7 +2382,7 @@ struct drm_i915_gem_request {
 	/** On Which ring this request was generated */
 	struct drm_i915_private *i915;
 	struct intel_engine_cs *engine;
-	unsigned reset_counter;
+	struct intel_signal_node signaling;
 
 	 /** GEM sequence number associated with the previous request,
 	  * when the HWS breadcrumb is equal to this the GPU is processing
@@ -2613,7 +2638,7 @@ struct drm_i915_cmd_table {
 #define INTEL_DEVID(p)	(INTEL_INFO(p)->device_id)
 
 #define REVID_FOREVER		0xff
-#define INTEL_REVID(p)	(__I915__(p)->dev->pdev->revision)
+#define INTEL_REVID(p)	(__I915__(p)->drm.pdev->revision)
 
 #define GEN_FOREVER (0)
 /*
@@ -2743,29 +2768,34 @@ struct drm_i915_cmd_table {
  * have their own (e.g. HAS_PCH_SPLIT for ILK+ display, IS_foo for particular
  * chips, etc.).
  */
-#define IS_GEN2(dev)	(INTEL_INFO(dev)->gen_mask & BIT(1))
-#define IS_GEN3(dev)	(INTEL_INFO(dev)->gen_mask & BIT(2))
-#define IS_GEN4(dev)	(INTEL_INFO(dev)->gen_mask & BIT(3))
-#define IS_GEN5(dev)	(INTEL_INFO(dev)->gen_mask & BIT(4))
-#define IS_GEN6(dev)	(INTEL_INFO(dev)->gen_mask & BIT(5))
-#define IS_GEN7(dev)	(INTEL_INFO(dev)->gen_mask & BIT(6))
-#define IS_GEN8(dev)	(INTEL_INFO(dev)->gen_mask & BIT(7))
-#define IS_GEN9(dev)	(INTEL_INFO(dev)->gen_mask & BIT(8))
-
-#define RENDER_RING		(1<<RCS)
-#define BSD_RING		(1<<VCS)
-#define BLT_RING		(1<<BCS)
-#define VEBOX_RING		(1<<VECS)
-#define BSD2_RING		(1<<VCS2)
-#define ALL_ENGINES		(~0)
-
-#define HAS_BSD(dev)		(INTEL_INFO(dev)->ring_mask & BSD_RING)
-#define HAS_BSD2(dev)		(INTEL_INFO(dev)->ring_mask & BSD2_RING)
-#define HAS_BLT(dev)		(INTEL_INFO(dev)->ring_mask & BLT_RING)
-#define HAS_VEBOX(dev)		(INTEL_INFO(dev)->ring_mask & VEBOX_RING)
+#define IS_GEN2(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(1)))
+#define IS_GEN3(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(2)))
+#define IS_GEN4(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(3)))
+#define IS_GEN5(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(4)))
+#define IS_GEN6(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(5)))
+#define IS_GEN7(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(6)))
+#define IS_GEN8(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(7)))
+#define IS_GEN9(dev)	(!!(INTEL_INFO(dev)->gen_mask & BIT(8)))
+
+#define ENGINE_MASK(id)	BIT(id)
+#define RENDER_RING	ENGINE_MASK(RCS)
+#define BSD_RING	ENGINE_MASK(VCS)
+#define BLT_RING	ENGINE_MASK(BCS)
+#define VEBOX_RING	ENGINE_MASK(VECS)
+#define BSD2_RING	ENGINE_MASK(VCS2)
+#define ALL_ENGINES	(~0)
+
+#define HAS_ENGINE(dev_priv, id) \
+	(!!(INTEL_INFO(dev_priv)->ring_mask & ENGINE_MASK(id)))
+
+#define HAS_BSD(dev_priv)	HAS_ENGINE(dev_priv, VCS)
+#define HAS_BSD2(dev_priv)	HAS_ENGINE(dev_priv, VCS2)
+#define HAS_BLT(dev_priv)	HAS_ENGINE(dev_priv, BCS)
+#define HAS_VEBOX(dev_priv)	HAS_ENGINE(dev_priv, VECS)
+
 #define HAS_LLC(dev)		(INTEL_INFO(dev)->has_llc)
 #define HAS_SNOOP(dev)		(INTEL_INFO(dev)->has_snoop)
-#define HAS_EDRAM(dev)		(__I915__(dev)->edram_cap & EDRAM_ENABLED)
+#define HAS_EDRAM(dev)		(!!(__I915__(dev)->edram_cap & EDRAM_ENABLED))
 #define HAS_WT(dev)		((IS_HASWELL(dev) || IS_BROADWELL(dev)) && \
 				 HAS_EDRAM(dev))
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
@@ -2783,9 +2813,10 @@ struct drm_i915_cmd_table {
 #define HAS_BROKEN_CS_TLB(dev)		(IS_I830(dev) || IS_845G(dev))
 
 /* WaRsDisableCoarsePowerGating:skl,bxt */
-#define NEEDS_WaRsDisableCoarsePowerGating(dev) (IS_BXT_REVID(dev, 0, BXT_REVID_A1) || \
-						 IS_SKL_GT3(dev) || \
-						 IS_SKL_GT4(dev))
+#define NEEDS_WaRsDisableCoarsePowerGating(dev_priv) \
+	(IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1) || \
+	 IS_SKL_GT3(dev_priv) || \
+	 IS_SKL_GT4(dev_priv))
 
 /*
  * dp aux and gmbus irq on gen4 seems to be able to generate legacy interrupts
@@ -2832,7 +2863,7 @@ struct drm_i915_cmd_table {
  * command submission once loaded. But these are logically independent
  * properties, so we have separate macros to test them.
  */
-#define HAS_GUC(dev)		(IS_GEN9(dev) && !IS_KABYLAKE(dev))
+#define HAS_GUC(dev)		(IS_GEN9(dev))
 #define HAS_GUC_UCODE(dev)	(HAS_GUC(dev))
 #define HAS_GUC_SCHED(dev)	(HAS_GUC(dev))
 
@@ -2853,11 +2884,13 @@ struct drm_i915_cmd_table {
 #define INTEL_PCH_LPT_LP_DEVICE_ID_TYPE		0x9c00
 #define INTEL_PCH_SPT_DEVICE_ID_TYPE		0xA100
 #define INTEL_PCH_SPT_LP_DEVICE_ID_TYPE		0x9D00
+#define INTEL_PCH_KBP_DEVICE_ID_TYPE		0xA200
 #define INTEL_PCH_P2X_DEVICE_ID_TYPE		0x7100
 #define INTEL_PCH_P3X_DEVICE_ID_TYPE		0x7000
 #define INTEL_PCH_QEMU_DEVICE_ID_TYPE		0x2900 /* qemu q35 has 2918 */
 
 #define INTEL_PCH_TYPE(dev) (__I915__(dev)->pch_type)
+#define HAS_PCH_KBP(dev) (INTEL_PCH_TYPE(dev) == PCH_KBP)
 #define HAS_PCH_SPT(dev) (INTEL_PCH_TYPE(dev) == PCH_SPT)
 #define HAS_PCH_LPT(dev) (INTEL_PCH_TYPE(dev) == PCH_LPT)
 #define HAS_PCH_LPT_LP(dev) (__I915__(dev)->pch_id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE)
@@ -2879,8 +2912,14 @@ struct drm_i915_cmd_table {
 
 #include "i915_trace.h"
 
-extern const struct drm_ioctl_desc i915_ioctls[];
-extern int i915_max_ioctl;
+static inline bool intel_scanout_needs_vtd_wa(struct drm_i915_private *dev_priv)
+{
+#ifdef CONFIG_INTEL_IOMMU
+	if (INTEL_GEN(dev_priv) >= 6 && intel_iommu_gfx_mapped)
+		return true;
+#endif
+	return false;
+}
 
 extern int i915_suspend_switcheroo(struct drm_device *dev, pm_message_t state);
 extern int i915_resume_switcheroo(struct drm_device *dev);
@@ -2888,7 +2927,7 @@ extern int i915_resume_switcheroo(struct drm_device *dev);
 int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv,
 			       	int enable_ppgtt);
 
-/* i915_dma.c */
+/* i915_drv.c */
 void __printf(3, 4)
 __i915_printk(struct drm_i915_private *dev_priv, const char *level,
 	      const char *fmt, ...);
@@ -2896,14 +2935,6 @@ __i915_printk(struct drm_i915_private *dev_priv, const char *level,
 #define i915_report_error(dev_priv, fmt, ...)				   \
 	__i915_printk(dev_priv, KERN_ERR, fmt, ##__VA_ARGS__)
 
-extern int i915_driver_load(struct drm_device *, unsigned long flags);
-extern int i915_driver_unload(struct drm_device *);
-extern int i915_driver_open(struct drm_device *dev, struct drm_file *file);
-extern void i915_driver_lastclose(struct drm_device * dev);
-extern void i915_driver_preclose(struct drm_device *dev,
-				 struct drm_file *file);
-extern void i915_driver_postclose(struct drm_device *dev,
-				  struct drm_file *file);
 #ifdef CONFIG_COMPAT
 extern long i915_compat_ioctl(struct file *filp, unsigned int cmd,
 			      unsigned long arg);
@@ -2928,7 +2959,23 @@ void intel_hpd_cancel_work(struct drm_i915_private *dev_priv);
 bool intel_hpd_pin_to_port(enum hpd_pin pin, enum port *port);
 
 /* i915_irq.c */
-void i915_queue_hangcheck(struct drm_i915_private *dev_priv);
+static inline void i915_queue_hangcheck(struct drm_i915_private *dev_priv)
+{
+	unsigned long delay;
+
+	if (unlikely(!i915.enable_hangcheck))
+		return;
+
+	/* Don't continually defer the hangcheck so that it is always run at
+	 * least once after work has been scheduled on any ring. Otherwise,
+	 * we will ignore a hung ring if a second ring is kept busy.
+	 */
+
+	delay = round_jiffies_up_relative(DRM_I915_HANGCHECK_JIFFIES);
+	queue_delayed_work(system_long_wq,
+			   &dev_priv->gpu_error.hangcheck_work, delay);
+}
+
 __printf(3, 4)
 void i915_handle_error(struct drm_i915_private *dev_priv,
 		       u32 engine_mask,
@@ -2963,6 +3010,17 @@ u64 intel_uncore_edram_size(struct drm_i915_private *dev_priv);
 
 void assert_forcewakes_inactive(struct drm_i915_private *dev_priv);
 
+int intel_wait_for_register(struct drm_i915_private *dev_priv,
+			    i915_reg_t reg,
+			    const u32 mask,
+			    const u32 value,
+			    const unsigned long timeout_ms);
+int intel_wait_for_register_fw(struct drm_i915_private *dev_priv,
+			       i915_reg_t reg,
+			       const u32 mask,
+			       const u32 value,
+			       const unsigned long timeout_ms);
+
 static inline bool intel_gvt_active(struct drm_i915_private *dev_priv)
 {
 	return dev_priv->gvt.initialized;
@@ -3027,7 +3085,6 @@ ibx_disable_display_interrupt(struct drm_i915_private *dev_priv, uint32_t bits)
 	ibx_display_interrupt_update(dev_priv, bits, 0);
 }
 
-
 /* i915_gem.c */
 int i915_gem_create_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv);
@@ -3244,31 +3301,34 @@ i915_seqno_passed(uint32_t seq1, uint32_t seq2)
 	return (int32_t)(seq1 - seq2) >= 0;
 }
 
-static inline bool i915_gem_request_started(struct drm_i915_gem_request *req,
-					   bool lazy_coherency)
+static inline bool i915_gem_request_started(const struct drm_i915_gem_request *req)
 {
-	if (!lazy_coherency && req->engine->irq_seqno_barrier)
-		req->engine->irq_seqno_barrier(req->engine);
-	return i915_seqno_passed(req->engine->get_seqno(req->engine),
+	return i915_seqno_passed(intel_engine_get_seqno(req->engine),
 				 req->previous_seqno);
 }
 
-static inline bool i915_gem_request_completed(struct drm_i915_gem_request *req,
-					      bool lazy_coherency)
+static inline bool i915_gem_request_completed(const struct drm_i915_gem_request *req)
 {
-	if (!lazy_coherency && req->engine->irq_seqno_barrier)
-		req->engine->irq_seqno_barrier(req->engine);
-	return i915_seqno_passed(req->engine->get_seqno(req->engine),
+	return i915_seqno_passed(intel_engine_get_seqno(req->engine),
 				 req->seqno);
 }
 
+bool __i915_spin_request(const struct drm_i915_gem_request *request,
+			 int state, unsigned long timeout_us);
+static inline bool i915_spin_request(const struct drm_i915_gem_request *request,
+				     int state, unsigned long timeout_us)
+{
+	return (i915_gem_request_started(request) &&
+		__i915_spin_request(request, state, timeout_us));
+}
+
 int __must_check i915_gem_get_seqno(struct drm_i915_private *dev_priv, u32 *seqno);
 int __must_check i915_gem_set_seqno(struct drm_device *dev, u32 seqno);
 
 struct drm_i915_gem_request *
 i915_gem_find_active_request(struct intel_engine_cs *engine);
 
-bool i915_gem_retire_requests(struct drm_i915_private *dev_priv);
+void i915_gem_retire_requests(struct drm_i915_private *dev_priv);
 void i915_gem_retire_requests_ring(struct intel_engine_cs *engine);
 
 static inline u32 i915_reset_counter(struct i915_gpu_error *error)
@@ -3311,18 +3371,6 @@ static inline u32 i915_reset_count(struct i915_gpu_error *error)
 	return ((i915_reset_counter(error) & ~I915_WEDGED) + 1) / 2;
 }
 
-static inline bool i915_stop_ring_allow_ban(struct drm_i915_private *dev_priv)
-{
-	return dev_priv->gpu_error.stop_rings == 0 ||
-		dev_priv->gpu_error.stop_rings & I915_STOP_RING_ALLOW_BAN;
-}
-
-static inline bool i915_stop_ring_allow_warn(struct drm_i915_private *dev_priv)
-{
-	return dev_priv->gpu_error.stop_rings == 0 ||
-		dev_priv->gpu_error.stop_rings & I915_STOP_RING_ALLOW_WARN;
-}
-
 void i915_gem_reset(struct drm_device *dev);
 bool i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force);
 int __must_check i915_gem_init(struct drm_device *dev);
@@ -3330,7 +3378,7 @@ int i915_gem_init_engines(struct drm_device *dev);
 int __must_check i915_gem_init_hw(struct drm_device *dev);
 void i915_gem_init_swizzling(struct drm_device *dev);
 void i915_gem_cleanup_engines(struct drm_device *dev);
-int __must_check i915_gpu_idle(struct drm_device *dev);
+int __must_check i915_gem_wait_for_idle(struct drm_i915_private *dev_priv);
 int __must_check i915_gem_suspend(struct drm_device *dev);
 void __i915_add_request(struct drm_i915_gem_request *req,
 			struct drm_i915_gem_object *batch_obj,
@@ -3484,7 +3532,7 @@ i915_gem_context_lookup(struct drm_i915_file_private *file_priv, u32 id)
 {
 	struct i915_gem_context *ctx;
 
-	lockdep_assert_held(&file_priv->dev_priv->dev->struct_mutex);
+	lockdep_assert_held(&file_priv->dev_priv->drm.struct_mutex);
 
 	ctx = idr_find(&file_priv->context_idr, id);
 	if (!ctx)
@@ -3500,7 +3548,7 @@ static inline void i915_gem_context_reference(struct i915_gem_context *ctx)
 
 static inline void i915_gem_context_unreference(struct i915_gem_context *ctx)
 {
-	lockdep_assert_held(&ctx->i915->dev->struct_mutex);
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
 	kref_put(&ctx->ref, i915_gem_context_free);
 }
 
@@ -3576,7 +3624,7 @@ void i915_gem_shrinker_cleanup(struct drm_i915_private *dev_priv);
 /* i915_gem_tiling.c */
 static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 
 	return dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
 		obj->tiling_mode != I915_TILING_NONE;
@@ -3590,12 +3638,14 @@ int i915_verify_lists(struct drm_device *dev);
 #endif
 
 /* i915_debugfs.c */
-int i915_debugfs_init(struct drm_minor *minor);
-void i915_debugfs_cleanup(struct drm_minor *minor);
 #ifdef CONFIG_DEBUG_FS
+int i915_debugfs_register(struct drm_i915_private *dev_priv);
+void i915_debugfs_unregister(struct drm_i915_private *dev_priv);
 int i915_debugfs_connector_add(struct drm_connector *connector);
 void intel_display_crc_init(struct drm_device *dev);
 #else
+static inline int i915_debugfs_register(struct drm_i915_private *) {return 0;}
+static inline void i915_debugfs_unregister(struct drm_i915_private *) {}
 static inline int i915_debugfs_connector_add(struct drm_connector *connector)
 { return 0; }
 static inline void intel_display_crc_init(struct drm_device *dev) {}
@@ -3686,8 +3736,8 @@ extern int intel_opregion_notify_adapter(struct drm_i915_private *dev_priv,
 extern int intel_opregion_get_panel_type(struct drm_i915_private *dev_priv);
 #else
 static inline int intel_opregion_setup(struct drm_i915_private *dev) { return 0; }
-static inline void intel_opregion_init(struct drm_i915_private *dev) { }
-static inline void intel_opregion_fini(struct drm_i915_private *dev) { }
+static inline void intel_opregion_register(struct drm_i915_private *dev_priv) { }
+static inline void intel_opregion_unregister(struct drm_i915_private *dev_priv) { }
 static inline void intel_opregion_asle_intr(struct drm_i915_private *dev_priv)
 {
 }
@@ -3716,11 +3766,22 @@ static inline void intel_register_dsm_handler(void) { return; }
 static inline void intel_unregister_dsm_handler(void) { return; }
 #endif /* CONFIG_ACPI */
 
+/* intel_device_info.c */
+static inline struct intel_device_info *
+mkwrite_device_info(struct drm_i915_private *dev_priv)
+{
+	return (struct intel_device_info *)&dev_priv->info;
+}
+
+void intel_device_info_runtime_init(struct drm_i915_private *dev_priv);
+void intel_device_info_dump(struct drm_i915_private *dev_priv);
+
 /* modesetting */
 extern void intel_modeset_init_hw(struct drm_device *dev);
 extern void intel_modeset_init(struct drm_device *dev);
 extern void intel_modeset_gem_init(struct drm_device *dev);
 extern void intel_modeset_cleanup(struct drm_device *dev);
+extern int intel_connector_register(struct drm_connector *);
 extern void intel_connector_unregister(struct drm_connector *);
 extern int intel_modeset_vga_set_state(struct drm_device *dev, bool state);
 extern void intel_display_resume(struct drm_device *dev);
@@ -3731,7 +3792,6 @@ extern void intel_init_pch_refclk(struct drm_device *dev);
 extern void intel_set_rps(struct drm_i915_private *dev_priv, u8 val);
 extern void intel_set_memory_cxsr(struct drm_i915_private *dev_priv,
 				  bool enable);
-extern void intel_detect_pch(struct drm_device *dev);
 
 extern bool i915_semaphore_is_enabled(struct drm_i915_private *dev_priv);
 int i915_reg_read_ioctl(struct drm_device *dev, void *data,
@@ -3864,6 +3924,7 @@ __raw_write(64, q)
  */
 #define I915_READ_FW(reg__) __raw_i915_read32(dev_priv, (reg__))
 #define I915_WRITE_FW(reg__, val__) __raw_i915_write32(dev_priv, (reg__), (val__))
+#define I915_WRITE64_FW(reg__, val__) __raw_i915_write64(dev_priv, (reg__), (val__))
 #define POSTING_READ_FW(reg__) (void)I915_READ_FW(reg__)
 
 /* "Broadcast RGB" property */
@@ -3927,12 +3988,80 @@ wait_remaining_ms_from_jiffies(unsigned long timestamp_jiffies, int to_wait_ms)
 			    schedule_timeout_uninterruptible(remaining_jiffies);
 	}
 }
-
-static inline void i915_trace_irq_get(struct intel_engine_cs *engine,
-				      struct drm_i915_gem_request *req)
+static inline bool __i915_request_irq_complete(struct drm_i915_gem_request *req)
 {
-	if (engine->trace_irq_req == NULL && engine->irq_get(engine))
-		i915_gem_request_assign(&engine->trace_irq_req, req);
+	struct intel_engine_cs *engine = req->engine;
+
+	/* Before we do the heavier coherent read of the seqno,
+	 * check the value (hopefully) in the CPU cacheline.
+	 */
+	if (i915_gem_request_completed(req))
+		return true;
+
+	/* Ensure our read of the seqno is coherent so that we
+	 * do not "miss an interrupt" (i.e. if this is the last
+	 * request and the seqno write from the GPU is not visible
+	 * by the time the interrupt fires, we will see that the
+	 * request is incomplete and go back to sleep awaiting
+	 * another interrupt that will never come.)
+	 *
+	 * Strictly, we only need to do this once after an interrupt,
+	 * but it is easier and safer to do it every time the waiter
+	 * is woken.
+	 */
+	if (engine->irq_seqno_barrier &&
+	    READ_ONCE(engine->breadcrumbs.irq_seqno_bh) == current &&
+	    cmpxchg_relaxed(&engine->breadcrumbs.irq_posted, 1, 0)) {
+		struct task_struct *tsk;
+
+		/* The ordering of irq_posted versus applying the barrier
+		 * is crucial. The clearing of the current irq_posted must
+		 * be visible before we perform the barrier operation,
+		 * such that if a subsequent interrupt arrives, irq_posted
+		 * is reasserted and our task rewoken (which causes us to
+		 * do another __i915_request_irq_complete() immediately
+		 * and reapply the barrier). Conversely, if the clear
+		 * occurs after the barrier, then an interrupt that arrived
+		 * whilst we waited on the barrier would not trigger a
+		 * barrier on the next pass, and the read may not see the
+		 * seqno update.
+		 */
+		engine->irq_seqno_barrier(engine);
+
+		/* If we consume the irq, but we are no longer the bottom-half,
+		 * the real bottom-half may not have serialised their own
+		 * seqno check with the irq-barrier (i.e. may have inspected
+		 * the seqno before we believe it coherent since they see
+		 * irq_posted == false but we are still running).
+		 */
+		rcu_read_lock();
+		tsk = READ_ONCE(engine->breadcrumbs.irq_seqno_bh);
+		if (tsk && tsk != current)
+			/* Note that if the bottom-half is changed as we
+			 * are sending the wake-up, the new bottom-half will
+			 * be woken by whomever made the change. We only have
+			 * to worry about when we steal the irq-posted for
+			 * ourself.
+			 */
+			wake_up_process(tsk);
+		rcu_read_unlock();
+
+		if (i915_gem_request_completed(req))
+			return true;
+	}
+
+	/* We need to check whether any gpu reset happened in between
+	 * the request being submitted and now. If a reset has occurred,
+	 * the seqno will have been advance past ours and our request
+	 * is complete. If we are in the process of handling a reset,
+	 * the request is effectively complete as the rendering will
+	 * be discarded, but we need to return in order to drop the
+	 * struct_mutex.
+	 */
+	if (i915_reset_in_progress(&req->i915->gpu_error))
+		return true;
+
+	return false;
 }
 
 #endif
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 21d0dea57312..8f50919ba9b4 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -128,7 +128,7 @@ i915_gem_wait_for_error(struct i915_gpu_error *error)
 
 int i915_mutex_lock_interruptible(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	ret = i915_gem_wait_for_error(&dev_priv->gpu_error);
@@ -377,13 +377,13 @@ out:
 
 void *i915_gem_object_alloc(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	return kmem_cache_zalloc(dev_priv->objects, GFP_KERNEL);
 }
 
 void i915_gem_object_free(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	kmem_cache_free(dev_priv->objects, obj);
 }
 
@@ -508,7 +508,7 @@ int i915_gem_obj_prepare_shmem_read(struct drm_i915_gem_object *obj,
 
 	*needs_clflush = 0;
 
-	if (WARN_ON((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0))
+	if (WARN_ON(!i915_gem_object_has_struct_page(obj)))
 		return -EINVAL;
 
 	if (!(obj->base.read_domains & I915_GEM_DOMAIN_CPU)) {
@@ -636,7 +636,7 @@ i915_gem_gtt_pread(struct drm_device *dev,
 		   struct drm_i915_gem_object *obj, uint64_t size,
 		   uint64_t data_offset, uint64_t data_ptr)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct drm_mm_node node;
 	char __user *user_data;
@@ -760,7 +760,7 @@ i915_gem_shmem_pread(struct drm_device *dev,
 	int needs_clflush = 0;
 	struct sg_page_iter sg_iter;
 
-	if (!obj->base.filp)
+	if (!i915_gem_object_has_struct_page(obj))
 		return -ENODEV;
 
 	user_data = u64_to_user_ptr(args->data_ptr);
@@ -1250,7 +1250,7 @@ int
 i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 		      struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_pwrite *args = data;
 	struct drm_i915_gem_object *obj;
 	int ret;
@@ -1298,7 +1298,8 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	 * pread/pwrite currently are reading and writing from the CPU
 	 * perspective, requiring manual detiling by the client.
 	 */
-	if (!obj->base.filp || cpu_write_needs_clflush(obj)) {
+	if (!i915_gem_object_has_struct_page(obj) ||
+	    cpu_write_needs_clflush(obj)) {
 		ret = i915_gem_gtt_pwrite_fast(dev_priv, obj, args, file);
 		/* Note that the gtt paths might fail with non-page-backed user
 		 * pointers (e.g. gtt mappings when moving data between
@@ -1308,7 +1309,7 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	if (ret == -EFAULT) {
 		if (obj->phys_handle)
 			ret = i915_gem_phys_pwrite(obj, args, file);
-		else if (obj->base.filp)
+		else if (i915_gem_object_has_struct_page(obj))
 			ret = i915_gem_shmem_pwrite(dev, obj, args, file);
 		else
 			ret = -ENODEV;
@@ -1342,17 +1343,6 @@ i915_gem_check_wedge(unsigned reset_counter, bool interruptible)
 	return 0;
 }
 
-static void fake_irq(unsigned long data)
-{
-	wake_up_process((struct task_struct *)data);
-}
-
-static bool missed_irq(struct drm_i915_private *dev_priv,
-		       struct intel_engine_cs *engine)
-{
-	return test_bit(engine->id, &dev_priv->gpu_error.missed_irq_rings);
-}
-
 static unsigned long local_clock_us(unsigned *cpu)
 {
 	unsigned long t;
@@ -1385,9 +1375,9 @@ static bool busywait_stop(unsigned long timeout, unsigned cpu)
 	return this_cpu != cpu;
 }
 
-static int __i915_spin_request(struct drm_i915_gem_request *req, int state)
+bool __i915_spin_request(const struct drm_i915_gem_request *req,
+			 int state, unsigned long timeout_us)
 {
-	unsigned long timeout;
 	unsigned cpu;
 
 	/* When waiting for high frequency requests, e.g. during synchronous
@@ -1400,31 +1390,21 @@ static int __i915_spin_request(struct drm_i915_gem_request *req, int state)
 	 * takes to sleep on a request, on the order of a microsecond.
 	 */
 
-	if (req->engine->irq_refcount)
-		return -EBUSY;
-
-	/* Only spin if we know the GPU is processing this request */
-	if (!i915_gem_request_started(req, true))
-		return -EAGAIN;
-
-	timeout = local_clock_us(&cpu) + 5;
-	while (!need_resched()) {
-		if (i915_gem_request_completed(req, true))
-			return 0;
+	timeout_us += local_clock_us(&cpu);
+	do {
+		if (i915_gem_request_completed(req))
+			return true;
 
 		if (signal_pending_state(state, current))
 			break;
 
-		if (busywait_stop(timeout, cpu))
+		if (busywait_stop(timeout_us, cpu))
 			break;
 
 		cpu_relax_lowlatency();
-	}
+	} while (!need_resched());
 
-	if (i915_gem_request_completed(req, false))
-		return 0;
-
-	return -EAGAIN;
+	return false;
 }
 
 /**
@@ -1449,25 +1429,22 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
 			s64 *timeout,
 			struct intel_rps_client *rps)
 {
-	struct intel_engine_cs *engine = i915_gem_request_get_engine(req);
-	struct drm_i915_private *dev_priv = req->i915;
-	const bool irq_test_in_progress =
-		ACCESS_ONCE(dev_priv->gpu_error.test_irq_rings) & intel_engine_flag(engine);
 	int state = interruptible ? TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
-	DEFINE_WAIT(wait);
-	unsigned long timeout_expire;
+	DEFINE_WAIT(reset);
+	struct intel_wait wait;
+	unsigned long timeout_remain;
 	s64 before = 0; /* Only to silence a compiler warning. */
-	int ret;
+	int ret = 0;
 
-	WARN(!intel_irqs_enabled(dev_priv), "IRQs disabled");
+	might_sleep();
 
 	if (list_empty(&req->list))
 		return 0;
 
-	if (i915_gem_request_completed(req, true))
+	if (i915_gem_request_completed(req))
 		return 0;
 
-	timeout_expire = 0;
+	timeout_remain = MAX_SCHEDULE_TIMEOUT;
 	if (timeout) {
 		if (WARN_ON(*timeout < 0))
 			return -EINVAL;
@@ -1475,7 +1452,7 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
 		if (*timeout == 0)
 			return -ETIME;
 
-		timeout_expire = jiffies + nsecs_to_jiffies_timeout(*timeout);
+		timeout_remain = nsecs_to_jiffies_timeout(*timeout);
 
 		/*
 		 * Record current time in case interrupted by signal, or wedged.
@@ -1483,75 +1460,85 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
 		before = ktime_get_raw_ns();
 	}
 
-	if (INTEL_INFO(dev_priv)->gen >= 6)
-		gen6_rps_boost(dev_priv, rps, req->emitted_jiffies);
-
 	trace_i915_gem_request_wait_begin(req);
 
-	/* Optimistic spin for the next jiffie before touching IRQs */
-	ret = __i915_spin_request(req, state);
-	if (ret == 0)
-		goto out;
-
-	if (!irq_test_in_progress && WARN_ON(!engine->irq_get(engine))) {
-		ret = -ENODEV;
-		goto out;
-	}
+	/* This client is about to stall waiting for the GPU. In many cases
+	 * this is undesirable and limits the throughput of the system, as
+	 * many clients cannot continue processing user input/output whilst
+	 * blocked. RPS autotuning may take tens of milliseconds to respond
+	 * to the GPU load and thus incurs additional latency for the client.
+	 * We can circumvent that by promoting the GPU frequency to maximum
+	 * before we wait. This makes the GPU throttle up much more quickly
+	 * (good for benchmarks and user experience, e.g. window animations),
+	 * but at a cost of spending more power processing the workload
+	 * (bad for battery). Not all clients even want their results
+	 * immediately and for them we should just let the GPU select its own
+	 * frequency to maximise efficiency. To prevent a single client from
+	 * forcing the clocks too high for the whole system, we only allow
+	 * each client to waitboost once in a busy period.
+	 */
+	if (INTEL_INFO(req->i915)->gen >= 6)
+		gen6_rps_boost(req->i915, rps, req->emitted_jiffies);
 
-	for (;;) {
-		struct timer_list timer;
+	/* Optimistic spin for the next ~jiffie before touching IRQs */
+	if (i915_spin_request(req, state, 5))
+		goto complete;
 
-		prepare_to_wait(&engine->irq_queue, &wait, state);
+	set_current_state(state);
+	add_wait_queue(&req->i915->gpu_error.wait_queue, &reset);
 
-		/* We need to check whether any gpu reset happened in between
-		 * the request being submitted and now. If a reset has occurred,
-		 * the request is effectively complete (we either are in the
-		 * process of or have discarded the rendering and completely
-		 * reset the GPU. The results of the request are lost and we
-		 * are free to continue on with the original operation.
+	intel_wait_init(&wait, req->seqno);
+	if (intel_engine_add_wait(req->engine, &wait))
+		/* In order to check that we haven't missed the interrupt
+		 * as we enabled it, we need to kick ourselves to do a
+		 * coherent check on the seqno before we sleep.
 		 */
-		if (req->reset_counter != i915_reset_counter(&dev_priv->gpu_error)) {
-			ret = 0;
-			break;
-		}
-
-		if (i915_gem_request_completed(req, false)) {
-			ret = 0;
-			break;
-		}
+		goto wakeup;
 
+	for (;;) {
 		if (signal_pending_state(state, current)) {
 			ret = -ERESTARTSYS;
 			break;
 		}
 
-		if (timeout && time_after_eq(jiffies, timeout_expire)) {
+		/* Ensure that even if the GPU hangs, we get woken up.
+		 *
+		 * However, note that if no one is waiting, we never notice
+		 * a gpu hang. Eventually, we will have to wait for a resource
+		 * held by the GPU and so trigger a hangcheck. In the most
+		 * pathological case, this will be upon memory starvation!
+		 */
+		i915_queue_hangcheck(req->i915);
+
+		timeout_remain = io_schedule_timeout(timeout_remain);
+		if (timeout_remain == 0) {
 			ret = -ETIME;
 			break;
 		}
 
-		timer.function = NULL;
-		if (timeout || missed_irq(dev_priv, engine)) {
-			unsigned long expire;
+		if (intel_wait_complete(&wait))
+			break;
 
-			setup_timer_on_stack(&timer, fake_irq, (unsigned long)current);
-			expire = missed_irq(dev_priv, engine) ? jiffies + 1 : timeout_expire;
-			mod_timer(&timer, expire);
-		}
+		set_current_state(state);
 
-		io_schedule();
+wakeup:
+		/* Carefully check if the request is complete, giving time
+		 * for the seqno to be visible following the interrupt.
+		 * We also have to check in case we are kicked by the GPU
+		 * reset in order to drop the struct_mutex.
+		 */
+		if (__i915_request_irq_complete(req))
+			break;
 
-		if (timer.function) {
-			del_singleshot_timer_sync(&timer);
-			destroy_timer_on_stack(&timer);
-		}
+		/* Only spin if we know the GPU is processing this request */
+		if (i915_spin_request(req, state, 2))
+			break;
 	}
-	if (!irq_test_in_progress)
-		engine->irq_put(engine);
-
-	finish_wait(&engine->irq_queue, &wait);
+	remove_wait_queue(&req->i915->gpu_error.wait_queue, &reset);
 
-out:
+	intel_engine_remove_wait(req->engine, &wait);
+	__set_current_state(TASK_RUNNING);
+complete:
 	trace_i915_gem_request_wait_end(req);
 
 	if (timeout) {
@@ -1570,6 +1557,22 @@ out:
 			*timeout = 0;
 	}
 
+	if (rps && req->seqno == req->engine->last_submitted_seqno) {
+		/* The GPU is now idle and this client has stalled.
+		 * Since no other client has submitted a request in the
+		 * meantime, assume that this client is the only one
+		 * supplying work to the GPU but is unable to keep that
+		 * work supplied because it is waiting. Since the GPU is
+		 * then never kept fully busy, RPS autoclocking will
+		 * keep the clocks relatively low, causing further delays.
+		 * Compensate by giving the synchronous client credit for
+		 * a waitboost next time.
+		 */
+		spin_lock(&req->i915->rps.client_lock);
+		list_del_init(&rps->link);
+		spin_unlock(&req->i915->rps.client_lock);
+	}
+
 	return ret;
 }
 
@@ -1648,7 +1651,7 @@ __i915_gem_request_retire__upto(struct drm_i915_gem_request *req)
 	struct intel_engine_cs *engine = req->engine;
 	struct drm_i915_gem_request *tmp;
 
-	lockdep_assert_held(&engine->i915->dev->struct_mutex);
+	lockdep_assert_held(&engine->i915->drm.struct_mutex);
 
 	if (list_empty(&req->list))
 		return;
@@ -1677,14 +1680,14 @@ i915_wait_request(struct drm_i915_gem_request *req)
 
 	interruptible = dev_priv->mm.interruptible;
 
-	BUG_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
+	BUG_ON(!mutex_is_locked(&dev_priv->drm.struct_mutex));
 
 	ret = __i915_wait_request(req, interruptible, NULL, NULL);
 	if (ret)
 		return ret;
 
 	/* If the GPU hung, we want to keep the requests to find the guilty. */
-	if (req->reset_counter == i915_reset_counter(&dev_priv->gpu_error))
+	if (!i915_reset_in_progress(&dev_priv->gpu_error))
 		__i915_gem_request_retire__upto(req);
 
 	return 0;
@@ -1745,7 +1748,7 @@ i915_gem_object_retire_request(struct drm_i915_gem_object *obj,
 	else if (obj->last_write_req == req)
 		i915_gem_object_retire__write(obj);
 
-	if (req->reset_counter == i915_reset_counter(&req->i915->gpu_error))
+	if (!i915_reset_in_progress(&req->i915->gpu_error))
 		__i915_gem_request_retire__upto(req);
 }
 
@@ -1758,7 +1761,7 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 					    bool readonly)
 {
 	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_request *requests[I915_NUM_ENGINES];
 	int ret, i, n = 0;
 
@@ -1809,6 +1812,13 @@ static struct intel_rps_client *to_rps_client(struct drm_file *file)
 	return &fpriv->rps;
 }
 
+static enum fb_op_origin
+write_origin(struct drm_i915_gem_object *obj, unsigned domain)
+{
+	return domain == I915_GEM_DOMAIN_GTT && !obj->has_wc_mmap ?
+	       ORIGIN_GTT : ORIGIN_CPU;
+}
+
 /**
  * Called when user space prepares to use an object with the CPU, either
  * through the mmap ioctl's mapping or a GTT mapping.
@@ -1865,9 +1875,7 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 		ret = i915_gem_object_set_to_cpu_domain(obj, write_domain != 0);
 
 	if (write_domain != 0)
-		intel_fb_obj_invalidate(obj,
-					write_domain == I915_GEM_DOMAIN_GTT ?
-					ORIGIN_GTT : ORIGIN_CPU);
+		intel_fb_obj_invalidate(obj, write_origin(obj, write_domain));
 
 unref:
 	drm_gem_object_unreference(&obj->base);
@@ -1974,6 +1982,9 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
 		else
 			addr = -ENOMEM;
 		up_write(&mm->mmap_sem);
+
+		/* This may race, but that's ok, it only gets set */
+		WRITE_ONCE(to_intel_bo(obj)->has_wc_mmap, true);
 	}
 	drm_gem_object_unreference_unlocked(obj);
 	if (IS_ERR((void *)addr))
@@ -2262,7 +2273,7 @@ i915_gem_get_gtt_alignment(struct drm_device *dev, uint32_t size,
 
 static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	int ret;
 
 	dev_priv->mm.shrinker_no_lock_stealing = true;
@@ -2478,7 +2489,7 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 static int
 i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	int page_count, i;
 	struct address_space *mapping;
 	struct sg_table *st;
@@ -2609,7 +2620,7 @@ err_pages:
 int
 i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	const struct drm_i915_gem_object_ops *ops = obj->ops;
 	int ret;
 
@@ -2773,6 +2784,13 @@ i915_gem_init_seqno(struct drm_i915_private *dev_priv, u32 seqno)
 	}
 	i915_gem_retire_requests(dev_priv);
 
+	/* If the seqno wraps around, we need to clear the breadcrumb rbtree */
+	if (!i915_seqno_passed(seqno, dev_priv->next_seqno)) {
+		while (intel_kick_waiters(dev_priv) ||
+		       intel_kick_signalers(dev_priv))
+			yield();
+	}
+
 	/* Finally reset hw state */
 	for_each_engine(engine, dev_priv)
 		intel_ring_init_seqno(engine, seqno);
@@ -2782,7 +2800,7 @@ i915_gem_init_seqno(struct drm_i915_private *dev_priv, u32 seqno)
 
 int i915_gem_set_seqno(struct drm_device *dev, u32 seqno)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	if (seqno == 0)
@@ -2822,6 +2840,26 @@ i915_gem_get_seqno(struct drm_i915_private *dev_priv, u32 *seqno)
 	return 0;
 }
 
+static void i915_gem_mark_busy(const struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+
+	dev_priv->gt.active_engines |= intel_engine_flag(engine);
+	if (dev_priv->gt.awake)
+		return;
+
+	intel_runtime_pm_get_noresume(dev_priv);
+	dev_priv->gt.awake = true;
+
+	i915_update_gfx_val(dev_priv);
+	if (INTEL_GEN(dev_priv) >= 6)
+		gen6_rps_busy(dev_priv);
+
+	queue_delayed_work(dev_priv->wq,
+			   &dev_priv->gt.retire_work,
+			   round_jiffies_up_relative(HZ));
+}
+
 /*
  * NB: This function is not allowed to fail. Doing so would mean the the
  * request is not being tracked for completion but the work itself is
@@ -2832,7 +2870,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 			bool flush_caches)
 {
 	struct intel_engine_cs *engine;
-	struct drm_i915_private *dev_priv;
 	struct intel_ringbuffer *ringbuf;
 	u32 request_start;
 	u32 reserved_tail;
@@ -2842,7 +2879,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 		return;
 
 	engine = request->engine;
-	dev_priv = request->i915;
 	ringbuf = request->ringbuf;
 
 	/*
@@ -2908,14 +2944,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 	}
 	/* Not allowed to fail! */
 	WARN(ret, "emit|add_request failed: %d!\n", ret);
-
-	i915_queue_hangcheck(engine->i915);
-
-	queue_delayed_work(dev_priv->wq,
-			   &dev_priv->mm.retire_work,
-			   round_jiffies_up_relative(HZ));
-	intel_mark_busy(dev_priv);
-
 	/* Sanity check that the reserved size was large enough. */
 	ret = intel_ring_get_tail(ringbuf) - request_start;
 	if (ret < 0)
@@ -2924,46 +2952,34 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 		  "Not enough space reserved (%d bytes) "
 		  "for adding the request (%d bytes)\n",
 		  reserved_tail, ret);
+
+	i915_gem_mark_busy(engine);
 }
 
-static bool i915_context_is_banned(struct drm_i915_private *dev_priv,
-				   const struct i915_gem_context *ctx)
+static bool i915_context_is_banned(const struct i915_gem_context *ctx)
 {
 	unsigned long elapsed;
 
-	elapsed = get_seconds() - ctx->hang_stats.guilty_ts;
-
 	if (ctx->hang_stats.banned)
 		return true;
 
+	elapsed = get_seconds() - ctx->hang_stats.guilty_ts;
 	if (ctx->hang_stats.ban_period_seconds &&
 	    elapsed <= ctx->hang_stats.ban_period_seconds) {
-		if (!i915_gem_context_is_default(ctx)) {
-			DRM_DEBUG("context hanging too fast, banning!\n");
-			return true;
-		} else if (i915_stop_ring_allow_ban(dev_priv)) {
-			if (i915_stop_ring_allow_warn(dev_priv))
-				DRM_ERROR("gpu hanging too fast, banning!\n");
-			return true;
-		}
+		DRM_DEBUG("context hanging too fast, banning!\n");
+		return true;
 	}
 
 	return false;
 }
 
-static void i915_set_reset_status(struct drm_i915_private *dev_priv,
-				  struct i915_gem_context *ctx,
+static void i915_set_reset_status(struct i915_gem_context *ctx,
 				  const bool guilty)
 {
-	struct i915_ctx_hang_stats *hs;
-
-	if (WARN_ON(!ctx))
-		return;
-
-	hs = &ctx->hang_stats;
+	struct i915_ctx_hang_stats *hs = &ctx->hang_stats;
 
 	if (guilty) {
-		hs->banned = i915_context_is_banned(dev_priv, ctx);
+		hs->banned = i915_context_is_banned(ctx);
 		hs->batch_active++;
 		hs->guilty_ts = get_seconds();
 	} else {
@@ -3012,7 +3028,6 @@ __i915_gem_request_alloc(struct intel_engine_cs *engine,
 	kref_init(&req->ref);
 	req->i915 = dev_priv;
 	req->engine = engine;
-	req->reset_counter = reset_counter;
 	req->ctx  = ctx;
 	i915_gem_context_reference(req->ctx);
 
@@ -3072,8 +3087,16 @@ i915_gem_find_active_request(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_request *request;
 
+	/* We are called by the error capture and reset at a random
+	 * point in time. In particular, note that neither is crucially
+	 * ordered with an interrupt. After a hang, the GPU is dead and we
+	 * assume that no more writes can happen (we waited long enough for
+	 * all writes that were in transaction to be flushed) - adding an
+	 * extra delay for a recent interrupt is pointless. Hence, we do
+	 * not need an engine->irq_seqno_barrier() before the seqno reads.
+	 */
 	list_for_each_entry(request, &engine->request_list, list) {
-		if (i915_gem_request_completed(request, false))
+		if (i915_gem_request_completed(request))
 			continue;
 
 		return request;
@@ -3082,27 +3105,23 @@ i915_gem_find_active_request(struct intel_engine_cs *engine)
 	return NULL;
 }
 
-static void i915_gem_reset_engine_status(struct drm_i915_private *dev_priv,
-				       struct intel_engine_cs *engine)
+static void i915_gem_reset_engine_status(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_request *request;
 	bool ring_hung;
 
 	request = i915_gem_find_active_request(engine);
-
 	if (request == NULL)
 		return;
 
 	ring_hung = engine->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG;
 
-	i915_set_reset_status(dev_priv, request->ctx, ring_hung);
-
+	i915_set_reset_status(request->ctx, ring_hung);
 	list_for_each_entry_continue(request, &engine->request_list, list)
-		i915_set_reset_status(dev_priv, request->ctx, false);
+		i915_set_reset_status(request->ctx, false);
 }
 
-static void i915_gem_reset_engine_cleanup(struct drm_i915_private *dev_priv,
-					struct intel_engine_cs *engine)
+static void i915_gem_reset_engine_cleanup(struct intel_engine_cs *engine)
 {
 	struct intel_ringbuffer *buffer;
 
@@ -3163,7 +3182,7 @@ static void i915_gem_reset_engine_cleanup(struct drm_i915_private *dev_priv,
 
 void i915_gem_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 
 	/*
@@ -3172,10 +3191,10 @@ void i915_gem_reset(struct drm_device *dev)
 	 * their reference to the objects, the inspection must be done first.
 	 */
 	for_each_engine(engine, dev_priv)
-		i915_gem_reset_engine_status(dev_priv, engine);
+		i915_gem_reset_engine_status(engine);
 
 	for_each_engine(engine, dev_priv)
-		i915_gem_reset_engine_cleanup(dev_priv, engine);
+		i915_gem_reset_engine_cleanup(engine);
 
 	i915_gem_context_reset(dev);
 
@@ -3205,7 +3224,7 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *engine)
 					   struct drm_i915_gem_request,
 					   list);
 
-		if (!i915_gem_request_completed(request, true))
+		if (!i915_gem_request_completed(request))
 			break;
 
 		i915_gem_request_retire(request);
@@ -3228,55 +3247,52 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *engine)
 		i915_gem_object_retire__read(obj, engine->id);
 	}
 
-	if (unlikely(engine->trace_irq_req &&
-		     i915_gem_request_completed(engine->trace_irq_req, true))) {
-		engine->irq_put(engine);
-		i915_gem_request_assign(&engine->trace_irq_req, NULL);
-	}
-
 	WARN_ON(i915_verify_lists(engine->dev));
 }
 
-bool
-i915_gem_retire_requests(struct drm_i915_private *dev_priv)
+void i915_gem_retire_requests(struct drm_i915_private *dev_priv)
 {
 	struct intel_engine_cs *engine;
-	bool idle = true;
+
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+
+	if (dev_priv->gt.active_engines == 0)
+		return;
+
+	GEM_BUG_ON(!dev_priv->gt.awake);
 
 	for_each_engine(engine, dev_priv) {
 		i915_gem_retire_requests_ring(engine);
-		idle &= list_empty(&engine->request_list);
-		if (i915.enable_execlists) {
-			spin_lock_bh(&engine->execlist_lock);
-			idle &= list_empty(&engine->execlist_queue);
-			spin_unlock_bh(&engine->execlist_lock);
-		}
+		if (list_empty(&engine->request_list))
+			dev_priv->gt.active_engines &= ~intel_engine_flag(engine);
 	}
 
-	if (idle)
-		mod_delayed_work(dev_priv->wq,
-				   &dev_priv->mm.idle_work,
+	if (dev_priv->gt.active_engines == 0)
+		queue_delayed_work(dev_priv->wq,
+				   &dev_priv->gt.idle_work,
 				   msecs_to_jiffies(100));
-
-	return idle;
 }
 
 static void
 i915_gem_retire_work_handler(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
-		container_of(work, typeof(*dev_priv), mm.retire_work.work);
-	struct drm_device *dev = dev_priv->dev;
-	bool idle;
+		container_of(work, typeof(*dev_priv), gt.retire_work.work);
+	struct drm_device *dev = &dev_priv->drm;
 
 	/* Come back later if the device is busy... */
-	idle = false;
 	if (mutex_trylock(&dev->struct_mutex)) {
-		idle = i915_gem_retire_requests(dev_priv);
+		i915_gem_retire_requests(dev_priv);
 		mutex_unlock(&dev->struct_mutex);
 	}
-	if (!idle)
-		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work,
+
+	/* Keep the retire handler running until we are finally idle.
+	 * We do not need to do this test under locking as in the worst-case
+	 * we queue the retire worker once too often.
+	 */
+	if (READ_ONCE(dev_priv->gt.awake))
+		queue_delayed_work(dev_priv->wq,
+				   &dev_priv->gt.retire_work,
 				   round_jiffies_up_relative(HZ));
 }
 
@@ -3284,25 +3300,55 @@ static void
 i915_gem_idle_work_handler(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
-		container_of(work, typeof(*dev_priv), mm.idle_work.work);
-	struct drm_device *dev = dev_priv->dev;
+		container_of(work, typeof(*dev_priv), gt.idle_work.work);
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_engine_cs *engine;
+	unsigned int stuck_engines;
+	bool rearm_hangcheck;
+
+	if (!READ_ONCE(dev_priv->gt.awake))
+		return;
+
+	if (READ_ONCE(dev_priv->gt.active_engines))
+		return;
+
+	rearm_hangcheck =
+		cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
+
+	if (!mutex_trylock(&dev->struct_mutex)) {
+		/* Currently busy, come back later */
+		mod_delayed_work(dev_priv->wq,
+				 &dev_priv->gt.idle_work,
+				 msecs_to_jiffies(50));
+		goto out_rearm;
+	}
+
+	if (dev_priv->gt.active_engines)
+		goto out_unlock;
 
 	for_each_engine(engine, dev_priv)
-		if (!list_empty(&engine->request_list))
-			return;
+		i915_gem_batch_pool_fini(&engine->batch_pool);
 
-	/* we probably should sync with hangcheck here, using cancel_work_sync.
-	 * Also locking seems to be fubar here, engine->request_list is protected
-	 * by dev->struct_mutex. */
+	GEM_BUG_ON(!dev_priv->gt.awake);
+	dev_priv->gt.awake = false;
+	rearm_hangcheck = false;
 
-	intel_mark_idle(dev_priv);
+	stuck_engines = intel_kick_waiters(dev_priv);
+	if (unlikely(stuck_engines)) {
+		DRM_DEBUG_DRIVER("kicked stuck waiters...missed irq\n");
+		dev_priv->gpu_error.missed_irq_rings |= stuck_engines;
+	}
 
-	if (mutex_trylock(&dev->struct_mutex)) {
-		for_each_engine(engine, dev_priv)
-			i915_gem_batch_pool_fini(&engine->batch_pool);
+	if (INTEL_GEN(dev_priv) >= 6)
+		gen6_rps_idle(dev_priv);
+	intel_runtime_pm_put(dev_priv);
+out_unlock:
+	mutex_unlock(&dev->struct_mutex);
 
-		mutex_unlock(&dev->struct_mutex);
+out_rearm:
+	if (rearm_hangcheck) {
+		GEM_BUG_ON(!dev_priv->gt.awake);
+		i915_queue_hangcheck(dev_priv);
 	}
 }
 
@@ -3327,7 +3373,7 @@ i915_gem_object_flush_active(struct drm_i915_gem_object *obj)
 		if (req == NULL)
 			continue;
 
-		if (i915_gem_request_completed(req, true))
+		if (i915_gem_request_completed(req))
 			i915_gem_object_retire__read(obj, i);
 	}
 
@@ -3435,7 +3481,7 @@ __i915_gem_object_sync(struct drm_i915_gem_object *obj,
 	if (to == from)
 		return 0;
 
-	if (i915_gem_request_completed(from_req, true))
+	if (i915_gem_request_completed(from_req))
 		return 0;
 
 	if (!i915_semaphore_is_enabled(to_i915(obj->base.dev))) {
@@ -3586,7 +3632,7 @@ static void __i915_vma_iounmap(struct i915_vma *vma)
 static int __i915_vma_unbind(struct i915_vma *vma, bool wait)
 {
 	struct drm_i915_gem_object *obj = vma->obj;
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	int ret;
 
 	if (list_empty(&vma->obj_link))
@@ -3662,26 +3708,16 @@ int __i915_vma_unbind_no_wait(struct i915_vma *vma)
 	return __i915_vma_unbind(vma, false);
 }
 
-int i915_gpu_idle(struct drm_device *dev)
+int i915_gem_wait_for_idle(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *engine;
 	int ret;
 
-	/* Flush everything onto the inactive list. */
-	for_each_engine(engine, dev_priv) {
-		if (!i915.enable_execlists) {
-			struct drm_i915_gem_request *req;
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
 
-			req = i915_gem_request_alloc(engine, NULL);
-			if (IS_ERR(req))
-				return PTR_ERR(req);
-
-			ret = i915_switch_context(req);
-			i915_add_request_no_flush(req);
-			if (ret)
-				return ret;
-		}
+	for_each_engine(engine, dev_priv) {
+		if (engine->last_context == NULL)
+			continue;
 
 		ret = intel_engine_idle(engine);
 		if (ret)
@@ -4214,7 +4250,7 @@ int i915_gem_get_caching_ioctl(struct drm_device *dev, void *data,
 int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data,
 			       struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_caching *args = data;
 	struct drm_i915_gem_object *obj;
 	enum i915_cache_level level;
@@ -4408,7 +4444,7 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 static int
 i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_file_private *file_priv = file->driver_priv;
 	unsigned long recent_enough = jiffies - DRM_I915_THROTTLE_JIFFIES;
 	struct drm_i915_gem_request *request, *target = NULL;
@@ -4444,9 +4480,6 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 		return 0;
 
 	ret = __i915_wait_request(target, true, NULL, NULL);
-	if (ret == 0)
-		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, 0);
-
 	i915_gem_request_unreference(target);
 
 	return ret;
@@ -4505,7 +4538,7 @@ i915_gem_object_do_pin(struct drm_i915_gem_object *obj,
 		       uint32_t alignment,
 		       uint64_t flags)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	struct i915_vma *vma;
 	unsigned bound;
 	int ret;
@@ -4669,7 +4702,7 @@ int
 i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
 		       struct drm_file *file_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_madvise *args = data;
 	struct drm_i915_gem_object *obj;
 	int ret;
@@ -4739,7 +4772,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 	obj->fence_reg = I915_FENCE_REG_NONE;
 	obj->madv = I915_MADV_WILLNEED;
 
-	i915_gem_info_add_obj(obj->base.dev->dev_private, obj->base.size);
+	i915_gem_info_add_obj(to_i915(obj->base.dev), obj->base.size);
 }
 
 static const struct drm_i915_gem_object_ops i915_gem_object_ops = {
@@ -4834,7 +4867,7 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 {
 	struct drm_i915_gem_object *obj = to_intel_bo(gem_obj);
 	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_vma *vma, *next;
 
 	intel_runtime_pm_get(dev_priv);
@@ -4938,7 +4971,7 @@ void i915_gem_vma_destroy(struct i915_vma *vma)
 static void
 i915_gem_stop_engines(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 
 	for_each_engine(engine, dev_priv)
@@ -4948,11 +4981,11 @@ i915_gem_stop_engines(struct drm_device *dev)
 int
 i915_gem_suspend(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
 	mutex_lock(&dev->struct_mutex);
-	ret = i915_gpu_idle(dev);
+	ret = i915_gem_wait_for_idle(dev_priv);
 	if (ret)
 		goto err;
 
@@ -4963,13 +4996,13 @@ i915_gem_suspend(struct drm_device *dev)
 	mutex_unlock(&dev->struct_mutex);
 
 	cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
-	cancel_delayed_work_sync(&dev_priv->mm.retire_work);
-	flush_delayed_work(&dev_priv->mm.idle_work);
+	cancel_delayed_work_sync(&dev_priv->gt.retire_work);
+	flush_delayed_work(&dev_priv->gt.idle_work);
 
 	/* Assert that we sucessfully flushed all the work and
 	 * reset the GPU back to its idle, low power state.
 	 */
-	WARN_ON(dev_priv->mm.busy);
+	WARN_ON(dev_priv->gt.awake);
 
 	return 0;
 
@@ -4980,7 +5013,7 @@ err:
 
 void i915_gem_init_swizzling(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_INFO(dev)->gen < 5 ||
 	    dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_NONE)
@@ -5005,7 +5038,7 @@ void i915_gem_init_swizzling(struct drm_device *dev)
 
 static void init_unused_ring(struct drm_device *dev, u32 base)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(RING_CTL(base), 0);
 	I915_WRITE(RING_HEAD(base), 0);
@@ -5032,7 +5065,7 @@ static void init_unused_rings(struct drm_device *dev)
 
 int i915_gem_init_engines(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	ret = intel_init_render_ring_buffer(dev);
@@ -5080,7 +5113,7 @@ cleanup_render_ring:
 int
 i915_gem_init_hw(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	int ret;
 
@@ -5138,12 +5171,6 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (ret)
 		goto out;
 
-	/*
-	 * Increment the next seqno by 0x100 so we have a visible break
-	 * on re-initialisation
-	 */
-	ret = i915_gem_set_seqno(dev, dev_priv->next_seqno+0x100);
-
 out:
 	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
 	return ret;
@@ -5151,7 +5178,7 @@ out:
 
 int i915_gem_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	mutex_lock(&dev->struct_mutex);
@@ -5208,7 +5235,7 @@ out_unlock:
 void
 i915_gem_cleanup_engines(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 
 	for_each_engine(engine, dev_priv)
@@ -5225,7 +5252,7 @@ init_engine_lists(struct intel_engine_cs *engine)
 void
 i915_gem_load_init_fences(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	if (INTEL_INFO(dev_priv)->gen >= 7 && !IS_VALLEYVIEW(dev_priv) &&
 	    !IS_CHERRYVIEW(dev_priv))
@@ -5249,7 +5276,7 @@ i915_gem_load_init_fences(struct drm_i915_private *dev_priv)
 void
 i915_gem_load_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	dev_priv->objects =
@@ -5277,22 +5304,15 @@ i915_gem_load_init(struct drm_device *dev)
 		init_engine_lists(&dev_priv->engine[i]);
 	for (i = 0; i < I915_MAX_NUM_FENCES; i++)
 		INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list);
-	INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
+	INIT_DELAYED_WORK(&dev_priv->gt.retire_work,
 			  i915_gem_retire_work_handler);
-	INIT_DELAYED_WORK(&dev_priv->mm.idle_work,
+	INIT_DELAYED_WORK(&dev_priv->gt.idle_work,
 			  i915_gem_idle_work_handler);
+	init_waitqueue_head(&dev_priv->gpu_error.wait_queue);
 	init_waitqueue_head(&dev_priv->gpu_error.reset_queue);
 
 	dev_priv->relative_constants_mode = I915_EXEC_CONSTANTS_REL_GENERAL;
 
-	/*
-	 * Set initial sequence number for requests.
-	 * Using this number allows the wraparound to happen early,
-	 * catching any obvious problems.
-	 */
-	dev_priv->next_seqno = ((u32)~0 - 0x1100);
-	dev_priv->last_seqno = ((u32)~0 - 0x1101);
-
 	INIT_LIST_HEAD(&dev_priv->mm.fence_list);
 
 	init_waitqueue_head(&dev_priv->pending_flip_queue);
@@ -5378,7 +5398,7 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file)
 		return -ENOMEM;
 
 	file->driver_priv = file_priv;
-	file_priv->dev_priv = dev->dev_private;
+	file_priv->dev_priv = to_i915(dev);
 	file_priv->file = file;
 	INIT_LIST_HEAD(&file_priv->rps.link);
 
@@ -5424,7 +5444,7 @@ void i915_gem_track_fb(struct drm_i915_gem_object *old,
 u64 i915_gem_obj_offset(struct drm_i915_gem_object *o,
 			struct i915_address_space *vm)
 {
-	struct drm_i915_private *dev_priv = o->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(o->base.dev);
 	struct i915_vma *vma;
 
 	WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
@@ -5528,7 +5548,7 @@ i915_gem_object_get_dirty_page(struct drm_i915_gem_object *obj, int n)
 	struct page *page;
 
 	/* Only default objects have per-page dirty tracking */
-	if (WARN_ON((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0))
+	if (WARN_ON(!i915_gem_object_has_struct_page(obj)))
 		return NULL;
 
 	page = i915_gem_object_get_page(obj, n);
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 30d9b4fd30f3..3c97f0e7a003 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -154,7 +154,7 @@ void i915_gem_context_free(struct kref *ctx_ref)
 	struct i915_gem_context *ctx = container_of(ctx_ref, typeof(*ctx), ref);
 	int i;
 
-	lockdep_assert_held(&ctx->i915->dev->struct_mutex);
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
 	trace_i915_context_free(ctx);
 
 	/*
@@ -250,7 +250,7 @@ static struct i915_gem_context *
 __create_hw_context(struct drm_device *dev,
 		    struct drm_i915_file_private *file_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_gem_context *ctx;
 	int ret;
 
@@ -268,6 +268,8 @@ __create_hw_context(struct drm_device *dev,
 	list_add_tail(&ctx->link, &dev_priv->context_list);
 	ctx->i915 = dev_priv;
 
+	ctx->ggtt_alignment = get_context_alignment(dev_priv);
+
 	if (dev_priv->hw_context_size) {
 		struct drm_i915_gem_object *obj =
 				i915_gem_alloc_context_obj(dev, dev_priv->hw_context_size);
@@ -394,7 +396,7 @@ static void i915_gem_context_unpin(struct i915_gem_context *ctx,
 
 void i915_gem_context_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	lockdep_assert_held(&dev->struct_mutex);
 
@@ -410,7 +412,7 @@ void i915_gem_context_reset(struct drm_device *dev)
 
 int i915_gem_context_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_gem_context *ctx;
 
 	/* Init should only be called once per module load. Eventually the
@@ -451,26 +453,6 @@ int i915_gem_context_init(struct drm_device *dev)
 		return PTR_ERR(ctx);
 	}
 
-	if (!i915.enable_execlists && ctx->engine[RCS].state) {
-		int ret;
-
-		/* We may need to do things with the shrinker which
-		 * require us to immediately switch back to the default
-		 * context. This can cause a problem as pinning the
-		 * default context also requires GTT space which may not
-		 * be available. To avoid this we always pin the default
-		 * context.
-		 */
-		ret = i915_gem_obj_ggtt_pin(ctx->engine[RCS].state,
-					    get_context_alignment(dev_priv), 0);
-		if (ret) {
-			DRM_ERROR("Failed to pinned default global context (error %d)\n",
-				  ret);
-			i915_gem_context_unreference(ctx);
-			return ret;
-		}
-	}
-
 	dev_priv->kernel_context = ctx;
 
 	DRM_DEBUG_DRIVER("%s context support initialized\n",
@@ -483,33 +465,45 @@ void i915_gem_context_lost(struct drm_i915_private *dev_priv)
 {
 	struct intel_engine_cs *engine;
 
-	lockdep_assert_held(&dev_priv->dev->struct_mutex);
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
 
 	for_each_engine(engine, dev_priv) {
 		if (engine->last_context) {
 			i915_gem_context_unpin(engine->last_context, engine);
 			engine->last_context = NULL;
 		}
-
-		/* Force the GPU state to be reinitialised on enabling */
-		dev_priv->kernel_context->engine[engine->id].initialised =
-			engine->init_context == NULL;
 	}
 
-	/* Force the GPU state to be reinitialised on enabling */
-	dev_priv->kernel_context->remap_slice = ALL_L3_SLICES(dev_priv);
+	/* Force the GPU state to be restored on enabling */
+	if (!i915.enable_execlists) {
+		struct i915_gem_context *ctx;
+
+		list_for_each_entry(ctx, &dev_priv->context_list, link) {
+			if (!i915_gem_context_is_default(ctx))
+				continue;
+
+			for_each_engine(engine, dev_priv)
+				ctx->engine[engine->id].initialised = false;
+
+			ctx->remap_slice = ALL_L3_SLICES(dev_priv);
+		}
+
+		for_each_engine(engine, dev_priv) {
+			struct intel_context *kce =
+				&dev_priv->kernel_context->engine[engine->id];
+
+			kce->initialised = true;
+		}
+	}
 }
 
 void i915_gem_context_fini(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_gem_context *dctx = dev_priv->kernel_context;
 
 	lockdep_assert_held(&dev->struct_mutex);
 
-	if (!i915.enable_execlists && dctx->engine[RCS].state)
-		i915_gem_object_ggtt_unpin(dctx->engine[RCS].state);
-
 	i915_gem_context_unreference(dctx);
 	dev_priv->kernel_context = NULL;
 
@@ -759,7 +753,7 @@ static int do_rcs_switch(struct drm_i915_gem_request *req)
 
 	/* Trying to pin first makes error handling easier. */
 	ret = i915_gem_obj_ggtt_pin(to->engine[RCS].state,
-				    get_context_alignment(engine->i915),
+				    to->ggtt_alignment,
 				    0);
 	if (ret)
 		return ret;
@@ -901,7 +895,7 @@ int i915_switch_context(struct drm_i915_gem_request *req)
 	struct intel_engine_cs *engine = req->engine;
 
 	WARN_ON(i915.enable_execlists);
-	lockdep_assert_held(&req->i915->dev->struct_mutex);
+	lockdep_assert_held(&req->i915->drm.struct_mutex);
 
 	if (!req->ctx->engine[engine->id].state) {
 		struct i915_gem_context *to = req->ctx;
@@ -1032,6 +1026,9 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
 		else
 			args->value = to_i915(dev)->ggtt.base.total;
 		break;
+	case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
+		args->value = !!(ctx->flags & CONTEXT_NO_ERROR_CAPTURE);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -1077,6 +1074,16 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 			ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0;
 		}
 		break;
+	case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
+		if (args->size) {
+			ret = -EINVAL;
+		} else {
+			if (args->value)
+				ctx->flags |= CONTEXT_NO_ERROR_CAPTURE;
+			else
+				ctx->flags &= ~CONTEXT_NO_ERROR_CAPTURE;
+		}
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -1089,7 +1096,7 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
 int i915_gem_context_reset_stats_ioctl(struct drm_device *dev,
 				       void *data, struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_reset_stats *args = data;
 	struct i915_ctx_hang_stats *hs;
 	struct i915_gem_context *ctx;
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index b144c3f5c650..3c1280ec7ff6 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -33,6 +33,37 @@
 #include "intel_drv.h"
 #include "i915_trace.h"
 
+static int switch_to_pinned_context(struct drm_i915_private *dev_priv)
+{
+	struct intel_engine_cs *engine;
+
+	if (i915.enable_execlists)
+		return 0;
+
+	for_each_engine(engine, dev_priv) {
+		struct drm_i915_gem_request *req;
+		int ret;
+
+		if (engine->last_context == NULL)
+			continue;
+
+		if (engine->last_context == dev_priv->kernel_context)
+			continue;
+
+		req = i915_gem_request_alloc(engine, dev_priv->kernel_context);
+		if (IS_ERR(req))
+			return PTR_ERR(req);
+
+		ret = i915_switch_context(req);
+		i915_add_request_no_flush(req);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+
 static bool
 mark_free(struct i915_vma *vma, struct list_head *unwind)
 {
@@ -150,11 +181,19 @@ none:
 
 	/* Only idle the GPU and repeat the search once */
 	if (pass++ == 0) {
-		ret = i915_gpu_idle(dev);
+		struct drm_i915_private *dev_priv = to_i915(dev);
+
+		if (i915_is_ggtt(vm)) {
+			ret = switch_to_pinned_context(dev_priv);
+			if (ret)
+				return ret;
+		}
+
+		ret = i915_gem_wait_for_idle(dev_priv);
 		if (ret)
 			return ret;
 
-		i915_gem_retire_requests(to_i915(dev));
+		i915_gem_retire_requests(dev_priv);
 		goto search_again;
 	}
 
@@ -261,11 +300,19 @@ int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle)
 	trace_i915_gem_evict_vm(vm);
 
 	if (do_idle) {
-		ret = i915_gpu_idle(vm->dev);
+		struct drm_i915_private *dev_priv = to_i915(vm->dev);
+
+		if (i915_is_ggtt(vm)) {
+			ret = switch_to_pinned_context(dev_priv);
+			if (ret)
+				return ret;
+		}
+
+		ret = i915_gem_wait_for_idle(dev_priv);
 		if (ret)
 			return ret;
 
-		i915_gem_retire_requests(to_i915(vm->dev));
+		i915_gem_retire_requests(dev_priv);
 
 		WARN_ON(!list_empty(&vm->active_list));
 	}
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 7941f1fe9cd2..1978633e7549 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -1142,7 +1142,7 @@ i915_reset_gen7_sol_offsets(struct drm_device *dev,
 			    struct drm_i915_gem_request *req)
 {
 	struct intel_engine_cs *engine = req->engine;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret, i;
 
 	if (!IS_GEN7(dev) || engine != &dev_priv->engine[RCS]) {
@@ -1225,7 +1225,7 @@ i915_gem_ringbuffer_submission(struct i915_execbuffer_params *params,
 {
 	struct drm_device *dev = params->dev;
 	struct intel_engine_cs *engine = params->engine;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u64 exec_start, exec_len;
 	int instp_mode;
 	u32 instp_mask;
@@ -1328,10 +1328,10 @@ gen8_dispatch_bsd_ring(struct drm_i915_private *dev_priv, struct drm_file *file)
 	/* Check whether the file_priv has already selected one ring. */
 	if ((int)file_priv->bsd_ring < 0) {
 		/* If not, use the ping-pong mechanism to select one. */
-		mutex_lock(&dev_priv->dev->struct_mutex);
+		mutex_lock(&dev_priv->drm.struct_mutex);
 		file_priv->bsd_ring = dev_priv->mm.bsd_ring_dispatch_index;
 		dev_priv->mm.bsd_ring_dispatch_index ^= 1;
-		mutex_unlock(&dev_priv->dev->struct_mutex);
+		mutex_unlock(&dev_priv->drm.struct_mutex);
 	}
 
 	return file_priv->bsd_ring;
@@ -1477,6 +1477,12 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		dispatch_flags |= I915_DISPATCH_RS;
 	}
 
+	/* Take a local wakeref for preparing to dispatch the execbuf as
+	 * we expect to access the hardware fairly frequently in the
+	 * process. Upon first dispatch, we acquire another prolonged
+	 * wakeref that we hold until the GPU has been idle for at least
+	 * 100ms.
+	 */
 	intel_runtime_pm_get(dev_priv);
 
 	ret = i915_mutex_lock_interruptible(dev);
diff --git a/drivers/gpu/drm/i915/i915_gem_fence.c b/drivers/gpu/drm/i915/i915_gem_fence.c
index 2b6bdc267fb5..251d7a95af89 100644
--- a/drivers/gpu/drm/i915/i915_gem_fence.c
+++ b/drivers/gpu/drm/i915/i915_gem_fence.c
@@ -58,7 +58,7 @@
 static void i965_write_fence_reg(struct drm_device *dev, int reg,
 				 struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t fence_reg_lo, fence_reg_hi;
 	int fence_pitch_shift;
 
@@ -117,7 +117,7 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 static void i915_write_fence_reg(struct drm_device *dev, int reg,
 				 struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val;
 
 	if (obj) {
@@ -156,7 +156,7 @@ static void i915_write_fence_reg(struct drm_device *dev, int reg,
 static void i830_write_fence_reg(struct drm_device *dev, int reg,
 				struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t val;
 
 	if (obj) {
@@ -193,7 +193,7 @@ inline static bool i915_gem_object_needs_mb(struct drm_i915_gem_object *obj)
 static void i915_gem_write_fence(struct drm_device *dev, int reg,
 				 struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Ensure that all CPU reads are completed before installing a fence
 	 * and all writes before removing the fence.
@@ -229,7 +229,7 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 					 struct drm_i915_fence_reg *fence,
 					 bool enable)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	int reg = fence_number(dev_priv, fence);
 
 	i915_gem_write_fence(obj->base.dev, reg, enable ? obj : NULL);
@@ -286,7 +286,7 @@ i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
 int
 i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	struct drm_i915_fence_reg *fence;
 	int ret;
 
@@ -311,7 +311,7 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 static struct drm_i915_fence_reg *
 i915_find_fence_reg(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_fence_reg *reg, *avail;
 	int i;
 
@@ -367,7 +367,7 @@ int
 i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool enable = obj->tiling_mode != I915_TILING_NONE;
 	struct drm_i915_fence_reg *reg;
 	int ret;
@@ -433,7 +433,7 @@ bool
 i915_gem_object_pin_fence(struct drm_i915_gem_object *obj)
 {
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
-		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 		struct i915_vma *ggtt_vma = i915_gem_obj_to_ggtt(obj);
 
 		WARN_ON(!ggtt_vma ||
@@ -457,7 +457,7 @@ void
 i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
 {
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
-		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 		WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count <= 0);
 		dev_priv->fence_regs[obj->fence_reg].pin_count--;
 	}
@@ -472,7 +472,7 @@ i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
  */
 void i915_gem_restore_fences(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	for (i = 0; i < dev_priv->num_fence_regs; i++) {
@@ -549,7 +549,7 @@ void i915_gem_restore_fences(struct drm_device *dev)
 void
 i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
 	uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 5890017b9832..10f1e32767e6 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -153,7 +153,7 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv,
 #endif
 
 	/* Early VLV doesn't have this */
-	if (IS_VALLEYVIEW(dev_priv) && dev_priv->dev->pdev->revision < 0xb) {
+	if (IS_VALLEYVIEW(dev_priv) && dev_priv->drm.pdev->revision < 0xb) {
 		DRM_DEBUG_DRIVER("disabling PPGTT on pre-B3 step VLV\n");
 		return 0;
 	}
@@ -1570,13 +1570,13 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 	struct i915_page_table *unused;
 	gen6_pte_t scratch_pte;
 	uint32_t pd_entry;
-	uint32_t  pte, pde, temp;
+	uint32_t  pte, pde;
 	uint32_t start = ppgtt->base.start, length = ppgtt->base.total;
 
 	scratch_pte = vm->pte_encode(px_dma(vm->scratch_page),
 				     I915_CACHE_LLC, true, 0);
 
-	gen6_for_each_pde(unused, &ppgtt->pd, start, length, temp, pde) {
+	gen6_for_each_pde(unused, &ppgtt->pd, start, length, pde) {
 		u32 expected;
 		gen6_pte_t *pt_vaddr;
 		const dma_addr_t pt_addr = px_dma(ppgtt->pd.page_table[pde]);
@@ -1640,9 +1640,9 @@ static void gen6_write_page_range(struct drm_i915_private *dev_priv,
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct i915_page_table *pt;
-	uint32_t pde, temp;
+	uint32_t pde;
 
-	gen6_for_each_pde(pt, pd, start, length, temp, pde)
+	gen6_for_each_pde(pt, pd, start, length, pde)
 		gen6_write_pde(pd, pde, pt);
 
 	/* Make sure write is complete before other code can use this page
@@ -1683,17 +1683,6 @@ static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	return 0;
 }
 
-static int vgpu_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct drm_i915_gem_request *req)
-{
-	struct intel_engine_cs *engine = req->engine;
-	struct drm_i915_private *dev_priv = to_i915(ppgtt->base.dev);
-
-	I915_WRITE(RING_PP_DIR_DCLV(engine), PP_DIR_DCLV_2G);
-	I915_WRITE(RING_PP_DIR_BASE(engine), get_pd_offset(ppgtt));
-	return 0;
-}
-
 static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 			  struct drm_i915_gem_request *req)
 {
@@ -1731,21 +1720,16 @@ static int gen6_mm_switch(struct i915_hw_ppgtt *ppgtt,
 			  struct drm_i915_gem_request *req)
 {
 	struct intel_engine_cs *engine = req->engine;
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
+	struct drm_i915_private *dev_priv = req->i915;
 
 	I915_WRITE(RING_PP_DIR_DCLV(engine), PP_DIR_DCLV_2G);
 	I915_WRITE(RING_PP_DIR_BASE(engine), get_pd_offset(ppgtt));
-
-	POSTING_READ(RING_PP_DIR_DCLV(engine));
-
 	return 0;
 }
 
 static void gen8_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 
 	for_each_engine(engine, dev_priv) {
@@ -1757,7 +1741,7 @@ static void gen8_ppgtt_enable(struct drm_device *dev)
 
 static void gen7_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine;
 	uint32_t ecochk, ecobits;
 
@@ -1782,7 +1766,7 @@ static void gen7_ppgtt_enable(struct drm_device *dev)
 
 static void gen6_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t ecochk, gab_ctl, ecobits;
 
 	ecobits = I915_READ(GAC_ECO_BITS);
@@ -1875,7 +1859,7 @@ static int gen6_alloc_va_range(struct i915_address_space *vm,
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_table *pt;
 	uint32_t start, length, start_save, length_save;
-	uint32_t pde, temp;
+	uint32_t pde;
 	int ret;
 
 	if (WARN_ON(start_in + length_in > ppgtt->base.total))
@@ -1891,7 +1875,7 @@ static int gen6_alloc_va_range(struct i915_address_space *vm,
 	 * need allocation. The second stage marks use ptes within the page
 	 * tables.
 	 */
-	gen6_for_each_pde(pt, &ppgtt->pd, start, length, temp, pde) {
+	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
 		if (pt != vm->scratch_pt) {
 			WARN_ON(bitmap_empty(pt->used_ptes, GEN6_PTES));
 			continue;
@@ -1916,7 +1900,7 @@ static int gen6_alloc_va_range(struct i915_address_space *vm,
 	start = start_save;
 	length = length_save;
 
-	gen6_for_each_pde(pt, &ppgtt->pd, start, length, temp, pde) {
+	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
 		DECLARE_BITMAP(tmp_bitmap, GEN6_PTES);
 
 		bitmap_zero(tmp_bitmap, GEN6_PTES);
@@ -1985,15 +1969,16 @@ static void gen6_free_scratch(struct i915_address_space *vm)
 static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct i915_page_directory *pd = &ppgtt->pd;
+	struct drm_device *dev = vm->dev;
 	struct i915_page_table *pt;
 	uint32_t pde;
 
 	drm_mm_remove_node(&ppgtt->node);
 
-	gen6_for_all_pdes(pt, ppgtt, pde) {
+	gen6_for_all_pdes(pt, pd, pde)
 		if (pt != vm->scratch_pt)
-			free_pt(ppgtt->base.dev, pt);
-	}
+			free_pt(dev, pt);
 
 	gen6_free_scratch(vm);
 }
@@ -2059,9 +2044,9 @@ static void gen6_scratch_va_range(struct i915_hw_ppgtt *ppgtt,
 				  uint64_t start, uint64_t length)
 {
 	struct i915_page_table *unused;
-	uint32_t pde, temp;
+	uint32_t pde;
 
-	gen6_for_each_pde(unused, &ppgtt->pd, start, length, temp, pde)
+	gen6_for_each_pde(unused, &ppgtt->pd, start, length, pde)
 		ppgtt->pd.page_table[pde] = ppgtt->base.scratch_pt;
 }
 
@@ -2073,18 +2058,15 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	int ret;
 
 	ppgtt->base.pte_encode = ggtt->base.pte_encode;
-	if (IS_GEN6(dev)) {
+	if (intel_vgpu_active(dev_priv) || IS_GEN6(dev))
 		ppgtt->switch_mm = gen6_mm_switch;
-	} else if (IS_HASWELL(dev)) {
+	else if (IS_HASWELL(dev))
 		ppgtt->switch_mm = hsw_mm_switch;
-	} else if (IS_GEN7(dev)) {
+	else if (IS_GEN7(dev))
 		ppgtt->switch_mm = gen7_mm_switch;
-	} else
+	else
 		BUG();
 
-	if (intel_vgpu_active(dev_priv))
-		ppgtt->switch_mm = vgpu_mm_switch;
-
 	ret = gen6_ppgtt_alloc(ppgtt);
 	if (ret)
 		return ret;
@@ -2133,7 +2115,7 @@ static void i915_address_space_init(struct i915_address_space *vm,
 				    struct drm_i915_private *dev_priv)
 {
 	drm_mm_init(&vm->mm, vm->start, vm->total);
-	vm->dev = dev_priv->dev;
+	vm->dev = &dev_priv->drm;
 	INIT_LIST_HEAD(&vm->active_list);
 	INIT_LIST_HEAD(&vm->inactive_list);
 	list_add_tail(&vm->global_link, &dev_priv->vm_list);
@@ -2141,7 +2123,7 @@ static void i915_address_space_init(struct i915_address_space *vm,
 
 static void gtt_write_workarounds(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* This function is for gtt related workarounds. This function is
 	 * called on driver load and after a GPU reset, so you can place
@@ -2160,7 +2142,7 @@ static void gtt_write_workarounds(struct drm_device *dev)
 
 static int i915_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
 	ret = __hw_ppgtt_init(dev, ppgtt);
@@ -2261,8 +2243,8 @@ static bool do_idling(struct drm_i915_private *dev_priv)
 
 	if (unlikely(ggtt->do_idle_maps)) {
 		dev_priv->mm.interruptible = false;
-		if (i915_gpu_idle(dev_priv->dev)) {
-			DRM_ERROR("Couldn't idle GPU\n");
+		if (i915_gem_wait_for_idle(dev_priv)) {
+			DRM_ERROR("Failed to wait for idle; VT'd may hang.\n");
 			/* Wait a bit, in hopes it avoids the hang */
 			udelay(10);
 		}
@@ -2610,7 +2592,7 @@ static void i915_ggtt_insert_entries(struct i915_address_space *vm,
 				     uint64_t start,
 				     enum i915_cache_level cache_level, u32 unused)
 {
-	struct drm_i915_private *dev_priv = vm->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(vm->dev);
 	unsigned int flags = (cache_level == I915_CACHE_NONE) ?
 		AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
 	int rpm_atomic_seq;
@@ -2628,7 +2610,7 @@ static void i915_ggtt_clear_range(struct i915_address_space *vm,
 				  uint64_t length,
 				  bool unused)
 {
-	struct drm_i915_private *dev_priv = vm->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(vm->dev);
 	unsigned first_entry = start >> PAGE_SHIFT;
 	unsigned num_entries = length >> PAGE_SHIFT;
 	int rpm_atomic_seq;
@@ -2709,7 +2691,7 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma,
 static void ggtt_unbind_vma(struct i915_vma *vma)
 {
 	struct drm_device *dev = vma->vm->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj = vma->obj;
 	const uint64_t size = min_t(uint64_t,
 				    obj->base.size,
@@ -2735,7 +2717,7 @@ static void ggtt_unbind_vma(struct i915_vma *vma)
 void i915_gem_gtt_finish_object(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool interruptible;
 
 	interruptible = do_idling(dev_priv);
@@ -3137,7 +3119,7 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)
 	ggtt->base.unbind_vma = ggtt_unbind_vma;
 	ggtt->base.insert_page = gen8_ggtt_insert_page;
 	ggtt->base.clear_range = nop_clear_range;
-	if (!USES_FULL_PPGTT(dev_priv))
+	if (!USES_FULL_PPGTT(dev_priv) || intel_scanout_needs_vtd_wa(dev_priv))
 		ggtt->base.clear_range = gen8_ggtt_clear_range;
 
 	ggtt->base.insert_entries = gen8_ggtt_insert_entries;
@@ -3197,7 +3179,7 @@ static int i915_gmch_probe(struct i915_ggtt *ggtt)
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
-	ret = intel_gmch_probe(dev_priv->bridge_dev, dev_priv->dev->pdev, NULL);
+	ret = intel_gmch_probe(dev_priv->bridge_dev, dev_priv->drm.pdev, NULL);
 	if (!ret) {
 		DRM_ERROR("failed to set up gmch\n");
 		return -EIO;
@@ -3206,7 +3188,7 @@ static int i915_gmch_probe(struct i915_ggtt *ggtt)
 	intel_gtt_get(&ggtt->base.total, &ggtt->stolen_size,
 		      &ggtt->mappable_base, &ggtt->mappable_end);
 
-	ggtt->do_idle_maps = needs_idle_maps(dev_priv->dev);
+	ggtt->do_idle_maps = needs_idle_maps(&dev_priv->drm);
 	ggtt->base.insert_page = i915_ggtt_insert_page;
 	ggtt->base.insert_entries = i915_ggtt_insert_entries;
 	ggtt->base.clear_range = i915_ggtt_clear_range;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 163b564fb87d..aa5f31d1c2ed 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -390,27 +390,27 @@ struct i915_hw_ppgtt {
 	void (*debug_dump)(struct i915_hw_ppgtt *ppgtt, struct seq_file *m);
 };
 
-/* For each pde iterates over every pde between from start until start + length.
- * If start, and start+length are not perfectly divisible, the macro will round
- * down, and up as needed. The macro modifies pde, start, and length. Dev is
- * only used to differentiate shift values. Temp is temp.  On gen6/7, start = 0,
- * and length = 2G effectively iterates over every PDE in the system.
- *
- * XXX: temp is not actually needed, but it saves doing the ALIGN operation.
+/*
+ * gen6_for_each_pde() iterates over every pde from start until start+length.
+ * If start and start+length are not perfectly divisible, the macro will round
+ * down and up as needed. Start=0 and length=2G effectively iterates over
+ * every PDE in the system. The macro modifies ALL its parameters except 'pd',
+ * so each of the other parameters should preferably be a simple variable, or
+ * at most an lvalue with no side-effects!
  */
-#define gen6_for_each_pde(pt, pd, start, length, temp, iter) \
-	for (iter = gen6_pde_index(start); \
-	     length > 0 && iter < I915_PDES ? \
-			(pt = (pd)->page_table[iter]), 1 : 0; \
-	     iter++, \
-	     temp = ALIGN(start+1, 1 << GEN6_PDE_SHIFT) - start, \
-	     temp = min_t(unsigned, temp, length), \
-	     start += temp, length -= temp)
-
-#define gen6_for_all_pdes(pt, ppgtt, iter)  \
-	for (iter = 0;		\
-	     pt = ppgtt->pd.page_table[iter], iter < I915_PDES;	\
-	     iter++)
+#define gen6_for_each_pde(pt, pd, start, length, iter)			\
+	for (iter = gen6_pde_index(start);				\
+	     length > 0 && iter < I915_PDES &&				\
+		(pt = (pd)->page_table[iter], true);			\
+	     ({ u32 temp = ALIGN(start+1, 1 << GEN6_PDE_SHIFT);		\
+		    temp = min(temp - start, length);			\
+		    start += temp, length -= temp; }), ++iter)
+
+#define gen6_for_all_pdes(pt, pd, iter)					\
+	for (iter = 0;							\
+	     iter < I915_PDES &&					\
+		(pt = (pd)->page_table[iter], true);			\
+	     ++iter)
 
 static inline uint32_t i915_pte_index(uint64_t address, uint32_t pde_shift)
 {
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c
index b7c1b5fb61ea..f75bbd67a13a 100644
--- a/drivers/gpu/drm/i915/i915_gem_render_state.c
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.c
@@ -58,7 +58,7 @@ static int render_state_init(struct render_state *so,
 	if (so->rodata->batch_items * 4 > 4096)
 		return -EINVAL;
 
-	so->obj = i915_gem_object_create(dev_priv->dev, 4096);
+	so->obj = i915_gem_object_create(&dev_priv->drm, 4096);
 	if (IS_ERR(so->obj))
 		return PTR_ERR(so->obj);
 
diff --git a/drivers/gpu/drm/i915/i915_gem_shrinker.c b/drivers/gpu/drm/i915/i915_gem_shrinker.c
index 538c30499848..067632ad2f29 100644
--- a/drivers/gpu/drm/i915/i915_gem_shrinker.c
+++ b/drivers/gpu/drm/i915/i915_gem_shrinker.c
@@ -257,7 +257,7 @@ i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc)
 {
 	struct drm_i915_private *dev_priv =
 		container_of(shrinker, struct drm_i915_private, mm.shrinker);
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_i915_gem_object *obj;
 	unsigned long count;
 	bool unlock;
@@ -265,6 +265,8 @@ i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc)
 	if (!i915_gem_shrinker_lock(dev, &unlock))
 		return 0;
 
+	i915_gem_retire_requests(dev_priv);
+
 	count = 0;
 	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)
 		if (can_release_pages(obj))
@@ -286,7 +288,7 @@ i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)
 {
 	struct drm_i915_private *dev_priv =
 		container_of(shrinker, struct drm_i915_private, mm.shrinker);
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	unsigned long freed;
 	bool unlock;
 
@@ -321,7 +323,7 @@ i915_gem_shrinker_lock_uninterruptible(struct drm_i915_private *dev_priv,
 {
 	unsigned long timeout = msecs_to_jiffies(timeout_ms) + 1;
 
-	while (!i915_gem_shrinker_lock(dev_priv->dev, &slu->unlock)) {
+	while (!i915_gem_shrinker_lock(&dev_priv->drm, &slu->unlock)) {
 		schedule_timeout_killable(1);
 		if (fatal_signal_pending(current))
 			return false;
@@ -342,7 +344,7 @@ i915_gem_shrinker_unlock_uninterruptible(struct drm_i915_private *dev_priv,
 {
 	dev_priv->mm.interruptible = slu->was_interruptible;
 	if (slu->unlock)
-		mutex_unlock(&dev_priv->dev->struct_mutex);
+		mutex_unlock(&dev_priv->drm.struct_mutex);
 }
 
 static int
@@ -408,7 +410,7 @@ i915_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr
 		return NOTIFY_DONE;
 
 	/* Force everything onto the inactive lists */
-	ret = i915_gpu_idle(dev_priv->dev);
+	ret = i915_gem_wait_for_idle(dev_priv);
 	if (ret)
 		goto out;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index e9cd82290408..66be299a1486 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -270,7 +270,7 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 
 void i915_gem_cleanup_stolen(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return;
@@ -550,7 +550,7 @@ static void i915_gem_object_put_pages_stolen(struct drm_i915_gem_object *obj)
 static void
 i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 
 	if (obj->stolen) {
 		i915_gem_stolen_remove_node(dev_priv, obj->stolen);
@@ -601,7 +601,7 @@ cleanup:
 struct drm_i915_gem_object *
 i915_gem_object_create_stolen(struct drm_device *dev, u32 size)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 	int ret;
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index a6eb5c47a49c..8030199731db 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -162,7 +162,7 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 		   struct drm_file *file)
 {
 	struct drm_i915_gem_set_tiling *args = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	int ret = 0;
 
@@ -294,7 +294,7 @@ i915_gem_get_tiling(struct drm_device *dev, void *data,
 		   struct drm_file *file)
 {
 	struct drm_i915_gem_get_tiling *args = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 
 	obj = to_intel_bo(drm_gem_object_lookup(file, args->handle));
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 34ff2459ceea..9d73d2216adc 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -332,7 +332,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 			    const struct i915_error_state_file_priv *error_priv)
 {
 	struct drm_device *dev = error_priv->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_error_state *error = error_priv->error;
 	struct drm_i915_error_object *obj;
 	int i, j, offset, elt;
@@ -463,6 +463,18 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 			}
 		}
 
+		if (error->ring[i].num_waiters) {
+			err_printf(m, "%s --- %d waiters\n",
+				   dev_priv->engine[i].name,
+				   error->ring[i].num_waiters);
+			for (j = 0; j < error->ring[i].num_waiters; j++) {
+				err_printf(m, " seqno 0x%08x for %s [%d]\n",
+					   error->ring[i].waiters[j].seqno,
+					   error->ring[i].waiters[j].comm,
+					   error->ring[i].waiters[j].pid);
+			}
+		}
+
 		if ((obj = error->ring[i].ringbuffer)) {
 			err_printf(m, "%s --- ringbuffer = 0x%08x\n",
 				   dev_priv->engine[i].name,
@@ -488,7 +500,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 					   hws_page[elt+1],
 					   hws_page[elt+2],
 					   hws_page[elt+3]);
-					offset += 16;
+				offset += 16;
 			}
 		}
 
@@ -605,8 +617,9 @@ static void i915_error_state_free(struct kref *error_ref)
 		i915_error_object_free(error->ring[i].ringbuffer);
 		i915_error_object_free(error->ring[i].hws_page);
 		i915_error_object_free(error->ring[i].ctx);
-		kfree(error->ring[i].requests);
 		i915_error_object_free(error->ring[i].wa_ctx);
+		kfree(error->ring[i].requests);
+		kfree(error->ring[i].waiters);
 	}
 
 	i915_error_object_free(error->semaphore_obj);
@@ -892,6 +905,48 @@ static void gen6_record_semaphore_state(struct drm_i915_private *dev_priv,
 	}
 }
 
+static void engine_record_waiters(struct intel_engine_cs *engine,
+				  struct drm_i915_error_ring *ering)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct drm_i915_error_waiter *waiter;
+	struct rb_node *rb;
+	int count;
+
+	ering->num_waiters = 0;
+	ering->waiters = NULL;
+
+	spin_lock(&b->lock);
+	count = 0;
+	for (rb = rb_first(&b->waiters); rb != NULL; rb = rb_next(rb))
+		count++;
+	spin_unlock(&b->lock);
+
+	waiter = NULL;
+	if (count)
+		waiter = kmalloc_array(count,
+				       sizeof(struct drm_i915_error_waiter),
+				       GFP_ATOMIC);
+	if (!waiter)
+		return;
+
+	ering->waiters = waiter;
+
+	spin_lock(&b->lock);
+	for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
+		struct intel_wait *w = container_of(rb, typeof(*w), node);
+
+		strcpy(waiter->comm, w->tsk->comm);
+		waiter->pid = w->tsk->pid;
+		waiter->seqno = w->seqno;
+		waiter++;
+
+		if (++ering->num_waiters == count)
+			break;
+	}
+	spin_unlock(&b->lock);
+}
+
 static void i915_record_ring_state(struct drm_i915_private *dev_priv,
 				   struct drm_i915_error_state *error,
 				   struct intel_engine_cs *engine,
@@ -926,10 +981,10 @@ static void i915_record_ring_state(struct drm_i915_private *dev_priv,
 		ering->instdone = I915_READ(GEN2_INSTDONE);
 	}
 
-	ering->waiting = waitqueue_active(&engine->irq_queue);
+	ering->waiting = intel_engine_has_waiter(engine);
 	ering->instpm = I915_READ(RING_INSTPM(engine->mmio_base));
 	ering->acthd = intel_ring_get_active_head(engine);
-	ering->seqno = engine->get_seqno(engine);
+	ering->seqno = intel_engine_get_seqno(engine);
 	ering->last_seqno = engine->last_submitted_seqno;
 	ering->start = I915_READ_START(engine);
 	ering->head = I915_READ_HEAD(engine);
@@ -1022,7 +1077,6 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 
 	for (i = 0; i < I915_NUM_ENGINES; i++) {
 		struct intel_engine_cs *engine = &dev_priv->engine[i];
-		struct intel_ringbuffer *rbuf;
 
 		error->ring[i].pid = -1;
 
@@ -1032,14 +1086,15 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 		error->ring[i].valid = true;
 
 		i915_record_ring_state(dev_priv, error, engine, &error->ring[i]);
+		engine_record_waiters(engine, &error->ring[i]);
 
 		request = i915_gem_find_active_request(engine);
 		if (request) {
 			struct i915_address_space *vm;
+			struct intel_ringbuffer *rb;
 
-			vm = request->ctx && request->ctx->ppgtt ?
-				&request->ctx->ppgtt->base :
-				&ggtt->base;
+			vm = request->ctx->ppgtt ?
+				&request->ctx->ppgtt->base : &ggtt->base;
 
 			/* We need to copy these to an anonymous buffer
 			 * as the simplest method to avoid being overwritten
@@ -1066,26 +1121,17 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 				}
 				rcu_read_unlock();
 			}
-		}
-
-		if (i915.enable_execlists) {
-			/* TODO: This is only a small fix to keep basic error
-			 * capture working, but we need to add more information
-			 * for it to be useful (e.g. dump the context being
-			 * executed).
-			 */
-			if (request)
-				rbuf = request->ctx->engine[engine->id].ringbuf;
-			else
-				rbuf = dev_priv->kernel_context->engine[engine->id].ringbuf;
-		} else
-			rbuf = engine->buffer;
 
-		error->ring[i].cpu_ring_head = rbuf->head;
-		error->ring[i].cpu_ring_tail = rbuf->tail;
+			error->simulated |=
+				request->ctx->flags & CONTEXT_NO_ERROR_CAPTURE;
 
-		error->ring[i].ringbuffer =
-			i915_error_ggtt_object_create(dev_priv, rbuf->obj);
+			rb = request->ringbuf;
+			error->ring[i].cpu_ring_head = rb->head;
+			error->ring[i].cpu_ring_tail = rb->tail;
+			error->ring[i].ringbuffer =
+				i915_error_ggtt_object_create(dev_priv,
+							      rb->obj);
+		}
 
 		error->ring[i].hws_page =
 			i915_error_ggtt_object_create(dev_priv,
@@ -1230,7 +1276,7 @@ static void i915_gem_capture_buffers(struct drm_i915_private *dev_priv,
 static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 				   struct drm_i915_error_state *error)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	int i;
 
 	/* General organization
@@ -1355,6 +1401,9 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	struct drm_i915_error_state *error;
 	unsigned long flags;
 
+	if (READ_ONCE(dev_priv->gpu_error.first_error))
+		return;
+
 	/* Account for pipe specific data like PIPE*STAT */
 	error = kzalloc(sizeof(*error), GFP_ATOMIC);
 	if (!error) {
@@ -1378,12 +1427,14 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	i915_error_capture_msg(dev_priv, error, engine_mask, error_msg);
 	DRM_INFO("%s\n", error->error_msg);
 
-	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
-	if (dev_priv->gpu_error.first_error == NULL) {
-		dev_priv->gpu_error.first_error = error;
-		error = NULL;
+	if (!error->simulated) {
+		spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+		if (!dev_priv->gpu_error.first_error) {
+			dev_priv->gpu_error.first_error = error;
+			error = NULL;
+		}
+		spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 	}
-	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
 	if (error) {
 		i915_error_state_free(&error->ref);
@@ -1395,7 +1446,8 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 		DRM_INFO("Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel\n");
 		DRM_INFO("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n");
 		DRM_INFO("The gpu crash dump is required to analyze gpu hangs, so please always attach it.\n");
-		DRM_INFO("GPU crash dump saved to /sys/class/drm/card%d/error\n", dev_priv->dev->primary->index);
+		DRM_INFO("GPU crash dump saved to /sys/class/drm/card%d/error\n",
+			 dev_priv->drm.primary->index);
 		warned = true;
 	}
 }
@@ -1403,7 +1455,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 void i915_error_state_get(struct drm_device *dev,
 			  struct i915_error_state_file_priv *error_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	spin_lock_irq(&dev_priv->gpu_error.lock);
 	error_priv->error = dev_priv->gpu_error.first_error;
@@ -1421,7 +1473,7 @@ void i915_error_state_put(struct i915_error_state_file_priv *error_priv)
 
 void i915_destroy_error_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_error_state *error;
 
 	spin_lock_irq(&dev_priv->gpu_error.lock);
diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 22a55ac4e51c..2112e029db6a 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -97,8 +97,14 @@ static int host2guc_action(struct intel_guc *guc, u32 *data, u32 len)
 
 	I915_WRITE(HOST2GUC_INTERRUPT, HOST2GUC_TRIGGER);
 
-	/* No HOST2GUC command should take longer than 10ms */
-	ret = wait_for_atomic(host2guc_action_response(dev_priv, &status), 10);
+	/*
+	 * Fast commands should complete in less than 10us, so sample quickly
+	 * up to that length of time, then switch to a slower sleep-wait loop.
+	 * No HOST2GUC command should ever take longer than 10ms.
+	 */
+	ret = wait_for_us(host2guc_action_response(dev_priv, &status), 10);
+	if (ret)
+		ret = wait_for(host2guc_action_response(dev_priv, &status), 10);
 	if (status != GUC2HOST_STATUS_SUCCESS) {
 		/*
 		 * Either the GuC explicitly returned an error (which
@@ -153,12 +159,11 @@ static int host2guc_sample_forcewake(struct intel_guc *guc,
 				     struct i915_guc_client *client)
 {
 	struct drm_i915_private *dev_priv = guc_to_i915(guc);
-	struct drm_device *dev = dev_priv->dev;
 	u32 data[2];
 
 	data[0] = HOST2GUC_ACTION_SAMPLE_FORCEWAKE;
 	/* WaRsDisableCoarsePowerGating:skl,bxt */
-	if (!intel_enable_rc6() || NEEDS_WaRsDisableCoarsePowerGating(dev))
+	if (!intel_enable_rc6() || NEEDS_WaRsDisableCoarsePowerGating(dev_priv))
 		data[1] = 0;
 	else
 		/* bit 0 and 1 are for Render and Media domain separately */
@@ -582,7 +587,7 @@ static int guc_ring_doorbell(struct i915_guc_client *gc)
  */
 int i915_guc_submit(struct drm_i915_gem_request *rq)
 {
-	unsigned int engine_id = rq->engine->guc_id;
+	unsigned int engine_id = rq->engine->id;
 	struct intel_guc *guc = &rq->i915->guc;
 	struct i915_guc_client *client = guc->execbuf_client;
 	int b_ret;
@@ -623,7 +628,7 @@ gem_allocate_guc_obj(struct drm_i915_private *dev_priv, u32 size)
 {
 	struct drm_i915_gem_object *obj;
 
-	obj = i915_gem_object_create(dev_priv->dev, size);
+	obj = i915_gem_object_create(&dev_priv->drm, size);
 	if (IS_ERR(obj))
 		return NULL;
 
@@ -1034,7 +1039,7 @@ void i915_guc_submission_fini(struct drm_i915_private *dev_priv)
  */
 int intel_guc_suspend(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc *guc = &dev_priv->guc;
 	struct i915_gem_context *ctx;
 	u32 data[3];
@@ -1060,7 +1065,7 @@ int intel_guc_suspend(struct drm_device *dev)
  */
 int intel_guc_resume(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc *guc = &dev_priv->guc;
 	struct i915_gem_context *ctx;
 	u32 data[3];
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 4378a659d962..1c2aec392412 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -259,12 +259,12 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv,
 	dev_priv->gt_irq_mask &= ~interrupt_mask;
 	dev_priv->gt_irq_mask |= (~enabled_irq_mask & interrupt_mask);
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
-	POSTING_READ(GTIMR);
 }
 
 void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	ilk_update_gt_irq(dev_priv, mask, mask);
+	POSTING_READ_FW(GTIMR);
 }
 
 void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
@@ -351,9 +351,8 @@ void gen6_reset_rps_interrupts(struct drm_i915_private *dev_priv)
 void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv)
 {
 	spin_lock_irq(&dev_priv->irq_lock);
-
-	WARN_ON(dev_priv->rps.pm_iir);
-	WARN_ON(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events);
+	WARN_ON_ONCE(dev_priv->rps.pm_iir);
+	WARN_ON_ONCE(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events);
 	dev_priv->rps.interrupts_enabled = true;
 	I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) |
 				dev_priv->pm_rps_events);
@@ -371,11 +370,6 @@ void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv)
 {
 	spin_lock_irq(&dev_priv->irq_lock);
 	dev_priv->rps.interrupts_enabled = false;
-	spin_unlock_irq(&dev_priv->irq_lock);
-
-	cancel_work_sync(&dev_priv->rps.work);
-
-	spin_lock_irq(&dev_priv->irq_lock);
 
 	I915_WRITE(GEN6_PMINTRMSK, gen6_sanitize_rps_pm_mask(dev_priv, ~0));
 
@@ -384,8 +378,15 @@ void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv)
 				~dev_priv->pm_rps_events);
 
 	spin_unlock_irq(&dev_priv->irq_lock);
+	synchronize_irq(dev_priv->drm.irq);
 
-	synchronize_irq(dev_priv->dev->irq);
+	/* Now that we will not be generating any more work, flush any
+	 * outsanding tasks. As we are called on the RPS idle path,
+	 * we will reset the GPU to minimum frequencies, so the current
+	 * state of the worker can be discarded.
+	 */
+	cancel_work_sync(&dev_priv->rps.work);
+	gen6_reset_rps_interrupts(dev_priv);
 }
 
 /**
@@ -565,7 +566,7 @@ i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	u32 enable_mask;
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		enable_mask = vlv_get_pipestat_enable_mask(dev_priv->dev,
+		enable_mask = vlv_get_pipestat_enable_mask(&dev_priv->drm,
 							   status_mask);
 	else
 		enable_mask = status_mask << 16;
@@ -579,7 +580,7 @@ i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	u32 enable_mask;
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		enable_mask = vlv_get_pipestat_enable_mask(dev_priv->dev,
+		enable_mask = vlv_get_pipestat_enable_mask(&dev_priv->drm,
 							   status_mask);
 	else
 		enable_mask = status_mask << 16;
@@ -666,7 +667,7 @@ static u32 i8xx_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
  */
 static u32 i915_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t high_frame, low_frame;
 	u32 high1, high2, low, pixel, vbl_start, hsync_start, htotal;
 	struct intel_crtc *intel_crtc =
@@ -713,7 +714,7 @@ static u32 i915_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
 
 static u32 g4x_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	return I915_READ(PIPE_FRMCOUNT_G4X(pipe));
 }
@@ -722,7 +723,7 @@ static u32 g4x_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
 static int __intel_get_crtc_scanline(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct drm_display_mode *mode = &crtc->base.hwmode;
 	enum pipe pipe = crtc->pipe;
 	int position, vtotal;
@@ -774,7 +775,7 @@ static int i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe,
 				    ktime_t *stime, ktime_t *etime,
 				    const struct drm_display_mode *mode)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int position;
@@ -895,7 +896,7 @@ static int i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe,
 
 int intel_get_crtc_scanline(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	unsigned long irqflags;
 	int position;
 
@@ -976,13 +977,11 @@ static void ironlake_rps_change_irq_handler(struct drm_i915_private *dev_priv)
 
 static void notify_ring(struct intel_engine_cs *engine)
 {
-	if (!intel_engine_initialized(engine))
-		return;
-
-	trace_i915_gem_request_notify(engine);
-	engine->user_interrupts++;
-
-	wake_up_all(&engine->irq_queue);
+	smp_store_mb(engine->breadcrumbs.irq_posted, true);
+	if (intel_engine_wakeup(engine)) {
+		trace_i915_gem_request_notify(engine);
+		engine->breadcrumbs.irq_wakeups++;
+	}
 }
 
 static void vlv_c0_read(struct drm_i915_private *dev_priv,
@@ -1063,7 +1062,7 @@ static bool any_waiters(struct drm_i915_private *dev_priv)
 	struct intel_engine_cs *engine;
 
 	for_each_engine(engine, dev_priv)
-		if (engine->irq_refcount)
+		if (intel_engine_has_waiter(engine))
 			return true;
 
 	return false;
@@ -1084,13 +1083,6 @@ static void gen6_pm_rps_work(struct work_struct *work)
 		return;
 	}
 
-	/*
-	 * The RPS work is synced during runtime suspend, we don't require a
-	 * wakeref. TODO: instead of disabling the asserts make sure that we
-	 * always hold an RPM reference while the work is running.
-	 */
-	DISABLE_RPM_WAKEREF_ASSERTS(dev_priv);
-
 	pm_iir = dev_priv->rps.pm_iir;
 	dev_priv->rps.pm_iir = 0;
 	/* Make sure not to corrupt PMIMR state used by ringbuffer on GEN6 */
@@ -1103,7 +1095,7 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	WARN_ON(pm_iir & ~dev_priv->pm_rps_events);
 
 	if ((pm_iir & dev_priv->pm_rps_events) == 0 && !client_boost)
-		goto out;
+		return;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
@@ -1158,8 +1150,6 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	intel_set_rps(dev_priv, new_delay);
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
-out:
-	ENABLE_RPM_WAKEREF_ASSERTS(dev_priv);
 }
 
 
@@ -1185,7 +1175,7 @@ static void ivybridge_parity_work(struct work_struct *work)
 	 * In order to prevent a get/put style interface, acquire struct mutex
 	 * any time we access those registers.
 	 */
-	mutex_lock(&dev_priv->dev->struct_mutex);
+	mutex_lock(&dev_priv->drm.struct_mutex);
 
 	/* If we've screwed up tracking, just let the interrupt fire again */
 	if (WARN_ON(!dev_priv->l3_parity.which_slice))
@@ -1221,7 +1211,7 @@ static void ivybridge_parity_work(struct work_struct *work)
 		parity_event[4] = kasprintf(GFP_KERNEL, "SLICE=%d", slice);
 		parity_event[5] = NULL;
 
-		kobject_uevent_env(&dev_priv->dev->primary->kdev->kobj,
+		kobject_uevent_env(&dev_priv->drm.primary->kdev->kobj,
 				   KOBJ_CHANGE, parity_event);
 
 		DRM_DEBUG("Parity error: Slice = %d, Row = %d, Bank = %d, Sub bank = %d.\n",
@@ -1241,7 +1231,7 @@ out:
 	gen5_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv));
 	spin_unlock_irq(&dev_priv->irq_lock);
 
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 }
 
 static void ivybridge_parity_error_irq_handler(struct drm_i915_private *dev_priv,
@@ -1267,8 +1257,7 @@ static void ivybridge_parity_error_irq_handler(struct drm_i915_private *dev_priv
 static void ilk_gt_irq_handler(struct drm_i915_private *dev_priv,
 			       u32 gt_iir)
 {
-	if (gt_iir &
-	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
+	if (gt_iir & GT_RENDER_USER_INTERRUPT)
 		notify_ring(&dev_priv->engine[RCS]);
 	if (gt_iir & ILK_BSD_USER_INTERRUPT)
 		notify_ring(&dev_priv->engine[VCS]);
@@ -1277,9 +1266,7 @@ static void ilk_gt_irq_handler(struct drm_i915_private *dev_priv,
 static void snb_gt_irq_handler(struct drm_i915_private *dev_priv,
 			       u32 gt_iir)
 {
-
-	if (gt_iir &
-	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
+	if (gt_iir & GT_RENDER_USER_INTERRUPT)
 		notify_ring(&dev_priv->engine[RCS]);
 	if (gt_iir & GT_BSD_USER_INTERRUPT)
 		notify_ring(&dev_priv->engine[VCS]);
@@ -1526,7 +1513,7 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
 
 	entry = &pipe_crc->entries[head];
 
-	entry->frame = dev_priv->dev->driver->get_vblank_counter(dev_priv->dev,
+	entry->frame = dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm,
 								 pipe);
 	entry->crc[0] = crc0;
 	entry->crc[1] = crc1;
@@ -1602,7 +1589,7 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 		gen6_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
 		if (dev_priv->rps.interrupts_enabled) {
 			dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
-			queue_work(dev_priv->wq, &dev_priv->rps.work);
+			schedule_work(&dev_priv->rps.work);
 		}
 		spin_unlock(&dev_priv->irq_lock);
 	}
@@ -1624,7 +1611,7 @@ static bool intel_pipe_handle_vblank(struct drm_i915_private *dev_priv,
 {
 	bool ret;
 
-	ret = drm_handle_vblank(dev_priv->dev, pipe);
+	ret = drm_handle_vblank(&dev_priv->drm, pipe);
 	if (ret)
 		intel_finish_page_flip_mmio(dev_priv, pipe);
 
@@ -1757,7 +1744,7 @@ static void i9xx_hpd_irq_handler(struct drm_i915_private *dev_priv,
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	irqreturn_t ret = IRQ_NONE;
 
 	if (!intel_irqs_enabled(dev_priv))
@@ -1840,7 +1827,7 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 static irqreturn_t cherryview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	irqreturn_t ret = IRQ_NONE;
 
 	if (!intel_irqs_enabled(dev_priv))
@@ -2225,7 +2212,7 @@ static void ivb_display_irq_handler(struct drm_i915_private *dev_priv,
 static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 de_iir, gt_iir, de_ier, sde_ier = 0;
 	irqreturn_t ret = IRQ_NONE;
 
@@ -2438,7 +2425,7 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 			I915_WRITE(SDEIIR, iir);
 			ret = IRQ_HANDLED;
 
-			if (HAS_PCH_SPT(dev_priv))
+			if (HAS_PCH_SPT(dev_priv) || HAS_PCH_KBP(dev_priv))
 				spt_irq_handler(dev_priv, iir);
 			else
 				cpt_irq_handler(dev_priv, iir);
@@ -2457,7 +2444,7 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 static irqreturn_t gen8_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 master_ctl;
 	u32 gt_iir[4] = {};
 	irqreturn_t ret;
@@ -2488,11 +2475,8 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 	return ret;
 }
 
-static void i915_error_wake_up(struct drm_i915_private *dev_priv,
-			       bool reset_completed)
+static void i915_error_wake_up(struct drm_i915_private *dev_priv)
 {
-	struct intel_engine_cs *engine;
-
 	/*
 	 * Notify all waiters for GPU completion events that reset state has
 	 * been changed, and that they need to restart their wait after
@@ -2501,18 +2485,10 @@ static void i915_error_wake_up(struct drm_i915_private *dev_priv,
 	 */
 
 	/* Wake up __wait_seqno, potentially holding dev->struct_mutex. */
-	for_each_engine(engine, dev_priv)
-		wake_up_all(&engine->irq_queue);
+	wake_up_all(&dev_priv->gpu_error.wait_queue);
 
 	/* Wake up intel_crtc_wait_for_pending_flips, holding crtc->mutex. */
 	wake_up_all(&dev_priv->pending_flip_queue);
-
-	/*
-	 * Signal tasks blocked in i915_gem_wait_for_error that the pending
-	 * reset state is cleared.
-	 */
-	if (reset_completed)
-		wake_up_all(&dev_priv->gpu_error.reset_queue);
 }
 
 /**
@@ -2524,7 +2500,7 @@ static void i915_error_wake_up(struct drm_i915_private *dev_priv,
  */
 static void i915_reset_and_wakeup(struct drm_i915_private *dev_priv)
 {
-	struct kobject *kobj = &dev_priv->dev->primary->kdev->kobj;
+	struct kobject *kobj = &dev_priv->drm.primary->kdev->kobj;
 	char *error_event[] = { I915_ERROR_UEVENT "=1", NULL };
 	char *reset_event[] = { I915_RESET_UEVENT "=1", NULL };
 	char *reset_done_event[] = { I915_ERROR_UEVENT "=0", NULL };
@@ -2577,7 +2553,7 @@ static void i915_reset_and_wakeup(struct drm_i915_private *dev_priv)
 		 * Note: The wake_up also serves as a memory barrier so that
 		 * waiters see the update value of the reset counter atomic_t.
 		 */
-		i915_error_wake_up(dev_priv, true);
+		wake_up_all(&dev_priv->gpu_error.reset_queue);
 	}
 }
 
@@ -2714,7 +2690,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv,
 		 * ensure that the waiters see the updated value of the reset
 		 * counter atomic_t.
 		 */
-		i915_error_wake_up(dev_priv, false);
+		i915_error_wake_up(dev_priv);
 	}
 
 	i915_reset_and_wakeup(dev_priv);
@@ -2725,7 +2701,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv,
  */
 static int i915_enable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2742,7 +2718,7 @@ static int i915_enable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static int ironlake_enable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 	uint32_t bit = (INTEL_INFO(dev)->gen >= 7) ? DE_PIPE_VBLANK_IVB(pipe) :
 						     DE_PIPE_VBLANK(pipe);
@@ -2756,7 +2732,7 @@ static int ironlake_enable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static int valleyview_enable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2769,7 +2745,7 @@ static int valleyview_enable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static int gen8_enable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2784,7 +2760,7 @@ static int gen8_enable_vblank(struct drm_device *dev, unsigned int pipe)
  */
 static void i915_disable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2796,7 +2772,7 @@ static void i915_disable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static void ironlake_disable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 	uint32_t bit = (INTEL_INFO(dev)->gen >= 7) ? DE_PIPE_VBLANK_IVB(pipe) :
 						     DE_PIPE_VBLANK(pipe);
@@ -2808,7 +2784,7 @@ static void ironlake_disable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static void valleyview_disable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2819,7 +2795,7 @@ static void valleyview_disable_vblank(struct drm_device *dev, unsigned int pipe)
 
 static void gen8_disable_vblank(struct drm_device *dev, unsigned int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
@@ -2835,9 +2811,9 @@ ring_idle(struct intel_engine_cs *engine, u32 seqno)
 }
 
 static bool
-ipehr_is_semaphore_wait(struct drm_i915_private *dev_priv, u32 ipehr)
+ipehr_is_semaphore_wait(struct intel_engine_cs *engine, u32 ipehr)
 {
-	if (INTEL_GEN(dev_priv) >= 8) {
+	if (INTEL_GEN(engine->i915) >= 8) {
 		return (ipehr >> 23) == 0x1c;
 	} else {
 		ipehr &= ~MI_SEMAPHORE_SYNC_MASK;
@@ -2908,7 +2884,7 @@ semaphore_waits_for(struct intel_engine_cs *engine, u32 *seqno)
 		return NULL;
 
 	ipehr = I915_READ(RING_IPEHR(engine->mmio_base));
-	if (!ipehr_is_semaphore_wait(engine->i915, ipehr))
+	if (!ipehr_is_semaphore_wait(engine, ipehr))
 		return NULL;
 
 	/*
@@ -2966,7 +2942,7 @@ static int semaphore_passed(struct intel_engine_cs *engine)
 	if (signaller->hangcheck.deadlock >= I915_NUM_ENGINES)
 		return -1;
 
-	if (i915_seqno_passed(signaller->get_seqno(signaller), seqno))
+	if (i915_seqno_passed(intel_engine_get_seqno(signaller), seqno))
 		return 1;
 
 	/* cursory check for an unkickable deadlock */
@@ -3078,23 +3054,21 @@ ring_stuck(struct intel_engine_cs *engine, u64 acthd)
 	return HANGCHECK_HUNG;
 }
 
-static unsigned kick_waiters(struct intel_engine_cs *engine)
+static unsigned long kick_waiters(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *i915 = engine->i915;
-	unsigned user_interrupts = READ_ONCE(engine->user_interrupts);
+	unsigned long irq_count = READ_ONCE(engine->breadcrumbs.irq_wakeups);
 
-	if (engine->hangcheck.user_interrupts == user_interrupts &&
+	if (engine->hangcheck.user_interrupts == irq_count &&
 	    !test_and_set_bit(engine->id, &i915->gpu_error.missed_irq_rings)) {
-		if (!(i915->gpu_error.test_irq_rings & intel_engine_flag(engine)))
+		if (!test_bit(engine->id, &i915->gpu_error.test_irq_rings))
 			DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
 				  engine->name);
-		else
-			DRM_INFO("Fake missed irq on %s\n",
-				 engine->name);
-		wake_up_all(&engine->irq_queue);
+
+		intel_engine_enable_fake_irq(engine);
 	}
 
-	return user_interrupts;
+	return irq_count;
 }
 /*
  * This is called when the chip hasn't reported back with completed
@@ -3110,9 +3084,8 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 		container_of(work, typeof(*dev_priv),
 			     gpu_error.hangcheck_work.work);
 	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	int busy_count = 0, rings_hung = 0;
-	bool stuck[I915_NUM_ENGINES] = { 0 };
+	unsigned int hung = 0, stuck = 0;
+	int busy_count = 0;
 #define BUSY 1
 #define KICK 5
 #define HUNG 20
@@ -3121,12 +3094,8 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	if (!i915.enable_hangcheck)
 		return;
 
-	/*
-	 * The hangcheck work is synced during runtime suspend, we don't
-	 * require a wakeref. TODO: instead of disabling the asserts make
-	 * sure that we hold a reference when this work is running.
-	 */
-	DISABLE_RPM_WAKEREF_ASSERTS(dev_priv);
+	if (!READ_ONCE(dev_priv->gt.awake))
+		return;
 
 	/* As enabling the GPU requires fairly extensive mmio access,
 	 * periodically arm the mmio checker to see if we are triggering
@@ -3134,11 +3103,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	 */
 	intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
 
-	for_each_engine_id(engine, dev_priv, id) {
+	for_each_engine(engine, dev_priv) {
+		bool busy = intel_engine_has_waiter(engine);
 		u64 acthd;
 		u32 seqno;
 		unsigned user_interrupts;
-		bool busy = true;
 
 		semaphore_clear_deadlocks(dev_priv);
 
@@ -3153,7 +3122,7 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 			engine->irq_seqno_barrier(engine);
 
 		acthd = intel_ring_get_active_head(engine);
-		seqno = engine->get_seqno(engine);
+		seqno = intel_engine_get_seqno(engine);
 
 		/* Reset stuck interrupts between batch advances */
 		user_interrupts = 0;
@@ -3161,12 +3130,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 		if (engine->hangcheck.seqno == seqno) {
 			if (ring_idle(engine, seqno)) {
 				engine->hangcheck.action = HANGCHECK_IDLE;
-				if (waitqueue_active(&engine->irq_queue)) {
+				if (busy) {
 					/* Safeguard against driver failure */
 					user_interrupts = kick_waiters(engine);
 					engine->hangcheck.score += BUSY;
-				} else
-					busy = false;
+				}
 			} else {
 				/* We always increment the hangcheck score
 				 * if the ring is busy and still processing
@@ -3198,10 +3166,15 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 					break;
 				case HANGCHECK_HUNG:
 					engine->hangcheck.score += HUNG;
-					stuck[id] = true;
 					break;
 				}
 			}
+
+			if (engine->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG) {
+				hung |= intel_engine_flag(engine);
+				if (engine->hangcheck.action != HANGCHECK_HUNG)
+					stuck |= intel_engine_flag(engine);
+			}
 		} else {
 			engine->hangcheck.action = HANGCHECK_ACTIVE;
 
@@ -3226,48 +3199,33 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 		busy_count += busy;
 	}
 
-	for_each_engine_id(engine, dev_priv, id) {
-		if (engine->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG) {
-			DRM_INFO("%s on %s\n",
-				 stuck[id] ? "stuck" : "no progress",
-				 engine->name);
-			rings_hung |= intel_engine_flag(engine);
-		}
-	}
+	if (hung) {
+		char msg[80];
+		int len;
 
-	if (rings_hung) {
-		i915_handle_error(dev_priv, rings_hung, "Engine(s) hung");
-		goto out;
+		/* If some rings hung but others were still busy, only
+		 * blame the hanging rings in the synopsis.
+		 */
+		if (stuck != hung)
+			hung &= ~stuck;
+		len = scnprintf(msg, sizeof(msg),
+				"%s on ", stuck == hung ? "No progress" : "Hang");
+		for_each_engine_masked(engine, dev_priv, hung)
+			len += scnprintf(msg + len, sizeof(msg) - len,
+					 "%s, ", engine->name);
+		msg[len-2] = '\0';
+
+		return i915_handle_error(dev_priv, hung, msg);
 	}
 
+	/* Reset timer in case GPU hangs without another request being added */
 	if (busy_count)
-		/* Reset timer case chip hangs without another request
-		 * being added */
 		i915_queue_hangcheck(dev_priv);
-
-out:
-	ENABLE_RPM_WAKEREF_ASSERTS(dev_priv);
-}
-
-void i915_queue_hangcheck(struct drm_i915_private *dev_priv)
-{
-	struct i915_gpu_error *e = &dev_priv->gpu_error;
-
-	if (!i915.enable_hangcheck)
-		return;
-
-	/* Don't continually defer the hangcheck so that it is always run at
-	 * least once after work has been scheduled on any ring. Otherwise,
-	 * we will ignore a hung ring if a second ring is kept busy.
-	 */
-
-	queue_delayed_work(e->hangcheck_wq, &e->hangcheck_work,
-			   round_jiffies_up_relative(DRM_I915_HANGCHECK_JIFFIES));
 }
 
 static void ibx_irq_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (HAS_PCH_NOP(dev))
 		return;
@@ -3288,7 +3246,7 @@ static void ibx_irq_reset(struct drm_device *dev)
  */
 static void ibx_irq_pre_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (HAS_PCH_NOP(dev))
 		return;
@@ -3300,7 +3258,7 @@ static void ibx_irq_pre_postinstall(struct drm_device *dev)
 
 static void gen5_gt_irq_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	GEN5_IRQ_RESET(GT);
 	if (INTEL_INFO(dev)->gen >= 6)
@@ -3360,7 +3318,7 @@ static void vlv_display_irq_postinstall(struct drm_i915_private *dev_priv)
 */
 static void ironlake_irq_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(HWSTAM, 0xffffffff);
 
@@ -3375,7 +3333,7 @@ static void ironlake_irq_reset(struct drm_device *dev)
 
 static void valleyview_irq_preinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(VLV_MASTER_IER, 0);
 	POSTING_READ(VLV_MASTER_IER);
@@ -3398,7 +3356,7 @@ static void gen8_gt_irq_reset(struct drm_i915_private *dev_priv)
 
 static void gen8_irq_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	I915_WRITE(GEN8_MASTER_IRQ, 0);
@@ -3444,12 +3402,12 @@ void gen8_irq_power_well_pre_disable(struct drm_i915_private *dev_priv,
 	spin_unlock_irq(&dev_priv->irq_lock);
 
 	/* make sure we're done processing display irqs */
-	synchronize_irq(dev_priv->dev->irq);
+	synchronize_irq(dev_priv->drm.irq);
 }
 
 static void cherryview_irq_preinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(GEN8_MASTER_IRQ, 0);
 	POSTING_READ(GEN8_MASTER_IRQ);
@@ -3470,7 +3428,7 @@ static u32 intel_hpd_enabled_irqs(struct drm_i915_private *dev_priv,
 	struct intel_encoder *encoder;
 	u32 enabled_irqs = 0;
 
-	for_each_intel_encoder(dev_priv->dev, encoder)
+	for_each_intel_encoder(&dev_priv->drm, encoder)
 		if (dev_priv->hotplug.stats[encoder->hpd_pin].state == HPD_ENABLED)
 			enabled_irqs |= hpd[encoder->hpd_pin];
 
@@ -3601,7 +3559,7 @@ static void bxt_hpd_irq_setup(struct drm_i915_private *dev_priv)
 
 static void ibx_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 mask;
 
 	if (HAS_PCH_NOP(dev))
@@ -3618,7 +3576,7 @@ static void ibx_irq_postinstall(struct drm_device *dev)
 
 static void gen5_gt_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pm_irqs, gt_irqs;
 
 	pm_irqs = gt_irqs = 0;
@@ -3632,8 +3590,7 @@ static void gen5_gt_irq_postinstall(struct drm_device *dev)
 
 	gt_irqs |= GT_RENDER_USER_INTERRUPT;
 	if (IS_GEN5(dev)) {
-		gt_irqs |= GT_RENDER_PIPECTL_NOTIFY_INTERRUPT |
-			   ILK_BSD_USER_INTERRUPT;
+		gt_irqs |= ILK_BSD_USER_INTERRUPT;
 	} else {
 		gt_irqs |= GT_BLT_USER_INTERRUPT | GT_BSD_USER_INTERRUPT;
 	}
@@ -3655,7 +3612,7 @@ static void gen5_gt_irq_postinstall(struct drm_device *dev)
 
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 display_mask, extra_mask;
 
 	if (INTEL_INFO(dev)->gen >= 7) {
@@ -3734,7 +3691,7 @@ void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv)
 
 static int valleyview_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	gen5_gt_irq_postinstall(dev);
 
@@ -3827,7 +3784,7 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 
 static int gen8_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (HAS_PCH_SPLIT(dev))
 		ibx_irq_pre_postinstall(dev);
@@ -3846,7 +3803,7 @@ static int gen8_irq_postinstall(struct drm_device *dev)
 
 static int cherryview_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	gen8_gt_irq_postinstall(dev_priv);
 
@@ -3863,7 +3820,7 @@ static int cherryview_irq_postinstall(struct drm_device *dev)
 
 static void gen8_irq_uninstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!dev_priv)
 		return;
@@ -3873,7 +3830,7 @@ static void gen8_irq_uninstall(struct drm_device *dev)
 
 static void valleyview_irq_uninstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!dev_priv)
 		return;
@@ -3893,7 +3850,7 @@ static void valleyview_irq_uninstall(struct drm_device *dev)
 
 static void cherryview_irq_uninstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!dev_priv)
 		return;
@@ -3913,7 +3870,7 @@ static void cherryview_irq_uninstall(struct drm_device *dev)
 
 static void ironlake_irq_uninstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!dev_priv)
 		return;
@@ -3923,7 +3880,7 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 
 static void i8xx_irq_preinstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	for_each_pipe(dev_priv, pipe)
@@ -3935,7 +3892,7 @@ static void i8xx_irq_preinstall(struct drm_device * dev)
 
 static int i8xx_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE16(EMR,
 		     ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
@@ -3998,7 +3955,7 @@ check_page_flip:
 static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u16 iir, new_iir;
 	u32 pipe_stats[2];
 	int pipe;
@@ -4075,7 +4032,7 @@ out:
 
 static void i8xx_irq_uninstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	for_each_pipe(dev_priv, pipe) {
@@ -4090,7 +4047,7 @@ static void i8xx_irq_uninstall(struct drm_device * dev)
 
 static void i915_irq_preinstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	if (I915_HAS_HOTPLUG(dev)) {
@@ -4108,7 +4065,7 @@ static void i915_irq_preinstall(struct drm_device * dev)
 
 static int i915_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 enable_mask;
 
 	I915_WRITE(EMR, ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
@@ -4187,7 +4144,7 @@ check_page_flip:
 static irqreturn_t i915_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 iir, new_iir, pipe_stats[I915_MAX_PIPES];
 	u32 flip_mask =
 		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
@@ -4292,7 +4249,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 
 static void i915_irq_uninstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	if (I915_HAS_HOTPLUG(dev)) {
@@ -4314,7 +4271,7 @@ static void i915_irq_uninstall(struct drm_device * dev)
 
 static void i965_irq_preinstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
@@ -4330,7 +4287,7 @@ static void i965_irq_preinstall(struct drm_device * dev)
 
 static int i965_irq_postinstall(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 enable_mask;
 	u32 error_mask;
 
@@ -4414,7 +4371,7 @@ static void i915_hpd_irq_setup(struct drm_i915_private *dev_priv)
 static irqreturn_t i965_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 iir, new_iir;
 	u32 pipe_stats[I915_MAX_PIPES];
 	int ret = IRQ_NONE, pipe;
@@ -4523,7 +4480,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 
 static void i965_irq_uninstall(struct drm_device * dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 
 	if (!dev_priv)
@@ -4553,7 +4510,7 @@ static void i965_irq_uninstall(struct drm_device * dev)
  */
 void intel_irq_init(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	intel_hpd_init_work(dev_priv);
 
@@ -4631,7 +4588,7 @@ void intel_irq_init(struct drm_i915_private *dev_priv)
 		dev->driver->disable_vblank = gen8_disable_vblank;
 		if (IS_BROXTON(dev))
 			dev_priv->display.hpd_irq_setup = bxt_hpd_irq_setup;
-		else if (HAS_PCH_SPT(dev))
+		else if (HAS_PCH_SPT(dev) || HAS_PCH_KBP(dev))
 			dev_priv->display.hpd_irq_setup = spt_hpd_irq_setup;
 		else
 			dev_priv->display.hpd_irq_setup = ilk_hpd_irq_setup;
@@ -4687,7 +4644,7 @@ int intel_irq_install(struct drm_i915_private *dev_priv)
 	 */
 	dev_priv->pm.irqs_enabled = true;
 
-	return drm_irq_install(dev_priv->dev, dev_priv->dev->pdev->irq);
+	return drm_irq_install(&dev_priv->drm, dev_priv->drm.pdev->irq);
 }
 
 /**
@@ -4699,7 +4656,7 @@ int intel_irq_install(struct drm_i915_private *dev_priv)
  */
 void intel_irq_uninstall(struct drm_i915_private *dev_priv)
 {
-	drm_irq_uninstall(dev_priv->dev);
+	drm_irq_uninstall(&dev_priv->drm);
 	intel_hpd_cancel_work(dev_priv);
 	dev_priv->pm.irqs_enabled = false;
 }
@@ -4713,9 +4670,9 @@ void intel_irq_uninstall(struct drm_i915_private *dev_priv)
  */
 void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv)
 {
-	dev_priv->dev->driver->irq_uninstall(dev_priv->dev);
+	dev_priv->drm.driver->irq_uninstall(&dev_priv->drm);
 	dev_priv->pm.irqs_enabled = false;
-	synchronize_irq(dev_priv->dev->irq);
+	synchronize_irq(dev_priv->drm.irq);
 }
 
 /**
@@ -4728,6 +4685,6 @@ void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv)
 void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv)
 {
 	dev_priv->pm.irqs_enabled = true;
-	dev_priv->dev->driver->irq_preinstall(dev_priv->dev);
-	dev_priv->dev->driver->irq_postinstall(dev_priv->dev);
+	dev_priv->drm.driver->irq_preinstall(&dev_priv->drm);
+	dev_priv->drm.driver->irq_postinstall(&dev_priv->drm);
 }
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index 7effe68d552c..8b13bfa47fba 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -224,6 +224,6 @@ module_param_named(enable_dpcd_backlight, i915.enable_dpcd_backlight, bool, 0600
 MODULE_PARM_DESC(enable_dpcd_backlight,
 	"Enable support for DPCD backlight control (default:false)");
 
-module_param_named(enable_gvt, i915.enable_gvt, bool, 0600);
+module_param_named(enable_gvt, i915.enable_gvt, bool, 0400);
 MODULE_PARM_DESC(enable_gvt,
 	"Enable support for Intel GVT-g graphics virtualization host support(default:false)");
diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
new file mode 100644
index 000000000000..949c01686a66
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pci.c
@@ -0,0 +1,503 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/console.h>
+#include <linux/vgaarb.h>
+#include <linux/vga_switcheroo.h>
+
+#include "i915_drv.h"
+
+#define GEN_DEFAULT_PIPEOFFSETS \
+	.pipe_offsets = { PIPE_A_OFFSET, PIPE_B_OFFSET, \
+			  PIPE_C_OFFSET, PIPE_EDP_OFFSET }, \
+	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
+			   TRANSCODER_C_OFFSET, TRANSCODER_EDP_OFFSET }, \
+	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET }
+
+#define GEN_CHV_PIPEOFFSETS \
+	.pipe_offsets = { PIPE_A_OFFSET, PIPE_B_OFFSET, \
+			  CHV_PIPE_C_OFFSET }, \
+	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
+			   CHV_TRANSCODER_C_OFFSET, }, \
+	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET, \
+			     CHV_PALETTE_C_OFFSET }
+
+#define CURSOR_OFFSETS \
+	.cursor_offsets = { CURSOR_A_OFFSET, CURSOR_B_OFFSET, CHV_CURSOR_C_OFFSET }
+
+#define IVB_CURSOR_OFFSETS \
+	.cursor_offsets = { CURSOR_A_OFFSET, IVB_CURSOR_B_OFFSET, IVB_CURSOR_C_OFFSET }
+
+#define BDW_COLORS \
+	.color = { .degamma_lut_size = 512, .gamma_lut_size = 512 }
+#define CHV_COLORS \
+	.color = { .degamma_lut_size = 65, .gamma_lut_size = 257 }
+
+static const struct intel_device_info intel_i830_info = {
+	.gen = 2, .is_mobile = 1, .cursor_needs_physical = 1, .num_pipes = 2,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_845g_info = {
+	.gen = 2, .num_pipes = 1,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_i85x_info = {
+	.gen = 2, .is_i85x = 1, .is_mobile = 1, .num_pipes = 2,
+	.cursor_needs_physical = 1,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_i865g_info = {
+	.gen = 2, .num_pipes = 1,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_i915g_info = {
+	.gen = 3, .is_i915g = 1, .cursor_needs_physical = 1, .num_pipes = 2,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+static const struct intel_device_info intel_i915gm_info = {
+	.gen = 3, .is_mobile = 1, .num_pipes = 2,
+	.cursor_needs_physical = 1,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.supports_tv = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+static const struct intel_device_info intel_i945g_info = {
+	.gen = 3, .has_hotplug = 1, .cursor_needs_physical = 1, .num_pipes = 2,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+static const struct intel_device_info intel_i945gm_info = {
+	.gen = 3, .is_i945gm = 1, .is_mobile = 1, .num_pipes = 2,
+	.has_hotplug = 1, .cursor_needs_physical = 1,
+	.has_overlay = 1, .overlay_needs_physical = 1,
+	.supports_tv = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_i965g_info = {
+	.gen = 4, .is_broadwater = 1, .num_pipes = 2,
+	.has_hotplug = 1,
+	.has_overlay = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_i965gm_info = {
+	.gen = 4, .is_crestline = 1, .num_pipes = 2,
+	.is_mobile = 1, .has_fbc = 1, .has_hotplug = 1,
+	.has_overlay = 1,
+	.supports_tv = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_g33_info = {
+	.gen = 3, .is_g33 = 1, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.has_overlay = 1,
+	.ring_mask = RENDER_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_g45_info = {
+	.gen = 4, .is_g4x = 1, .need_gfx_hws = 1, .num_pipes = 2,
+	.has_pipe_cxsr = 1, .has_hotplug = 1,
+	.ring_mask = RENDER_RING | BSD_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_gm45_info = {
+	.gen = 4, .is_g4x = 1, .num_pipes = 2,
+	.is_mobile = 1, .need_gfx_hws = 1, .has_fbc = 1,
+	.has_pipe_cxsr = 1, .has_hotplug = 1,
+	.supports_tv = 1,
+	.ring_mask = RENDER_RING | BSD_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_pineview_info = {
+	.gen = 3, .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.has_overlay = 1,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_ironlake_d_info = {
+	.gen = 5, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.ring_mask = RENDER_RING | BSD_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_ironlake_m_info = {
+	.gen = 5, .is_mobile = 1, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING | BSD_RING,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_sandybridge_d_info = {
+	.gen = 6, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING,
+	.has_llc = 1,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+static const struct intel_device_info intel_sandybridge_m_info = {
+	.gen = 6, .is_mobile = 1, .num_pipes = 2,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.has_fbc = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING,
+	.has_llc = 1,
+	GEN_DEFAULT_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+};
+
+#define GEN7_FEATURES  \
+	.gen = 7, .num_pipes = 3, \
+	.need_gfx_hws = 1, .has_hotplug = 1, \
+	.has_fbc = 1, \
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING, \
+	.has_llc = 1, \
+	GEN_DEFAULT_PIPEOFFSETS, \
+	IVB_CURSOR_OFFSETS
+
+static const struct intel_device_info intel_ivybridge_d_info = {
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
+};
+
+static const struct intel_device_info intel_ivybridge_m_info = {
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
+	.is_mobile = 1,
+};
+
+static const struct intel_device_info intel_ivybridge_q_info = {
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
+	.num_pipes = 0, /* legal, last one wins */
+};
+
+#define VLV_FEATURES  \
+	.gen = 7, .num_pipes = 2, \
+	.need_gfx_hws = 1, .has_hotplug = 1, \
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING, \
+	.display_mmio_offset = VLV_DISPLAY_BASE, \
+	GEN_DEFAULT_PIPEOFFSETS, \
+	CURSOR_OFFSETS
+
+static const struct intel_device_info intel_valleyview_m_info = {
+	VLV_FEATURES,
+	.is_valleyview = 1,
+	.is_mobile = 1,
+};
+
+static const struct intel_device_info intel_valleyview_d_info = {
+	VLV_FEATURES,
+	.is_valleyview = 1,
+};
+
+#define HSW_FEATURES  \
+	GEN7_FEATURES, \
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING, \
+	.has_ddi = 1, \
+	.has_fpga_dbg = 1
+
+static const struct intel_device_info intel_haswell_d_info = {
+	HSW_FEATURES,
+	.is_haswell = 1,
+};
+
+static const struct intel_device_info intel_haswell_m_info = {
+	HSW_FEATURES,
+	.is_haswell = 1,
+	.is_mobile = 1,
+};
+
+#define BDW_FEATURES \
+	HSW_FEATURES, \
+	BDW_COLORS
+
+static const struct intel_device_info intel_broadwell_d_info = {
+	BDW_FEATURES,
+	.gen = 8,
+	.is_broadwell = 1,
+};
+
+static const struct intel_device_info intel_broadwell_m_info = {
+	BDW_FEATURES,
+	.gen = 8, .is_mobile = 1,
+	.is_broadwell = 1,
+};
+
+static const struct intel_device_info intel_broadwell_gt3d_info = {
+	BDW_FEATURES,
+	.gen = 8,
+	.is_broadwell = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
+};
+
+static const struct intel_device_info intel_broadwell_gt3m_info = {
+	BDW_FEATURES,
+	.gen = 8, .is_mobile = 1,
+	.is_broadwell = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
+};
+
+static const struct intel_device_info intel_cherryview_info = {
+	.gen = 8, .num_pipes = 3,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
+	.is_cherryview = 1,
+	.display_mmio_offset = VLV_DISPLAY_BASE,
+	GEN_CHV_PIPEOFFSETS,
+	CURSOR_OFFSETS,
+	CHV_COLORS,
+};
+
+static const struct intel_device_info intel_skylake_info = {
+	BDW_FEATURES,
+	.is_skylake = 1,
+	.gen = 9,
+};
+
+static const struct intel_device_info intel_skylake_gt3_info = {
+	BDW_FEATURES,
+	.is_skylake = 1,
+	.gen = 9,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
+};
+
+static const struct intel_device_info intel_broxton_info = {
+	.is_broxton = 1,
+	.gen = 9,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
+	.num_pipes = 3,
+	.has_ddi = 1,
+	.has_fpga_dbg = 1,
+	.has_fbc = 1,
+	.has_pooled_eu = 0,
+	GEN_DEFAULT_PIPEOFFSETS,
+	IVB_CURSOR_OFFSETS,
+	BDW_COLORS,
+};
+
+static const struct intel_device_info intel_kabylake_info = {
+	BDW_FEATURES,
+	.is_kabylake = 1,
+	.gen = 9,
+};
+
+static const struct intel_device_info intel_kabylake_gt3_info = {
+	BDW_FEATURES,
+	.is_kabylake = 1,
+	.gen = 9,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
+};
+
+/*
+ * Make sure any device matches here are from most specific to most
+ * general.  For example, since the Quanta match is based on the subsystem
+ * and subvendor IDs, we need it to come before the more general IVB
+ * PCI ID matches, otherwise we'll use the wrong info struct above.
+ */
+static const struct pci_device_id pciidlist[] = {
+	INTEL_I830_IDS(&intel_i830_info),
+	INTEL_I845G_IDS(&intel_845g_info),
+	INTEL_I85X_IDS(&intel_i85x_info),
+	INTEL_I865G_IDS(&intel_i865g_info),
+	INTEL_I915G_IDS(&intel_i915g_info),
+	INTEL_I915GM_IDS(&intel_i915gm_info),
+	INTEL_I945G_IDS(&intel_i945g_info),
+	INTEL_I945GM_IDS(&intel_i945gm_info),
+	INTEL_I965G_IDS(&intel_i965g_info),
+	INTEL_G33_IDS(&intel_g33_info),
+	INTEL_I965GM_IDS(&intel_i965gm_info),
+	INTEL_GM45_IDS(&intel_gm45_info),
+	INTEL_G45_IDS(&intel_g45_info),
+	INTEL_PINEVIEW_IDS(&intel_pineview_info),
+	INTEL_IRONLAKE_D_IDS(&intel_ironlake_d_info),
+	INTEL_IRONLAKE_M_IDS(&intel_ironlake_m_info),
+	INTEL_SNB_D_IDS(&intel_sandybridge_d_info),
+	INTEL_SNB_M_IDS(&intel_sandybridge_m_info),
+	INTEL_IVB_Q_IDS(&intel_ivybridge_q_info), /* must be first IVB */
+	INTEL_IVB_M_IDS(&intel_ivybridge_m_info),
+	INTEL_IVB_D_IDS(&intel_ivybridge_d_info),
+	INTEL_HSW_D_IDS(&intel_haswell_d_info),
+	INTEL_HSW_M_IDS(&intel_haswell_m_info),
+	INTEL_VLV_M_IDS(&intel_valleyview_m_info),
+	INTEL_VLV_D_IDS(&intel_valleyview_d_info),
+	INTEL_BDW_GT12M_IDS(&intel_broadwell_m_info),
+	INTEL_BDW_GT12D_IDS(&intel_broadwell_d_info),
+	INTEL_BDW_GT3M_IDS(&intel_broadwell_gt3m_info),
+	INTEL_BDW_GT3D_IDS(&intel_broadwell_gt3d_info),
+	INTEL_CHV_IDS(&intel_cherryview_info),
+	INTEL_SKL_GT1_IDS(&intel_skylake_info),
+	INTEL_SKL_GT2_IDS(&intel_skylake_info),
+	INTEL_SKL_GT3_IDS(&intel_skylake_gt3_info),
+	INTEL_SKL_GT4_IDS(&intel_skylake_gt3_info),
+	INTEL_BXT_IDS(&intel_broxton_info),
+	INTEL_KBL_GT1_IDS(&intel_kabylake_info),
+	INTEL_KBL_GT2_IDS(&intel_kabylake_info),
+	INTEL_KBL_GT3_IDS(&intel_kabylake_gt3_info),
+	INTEL_KBL_GT4_IDS(&intel_kabylake_gt3_info),
+	{0, 0, 0}
+};
+MODULE_DEVICE_TABLE(pci, pciidlist);
+
+extern int i915_driver_load(struct pci_dev *pdev,
+			    const struct pci_device_id *ent);
+
+static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	struct intel_device_info *intel_info =
+		(struct intel_device_info *) ent->driver_data;
+
+	if (IS_PRELIMINARY_HW(intel_info) && !i915.preliminary_hw_support) {
+		DRM_INFO("This hardware requires preliminary hardware support.\n"
+			 "See CONFIG_DRM_I915_PRELIMINARY_HW_SUPPORT, and/or modparam preliminary_hw_support\n");
+		return -ENODEV;
+	}
+
+	/* Only bind to function 0 of the device. Early generations
+	 * used function 1 as a placeholder for multi-head. This causes
+	 * us confusion instead, especially on the systems where both
+	 * functions have the same PCI-ID!
+	 */
+	if (PCI_FUNC(pdev->devfn))
+		return -ENODEV;
+
+	/*
+	 * apple-gmux is needed on dual GPU MacBook Pro
+	 * to probe the panel if we're the inactive GPU.
+	 */
+	if (vga_switcheroo_client_probe_defer(pdev))
+		return -EPROBE_DEFER;
+
+	return i915_driver_load(pdev, ent);
+}
+
+extern void i915_driver_unload(struct drm_device *dev);
+
+static void i915_pci_remove(struct pci_dev *pdev)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+
+	i915_driver_unload(dev);
+	drm_dev_unref(dev);
+}
+
+extern const struct dev_pm_ops i915_pm_ops;
+
+static struct pci_driver i915_pci_driver = {
+	.name = DRIVER_NAME,
+	.id_table = pciidlist,
+	.probe = i915_pci_probe,
+	.remove = i915_pci_remove,
+	.driver.pm = &i915_pm_ops,
+};
+
+static int __init i915_init(void)
+{
+	bool use_kms = true;
+
+	/*
+	 * Enable KMS by default, unless explicitly overriden by
+	 * either the i915.modeset prarameter or by the
+	 * vga_text_mode_force boot option.
+	 */
+
+	if (i915.modeset == 0)
+		use_kms = false;
+
+	if (vgacon_text_force() && i915.modeset == -1)
+		use_kms = false;
+
+	if (!use_kms) {
+		/* Silently fail loading to not upset userspace. */
+		DRM_DEBUG_DRIVER("KMS disabled.\n");
+		return 0;
+	}
+
+	return pci_register_driver(&i915_pci_driver);
+}
+
+static void __exit i915_exit(void)
+{
+	if (!i915_pci_driver.driver.owner)
+		return;
+
+	pci_unregister_driver(&i915_pci_driver);
+}
+
+module_init(i915_init);
+module_exit(i915_exit);
+
+MODULE_AUTHOR("Tungsten Graphics, Inc.");
+MODULE_AUTHOR("Intel Corporation");
+
+MODULE_DESCRIPTION(DRIVER_DESC);
+MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index c6bfbf8d7cca..8bfde75789f6 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -7070,7 +7070,8 @@ enum {
 #define GEN6_RPDEUC				_MMIO(0xA084)
 #define GEN6_RPDEUCSW				_MMIO(0xA088)
 #define GEN6_RC_STATE				_MMIO(0xA094)
-#define   RC6_STATE				(1 << 18)
+#define   RC_SW_TARGET_STATE_SHIFT		16
+#define   RC_SW_TARGET_STATE_MASK		(7 << RC_SW_TARGET_STATE_SHIFT)
 #define GEN6_RC1_WAKE_RATE_LIMIT		_MMIO(0xA098)
 #define GEN6_RC6_WAKE_RATE_LIMIT		_MMIO(0xA09C)
 #define GEN6_RC6pp_WAKE_RATE_LIMIT		_MMIO(0xA0A0)
@@ -7085,12 +7086,16 @@ enum {
 #define GEN6_RC6pp_THRESHOLD			_MMIO(0xA0C0)
 #define GEN6_PMINTRMSK				_MMIO(0xA168)
 #define   GEN8_PMINTR_REDIRECT_TO_NON_DISP	(1<<31)
+#define GEN8_MISC_CTRL0				_MMIO(0xA180)
 #define VLV_PWRDWNUPCTL				_MMIO(0xA294)
 #define GEN9_MEDIA_PG_IDLE_HYSTERESIS		_MMIO(0xA0C4)
 #define GEN9_RENDER_PG_IDLE_HYSTERESIS		_MMIO(0xA0C8)
 #define GEN9_PG_ENABLE				_MMIO(0xA210)
 #define GEN9_RENDER_PG_ENABLE			(1<<0)
 #define GEN9_MEDIA_PG_ENABLE			(1<<1)
+#define GEN8_PUSHBUS_CONTROL			_MMIO(0xA248)
+#define GEN8_PUSHBUS_ENABLE			_MMIO(0xA250)
+#define GEN8_PUSHBUS_SHIFT			_MMIO(0xA25C)
 
 #define VLV_CHICKEN_3				_MMIO(VLV_DISPLAY_BASE + 0x7040C)
 #define  PIXEL_OVERLAP_CNT_MASK			(3 << 30)
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 34e061a9ef06..5cfe4c7716b4 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -31,7 +31,7 @@
 
 static void i915_save_display(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Display arbitration control */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -63,7 +63,7 @@ static void i915_save_display(struct drm_device *dev)
 
 static void i915_restore_display(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 mask = 0xffffffff;
 
 	/* Display arbitration */
@@ -103,7 +103,7 @@ static void i915_restore_display(struct drm_device *dev)
 
 int i915_save_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	mutex_lock(&dev->struct_mutex);
@@ -148,7 +148,7 @@ int i915_save_state(struct drm_device *dev)
 
 int i915_restore_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	mutex_lock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index 02507bfc8def..d61829e54f93 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -38,7 +38,7 @@
 static u32 calc_residency(struct drm_device *dev,
 			  i915_reg_t reg)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u64 raw_time; /* 32b value may overflow during fixed point math */
 	u64 units = 128ULL, div = 100000ULL;
 	u32 ret;
@@ -166,7 +166,7 @@ i915_l3_read(struct file *filp, struct kobject *kobj,
 	struct device *dev = kobj_to_dev(kobj);
 	struct drm_minor *dminor = dev_to_drm_minor(dev);
 	struct drm_device *drm_dev = dminor->dev;
-	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(drm_dev);
 	int slice = (int)(uintptr_t)attr->private;
 	int ret;
 
@@ -202,7 +202,7 @@ i915_l3_write(struct file *filp, struct kobject *kobj,
 	struct device *dev = kobj_to_dev(kobj);
 	struct drm_minor *dminor = dev_to_drm_minor(dev);
 	struct drm_device *drm_dev = dminor->dev;
-	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(drm_dev);
 	struct i915_gem_context *ctx;
 	u32 *temp = NULL; /* Just here to make handling failures easy */
 	int slice = (int)(uintptr_t)attr->private;
@@ -227,13 +227,6 @@ i915_l3_write(struct file *filp, struct kobject *kobj,
 		}
 	}
 
-	ret = i915_gpu_idle(drm_dev);
-	if (ret) {
-		kfree(temp);
-		mutex_unlock(&drm_dev->struct_mutex);
-		return ret;
-	}
-
 	/* TODO: Ideally we really want a GPU reset here to make sure errors
 	 * aren't propagated. Since I cannot find a stable way to reset the GPU
 	 * at this point it is left as a TODO.
@@ -275,7 +268,7 @@ static ssize_t gt_act_freq_mhz_show(struct device *kdev,
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -309,7 +302,7 @@ static ssize_t gt_cur_freq_mhz_show(struct device *kdev,
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -330,7 +323,7 @@ static ssize_t vlv_rpe_freq_mhz_show(struct device *kdev,
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	return snprintf(buf, PAGE_SIZE,
 			"%d\n",
@@ -341,7 +334,7 @@ static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -359,7 +352,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val;
 	ssize_t ret;
 
@@ -409,7 +402,7 @@ static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -427,7 +420,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev,
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val;
 	ssize_t ret;
 
@@ -487,7 +480,7 @@ static ssize_t gt_rp_mhz_show(struct device *kdev, struct device_attribute *attr
 {
 	struct drm_minor *minor = dev_to_drm_minor(kdev);
 	struct drm_device *dev = minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val;
 
 	if (attr == &dev_attr_gt_RP0_freq_mhz)
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index 6768db032f84..534154e05fbe 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -118,7 +118,7 @@ TRACE_EVENT(i915_gem_shrink,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = i915->dev->primary->index;
+			   __entry->dev = i915->drm.primary->index;
 			   __entry->target = target;
 			   __entry->flags = flags;
 			   ),
@@ -462,7 +462,7 @@ TRACE_EVENT(i915_gem_ring_sync_to,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = from->i915->dev->primary->index;
+			   __entry->dev = from->i915->drm.primary->index;
 			   __entry->sync_from = from->id;
 			   __entry->sync_to = to_req->engine->id;
 			   __entry->seqno = i915_gem_request_get_seqno(req);
@@ -486,11 +486,11 @@ TRACE_EVENT(i915_gem_ring_dispatch,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = req->i915->dev->primary->index;
+			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
 			   __entry->seqno = req->seqno;
 			   __entry->flags = flags;
-			   i915_trace_irq_get(req->engine, req);
+			   intel_engine_enable_signaling(req);
 			   ),
 
 	    TP_printk("dev=%u, ring=%u, seqno=%u, flags=%x",
@@ -509,7 +509,7 @@ TRACE_EVENT(i915_gem_ring_flush,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = req->i915->dev->primary->index;
+			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
 			   __entry->invalidate = invalidate;
 			   __entry->flush = flush;
@@ -531,7 +531,7 @@ DECLARE_EVENT_CLASS(i915_gem_request,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = req->i915->dev->primary->index;
+			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
 			   __entry->seqno = req->seqno;
 			   ),
@@ -556,9 +556,9 @@ TRACE_EVENT(i915_gem_request_notify,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->dev = engine->i915->dev->primary->index;
+			   __entry->dev = engine->i915->drm.primary->index;
 			   __entry->ring = engine->id;
-			   __entry->seqno = engine->get_seqno(engine);
+			   __entry->seqno = intel_engine_get_seqno(engine);
 			   ),
 
 	    TP_printk("dev=%u, ring=%u, seqno=%u",
@@ -593,11 +593,11 @@ TRACE_EVENT(i915_gem_request_wait_begin,
 	     * less desirable.
 	     */
 	    TP_fast_assign(
-			   __entry->dev = req->i915->dev->primary->index;
+			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
 			   __entry->seqno = req->seqno;
 			   __entry->blocking =
-				     mutex_is_locked(&req->i915->dev->struct_mutex);
+				     mutex_is_locked(&req->i915->drm.struct_mutex);
 			   ),
 
 	    TP_printk("dev=%u, ring=%u, seqno=%u, blocking=%s",
@@ -746,7 +746,7 @@ DECLARE_EVENT_CLASS(i915_context,
 	TP_fast_assign(
 			__entry->ctx = ctx;
 			__entry->vm = ctx->ppgtt ? &ctx->ppgtt->base : NULL;
-			__entry->dev = ctx->i915->dev->primary->index;
+			__entry->dev = ctx->i915->drm.primary->index;
 	),
 
 	TP_printk("dev=%u, ctx=%p, ctx_vm=%p",
@@ -786,7 +786,7 @@ TRACE_EVENT(switch_mm,
 			__entry->ring = engine->id;
 			__entry->to = to;
 			__entry->vm = to->ppgtt? &to->ppgtt->base : NULL;
-			__entry->dev = engine->i915->dev->primary->index;
+			__entry->dev = engine->i915->drm.primary->index;
 	),
 
 	TP_printk("dev=%u, ring=%u, ctx=%p, ctx_vm=%p",
diff --git a/drivers/gpu/drm/i915/intel_audio.c b/drivers/gpu/drm/i915/intel_audio.c
index b9329c2a670a..6700a7be7f78 100644
--- a/drivers/gpu/drm/i915/intel_audio.c
+++ b/drivers/gpu/drm/i915/intel_audio.c
@@ -154,7 +154,7 @@ static bool audio_rate_need_prog(struct intel_crtc *crtc,
 {
 	if (((mode->clock == TMDS_297M) ||
 		 (mode->clock == TMDS_296M)) &&
-		intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI))
+		intel_crtc_has_type(crtc->config, INTEL_OUTPUT_HDMI))
 		return true;
 	else
 		return false;
@@ -165,7 +165,7 @@ static bool intel_eld_uptodate(struct drm_connector *connector,
 			       i915_reg_t reg_elda, uint32_t bits_elda,
 			       i915_reg_t reg_edid)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	uint8_t *eld = connector->eld;
 	uint32_t tmp;
 	int i;
@@ -189,7 +189,7 @@ static bool intel_eld_uptodate(struct drm_connector *connector,
 
 static void g4x_audio_codec_disable(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	uint32_t eldv, tmp;
 
 	DRM_DEBUG_KMS("Disable audio codec\n");
@@ -210,7 +210,7 @@ static void g4x_audio_codec_enable(struct drm_connector *connector,
 				   struct intel_encoder *encoder,
 				   const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	uint8_t *eld = connector->eld;
 	uint32_t eldv;
 	uint32_t tmp;
@@ -247,7 +247,7 @@ static void g4x_audio_codec_enable(struct drm_connector *connector,
 
 static void hsw_audio_codec_disable(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	uint32_t tmp;
@@ -262,7 +262,7 @@ static void hsw_audio_codec_disable(struct intel_encoder *encoder)
 	tmp |= AUD_CONFIG_N_PROG_ENABLE;
 	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
 	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
-	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		tmp |= AUD_CONFIG_N_VALUE_INDEX;
 	I915_WRITE(HSW_AUD_CFG(pipe), tmp);
 
@@ -279,7 +279,7 @@ static void hsw_audio_codec_enable(struct drm_connector *connector,
 				   struct intel_encoder *encoder,
 				   const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	struct i915_audio_component *acomp = dev_priv->audio_component;
@@ -328,7 +328,7 @@ static void hsw_audio_codec_enable(struct drm_connector *connector,
 	tmp = I915_READ(HSW_AUD_CFG(pipe));
 	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
 	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
-	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		tmp |= AUD_CONFIG_N_VALUE_INDEX;
 	else
 		tmp |= audio_config_hdmi_pixel_clock(adjusted_mode);
@@ -357,7 +357,7 @@ static void hsw_audio_codec_enable(struct drm_connector *connector,
 
 static void ilk_audio_codec_disable(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_digital_port *intel_dig_port =
 		enc_to_dig_port(&encoder->base);
@@ -389,7 +389,7 @@ static void ilk_audio_codec_disable(struct intel_encoder *encoder)
 	tmp |= AUD_CONFIG_N_PROG_ENABLE;
 	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
 	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
-	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		tmp |= AUD_CONFIG_N_VALUE_INDEX;
 	I915_WRITE(aud_config, tmp);
 
@@ -405,7 +405,7 @@ static void ilk_audio_codec_enable(struct drm_connector *connector,
 				   struct intel_encoder *encoder,
 				   const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_digital_port *intel_dig_port =
 		enc_to_dig_port(&encoder->base);
@@ -475,7 +475,7 @@ static void ilk_audio_codec_enable(struct drm_connector *connector,
 	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
 	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
 	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
-	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		tmp |= AUD_CONFIG_N_VALUE_INDEX;
 	else
 		tmp |= audio_config_hdmi_pixel_clock(adjusted_mode);
@@ -496,7 +496,7 @@ void intel_audio_codec_enable(struct intel_encoder *intel_encoder)
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
 	struct drm_connector *connector;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_audio_component *acomp = dev_priv->audio_component;
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	enum port port = intel_dig_port->port;
@@ -513,7 +513,7 @@ void intel_audio_codec_enable(struct intel_encoder *intel_encoder)
 
 	/* ELD Conn_Type */
 	connector->eld[5] &= ~(3 << 2);
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_crtc_has_dp_encoder(crtc->config))
 		connector->eld[5] |= (1 << 2);
 
 	connector->eld[6] = drm_av_sync_delay(connector, adjusted_mode) / 2;
@@ -543,7 +543,7 @@ void intel_audio_codec_disable(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_audio_component *acomp = dev_priv->audio_component;
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	enum port port = intel_dig_port->port;
@@ -749,14 +749,14 @@ static int i915_audio_component_bind(struct device *i915_dev,
 	if (WARN_ON(acomp->ops || acomp->dev))
 		return -EEXIST;
 
-	drm_modeset_lock_all(dev_priv->dev);
+	drm_modeset_lock_all(&dev_priv->drm);
 	acomp->ops = &i915_audio_component_ops;
 	acomp->dev = i915_dev;
 	BUILD_BUG_ON(MAX_PORTS != I915_MAX_PORTS);
 	for (i = 0; i < ARRAY_SIZE(acomp->aud_sample_rate); i++)
 		acomp->aud_sample_rate[i] = 0;
 	dev_priv->audio_component = acomp;
-	drm_modeset_unlock_all(dev_priv->dev);
+	drm_modeset_unlock_all(&dev_priv->drm);
 
 	return 0;
 }
@@ -767,11 +767,11 @@ static void i915_audio_component_unbind(struct device *i915_dev,
 	struct i915_audio_component *acomp = data;
 	struct drm_i915_private *dev_priv = dev_to_i915(i915_dev);
 
-	drm_modeset_lock_all(dev_priv->dev);
+	drm_modeset_lock_all(&dev_priv->drm);
 	acomp->ops = NULL;
 	acomp->dev = NULL;
 	dev_priv->audio_component = NULL;
-	drm_modeset_unlock_all(dev_priv->dev);
+	drm_modeset_unlock_all(&dev_priv->drm);
 }
 
 static const struct component_ops i915_audio_component_bind_ops = {
@@ -799,7 +799,7 @@ void i915_audio_component_init(struct drm_i915_private *dev_priv)
 {
 	int ret;
 
-	ret = component_add(dev_priv->dev->dev, &i915_audio_component_bind_ops);
+	ret = component_add(dev_priv->drm.dev, &i915_audio_component_bind_ops);
 	if (ret < 0) {
 		DRM_ERROR("failed to add audio component (%d)\n", ret);
 		/* continue with reduced functionality */
@@ -821,6 +821,6 @@ void i915_audio_component_cleanup(struct drm_i915_private *dev_priv)
 	if (!dev_priv->audio_component_registered)
 		return;
 
-	component_del(dev_priv->dev->dev, &i915_audio_component_bind_ops);
+	component_del(dev_priv->drm.dev, &i915_audio_component_bind_ops);
 	dev_priv->audio_component_registered = false;
 }
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index da5ed4a850b9..c6e69e4cfa83 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -1426,7 +1426,7 @@ static const struct vbt_header *find_vbt(void __iomem *bios, size_t size)
 int
 intel_bios_init(struct drm_i915_private *dev_priv)
 {
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	const struct vbt_header *vbt = dev_priv->opregion.vbt;
 	const struct bdb_header *bdb;
 	u8 __iomem *bios = NULL;
diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c
new file mode 100644
index 000000000000..d89b2c963618
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c
@@ -0,0 +1,586 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/kthread.h>
+
+#include "i915_drv.h"
+
+static void intel_breadcrumbs_fake_irq(unsigned long data)
+{
+	struct intel_engine_cs *engine = (struct intel_engine_cs *)data;
+
+	/*
+	 * The timer persists in case we cannot enable interrupts,
+	 * or if we have previously seen seqno/interrupt incoherency
+	 * ("missed interrupt" syndrome). Here the worker will wake up
+	 * every jiffie in order to kick the oldest waiter to do the
+	 * coherent seqno check.
+	 */
+	rcu_read_lock();
+	if (intel_engine_wakeup(engine))
+		mod_timer(&engine->breadcrumbs.fake_irq, jiffies + 1);
+	rcu_read_unlock();
+}
+
+static void irq_enable(struct intel_engine_cs *engine)
+{
+	/* Enabling the IRQ may miss the generation of the interrupt, but
+	 * we still need to force the barrier before reading the seqno,
+	 * just in case.
+	 */
+	engine->breadcrumbs.irq_posted = true;
+
+	spin_lock_irq(&engine->i915->irq_lock);
+	engine->irq_enable(engine);
+	spin_unlock_irq(&engine->i915->irq_lock);
+}
+
+static void irq_disable(struct intel_engine_cs *engine)
+{
+	spin_lock_irq(&engine->i915->irq_lock);
+	engine->irq_disable(engine);
+	spin_unlock_irq(&engine->i915->irq_lock);
+
+	engine->breadcrumbs.irq_posted = false;
+}
+
+static void __intel_breadcrumbs_enable_irq(struct intel_breadcrumbs *b)
+{
+	struct intel_engine_cs *engine =
+		container_of(b, struct intel_engine_cs, breadcrumbs);
+	struct drm_i915_private *i915 = engine->i915;
+
+	assert_spin_locked(&b->lock);
+	if (b->rpm_wakelock)
+		return;
+
+	/* Since we are waiting on a request, the GPU should be busy
+	 * and should have its own rpm reference. For completeness,
+	 * record an rpm reference for ourselves to cover the
+	 * interrupt we unmask.
+	 */
+	intel_runtime_pm_get_noresume(i915);
+	b->rpm_wakelock = true;
+
+	/* No interrupts? Kick the waiter every jiffie! */
+	if (intel_irqs_enabled(i915)) {
+		if (!test_bit(engine->id, &i915->gpu_error.test_irq_rings))
+			irq_enable(engine);
+		b->irq_enabled = true;
+	}
+
+	if (!b->irq_enabled ||
+	    test_bit(engine->id, &i915->gpu_error.missed_irq_rings))
+		mod_timer(&b->fake_irq, jiffies + 1);
+}
+
+static void __intel_breadcrumbs_disable_irq(struct intel_breadcrumbs *b)
+{
+	struct intel_engine_cs *engine =
+		container_of(b, struct intel_engine_cs, breadcrumbs);
+
+	assert_spin_locked(&b->lock);
+	if (!b->rpm_wakelock)
+		return;
+
+	if (b->irq_enabled) {
+		irq_disable(engine);
+		b->irq_enabled = false;
+	}
+
+	intel_runtime_pm_put(engine->i915);
+	b->rpm_wakelock = false;
+}
+
+static inline struct intel_wait *to_wait(struct rb_node *node)
+{
+	return container_of(node, struct intel_wait, node);
+}
+
+static inline void __intel_breadcrumbs_finish(struct intel_breadcrumbs *b,
+					      struct intel_wait *wait)
+{
+	assert_spin_locked(&b->lock);
+
+	/* This request is completed, so remove it from the tree, mark it as
+	 * complete, and *then* wake up the associated task.
+	 */
+	rb_erase(&wait->node, &b->waiters);
+	RB_CLEAR_NODE(&wait->node);
+
+	wake_up_process(wait->tsk); /* implicit smp_wmb() */
+}
+
+static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
+				    struct intel_wait *wait)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct rb_node **p, *parent, *completed;
+	bool first;
+	u32 seqno;
+
+	/* Insert the request into the retirement ordered list
+	 * of waiters by walking the rbtree. If we are the oldest
+	 * seqno in the tree (the first to be retired), then
+	 * set ourselves as the bottom-half.
+	 *
+	 * As we descend the tree, prune completed branches since we hold the
+	 * spinlock we know that the first_waiter must be delayed and can
+	 * reduce some of the sequential wake up latency if we take action
+	 * ourselves and wake up the completed tasks in parallel. Also, by
+	 * removing stale elements in the tree, we may be able to reduce the
+	 * ping-pong between the old bottom-half and ourselves as first-waiter.
+	 */
+	first = true;
+	parent = NULL;
+	completed = NULL;
+	seqno = intel_engine_get_seqno(engine);
+
+	 /* If the request completed before we managed to grab the spinlock,
+	  * return now before adding ourselves to the rbtree. We let the
+	  * current bottom-half handle any pending wakeups and instead
+	  * try and get out of the way quickly.
+	  */
+	if (i915_seqno_passed(seqno, wait->seqno)) {
+		RB_CLEAR_NODE(&wait->node);
+		return first;
+	}
+
+	p = &b->waiters.rb_node;
+	while (*p) {
+		parent = *p;
+		if (wait->seqno == to_wait(parent)->seqno) {
+			/* We have multiple waiters on the same seqno, select
+			 * the highest priority task (that with the smallest
+			 * task->prio) to serve as the bottom-half for this
+			 * group.
+			 */
+			if (wait->tsk->prio > to_wait(parent)->tsk->prio) {
+				p = &parent->rb_right;
+				first = false;
+			} else {
+				p = &parent->rb_left;
+			}
+		} else if (i915_seqno_passed(wait->seqno,
+					     to_wait(parent)->seqno)) {
+			p = &parent->rb_right;
+			if (i915_seqno_passed(seqno, to_wait(parent)->seqno))
+				completed = parent;
+			else
+				first = false;
+		} else {
+			p = &parent->rb_left;
+		}
+	}
+	rb_link_node(&wait->node, parent, p);
+	rb_insert_color(&wait->node, &b->waiters);
+	GEM_BUG_ON(!first && !b->irq_seqno_bh);
+
+	if (completed) {
+		struct rb_node *next = rb_next(completed);
+
+		GEM_BUG_ON(!next && !first);
+		if (next && next != &wait->node) {
+			GEM_BUG_ON(first);
+			b->first_wait = to_wait(next);
+			smp_store_mb(b->irq_seqno_bh, b->first_wait->tsk);
+			/* As there is a delay between reading the current
+			 * seqno, processing the completed tasks and selecting
+			 * the next waiter, we may have missed the interrupt
+			 * and so need for the next bottom-half to wakeup.
+			 *
+			 * Also as we enable the IRQ, we may miss the
+			 * interrupt for that seqno, so we have to wake up
+			 * the next bottom-half in order to do a coherent check
+			 * in case the seqno passed.
+			 */
+			__intel_breadcrumbs_enable_irq(b);
+			if (READ_ONCE(b->irq_posted))
+				wake_up_process(to_wait(next)->tsk);
+		}
+
+		do {
+			struct intel_wait *crumb = to_wait(completed);
+			completed = rb_prev(completed);
+			__intel_breadcrumbs_finish(b, crumb);
+		} while (completed);
+	}
+
+	if (first) {
+		GEM_BUG_ON(rb_first(&b->waiters) != &wait->node);
+		b->first_wait = wait;
+		smp_store_mb(b->irq_seqno_bh, wait->tsk);
+		/* After assigning ourselves as the new bottom-half, we must
+		 * perform a cursory check to prevent a missed interrupt.
+		 * Either we miss the interrupt whilst programming the hardware,
+		 * or if there was a previous waiter (for a later seqno) they
+		 * may be woken instead of us (due to the inherent race
+		 * in the unlocked read of b->irq_seqno_bh in the irq handler)
+		 * and so we miss the wake up.
+		 */
+		__intel_breadcrumbs_enable_irq(b);
+	}
+	GEM_BUG_ON(!b->irq_seqno_bh);
+	GEM_BUG_ON(!b->first_wait);
+	GEM_BUG_ON(rb_first(&b->waiters) != &b->first_wait->node);
+
+	return first;
+}
+
+bool intel_engine_add_wait(struct intel_engine_cs *engine,
+			   struct intel_wait *wait)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	bool first;
+
+	spin_lock(&b->lock);
+	first = __intel_engine_add_wait(engine, wait);
+	spin_unlock(&b->lock);
+
+	return first;
+}
+
+void intel_engine_enable_fake_irq(struct intel_engine_cs *engine)
+{
+	mod_timer(&engine->breadcrumbs.fake_irq, jiffies + 1);
+}
+
+static inline bool chain_wakeup(struct rb_node *rb, int priority)
+{
+	return rb && to_wait(rb)->tsk->prio <= priority;
+}
+
+static inline int wakeup_priority(struct intel_breadcrumbs *b,
+				  struct task_struct *tsk)
+{
+	if (tsk == b->signaler)
+		return INT_MIN;
+	else
+		return tsk->prio;
+}
+
+void intel_engine_remove_wait(struct intel_engine_cs *engine,
+			      struct intel_wait *wait)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	/* Quick check to see if this waiter was already decoupled from
+	 * the tree by the bottom-half to avoid contention on the spinlock
+	 * by the herd.
+	 */
+	if (RB_EMPTY_NODE(&wait->node))
+		return;
+
+	spin_lock(&b->lock);
+
+	if (RB_EMPTY_NODE(&wait->node))
+		goto out_unlock;
+
+	if (b->first_wait == wait) {
+		const int priority = wakeup_priority(b, wait->tsk);
+		struct rb_node *next;
+
+		GEM_BUG_ON(b->irq_seqno_bh != wait->tsk);
+
+		/* We are the current bottom-half. Find the next candidate,
+		 * the first waiter in the queue on the remaining oldest
+		 * request. As multiple seqnos may complete in the time it
+		 * takes us to wake up and find the next waiter, we have to
+		 * wake up that waiter for it to perform its own coherent
+		 * completion check.
+		 */
+		next = rb_next(&wait->node);
+		if (chain_wakeup(next, priority)) {
+			/* If the next waiter is already complete,
+			 * wake it up and continue onto the next waiter. So
+			 * if have a small herd, they will wake up in parallel
+			 * rather than sequentially, which should reduce
+			 * the overall latency in waking all the completed
+			 * clients.
+			 *
+			 * However, waking up a chain adds extra latency to
+			 * the first_waiter. This is undesirable if that
+			 * waiter is a high priority task.
+			 */
+			u32 seqno = intel_engine_get_seqno(engine);
+
+			while (i915_seqno_passed(seqno, to_wait(next)->seqno)) {
+				struct rb_node *n = rb_next(next);
+
+				__intel_breadcrumbs_finish(b, to_wait(next));
+				next = n;
+				if (!chain_wakeup(next, priority))
+					break;
+			}
+		}
+
+		if (next) {
+			/* In our haste, we may have completed the first waiter
+			 * before we enabled the interrupt. Do so now as we
+			 * have a second waiter for a future seqno. Afterwards,
+			 * we have to wake up that waiter in case we missed
+			 * the interrupt, or if we have to handle an
+			 * exception rather than a seqno completion.
+			 */
+			b->first_wait = to_wait(next);
+			smp_store_mb(b->irq_seqno_bh, b->first_wait->tsk);
+			if (b->first_wait->seqno != wait->seqno)
+				__intel_breadcrumbs_enable_irq(b);
+			wake_up_process(b->irq_seqno_bh);
+		} else {
+			b->first_wait = NULL;
+			WRITE_ONCE(b->irq_seqno_bh, NULL);
+			__intel_breadcrumbs_disable_irq(b);
+		}
+	} else {
+		GEM_BUG_ON(rb_first(&b->waiters) == &wait->node);
+	}
+
+	GEM_BUG_ON(RB_EMPTY_NODE(&wait->node));
+	rb_erase(&wait->node, &b->waiters);
+
+out_unlock:
+	GEM_BUG_ON(b->first_wait == wait);
+	GEM_BUG_ON(rb_first(&b->waiters) !=
+		   (b->first_wait ? &b->first_wait->node : NULL));
+	GEM_BUG_ON(!b->irq_seqno_bh ^ RB_EMPTY_ROOT(&b->waiters));
+	spin_unlock(&b->lock);
+}
+
+static bool signal_complete(struct drm_i915_gem_request *request)
+{
+	if (!request)
+		return false;
+
+	/* If another process served as the bottom-half it may have already
+	 * signalled that this wait is already completed.
+	 */
+	if (intel_wait_complete(&request->signaling.wait))
+		return true;
+
+	/* Carefully check if the request is complete, giving time for the
+	 * seqno to be visible or if the GPU hung.
+	 */
+	if (__i915_request_irq_complete(request))
+		return true;
+
+	return false;
+}
+
+static struct drm_i915_gem_request *to_signaler(struct rb_node *rb)
+{
+	return container_of(rb, struct drm_i915_gem_request, signaling.node);
+}
+
+static void signaler_set_rtpriority(void)
+{
+	 struct sched_param param = { .sched_priority = 1 };
+
+	 sched_setscheduler_nocheck(current, SCHED_FIFO, &param);
+}
+
+static int intel_breadcrumbs_signaler(void *arg)
+{
+	struct intel_engine_cs *engine = arg;
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct drm_i915_gem_request *request;
+
+	/* Install ourselves with high priority to reduce signalling latency */
+	signaler_set_rtpriority();
+
+	do {
+		set_current_state(TASK_INTERRUPTIBLE);
+
+		/* We are either woken up by the interrupt bottom-half,
+		 * or by a client adding a new signaller. In both cases,
+		 * the GPU seqno may have advanced beyond our oldest signal.
+		 * If it has, propagate the signal, remove the waiter and
+		 * check again with the next oldest signal. Otherwise we
+		 * need to wait for a new interrupt from the GPU or for
+		 * a new client.
+		 */
+		request = READ_ONCE(b->first_signal);
+		if (signal_complete(request)) {
+			/* Wake up all other completed waiters and select the
+			 * next bottom-half for the next user interrupt.
+			 */
+			intel_engine_remove_wait(engine,
+						 &request->signaling.wait);
+
+			/* Find the next oldest signal. Note that as we have
+			 * not been holding the lock, another client may
+			 * have installed an even older signal than the one
+			 * we just completed - so double check we are still
+			 * the oldest before picking the next one.
+			 */
+			spin_lock(&b->lock);
+			if (request == b->first_signal) {
+				struct rb_node *rb =
+					rb_next(&request->signaling.node);
+				b->first_signal = rb ? to_signaler(rb) : NULL;
+			}
+			rb_erase(&request->signaling.node, &b->signals);
+			spin_unlock(&b->lock);
+
+			i915_gem_request_unreference(request);
+		} else {
+			if (kthread_should_stop())
+				break;
+
+			schedule();
+		}
+	} while (1);
+	__set_current_state(TASK_RUNNING);
+
+	return 0;
+}
+
+void intel_engine_enable_signaling(struct drm_i915_gem_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct rb_node *parent, **p;
+	bool first, wakeup;
+
+	if (unlikely(READ_ONCE(request->signaling.wait.tsk)))
+		return;
+
+	spin_lock(&b->lock);
+	if (unlikely(request->signaling.wait.tsk)) {
+		wakeup = false;
+		goto unlock;
+	}
+
+	request->signaling.wait.tsk = b->signaler;
+	request->signaling.wait.seqno = request->seqno;
+	i915_gem_request_reference(request);
+
+	/* First add ourselves into the list of waiters, but register our
+	 * bottom-half as the signaller thread. As per usual, only the oldest
+	 * waiter (not just signaller) is tasked as the bottom-half waking
+	 * up all completed waiters after the user interrupt.
+	 *
+	 * If we are the oldest waiter, enable the irq (after which we
+	 * must double check that the seqno did not complete).
+	 */
+	wakeup = __intel_engine_add_wait(engine, &request->signaling.wait);
+
+	/* Now insert ourselves into the retirement ordered list of signals
+	 * on this engine. We track the oldest seqno as that will be the
+	 * first signal to complete.
+	 */
+	parent = NULL;
+	first = true;
+	p = &b->signals.rb_node;
+	while (*p) {
+		parent = *p;
+		if (i915_seqno_passed(request->seqno,
+				      to_signaler(parent)->seqno)) {
+			p = &parent->rb_right;
+			first = false;
+		} else {
+			p = &parent->rb_left;
+		}
+	}
+	rb_link_node(&request->signaling.node, parent, p);
+	rb_insert_color(&request->signaling.node, &b->signals);
+	if (first)
+		smp_store_mb(b->first_signal, request);
+
+unlock:
+	spin_unlock(&b->lock);
+
+	if (wakeup)
+		wake_up_process(b->signaler);
+}
+
+int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct task_struct *tsk;
+
+	spin_lock_init(&b->lock);
+	setup_timer(&b->fake_irq,
+		    intel_breadcrumbs_fake_irq,
+		    (unsigned long)engine);
+
+	/* Spawn a thread to provide a common bottom-half for all signals.
+	 * As this is an asynchronous interface we cannot steal the current
+	 * task for handling the bottom-half to the user interrupt, therefore
+	 * we create a thread to do the coherent seqno dance after the
+	 * interrupt and then signal the waitqueue (via the dma-buf/fence).
+	 */
+	tsk = kthread_run(intel_breadcrumbs_signaler, engine,
+			  "i915/signal:%d", engine->id);
+	if (IS_ERR(tsk))
+		return PTR_ERR(tsk);
+
+	b->signaler = tsk;
+
+	return 0;
+}
+
+void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	if (!IS_ERR_OR_NULL(b->signaler))
+		kthread_stop(b->signaler);
+
+	del_timer_sync(&b->fake_irq);
+}
+
+unsigned int intel_kick_waiters(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	unsigned int mask = 0;
+
+	/* To avoid the task_struct disappearing beneath us as we wake up
+	 * the process, we must first inspect the task_struct->state under the
+	 * RCU lock, i.e. as we call wake_up_process() we must be holding the
+	 * rcu_read_lock().
+	 */
+	rcu_read_lock();
+	for_each_engine(engine, i915)
+		if (unlikely(intel_engine_wakeup(engine)))
+			mask |= intel_engine_flag(engine);
+	rcu_read_unlock();
+
+	return mask;
+}
+
+unsigned int intel_kick_signalers(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	unsigned int mask = 0;
+
+	for_each_engine(engine, i915) {
+		if (unlikely(READ_ONCE(engine->breadcrumbs.first_signal))) {
+			wake_up_process(engine->breadcrumbs.signaler);
+			mask |= intel_engine_flag(engine);
+		}
+	}
+
+	return mask;
+}
diff --git a/drivers/gpu/drm/i915/intel_color.c b/drivers/gpu/drm/i915/intel_color.c
index 522f5a2de015..bc0fef3d3335 100644
--- a/drivers/gpu/drm/i915/intel_color.c
+++ b/drivers/gpu/drm/i915/intel_color.c
@@ -96,7 +96,7 @@ static void i9xx_load_csc_matrix(struct drm_crtc_state *crtc_state)
 {
 	struct drm_crtc *crtc = crtc_state->crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int i, pipe = intel_crtc->pipe;
 	uint16_t coeffs[9] = { 0, };
@@ -207,7 +207,7 @@ static void cherryview_load_csc_matrix(struct drm_crtc_state *state)
 {
 	struct drm_crtc *crtc = state->crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = to_intel_crtc(crtc)->pipe;
 	uint32_t mode;
 
@@ -255,7 +255,7 @@ static void cherryview_load_csc_matrix(struct drm_crtc_state *state)
 void intel_color_set_csc(struct drm_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc_state->crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (dev_priv->display.load_csc_matrix)
 		dev_priv->display.load_csc_matrix(crtc_state);
@@ -266,13 +266,13 @@ static void i9xx_load_luts_internal(struct drm_crtc *crtc,
 				    struct drm_property_blob *blob)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	int i;
 
 	if (HAS_GMCH_DISPLAY(dev)) {
-		if (intel_crtc->config->has_dsi_encoder)
+		if (intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_DSI))
 			assert_dsi_pll_enabled(dev_priv);
 		else
 			assert_pll_enabled(dev_priv, pipe);
@@ -313,7 +313,7 @@ static void haswell_load_luts(struct drm_crtc_state *crtc_state)
 {
 	struct drm_crtc *crtc = crtc_state->crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_crtc_state *intel_crtc_state =
 		to_intel_crtc_state(crtc_state);
@@ -343,7 +343,7 @@ static void broadwell_load_luts(struct drm_crtc_state *state)
 {
 	struct drm_crtc *crtc = state->crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_state *intel_state = to_intel_crtc_state(state);
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 	uint32_t i, lut_size = INTEL_INFO(dev)->color.degamma_lut_size;
@@ -426,7 +426,7 @@ static void cherryview_load_luts(struct drm_crtc_state *state)
 {
 	struct drm_crtc *crtc = state->crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 	struct drm_color_lut *lut;
 	uint32_t i, lut_size;
@@ -485,7 +485,7 @@ static void cherryview_load_luts(struct drm_crtc_state *state)
 void intel_color_load_luts(struct drm_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc_state->crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->display.load_luts(crtc_state);
 }
@@ -526,7 +526,7 @@ int intel_color_check(struct drm_crtc *crtc,
 void intel_color_init(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	drm_mode_crtc_set_gamma_size(crtc, 256);
 
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index e115bcc6766f..5819d524d917 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -67,7 +67,7 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 				   enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_encoder_to_crt(encoder);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
@@ -98,7 +98,7 @@ out:
 
 static unsigned int intel_crt_get_flags(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_crt *crt = intel_encoder_to_crt(encoder);
 	u32 tmp, flags = 0;
 
@@ -146,7 +146,7 @@ static void hsw_crt_get_config(struct intel_encoder *encoder,
 static void intel_crt_set_dpms(struct intel_encoder *encoder, int mode)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_encoder_to_crt(encoder);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
@@ -281,7 +281,7 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 	struct intel_crt *crt = intel_attached_crt(connector);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 adpa;
 	bool ret;
 
@@ -301,8 +301,10 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector)
 
 		I915_WRITE(crt->adpa_reg, adpa);
 
-		if (wait_for((I915_READ(crt->adpa_reg) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
-			     1000))
+		if (intel_wait_for_register(dev_priv,
+					    crt->adpa_reg,
+					    ADPA_CRT_HOTPLUG_FORCE_TRIGGER, 0,
+					    1000))
 			DRM_DEBUG_KMS("timed out waiting for FORCE_TRIGGER");
 
 		if (turn_off_dac) {
@@ -326,7 +328,7 @@ static bool valleyview_crt_detect_hotplug(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 	struct intel_crt *crt = intel_attached_crt(connector);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 adpa;
 	bool ret;
 	u32 save_adpa;
@@ -338,8 +340,10 @@ static bool valleyview_crt_detect_hotplug(struct drm_connector *connector)
 
 	I915_WRITE(crt->adpa_reg, adpa);
 
-	if (wait_for((I915_READ(crt->adpa_reg) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
-		     1000)) {
+	if (intel_wait_for_register(dev_priv,
+				    crt->adpa_reg,
+				    ADPA_CRT_HOTPLUG_FORCE_TRIGGER, 0,
+				    1000)) {
 		DRM_DEBUG_KMS("timed out waiting for FORCE_TRIGGER");
 		I915_WRITE(crt->adpa_reg, save_adpa);
 	}
@@ -367,7 +371,7 @@ static bool valleyview_crt_detect_hotplug(struct drm_connector *connector)
 static bool intel_crt_detect_hotplug(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 stat;
 	bool ret = false;
 	int i, tries = 0;
@@ -394,9 +398,9 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector)
 					      CRT_HOTPLUG_FORCE_DETECT,
 					      CRT_HOTPLUG_FORCE_DETECT);
 		/* wait for FORCE_DETECT to go off */
-		if (wait_for((I915_READ(PORT_HOTPLUG_EN) &
-			      CRT_HOTPLUG_FORCE_DETECT) == 0,
-			     1000))
+		if (intel_wait_for_register(dev_priv, PORT_HOTPLUG_EN,
+					    CRT_HOTPLUG_FORCE_DETECT, 0,
+					    1000))
 			DRM_DEBUG_KMS("timed out waiting for FORCE_DETECT to go off");
 	}
 
@@ -449,7 +453,7 @@ static int intel_crt_ddc_get_modes(struct drm_connector *connector,
 static bool intel_crt_detect_ddc(struct drm_connector *connector)
 {
 	struct intel_crt *crt = intel_attached_crt(connector);
-	struct drm_i915_private *dev_priv = crt->base.base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crt->base.base.dev);
 	struct edid *edid;
 	struct i2c_adapter *i2c;
 
@@ -485,7 +489,7 @@ static enum drm_connector_status
 intel_crt_load_detect(struct intel_crt *crt, uint32_t pipe)
 {
 	struct drm_device *dev = crt->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t save_bclrpat;
 	uint32_t save_vtotal;
 	uint32_t vtotal, vactive;
@@ -600,7 +604,7 @@ static enum drm_connector_status
 intel_crt_detect(struct drm_connector *connector, bool force)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_attached_crt(connector);
 	struct intel_encoder *intel_encoder = &crt->base;
 	enum intel_display_power_domain power_domain;
@@ -681,7 +685,7 @@ static void intel_crt_destroy(struct drm_connector *connector)
 static int intel_crt_get_modes(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_attached_crt(connector);
 	struct intel_encoder *intel_encoder = &crt->base;
 	enum intel_display_power_domain power_domain;
@@ -716,7 +720,7 @@ static int intel_crt_set_property(struct drm_connector *connector,
 static void intel_crt_reset(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_attached_crt(connector);
 
 	if (INTEL_INFO(dev)->gen >= 5) {
@@ -743,6 +747,7 @@ static const struct drm_connector_funcs intel_crt_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.detect = intel_crt_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_crt_destroy,
 	.set_property = intel_crt_set_property,
@@ -791,7 +796,7 @@ void intel_crt_init(struct drm_device *dev)
 	struct drm_connector *connector;
 	struct intel_crt *crt;
 	struct intel_connector *intel_connector;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t adpa_reg;
 	u32 adpa;
 
@@ -879,8 +884,6 @@ void intel_crt_init(struct drm_device *dev)
 
 	drm_connector_helper_add(connector, &intel_crt_connector_helper_funcs);
 
-	drm_connector_register(connector);
-
 	if (!I915_HAS_HOTPLUG(dev))
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 
diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c
index 2b3b428d9cd2..c3b33a10c15c 100644
--- a/drivers/gpu/drm/i915/intel_csr.c
+++ b/drivers/gpu/drm/i915/intel_csr.c
@@ -41,15 +41,15 @@
  * be moved to FW_FAILED.
  */
 
-#define I915_CSR_KBL "i915/kbl_dmc_ver1.bin"
+#define I915_CSR_KBL "i915/kbl_dmc_ver1_01.bin"
 MODULE_FIRMWARE(I915_CSR_KBL);
 #define KBL_CSR_VERSION_REQUIRED	CSR_VERSION(1, 1)
 
-#define I915_CSR_SKL "i915/skl_dmc_ver1.bin"
+#define I915_CSR_SKL "i915/skl_dmc_ver1_26.bin"
 MODULE_FIRMWARE(I915_CSR_SKL);
-#define SKL_CSR_VERSION_REQUIRED	CSR_VERSION(1, 23)
+#define SKL_CSR_VERSION_REQUIRED	CSR_VERSION(1, 26)
 
-#define I915_CSR_BXT "i915/bxt_dmc_ver1.bin"
+#define I915_CSR_BXT "i915/bxt_dmc_ver1_07.bin"
 MODULE_FIRMWARE(I915_CSR_BXT);
 #define BXT_CSR_VERSION_REQUIRED	CSR_VERSION(1, 7)
 
@@ -286,7 +286,7 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv,
 	uint32_t dmc_offset = CSR_DEFAULT_FW_OFFSET, readcount = 0, nbytes;
 	uint32_t i;
 	uint32_t *dmc_payload;
-	uint32_t required_min_version;
+	uint32_t required_version;
 
 	if (!fw)
 		return NULL;
@@ -303,24 +303,23 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv,
 	csr->version = css_header->version;
 
 	if (IS_KABYLAKE(dev_priv)) {
-		required_min_version = KBL_CSR_VERSION_REQUIRED;
+		required_version = KBL_CSR_VERSION_REQUIRED;
 	} else if (IS_SKYLAKE(dev_priv)) {
-		required_min_version = SKL_CSR_VERSION_REQUIRED;
+		required_version = SKL_CSR_VERSION_REQUIRED;
 	} else if (IS_BROXTON(dev_priv)) {
-		required_min_version = BXT_CSR_VERSION_REQUIRED;
+		required_version = BXT_CSR_VERSION_REQUIRED;
 	} else {
 		MISSING_CASE(INTEL_REVID(dev_priv));
-		required_min_version = 0;
+		required_version = 0;
 	}
 
-	if (csr->version < required_min_version) {
-		DRM_INFO("Refusing to load old DMC firmware v%u.%u,"
-			 " please upgrade to v%u.%u or later"
-			   " [" FIRMWARE_URL "].\n",
+	if (csr->version != required_version) {
+		DRM_INFO("Refusing to load DMC firmware v%u.%u,"
+			 " please use v%u.%u [" FIRMWARE_URL "].\n",
 			 CSR_VERSION_MAJOR(csr->version),
 			 CSR_VERSION_MINOR(csr->version),
-			 CSR_VERSION_MAJOR(required_min_version),
-			 CSR_VERSION_MINOR(required_min_version));
+			 CSR_VERSION_MAJOR(required_version),
+			 CSR_VERSION_MINOR(required_version));
 		return NULL;
 	}
 
@@ -413,7 +412,7 @@ static void csr_load_work_fn(struct work_struct *work)
 	csr = &dev_priv->csr;
 
 	ret = request_firmware(&fw, dev_priv->csr.fw_path,
-			       &dev_priv->dev->pdev->dev);
+			       &dev_priv->drm.pdev->dev);
 	if (fw)
 		dev_priv->csr.dmc_payload = parse_csr_fw(dev_priv, fw);
 
@@ -427,7 +426,7 @@ static void csr_load_work_fn(struct work_struct *work)
 			 CSR_VERSION_MAJOR(csr->version),
 			 CSR_VERSION_MINOR(csr->version));
 	} else {
-		dev_notice(dev_priv->dev->dev,
+		dev_notice(dev_priv->drm.dev,
 			   "Failed to load DMC firmware"
 			   " [" FIRMWARE_URL "],"
 			   " disabling runtime power management.\n");
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index ad3b0ee5e55b..dd1d6fe12297 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -318,7 +318,7 @@ static void ddi_get_encoder_port(struct intel_encoder *intel_encoder,
 	default:
 		WARN(1, "Invalid DDI encoder type %d\n", intel_encoder->type);
 		/* fallthrough and treat as unknown */
-	case INTEL_OUTPUT_DISPLAYPORT:
+	case INTEL_OUTPUT_DP:
 	case INTEL_OUTPUT_EDP:
 	case INTEL_OUTPUT_HDMI:
 	case INTEL_OUTPUT_UNKNOWN:
@@ -482,7 +482,7 @@ void intel_prepare_ddi_buffer(struct intel_encoder *encoder)
 		ddi_translations = ddi_translations_edp;
 		size = n_edp_entries;
 		break;
-	case INTEL_OUTPUT_DISPLAYPORT:
+	case INTEL_OUTPUT_DP:
 	case INTEL_OUTPUT_HDMI:
 		ddi_translations = ddi_translations_dp;
 		size = n_dp_entries;
@@ -543,7 +543,7 @@ static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv,
 void hsw_fdi_link_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	u32 temp, i, rx_ctl_val;
@@ -834,7 +834,7 @@ static void ddi_dotclock_get(struct intel_crtc_state *pipe_config)
 	if (pipe_config->has_pch_encoder)
 		dotclock = intel_dotclock_calculate(pipe_config->port_clock,
 						    &pipe_config->fdi_m_n);
-	else if (pipe_config->has_dp_encoder)
+	else if (intel_crtc_has_dp_encoder(pipe_config))
 		dotclock = intel_dotclock_calculate(pipe_config->port_clock,
 						    &pipe_config->dp_m_n);
 	else if (pipe_config->has_hdmi_sink && pipe_config->pipe_bpp == 36)
@@ -851,7 +851,7 @@ static void ddi_dotclock_get(struct intel_crtc_state *pipe_config)
 static void skl_ddi_clock_get(struct intel_encoder *encoder,
 				struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	int link_clock = 0;
 	uint32_t dpll_ctl1, dpll;
 
@@ -899,7 +899,7 @@ static void skl_ddi_clock_get(struct intel_encoder *encoder,
 static void hsw_ddi_clock_get(struct intel_encoder *encoder,
 			      struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	int link_clock = 0;
 	u32 val, pll;
 
@@ -971,7 +971,7 @@ static int bxt_calc_pll_link(struct drm_i915_private *dev_priv,
 static void bxt_ddi_clock_get(struct intel_encoder *encoder,
 				struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	enum port port = intel_ddi_get_encoder_port(encoder);
 	uint32_t dpll = port;
 
@@ -1061,14 +1061,14 @@ bool intel_ddi_pll_select(struct intel_crtc *intel_crtc,
 
 void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	int type = intel_encoder->type;
 	uint32_t temp;
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP || type == INTEL_OUTPUT_DP_MST) {
+	if (type == INTEL_OUTPUT_DP || type == INTEL_OUTPUT_EDP || type == INTEL_OUTPUT_DP_MST) {
 		WARN_ON(transcoder_is_dsi(cpu_transcoder));
 
 		temp = TRANS_MSA_SYNC_CLK;
@@ -1096,7 +1096,7 @@ void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	uint32_t temp;
 	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
@@ -1113,7 +1113,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
@@ -1182,7 +1182,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 		temp |= TRANS_DDI_MODE_SELECT_FDI;
 		temp |= (intel_crtc->config->fdi_lanes - 1) << 1;
 
-	} else if (type == INTEL_OUTPUT_DISPLAYPORT ||
+	} else if (type == INTEL_OUTPUT_DP ||
 		   type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
@@ -1223,7 +1223,7 @@ void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 {
 	struct drm_device *dev = intel_connector->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *intel_encoder = intel_connector->encoder;
 	int type = intel_connector->base.connector_type;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
@@ -1285,7 +1285,7 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 			    enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_ddi_get_encoder_port(encoder);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
@@ -1359,7 +1359,7 @@ void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
@@ -1371,7 +1371,7 @@ void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
 
 void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 
 	if (cpu_transcoder != TRANSCODER_EDP)
@@ -1392,7 +1392,7 @@ static void skl_ddi_set_iboost(struct drm_i915_private *dev_priv,
 	dp_iboost = dev_priv->vbt.ddi_port_info[port].dp_boost_level;
 	hdmi_iboost = dev_priv->vbt.ddi_port_info[port].hdmi_boost_level;
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT) {
+	if (type == INTEL_OUTPUT_DP) {
 		if (dp_iboost) {
 			iboost = dp_iboost;
 		} else {
@@ -1450,7 +1450,7 @@ static void bxt_ddi_vswing_sequence(struct drm_i915_private *dev_priv,
 	if (type == INTEL_OUTPUT_EDP && dev_priv->vbt.edp.low_vswing) {
 		n_entries = ARRAY_SIZE(bxt_ddi_translations_edp);
 		ddi_translations = bxt_ddi_translations_edp;
-	} else if (type == INTEL_OUTPUT_DISPLAYPORT
+	} else if (type == INTEL_OUTPUT_DP
 			|| type == INTEL_OUTPUT_EDP) {
 		n_entries = ARRAY_SIZE(bxt_ddi_translations_dp);
 		ddi_translations = bxt_ddi_translations_dp;
@@ -1624,7 +1624,7 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder)
 
 	intel_ddi_clk_select(intel_encoder, crtc->config);
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
+	if (type == INTEL_OUTPUT_DP || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
 		intel_dp_set_link_params(intel_dp, crtc->config);
@@ -1648,7 +1648,7 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 	uint32_t val;
@@ -1669,7 +1669,7 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
 	if (wait)
 		intel_wait_ddi_buf_idle(dev_priv, port);
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
+	if (type == INTEL_OUTPUT_DP || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF);
 		intel_edp_panel_vdd_on(intel_dp);
@@ -1695,7 +1695,7 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_crtc *crtc = encoder->crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 
@@ -1734,7 +1734,7 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int type = intel_encoder->type;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (intel_crtc->config->has_audio) {
 		intel_audio_codec_disable(intel_encoder);
@@ -1808,7 +1808,10 @@ static u32 bxt_get_grc(struct drm_i915_private *dev_priv, enum dpio_phy phy)
 static void bxt_phy_wait_grc_done(struct drm_i915_private *dev_priv,
 				  enum dpio_phy phy)
 {
-	if (wait_for(I915_READ(BXT_PORT_REF_DW3(phy)) & GRC_DONE, 10))
+	if (intel_wait_for_register(dev_priv,
+				    BXT_PORT_REF_DW3(phy),
+				    GRC_DONE, GRC_DONE,
+				    10))
 		DRM_ERROR("timeout waiting for PHY%d GRC\n", phy);
 }
 
@@ -2121,7 +2124,7 @@ void intel_ddi_prepare_link_retrain(struct intel_dp *intel_dp)
 
 void intel_ddi_fdi_disable(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	uint32_t val;
 
@@ -2154,7 +2157,7 @@ void intel_ddi_fdi_disable(struct drm_crtc *crtc)
 void intel_ddi_get_config(struct intel_encoder *encoder,
 			  struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
 	struct intel_hdmi *intel_hdmi;
@@ -2208,7 +2211,6 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 		break;
 	case TRANS_DDI_MODE_SELECT_DP_SST:
 	case TRANS_DDI_MODE_SELECT_DP_MST:
-		pipe_config->has_dp_encoder = true;
 		pipe_config->lane_count =
 			((temp & DDI_PORT_WIDTH_MASK) >> DDI_PORT_WIDTH_SHIFT) + 1;
 		intel_dp_get_m_n(intel_crtc, pipe_config);
@@ -2253,7 +2255,7 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 static bool intel_ddi_compute_config(struct intel_encoder *encoder,
 				     struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	int type = encoder->type;
 	int port = intel_ddi_get_encoder_port(encoder);
 	int ret;
@@ -2319,7 +2321,7 @@ intel_ddi_init_hdmi_connector(struct intel_digital_port *intel_dig_port)
 
 void intel_ddi_init(struct drm_device *dev, enum port port)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
 	struct drm_encoder *encoder;
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
new file mode 100644
index 000000000000..cba137f9ad3e
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -0,0 +1,388 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "i915_drv.h"
+
+void intel_device_info_dump(struct drm_i915_private *dev_priv)
+{
+	const struct intel_device_info *info = &dev_priv->info;
+
+#define PRINT_S(name) "%s"
+#define SEP_EMPTY
+#define PRINT_FLAG(name) info->name ? #name "," : ""
+#define SEP_COMMA ,
+	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x rev=0x%02x flags="
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_S, SEP_EMPTY),
+			 info->gen,
+			 dev_priv->drm.pdev->device,
+			 dev_priv->drm.pdev->revision,
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_COMMA));
+#undef PRINT_S
+#undef SEP_EMPTY
+#undef PRINT_FLAG
+#undef SEP_COMMA
+}
+
+static void cherryview_sseu_info_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *info = mkwrite_device_info(dev_priv);
+	u32 fuse, eu_dis;
+
+	fuse = I915_READ(CHV_FUSE_GT);
+
+	info->slice_total = 1;
+
+	if (!(fuse & CHV_FGT_DISABLE_SS0)) {
+		info->subslice_per_slice++;
+		eu_dis = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK |
+				 CHV_FGT_EU_DIS_SS0_R1_MASK);
+		info->eu_total += 8 - hweight32(eu_dis);
+	}
+
+	if (!(fuse & CHV_FGT_DISABLE_SS1)) {
+		info->subslice_per_slice++;
+		eu_dis = fuse & (CHV_FGT_EU_DIS_SS1_R0_MASK |
+				 CHV_FGT_EU_DIS_SS1_R1_MASK);
+		info->eu_total += 8 - hweight32(eu_dis);
+	}
+
+	info->subslice_total = info->subslice_per_slice;
+	/*
+	 * CHV expected to always have a uniform distribution of EU
+	 * across subslices.
+	*/
+	info->eu_per_subslice = info->subslice_total ?
+				info->eu_total / info->subslice_total :
+				0;
+	/*
+	 * CHV supports subslice power gating on devices with more than
+	 * one subslice, and supports EU power gating on devices with
+	 * more than one EU pair per subslice.
+	*/
+	info->has_slice_pg = 0;
+	info->has_subslice_pg = (info->subslice_total > 1);
+	info->has_eu_pg = (info->eu_per_subslice > 2);
+}
+
+static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *info = mkwrite_device_info(dev_priv);
+	int s_max = 3, ss_max = 4, eu_max = 8;
+	int s, ss;
+	u32 fuse2, s_enable, ss_disable, eu_disable;
+	u8 eu_mask = 0xff;
+
+	fuse2 = I915_READ(GEN8_FUSE2);
+	s_enable = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
+	ss_disable = (fuse2 & GEN9_F2_SS_DIS_MASK) >> GEN9_F2_SS_DIS_SHIFT;
+
+	info->slice_total = hweight32(s_enable);
+	/*
+	 * The subslice disable field is global, i.e. it applies
+	 * to each of the enabled slices.
+	*/
+	info->subslice_per_slice = ss_max - hweight32(ss_disable);
+	info->subslice_total = info->slice_total * info->subslice_per_slice;
+
+	/*
+	 * Iterate through enabled slices and subslices to
+	 * count the total enabled EU.
+	*/
+	for (s = 0; s < s_max; s++) {
+		if (!(s_enable & BIT(s)))
+			/* skip disabled slice */
+			continue;
+
+		eu_disable = I915_READ(GEN9_EU_DISABLE(s));
+		for (ss = 0; ss < ss_max; ss++) {
+			int eu_per_ss;
+
+			if (ss_disable & BIT(ss))
+				/* skip disabled subslice */
+				continue;
+
+			eu_per_ss = eu_max - hweight8((eu_disable >> (ss*8)) &
+						      eu_mask);
+
+			/*
+			 * Record which subslice(s) has(have) 7 EUs. we
+			 * can tune the hash used to spread work among
+			 * subslices if they are unbalanced.
+			 */
+			if (eu_per_ss == 7)
+				info->subslice_7eu[s] |= BIT(ss);
+
+			info->eu_total += eu_per_ss;
+		}
+	}
+
+	/*
+	 * SKL is expected to always have a uniform distribution
+	 * of EU across subslices with the exception that any one
+	 * EU in any one subslice may be fused off for die
+	 * recovery. BXT is expected to be perfectly uniform in EU
+	 * distribution.
+	*/
+	info->eu_per_subslice = info->subslice_total ?
+				DIV_ROUND_UP(info->eu_total,
+					     info->subslice_total) : 0;
+	/*
+	 * SKL supports slice power gating on devices with more than
+	 * one slice, and supports EU power gating on devices with
+	 * more than one EU pair per subslice. BXT supports subslice
+	 * power gating on devices with more than one subslice, and
+	 * supports EU power gating on devices with more than one EU
+	 * pair per subslice.
+	*/
+	info->has_slice_pg =
+		(IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) &&
+		info->slice_total > 1;
+	info->has_subslice_pg =
+		IS_BROXTON(dev_priv) && info->subslice_total > 1;
+	info->has_eu_pg = info->eu_per_subslice > 2;
+
+	if (IS_BROXTON(dev_priv)) {
+#define IS_SS_DISABLED(_ss_disable, ss)    (_ss_disable & BIT(ss))
+		/*
+		 * There is a HW issue in 2x6 fused down parts that requires
+		 * Pooled EU to be enabled as a WA. The pool configuration
+		 * changes depending upon which subslice is fused down. This
+		 * doesn't affect if the device has all 3 subslices enabled.
+		 */
+		/* WaEnablePooledEuFor2x6:bxt */
+		info->has_pooled_eu = ((info->subslice_per_slice == 3) ||
+				       (info->subslice_per_slice == 2 &&
+					INTEL_REVID(dev_priv) < BXT_REVID_C0));
+
+		info->min_eu_in_pool = 0;
+		if (info->has_pooled_eu) {
+			if (IS_SS_DISABLED(ss_disable, 0) ||
+			    IS_SS_DISABLED(ss_disable, 2))
+				info->min_eu_in_pool = 3;
+			else if (IS_SS_DISABLED(ss_disable, 1))
+				info->min_eu_in_pool = 6;
+			else
+				info->min_eu_in_pool = 9;
+		}
+#undef IS_SS_DISABLED
+	}
+}
+
+static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *info = mkwrite_device_info(dev_priv);
+	const int s_max = 3, ss_max = 3, eu_max = 8;
+	int s, ss;
+	u32 fuse2, eu_disable[s_max], s_enable, ss_disable;
+
+	fuse2 = I915_READ(GEN8_FUSE2);
+	s_enable = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
+	ss_disable = (fuse2 & GEN8_F2_SS_DIS_MASK) >> GEN8_F2_SS_DIS_SHIFT;
+
+	eu_disable[0] = I915_READ(GEN8_EU_DISABLE0) & GEN8_EU_DIS0_S0_MASK;
+	eu_disable[1] = (I915_READ(GEN8_EU_DISABLE0) >> GEN8_EU_DIS0_S1_SHIFT) |
+			((I915_READ(GEN8_EU_DISABLE1) & GEN8_EU_DIS1_S1_MASK) <<
+			 (32 - GEN8_EU_DIS0_S1_SHIFT));
+	eu_disable[2] = (I915_READ(GEN8_EU_DISABLE1) >> GEN8_EU_DIS1_S2_SHIFT) |
+			((I915_READ(GEN8_EU_DISABLE2) & GEN8_EU_DIS2_S2_MASK) <<
+			 (32 - GEN8_EU_DIS1_S2_SHIFT));
+
+	info->slice_total = hweight32(s_enable);
+
+	/*
+	 * The subslice disable field is global, i.e. it applies
+	 * to each of the enabled slices.
+	 */
+	info->subslice_per_slice = ss_max - hweight32(ss_disable);
+	info->subslice_total = info->slice_total * info->subslice_per_slice;
+
+	/*
+	 * Iterate through enabled slices and subslices to
+	 * count the total enabled EU.
+	 */
+	for (s = 0; s < s_max; s++) {
+		if (!(s_enable & (0x1 << s)))
+			/* skip disabled slice */
+			continue;
+
+		for (ss = 0; ss < ss_max; ss++) {
+			u32 n_disabled;
+
+			if (ss_disable & (0x1 << ss))
+				/* skip disabled subslice */
+				continue;
+
+			n_disabled = hweight8(eu_disable[s] >> (ss * eu_max));
+
+			/*
+			 * Record which subslices have 7 EUs.
+			 */
+			if (eu_max - n_disabled == 7)
+				info->subslice_7eu[s] |= 1 << ss;
+
+			info->eu_total += eu_max - n_disabled;
+		}
+	}
+
+	/*
+	 * BDW is expected to always have a uniform distribution of EU across
+	 * subslices with the exception that any one EU in any one subslice may
+	 * be fused off for die recovery.
+	 */
+	info->eu_per_subslice = info->subslice_total ?
+		DIV_ROUND_UP(info->eu_total, info->subslice_total) : 0;
+
+	/*
+	 * BDW supports slice power gating on devices with more than
+	 * one slice.
+	 */
+	info->has_slice_pg = (info->slice_total > 1);
+	info->has_subslice_pg = 0;
+	info->has_eu_pg = 0;
+}
+
+/*
+ * Determine various intel_device_info fields at runtime.
+ *
+ * Use it when either:
+ *   - it's judged too laborious to fill n static structures with the limit
+ *     when a simple if statement does the job,
+ *   - run-time checks (eg read fuse/strap registers) are needed.
+ *
+ * This function needs to be called:
+ *   - after the MMIO has been setup as we are reading registers,
+ *   - after the PCH has been detected,
+ *   - before the first usage of the fields it can tweak.
+ */
+void intel_device_info_runtime_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *info = mkwrite_device_info(dev_priv);
+	enum pipe pipe;
+
+	/*
+	 * Skylake and Broxton currently don't expose the topmost plane as its
+	 * use is exclusive with the legacy cursor and we only want to expose
+	 * one of those, not both. Until we can safely expose the topmost plane
+	 * as a DRM_PLANE_TYPE_CURSOR with all the features exposed/supported,
+	 * we don't expose the topmost plane at all to prevent ABI breakage
+	 * down the line.
+	 */
+	if (IS_BROXTON(dev_priv)) {
+		info->num_sprites[PIPE_A] = 2;
+		info->num_sprites[PIPE_B] = 2;
+		info->num_sprites[PIPE_C] = 1;
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		for_each_pipe(dev_priv, pipe)
+			info->num_sprites[pipe] = 2;
+	else
+		for_each_pipe(dev_priv, pipe)
+			info->num_sprites[pipe] = 1;
+
+	if (i915.disable_display) {
+		DRM_INFO("Display disabled (module parameter)\n");
+		info->num_pipes = 0;
+	} else if (info->num_pipes > 0 &&
+		   (IS_GEN7(dev_priv) || IS_GEN8(dev_priv)) &&
+		   HAS_PCH_SPLIT(dev_priv)) {
+		u32 fuse_strap = I915_READ(FUSE_STRAP);
+		u32 sfuse_strap = I915_READ(SFUSE_STRAP);
+
+		/*
+		 * SFUSE_STRAP is supposed to have a bit signalling the display
+		 * is fused off. Unfortunately it seems that, at least in
+		 * certain cases, fused off display means that PCH display
+		 * reads don't land anywhere. In that case, we read 0s.
+		 *
+		 * On CPT/PPT, we can detect this case as SFUSE_STRAP_FUSE_LOCK
+		 * should be set when taking over after the firmware.
+		 */
+		if (fuse_strap & ILK_INTERNAL_DISPLAY_DISABLE ||
+		    sfuse_strap & SFUSE_STRAP_DISPLAY_DISABLED ||
+		    (dev_priv->pch_type == PCH_CPT &&
+		     !(sfuse_strap & SFUSE_STRAP_FUSE_LOCK))) {
+			DRM_INFO("Display fused off, disabling\n");
+			info->num_pipes = 0;
+		} else if (fuse_strap & IVB_PIPE_C_DISABLE) {
+			DRM_INFO("PipeC fused off\n");
+			info->num_pipes -= 1;
+		}
+	} else if (info->num_pipes > 0 && IS_GEN9(dev_priv)) {
+		u32 dfsm = I915_READ(SKL_DFSM);
+		u8 disabled_mask = 0;
+		bool invalid;
+		int num_bits;
+
+		if (dfsm & SKL_DFSM_PIPE_A_DISABLE)
+			disabled_mask |= BIT(PIPE_A);
+		if (dfsm & SKL_DFSM_PIPE_B_DISABLE)
+			disabled_mask |= BIT(PIPE_B);
+		if (dfsm & SKL_DFSM_PIPE_C_DISABLE)
+			disabled_mask |= BIT(PIPE_C);
+
+		num_bits = hweight8(disabled_mask);
+
+		switch (disabled_mask) {
+		case BIT(PIPE_A):
+		case BIT(PIPE_B):
+		case BIT(PIPE_A) | BIT(PIPE_B):
+		case BIT(PIPE_A) | BIT(PIPE_C):
+			invalid = true;
+			break;
+		default:
+			invalid = false;
+		}
+
+		if (num_bits > info->num_pipes || invalid)
+			DRM_ERROR("invalid pipe fuse configuration: 0x%x\n",
+				  disabled_mask);
+		else
+			info->num_pipes -= num_bits;
+	}
+
+	/* Initialize slice/subslice/EU info */
+	if (IS_CHERRYVIEW(dev_priv))
+		cherryview_sseu_info_init(dev_priv);
+	else if (IS_BROADWELL(dev_priv))
+		broadwell_sseu_info_init(dev_priv);
+	else if (INTEL_INFO(dev_priv)->gen >= 9)
+		gen9_sseu_info_init(dev_priv);
+
+	info->has_snoop = !info->has_llc;
+
+	/* Snooping is broken on BXT A stepping. */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
+		info->has_snoop = false;
+
+	DRM_DEBUG_DRIVER("slice total: %u\n", info->slice_total);
+	DRM_DEBUG_DRIVER("subslice total: %u\n", info->subslice_total);
+	DRM_DEBUG_DRIVER("subslice per slice: %u\n", info->subslice_per_slice);
+	DRM_DEBUG_DRIVER("EU total: %u\n", info->eu_total);
+	DRM_DEBUG_DRIVER("EU per subslice: %u\n", info->eu_per_subslice);
+	DRM_DEBUG_DRIVER("has slice power gating: %s\n",
+			 info->has_slice_pg ? "y" : "n");
+	DRM_DEBUG_DRIVER("has subslice power gating: %s\n",
+			 info->has_subslice_pg ? "y" : "n");
+	DRM_DEBUG_DRIVER("has EU power gating: %s\n",
+			 info->has_eu_pg ? "y" : "n");
+}
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0b2cd669ac05..be3b2cab2640 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -530,52 +530,6 @@ needs_modeset(struct drm_crtc_state *state)
 	return drm_atomic_crtc_needs_modeset(state);
 }
 
-/**
- * Returns whether any output on the specified pipe is of the specified type
- */
-bool intel_pipe_has_type(struct intel_crtc *crtc, enum intel_output_type type)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct intel_encoder *encoder;
-
-	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
-		if (encoder->type == type)
-			return true;
-
-	return false;
-}
-
-/**
- * Returns whether any output on the specified pipe will have the specified
- * type after a staged modeset is complete, i.e., the same as
- * intel_pipe_has_type() but looking at encoder->new_crtc instead of
- * encoder->crtc.
- */
-static bool intel_pipe_will_have_type(const struct intel_crtc_state *crtc_state,
-				      int type)
-{
-	struct drm_atomic_state *state = crtc_state->base.state;
-	struct drm_connector *connector;
-	struct drm_connector_state *connector_state;
-	struct intel_encoder *encoder;
-	int i, num_connectors = 0;
-
-	for_each_connector_in_state(state, connector, connector_state, i) {
-		if (connector_state->crtc != crtc_state->base.crtc)
-			continue;
-
-		num_connectors++;
-
-		encoder = to_intel_encoder(connector_state->best_encoder);
-		if (encoder->type == type)
-			return true;
-	}
-
-	WARN_ON(num_connectors == 0);
-
-	return false;
-}
-
 /*
  * Platform specific helpers to calculate the port PLL loopback- (clock.m),
  * and post-divider (clock.p) values, pre- (clock.vco) and post-divided fast
@@ -688,7 +642,7 @@ i9xx_select_p2_div(const struct intel_limit *limit,
 {
 	struct drm_device *dev = crtc_state->base.crtc->dev;
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		/*
 		 * For LVDS just rely on its current settings for dual-channel.
 		 * We haven't figured out how to reliably set up different
@@ -1080,7 +1034,7 @@ enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 
 static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t reg = PIPEDSL(pipe);
 	u32 line1, line2;
 	u32 line_mask;
@@ -1116,7 +1070,7 @@ static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
 static void intel_wait_for_pipe_off(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 	enum pipe pipe = crtc->pipe;
 
@@ -1124,8 +1078,9 @@ static void intel_wait_for_pipe_off(struct intel_crtc *crtc)
 		i915_reg_t reg = PIPECONF(cpu_transcoder);
 
 		/* Wait for the Pipe State to go off */
-		if (wait_for((I915_READ(reg) & I965_PIPECONF_ACTIVE) == 0,
-			     100))
+		if (intel_wait_for_register(dev_priv,
+					    reg, I965_PIPECONF_ACTIVE, 0,
+					    100))
 			WARN(1, "pipe_off wait timed out\n");
 	} else {
 		/* Wait for the display line to settle */
@@ -1234,7 +1189,7 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv,
 void assert_panel_unlocked(struct drm_i915_private *dev_priv,
 			   enum pipe pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	i915_reg_t pp_reg;
 	u32 val;
 	enum pipe panel_pipe = PIPE_A;
@@ -1276,7 +1231,7 @@ void assert_panel_unlocked(struct drm_i915_private *dev_priv,
 static void assert_cursor(struct drm_i915_private *dev_priv,
 			  enum pipe pipe, bool state)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	bool cur_state;
 
 	if (IS_845G(dev) || IS_I865G(dev))
@@ -1338,7 +1293,7 @@ static void assert_plane(struct drm_i915_private *dev_priv,
 static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 				   enum pipe pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	int i;
 
 	/* Primary planes are fixed to pipes on gen4+ */
@@ -1364,7 +1319,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	int sprite;
 
 	if (INTEL_INFO(dev)->gen >= 9) {
@@ -1544,7 +1499,11 @@ static void _vlv_enable_pll(struct intel_crtc *crtc,
 	POSTING_READ(DPLL(pipe));
 	udelay(150);
 
-	if (wait_for(((I915_READ(DPLL(pipe)) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
+	if (intel_wait_for_register(dev_priv,
+				    DPLL(pipe),
+				    DPLL_LOCK_VLV,
+				    DPLL_LOCK_VLV,
+				    1))
 		DRM_ERROR("DPLL %d failed to lock\n", pipe);
 }
 
@@ -1593,7 +1552,9 @@ static void _chv_enable_pll(struct intel_crtc *crtc,
 	I915_WRITE(DPLL(pipe), pipe_config->dpll_hw_state.dpll);
 
 	/* Check PLL is locked */
-	if (wait_for(((I915_READ(DPLL(pipe)) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
+	if (intel_wait_for_register(dev_priv,
+				    DPLL(pipe), DPLL_LOCK_VLV, DPLL_LOCK_VLV,
+				    1))
 		DRM_ERROR("PLL %d failed to lock\n", pipe);
 }
 
@@ -1639,9 +1600,10 @@ static int intel_num_dvo_pipes(struct drm_device *dev)
 	struct intel_crtc *crtc;
 	int count = 0;
 
-	for_each_intel_crtc(dev, crtc)
+	for_each_intel_crtc(dev, crtc) {
 		count += crtc->base.state->active &&
-			intel_pipe_has_type(crtc, INTEL_OUTPUT_DVO);
+			intel_crtc_has_type(crtc->config, INTEL_OUTPUT_DVO);
+	}
 
 	return count;
 }
@@ -1649,7 +1611,7 @@ static int intel_num_dvo_pipes(struct drm_device *dev)
 static void i9xx_enable_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t reg = DPLL(crtc->pipe);
 	u32 dpll = crtc->config->dpll_hw_state.dpll;
 
@@ -1721,12 +1683,12 @@ static void i9xx_enable_pll(struct intel_crtc *crtc)
 static void i9xx_disable_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 
 	/* Disable DVO 2x clock on both PLLs if necessary */
 	if (IS_I830(dev) &&
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_DVO) &&
+	    intel_crtc_has_type(crtc->config, INTEL_OUTPUT_DVO) &&
 	    !intel_num_dvo_pipes(dev)) {
 		I915_WRITE(DPLL(PIPE_B),
 			   I915_READ(DPLL(PIPE_B)) & ~DPLL_DVO_2X_MODE);
@@ -1813,7 +1775,9 @@ void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
 		BUG();
 	}
 
-	if (wait_for((I915_READ(dpll_reg) & port_mask) == expected_mask, 1000))
+	if (intel_wait_for_register(dev_priv,
+				    dpll_reg, port_mask, expected_mask,
+				    1000))
 		WARN(1, "timed out waiting for port %c ready: got 0x%x, expected 0x%x\n",
 		     port_name(dport->port), I915_READ(dpll_reg) & port_mask, expected_mask);
 }
@@ -1821,7 +1785,7 @@ void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
 static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 					   enum pipe pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	i915_reg_t reg;
@@ -1854,7 +1818,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		 * here for both 8bpc and 12bpc.
 		 */
 		val &= ~PIPECONF_BPC_MASK;
-		if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_HDMI))
+		if (intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_HDMI))
 			val |= PIPECONF_8BPC;
 		else
 			val |= pipeconf_val & PIPECONF_BPC_MASK;
@@ -1863,7 +1827,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	val &= ~TRANS_INTERLACE_MASK;
 	if ((pipeconf_val & PIPECONF_INTERLACE_MASK) == PIPECONF_INTERLACED_ILK)
 		if (HAS_PCH_IBX(dev_priv) &&
-		    intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
+		    intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_SDVO))
 			val |= TRANS_LEGACY_INTERLACED_ILK;
 		else
 			val |= TRANS_INTERLACED;
@@ -1871,7 +1835,9 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		val |= TRANS_PROGRESSIVE;
 
 	I915_WRITE(reg, val | TRANS_ENABLE);
-	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
+	if (intel_wait_for_register(dev_priv,
+				    reg, TRANS_STATE_ENABLE, TRANS_STATE_ENABLE,
+				    100))
 		DRM_ERROR("failed to enable transcoder %c\n", pipe_name(pipe));
 }
 
@@ -1899,14 +1865,18 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		val |= TRANS_PROGRESSIVE;
 
 	I915_WRITE(LPT_TRANSCONF, val);
-	if (wait_for(I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE, 100))
+	if (intel_wait_for_register(dev_priv,
+				    LPT_TRANSCONF,
+				    TRANS_STATE_ENABLE,
+				    TRANS_STATE_ENABLE,
+				    100))
 		DRM_ERROR("Failed to enable PCH transcoder\n");
 }
 
 static void ironlake_disable_pch_transcoder(struct drm_i915_private *dev_priv,
 					    enum pipe pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	i915_reg_t reg;
 	uint32_t val;
 
@@ -1922,7 +1892,9 @@ static void ironlake_disable_pch_transcoder(struct drm_i915_private *dev_priv,
 	val &= ~TRANS_ENABLE;
 	I915_WRITE(reg, val);
 	/* wait for PCH transcoder off, transcoder state */
-	if (wait_for((I915_READ(reg) & TRANS_STATE_ENABLE) == 0, 50))
+	if (intel_wait_for_register(dev_priv,
+				    reg, TRANS_STATE_ENABLE, 0,
+				    50))
 		DRM_ERROR("failed to disable transcoder %c\n", pipe_name(pipe));
 
 	if (HAS_PCH_CPT(dev)) {
@@ -1942,7 +1914,9 @@ static void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv)
 	val &= ~TRANS_ENABLE;
 	I915_WRITE(LPT_TRANSCONF, val);
 	/* wait for PCH transcoder off, transcoder state */
-	if (wait_for((I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE) == 0, 50))
+	if (intel_wait_for_register(dev_priv,
+				    LPT_TRANSCONF, TRANS_STATE_ENABLE, 0,
+				    50))
 		DRM_ERROR("Failed to disable PCH transcoder\n");
 
 	/* Workaround: clear timing override bit. */
@@ -1961,7 +1935,7 @@ static void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv)
 static void intel_enable_pipe(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 	enum pipe pch_transcoder;
@@ -1985,7 +1959,7 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 	 * need the check.
 	 */
 	if (HAS_GMCH_DISPLAY(dev_priv))
-		if (crtc->config->has_dsi_encoder)
+		if (intel_crtc_has_type(crtc->config, INTEL_OUTPUT_DSI))
 			assert_dsi_pll_enabled(dev_priv);
 		else
 			assert_pll_enabled(dev_priv, pipe);
@@ -2034,7 +2008,7 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
  */
 static void intel_disable_pipe(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 	enum pipe pipe = crtc->pipe;
 	i915_reg_t reg;
@@ -2072,15 +2046,6 @@ static void intel_disable_pipe(struct intel_crtc *crtc)
 		intel_wait_for_pipe_off(crtc);
 }
 
-static bool need_vtd_wa(struct drm_device *dev)
-{
-#ifdef CONFIG_INTEL_IOMMU
-	if (INTEL_INFO(dev)->gen >= 6 && intel_iommu_gfx_mapped)
-		return true;
-#endif
-	return false;
-}
-
 static unsigned int intel_tile_size(const struct drm_i915_private *dev_priv)
 {
 	return IS_GEN2(dev_priv) ? 2048 : 4096;
@@ -2245,7 +2210,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb,
 			   unsigned int rotation)
 {
 	struct drm_device *dev = fb->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct i915_ggtt_view view;
 	u32 alignment;
@@ -2262,7 +2227,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb,
 	 * we should always have valid PTE following the scanout preventing
 	 * the VT-d warning.
 	 */
-	if (need_vtd_wa(dev) && alignment < 256 * 1024)
+	if (intel_scanout_needs_vtd_wa(dev_priv) && alignment < 256 * 1024)
 		alignment = 256 * 1024;
 
 	/*
@@ -2547,7 +2512,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc,
 			     struct intel_initial_plane_config *plane_config)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *c;
 	struct intel_crtc *i;
 	struct drm_i915_gem_object *obj;
@@ -2643,7 +2608,7 @@ static void i9xx_update_primary_plane(struct drm_plane *primary,
 				      const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = primary->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -2756,7 +2721,7 @@ static void i9xx_disable_primary_plane(struct drm_plane *primary,
 				       struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int plane = intel_crtc->plane;
 
@@ -2773,7 +2738,7 @@ static void ironlake_update_primary_plane(struct drm_plane *primary,
 					  const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = primary->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -2901,7 +2866,7 @@ u32 intel_plane_obj_offset(struct intel_plane *intel_plane,
 static void skl_detach_scaler(struct intel_crtc *intel_crtc, int id)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(SKL_PS_CTRL(intel_crtc->pipe, id), 0);
 	I915_WRITE(SKL_PS_WIN_POS(intel_crtc->pipe, id), 0);
@@ -3011,7 +2976,7 @@ static void skylake_update_primary_plane(struct drm_plane *plane,
 					 const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -3095,7 +3060,7 @@ static void skylake_disable_primary_plane(struct drm_plane *primary,
 					  struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = to_intel_crtc(crtc)->pipe;
 
 	I915_WRITE(PLANE_CTL(pipe, 0), 0);
@@ -3118,7 +3083,7 @@ static void intel_complete_page_flips(struct drm_i915_private *dev_priv)
 {
 	struct intel_crtc *crtc;
 
-	for_each_intel_crtc(dev_priv->dev, crtc)
+	for_each_intel_crtc(&dev_priv->drm, crtc)
 		intel_finish_page_flip_cs(dev_priv, crtc->pipe);
 }
 
@@ -3152,12 +3117,12 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv)
 	if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv))
 		return;
 
-	drm_modeset_lock_all(dev_priv->dev);
+	drm_modeset_lock_all(&dev_priv->drm);
 	/*
 	 * Disabling the crtcs gracefully seems nicer. Also the
 	 * g33 docs say we should at least disable all the planes.
 	 */
-	intel_display_suspend(dev_priv->dev);
+	intel_display_suspend(&dev_priv->drm);
 }
 
 void intel_finish_reset(struct drm_i915_private *dev_priv)
@@ -3184,7 +3149,7 @@ void intel_finish_reset(struct drm_i915_private *dev_priv)
 		 * FIXME: Atomic will make this obsolete since we won't schedule
 		 * CS-based flips (which might get lost in gpu resets) any more.
 		 */
-		intel_update_primary_planes(dev_priv->dev);
+		intel_update_primary_planes(&dev_priv->drm);
 		return;
 	}
 
@@ -3195,18 +3160,18 @@ void intel_finish_reset(struct drm_i915_private *dev_priv)
 	intel_runtime_pm_disable_interrupts(dev_priv);
 	intel_runtime_pm_enable_interrupts(dev_priv);
 
-	intel_modeset_init_hw(dev_priv->dev);
+	intel_modeset_init_hw(&dev_priv->drm);
 
 	spin_lock_irq(&dev_priv->irq_lock);
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev_priv);
 	spin_unlock_irq(&dev_priv->irq_lock);
 
-	intel_display_resume(dev_priv->dev);
+	intel_display_resume(&dev_priv->drm);
 
 	intel_hpd_init(dev_priv);
 
-	drm_modeset_unlock_all(dev_priv->dev);
+	drm_modeset_unlock_all(&dev_priv->drm);
 }
 
 static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
@@ -3231,7 +3196,7 @@ static void intel_update_pipe_config(struct intel_crtc *crtc,
 				     struct intel_crtc_state *old_crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_state *pipe_config =
 		to_intel_crtc_state(crtc->base.state);
 
@@ -3272,7 +3237,7 @@ static void intel_update_pipe_config(struct intel_crtc *crtc,
 static void intel_fdi_normal_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
@@ -3315,7 +3280,7 @@ static void intel_fdi_normal_train(struct drm_crtc *crtc)
 static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
@@ -3416,7 +3381,7 @@ static const int snb_b_fdi_train_param[] = {
 static void gen6_fdi_link_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
@@ -3549,7 +3514,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
@@ -3668,7 +3633,7 @@ train_done:
 static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
 	u32 temp;
@@ -3705,7 +3670,7 @@ static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 static void ironlake_fdi_pll_disable(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
 	u32 temp;
@@ -3735,7 +3700,7 @@ static void ironlake_fdi_pll_disable(struct intel_crtc *intel_crtc)
 static void ironlake_fdi_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	i915_reg_t reg;
@@ -3831,7 +3796,7 @@ static void page_flip_completed(struct intel_crtc *intel_crtc)
 static int intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	long ret;
 
 	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
@@ -3994,7 +3959,7 @@ static void ironlake_pch_transcoder_set_timings(struct intel_crtc *crtc,
 						enum pipe pch_transcoder)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 
 	I915_WRITE(PCH_TRANS_HTOTAL(pch_transcoder),
@@ -4016,7 +3981,7 @@ static void ironlake_pch_transcoder_set_timings(struct intel_crtc *crtc,
 
 static void cpt_set_fdi_bc_bifurcation(struct drm_device *dev, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t temp;
 
 	temp = I915_READ(SOUTH_CHICKEN1);
@@ -4066,7 +4031,7 @@ intel_trans_dp_port_sel(struct drm_crtc *crtc)
 	struct intel_encoder *encoder;
 
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
-		if (encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
+		if (encoder->type == INTEL_OUTPUT_DP ||
 		    encoder->type == INTEL_OUTPUT_EDP)
 			return enc_to_dig_port(&encoder->base)->port;
 	}
@@ -4085,7 +4050,7 @@ intel_trans_dp_port_sel(struct drm_crtc *crtc)
 static void ironlake_pch_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	u32 temp;
@@ -4135,7 +4100,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	intel_fdi_normal_train(crtc);
 
 	/* For PCH DP, enable TRANS_DP_CTL */
-	if (HAS_PCH_CPT(dev) && intel_crtc->config->has_dp_encoder) {
+	if (HAS_PCH_CPT(dev) && intel_crtc_has_dp_encoder(intel_crtc->config)) {
 		const struct drm_display_mode *adjusted_mode =
 			&intel_crtc->config->base.adjusted_mode;
 		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) >> 5;
@@ -4175,7 +4140,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 static void lpt_pch_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 
@@ -4191,7 +4156,7 @@ static void lpt_pch_enable(struct drm_crtc *crtc)
 
 static void cpt_verify_modeset(struct drm_device *dev, int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t dslreg = PIPEDSL(pipe);
 	u32 temp;
 
@@ -4369,7 +4334,7 @@ static void skylake_scaler_disable(struct intel_crtc *crtc)
 static void skylake_pfit_enable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = crtc->pipe;
 	struct intel_crtc_scaler_state *scaler_state =
 		&crtc->config->scaler_state;
@@ -4397,7 +4362,7 @@ static void skylake_pfit_enable(struct intel_crtc *crtc)
 static void ironlake_pfit_enable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = crtc->pipe;
 
 	if (crtc->config->pch_pfit.enabled) {
@@ -4418,7 +4383,7 @@ static void ironlake_pfit_enable(struct intel_crtc *crtc)
 void hsw_enable_ips(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!crtc->config->ips_enabled)
 		return;
@@ -4446,7 +4411,9 @@ void hsw_enable_ips(struct intel_crtc *crtc)
 		 * and don't wait for vblanks until the end of crtc_enable, then
 		 * the HW state readout code will complain that the expected
 		 * IPS_CTL value is not the one we read. */
-		if (wait_for(I915_READ_NOTRACE(IPS_CTL) & IPS_ENABLE, 50))
+		if (intel_wait_for_register(dev_priv,
+					    IPS_CTL, IPS_ENABLE, IPS_ENABLE,
+					    50))
 			DRM_ERROR("Timed out waiting for IPS enable\n");
 	}
 }
@@ -4454,7 +4421,7 @@ void hsw_enable_ips(struct intel_crtc *crtc)
 void hsw_disable_ips(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!crtc->config->ips_enabled)
 		return;
@@ -4465,7 +4432,9 @@ void hsw_disable_ips(struct intel_crtc *crtc)
 		WARN_ON(sandybridge_pcode_write(dev_priv, DISPLAY_IPS_CONTROL, 0));
 		mutex_unlock(&dev_priv->rps.hw_lock);
 		/* wait for pcode to finish disabling IPS, which may take up to 42ms */
-		if (wait_for((I915_READ(IPS_CTL) & IPS_ENABLE) == 0, 42))
+		if (intel_wait_for_register(dev_priv,
+					    IPS_CTL, IPS_ENABLE, 0,
+					    42))
 			DRM_ERROR("Timed out waiting for IPS disable\n");
 	} else {
 		I915_WRITE(IPS_CTL, 0);
@@ -4480,7 +4449,7 @@ static void intel_crtc_dpms_overlay_disable(struct intel_crtc *intel_crtc)
 {
 	if (intel_crtc->overlay) {
 		struct drm_device *dev = intel_crtc->base.dev;
-		struct drm_i915_private *dev_priv = dev->dev_private;
+		struct drm_i915_private *dev_priv = to_i915(dev);
 
 		mutex_lock(&dev->struct_mutex);
 		dev_priv->mm.interruptible = false;
@@ -4508,7 +4477,7 @@ static void
 intel_post_enable_primary(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 
@@ -4540,7 +4509,7 @@ static void
 intel_pre_disable_primary(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 
@@ -4567,7 +4536,7 @@ static void
 intel_pre_disable_primary_noatomic(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 
@@ -4626,7 +4595,7 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->base.crtc);
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_state *pipe_config =
 		to_intel_crtc_state(crtc->base.state);
 	struct drm_atomic_state *old_state = old_crtc_state->base.state;
@@ -4729,7 +4698,7 @@ static void intel_crtc_disable_planes(struct drm_crtc *crtc, unsigned plane_mask
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
@@ -4757,7 +4726,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	if (intel_crtc->config->has_pch_encoder)
 		intel_prepare_shared_dpll(intel_crtc);
 
-	if (intel_crtc->config->has_dp_encoder)
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		intel_dp_set_m_n(intel_crtc, M1_N1);
 
 	intel_set_pipe_timings(intel_crtc);
@@ -4826,7 +4795,7 @@ static bool hsw_crtc_supports_ips(struct intel_crtc *crtc)
 static void haswell_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe, hsw_workaround_pipe;
@@ -4848,10 +4817,10 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	if (intel_crtc->config->shared_dpll)
 		intel_enable_shared_dpll(intel_crtc);
 
-	if (intel_crtc->config->has_dp_encoder)
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		intel_dp_set_m_n(intel_crtc, M1_N1);
 
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_set_pipe_timings(intel_crtc);
 
 	intel_set_pipe_src_size(intel_crtc);
@@ -4867,7 +4836,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 				     &intel_crtc->config->fdi_m_n, NULL);
 	}
 
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		haswell_set_pipeconf(crtc);
 
 	haswell_set_pipemisc(crtc);
@@ -4889,7 +4858,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	if (intel_crtc->config->has_pch_encoder)
 		dev_priv->display.fdi_link_train(crtc);
 
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_ddi_enable_pipe_clock(intel_crtc);
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -4904,7 +4873,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_color_load_luts(&pipe_config->base);
 
 	intel_ddi_set_pipe_settings(crtc);
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_ddi_enable_transcoder_func(crtc);
 
 	if (dev_priv->display.initial_watermarks != NULL)
@@ -4913,7 +4882,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 		intel_update_watermarks(crtc);
 
 	/* XXX: Do the pipe assertions at the right place for BXT DSI. */
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_enable_pipe(intel_crtc);
 
 	if (intel_crtc->config->has_pch_encoder)
@@ -4950,7 +4919,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 static void ironlake_pfit_disable(struct intel_crtc *crtc, bool force)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = crtc->pipe;
 
 	/* To avoid upsetting the power well on haswell only disable the pfit if
@@ -4965,7 +4934,7 @@ static void ironlake_pfit_disable(struct intel_crtc *crtc, bool force)
 static void ironlake_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
@@ -5028,7 +4997,7 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 static void haswell_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
@@ -5046,13 +5015,13 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	assert_vblank_disabled(crtc);
 
 	/* XXX: Do the pipe assertions at the right place for BXT DSI. */
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_disable_pipe(intel_crtc);
 
 	if (intel_crtc->config->dp_encoder_is_mst)
 		intel_ddi_set_vc_payload_alloc(crtc, false);
 
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
 	if (INTEL_INFO(dev)->gen >= 9)
@@ -5060,7 +5029,7 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	else
 		ironlake_pfit_disable(intel_crtc, false);
 
-	if (!intel_crtc->config->has_dsi_encoder)
+	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_ddi_disable_pipe_clock(intel_crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
@@ -5080,7 +5049,7 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 static void i9xx_pfit_enable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_state *pipe_config = crtc->config;
 
 	if (!pipe_config->gmch_pfit.control)
@@ -5150,7 +5119,7 @@ intel_display_port_power_domain(struct intel_encoder *intel_encoder)
 	case INTEL_OUTPUT_UNKNOWN:
 		/* Only DDI platforms should ever use this output type */
 		WARN_ON_ONCE(!HAS_DDI(dev));
-	case INTEL_OUTPUT_DISPLAYPORT:
+	case INTEL_OUTPUT_DP:
 	case INTEL_OUTPUT_HDMI:
 	case INTEL_OUTPUT_EDP:
 		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
@@ -5184,7 +5153,7 @@ intel_display_port_aux_power_domain(struct intel_encoder *intel_encoder)
 		 * run the DP detection too.
 		 */
 		WARN_ON_ONCE(!HAS_DDI(dev));
-	case INTEL_OUTPUT_DISPLAYPORT:
+	case INTEL_OUTPUT_DP:
 	case INTEL_OUTPUT_EDP:
 		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
 		return port_to_aux_power_domain(intel_dig_port->port);
@@ -5232,7 +5201,7 @@ static unsigned long
 modeset_get_crtc_power_domains(struct drm_crtc *crtc,
 			       struct intel_crtc_state *crtc_state)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum intel_display_power_domain domain;
 	unsigned long domains, new_domains, old_domains;
@@ -5277,7 +5246,7 @@ static int skl_calc_cdclk(int max_pixclk, int vco);
 
 static void intel_update_max_cdclk(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) {
 		u32 limit = I915_READ(SKL_DFSM) & SKL_DFSM_CDCLK_LIMIT_MASK;
@@ -5338,7 +5307,7 @@ static void intel_update_max_cdclk(struct drm_device *dev)
 
 static void intel_update_cdclk(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->cdclk_freq = dev_priv->display.get_display_clock_speed(dev);
 
@@ -5395,7 +5364,9 @@ static void bxt_de_pll_disable(struct drm_i915_private *dev_priv)
 	I915_WRITE(BXT_DE_PLL_ENABLE, 0);
 
 	/* Timeout 200us */
-	if (wait_for((I915_READ(BXT_DE_PLL_ENABLE) & BXT_DE_PLL_LOCK) == 0, 1))
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 0,
+				    1))
 		DRM_ERROR("timeout waiting for DE PLL unlock\n");
 
 	dev_priv->cdclk_pll.vco = 0;
@@ -5414,7 +5385,11 @@ static void bxt_de_pll_enable(struct drm_i915_private *dev_priv, int vco)
 	I915_WRITE(BXT_DE_PLL_ENABLE, BXT_DE_PLL_PLL_ENABLE);
 
 	/* Timeout 200us */
-	if (wait_for((I915_READ(BXT_DE_PLL_ENABLE) & BXT_DE_PLL_LOCK) != 0, 1))
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DE_PLL_ENABLE,
+				    BXT_DE_PLL_LOCK,
+				    BXT_DE_PLL_LOCK,
+				    1))
 		DRM_ERROR("timeout waiting for DE PLL lock\n");
 
 	dev_priv->cdclk_pll.vco = vco;
@@ -5495,14 +5470,14 @@ static void bxt_set_cdclk(struct drm_i915_private *dev_priv, int cdclk)
 		return;
 	}
 
-	intel_update_cdclk(dev_priv->dev);
+	intel_update_cdclk(&dev_priv->drm);
 }
 
 static void bxt_sanitize_cdclk(struct drm_i915_private *dev_priv)
 {
 	u32 cdctl, expected;
 
-	intel_update_cdclk(dev_priv->dev);
+	intel_update_cdclk(&dev_priv->drm);
 
 	if (dev_priv->cdclk_pll.vco == 0 ||
 	    dev_priv->cdclk_freq == dev_priv->cdclk_pll.ref)
@@ -5635,7 +5610,7 @@ void skl_set_preferred_cdclk_vco(struct drm_i915_private *dev_priv, int vco)
 	dev_priv->skl_preferred_vco_freq = vco;
 
 	if (changed)
-		intel_update_max_cdclk(dev_priv->dev);
+		intel_update_max_cdclk(&dev_priv->drm);
 }
 
 static void
@@ -5677,7 +5652,9 @@ skl_dpll0_enable(struct drm_i915_private *dev_priv, int vco)
 
 	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) | LCPLL_PLL_ENABLE);
 
-	if (wait_for(I915_READ(LCPLL1_CTL) & LCPLL_PLL_LOCK, 5))
+	if (intel_wait_for_register(dev_priv,
+				    LCPLL1_CTL, LCPLL_PLL_LOCK, LCPLL_PLL_LOCK,
+				    5))
 		DRM_ERROR("DPLL0 not locked\n");
 
 	dev_priv->cdclk_pll.vco = vco;
@@ -5690,7 +5667,9 @@ static void
 skl_dpll0_disable(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) & ~LCPLL_PLL_ENABLE);
-	if (wait_for(!(I915_READ(LCPLL1_CTL) & LCPLL_PLL_LOCK), 1))
+	if (intel_wait_for_register(dev_priv,
+				   LCPLL1_CTL, LCPLL_PLL_LOCK, 0,
+				   1))
 		DRM_ERROR("Couldn't disable DPLL0\n");
 
 	dev_priv->cdclk_pll.vco = 0;
@@ -5725,7 +5704,7 @@ static bool skl_cdclk_wait_for_pcu_ready(struct drm_i915_private *dev_priv)
 
 static void skl_set_cdclk(struct drm_i915_private *dev_priv, int cdclk, int vco)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	u32 freq_select, pcu_ack;
 
 	WARN_ON((cdclk == 24000) != (vco == 0));
@@ -5823,7 +5802,7 @@ static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv)
 	if ((I915_READ(SWF_ILK(0x18)) & 0x00FFFFFF) == 0)
 		goto sanitize;
 
-	intel_update_cdclk(dev_priv->dev);
+	intel_update_cdclk(&dev_priv->drm);
 	/* Is PLL enabled and locked ? */
 	if (dev_priv->cdclk_pll.vco == 0 ||
 	    dev_priv->cdclk_freq == dev_priv->cdclk_pll.ref)
@@ -5854,7 +5833,7 @@ sanitize:
 /* Adjust CDclk dividers to allow high res or save power if possible */
 static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val, cmd;
 
 	WARN_ON(dev_priv->display.get_display_clock_speed(dev)
@@ -5919,7 +5898,7 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 
 static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val, cmd;
 
 	WARN_ON(dev_priv->display.get_display_clock_speed(dev)
@@ -6007,7 +5986,7 @@ static int intel_mode_max_pixclk(struct drm_device *dev,
 				 struct drm_atomic_state *state)
 {
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *crtc_state;
 	unsigned max_pixclk = 0, i;
@@ -6034,7 +6013,7 @@ static int intel_mode_max_pixclk(struct drm_device *dev,
 static int valleyview_modeset_calc_cdclk(struct drm_atomic_state *state)
 {
 	struct drm_device *dev = state->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int max_pixclk = intel_mode_max_pixclk(dev, state);
 	struct intel_atomic_state *intel_state =
 		to_intel_atomic_state(state);
@@ -6102,7 +6081,7 @@ static void vlv_program_pfi_credits(struct drm_i915_private *dev_priv)
 static void valleyview_modeset_commit_cdclk(struct drm_atomic_state *old_state)
 {
 	struct drm_device *dev = old_state->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_atomic_state *old_intel_state =
 		to_intel_atomic_state(old_state);
 	unsigned req_cdclk = old_intel_state->dev_cdclk;
@@ -6141,14 +6120,14 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 	if (WARN_ON(intel_crtc->active))
 		return;
 
-	if (intel_crtc->config->has_dp_encoder)
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		intel_dp_set_m_n(intel_crtc, M1_N1);
 
 	intel_set_pipe_timings(intel_crtc);
 	intel_set_pipe_src_size(intel_crtc);
 
 	if (IS_CHERRYVIEW(dev) && pipe == PIPE_B) {
-		struct drm_i915_private *dev_priv = dev->dev_private;
+		struct drm_i915_private *dev_priv = to_i915(dev);
 
 		I915_WRITE(CHV_BLEND(pipe), CHV_BLEND_LEGACY);
 		I915_WRITE(CHV_CANVAS(pipe), 0);
@@ -6193,7 +6172,7 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 static void i9xx_set_pll_dividers(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(FP0(crtc->pipe), crtc->config->dpll_hw_state.fp0);
 	I915_WRITE(FP1(crtc->pipe), crtc->config->dpll_hw_state.fp1);
@@ -6214,7 +6193,7 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 
 	i9xx_set_pll_dividers(intel_crtc);
 
-	if (intel_crtc->config->has_dp_encoder)
+	if (intel_crtc_has_dp_encoder(intel_crtc->config))
 		intel_dp_set_m_n(intel_crtc, M1_N1);
 
 	intel_set_pipe_timings(intel_crtc);
@@ -6250,7 +6229,7 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 static void i9xx_pfit_disable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!crtc->config->gmch_pfit.control)
 		return;
@@ -6265,7 +6244,7 @@ static void i9xx_pfit_disable(struct intel_crtc *crtc)
 static void i9xx_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
@@ -6291,7 +6270,7 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
 
-	if (!intel_crtc->config->has_dsi_encoder) {
+	if (!intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_DSI)) {
 		if (IS_CHERRYVIEW(dev))
 			chv_disable_pll(dev_priv, pipe);
 		else if (IS_VALLEYVIEW(dev))
@@ -6609,7 +6588,7 @@ static void hsw_compute_ips_config(struct intel_crtc *crtc,
 				   struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	pipe_config->ips_enabled = i915.enable_ips &&
 		hsw_crtc_supports_ips(crtc) &&
@@ -6629,7 +6608,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 				     struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode;
 	int clock_limit = dev_priv->max_dotclk_freq;
 
@@ -6660,7 +6639,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	 * - LVDS dual channel mode
 	 * - Double wide pipe
 	 */
-	if ((intel_pipe_will_have_type(pipe_config, INTEL_OUTPUT_LVDS) &&
+	if ((intel_crtc_has_type(pipe_config, INTEL_OUTPUT_LVDS) &&
 	     intel_is_dual_link_lvds(dev)) || pipe_config->double_wide)
 		pipe_config->pipe_src_w &= ~1;
 
@@ -6779,7 +6758,7 @@ static int broxton_get_display_clock_speed(struct drm_device *dev)
 
 static int broadwell_get_display_clock_speed(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t lcpll = I915_READ(LCPLL_CTL);
 	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
 
@@ -6799,7 +6778,7 @@ static int broadwell_get_display_clock_speed(struct drm_device *dev)
 
 static int haswell_get_display_clock_speed(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t lcpll = I915_READ(LCPLL_CTL);
 	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
 
@@ -6933,7 +6912,7 @@ static int i830_get_display_clock_speed(struct drm_device *dev)
 
 static unsigned int intel_hpll_vco(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	static const unsigned int blb_vco[8] = {
 		[0] = 3200000,
 		[1] = 4000000,
@@ -7171,7 +7150,7 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	crtc_state->dpll_hw_state.fp0 = fp;
 
 	crtc->lowfreq_avail = false;
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
 	    reduced_clock) {
 		crtc_state->dpll_hw_state.fp1 = fp2;
 		crtc->lowfreq_avail = true;
@@ -7213,7 +7192,7 @@ static void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
 					 struct intel_link_m_n *m_n)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = crtc->pipe;
 
 	I915_WRITE(PCH_TRANS_DATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
@@ -7227,7 +7206,7 @@ static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
 					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = crtc->pipe;
 	enum transcoder transcoder = crtc->config->cpu_transcoder;
 
@@ -7290,7 +7269,7 @@ static void vlv_compute_dpll(struct intel_crtc *crtc,
 		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
 
 	/* DPLL not used with DSI, but still need the rest set up */
-	if (!pipe_config->has_dsi_encoder)
+	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
 		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE |
 			DPLL_EXT_BUFFER_ENABLE_VLV;
 
@@ -7307,7 +7286,7 @@ static void chv_compute_dpll(struct intel_crtc *crtc,
 		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
 
 	/* DPLL not used with DSI, but still need the rest set up */
-	if (!pipe_config->has_dsi_encoder)
+	if (!intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DSI))
 		pipe_config->dpll_hw_state.dpll |= DPLL_VCO_ENABLE;
 
 	pipe_config->dpll_hw_state.dpll_md =
@@ -7318,7 +7297,7 @@ static void vlv_prepare_pll(struct intel_crtc *crtc,
 			    const struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 	u32 mdiv;
 	u32 bestn, bestm1, bestm2, bestp1, bestp2;
@@ -7377,15 +7356,15 @@ static void vlv_prepare_pll(struct intel_crtc *crtc,
 
 	/* Set HBR and RBR LPF coefficients */
 	if (pipe_config->port_clock == 162000 ||
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_ANALOG) ||
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI))
+	    intel_crtc_has_type(crtc->config, INTEL_OUTPUT_ANALOG) ||
+	    intel_crtc_has_type(crtc->config, INTEL_OUTPUT_HDMI))
 		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
 				 0x009f0003);
 	else
 		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
 				 0x00d0000f);
 
-	if (pipe_config->has_dp_encoder) {
+	if (intel_crtc_has_dp_encoder(pipe_config)) {
 		/* Use SSC source */
 		if (pipe == PIPE_A)
 			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
@@ -7405,8 +7384,7 @@ static void vlv_prepare_pll(struct intel_crtc *crtc,
 
 	coreclk = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW7(pipe));
 	coreclk = (coreclk & 0x0000ff00) | 0x01c00000;
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
+	if (intel_crtc_has_dp_encoder(crtc->config))
 		coreclk |= 0x01000000;
 	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW7(pipe), coreclk);
 
@@ -7418,7 +7396,7 @@ static void chv_prepare_pll(struct intel_crtc *crtc,
 			    const struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 	enum dpio_channel port = vlv_pipe_to_channel(pipe);
 	u32 loopfilter, tribuf_calcntr;
@@ -7580,19 +7558,15 @@ static void i9xx_compute_dpll(struct intel_crtc *crtc,
 			      struct dpll *reduced_clock)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 dpll;
-	bool is_sdvo;
 	struct dpll *clock = &crtc_state->dpll;
 
 	i9xx_update_pll_dividers(crtc, crtc_state, reduced_clock);
 
-	is_sdvo = intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_HDMI);
-
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS))
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
@@ -7602,10 +7576,11 @@ static void i9xx_compute_dpll(struct intel_crtc *crtc,
 			<< SDVO_MULTIPLIER_SHIFT_HIRES;
 	}
 
-	if (is_sdvo)
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
-	if (crtc_state->has_dp_encoder)
+	if (intel_crtc_has_dp_encoder(crtc_state))
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
@@ -7635,7 +7610,7 @@ static void i9xx_compute_dpll(struct intel_crtc *crtc,
 
 	if (crtc_state->sdvo_tv_clock)
 		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv))
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
@@ -7656,7 +7631,7 @@ static void i8xx_compute_dpll(struct intel_crtc *crtc,
 			      struct dpll *reduced_clock)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 dpll;
 	struct dpll *clock = &crtc_state->dpll;
 
@@ -7664,7 +7639,7 @@ static void i8xx_compute_dpll(struct intel_crtc *crtc,
 
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	} else {
 		if (clock->p1 == 2)
@@ -7675,10 +7650,10 @@ static void i8xx_compute_dpll(struct intel_crtc *crtc,
 			dpll |= PLL_P2_DIVIDE_BY_4;
 	}
 
-	if (!IS_I830(dev) && intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_DVO))
+	if (!IS_I830(dev) && intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO))
 		dpll |= DPLL_DVO_2X_MODE;
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
 	    intel_panel_use_ssc(dev_priv))
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
@@ -7691,7 +7666,7 @@ static void i8xx_compute_dpll(struct intel_crtc *crtc,
 static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	const struct drm_display_mode *adjusted_mode = &intel_crtc->config->base.adjusted_mode;
@@ -7708,7 +7683,7 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 		crtc_vtotal -= 1;
 		crtc_vblank_end -= 1;
 
-		if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
+		if (intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_SDVO))
 			vsyncshift = (adjusted_mode->crtc_htotal - 1) / 2;
 		else
 			vsyncshift = adjusted_mode->crtc_hsync_start -
@@ -7753,7 +7728,7 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 static void intel_set_pipe_src_size(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = intel_crtc->pipe;
 
 	/* pipesrc controls the size that is scaled from, which should
@@ -7768,7 +7743,7 @@ static void intel_get_pipe_timings(struct intel_crtc *crtc,
 				   struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
 	uint32_t tmp;
 
@@ -7803,7 +7778,7 @@ static void intel_get_pipe_src_size(struct intel_crtc *crtc,
 				    struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tmp;
 
 	tmp = I915_READ(PIPESRC(crtc->pipe));
@@ -7841,7 +7816,7 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode,
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t pipeconf;
 
 	pipeconf = 0;
@@ -7887,7 +7862,7 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 
 	if (intel_crtc->config->base.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
 		if (INTEL_INFO(dev)->gen < 4 ||
-		    intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
+		    intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_SDVO))
 			pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 		else
 			pipeconf |= PIPECONF_INTERLACE_W_SYNC_SHIFT;
@@ -7906,21 +7881,21 @@ static int i8xx_crtc_compute_clock(struct intel_crtc *crtc,
 				   struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_limit *limit;
 	int refclk = 48000;
 
 	memset(&crtc_state->dpll_hw_state, 0,
 	       sizeof(crtc_state->dpll_hw_state));
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if (intel_panel_use_ssc(dev_priv)) {
 			refclk = dev_priv->vbt.lvds_ssc_freq;
 			DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n", refclk);
 		}
 
 		limit = &intel_limits_i8xx_lvds;
-	} else if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_DVO)) {
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO)) {
 		limit = &intel_limits_i8xx_dvo;
 	} else {
 		limit = &intel_limits_i8xx_dac;
@@ -7942,14 +7917,14 @@ static int g4x_crtc_compute_clock(struct intel_crtc *crtc,
 				  struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_limit *limit;
 	int refclk = 96000;
 
 	memset(&crtc_state->dpll_hw_state, 0,
 	       sizeof(crtc_state->dpll_hw_state));
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if (intel_panel_use_ssc(dev_priv)) {
 			refclk = dev_priv->vbt.lvds_ssc_freq;
 			DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n", refclk);
@@ -7959,10 +7934,10 @@ static int g4x_crtc_compute_clock(struct intel_crtc *crtc,
 			limit = &intel_limits_g4x_dual_channel_lvds;
 		else
 			limit = &intel_limits_g4x_single_channel_lvds;
-	} else if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_HDMI) ||
-		   intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) ||
+		   intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
 		limit = &intel_limits_g4x_hdmi;
-	} else if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_SDVO)) {
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO)) {
 		limit = &intel_limits_g4x_sdvo;
 	} else {
 		/* The option is for other outputs */
@@ -7985,14 +7960,14 @@ static int pnv_crtc_compute_clock(struct intel_crtc *crtc,
 				  struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_limit *limit;
 	int refclk = 96000;
 
 	memset(&crtc_state->dpll_hw_state, 0,
 	       sizeof(crtc_state->dpll_hw_state));
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if (intel_panel_use_ssc(dev_priv)) {
 			refclk = dev_priv->vbt.lvds_ssc_freq;
 			DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n", refclk);
@@ -8019,14 +7994,14 @@ static int i9xx_crtc_compute_clock(struct intel_crtc *crtc,
 				   struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_limit *limit;
 	int refclk = 96000;
 
 	memset(&crtc_state->dpll_hw_state, 0,
 	       sizeof(crtc_state->dpll_hw_state));
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if (intel_panel_use_ssc(dev_priv)) {
 			refclk = dev_priv->vbt.lvds_ssc_freq;
 			DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n", refclk);
@@ -8095,7 +8070,7 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc,
 				 struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t tmp;
 
 	if (INTEL_INFO(dev)->gen <= 3 && (IS_I830(dev) || !IS_MOBILE(dev)))
@@ -8122,7 +8097,7 @@ static void vlv_crtc_clock_get(struct intel_crtc *crtc,
 			       struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = pipe_config->cpu_transcoder;
 	struct dpll clock;
 	u32 mdiv;
@@ -8150,7 +8125,7 @@ i9xx_get_initial_plane_config(struct intel_crtc *crtc,
 			      struct intel_initial_plane_config *plane_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val, base, offset;
 	int pipe = crtc->pipe, plane = crtc->plane;
 	int fourcc, pixel_format;
@@ -8218,7 +8193,7 @@ static void chv_crtc_clock_get(struct intel_crtc *crtc,
 			       struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = pipe_config->cpu_transcoder;
 	enum dpio_channel port = vlv_pipe_to_channel(pipe);
 	struct dpll clock;
@@ -8252,7 +8227,7 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 				 struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	uint32_t tmp;
 	bool ret;
@@ -8363,7 +8338,7 @@ out:
 
 static void ironlake_init_pch_refclk(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 	int i;
 	u32 val, final;
@@ -8537,16 +8512,16 @@ static void lpt_reset_fdi_mphy(struct drm_i915_private *dev_priv)
 	tmp |= FDI_MPHY_IOSFSB_RESET_CTL;
 	I915_WRITE(SOUTH_CHICKEN2, tmp);
 
-	if (wait_for_atomic_us(I915_READ(SOUTH_CHICKEN2) &
-			       FDI_MPHY_IOSFSB_RESET_STATUS, 100))
+	if (wait_for_us(I915_READ(SOUTH_CHICKEN2) &
+			FDI_MPHY_IOSFSB_RESET_STATUS, 100))
 		DRM_ERROR("FDI mPHY reset assert timeout\n");
 
 	tmp = I915_READ(SOUTH_CHICKEN2);
 	tmp &= ~FDI_MPHY_IOSFSB_RESET_CTL;
 	I915_WRITE(SOUTH_CHICKEN2, tmp);
 
-	if (wait_for_atomic_us((I915_READ(SOUTH_CHICKEN2) &
-				FDI_MPHY_IOSFSB_RESET_STATUS) == 0, 100))
+	if (wait_for_us((I915_READ(SOUTH_CHICKEN2) &
+			 FDI_MPHY_IOSFSB_RESET_STATUS) == 0, 100))
 		DRM_ERROR("FDI mPHY reset de-assert timeout\n");
 }
 
@@ -8634,7 +8609,7 @@ static void lpt_program_fdi_mphy(struct drm_i915_private *dev_priv)
 static void lpt_enable_clkout_dp(struct drm_device *dev, bool with_spread,
 				 bool with_fdi)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t reg, tmp;
 
 	if (WARN(with_fdi && !with_spread, "FDI requires downspread\n"))
@@ -8673,7 +8648,7 @@ static void lpt_enable_clkout_dp(struct drm_device *dev, bool with_spread,
 /* Sequence to disable CLKOUT_DP */
 static void lpt_disable_clkout_dp(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t reg, tmp;
 
 	mutex_lock(&dev_priv->sb_lock);
@@ -8794,7 +8769,7 @@ void intel_init_pch_refclk(struct drm_device *dev)
 
 static void ironlake_set_pipeconf(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	uint32_t val;
@@ -8836,7 +8811,7 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc)
 
 static void haswell_set_pipeconf(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	u32 val = 0;
@@ -8855,7 +8830,7 @@ static void haswell_set_pipeconf(struct drm_crtc *crtc)
 
 static void haswell_set_pipemisc(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
 	if (IS_BROADWELL(dev_priv) || INTEL_INFO(dev_priv)->gen >= 9) {
@@ -8908,37 +8883,13 @@ static void ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_atomic_state *state = crtc_state->base.state;
-	struct drm_connector *connector;
-	struct drm_connector_state *connector_state;
-	struct intel_encoder *encoder;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 dpll, fp, fp2;
-	int factor, i;
-	bool is_lvds = false, is_sdvo = false;
-
-	for_each_connector_in_state(state, connector, connector_state, i) {
-		if (connector_state->crtc != crtc_state->base.crtc)
-			continue;
-
-		encoder = to_intel_encoder(connector_state->best_encoder);
-
-		switch (encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			is_lvds = true;
-			break;
-		case INTEL_OUTPUT_SDVO:
-		case INTEL_OUTPUT_HDMI:
-			is_sdvo = true;
-			break;
-		default:
-			break;
-		}
-	}
+	int factor;
 
 	/* Enable autotuning of the PLL clock (if permissible) */
 	factor = 21;
-	if (is_lvds) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if ((intel_panel_use_ssc(dev_priv) &&
 		     dev_priv->vbt.lvds_ssc_freq == 100000) ||
 		    (HAS_PCH_IBX(dev) && intel_is_dual_link_lvds(dev)))
@@ -8962,7 +8913,7 @@ static void ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 
 	dpll = 0;
 
-	if (is_lvds)
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
@@ -8970,9 +8921,11 @@ static void ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	dpll |= (crtc_state->pixel_multiplier - 1)
 		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
 
-	if (is_sdvo)
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
 		dpll |= DPLL_SDVO_HIGH_SPEED;
-	if (crtc_state->has_dp_encoder)
+
+	if (intel_crtc_has_dp_encoder(crtc_state))
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
@@ -8995,7 +8948,8 @@ static void ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		break;
 	}
 
-	if (is_lvds && intel_panel_use_ssc(dev_priv))
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv))
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
 		dpll |= PLL_REF_INPUT_DREFCLK;
@@ -9011,7 +8965,7 @@ static int ironlake_crtc_compute_clock(struct intel_crtc *crtc,
 				       struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct dpll reduced_clock;
 	bool has_reduced_clock = false;
 	struct intel_shared_dpll *pll;
@@ -9027,7 +8981,7 @@ static int ironlake_crtc_compute_clock(struct intel_crtc *crtc,
 	if (!crtc_state->has_pch_encoder)
 		return 0;
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
 		if (intel_panel_use_ssc(dev_priv)) {
 			DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n",
 				      dev_priv->vbt.lvds_ssc_freq);
@@ -9066,7 +9020,7 @@ static int ironlake_crtc_compute_clock(struct intel_crtc *crtc,
 		return -EINVAL;
 	}
 
-	if (intel_pipe_will_have_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
 	    has_reduced_clock)
 		crtc->lowfreq_avail = true;
 
@@ -9077,7 +9031,7 @@ static void intel_pch_transcoder_get_m_n(struct intel_crtc *crtc,
 					 struct intel_link_m_n *m_n)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 
 	m_n->link_m = I915_READ(PCH_TRANS_LINK_M1(pipe));
@@ -9095,7 +9049,7 @@ static void intel_cpu_transcoder_get_m_n(struct intel_crtc *crtc,
 					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = crtc->pipe;
 
 	if (INTEL_INFO(dev)->gen >= 5) {
@@ -9153,7 +9107,7 @@ static void skylake_get_pfit_config(struct intel_crtc *crtc,
 				    struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_scaler_state *scaler_state = &pipe_config->scaler_state;
 	uint32_t ps_ctrl = 0;
 	int id = -1;
@@ -9184,7 +9138,7 @@ skylake_get_initial_plane_config(struct intel_crtc *crtc,
 				 struct intel_initial_plane_config *plane_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val, base, offset, stride_mult, tiling;
 	int pipe = crtc->pipe;
 	int fourcc, pixel_format;
@@ -9267,7 +9221,7 @@ static void ironlake_get_pfit_config(struct intel_crtc *crtc,
 				     struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t tmp;
 
 	tmp = I915_READ(PF_CTL(crtc->pipe));
@@ -9292,7 +9246,7 @@ ironlake_get_initial_plane_config(struct intel_crtc *crtc,
 				  struct intel_initial_plane_config *plane_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val, base, offset;
 	int pipe = crtc->pipe;
 	int fourcc, pixel_format;
@@ -9360,7 +9314,7 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 				     struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	uint32_t tmp;
 	bool ret;
@@ -9455,7 +9409,7 @@ out:
 
 static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_crtc *crtc;
 
 	for_each_intel_crtc(dev, crtc)
@@ -9489,7 +9443,7 @@ static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
 
 static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	if (IS_HASWELL(dev))
 		return I915_READ(D_COMP_HSW);
@@ -9499,7 +9453,7 @@ static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv)
 
 static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	if (IS_HASWELL(dev)) {
 		mutex_lock(&dev_priv->rps.hw_lock);
@@ -9534,8 +9488,8 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
 		val |= LCPLL_CD_SOURCE_FCLK;
 		I915_WRITE(LCPLL_CTL, val);
 
-		if (wait_for_atomic_us(I915_READ(LCPLL_CTL) &
-				       LCPLL_CD_SOURCE_FCLK_DONE, 1))
+		if (wait_for_us(I915_READ(LCPLL_CTL) &
+				LCPLL_CD_SOURCE_FCLK_DONE, 1))
 			DRM_ERROR("Switching to FCLK failed\n");
 
 		val = I915_READ(LCPLL_CTL);
@@ -9545,7 +9499,7 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
 	I915_WRITE(LCPLL_CTL, val);
 	POSTING_READ(LCPLL_CTL);
 
-	if (wait_for((I915_READ(LCPLL_CTL) & LCPLL_PLL_LOCK) == 0, 1))
+	if (intel_wait_for_register(dev_priv, LCPLL_CTL, LCPLL_PLL_LOCK, 0, 1))
 		DRM_ERROR("LCPLL still locked\n");
 
 	val = hsw_read_dcomp(dev_priv);
@@ -9600,7 +9554,9 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 	val &= ~LCPLL_PLL_DISABLE;
 	I915_WRITE(LCPLL_CTL, val);
 
-	if (wait_for(I915_READ(LCPLL_CTL) & LCPLL_PLL_LOCK, 5))
+	if (intel_wait_for_register(dev_priv,
+				    LCPLL_CTL, LCPLL_PLL_LOCK, LCPLL_PLL_LOCK,
+				    5))
 		DRM_ERROR("LCPLL not locked yet\n");
 
 	if (val & LCPLL_CD_SOURCE_FCLK) {
@@ -9608,13 +9564,13 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 		val &= ~LCPLL_CD_SOURCE_FCLK;
 		I915_WRITE(LCPLL_CTL, val);
 
-		if (wait_for_atomic_us((I915_READ(LCPLL_CTL) &
-					LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
+		if (wait_for_us((I915_READ(LCPLL_CTL) &
+				 LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
 			DRM_ERROR("Switching back to LCPLL failed\n");
 	}
 
 	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
-	intel_update_cdclk(dev_priv->dev);
+	intel_update_cdclk(&dev_priv->drm);
 }
 
 /*
@@ -9642,7 +9598,7 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
  */
 void hsw_enable_pc8(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	uint32_t val;
 
 	DRM_DEBUG_KMS("Enabling package C8+\n");
@@ -9659,7 +9615,7 @@ void hsw_enable_pc8(struct drm_i915_private *dev_priv)
 
 void hsw_disable_pc8(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	uint32_t val;
 
 	DRM_DEBUG_KMS("Disabling package C8+\n");
@@ -9688,7 +9644,7 @@ static void bxt_modeset_commit_cdclk(struct drm_atomic_state *old_state)
 static int ilk_max_pixel_rate(struct drm_atomic_state *state)
 {
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = state->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *cstate;
 	struct intel_crtc_state *crtc_state;
@@ -9724,7 +9680,7 @@ static int ilk_max_pixel_rate(struct drm_atomic_state *state)
 
 static void broadwell_set_cdclk(struct drm_device *dev, int cdclk)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t val, data;
 	int ret;
 
@@ -9893,10 +9849,7 @@ static void skl_modeset_commit_cdclk(struct drm_atomic_state *old_state)
 static int haswell_crtc_compute_clock(struct intel_crtc *crtc,
 				      struct intel_crtc_state *crtc_state)
 {
-	struct intel_encoder *intel_encoder =
-		intel_ddi_get_crtc_new_encoder(crtc_state);
-
-	if (intel_encoder->type != INTEL_OUTPUT_DSI) {
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI)) {
 		if (!intel_ddi_pll_select(crtc, crtc_state))
 			return -EINVAL;
 	}
@@ -10006,7 +9959,7 @@ static bool hsw_get_transcoder_state(struct intel_crtc *crtc,
 				     unsigned long *power_domain_mask)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
 
@@ -10057,14 +10010,12 @@ static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc,
 					 unsigned long *power_domain_mask)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	enum port port;
 	enum transcoder cpu_transcoder;
 	u32 tmp;
 
-	pipe_config->has_dsi_encoder = false;
-
 	for_each_port_masked(port, BIT(PORT_A) | BIT(PORT_C)) {
 		if (port == PORT_A)
 			cpu_transcoder = TRANSCODER_DSI_A;
@@ -10096,18 +10047,17 @@ static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc,
 			continue;
 
 		pipe_config->cpu_transcoder = cpu_transcoder;
-		pipe_config->has_dsi_encoder = true;
 		break;
 	}
 
-	return pipe_config->has_dsi_encoder;
+	return transcoder_is_dsi(pipe_config->cpu_transcoder);
 }
 
 static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
 				       struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_shared_dpll *pll;
 	enum port port;
 	uint32_t tmp;
@@ -10150,7 +10100,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 				    struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	unsigned long power_domain_mask;
 	bool active;
@@ -10164,18 +10114,16 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 
 	active = hsw_get_transcoder_state(crtc, pipe_config, &power_domain_mask);
 
-	if (IS_BROXTON(dev_priv)) {
-		bxt_get_dsi_transcoder_state(crtc, pipe_config,
-					     &power_domain_mask);
-		WARN_ON(active && pipe_config->has_dsi_encoder);
-		if (pipe_config->has_dsi_encoder)
-			active = true;
+	if (IS_BROXTON(dev_priv) &&
+	    bxt_get_dsi_transcoder_state(crtc, pipe_config, &power_domain_mask)) {
+		WARN_ON(active);
+		active = true;
 	}
 
 	if (!active)
 		goto out;
 
-	if (!pipe_config->has_dsi_encoder) {
+	if (!transcoder_is_dsi(pipe_config->cpu_transcoder)) {
 		haswell_get_ddi_port_state(crtc, pipe_config);
 		intel_get_pipe_timings(crtc, pipe_config);
 	}
@@ -10226,7 +10174,7 @@ static void i845_update_cursor(struct drm_crtc *crtc, u32 base,
 			       const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t cntl = 0, size = 0;
 
@@ -10289,7 +10237,7 @@ static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base,
 			       const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	uint32_t cntl = 0;
@@ -10337,7 +10285,7 @@ static void intel_crtc_update_cursor(struct drm_crtc *crtc,
 				     const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	u32 base = intel_crtc->cursor_addr;
@@ -10504,7 +10452,7 @@ mode_fits_in_fbdev(struct drm_device *dev,
 		   struct drm_display_mode *mode)
 {
 #ifdef CONFIG_DRM_FBDEV_EMULATION
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_gem_object *obj;
 	struct drm_framebuffer *fb;
 
@@ -10774,7 +10722,7 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 static int i9xx_pll_refclk(struct drm_device *dev,
 			   const struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 dpll = pipe_config->dpll_hw_state.dpll;
 
 	if ((dpll & PLL_REF_INPUT_MASK) == PLLB_REF_INPUT_SPREADSPECTRUMIN)
@@ -10792,7 +10740,7 @@ static void i9xx_crtc_clock_get(struct intel_crtc *crtc,
 				struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe = pipe_config->cpu_transcoder;
 	u32 dpll = pipe_config->dpll_hw_state.dpll;
 	u32 fp;
@@ -10918,7 +10866,7 @@ static void ironlake_pch_clock_get(struct intel_crtc *crtc,
 struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 					     struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	struct drm_display_mode *mode;
@@ -10970,31 +10918,6 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 	return mode;
 }
 
-void intel_mark_busy(struct drm_i915_private *dev_priv)
-{
-	if (dev_priv->mm.busy)
-		return;
-
-	intel_runtime_pm_get(dev_priv);
-	i915_update_gfx_val(dev_priv);
-	if (INTEL_GEN(dev_priv) >= 6)
-		gen6_rps_busy(dev_priv);
-	dev_priv->mm.busy = true;
-}
-
-void intel_mark_idle(struct drm_i915_private *dev_priv)
-{
-	if (!dev_priv->mm.busy)
-		return;
-
-	dev_priv->mm.busy = false;
-
-	if (INTEL_GEN(dev_priv) >= 6)
-		gen6_rps_idle(dev_priv);
-
-	intel_runtime_pm_put(dev_priv);
-}
-
 static void intel_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -11056,7 +10979,7 @@ static bool __pageflip_finished_cs(struct intel_crtc *crtc,
 				   struct intel_flip_work *work)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned reset_counter;
 
 	reset_counter = i915_reset_counter(&dev_priv->gpu_error);
@@ -11132,7 +11055,7 @@ static bool pageflip_finished(struct intel_crtc *crtc,
 
 void intel_finish_page_flip_cs(struct drm_i915_private *dev_priv, int pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_flip_work *work;
@@ -11159,7 +11082,7 @@ void intel_finish_page_flip_cs(struct drm_i915_private *dev_priv, int pipe)
 
 void intel_finish_page_flip_mmio(struct drm_i915_private *dev_priv, int pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_flip_work *work;
@@ -11267,7 +11190,7 @@ static int intel_gen4_queue_flip(struct drm_device *dev,
 				 uint32_t flags)
 {
 	struct intel_engine_cs *engine = req->engine;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t pf, pipesrc;
 	int ret;
@@ -11305,7 +11228,7 @@ static int intel_gen6_queue_flip(struct drm_device *dev,
 				 uint32_t flags)
 {
 	struct intel_engine_cs *engine = req->engine;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t pf, pipesrc;
 	int ret;
@@ -11464,7 +11387,7 @@ static void skl_do_mmio_flip(struct intel_crtc *intel_crtc,
 			     struct intel_flip_work *work)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_framebuffer *fb = intel_crtc->base.primary->fb;
 	const enum pipe pipe = intel_crtc->pipe;
 	u32 ctl, stride, tile_height;
@@ -11516,7 +11439,7 @@ static void ilk_do_mmio_flip(struct intel_crtc *intel_crtc,
 			     struct intel_flip_work *work)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_framebuffer *intel_fb =
 		to_intel_framebuffer(intel_crtc->base.primary->fb);
 	struct drm_i915_gem_object *obj = intel_fb->obj;
@@ -11593,7 +11516,7 @@ static bool __pageflip_stall_check_cs(struct drm_i915_private *dev_priv,
 	vblank = intel_crtc_get_vblank_counter(intel_crtc);
 	if (work->flip_ready_vblank == 0) {
 		if (work->flip_queued_req &&
-		    !i915_gem_request_completed(work->flip_queued_req, true))
+		    !i915_gem_request_completed(work->flip_queued_req))
 			return false;
 
 		work->flip_ready_vblank = vblank;
@@ -11618,7 +11541,7 @@ static bool __pageflip_stall_check_cs(struct drm_i915_private *dev_priv,
 
 void intel_check_page_flip(struct drm_i915_private *dev_priv, int pipe)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_flip_work *work;
@@ -11646,14 +11569,13 @@ void intel_check_page_flip(struct drm_i915_private *dev_priv, int pipe)
 	spin_unlock(&dev->event_lock);
 }
 
-__maybe_unused
 static int intel_crtc_page_flip(struct drm_crtc *crtc,
 				struct drm_framebuffer *fb,
 				struct drm_pending_vblank_event *event,
 				uint32_t page_flip_flags)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_framebuffer *old_fb = crtc->primary->fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -11949,8 +11871,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	struct drm_framebuffer *fb = plane_state->fb;
 	int ret;
 
-	if (crtc_state && INTEL_INFO(dev)->gen >= 9 &&
-	    plane->type != DRM_PLANE_TYPE_CURSOR) {
+	if (INTEL_GEN(dev) >= 9 && plane->type != DRM_PLANE_TYPE_CURSOR) {
 		ret = skl_update_scaler_plane(
 			to_intel_crtc_state(crtc_state),
 			to_intel_plane_state(plane_state));
@@ -12067,31 +11988,11 @@ static bool check_single_encoder_cloning(struct drm_atomic_state *state,
 	return true;
 }
 
-static bool check_encoder_cloning(struct drm_atomic_state *state,
-				  struct intel_crtc *crtc)
-{
-	struct intel_encoder *encoder;
-	struct drm_connector *connector;
-	struct drm_connector_state *connector_state;
-	int i;
-
-	for_each_connector_in_state(state, connector, connector_state, i) {
-		if (connector_state->crtc != &crtc->base)
-			continue;
-
-		encoder = to_intel_encoder(connector_state->best_encoder);
-		if (!check_single_encoder_cloning(state, crtc, encoder))
-			return false;
-	}
-
-	return true;
-}
-
 static int intel_crtc_atomic_check(struct drm_crtc *crtc,
 				   struct drm_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_crtc_state *pipe_config =
 		to_intel_crtc_state(crtc_state);
@@ -12099,11 +12000,6 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc,
 	int ret;
 	bool mode_changed = needs_modeset(crtc_state);
 
-	if (mode_changed && !check_encoder_cloning(state, intel_crtc)) {
-		DRM_DEBUG_KMS("rejecting invalid cloning configuration\n");
-		return -EINVAL;
-	}
-
 	if (mode_changed && !crtc_state->active)
 		pipe_config->update_wm_post = true;
 
@@ -12299,14 +12195,14 @@ static void intel_dump_pipe_config(struct intel_crtc *crtc,
 		      pipe_config->fdi_m_n.link_m, pipe_config->fdi_m_n.link_n,
 		      pipe_config->fdi_m_n.tu);
 	DRM_DEBUG_KMS("dp: %i, lanes: %i, gmch_m: %u, gmch_n: %u, link_m: %u, link_n: %u, tu: %u\n",
-		      pipe_config->has_dp_encoder,
+		      intel_crtc_has_dp_encoder(pipe_config),
 		      pipe_config->lane_count,
 		      pipe_config->dp_m_n.gmch_m, pipe_config->dp_m_n.gmch_n,
 		      pipe_config->dp_m_n.link_m, pipe_config->dp_m_n.link_n,
 		      pipe_config->dp_m_n.tu);
 
 	DRM_DEBUG_KMS("dp: %i, lanes: %i, gmch_m2: %u, gmch_n2: %u, link_m2: %u, link_n2: %u, tu2: %u\n",
-		      pipe_config->has_dp_encoder,
+		      intel_crtc_has_dp_encoder(pipe_config),
 		      pipe_config->lane_count,
 		      pipe_config->dp_m2_n2.gmch_m,
 		      pipe_config->dp_m2_n2.gmch_n,
@@ -12439,7 +12335,7 @@ static bool check_digital_port_conflicts(struct drm_atomic_state *state)
 		case INTEL_OUTPUT_UNKNOWN:
 			if (WARN_ON(!HAS_DDI(dev)))
 				break;
-		case INTEL_OUTPUT_DISPLAYPORT:
+		case INTEL_OUTPUT_DP:
 		case INTEL_OUTPUT_HDMI:
 		case INTEL_OUTPUT_EDP:
 			port_mask = 1 << enc_to_dig_port(&encoder->base)->port;
@@ -12536,6 +12432,24 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 			       &pipe_config->pipe_src_w,
 			       &pipe_config->pipe_src_h);
 
+	for_each_connector_in_state(state, connector, connector_state, i) {
+		if (connector_state->crtc != crtc)
+			continue;
+
+		encoder = to_intel_encoder(connector_state->best_encoder);
+
+		if (!check_single_encoder_cloning(state, to_intel_crtc(crtc), encoder)) {
+			DRM_DEBUG_KMS("rejecting invalid cloning configuration\n");
+			goto fail;
+		}
+
+		/*
+		 * Determine output_types before calling the .compute_config()
+		 * hooks so that the hooks can use this information safely.
+		 */
+		pipe_config->output_types |= 1 << encoder->type;
+	}
+
 encoder_retry:
 	/* Ensure the port clock defaults are reset when retrying. */
 	pipe_config->port_clock = 0;
@@ -12821,7 +12735,6 @@ intel_pipe_config_compare(struct drm_device *dev,
 	PIPE_CONF_CHECK_I(fdi_lanes);
 	PIPE_CONF_CHECK_M_N(fdi_m_n);
 
-	PIPE_CONF_CHECK_I(has_dp_encoder);
 	PIPE_CONF_CHECK_I(lane_count);
 	PIPE_CONF_CHECK_X(lane_lat_optim_mask);
 
@@ -12833,7 +12746,7 @@ intel_pipe_config_compare(struct drm_device *dev,
 	} else
 		PIPE_CONF_CHECK_M_N_ALT(dp_m_n, dp_m2_n2);
 
-	PIPE_CONF_CHECK_I(has_dsi_encoder);
+	PIPE_CONF_CHECK_X(output_types);
 
 	PIPE_CONF_CHECK_I(base.adjusted_mode.crtc_hdisplay);
 	PIPE_CONF_CHECK_I(base.adjusted_mode.crtc_htotal);
@@ -12952,7 +12865,7 @@ static void verify_wm_state(struct drm_crtc *crtc,
 			    struct drm_crtc_state *new_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct skl_ddb_allocation hw_ddb, *sw_ddb;
 	struct skl_ddb_entry *hw_entry, *sw_entry;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -13058,7 +12971,7 @@ verify_crtc_state(struct drm_crtc *crtc,
 		  struct drm_crtc_state *new_crtc_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_crtc_state *pipe_config, *sw_config;
@@ -13101,8 +13014,10 @@ verify_crtc_state(struct drm_crtc *crtc,
 				"Encoder connected to wrong pipe %c\n",
 				pipe_name(pipe));
 
-		if (active)
+		if (active) {
+			pipe_config->output_types |= 1 << encoder->type;
 			encoder->get_config(encoder, pipe_config);
+		}
 	}
 
 	if (!new_crtc_state->active)
@@ -13181,7 +13096,7 @@ verify_shared_dpll_state(struct drm_device *dev, struct drm_crtc *crtc,
 			 struct drm_crtc_state *old_crtc_state,
 			 struct drm_crtc_state *new_crtc_state)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc_state *old_state = to_intel_crtc_state(old_crtc_state);
 	struct intel_crtc_state *new_state = to_intel_crtc_state(new_crtc_state);
 
@@ -13220,7 +13135,7 @@ intel_modeset_verify_crtc(struct drm_crtc *crtc,
 static void
 verify_disabled_dpll_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	for (i = 0; i < dev_priv->num_shared_dpll; i++)
@@ -13267,7 +13182,7 @@ static void update_scanline_offset(struct intel_crtc *crtc)
 
 		crtc->scanline_offset = vtotal - 1;
 	} else if (HAS_DDI(dev) &&
-		   intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI)) {
+		   intel_crtc_has_type(crtc->config, INTEL_OUTPUT_HDMI)) {
 		crtc->scanline_offset = 2;
 	} else
 		crtc->scanline_offset = 1;
@@ -13402,7 +13317,7 @@ static int intel_modeset_all_pipes(struct drm_atomic_state *state)
 static int intel_modeset_checks(struct drm_atomic_state *state)
 {
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = state->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *crtc_state;
 	int ret = 0, i;
@@ -13568,7 +13483,7 @@ static int intel_atomic_prepare_commit(struct drm_device *dev,
 				       struct drm_atomic_state *state,
 				       bool nonblock)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_plane_state *plane_state;
 	struct drm_crtc_state *crtc_state;
 	struct drm_plane *plane;
@@ -13697,7 +13612,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 {
 	struct drm_device *dev = state->dev;
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc_state *old_crtc_state;
 	struct drm_crtc *crtc;
 	struct intel_crtc_state *intel_cstate;
@@ -13929,7 +13844,7 @@ static int intel_atomic_commit(struct drm_device *dev,
 			       bool nonblock)
 {
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
 	if (intel_state->modeset && nonblock) {
@@ -14008,7 +13923,7 @@ static const struct drm_crtc_funcs intel_crtc_funcs = {
 	.set_config = drm_atomic_helper_set_config,
 	.set_property = drm_atomic_helper_crtc_set_property,
 	.destroy = intel_crtc_destroy,
-	.page_flip = drm_atomic_helper_page_flip,
+	.page_flip = intel_crtc_page_flip,
 	.atomic_duplicate_state = intel_crtc_duplicate_state,
 	.atomic_destroy_state = intel_crtc_destroy_state,
 };
@@ -14136,15 +14051,11 @@ int
 skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state)
 {
 	int max_scale;
-	struct drm_device *dev;
-	struct drm_i915_private *dev_priv;
 	int crtc_clock, cdclk;
 
 	if (!intel_crtc || !crtc_state->base.enable)
 		return DRM_PLANE_HELPER_NO_SCALING;
 
-	dev = intel_crtc->base.dev;
-	dev_priv = dev->dev_private;
 	crtc_clock = crtc_state->base.adjusted_mode.crtc_clock;
 	cdclk = to_intel_atomic_state(crtc_state->base.state)->cdclk;
 
@@ -14534,7 +14445,7 @@ static void skl_init_scalers(struct drm_device *dev, struct intel_crtc *intel_cr
 
 static void intel_crtc_init(struct drm_device *dev, int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc;
 	struct intel_crtc_state *crtc_state = NULL;
 	struct drm_plane *primary = NULL;
@@ -14633,11 +14544,8 @@ int intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data,
 	struct intel_crtc *crtc;
 
 	drmmode_crtc = drm_crtc_find(dev, pipe_from_crtc_id->crtc_id);
-
-	if (!drmmode_crtc) {
-		DRM_ERROR("no such CRTC id\n");
+	if (!drmmode_crtc)
 		return -ENOENT;
-	}
 
 	crtc = to_intel_crtc(drmmode_crtc);
 	pipe_from_crtc_id->pipe = crtc->pipe;
@@ -14664,7 +14572,7 @@ static int intel_encoder_clones(struct intel_encoder *encoder)
 
 static bool has_edp_a(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!IS_MOBILE(dev))
 		return false;
@@ -14680,7 +14588,7 @@ static bool has_edp_a(struct drm_device *dev)
 
 static bool intel_crt_present(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_INFO(dev)->gen >= 9)
 		return false;
@@ -14706,10 +14614,15 @@ static bool intel_crt_present(struct drm_device *dev)
 
 static void intel_setup_outputs(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
 
+	/*
+	 * intel_edp_init_connector() depends on this completing first, to
+	 * prevent the registeration of both eDP and LVDS and the incorrect
+	 * sharing of the PPS.
+	 */
 	intel_lvds_init(dev);
 
 	if (intel_crt_present(dev))
@@ -15354,7 +15267,7 @@ void intel_init_display_hooks(struct drm_i915_private *dev_priv)
  */
 static void quirk_pipea_force(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->quirks |= QUIRK_PIPEA_FORCE;
 	DRM_INFO("applying pipe a force quirk\n");
@@ -15362,7 +15275,7 @@ static void quirk_pipea_force(struct drm_device *dev)
 
 static void quirk_pipeb_force(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->quirks |= QUIRK_PIPEB_FORCE;
 	DRM_INFO("applying pipe b force quirk\n");
@@ -15373,7 +15286,7 @@ static void quirk_pipeb_force(struct drm_device *dev)
  */
 static void quirk_ssc_force_disable(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	dev_priv->quirks |= QUIRK_LVDS_SSC_DISABLE;
 	DRM_INFO("applying lvds SSC disable quirk\n");
 }
@@ -15384,7 +15297,7 @@ static void quirk_ssc_force_disable(struct drm_device *dev)
  */
 static void quirk_invert_brightness(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	dev_priv->quirks |= QUIRK_INVERT_BRIGHTNESS;
 	DRM_INFO("applying inverted panel brightness quirk\n");
 }
@@ -15392,7 +15305,7 @@ static void quirk_invert_brightness(struct drm_device *dev)
 /* Some VBT's incorrectly indicate no backlight is present */
 static void quirk_backlight_present(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	dev_priv->quirks |= QUIRK_BACKLIGHT_PRESENT;
 	DRM_INFO("applying backlight present quirk\n");
 }
@@ -15518,7 +15431,7 @@ static void intel_init_quirks(struct drm_device *dev)
 /* Disable the VGA plane that we never use */
 static void i915_disable_vga(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u8 sr1;
 	i915_reg_t vga_reg = i915_vgacntrl_reg(dev);
 
@@ -15536,7 +15449,7 @@ static void i915_disable_vga(struct drm_device *dev)
 
 void intel_modeset_init_hw(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_update_cdclk(dev);
 
@@ -15784,7 +15697,7 @@ static bool
 intel_check_plane_mapping(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val;
 
 	if (INTEL_INFO(dev)->num_pipes == 1)
@@ -15824,7 +15737,7 @@ static bool intel_encoder_has_connectors(struct intel_encoder *encoder)
 static void intel_sanitize_crtc(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 
 	/* Clear any frame start delays used for debugging left by the BIOS */
@@ -15949,7 +15862,7 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder)
 
 void i915_redisable_vga_power_on(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t vga_reg = i915_vgacntrl_reg(dev);
 
 	if (!(I915_READ(vga_reg) & VGA_DISP_DISABLE)) {
@@ -15960,7 +15873,7 @@ void i915_redisable_vga_power_on(struct drm_device *dev)
 
 void i915_redisable_vga(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* This function can be called both from intel_modeset_setup_hw_state or
 	 * at a very early point in our resume sequence, where the power well
@@ -16000,7 +15913,7 @@ static void readout_plane_state(struct intel_crtc *crtc)
 
 static void intel_modeset_readout_hw_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
@@ -16069,6 +15982,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 		if (encoder->get_hw_state(encoder, &pipe)) {
 			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 			encoder->base.crtc = &crtc->base;
+			crtc->config->output_types |= 1 << encoder->type;
 			encoder->get_config(encoder, crtc->config);
 		} else {
 			encoder->base.crtc = NULL;
@@ -16153,7 +16067,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 static void
 intel_modeset_setup_hw_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
@@ -16309,8 +16223,21 @@ void intel_modeset_gem_init(struct drm_device *dev)
 			c->state->plane_mask &= ~(1 << drm_plane_index(c->primary));
 		}
 	}
+}
+
+int intel_connector_register(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	int ret;
+
+	ret = intel_backlight_device_register(intel_connector);
+	if (ret)
+		goto err;
 
-	intel_backlight_register(dev);
+	return 0;
+
+err:
+	return ret;
 }
 
 void intel_connector_unregister(struct drm_connector *connector)
@@ -16323,7 +16250,7 @@ void intel_connector_unregister(struct drm_connector *connector)
 
 void intel_modeset_cleanup(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_disable_gt_powersave(dev_priv);
 
@@ -16347,8 +16274,6 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	/* flush any delayed tasks or pending work */
 	flush_scheduled_work();
 
-	drm_connector_unregister_all(dev);
-
 	drm_mode_config_cleanup(dev);
 
 	intel_cleanup_overlay(dev_priv);
@@ -16371,7 +16296,7 @@ void intel_connector_attach_encoder(struct intel_connector *connector,
  */
 int intel_modeset_vga_set_state(struct drm_device *dev, bool state)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	unsigned reg = INTEL_INFO(dev)->gen >= 6 ? SNB_GMCH_CTRL : INTEL_GMCH_CTRL;
 	u16 gmch_ctrl;
 
@@ -16527,7 +16452,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	if (!error)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 0b84f8e5df50..0c5ba3410a1e 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -262,7 +262,7 @@ static void pps_lock(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *encoder = &intel_dig_port->base;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 
 	/*
@@ -280,7 +280,7 @@ static void pps_unlock(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *encoder = &intel_dig_port->base;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 
 	mutex_unlock(&dev_priv->pps_mutex);
@@ -294,7 +294,7 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe = intel_dp->pps_pipe;
 	bool pll_enabled, release_cl_override = false;
 	enum dpio_phy phy = DPIO_PHY(pipe);
@@ -368,7 +368,7 @@ vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
 	enum pipe pipe;
@@ -426,6 +426,37 @@ vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
 	return intel_dp->pps_pipe;
 }
 
+static int
+bxt_power_sequencer_idx(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* We should never land here with regular DP ports */
+	WARN_ON(!is_edp(intel_dp));
+
+	/*
+	 * TODO: BXT has 2 PPS instances. The correct port->PPS instance
+	 * mapping needs to be retrieved from VBT, for now just hard-code to
+	 * use instance #0 always.
+	 */
+	if (!intel_dp->pps_reset)
+		return 0;
+
+	intel_dp->pps_reset = false;
+
+	/*
+	 * Only the HW needs to be reprogrammed, the SW state is fixed and
+	 * has been setup during connector init.
+	 */
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+
+	return 0;
+}
+
 typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
 			       enum pipe pipe);
 
@@ -475,7 +506,7 @@ vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
@@ -507,12 +538,13 @@ vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
 	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
 }
 
-void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv)
+void intel_power_sequencer_reset(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_encoder *encoder;
 
-	if (WARN_ON(!IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev)))
+	if (WARN_ON(!IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev) &&
+		    !IS_BROXTON(dev)))
 		return;
 
 	/*
@@ -532,34 +564,71 @@ void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv)
 			continue;
 
 		intel_dp = enc_to_intel_dp(&encoder->base);
-		intel_dp->pps_pipe = INVALID_PIPE;
+		if (IS_BROXTON(dev))
+			intel_dp->pps_reset = true;
+		else
+			intel_dp->pps_pipe = INVALID_PIPE;
+	}
+}
+
+struct pps_registers {
+	i915_reg_t pp_ctrl;
+	i915_reg_t pp_stat;
+	i915_reg_t pp_on;
+	i915_reg_t pp_off;
+	i915_reg_t pp_div;
+};
+
+static void intel_pps_get_registers(struct drm_i915_private *dev_priv,
+				    struct intel_dp *intel_dp,
+				    struct pps_registers *regs)
+{
+	memset(regs, 0, sizeof(*regs));
+
+	if (IS_BROXTON(dev_priv)) {
+		int idx = bxt_power_sequencer_idx(intel_dp);
+
+		regs->pp_ctrl = BXT_PP_CONTROL(idx);
+		regs->pp_stat = BXT_PP_STATUS(idx);
+		regs->pp_on = BXT_PP_ON_DELAYS(idx);
+		regs->pp_off = BXT_PP_OFF_DELAYS(idx);
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		regs->pp_ctrl = PCH_PP_CONTROL;
+		regs->pp_stat = PCH_PP_STATUS;
+		regs->pp_on = PCH_PP_ON_DELAYS;
+		regs->pp_off = PCH_PP_OFF_DELAYS;
+		regs->pp_div = PCH_PP_DIVISOR;
+	} else {
+		enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
+
+		regs->pp_ctrl = VLV_PIPE_PP_CONTROL(pipe);
+		regs->pp_stat = VLV_PIPE_PP_STATUS(pipe);
+		regs->pp_on = VLV_PIPE_PP_ON_DELAYS(pipe);
+		regs->pp_off = VLV_PIPE_PP_OFF_DELAYS(pipe);
+		regs->pp_div = VLV_PIPE_PP_DIVISOR(pipe);
 	}
 }
 
 static i915_reg_t
 _pp_ctrl_reg(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct pps_registers regs;
 
-	if (IS_BROXTON(dev))
-		return BXT_PP_CONTROL(0);
-	else if (HAS_PCH_SPLIT(dev))
-		return PCH_PP_CONTROL;
-	else
-		return VLV_PIPE_PP_CONTROL(vlv_power_sequencer_pipe(intel_dp));
+	intel_pps_get_registers(to_i915(intel_dp_to_dev(intel_dp)), intel_dp,
+				&regs);
+
+	return regs.pp_ctrl;
 }
 
 static i915_reg_t
 _pp_stat_reg(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct pps_registers regs;
 
-	if (IS_BROXTON(dev))
-		return BXT_PP_STATUS(0);
-	else if (HAS_PCH_SPLIT(dev))
-		return PCH_PP_STATUS;
-	else
-		return VLV_PIPE_PP_STATUS(vlv_power_sequencer_pipe(intel_dp));
+	intel_pps_get_registers(to_i915(intel_dp_to_dev(intel_dp)), intel_dp,
+				&regs);
+
+	return regs.pp_stat;
 }
 
 /* Reboot notifier handler to shutdown panel power to guarantee T12 timing
@@ -570,7 +639,7 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 	struct intel_dp *intel_dp = container_of(this, typeof(* intel_dp),
 						 edp_notifier);
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!is_edp(intel_dp) || code != SYS_RESTART)
 		return 0;
@@ -601,7 +670,7 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 static bool edp_have_panel_power(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
@@ -615,7 +684,7 @@ static bool edp_have_panel_power(struct intel_dp *intel_dp)
 static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
@@ -630,7 +699,7 @@ static void
 intel_dp_check_edp(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (!is_edp(intel_dp))
 		return;
@@ -648,7 +717,7 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp, bool has_aux_irq)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg;
 	uint32_t status;
 	bool done;
@@ -658,7 +727,7 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp, bool has_aux_irq)
 		done = wait_event_timeout(dev_priv->gmbus_wait_queue, C,
 					  msecs_to_jiffies_timeout(10));
 	else
-		done = wait_for_atomic(C, 10) == 0;
+		done = wait_for(C, 10) == 0;
 	if (!done)
 		DRM_ERROR("dp aux hw did not signal timeout (has irq: %i)!\n",
 			  has_aux_irq);
@@ -781,7 +850,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg;
 	uint32_t aux_clock_divider;
 	int i, ret, recv_bytes;
@@ -1180,35 +1249,18 @@ intel_dp_aux_fini(struct intel_dp *intel_dp)
 	kfree(intel_dp->aux.name);
 }
 
-static int
+static void
 intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	enum port port = intel_dig_port->port;
-	int ret;
 
 	intel_aux_reg_init(intel_dp);
+	drm_dp_aux_init(&intel_dp->aux);
 
+	/* Failure to allocate our preferred name is not critical */
 	intel_dp->aux.name = kasprintf(GFP_KERNEL, "DPDDC-%c", port_name(port));
-	if (!intel_dp->aux.name)
-		return -ENOMEM;
-
-	intel_dp->aux.dev = connector->base.kdev;
 	intel_dp->aux.transfer = intel_dp_aux_transfer;
-
-	DRM_DEBUG_KMS("registering %s bus for %s\n",
-		      intel_dp->aux.name,
-		      connector->base.kdev->kobj.name);
-
-	ret = drm_dp_aux_register(&intel_dp->aux);
-	if (ret < 0) {
-		DRM_ERROR("drm_dp_aux_register() for %s failed (%d)\n",
-			  intel_dp->aux.name, ret);
-		kfree(intel_dp->aux.name);
-		return ret;
-	}
-
-	return 0;
 }
 
 static int
@@ -1421,7 +1473,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 			struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode;
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	enum port port = dp_to_dig_port(intel_dp)->port;
@@ -1449,7 +1501,6 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && port != PORT_A)
 		pipe_config->has_pch_encoder = true;
 
-	pipe_config->has_dp_encoder = true;
 	pipe_config->has_drrs = false;
 	pipe_config->has_audio = intel_dp->has_audio && port != PORT_A;
 
@@ -1605,7 +1656,7 @@ void intel_dp_set_link_params(struct intel_dp *intel_dp,
 static void intel_dp_prepare(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
@@ -1693,16 +1744,21 @@ static void intel_dp_prepare(struct intel_encoder *encoder)
 #define IDLE_CYCLE_MASK		(PP_ON | PP_SEQUENCE_MASK | PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK)
 #define IDLE_CYCLE_VALUE	(0     | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
 
+static void intel_pps_verify_state(struct drm_i915_private *dev_priv,
+				   struct intel_dp *intel_dp);
+
 static void wait_panel_status(struct intel_dp *intel_dp,
 				       u32 mask,
 				       u32 value)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t pp_stat_reg, pp_ctrl_reg;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
+	intel_pps_verify_state(dev_priv, intel_dp);
+
 	pp_stat_reg = _pp_stat_reg(intel_dp);
 	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
 
@@ -1711,8 +1767,9 @@ static void wait_panel_status(struct intel_dp *intel_dp,
 			I915_READ(pp_stat_reg),
 			I915_READ(pp_ctrl_reg));
 
-	if (_wait_for((I915_READ(pp_stat_reg) & mask) == value,
-		      5 * USEC_PER_SEC, 10 * USEC_PER_MSEC))
+	if (intel_wait_for_register(dev_priv,
+				    pp_stat_reg, mask, value,
+				    5000))
 		DRM_ERROR("Panel status timeout: status %08x control %08x\n",
 				I915_READ(pp_stat_reg),
 				I915_READ(pp_ctrl_reg));
@@ -1772,7 +1829,7 @@ static void edp_wait_backlight_off(struct intel_dp *intel_dp)
 static  u32 ironlake_get_pp_control(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 control;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
@@ -1795,7 +1852,7 @@ static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	u32 pp;
 	i915_reg_t pp_stat_reg, pp_ctrl_reg;
@@ -1868,7 +1925,7 @@ void intel_edp_panel_vdd_on(struct intel_dp *intel_dp)
 static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *intel_dig_port =
 		dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
@@ -1937,8 +1994,7 @@ static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
  */
 static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 {
-	struct drm_i915_private *dev_priv =
-		intel_dp_to_dev(intel_dp)->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(intel_dp_to_dev(intel_dp));
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
@@ -1959,7 +2015,7 @@ static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 static void edp_panel_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pp;
 	i915_reg_t pp_ctrl_reg;
 
@@ -2020,7 +2076,7 @@ static void edp_panel_off(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	u32 pp;
 	i915_reg_t pp_ctrl_reg;
@@ -2072,7 +2128,7 @@ static void _intel_edp_backlight_on(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pp;
 	i915_reg_t pp_ctrl_reg;
 
@@ -2113,7 +2169,7 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 static void _intel_edp_backlight_off(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pp;
 	i915_reg_t pp_ctrl_reg;
 
@@ -2229,7 +2285,7 @@ static void ironlake_edp_pll_on(struct intel_dp *intel_dp)
 	 * 2. Program DP PLL enable
 	 */
 	if (IS_GEN5(dev_priv))
-		intel_wait_for_vblank_if_active(dev_priv->dev, !crtc->pipe);
+		intel_wait_for_vblank_if_active(&dev_priv->drm, !crtc->pipe);
 
 	intel_dp->DP |= DP_PLL_ENABLE;
 
@@ -2294,7 +2350,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	bool ret;
@@ -2347,7 +2403,7 @@ static void intel_dp_get_config(struct intel_encoder *encoder,
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	u32 tmp, flags = 0;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 
@@ -2385,8 +2441,6 @@ static void intel_dp_get_config(struct intel_encoder *encoder,
 	    !IS_CHERRYVIEW(dev) && tmp & DP_COLOR_RANGE_16_235)
 		pipe_config->limited_color_range = true;
 
-	pipe_config->has_dp_encoder = true;
-
 	pipe_config->lane_count =
 		((tmp & DP_PORT_WIDTH_MASK) >> DP_PORT_WIDTH_SHIFT) + 1;
 
@@ -2471,7 +2525,7 @@ static void chv_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_dp_link_down(intel_dp);
 
@@ -2490,7 +2544,7 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp,
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 
 	if (HAS_DDI(dev)) {
@@ -2570,7 +2624,7 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp,
 static void intel_dp_enable_port(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc =
 		to_intel_crtc(dp_to_dig_port(intel_dp)->base.base.crtc);
 
@@ -2599,7 +2653,7 @@ static void intel_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	uint32_t dp_reg = I915_READ(intel_dp->output_reg);
 	enum pipe pipe = crtc->pipe;
@@ -2672,7 +2726,7 @@ static void g4x_pre_enable_dp(struct intel_encoder *encoder)
 static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct drm_i915_private *dev_priv = intel_dig_port->base.base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev);
 	enum pipe pipe = intel_dp->pps_pipe;
 	i915_reg_t pp_on_reg = VLV_PIPE_PP_ON_DELAYS(pipe);
 
@@ -2698,7 +2752,7 @@ static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
 static void vlv_steal_power_sequencer(struct drm_device *dev,
 				      enum pipe pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
@@ -2736,7 +2790,7 @@ static void vlv_init_panel_power_sequencer(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *encoder = &intel_dig_port->base;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
@@ -2824,7 +2878,7 @@ uint8_t
 intel_dp_voltage_max(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
 	if (IS_BROXTON(dev))
@@ -3242,7 +3296,7 @@ void intel_dp_set_idle_link_train(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 	uint32_t val;
 
@@ -3264,8 +3318,10 @@ void intel_dp_set_idle_link_train(struct intel_dp *intel_dp)
 	if (port == PORT_A)
 		return;
 
-	if (wait_for((I915_READ(DP_TP_STATUS(port)) & DP_TP_STATUS_IDLE_DONE),
-		     1))
+	if (intel_wait_for_register(dev_priv,DP_TP_STATUS(port),
+				    DP_TP_STATUS_IDLE_DONE,
+				    DP_TP_STATUS_IDLE_DONE,
+				    1))
 		DRM_ERROR("Timed out waiting for DP idle patterns\n");
 }
 
@@ -3276,7 +3332,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 	struct intel_crtc *crtc = to_intel_crtc(intel_dig_port->base.base.crtc);
 	enum port port = intel_dig_port->port;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t DP = intel_dp->DP;
 
 	if (WARN_ON(HAS_DDI(dev)))
@@ -3328,7 +3384,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 		I915_WRITE(intel_dp->output_reg, DP);
 		POSTING_READ(intel_dp->output_reg);
 
-		intel_wait_for_vblank_if_active(dev_priv->dev, PIPE_A);
+		intel_wait_for_vblank_if_active(&dev_priv->drm, PIPE_A);
 		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 	}
@@ -3343,7 +3399,7 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (drm_dp_dpcd_read(&intel_dp->aux, 0x000, intel_dp->dpcd,
 			     sizeof(intel_dp->dpcd)) < 0)
@@ -4194,7 +4250,7 @@ intel_dp_long_pulse(struct intel_connector *intel_connector)
 	}
 
 	if (intel_encoder->type != INTEL_OUTPUT_EDP)
-		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+		intel_encoder->type = INTEL_OUTPUT_DP;
 
 	intel_dp_probe_oui(intel_dp);
 
@@ -4270,7 +4326,7 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 		/* MST devices are disconnected from a monitor POV */
 		intel_dp_unset_edid(intel_dp);
 		if (intel_encoder->type != INTEL_OUTPUT_EDP)
-			intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+			intel_encoder->type = INTEL_OUTPUT_DP;
 		return connector_status_disconnected;
 	}
 
@@ -4309,7 +4365,7 @@ intel_dp_force(struct drm_connector *connector)
 	intel_display_power_put(dev_priv, power_domain);
 
 	if (intel_encoder->type != INTEL_OUTPUT_EDP)
-		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+		intel_encoder->type = INTEL_OUTPUT_DP;
 }
 
 static int intel_dp_get_modes(struct drm_connector *connector)
@@ -4358,7 +4414,7 @@ intel_dp_set_property(struct drm_connector *connector,
 		      struct drm_property *property,
 		      uint64_t val)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct intel_encoder *intel_encoder = intel_attached_encoder(connector);
 	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
@@ -4446,6 +4502,25 @@ done:
 	return 0;
 }
 
+static int
+intel_dp_connector_register(struct drm_connector *connector)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	int ret;
+
+	ret = intel_connector_register(connector);
+	if (ret)
+		return ret;
+
+	i915_debugfs_connector_add(connector);
+
+	DRM_DEBUG_KMS("registering %s bus for %s\n",
+		      intel_dp->aux.name, connector->kdev->kobj.name);
+
+	intel_dp->aux.dev = connector->kdev;
+	return drm_dp_aux_register(&intel_dp->aux);
+}
+
 static void
 intel_dp_connector_unregister(struct drm_connector *connector)
 {
@@ -4521,7 +4596,7 @@ static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
@@ -4544,13 +4619,15 @@ static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
 
 void intel_dp_encoder_reset(struct drm_encoder *encoder)
 {
-	struct intel_dp *intel_dp;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	if (!HAS_DDI(dev_priv))
+		intel_dp->DP = I915_READ(intel_dp->output_reg);
 
 	if (to_intel_encoder(encoder)->type != INTEL_OUTPUT_EDP)
 		return;
 
-	intel_dp = enc_to_intel_dp(encoder);
-
 	pps_lock(intel_dp);
 
 	/*
@@ -4572,6 +4649,7 @@ static const struct drm_connector_funcs intel_dp_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_set_property,
 	.atomic_get_property = intel_connector_atomic_get_property,
+	.late_register = intel_dp_connector_register,
 	.early_unregister = intel_dp_connector_unregister,
 	.destroy = intel_dp_connector_destroy,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
@@ -4594,13 +4672,13 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum intel_display_power_domain power_domain;
 	enum irqreturn ret = IRQ_NONE;
 
 	if (intel_dig_port->base.type != INTEL_OUTPUT_EDP &&
 	    intel_dig_port->base.type != INTEL_OUTPUT_HDMI)
-		intel_dig_port->base.type = INTEL_OUTPUT_DISPLAYPORT;
+		intel_dig_port->base.type = INTEL_OUTPUT_DP;
 
 	if (long_hpd && intel_dig_port->base.type == INTEL_OUTPUT_EDP) {
 		/*
@@ -4622,9 +4700,6 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 	intel_display_power_get(dev_priv, power_domain);
 
 	if (long_hpd) {
-		/* indicate that we need to restart link training */
-		intel_dp->train_set_valid = false;
-
 		intel_dp_long_pulse(intel_dp->attached_connector);
 		if (intel_dp->is_mst)
 			ret = IRQ_HANDLED;
@@ -4665,7 +4740,7 @@ put_power:
 /* check the VBT to see whether the eDP is on another port */
 bool intel_dp_is_edp(struct drm_device *dev, enum port port)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/*
 	 * eDP not supported on g4x. so bail out early just
@@ -4707,82 +4782,93 @@ static void intel_dp_init_panel_power_timestamps(struct intel_dp *intel_dp)
 }
 
 static void
-intel_dp_init_panel_power_sequencer(struct drm_device *dev,
-				    struct intel_dp *intel_dp)
+intel_pps_readout_hw_state(struct drm_i915_private *dev_priv,
+			   struct intel_dp *intel_dp, struct edp_power_seq *seq)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edp_power_seq cur, vbt, spec,
-		*final = &intel_dp->pps_delays;
 	u32 pp_on, pp_off, pp_div = 0, pp_ctl = 0;
-	i915_reg_t pp_ctrl_reg, pp_on_reg, pp_off_reg, pp_div_reg;
-
-	lockdep_assert_held(&dev_priv->pps_mutex);
-
-	/* already initialized? */
-	if (final->t11_t12 != 0)
-		return;
-
-	if (IS_BROXTON(dev)) {
-		/*
-		 * TODO: BXT has 2 sets of PPS registers.
-		 * Correct Register for Broxton need to be identified
-		 * using VBT. hardcoding for now
-		 */
-		pp_ctrl_reg = BXT_PP_CONTROL(0);
-		pp_on_reg = BXT_PP_ON_DELAYS(0);
-		pp_off_reg = BXT_PP_OFF_DELAYS(0);
-	} else if (HAS_PCH_SPLIT(dev)) {
-		pp_ctrl_reg = PCH_PP_CONTROL;
-		pp_on_reg = PCH_PP_ON_DELAYS;
-		pp_off_reg = PCH_PP_OFF_DELAYS;
-		pp_div_reg = PCH_PP_DIVISOR;
-	} else {
-		enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
+	struct pps_registers regs;
 
-		pp_ctrl_reg = VLV_PIPE_PP_CONTROL(pipe);
-		pp_on_reg = VLV_PIPE_PP_ON_DELAYS(pipe);
-		pp_off_reg = VLV_PIPE_PP_OFF_DELAYS(pipe);
-		pp_div_reg = VLV_PIPE_PP_DIVISOR(pipe);
-	}
+	intel_pps_get_registers(dev_priv, intel_dp, &regs);
 
 	/* Workaround: Need to write PP_CONTROL with the unlock key as
 	 * the very first thing. */
 	pp_ctl = ironlake_get_pp_control(intel_dp);
 
-	pp_on = I915_READ(pp_on_reg);
-	pp_off = I915_READ(pp_off_reg);
-	if (!IS_BROXTON(dev)) {
-		I915_WRITE(pp_ctrl_reg, pp_ctl);
-		pp_div = I915_READ(pp_div_reg);
+	pp_on = I915_READ(regs.pp_on);
+	pp_off = I915_READ(regs.pp_off);
+	if (!IS_BROXTON(dev_priv)) {
+		I915_WRITE(regs.pp_ctrl, pp_ctl);
+		pp_div = I915_READ(regs.pp_div);
 	}
 
 	/* Pull timing values out of registers */
-	cur.t1_t3 = (pp_on & PANEL_POWER_UP_DELAY_MASK) >>
-		PANEL_POWER_UP_DELAY_SHIFT;
+	seq->t1_t3 = (pp_on & PANEL_POWER_UP_DELAY_MASK) >>
+		     PANEL_POWER_UP_DELAY_SHIFT;
 
-	cur.t8 = (pp_on & PANEL_LIGHT_ON_DELAY_MASK) >>
-		PANEL_LIGHT_ON_DELAY_SHIFT;
+	seq->t8 = (pp_on & PANEL_LIGHT_ON_DELAY_MASK) >>
+		  PANEL_LIGHT_ON_DELAY_SHIFT;
 
-	cur.t9 = (pp_off & PANEL_LIGHT_OFF_DELAY_MASK) >>
-		PANEL_LIGHT_OFF_DELAY_SHIFT;
+	seq->t9 = (pp_off & PANEL_LIGHT_OFF_DELAY_MASK) >>
+		  PANEL_LIGHT_OFF_DELAY_SHIFT;
 
-	cur.t10 = (pp_off & PANEL_POWER_DOWN_DELAY_MASK) >>
-		PANEL_POWER_DOWN_DELAY_SHIFT;
+	seq->t10 = (pp_off & PANEL_POWER_DOWN_DELAY_MASK) >>
+		   PANEL_POWER_DOWN_DELAY_SHIFT;
 
-	if (IS_BROXTON(dev)) {
+	if (IS_BROXTON(dev_priv)) {
 		u16 tmp = (pp_ctl & BXT_POWER_CYCLE_DELAY_MASK) >>
 			BXT_POWER_CYCLE_DELAY_SHIFT;
 		if (tmp > 0)
-			cur.t11_t12 = (tmp - 1) * 1000;
+			seq->t11_t12 = (tmp - 1) * 1000;
 		else
-			cur.t11_t12 = 0;
+			seq->t11_t12 = 0;
 	} else {
-		cur.t11_t12 = ((pp_div & PANEL_POWER_CYCLE_DELAY_MASK) >>
+		seq->t11_t12 = ((pp_div & PANEL_POWER_CYCLE_DELAY_MASK) >>
 		       PANEL_POWER_CYCLE_DELAY_SHIFT) * 1000;
 	}
+}
+
+static void
+intel_pps_dump_state(const char *state_name, const struct edp_power_seq *seq)
+{
+	DRM_DEBUG_KMS("%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
+		      state_name,
+		      seq->t1_t3, seq->t8, seq->t9, seq->t10, seq->t11_t12);
+}
+
+static void
+intel_pps_verify_state(struct drm_i915_private *dev_priv,
+		       struct intel_dp *intel_dp)
+{
+	struct edp_power_seq hw;
+	struct edp_power_seq *sw = &intel_dp->pps_delays;
+
+	intel_pps_readout_hw_state(dev_priv, intel_dp, &hw);
+
+	if (hw.t1_t3 != sw->t1_t3 || hw.t8 != sw->t8 || hw.t9 != sw->t9 ||
+	    hw.t10 != sw->t10 || hw.t11_t12 != sw->t11_t12) {
+		DRM_ERROR("PPS state mismatch\n");
+		intel_pps_dump_state("sw", sw);
+		intel_pps_dump_state("hw", &hw);
+	}
+}
+
+static void
+intel_dp_init_panel_power_sequencer(struct drm_device *dev,
+				    struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct edp_power_seq cur, vbt, spec,
+		*final = &intel_dp->pps_delays;
 
-	DRM_DEBUG_KMS("cur t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
-		      cur.t1_t3, cur.t8, cur.t9, cur.t10, cur.t11_t12);
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* already initialized? */
+	if (final->t11_t12 != 0)
+		return;
+
+	intel_pps_readout_hw_state(dev_priv, intel_dp, &cur);
+
+	intel_pps_dump_state("cur", &cur);
 
 	vbt = dev_priv->vbt.edp.pps;
 
@@ -4798,8 +4884,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	 * too. */
 	spec.t11_t12 = (510 + 100) * 10;
 
-	DRM_DEBUG_KMS("vbt t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
-		      vbt.t1_t3, vbt.t8, vbt.t9, vbt.t10, vbt.t11_t12);
+	intel_pps_dump_state("vbt", &vbt);
 
 	/* Use the max of the register settings and vbt. If both are
 	 * unset, fall back to the spec limits. */
@@ -4827,59 +4912,41 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 
 	DRM_DEBUG_KMS("backlight on delay %d, off delay %d\n",
 		      intel_dp->backlight_on_delay, intel_dp->backlight_off_delay);
+
+	/*
+	 * We override the HW backlight delays to 1 because we do manual waits
+	 * on them. For T8, even BSpec recommends doing it. For T9, if we
+	 * don't do this, we'll end up waiting for the backlight off delay
+	 * twice: once when we do the manual sleep, and once when we disable
+	 * the panel and wait for the PP_STATUS bit to become zero.
+	 */
+	final->t8 = 1;
+	final->t9 = 1;
 }
 
 static void
 intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 					      struct intel_dp *intel_dp)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 pp_on, pp_off, pp_div, port_sel = 0;
 	int div = dev_priv->rawclk_freq / 1000;
-	i915_reg_t pp_on_reg, pp_off_reg, pp_div_reg, pp_ctrl_reg;
+	struct pps_registers regs;
 	enum port port = dp_to_dig_port(intel_dp)->port;
 	const struct edp_power_seq *seq = &intel_dp->pps_delays;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
-	if (IS_BROXTON(dev)) {
-		/*
-		 * TODO: BXT has 2 sets of PPS registers.
-		 * Correct Register for Broxton need to be identified
-		 * using VBT. hardcoding for now
-		 */
-		pp_ctrl_reg = BXT_PP_CONTROL(0);
-		pp_on_reg = BXT_PP_ON_DELAYS(0);
-		pp_off_reg = BXT_PP_OFF_DELAYS(0);
-
-	} else if (HAS_PCH_SPLIT(dev)) {
-		pp_on_reg = PCH_PP_ON_DELAYS;
-		pp_off_reg = PCH_PP_OFF_DELAYS;
-		pp_div_reg = PCH_PP_DIVISOR;
-	} else {
-		enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
+	intel_pps_get_registers(dev_priv, intel_dp, &regs);
 
-		pp_on_reg = VLV_PIPE_PP_ON_DELAYS(pipe);
-		pp_off_reg = VLV_PIPE_PP_OFF_DELAYS(pipe);
-		pp_div_reg = VLV_PIPE_PP_DIVISOR(pipe);
-	}
-
-	/*
-	 * And finally store the new values in the power sequencer. The
-	 * backlight delays are set to 1 because we do manual waits on them. For
-	 * T8, even BSpec recommends doing it. For T9, if we don't do this,
-	 * we'll end up waiting for the backlight off delay twice: once when we
-	 * do the manual sleep, and once when we disable the panel and wait for
-	 * the PP_STATUS bit to become zero.
-	 */
 	pp_on = (seq->t1_t3 << PANEL_POWER_UP_DELAY_SHIFT) |
-		(1 << PANEL_LIGHT_ON_DELAY_SHIFT);
-	pp_off = (1 << PANEL_LIGHT_OFF_DELAY_SHIFT) |
+		(seq->t8 << PANEL_LIGHT_ON_DELAY_SHIFT);
+	pp_off = (seq->t9 << PANEL_LIGHT_OFF_DELAY_SHIFT) |
 		 (seq->t10 << PANEL_POWER_DOWN_DELAY_SHIFT);
 	/* Compute the divisor for the pp clock, simply match the Bspec
 	 * formula. */
 	if (IS_BROXTON(dev)) {
-		pp_div = I915_READ(pp_ctrl_reg);
+		pp_div = I915_READ(regs.pp_ctrl);
 		pp_div &= ~BXT_POWER_CYCLE_DELAY_MASK;
 		pp_div |= (DIV_ROUND_UP((seq->t11_t12 + 1), 1000)
 				<< BXT_POWER_CYCLE_DELAY_SHIFT);
@@ -4902,19 +4969,19 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 
 	pp_on |= port_sel;
 
-	I915_WRITE(pp_on_reg, pp_on);
-	I915_WRITE(pp_off_reg, pp_off);
+	I915_WRITE(regs.pp_on, pp_on);
+	I915_WRITE(regs.pp_off, pp_off);
 	if (IS_BROXTON(dev))
-		I915_WRITE(pp_ctrl_reg, pp_div);
+		I915_WRITE(regs.pp_ctrl, pp_div);
 	else
-		I915_WRITE(pp_div_reg, pp_div);
+		I915_WRITE(regs.pp_div, pp_div);
 
 	DRM_DEBUG_KMS("panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n",
-		      I915_READ(pp_on_reg),
-		      I915_READ(pp_off_reg),
+		      I915_READ(regs.pp_on),
+		      I915_READ(regs.pp_off),
 		      IS_BROXTON(dev) ?
-		      (I915_READ(pp_ctrl_reg) & BXT_POWER_CYCLE_DELAY_MASK) :
-		      I915_READ(pp_div_reg));
+		      (I915_READ(regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK) :
+		      I915_READ(regs.pp_div));
 }
 
 /**
@@ -4931,7 +4998,7 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
  */
 static void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder;
 	struct intel_digital_port *dig_port = NULL;
 	struct intel_dp *intel_dp = dev_priv->drrs.dp;
@@ -5030,7 +5097,7 @@ static void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 void intel_edp_drrs_enable(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -5062,7 +5129,7 @@ unlock:
 void intel_edp_drrs_disable(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -5077,9 +5144,9 @@ void intel_edp_drrs_disable(struct intel_dp *intel_dp)
 	}
 
 	if (dev_priv->drrs.refresh_rate_type == DRRS_LOW_RR)
-		intel_dp_set_drrs_state(dev_priv->dev,
-			intel_dp->attached_connector->panel.
-			fixed_mode->vrefresh);
+		intel_dp_set_drrs_state(&dev_priv->drm,
+					intel_dp->attached_connector->panel.
+					fixed_mode->vrefresh);
 
 	dev_priv->drrs.dp = NULL;
 	mutex_unlock(&dev_priv->drrs.mutex);
@@ -5109,9 +5176,9 @@ static void intel_edp_drrs_downclock_work(struct work_struct *work)
 		goto unlock;
 
 	if (dev_priv->drrs.refresh_rate_type != DRRS_LOW_RR)
-		intel_dp_set_drrs_state(dev_priv->dev,
-			intel_dp->attached_connector->panel.
-			downclock_mode->vrefresh);
+		intel_dp_set_drrs_state(&dev_priv->drm,
+					intel_dp->attached_connector->panel.
+					downclock_mode->vrefresh);
 
 unlock:
 	mutex_unlock(&dev_priv->drrs.mutex);
@@ -5130,7 +5197,7 @@ unlock:
 void intel_edp_drrs_invalidate(struct drm_device *dev,
 		unsigned frontbuffer_bits)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	enum pipe pipe;
 
@@ -5153,9 +5220,9 @@ void intel_edp_drrs_invalidate(struct drm_device *dev,
 
 	/* invalidate means busy screen hence upclock */
 	if (frontbuffer_bits && dev_priv->drrs.refresh_rate_type == DRRS_LOW_RR)
-		intel_dp_set_drrs_state(dev_priv->dev,
-				dev_priv->drrs.dp->attached_connector->panel.
-				fixed_mode->vrefresh);
+		intel_dp_set_drrs_state(&dev_priv->drm,
+					dev_priv->drrs.dp->attached_connector->panel.
+					fixed_mode->vrefresh);
 
 	mutex_unlock(&dev_priv->drrs.mutex);
 }
@@ -5175,7 +5242,7 @@ void intel_edp_drrs_invalidate(struct drm_device *dev,
 void intel_edp_drrs_flush(struct drm_device *dev,
 		unsigned frontbuffer_bits)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	enum pipe pipe;
 
@@ -5198,9 +5265,9 @@ void intel_edp_drrs_flush(struct drm_device *dev,
 
 	/* flush means busy screen hence upclock */
 	if (frontbuffer_bits && dev_priv->drrs.refresh_rate_type == DRRS_LOW_RR)
-		intel_dp_set_drrs_state(dev_priv->dev,
-				dev_priv->drrs.dp->attached_connector->panel.
-				fixed_mode->vrefresh);
+		intel_dp_set_drrs_state(&dev_priv->drm,
+					dev_priv->drrs.dp->attached_connector->panel.
+					fixed_mode->vrefresh);
 
 	/*
 	 * flush also means no more activity hence schedule downclock, if all
@@ -5268,7 +5335,7 @@ intel_dp_drrs_init(struct intel_connector *intel_connector,
 {
 	struct drm_connector *connector = &intel_connector->base;
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_display_mode *downclock_mode = NULL;
 
 	INIT_DELAYED_WORK(&dev_priv->drrs.work, intel_edp_drrs_downclock_work);
@@ -5306,7 +5373,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_display_mode *fixed_mode = NULL;
 	struct drm_display_mode *downclock_mode = NULL;
 	bool has_dpcd;
@@ -5317,8 +5384,32 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	if (!is_edp(intel_dp))
 		return true;
 
+	/*
+	 * On IBX/CPT we may get here with LVDS already registered. Since the
+	 * driver uses the only internal power sequencer available for both
+	 * eDP and LVDS bail out early in this case to prevent interfering
+	 * with an already powered-on LVDS power sequencer.
+	 */
+	if (intel_get_lvds_encoder(dev)) {
+		WARN_ON(!(HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)));
+		DRM_INFO("LVDS was detected, not registering eDP\n");
+
+		return false;
+	}
+
 	pps_lock(intel_dp);
+
+	intel_dp_init_panel_power_timestamps(intel_dp);
+
+	if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)) {
+		vlv_initial_power_sequencer_setup(intel_dp);
+	} else {
+		intel_dp_init_panel_power_sequencer(dev, intel_dp);
+		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+	}
+
 	intel_edp_panel_vdd_sanitize(intel_dp);
+
 	pps_unlock(intel_dp);
 
 	/* Cache DPCD and EDID for edp. */
@@ -5332,14 +5423,9 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	} else {
 		/* if this fails, presume the device is a ghost */
 		DRM_INFO("failed to retrieve link info, disabling eDP\n");
-		return false;
+		goto out_vdd_off;
 	}
 
-	/* We now know it's not a ghost, init power sequence regs. */
-	pps_lock(intel_dp);
-	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
-	pps_unlock(intel_dp);
-
 	mutex_lock(&dev->mode_config.mutex);
 	edid = drm_get_edid(connector, &intel_dp->aux.ddc);
 	if (edid) {
@@ -5407,6 +5493,18 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	intel_panel_setup_backlight(connector, pipe);
 
 	return true;
+
+out_vdd_off:
+	cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
+	/*
+	 * vdd might still be enabled do to the delayed vdd off.
+	 * Make sure vdd is actually turned off here.
+	 */
+	pps_lock(intel_dp);
+	edp_panel_vdd_off_sync(intel_dp);
+	pps_unlock(intel_dp);
+
+	return false;
 }
 
 bool
@@ -5417,9 +5515,9 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
-	int type, ret;
+	int type;
 
 	if (WARN(intel_dig_port->max_lanes < 1,
 		 "Not enough lanes (%d) for DP on port %c\n",
@@ -5478,11 +5576,12 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	connector->interlace_allowed = true;
 	connector->doublescan_allowed = 0;
 
+	intel_dp_aux_init(intel_dp, intel_connector);
+
 	INIT_DELAYED_WORK(&intel_dp->panel_vdd_work,
 			  edp_panel_vdd_work);
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
-	drm_connector_register(connector);
 
 	if (HAS_DDI(dev))
 		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
@@ -5512,22 +5611,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		BUG();
 	}
 
-	if (is_edp(intel_dp)) {
-		pps_lock(intel_dp);
-		intel_dp_init_panel_power_timestamps(intel_dp);
-		if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev))
-			vlv_initial_power_sequencer_setup(intel_dp);
-		else
-			intel_dp_init_panel_power_sequencer(dev, intel_dp);
-		pps_unlock(intel_dp);
-	}
-
-	ret = intel_dp_aux_init(intel_dp, intel_connector);
-	if (ret)
-		goto fail;
-
 	/* init MST on ports that can support it */
-	if (HAS_DP_MST(dev) &&
+	if (HAS_DP_MST(dev) && !is_edp(intel_dp) &&
 	    (port == PORT_B || port == PORT_C || port == PORT_D))
 		intel_dp_mst_encoder_init(intel_dig_port,
 					  intel_connector->base.base.id);
@@ -5549,22 +5634,9 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		I915_WRITE(PEG_BAND_GAP_DATA, (temp & ~0xf) | 0xd);
 	}
 
-	i915_debugfs_connector_add(connector);
-
 	return true;
 
 fail:
-	if (is_edp(intel_dp)) {
-		cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-		/*
-		 * vdd might still be enabled do to the delayed vdd off.
-		 * Make sure vdd is actually turned off here.
-		 */
-		pps_lock(intel_dp);
-		edp_panel_vdd_off_sync(intel_dp);
-		pps_unlock(intel_dp);
-	}
-	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 
 	return false;
@@ -5574,7 +5646,7 @@ bool intel_dp_init(struct drm_device *dev,
 		   i915_reg_t output_reg,
 		   enum port port)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
 	struct drm_encoder *encoder;
@@ -5622,7 +5694,7 @@ bool intel_dp_init(struct drm_device *dev,
 	intel_dig_port->dp.output_reg = output_reg;
 	intel_dig_port->max_lanes = 4;
 
-	intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+	intel_encoder->type = INTEL_OUTPUT_DP;
 	if (IS_CHERRYVIEW(dev)) {
 		if (port == PORT_D)
 			intel_encoder->crtc_mask = 1 << 2;
@@ -5652,43 +5724,35 @@ err_connector_alloc:
 
 void intel_dp_mst_suspend(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	/* disable MST */
 	for (i = 0; i < I915_MAX_PORTS; i++) {
 		struct intel_digital_port *intel_dig_port = dev_priv->hotplug.irq_port[i];
-		if (!intel_dig_port)
+
+		if (!intel_dig_port || !intel_dig_port->dp.can_mst)
 			continue;
 
-		if (intel_dig_port->base.type == INTEL_OUTPUT_DISPLAYPORT) {
-			if (!intel_dig_port->dp.can_mst)
-				continue;
-			if (intel_dig_port->dp.is_mst)
-				drm_dp_mst_topology_mgr_suspend(&intel_dig_port->dp.mst_mgr);
-		}
+		if (intel_dig_port->dp.is_mst)
+			drm_dp_mst_topology_mgr_suspend(&intel_dig_port->dp.mst_mgr);
 	}
 }
 
 void intel_dp_mst_resume(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int i;
 
 	for (i = 0; i < I915_MAX_PORTS; i++) {
 		struct intel_digital_port *intel_dig_port = dev_priv->hotplug.irq_port[i];
-		if (!intel_dig_port)
-			continue;
-		if (intel_dig_port->base.type == INTEL_OUTPUT_DISPLAYPORT) {
-			int ret;
+		int ret;
 
-			if (!intel_dig_port->dp.can_mst)
-				continue;
+		if (!intel_dig_port || !intel_dig_port->dp.can_mst)
+			continue;
 
-			ret = drm_dp_mst_topology_mgr_resume(&intel_dig_port->dp.mst_mgr);
-			if (ret != 0) {
-				intel_dp_check_mst_status(&intel_dig_port->dp);
-			}
-		}
+		ret = drm_dp_mst_topology_mgr_resume(&intel_dig_port->dp.mst_mgr);
+		if (ret)
+			intel_dp_check_mst_status(&intel_dig_port->dp);
 	}
 }
diff --git a/drivers/gpu/drm/i915/intel_dp_link_training.c b/drivers/gpu/drm/i915/intel_dp_link_training.c
index 0b8eefc2acc5..60fb39cd220b 100644
--- a/drivers/gpu/drm/i915/intel_dp_link_training.c
+++ b/drivers/gpu/drm/i915/intel_dp_link_training.c
@@ -85,8 +85,7 @@ static bool
 intel_dp_reset_link_train(struct intel_dp *intel_dp,
 			uint8_t dp_train_pat)
 {
-	if (!intel_dp->train_set_valid)
-		memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set));
+	memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set));
 	intel_dp_set_signal_levels(intel_dp);
 	return intel_dp_set_link_train(intel_dp, dp_train_pat);
 }
@@ -161,23 +160,6 @@ intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp)
 			break;
 		}
 
-		/*
-		 * if we used previously trained voltage and pre-emphasis values
-		 * and we don't get clock recovery, reset link training values
-		 */
-		if (intel_dp->train_set_valid) {
-			DRM_DEBUG_KMS("clock recovery not ok, reset");
-			/* clear the flag as we are not reusing train set */
-			intel_dp->train_set_valid = false;
-			if (!intel_dp_reset_link_train(intel_dp,
-						       DP_TRAINING_PATTERN_1 |
-						       DP_LINK_SCRAMBLING_DISABLE)) {
-				DRM_ERROR("failed to enable link training\n");
-				return;
-			}
-			continue;
-		}
-
 		/* Check to see if we've tried the max voltage */
 		for (i = 0; i < intel_dp->lane_count; i++)
 			if ((intel_dp->train_set[i] & DP_TRAIN_MAX_SWING_REACHED) == 0)
@@ -284,7 +266,6 @@ intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp)
 		/* Make sure clock is still ok */
 		if (!drm_dp_clock_recovery_ok(link_status,
 					      intel_dp->lane_count)) {
-			intel_dp->train_set_valid = false;
 			intel_dp_link_training_clock_recovery(intel_dp);
 			intel_dp_set_link_train(intel_dp,
 						training_pattern |
@@ -301,7 +282,6 @@ intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp)
 
 		/* Try 5 times, then try clock recovery if that fails */
 		if (tries > 5) {
-			intel_dp->train_set_valid = false;
 			intel_dp_link_training_clock_recovery(intel_dp);
 			intel_dp_set_link_train(intel_dp,
 						training_pattern |
@@ -322,10 +302,8 @@ intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp)
 
 	intel_dp_set_idle_link_train(intel_dp);
 
-	if (channel_eq) {
-		intel_dp->train_set_valid = true;
+	if (channel_eq)
 		DRM_DEBUG_KMS("Channel EQ done. DP Training successful\n");
-	}
 }
 
 void intel_dp_stop_link_train(struct intel_dp *intel_dp)
diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index 9646816604be..68a005d729e9 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -47,7 +47,6 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 
 	pipe_config->dp_encoder_is_mst = true;
 	pipe_config->has_pch_encoder = false;
-	pipe_config->has_dp_encoder = true;
 	bpp = 24;
 	/*
 	 * for MST we always configure max link bw - the spec doesn't
@@ -140,7 +139,7 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder)
 	struct intel_digital_port *intel_dig_port = intel_mst->primary;
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 	int ret;
 	uint32_t temp;
@@ -207,14 +206,17 @@ static void intel_mst_enable_dp(struct intel_encoder *encoder)
 	struct intel_digital_port *intel_dig_port = intel_mst->primary;
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 	int ret;
 
 	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
 
-	if (wait_for((I915_READ(DP_TP_STATUS(port)) & DP_TP_STATUS_ACT_SENT),
-		     1))
+	if (intel_wait_for_register(dev_priv,
+				    DP_TP_STATUS(port),
+				    DP_TP_STATUS_ACT_SENT,
+				    DP_TP_STATUS_ACT_SENT,
+				    1))
 		DRM_ERROR("Timed out waiting for ACT sent\n");
 
 	ret = drm_dp_check_act_status(&intel_dp->mst_mgr);
@@ -239,12 +241,10 @@ static void intel_dp_mst_enc_get_config(struct intel_encoder *encoder,
 	struct intel_digital_port *intel_dig_port = intel_mst->primary;
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
 	u32 temp, flags = 0;
 
-	pipe_config->has_dp_encoder = true;
-
 	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
 	if (temp & TRANS_DDI_PHSYNC)
 		flags |= DRM_MODE_FLAG_PHSYNC;
@@ -336,6 +336,7 @@ static const struct drm_connector_funcs intel_dp_mst_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_mst_set_property,
 	.atomic_get_property = intel_connector_atomic_get_property,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_dp_mst_connector_destroy,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
@@ -477,9 +478,11 @@ static void intel_dp_register_mst_connector(struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct drm_device *dev = connector->dev;
+
 	drm_modeset_lock_all(dev);
 	intel_connector_add_to_fbdev(intel_connector);
 	drm_modeset_unlock_all(dev);
+
 	drm_connector_register(&intel_connector->base);
 }
 
diff --git a/drivers/gpu/drm/i915/intel_dpio_phy.c b/drivers/gpu/drm/i915/intel_dpio_phy.c
index 288da35572b4..047f48748944 100644
--- a/drivers/gpu/drm/i915/intel_dpio_phy.c
+++ b/drivers/gpu/drm/i915/intel_dpio_phy.c
@@ -168,7 +168,7 @@ void chv_phy_pre_pll_enable(struct intel_encoder *encoder)
 {
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel ch = vlv_dport_to_channel(dport);
@@ -250,7 +250,7 @@ void chv_phy_pre_encoder_enable(struct intel_encoder *encoder)
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel ch = vlv_dport_to_channel(dport);
@@ -400,7 +400,7 @@ void vlv_phy_pre_pll_enable(struct intel_encoder *encoder)
 {
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel port = vlv_dport_to_channel(dport);
@@ -429,7 +429,7 @@ void vlv_phy_pre_encoder_enable(struct intel_encoder *encoder)
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel port = vlv_dport_to_channel(dport);
 	int pipe = intel_crtc->pipe;
@@ -457,7 +457,7 @@ void vlv_phy_pre_encoder_enable(struct intel_encoder *encoder)
 void vlv_phy_reset_lanes(struct intel_encoder *encoder)
 {
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel port = vlv_dport_to_channel(dport);
diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c
index c0eff1571731..5c1f2d235ffa 100644
--- a/drivers/gpu/drm/i915/intel_dpll_mgr.c
+++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c
@@ -83,7 +83,7 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv,
 void intel_prepare_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_shared_dpll *pll = crtc->config->shared_dpll;
 
 	if (WARN_ON(pll == NULL))
@@ -112,7 +112,7 @@ void intel_prepare_shared_dpll(struct intel_crtc *crtc)
 void intel_enable_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_shared_dpll *pll = crtc->config->shared_dpll;
 	unsigned crtc_mask = 1 << drm_crtc_index(&crtc->base);
 	unsigned old_mask;
@@ -151,7 +151,7 @@ out:
 void intel_disable_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_shared_dpll *pll = crtc->config->shared_dpll;
 	unsigned crtc_mask = 1 << drm_crtc_index(&crtc->base);
 
@@ -191,7 +191,7 @@ intel_find_shared_dpll(struct intel_crtc *crtc,
 		       enum intel_dpll_id range_min,
 		       enum intel_dpll_id range_max)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_shared_dpll *pll;
 	struct intel_shared_dpll_config *shared_dpll;
 	enum intel_dpll_id i;
@@ -331,7 +331,7 @@ static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
 static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv,
 				 struct intel_shared_dpll *pll)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_crtc *crtc;
 
 	/* Make sure no transcoder isn't still depending on us. */
@@ -713,7 +713,7 @@ hsw_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state,
 		pll = intel_find_shared_dpll(crtc, crtc_state,
 					     DPLL_ID_WRPLL1, DPLL_ID_WRPLL2);
 
-	} else if (encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
+	} else if (encoder->type == INTEL_OUTPUT_DP ||
 		   encoder->type == INTEL_OUTPUT_DP_MST ||
 		   encoder->type == INTEL_OUTPUT_EDP) {
 		enum intel_dpll_id pll_id;
@@ -856,7 +856,11 @@ static void skl_ddi_pll_enable(struct drm_i915_private *dev_priv,
 	I915_WRITE(regs[pll->id].ctl,
 		   I915_READ(regs[pll->id].ctl) | LCPLL_PLL_ENABLE);
 
-	if (wait_for(I915_READ(DPLL_STATUS) & DPLL_LOCK(pll->id), 5))
+	if (intel_wait_for_register(dev_priv,
+				    DPLL_STATUS,
+				    DPLL_LOCK(pll->id),
+				    DPLL_LOCK(pll->id),
+				    5))
 		DRM_ERROR("DPLL %d not locked\n", pll->id);
 }
 
@@ -1222,7 +1226,7 @@ skl_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state,
 			 DPLL_CFGCR2_KDIV(wrpll_params.kdiv) |
 			 DPLL_CFGCR2_PDIV(wrpll_params.pdiv) |
 			 wrpll_params.central_freq;
-	} else if (encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
+	} else if (encoder->type == INTEL_OUTPUT_DP ||
 		   encoder->type == INTEL_OUTPUT_DP_MST ||
 		   encoder->type == INTEL_OUTPUT_EDP) {
 		switch (crtc_state->port_clock / 2) {
@@ -1374,8 +1378,8 @@ static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv,
 	I915_WRITE(BXT_PORT_PLL_ENABLE(port), temp);
 	POSTING_READ(BXT_PORT_PLL_ENABLE(port));
 
-	if (wait_for_atomic_us((I915_READ(BXT_PORT_PLL_ENABLE(port)) &
-			PORT_PLL_LOCK), 200))
+	if (wait_for_us((I915_READ(BXT_PORT_PLL_ENABLE(port)) & PORT_PLL_LOCK),
+			200))
 		DRM_ERROR("PLL %d not locked\n", port);
 
 	/*
@@ -1530,7 +1534,7 @@ bxt_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state,
 		clk_div.m2_frac_en = clk_div.m2_frac != 0;
 
 		vco = best_clock.vco;
-	} else if (encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
+	} else if (encoder->type == INTEL_OUTPUT_DP ||
 		   encoder->type == INTEL_OUTPUT_EDP) {
 		int i;
 
@@ -1632,7 +1636,7 @@ static const struct intel_shared_dpll_funcs bxt_ddi_pll_funcs = {
 
 static void intel_ddi_pll_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (INTEL_GEN(dev_priv) < 9) {
 		uint32_t val = I915_READ(LCPLL_CTL);
@@ -1719,7 +1723,7 @@ static const struct intel_dpll_mgr bxt_pll_mgr = {
 
 void intel_shared_dpll_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_dpll_mgr *dpll_mgr = NULL;
 	const struct dpll_info *dpll_info;
 	int i;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 7d0e071fe355..55aeaf041749 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -69,39 +69,63 @@
 })
 
 #define wait_for(COND, MS)	  	_wait_for((COND), (MS) * 1000, 1000)
-#define wait_for_us(COND, US)	  	_wait_for((COND), (US), 1)
 
 /* If CONFIG_PREEMPT_COUNT is disabled, in_atomic() always reports false. */
 #if defined(CONFIG_DRM_I915_DEBUG) && defined(CONFIG_PREEMPT_COUNT)
-# define _WAIT_FOR_ATOMIC_CHECK WARN_ON_ONCE(!in_atomic())
+# define _WAIT_FOR_ATOMIC_CHECK(ATOMIC) WARN_ON_ONCE((ATOMIC) && !in_atomic())
 #else
-# define _WAIT_FOR_ATOMIC_CHECK do { } while (0)
+# define _WAIT_FOR_ATOMIC_CHECK(ATOMIC) do { } while (0)
 #endif
 
-#define _wait_for_atomic(COND, US) ({ \
-	unsigned long end__; \
-	int ret__ = 0; \
-	_WAIT_FOR_ATOMIC_CHECK; \
+#define _wait_for_atomic(COND, US, ATOMIC) \
+({ \
+	int cpu, ret, timeout = (US) * 1000; \
+	u64 base; \
+	_WAIT_FOR_ATOMIC_CHECK(ATOMIC); \
 	BUILD_BUG_ON((US) > 50000); \
-	end__ = (local_clock() >> 10) + (US) + 1; \
-	while (!(COND)) { \
-		if (time_after((unsigned long)(local_clock() >> 10), end__)) { \
-			/* Unlike the regular wait_for(), this atomic variant \
-			 * cannot be preempted (and we'll just ignore the issue\
-			 * of irq interruptions) and so we know that no time \
-			 * has passed since the last check of COND and can \
-			 * immediately report the timeout. \
-			 */ \
-			ret__ = -ETIMEDOUT; \
+	if (!(ATOMIC)) { \
+		preempt_disable(); \
+		cpu = smp_processor_id(); \
+	} \
+	base = local_clock(); \
+	for (;;) { \
+		u64 now = local_clock(); \
+		if (!(ATOMIC)) \
+			preempt_enable(); \
+		if (COND) { \
+			ret = 0; \
+			break; \
+		} \
+		if (now - base >= timeout) { \
+			ret = -ETIMEDOUT; \
 			break; \
 		} \
 		cpu_relax(); \
+		if (!(ATOMIC)) { \
+			preempt_disable(); \
+			if (unlikely(cpu != smp_processor_id())) { \
+				timeout -= now - base; \
+				cpu = smp_processor_id(); \
+				base = local_clock(); \
+			} \
+		} \
 	} \
+	ret; \
+})
+
+#define wait_for_us(COND, US) \
+({ \
+	int ret__; \
+	BUILD_BUG_ON(!__builtin_constant_p(US)); \
+	if ((US) > 10) \
+		ret__ = _wait_for((COND), (US), 10); \
+	else \
+		ret__ = _wait_for_atomic((COND), (US), 0); \
 	ret__; \
 })
 
-#define wait_for_atomic(COND, MS)	_wait_for_atomic((COND), (MS) * 1000)
-#define wait_for_atomic_us(COND, US)	_wait_for_atomic((COND), (US))
+#define wait_for_atomic(COND, MS)	_wait_for_atomic((COND), (MS) * 1000, 1)
+#define wait_for_atomic_us(COND, US)	_wait_for_atomic((COND), (US), 1)
 
 #define KHz(x) (1000 * (x))
 #define MHz(x) KHz(1000 * (x))
@@ -135,7 +159,7 @@ enum intel_output_type {
 	INTEL_OUTPUT_LVDS = 4,
 	INTEL_OUTPUT_TVOUT = 5,
 	INTEL_OUTPUT_HDMI = 6,
-	INTEL_OUTPUT_DISPLAYPORT = 7,
+	INTEL_OUTPUT_DP = 7,
 	INTEL_OUTPUT_EDP = 8,
 	INTEL_OUTPUT_DSI = 9,
 	INTEL_OUTPUT_UNKNOWN = 10,
@@ -159,6 +183,7 @@ struct intel_framebuffer {
 struct intel_fbdev {
 	struct drm_fb_helper helper;
 	struct intel_framebuffer *fb;
+	async_cookie_t cookie;
 	int preferred_bpp;
 };
 
@@ -497,12 +522,10 @@ struct intel_crtc_state {
 	 */
 	bool limited_color_range;
 
-	/* DP has a bunch of special case unfortunately, so mark the pipe
-	 * accordingly. */
-	bool has_dp_encoder;
-
-	/* DSI has special cases */
-	bool has_dsi_encoder;
+	/* Bitmask of encoder types (enum intel_output_type)
+	 * driven by the pipe.
+	 */
+	unsigned int output_types;
 
 	/* Whether we should send NULL infoframes. Required for audio. */
 	bool has_hdmi_sink;
@@ -861,6 +884,11 @@ struct intel_dp {
 	 * this port. Only relevant on VLV/CHV.
 	 */
 	enum pipe pps_pipe;
+	/*
+	 * Set if the sequencer may be reset due to a power transition,
+	 * requiring a reinitialization. Only relevant on BXT.
+	 */
+	bool pps_reset;
 	struct edp_power_seq pps_delays;
 
 	bool can_mst; /* this port supports mst */
@@ -886,8 +914,6 @@ struct intel_dp {
 	/* This is called before a link training is starterd */
 	void (*prepare_link_retrain)(struct intel_dp *intel_dp);
 
-	bool train_set_valid;
-
 	/* Displayport compliance testing */
 	unsigned long compliance_test_type;
 	unsigned long compliance_test_data;
@@ -959,14 +985,14 @@ vlv_pipe_to_channel(enum pipe pipe)
 static inline struct drm_crtc *
 intel_get_crtc_for_pipe(struct drm_device *dev, int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	return dev_priv->pipe_to_crtc_mapping[pipe];
 }
 
 static inline struct drm_crtc *
 intel_get_crtc_for_plane(struct drm_device *dev, int plane)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	return dev_priv->plane_to_crtc_mapping[plane];
 }
 
@@ -1159,7 +1185,20 @@ int intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
 enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 					     enum pipe pipe);
-bool intel_pipe_has_type(struct intel_crtc *crtc, enum intel_output_type type);
+static inline bool
+intel_crtc_has_type(const struct intel_crtc_state *crtc_state,
+		    enum intel_output_type type)
+{
+	return crtc_state->output_types & (1 << type);
+}
+static inline bool
+intel_crtc_has_dp_encoder(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->output_types &
+		((1 << INTEL_OUTPUT_DP) |
+		 (1 << INTEL_OUTPUT_DP_MST) |
+		 (1 << INTEL_OUTPUT_EDP));
+}
 static inline void
 intel_wait_for_vblank(struct drm_device *dev, int pipe)
 {
@@ -1340,7 +1379,7 @@ void intel_dp_mst_resume(struct drm_device *dev);
 int intel_dp_max_link_rate(struct intel_dp *intel_dp);
 int intel_dp_rate_select(struct intel_dp *intel_dp, int rate);
 void intel_dp_hot_plug(struct intel_encoder *intel_encoder);
-void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv);
+void intel_power_sequencer_reset(struct drm_i915_private *dev_priv);
 uint32_t intel_dp_pack_aux(const uint8_t *src, int src_bytes);
 void intel_plane_destroy(struct drm_plane *plane);
 void intel_edp_drrs_enable(struct intel_dp *intel_dp);
@@ -1453,6 +1492,7 @@ void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable);
 
 /* intel_lvds.c */
 void intel_lvds_init(struct drm_device *dev);
+struct intel_encoder *intel_get_lvds_encoder(struct drm_device *dev);
 bool intel_is_dual_link_lvds(struct drm_device *dev);
 
 
@@ -1491,7 +1531,8 @@ void intel_gmch_panel_fitting(struct intel_crtc *crtc,
 			      int fitting_mode);
 void intel_panel_set_backlight_acpi(struct intel_connector *connector,
 				    u32 level, u32 max);
-int intel_panel_setup_backlight(struct drm_connector *connector, enum pipe pipe);
+int intel_panel_setup_backlight(struct drm_connector *connector,
+				enum pipe pipe);
 void intel_panel_enable_backlight(struct intel_connector *connector);
 void intel_panel_disable_backlight(struct intel_connector *connector);
 void intel_panel_destroy_backlight(struct drm_connector *connector);
@@ -1500,11 +1541,15 @@ extern struct drm_display_mode *intel_find_panel_downclock(
 				struct drm_device *dev,
 				struct drm_display_mode *fixed_mode,
 				struct drm_connector *connector);
-void intel_backlight_register(struct drm_device *dev);
 
 #if IS_ENABLED(CONFIG_BACKLIGHT_CLASS_DEVICE)
+int intel_backlight_device_register(struct intel_connector *connector);
 void intel_backlight_device_unregister(struct intel_connector *connector);
 #else /* CONFIG_BACKLIGHT_CLASS_DEVICE */
+static int intel_backlight_device_register(struct intel_connector *connector)
+{
+	return 0;
+}
 static inline void intel_backlight_device_unregister(struct intel_connector *connector)
 {
 }
diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c
index b444d0e35a98..de8e9fb51595 100644
--- a/drivers/gpu/drm/i915/intel_dsi.c
+++ b/drivers/gpu/drm/i915/intel_dsi.c
@@ -84,13 +84,15 @@ static void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi, enum port port)
 {
 	struct drm_encoder *encoder = &intel_dsi->base.base;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 mask;
 
 	mask = LP_CTRL_FIFO_EMPTY | HS_CTRL_FIFO_EMPTY |
 		LP_DATA_FIFO_EMPTY | HS_DATA_FIFO_EMPTY;
 
-	if (wait_for((I915_READ(MIPI_GEN_FIFO_STAT(port)) & mask) == mask, 100))
+	if (intel_wait_for_register(dev_priv,
+				    MIPI_GEN_FIFO_STAT(port), mask, mask,
+				    100))
 		DRM_ERROR("DPI FIFOs are not empty\n");
 }
 
@@ -129,7 +131,7 @@ static ssize_t intel_dsi_host_transfer(struct mipi_dsi_host *host,
 {
 	struct intel_dsi_host *intel_dsi_host = to_intel_dsi_host(host);
 	struct drm_device *dev = intel_dsi_host->intel_dsi->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dsi_host->port;
 	struct mipi_dsi_packet packet;
 	ssize_t ret;
@@ -158,8 +160,10 @@ static ssize_t intel_dsi_host_transfer(struct mipi_dsi_host *host,
 
 	/* note: this is never true for reads */
 	if (packet.payload_length) {
-
-		if (wait_for((I915_READ(MIPI_GEN_FIFO_STAT(port)) & data_mask) == 0, 50))
+		if (intel_wait_for_register(dev_priv,
+					    MIPI_GEN_FIFO_STAT(port),
+					    data_mask, 0,
+					    50))
 			DRM_ERROR("Timeout waiting for HS/LP DATA FIFO !full\n");
 
 		write_data(dev_priv, data_reg, packet.payload,
@@ -170,7 +174,10 @@ static ssize_t intel_dsi_host_transfer(struct mipi_dsi_host *host,
 		I915_WRITE(MIPI_INTR_STAT(port), GEN_READ_DATA_AVAIL);
 	}
 
-	if (wait_for((I915_READ(MIPI_GEN_FIFO_STAT(port)) & ctrl_mask) == 0, 50)) {
+	if (intel_wait_for_register(dev_priv,
+				    MIPI_GEN_FIFO_STAT(port),
+				    ctrl_mask, 0,
+				    50)) {
 		DRM_ERROR("Timeout waiting for HS/LP CTRL FIFO !full\n");
 	}
 
@@ -179,7 +186,10 @@ static ssize_t intel_dsi_host_transfer(struct mipi_dsi_host *host,
 	/* ->rx_len is set only for reads */
 	if (msg->rx_len) {
 		data_mask = GEN_READ_DATA_AVAIL;
-		if (wait_for((I915_READ(MIPI_INTR_STAT(port)) & data_mask) == data_mask, 50))
+		if (intel_wait_for_register(dev_priv,
+					    MIPI_INTR_STAT(port),
+					    data_mask, data_mask,
+					    50))
 			DRM_ERROR("Timeout waiting for read data.\n");
 
 		read_data(dev_priv, data_reg, msg->rx_buf, msg->rx_len);
@@ -250,7 +260,7 @@ static int dpi_send_cmd(struct intel_dsi *intel_dsi, u32 cmd, bool hs,
 {
 	struct drm_encoder *encoder = &intel_dsi->base.base;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 mask;
 
 	/* XXX: pipe, hs */
@@ -269,7 +279,9 @@ static int dpi_send_cmd(struct intel_dsi *intel_dsi, u32 cmd, bool hs,
 	I915_WRITE(MIPI_DPI_CONTROL(port), cmd);
 
 	mask = SPL_PKT_SENT_INTERRUPT;
-	if (wait_for((I915_READ(MIPI_INTR_STAT(port)) & mask) == mask, 100))
+	if (intel_wait_for_register(dev_priv,
+				    MIPI_INTR_STAT(port), mask, mask,
+				    100))
 		DRM_ERROR("Video mode command 0x%08x send failed.\n", cmd);
 
 	return 0;
@@ -302,7 +314,7 @@ static inline bool is_cmd_mode(struct intel_dsi *intel_dsi)
 static bool intel_dsi_compute_config(struct intel_encoder *encoder,
 				     struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = container_of(encoder, struct intel_dsi,
 						   base);
 	struct intel_connector *intel_connector = intel_dsi->attached_connector;
@@ -313,8 +325,6 @@ static bool intel_dsi_compute_config(struct intel_encoder *encoder,
 
 	DRM_DEBUG_KMS("\n");
 
-	pipe_config->has_dsi_encoder = true;
-
 	if (fixed_mode) {
 		intel_fixed_panel_mode(fixed_mode, adjusted_mode);
 
@@ -348,7 +358,7 @@ static bool intel_dsi_compute_config(struct intel_encoder *encoder,
 
 static void bxt_dsi_device_ready(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 	u32 val;
@@ -387,7 +397,7 @@ static void bxt_dsi_device_ready(struct intel_encoder *encoder)
 
 static void vlv_dsi_device_ready(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 	u32 val;
@@ -437,7 +447,7 @@ static void intel_dsi_device_ready(struct intel_encoder *encoder)
 static void intel_dsi_port_enable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
@@ -478,7 +488,7 @@ static void intel_dsi_port_enable(struct intel_encoder *encoder)
 static void intel_dsi_port_disable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 
@@ -497,7 +507,7 @@ static void intel_dsi_port_disable(struct intel_encoder *encoder)
 static void intel_dsi_enable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 
@@ -528,7 +538,7 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder);
 static void intel_dsi_pre_enable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	enum port port;
@@ -602,7 +612,7 @@ static void intel_dsi_pre_disable(struct intel_encoder *encoder)
 static void intel_dsi_disable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 	u32 temp;
@@ -641,7 +651,7 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 
@@ -667,8 +677,9 @@ static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 		/* Wait till Clock lanes are in LP-00 state for MIPI Port A
 		 * only. MIPI Port C has no similar bit for checking
 		 */
-		if (wait_for(((I915_READ(port_ctrl) & AFE_LATCHOUT)
-						== 0x00000), 30))
+		if (intel_wait_for_register(dev_priv,
+					    port_ctrl, AFE_LATCHOUT, 0,
+					    30))
 			DRM_ERROR("DSI LP not going Low\n");
 
 		/* Disable MIPI PHY transparent latch */
@@ -685,7 +696,7 @@ static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 
 static void intel_dsi_post_disable(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 
 	DRM_DEBUG_KMS("\n");
@@ -720,7 +731,7 @@ static void intel_dsi_post_disable(struct intel_encoder *encoder)
 static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
 				   enum pipe *pipe)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
 	enum intel_display_power_domain power_domain;
@@ -794,7 +805,7 @@ static void bxt_dsi_get_pipe_config(struct intel_encoder *encoder,
 				 struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_display_mode *adjusted_mode =
 					&pipe_config->base.adjusted_mode;
 	struct drm_display_mode *adjusted_mode_sw;
@@ -954,8 +965,6 @@ static void intel_dsi_get_config(struct intel_encoder *encoder,
 	u32 pclk;
 	DRM_DEBUG_KMS("\n");
 
-	pipe_config->has_dsi_encoder = true;
-
 	if (IS_BROXTON(dev))
 		bxt_dsi_get_pipe_config(encoder, pipe_config);
 
@@ -1013,7 +1022,7 @@ static void set_dsi_timings(struct drm_encoder *encoder,
 			    const struct drm_display_mode *adjusted_mode)
 {
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
 	enum port port;
 	unsigned int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
@@ -1099,7 +1108,7 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
 	const struct drm_display_mode *adjusted_mode = &intel_crtc->config->base.adjusted_mode;
@@ -1390,6 +1399,7 @@ static const struct drm_connector_helper_funcs intel_dsi_connector_helper_funcs
 static const struct drm_connector_funcs intel_dsi_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.detect = intel_dsi_detect,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_dsi_connector_destroy,
 	.fill_modes = drm_helper_probe_single_connector_modes,
@@ -1420,7 +1430,7 @@ void intel_dsi_init(struct drm_device *dev)
 	struct intel_connector *intel_connector;
 	struct drm_connector *connector;
 	struct drm_display_mode *scan, *fixed_mode = NULL;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port;
 	unsigned int i;
 
@@ -1587,13 +1597,10 @@ void intel_dsi_init(struct drm_device *dev)
 	connector->display_info.height_mm = fixed_mode->height_mm;
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, NULL);
+	intel_panel_setup_backlight(connector, INVALID_PIPE);
 
 	intel_dsi_add_properties(intel_connector);
 
-	drm_connector_register(connector);
-
-	intel_panel_setup_backlight(connector, INVALID_PIPE);
-
 	return;
 
 err:
diff --git a/drivers/gpu/drm/i915/intel_dsi_dcs_backlight.c b/drivers/gpu/drm/i915/intel_dsi_dcs_backlight.c
index f0dc427743f8..ac7c6020c443 100644
--- a/drivers/gpu/drm/i915/intel_dsi_dcs_backlight.c
+++ b/drivers/gpu/drm/i915/intel_dsi_dcs_backlight.c
@@ -159,7 +159,7 @@ static int dcs_setup_backlight(struct intel_connector *connector,
 int intel_dsi_dcs_init_backlight_funcs(struct intel_connector *intel_connector)
 {
 	struct drm_device *dev = intel_connector->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *encoder = intel_connector->encoder;
 	struct intel_panel *panel = &intel_connector->panel;
 
diff --git a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
index f122484bedfc..cd154ce6b6c1 100644
--- a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
+++ b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
@@ -303,7 +303,7 @@ static void chv_exec_gpio(struct drm_i915_private *dev_priv,
 static const u8 *mipi_exec_gpio(struct intel_dsi *intel_dsi, const u8 *data)
 {
 	struct drm_device *dev = intel_dsi->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u8 gpio_source, gpio_index;
 	bool value;
 
@@ -469,7 +469,7 @@ static int vbt_panel_get_modes(struct drm_panel *panel)
 	struct vbt_panel *vbt_panel = to_vbt_panel(panel);
 	struct intel_dsi *intel_dsi = vbt_panel->intel_dsi;
 	struct drm_device *dev = intel_dsi->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_display_mode *mode;
 
 	if (!panel->connector)
@@ -497,7 +497,7 @@ static const struct drm_panel_funcs vbt_panel_funcs = {
 struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 {
 	struct drm_device *dev = intel_dsi->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct mipi_config *mipi_config = dev_priv->vbt.dsi.config;
 	struct mipi_pps_data *pps = dev_priv->vbt.dsi.pps;
 	struct drm_display_mode *mode = dev_priv->vbt.lfp_lvds_vbt_mode;
@@ -649,14 +649,13 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 				);
 
 	/*
-	 * Exit zero  is unified val ths_zero and ths_exit
+	 * Exit zero is unified val ths_zero and ths_exit
 	 * minimum value for ths_exit = 110ns
 	 * min (exit_zero_cnt * 2) = 110/UI
 	 * exit_zero_cnt = 55/UI
 	 */
-	 if (exit_zero_cnt < (55 * ui_den / ui_num))
-		if ((55 * ui_den) % ui_num)
-			exit_zero_cnt += 1;
+	if (exit_zero_cnt < (55 * ui_den / ui_num) && (55 * ui_den) % ui_num)
+		exit_zero_cnt += 1;
 
 	/* clk zero count */
 	clk_zero_cnt = DIV_ROUND_UP(
diff --git a/drivers/gpu/drm/i915/intel_dsi_pll.c b/drivers/gpu/drm/i915/intel_dsi_pll.c
index 1765e6e18f2c..6ab58a01b18e 100644
--- a/drivers/gpu/drm/i915/intel_dsi_pll.c
+++ b/drivers/gpu/drm/i915/intel_dsi_pll.c
@@ -55,12 +55,10 @@ static int dsi_calc_mnp(struct drm_i915_private *dev_priv,
 			struct intel_crtc_state *config,
 			int target_dsi_clk)
 {
-	unsigned int calc_m = 0, calc_p = 0;
 	unsigned int m_min, m_max, p_min = 2, p_max = 6;
 	unsigned int m, n, p;
-	int ref_clk;
-	int delta = target_dsi_clk;
-	u32 m_seed;
+	unsigned int calc_m, calc_p;
+	int delta, ref_clk;
 
 	/* target_dsi_clk is expected in kHz */
 	if (target_dsi_clk < 300000 || target_dsi_clk > 1150000) {
@@ -80,6 +78,10 @@ static int dsi_calc_mnp(struct drm_i915_private *dev_priv,
 		m_max = 92;
 	}
 
+	calc_p = p_min;
+	calc_m = m_min;
+	delta = abs(target_dsi_clk - (m_min * ref_clk) / (p_min * n));
+
 	for (m = m_min; m <= m_max && delta; m++) {
 		for (p = p_min; p <= p_max && delta; p++) {
 			/*
@@ -97,11 +99,10 @@ static int dsi_calc_mnp(struct drm_i915_private *dev_priv,
 	}
 
 	/* register has log2(N1), this works fine for powers of two */
-	n = ffs(n) - 1;
-	m_seed = lfsr_converts[calc_m - 62];
 	config->dsi_pll.ctrl = 1 << (DSI_PLL_P1_POST_DIV_SHIFT + calc_p - 2);
-	config->dsi_pll.div = n << DSI_PLL_N1_DIV_SHIFT |
-		m_seed << DSI_PLL_M1_DIV_SHIFT;
+	config->dsi_pll.div =
+		(ffs(n) - 1) << DSI_PLL_N1_DIV_SHIFT |
+		(u32)lfsr_converts[calc_m - 62] << DSI_PLL_M1_DIV_SHIFT;
 
 	return 0;
 }
@@ -113,7 +114,7 @@ static int dsi_calc_mnp(struct drm_i915_private *dev_priv,
 static int vlv_compute_dsi_pll(struct intel_encoder *encoder,
 			       struct intel_crtc_state *config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	int ret;
 	u32 dsi_clk;
@@ -234,8 +235,11 @@ static void bxt_disable_dsi_pll(struct intel_encoder *encoder)
 	 * PLL lock should deassert within 200us.
 	 * Wait up to 1ms before timing out.
 	 */
-	if (wait_for((I915_READ(BXT_DSI_PLL_ENABLE)
-					& BXT_DSI_PLL_LOCKED) == 0, 1))
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DSI_PLL_ENABLE,
+				    BXT_DSI_PLL_LOCKED,
+				    0,
+				    1))
 		DRM_ERROR("Timeout waiting for PLL lock deassertion\n");
 }
 
@@ -321,7 +325,7 @@ static u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp,
 	u32 dsi_clk;
 	u32 dsi_ratio;
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 
 	/* Divide by zero */
 	if (!pipe_bpp) {
@@ -356,7 +360,7 @@ u32 intel_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp,
 static void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
 {
 	u32 temp;
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 
 	temp = I915_READ(MIPI_CTRL(port));
@@ -370,7 +374,7 @@ static void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
 static void bxt_dsi_program_clocks(struct drm_device *dev, enum port port,
 				   const struct intel_crtc_state *config)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tmp;
 	u32 dsi_rate = 0;
 	u32 pll_ratio = 0;
@@ -465,7 +469,7 @@ static int bxt_compute_dsi_pll(struct intel_encoder *encoder,
 static void bxt_enable_dsi_pll(struct intel_encoder *encoder,
 			       const struct intel_crtc_state *config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 	u32 val;
@@ -486,7 +490,11 @@ static void bxt_enable_dsi_pll(struct intel_encoder *encoder,
 	I915_WRITE(BXT_DSI_PLL_ENABLE, val);
 
 	/* Timeout and fail if PLL not locked */
-	if (wait_for(I915_READ(BXT_DSI_PLL_ENABLE) & BXT_DSI_PLL_LOCKED, 1)) {
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DSI_PLL_ENABLE,
+				    BXT_DSI_PLL_LOCKED,
+				    BXT_DSI_PLL_LOCKED,
+				    1)) {
 		DRM_ERROR("Timed out waiting for DSI PLL to lock\n");
 		return;
 	}
@@ -542,7 +550,7 @@ static void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
 {
 	u32 tmp;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Clear old configurations */
 	tmp = I915_READ(BXT_MIPI_CLOCK_CTL);
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 60e4ddf2ec6d..47bdf9dad0d3 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -122,7 +122,7 @@ static struct intel_dvo *intel_attached_dvo(struct drm_connector *connector)
 static bool intel_dvo_connector_get_hw_state(struct intel_connector *connector)
 {
 	struct drm_device *dev = connector->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dvo *intel_dvo = intel_attached_dvo(&connector->base);
 	u32 tmp;
 
@@ -138,7 +138,7 @@ static bool intel_dvo_get_hw_state(struct intel_encoder *encoder,
 				   enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
 	u32 tmp;
 
@@ -155,7 +155,7 @@ static bool intel_dvo_get_hw_state(struct intel_encoder *encoder,
 static void intel_dvo_get_config(struct intel_encoder *encoder,
 				 struct intel_crtc_state *pipe_config)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
 	u32 tmp, flags = 0;
 
@@ -176,7 +176,7 @@ static void intel_dvo_get_config(struct intel_encoder *encoder,
 
 static void intel_disable_dvo(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
 	i915_reg_t dvo_reg = intel_dvo->dev.dvo_reg;
 	u32 temp = I915_READ(dvo_reg);
@@ -188,7 +188,7 @@ static void intel_disable_dvo(struct intel_encoder *encoder)
 
 static void intel_enable_dvo(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	i915_reg_t dvo_reg = intel_dvo->dev.dvo_reg;
@@ -256,7 +256,7 @@ static bool intel_dvo_compute_config(struct intel_encoder *encoder,
 static void intel_dvo_pre_enable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
 	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
@@ -305,7 +305,7 @@ intel_dvo_detect(struct drm_connector *connector, bool force)
 
 static int intel_dvo_get_modes(struct drm_connector *connector)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	const struct drm_display_mode *fixed_mode =
 		to_intel_connector(connector)->panel.fixed_mode;
 
@@ -341,6 +341,7 @@ static void intel_dvo_destroy(struct drm_connector *connector)
 static const struct drm_connector_funcs intel_dvo_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.detect = intel_dvo_detect,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_dvo_destroy,
 	.fill_modes = drm_helper_probe_single_connector_modes,
@@ -378,7 +379,7 @@ static struct drm_display_mode *
 intel_dvo_get_current_mode(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dvo *intel_dvo = intel_attached_dvo(connector);
 	uint32_t dvo_val = I915_READ(intel_dvo->dev.dvo_reg);
 	struct drm_display_mode *mode = NULL;
@@ -420,7 +421,7 @@ static char intel_dvo_port_name(i915_reg_t dvo_reg)
 
 void intel_dvo_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *intel_encoder;
 	struct intel_dvo *intel_dvo;
 	struct intel_connector *intel_connector;
@@ -550,7 +551,6 @@ void intel_dvo_init(struct drm_device *dev)
 			intel_dvo->panel_wants_dither = true;
 		}
 
-		drm_connector_register(connector);
 		return;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c
index ecabd59ffbaf..6a7ad3ed1463 100644
--- a/drivers/gpu/drm/i915/intel_fbc.c
+++ b/drivers/gpu/drm/i915/intel_fbc.c
@@ -124,7 +124,9 @@ static void i8xx_fbc_deactivate(struct drm_i915_private *dev_priv)
 	I915_WRITE(FBC_CONTROL, fbc_ctl);
 
 	/* Wait for compressing bit to clear */
-	if (wait_for((I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) == 0, 10)) {
+	if (intel_wait_for_register(dev_priv,
+				    FBC_STATUS, FBC_STAT_COMPRESSING, 0,
+				    10)) {
 		DRM_DEBUG_KMS("FBC idle timed out\n");
 		return;
 	}
@@ -390,7 +392,7 @@ static void intel_fbc_work_fn(struct work_struct *__work)
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_work *work = &fbc->work;
 	struct intel_crtc *crtc = fbc->crtc;
-	struct drm_vblank_crtc *vblank = &dev_priv->dev->vblank[crtc->pipe];
+	struct drm_vblank_crtc *vblank = &dev_priv->drm.vblank[crtc->pipe];
 
 	if (drm_crtc_vblank_get(&crtc->base)) {
 		DRM_ERROR("vblank not available for FBC on pipe %c\n",
@@ -443,7 +445,7 @@ out:
 
 static void intel_fbc_schedule_activation(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_work *work = &fbc->work;
 
@@ -553,7 +555,7 @@ again:
 
 static int intel_fbc_alloc_cfb(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct drm_mm_node *uninitialized_var(compressed_llb);
 	int size, fb_cpp, ret;
@@ -684,7 +686,7 @@ static bool pixel_format_is_valid(struct drm_i915_private *dev_priv,
  */
 static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	unsigned int effective_w, effective_h, max_w, max_h;
 
@@ -711,7 +713,7 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc,
 					 struct intel_crtc_state *crtc_state,
 					 struct intel_plane_state *plane_state)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_state_cache *cache = &fbc->state_cache;
 	struct drm_framebuffer *fb = plane_state->base.fb;
@@ -744,7 +746,7 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc,
 
 static bool intel_fbc_can_activate(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_state_cache *cache = &fbc->state_cache;
 
@@ -816,23 +818,16 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc)
 
 static bool intel_fbc_can_choose(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
-	bool enable_by_default = IS_HASWELL(dev_priv) ||
-				 IS_BROADWELL(dev_priv);
 
 	if (intel_vgpu_active(dev_priv)) {
 		fbc->no_fbc_reason = "VGPU is active";
 		return false;
 	}
 
-	if (i915.enable_fbc < 0 && !enable_by_default) {
-		fbc->no_fbc_reason = "disabled per chip default";
-		return false;
-	}
-
 	if (!i915.enable_fbc) {
-		fbc->no_fbc_reason = "disabled per module param";
+		fbc->no_fbc_reason = "disabled per module param or by default";
 		return false;
 	}
 
@@ -852,7 +847,7 @@ static bool intel_fbc_can_choose(struct intel_crtc *crtc)
 static void intel_fbc_get_reg_params(struct intel_crtc *crtc,
 				     struct intel_fbc_reg_params *params)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_state_cache *cache = &fbc->state_cache;
 
@@ -885,7 +880,7 @@ void intel_fbc_pre_update(struct intel_crtc *crtc,
 			  struct intel_crtc_state *crtc_state,
 			  struct intel_plane_state *plane_state)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 
 	if (!fbc_supported(dev_priv))
@@ -911,7 +906,7 @@ unlock:
 
 static void __intel_fbc_post_update(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_reg_params old_params;
 
@@ -944,7 +939,7 @@ static void __intel_fbc_post_update(struct intel_crtc *crtc)
 
 void intel_fbc_post_update(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 
 	if (!fbc_supported(dev_priv))
@@ -993,13 +988,13 @@ void intel_fbc_flush(struct drm_i915_private *dev_priv,
 	if (!fbc_supported(dev_priv))
 		return;
 
-	if (origin == ORIGIN_GTT || origin == ORIGIN_FLIP)
-		return;
-
 	mutex_lock(&fbc->lock);
 
 	fbc->busy_bits &= ~frontbuffer_bits;
 
+	if (origin == ORIGIN_GTT || origin == ORIGIN_FLIP)
+		goto out;
+
 	if (!fbc->busy_bits && fbc->enabled &&
 	    (frontbuffer_bits & intel_fbc_get_frontbuffer_bit(fbc))) {
 		if (fbc->active)
@@ -1008,6 +1003,7 @@ void intel_fbc_flush(struct drm_i915_private *dev_priv,
 			__intel_fbc_post_update(fbc->crtc);
 	}
 
+out:
 	mutex_unlock(&fbc->lock);
 }
 
@@ -1089,7 +1085,7 @@ void intel_fbc_enable(struct intel_crtc *crtc,
 		      struct intel_crtc_state *crtc_state,
 		      struct intel_plane_state *plane_state)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 
 	if (!fbc_supported(dev_priv))
@@ -1160,7 +1156,7 @@ static void __intel_fbc_disable(struct drm_i915_private *dev_priv)
  */
 void intel_fbc_disable(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_fbc *fbc = &dev_priv->fbc;
 
 	if (!fbc_supported(dev_priv))
@@ -1214,12 +1210,32 @@ void intel_fbc_init_pipe_state(struct drm_i915_private *dev_priv)
 	if (!no_fbc_on_multiple_pipes(dev_priv))
 		return;
 
-	for_each_intel_crtc(dev_priv->dev, crtc)
+	for_each_intel_crtc(&dev_priv->drm, crtc)
 		if (intel_crtc_active(&crtc->base) &&
 		    to_intel_plane_state(crtc->base.primary->state)->visible)
 			dev_priv->fbc.visible_pipes_mask |= (1 << crtc->pipe);
 }
 
+/*
+ * The DDX driver changes its behavior depending on the value it reads from
+ * i915.enable_fbc, so sanitize it by translating the default value into either
+ * 0 or 1 in order to allow it to know what's going on.
+ *
+ * Notice that this is done at driver initialization and we still allow user
+ * space to change the value during runtime without sanitizing it again. IGT
+ * relies on being able to change i915.enable_fbc at runtime.
+ */
+static int intel_sanitize_fbc_option(struct drm_i915_private *dev_priv)
+{
+	if (i915.enable_fbc >= 0)
+		return !!i915.enable_fbc;
+
+	if (IS_BROADWELL(dev_priv))
+		return 1;
+
+	return 0;
+}
+
 /**
  * intel_fbc_init - Initialize FBC
  * @dev_priv: the i915 device
@@ -1237,6 +1253,9 @@ void intel_fbc_init(struct drm_i915_private *dev_priv)
 	fbc->active = false;
 	fbc->work.scheduled = false;
 
+	i915.enable_fbc = intel_sanitize_fbc_option(dev_priv);
+	DRM_DEBUG_KMS("Sanitized enable_fbc value: %d\n", i915.enable_fbc);
+
 	if (!HAS_FBC(dev_priv)) {
 		fbc->no_fbc_reason = "unsupported by this chipset";
 		return;
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c
index 4babefc51eb2..86b00c6db1a6 100644
--- a/drivers/gpu/drm/i915/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/intel_fbdev.c
@@ -362,23 +362,24 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 				    bool *enabled, int width, int height)
 {
 	struct drm_device *dev = fb_helper->dev;
+	unsigned long conn_configured, mask;
+	unsigned int count = min(fb_helper->connector_count, BITS_PER_LONG);
 	int i, j;
 	bool *save_enabled;
 	bool fallback = true;
 	int num_connectors_enabled = 0;
 	int num_connectors_detected = 0;
-	uint64_t conn_configured = 0, mask;
 	int pass = 0;
 
-	save_enabled = kcalloc(fb_helper->connector_count, sizeof(bool),
-			       GFP_KERNEL);
+	save_enabled = kcalloc(count, sizeof(bool), GFP_KERNEL);
 	if (!save_enabled)
 		return false;
 
-	memcpy(save_enabled, enabled, fb_helper->connector_count);
-	mask = (1 << fb_helper->connector_count) - 1;
+	memcpy(save_enabled, enabled, count);
+	mask = BIT(count) - 1;
+	conn_configured = 0;
 retry:
-	for (i = 0; i < fb_helper->connector_count; i++) {
+	for (i = 0; i < count; i++) {
 		struct drm_fb_helper_connector *fb_conn;
 		struct drm_connector *connector;
 		struct drm_encoder *encoder;
@@ -388,7 +389,7 @@ retry:
 		fb_conn = fb_helper->connector_info[i];
 		connector = fb_conn->connector;
 
-		if (conn_configured & (1 << i))
+		if (conn_configured & BIT(i))
 			continue;
 
 		if (pass == 0 && !connector->has_tile)
@@ -400,7 +401,7 @@ retry:
 		if (!enabled[i]) {
 			DRM_DEBUG_KMS("connector %s not enabled, skipping\n",
 				      connector->name);
-			conn_configured |= (1 << i);
+			conn_configured |= BIT(i);
 			continue;
 		}
 
@@ -419,7 +420,7 @@ retry:
 			DRM_DEBUG_KMS("connector %s has no encoder or crtc, skipping\n",
 				      connector->name);
 			enabled[i] = false;
-			conn_configured |= (1 << i);
+			conn_configured |= BIT(i);
 			continue;
 		}
 
@@ -432,14 +433,15 @@ retry:
 			intel_crtc->lut_b[j] = j;
 		}
 
-		new_crtc = intel_fb_helper_crtc(fb_helper, connector->state->crtc);
+		new_crtc = intel_fb_helper_crtc(fb_helper,
+						connector->state->crtc);
 
 		/*
 		 * Make sure we're not trying to drive multiple connectors
 		 * with a single CRTC, since our cloning support may not
 		 * match the BIOS.
 		 */
-		for (j = 0; j < fb_helper->connector_count; j++) {
+		for (j = 0; j < count; j++) {
 			if (crtcs[j] == new_crtc) {
 				DRM_DEBUG_KMS("fallback: cloned configuration\n");
 				goto bail;
@@ -498,7 +500,7 @@ retry:
 			      modes[i]->flags & DRM_MODE_FLAG_INTERLACE ? "i" :"");
 
 		fallback = false;
-		conn_configured |= (1 << i);
+		conn_configured |= BIT(i);
 	}
 
 	if ((conn_configured & mask) != mask) {
@@ -522,7 +524,7 @@ retry:
 	if (fallback) {
 bail:
 		DRM_DEBUG_KMS("Not using firmware configuration\n");
-		memcpy(enabled, save_enabled, fb_helper->connector_count);
+		memcpy(enabled, save_enabled, count);
 		kfree(save_enabled);
 		return false;
 	}
@@ -538,8 +540,7 @@ static const struct drm_fb_helper_funcs intel_fb_helper_funcs = {
 	.fb_probe = intelfb_create,
 };
 
-static void intel_fbdev_destroy(struct drm_device *dev,
-				struct intel_fbdev *ifbdev)
+static void intel_fbdev_destroy(struct intel_fbdev *ifbdev)
 {
 	/* We rely on the object-free to release the VMA pinning for
 	 * the info->screen_base mmaping. Leaking the VMA is simpler than
@@ -552,12 +553,14 @@ static void intel_fbdev_destroy(struct drm_device *dev,
 	drm_fb_helper_fini(&ifbdev->helper);
 
 	if (ifbdev->fb) {
-		mutex_lock(&dev->struct_mutex);
+		mutex_lock(&ifbdev->helper.dev->struct_mutex);
 		intel_unpin_fb_obj(&ifbdev->fb->base, BIT(DRM_ROTATE_0));
-		mutex_unlock(&dev->struct_mutex);
+		mutex_unlock(&ifbdev->helper.dev->struct_mutex);
 
 		drm_framebuffer_remove(&ifbdev->fb->base);
 	}
+
+	kfree(ifbdev);
 }
 
 /*
@@ -690,9 +693,9 @@ out:
 
 static void intel_fbdev_suspend_worker(struct work_struct *work)
 {
-	intel_fbdev_set_suspend(container_of(work,
-					     struct drm_i915_private,
-					     fbdev_suspend_work)->dev,
+	intel_fbdev_set_suspend(&container_of(work,
+					      struct drm_i915_private,
+					      fbdev_suspend_work)->drm,
 				FBINFO_STATE_RUNNING,
 				true);
 }
@@ -700,7 +703,7 @@ static void intel_fbdev_suspend_worker(struct work_struct *work)
 int intel_fbdev_init(struct drm_device *dev)
 {
 	struct intel_fbdev *ifbdev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret;
 
 	if (WARN_ON(INTEL_INFO(dev)->num_pipes == 0))
@@ -732,38 +735,50 @@ int intel_fbdev_init(struct drm_device *dev)
 
 static void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
-	struct drm_i915_private *dev_priv = data;
-	struct intel_fbdev *ifbdev = dev_priv->fbdev;
+	struct intel_fbdev *ifbdev = data;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	if (drm_fb_helper_initial_config(&ifbdev->helper,
 					 ifbdev->preferred_bpp))
-		intel_fbdev_fini(dev_priv->dev);
+		intel_fbdev_fini(ifbdev->helper.dev);
 }
 
 void intel_fbdev_initial_config_async(struct drm_device *dev)
 {
-	async_schedule(intel_fbdev_initial_config, to_i915(dev));
+	struct intel_fbdev *ifbdev = to_i915(dev)->fbdev;
+
+	ifbdev->cookie = async_schedule(intel_fbdev_initial_config, ifbdev);
+}
+
+static void intel_fbdev_sync(struct intel_fbdev *ifbdev)
+{
+	if (!ifbdev->cookie)
+		return;
+
+	/* Only serialises with all preceding async calls, hence +1 */
+	async_synchronize_cookie(ifbdev->cookie + 1);
+	ifbdev->cookie = 0;
 }
 
 void intel_fbdev_fini(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (!dev_priv->fbdev)
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_fbdev *ifbdev = dev_priv->fbdev;
+
+	if (!ifbdev)
 		return;
 
 	flush_work(&dev_priv->fbdev_suspend_work);
-
 	if (!current_is_async())
-		async_synchronize_full();
-	intel_fbdev_destroy(dev, dev_priv->fbdev);
-	kfree(dev_priv->fbdev);
+		intel_fbdev_sync(ifbdev);
+
+	intel_fbdev_destroy(ifbdev);
 	dev_priv->fbdev = NULL;
 }
 
 void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 	struct fb_info *info;
 
@@ -812,7 +827,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous
 
 void intel_fbdev_output_poll_changed(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	if (dev_priv->fbdev)
 		drm_fb_helper_hotplug_event(&dev_priv->fbdev->helper);
 }
@@ -820,13 +835,15 @@ void intel_fbdev_output_poll_changed(struct drm_device *dev)
 void intel_fbdev_restore_mode(struct drm_device *dev)
 {
 	int ret;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 	struct drm_fb_helper *fb_helper;
 
 	if (!ifbdev)
 		return;
 
+	intel_fbdev_sync(ifbdev);
+
 	fb_helper = &ifbdev->helper;
 
 	ret = drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper);
diff --git a/drivers/gpu/drm/i915/intel_fifo_underrun.c b/drivers/gpu/drm/i915/intel_fifo_underrun.c
index 9be839a242f9..2aa744081f09 100644
--- a/drivers/gpu/drm/i915/intel_fifo_underrun.c
+++ b/drivers/gpu/drm/i915/intel_fifo_underrun.c
@@ -50,7 +50,7 @@
 
 static bool ivb_can_enable_err_int(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc;
 	enum pipe pipe;
 
@@ -68,7 +68,7 @@ static bool ivb_can_enable_err_int(struct drm_device *dev)
 
 static bool cpt_can_enable_serr_int(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 	struct intel_crtc *crtc;
 
@@ -105,7 +105,7 @@ static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
 					     enum pipe pipe,
 					     bool enable, bool old)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t reg = PIPESTAT(pipe);
 	u32 pipestat = I915_READ(reg) & 0xffff0000;
 
@@ -123,7 +123,7 @@ static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
 static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev,
 						 enum pipe pipe, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN :
 					  DE_PIPEB_FIFO_UNDERRUN;
 
@@ -154,7 +154,7 @@ static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
 						  enum pipe pipe,
 						  bool enable, bool old)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	if (enable) {
 		I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN(pipe));
 
@@ -176,7 +176,7 @@ static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
 static void broadwell_set_fifo_underrun_reporting(struct drm_device *dev,
 						  enum pipe pipe, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (enable)
 		bdw_enable_pipe_irq(dev_priv, pipe, GEN8_PIPE_FIFO_UNDERRUN);
@@ -188,7 +188,7 @@ static void ibx_set_fifo_underrun_reporting(struct drm_device *dev,
 					    enum transcoder pch_transcoder,
 					    bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t bit = (pch_transcoder == TRANSCODER_A) ?
 		       SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER;
 
@@ -220,7 +220,7 @@ static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
 					    enum transcoder pch_transcoder,
 					    bool enable, bool old)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (enable) {
 		I915_WRITE(SERR_INT,
@@ -244,7 +244,7 @@ static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
 static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
 						    enum pipe pipe, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	bool old;
@@ -289,7 +289,7 @@ bool intel_set_cpu_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
 	bool ret;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	ret = __intel_set_cpu_fifo_underrun_reporting(dev_priv->dev, pipe,
+	ret = __intel_set_cpu_fifo_underrun_reporting(&dev_priv->drm, pipe,
 						      enable);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -334,10 +334,12 @@ bool intel_set_pch_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
 	intel_crtc->pch_fifo_underrun_disabled = !enable;
 
 	if (HAS_PCH_IBX(dev_priv))
-		ibx_set_fifo_underrun_reporting(dev_priv->dev, pch_transcoder,
+		ibx_set_fifo_underrun_reporting(&dev_priv->drm,
+						pch_transcoder,
 						enable);
 	else
-		cpt_set_fifo_underrun_reporting(dev_priv->dev, pch_transcoder,
+		cpt_set_fifo_underrun_reporting(&dev_priv->drm,
+						pch_transcoder,
 						enable, old);
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
@@ -405,7 +407,7 @@ void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv)
 
 	spin_lock_irq(&dev_priv->irq_lock);
 
-	for_each_intel_crtc(dev_priv->dev, crtc) {
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
 		if (crtc->cpu_fifo_underrun_disabled)
 			continue;
 
@@ -432,7 +434,7 @@ void intel_check_pch_fifo_underruns(struct drm_i915_private *dev_priv)
 
 	spin_lock_irq(&dev_priv->irq_lock);
 
-	for_each_intel_crtc(dev_priv->dev, crtc) {
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
 		if (crtc->pch_fifo_underrun_disabled)
 			continue;
 
diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h
index 4df80cc9a291..3e3e743740c0 100644
--- a/drivers/gpu/drm/i915/intel_guc.h
+++ b/drivers/gpu/drm/i915/intel_guc.h
@@ -26,6 +26,7 @@
 
 #include "intel_guc_fwif.h"
 #include "i915_guc_reg.h"
+#include "intel_ringbuffer.h"
 
 struct drm_i915_gem_request;
 
@@ -86,7 +87,7 @@ struct i915_guc_client {
 	int retcode;
 
 	/* Per-engine counts of GuC submissions */
-	uint64_t submissions[GUC_MAX_ENGINES_NUM];
+	uint64_t submissions[I915_NUM_ENGINES];
 };
 
 enum intel_guc_fw_status {
@@ -143,8 +144,8 @@ struct intel_guc {
 	uint32_t action_fail;		/* Total number of failures	*/
 	int32_t action_err;		/* Last error code		*/
 
-	uint64_t submissions[GUC_MAX_ENGINES_NUM];
-	uint32_t last_seqno[GUC_MAX_ENGINES_NUM];
+	uint64_t submissions[I915_NUM_ENGINES];
+	uint32_t last_seqno[I915_NUM_ENGINES];
 };
 
 /* intel_guc_loader.c */
diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c
index 8fe96a2d989e..605c69658d2c 100644
--- a/drivers/gpu/drm/i915/intel_guc_loader.c
+++ b/drivers/gpu/drm/i915/intel_guc_loader.c
@@ -65,6 +65,9 @@ MODULE_FIRMWARE(I915_SKL_GUC_UCODE);
 #define I915_BXT_GUC_UCODE "i915/bxt_guc_ver8_7.bin"
 MODULE_FIRMWARE(I915_BXT_GUC_UCODE);
 
+#define I915_KBL_GUC_UCODE "i915/kbl_guc_ver9_14.bin"
+MODULE_FIRMWARE(I915_KBL_GUC_UCODE);
+
 /* User-friendly representation of an enum */
 const char *intel_guc_fw_status_repr(enum intel_guc_fw_status status)
 {
@@ -87,7 +90,7 @@ static void direct_interrupts_to_host(struct drm_i915_private *dev_priv)
 	struct intel_engine_cs *engine;
 	int irqs;
 
-	/* tell all command streamers NOT to forward interrupts and vblank to GuC */
+	/* tell all command streamers NOT to forward interrupts or vblank to GuC */
 	irqs = _MASKED_FIELD(GFX_FORWARD_VBLANK_MASK, GFX_FORWARD_VBLANK_NEVER);
 	irqs |= _MASKED_BIT_DISABLE(GFX_INTERRUPT_STEERING);
 	for_each_engine(engine, dev_priv)
@@ -105,9 +108,8 @@ static void direct_interrupts_to_guc(struct drm_i915_private *dev_priv)
 	int irqs;
 	u32 tmp;
 
-	/* tell all command streamers to forward interrupts and vblank to GuC */
-	irqs = _MASKED_FIELD(GFX_FORWARD_VBLANK_MASK, GFX_FORWARD_VBLANK_ALWAYS);
-	irqs |= _MASKED_BIT_ENABLE(GFX_INTERRUPT_STEERING);
+	/* tell all command streamers to forward interrupts (but not vblank) to GuC */
+	irqs = _MASKED_BIT_ENABLE(GFX_INTERRUPT_STEERING);
 	for_each_engine(engine, dev_priv)
 		I915_WRITE(RING_MODE_GEN7(engine), irqs);
 
@@ -312,7 +314,7 @@ static u32 guc_wopcm_size(struct drm_i915_private *dev_priv)
 static int guc_ucode_xfer(struct drm_i915_private *dev_priv)
 {
 	struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw;
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	int ret;
 
 	ret = i915_gem_object_set_to_gtt_domain(guc_fw->guc_fw_obj, false);
@@ -411,7 +413,7 @@ static int i915_reset_guc(struct drm_i915_private *dev_priv)
  */
 int intel_guc_setup(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw;
 	const char *fw_path = guc_fw->guc_fw_path;
 	int retries, ret, err;
@@ -606,7 +608,7 @@ static void guc_fw_fetch(struct drm_device *dev, struct intel_guc_fw *guc_fw)
 
 	/* Header and uCode will be loaded to WOPCM. Size of the two. */
 	size = guc_fw->header_size + guc_fw->ucode_size;
-	if (size > guc_wopcm_size(dev->dev_private)) {
+	if (size > guc_wopcm_size(to_i915(dev))) {
 		DRM_ERROR("Firmware is too large to fit in WOPCM\n");
 		goto fail;
 	}
@@ -679,7 +681,7 @@ fail:
  */
 void intel_guc_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw;
 	const char *fw_path;
 
@@ -699,6 +701,10 @@ void intel_guc_init(struct drm_device *dev)
 		fw_path = I915_BXT_GUC_UCODE;
 		guc_fw->guc_fw_major_wanted = 8;
 		guc_fw->guc_fw_minor_wanted = 7;
+	} else if (IS_KABYLAKE(dev)) {
+		fw_path = I915_KBL_GUC_UCODE;
+		guc_fw->guc_fw_major_wanted = 9;
+		guc_fw->guc_fw_minor_wanted = 14;
 	} else {
 		fw_path = "";	/* unknown device */
 	}
@@ -728,7 +734,7 @@ void intel_guc_init(struct drm_device *dev)
  */
 void intel_guc_fini(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw;
 
 	mutex_lock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/intel_gvt.c b/drivers/gpu/drm/i915/intel_gvt.c
index 9fa458ce40a6..434f4d5c553d 100644
--- a/drivers/gpu/drm/i915/intel_gvt.c
+++ b/drivers/gpu/drm/i915/intel_gvt.c
@@ -63,7 +63,7 @@ int intel_gvt_init(struct drm_i915_private *dev_priv)
 
 	if (!is_supported_device(dev_priv)) {
 		DRM_DEBUG_DRIVER("Unsupported device. GVT-g is disabled\n");
-		return 0;
+		goto bail;
 	}
 
 	/*
@@ -72,16 +72,20 @@ int intel_gvt_init(struct drm_i915_private *dev_priv)
 	ret = intel_gvt_init_host();
 	if (ret) {
 		DRM_DEBUG_DRIVER("Not in host or MPT modules not found\n");
-		return 0;
+		goto bail;
 	}
 
 	ret = intel_gvt_init_device(dev_priv);
 	if (ret) {
 		DRM_DEBUG_DRIVER("Fail to init GVT device\n");
-		return 0;
+		goto bail;
 	}
 
 	return 0;
+
+bail:
+	i915.enable_gvt = 0;
+	return 0;
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index fb21626ada64..4df9f384910c 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -47,7 +47,7 @@ static void
 assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
 {
 	struct drm_device *dev = intel_hdmi_to_dev(intel_hdmi);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t enabled_bits;
 
 	enabled_bits = HAS_DDI(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
@@ -138,7 +138,7 @@ static void g4x_write_infoframe(struct drm_encoder *encoder,
 {
 	const uint32_t *data = frame;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 val = I915_READ(VIDEO_DIP_CTL);
 	int i;
 
@@ -192,7 +192,7 @@ static void ibx_write_infoframe(struct drm_encoder *encoder,
 {
 	const uint32_t *data = frame;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	i915_reg_t reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
@@ -251,7 +251,7 @@ static void cpt_write_infoframe(struct drm_encoder *encoder,
 {
 	const uint32_t *data = frame;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	i915_reg_t reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
@@ -308,7 +308,7 @@ static void vlv_write_infoframe(struct drm_encoder *encoder,
 {
 	const uint32_t *data = frame;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	i915_reg_t reg = VLV_TVIDEO_DIP_CTL(intel_crtc->pipe);
 	u32 val = I915_READ(reg);
@@ -366,7 +366,7 @@ static void hsw_write_infoframe(struct drm_encoder *encoder,
 {
 	const uint32_t *data = frame;
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
 	i915_reg_t ctl_reg = HSW_TVIDEO_DIP_CTL(cpu_transcoder);
@@ -508,7 +508,7 @@ static void g4x_set_infoframes(struct drm_encoder *encoder,
 			       bool enable,
 			       const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	struct intel_hdmi *intel_hdmi = &intel_dig_port->hdmi;
 	i915_reg_t reg = VIDEO_DIP_CTL;
@@ -629,7 +629,7 @@ static bool gcp_default_phase_possible(int pipe_bpp,
 
 static bool intel_hdmi_set_gcp_infoframe(struct drm_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
 	i915_reg_t reg;
 	u32 val = 0;
@@ -661,7 +661,7 @@ static void ibx_set_infoframes(struct drm_encoder *encoder,
 			       bool enable,
 			       const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	struct intel_hdmi *intel_hdmi = &intel_dig_port->hdmi;
@@ -713,7 +713,7 @@ static void cpt_set_infoframes(struct drm_encoder *encoder,
 			       bool enable,
 			       const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
 	i915_reg_t reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
@@ -755,7 +755,7 @@ static void vlv_set_infoframes(struct drm_encoder *encoder,
 			       bool enable,
 			       const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
@@ -807,7 +807,7 @@ static void hsw_set_infoframes(struct drm_encoder *encoder,
 			       bool enable,
 			       const struct drm_display_mode *adjusted_mode)
 {
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
 	i915_reg_t reg = HSW_TVIDEO_DIP_CTL(intel_crtc->config->cpu_transcoder);
@@ -855,7 +855,7 @@ void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable)
 static void intel_hdmi_prepare(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
@@ -894,7 +894,7 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 				    enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
@@ -931,7 +931,7 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder,
 {
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tmp, flags = 0;
 	int dotclock;
 
@@ -988,7 +988,7 @@ static void intel_enable_hdmi_audio(struct intel_encoder *encoder)
 static void g4x_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	u32 temp;
@@ -1009,7 +1009,7 @@ static void g4x_enable_hdmi(struct intel_encoder *encoder)
 static void ibx_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	u32 temp;
@@ -1058,7 +1058,7 @@ static void ibx_enable_hdmi(struct intel_encoder *encoder)
 static void cpt_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	enum pipe pipe = crtc->pipe;
@@ -1115,7 +1115,7 @@ static void vlv_enable_hdmi(struct intel_encoder *encoder)
 static void intel_disable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	u32 temp;
@@ -1154,7 +1154,7 @@ static void intel_disable_hdmi(struct intel_encoder *encoder)
 		I915_WRITE(intel_hdmi->hdmi_reg, temp);
 		POSTING_READ(intel_hdmi->hdmi_reg);
 
-		intel_wait_for_vblank_if_active(dev_priv->dev, PIPE_A);
+		intel_wait_for_vblank_if_active(&dev_priv->drm, PIPE_A);
 		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 	}
@@ -1273,33 +1273,15 @@ intel_hdmi_mode_valid(struct drm_connector *connector,
 static bool hdmi_12bpc_possible(struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc_state->base.crtc->dev;
-	struct drm_atomic_state *state;
-	struct intel_encoder *encoder;
-	struct drm_connector *connector;
-	struct drm_connector_state *connector_state;
-	int count = 0, count_hdmi = 0;
-	int i;
 
 	if (HAS_GMCH_DISPLAY(dev))
 		return false;
 
-	state = crtc_state->base.state;
-
-	for_each_connector_in_state(state, connector, connector_state, i) {
-		if (connector_state->crtc != crtc_state->base.crtc)
-			continue;
-
-		encoder = to_intel_encoder(connector_state->best_encoder);
-
-		count_hdmi += encoder->type == INTEL_OUTPUT_HDMI;
-		count++;
-	}
-
 	/*
 	 * HDMI 12bpc affects the clocks, so it's only possible
 	 * when not cloning with other encoder types.
 	 */
-	return count_hdmi > 0 && count_hdmi == count;
+	return crtc_state->output_types == 1 << INTEL_OUTPUT_HDMI;
 }
 
 bool intel_hdmi_compute_config(struct intel_encoder *encoder,
@@ -1575,7 +1557,7 @@ intel_hdmi_set_property(struct drm_connector *connector,
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 	struct intel_digital_port *intel_dig_port =
 		hdmi_to_dig_port(intel_hdmi);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	int ret;
 
 	ret = drm_object_property_set_value(&connector->base, property, val);
@@ -1674,7 +1656,7 @@ static void vlv_hdmi_pre_enable(struct intel_encoder *encoder)
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
 	struct intel_hdmi *intel_hdmi = &dport->hdmi;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &intel_crtc->config->base.adjusted_mode;
@@ -1722,7 +1704,7 @@ static void vlv_hdmi_post_disable(struct intel_encoder *encoder)
 static void chv_hdmi_post_disable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	mutex_lock(&dev_priv->sb_lock);
 
@@ -1737,7 +1719,7 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
 	struct intel_hdmi *intel_hdmi = &dport->hdmi;
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &intel_crtc->config->base.adjusted_mode;
@@ -1774,6 +1756,7 @@ static const struct drm_connector_funcs intel_hdmi_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_hdmi_set_property,
 	.atomic_get_property = intel_connector_atomic_get_property,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_hdmi_destroy,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
@@ -1806,7 +1789,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	struct intel_hdmi *intel_hdmi = &intel_dig_port->hdmi;
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_dig_port->port;
 	uint8_t alternate_ddc_pin;
 
@@ -1914,7 +1897,6 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
-	drm_connector_register(connector);
 	intel_hdmi->attached_connector = intel_connector;
 
 	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c
index 38eeca7a6e72..51434ec871f2 100644
--- a/drivers/gpu/drm/i915/intel_hotplug.c
+++ b/drivers/gpu/drm/i915/intel_hotplug.c
@@ -144,7 +144,7 @@ static bool intel_hpd_irq_storm_detect(struct drm_i915_private *dev_priv,
 
 static void intel_hpd_irq_storm_disable(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_connector *intel_connector;
 	struct intel_encoder *intel_encoder;
@@ -191,7 +191,7 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work)
 	struct drm_i915_private *dev_priv =
 		container_of(work, typeof(*dev_priv),
 			     hotplug.reenable_work.work);
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	int i;
 
@@ -302,7 +302,7 @@ static void i915_hotplug_work_func(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
 		container_of(work, struct drm_i915_private, hotplug.hotplug_work);
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_connector *intel_connector;
 	struct intel_encoder *intel_encoder;
@@ -455,7 +455,7 @@ void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
  */
 void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
 	int i;
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 81de23098be7..1f266d7df2ec 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -113,7 +113,7 @@ to_intel_gmbus(struct i2c_adapter *i2c)
 void
 intel_i2c_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(GMBUS0, 0);
 	I915_WRITE(GMBUS4, 0);
@@ -138,7 +138,7 @@ static void intel_i2c_quirk_set(struct drm_i915_private *dev_priv, bool enable)
 static u32 get_reserved(struct intel_gmbus *bus)
 {
 	struct drm_i915_private *dev_priv = bus->dev_priv;
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	u32 reserved = 0;
 
 	/* On most chips, these bits must be preserved in software. */
@@ -212,7 +212,7 @@ intel_gpio_pre_xfer(struct i2c_adapter *adapter)
 					       adapter);
 	struct drm_i915_private *dev_priv = bus->dev_priv;
 
-	intel_i2c_reset(dev_priv->dev);
+	intel_i2c_reset(&dev_priv->drm);
 	intel_i2c_quirk_set(dev_priv, true);
 	set_data(bus, 1);
 	set_clock(bus, 1);
@@ -298,15 +298,16 @@ gmbus_wait_idle(struct drm_i915_private *dev_priv)
 {
 	int ret;
 
-#define C ((I915_READ_NOTRACE(GMBUS2) & GMBUS_ACTIVE) == 0)
-
 	if (!HAS_GMBUS_IRQ(dev_priv))
-		return wait_for(C, 10);
+		return intel_wait_for_register(dev_priv,
+					       GMBUS2, GMBUS_ACTIVE, 0,
+					       10);
 
 	/* Important: The hw handles only the first bit, so set only one! */
 	I915_WRITE(GMBUS4, GMBUS_IDLE_EN);
 
-	ret = wait_event_timeout(dev_priv->gmbus_wait_queue, C,
+	ret = wait_event_timeout(dev_priv->gmbus_wait_queue,
+				 (I915_READ_NOTRACE(GMBUS2) & GMBUS_ACTIVE) == 0,
 				 msecs_to_jiffies_timeout(10));
 
 	I915_WRITE(GMBUS4, 0);
@@ -315,7 +316,6 @@ gmbus_wait_idle(struct drm_i915_private *dev_priv)
 		return 0;
 	else
 		return -ETIMEDOUT;
-#undef C
 }
 
 static int
@@ -632,7 +632,7 @@ static const struct i2c_algorithm gmbus_algorithm = {
  */
 int intel_setup_gmbus(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_gmbus *bus;
 	unsigned int pin;
 	int ret;
@@ -688,7 +688,7 @@ int intel_setup_gmbus(struct drm_device *dev)
 			goto err;
 	}
 
-	intel_i2c_reset(dev_priv->dev);
+	intel_i2c_reset(&dev_priv->drm);
 
 	return 0;
 
@@ -736,7 +736,7 @@ void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit)
 
 void intel_teardown_gmbus(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_gmbus *bus;
 	unsigned int pin;
 
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index debed011a958..70c699043d0e 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -789,9 +789,6 @@ intel_logical_ring_advance_and_submit(struct drm_i915_gem_request *request)
 	intel_logical_ring_emit(ringbuf, MI_NOOP);
 	intel_logical_ring_advance(ringbuf);
 
-	if (intel_engine_stopped(engine))
-		return 0;
-
 	/* We keep the previous context alive until we retire the following
 	 * request. This ensures that any the context object is still pinned
 	 * for any residual writes the HW makes into it on the context switch
@@ -826,7 +823,7 @@ int intel_execlists_submission(struct i915_execbuffer_params *params,
 {
 	struct drm_device       *dev = params->dev;
 	struct intel_engine_cs *engine = params->engine;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_ringbuffer *ringbuf = params->ctx->engine[engine->id].ringbuf;
 	u64 exec_start;
 	int instp_mode;
@@ -902,7 +899,7 @@ void intel_execlists_cancel_requests(struct intel_engine_cs *engine)
 	struct drm_i915_gem_request *req, *tmp;
 	LIST_HEAD(cancel_list);
 
-	WARN_ON(!mutex_is_locked(&engine->i915->dev->struct_mutex));
+	WARN_ON(!mutex_is_locked(&engine->i915->drm.struct_mutex));
 
 	spin_lock_bh(&engine->execlist_lock);
 	list_replace_init(&engine->execlist_queue, &cancel_list);
@@ -929,7 +926,10 @@ void intel_logical_ring_stop(struct intel_engine_cs *engine)
 
 	/* TODO: Is this correct with Execlists enabled? */
 	I915_WRITE_MODE(engine, _MASKED_BIT_ENABLE(STOP_RING));
-	if (wait_for((I915_READ_MODE(engine) & MODE_IDLE) != 0, 1000)) {
+	if (intel_wait_for_register(dev_priv,
+				    RING_MI_MODE(engine->mmio_base),
+				    MODE_IDLE, MODE_IDLE,
+				    1000)) {
 		DRM_ERROR("%s :timed out trying to stop ring\n", engine->name);
 		return;
 	}
@@ -961,7 +961,7 @@ static int intel_lr_context_pin(struct i915_gem_context *ctx,
 	u32 *lrc_reg_state;
 	int ret;
 
-	lockdep_assert_held(&ctx->i915->dev->struct_mutex);
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
 
 	if (ce->pin_count++)
 		return 0;
@@ -1011,7 +1011,7 @@ void intel_lr_context_unpin(struct i915_gem_context *ctx,
 {
 	struct intel_context *ce = &ctx->engine[engine->id];
 
-	lockdep_assert_held(&ctx->i915->dev->struct_mutex);
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
 	GEM_BUG_ON(ce->pin_count == 0);
 
 	if (--ce->pin_count)
@@ -1296,6 +1296,31 @@ static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine,
 		wa_ctx_emit(batch, index, 0);
 		wa_ctx_emit(batch, index, 0);
 	}
+
+	/* WaMediaPoolStateCmdInWABB:bxt */
+	if (HAS_POOLED_EU(engine->i915)) {
+		/*
+		 * EU pool configuration is setup along with golden context
+		 * during context initialization. This value depends on
+		 * device type (2x6 or 3x6) and needs to be updated based
+		 * on which subslice is disabled especially for 2x6
+		 * devices, however it is safe to load default
+		 * configuration of 3x6 device instead of masking off
+		 * corresponding bits because HW ignores bits of a disabled
+		 * subslice and drops down to appropriate config. Please
+		 * see render_state_setup() in i915_gem_render_state.c for
+		 * possible configurations, to avoid duplication they are
+		 * not shown here again.
+		 */
+		u32 eu_pool_config = 0x00777000;
+		wa_ctx_emit(batch, index, GEN9_MEDIA_POOL_STATE);
+		wa_ctx_emit(batch, index, GEN9_MEDIA_POOL_ENABLE);
+		wa_ctx_emit(batch, index, eu_pool_config);
+		wa_ctx_emit(batch, index, 0);
+		wa_ctx_emit(batch, index, 0);
+		wa_ctx_emit(batch, index, 0);
+	}
+
 	/* Pad to end of cacheline */
 	while (index % CACHELINE_DWORDS)
 		wa_ctx_emit(batch, index, MI_NOOP);
@@ -1353,8 +1378,8 @@ static int lrc_setup_wa_ctx_obj(struct intel_engine_cs *engine, u32 size)
 {
 	int ret;
 
-	engine->wa_ctx.obj = i915_gem_object_create(engine->i915->dev,
-						   PAGE_ALIGN(size));
+	engine->wa_ctx.obj = i915_gem_object_create(&engine->i915->drm,
+						    PAGE_ALIGN(size));
 	if (IS_ERR(engine->wa_ctx.obj)) {
 		DRM_DEBUG_DRIVER("alloc LRC WA ctx backing obj failed.\n");
 		ret = PTR_ERR(engine->wa_ctx.obj);
@@ -1614,36 +1639,18 @@ static int gen8_emit_bb_start(struct drm_i915_gem_request *req,
 	return 0;
 }
 
-static bool gen8_logical_ring_get_irq(struct intel_engine_cs *engine)
+static void gen8_logical_ring_enable_irq(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return false;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		I915_WRITE_IMR(engine,
-			       ~(engine->irq_enable_mask | engine->irq_keep_mask));
-		POSTING_READ(RING_IMR(engine->mmio_base));
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	I915_WRITE_IMR(engine,
+		       ~(engine->irq_enable_mask | engine->irq_keep_mask));
+	POSTING_READ_FW(RING_IMR(engine->mmio_base));
 }
 
-static void gen8_logical_ring_put_irq(struct intel_engine_cs *engine)
+static void gen8_logical_ring_disable_irq(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
-		POSTING_READ(RING_IMR(engine->mmio_base));
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
 }
 
 static int gen8_emit_flush(struct drm_i915_gem_request *request,
@@ -1780,16 +1787,6 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 	return 0;
 }
 
-static u32 gen8_get_seqno(struct intel_engine_cs *engine)
-{
-	return intel_read_status_page(engine, I915_GEM_HWS_INDEX);
-}
-
-static void gen8_set_seqno(struct intel_engine_cs *engine, u32 seqno)
-{
-	intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
-}
-
 static void bxt_a_seqno_barrier(struct intel_engine_cs *engine)
 {
 	/*
@@ -1805,14 +1802,6 @@ static void bxt_a_seqno_barrier(struct intel_engine_cs *engine)
 	intel_flush_status_page(engine, I915_GEM_HWS_INDEX);
 }
 
-static void bxt_a_set_seqno(struct intel_engine_cs *engine, u32 seqno)
-{
-	intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
-
-	/* See bxt_a_get_seqno() explaining the reason for the clflush. */
-	intel_flush_status_page(engine, I915_GEM_HWS_INDEX);
-}
-
 /*
  * Reserve space for 2 NOOPs at the end of each request to be
  * used as a workaround for not being allowed to do lite
@@ -1838,7 +1827,7 @@ static int gen8_emit_request(struct drm_i915_gem_request *request)
 				intel_hws_seqno_address(request->engine) |
 				MI_FLUSH_DW_USE_GTT);
 	intel_logical_ring_emit(ringbuf, 0);
-	intel_logical_ring_emit(ringbuf, i915_gem_request_get_seqno(request));
+	intel_logical_ring_emit(ringbuf, request->seqno);
 	intel_logical_ring_emit(ringbuf, MI_USER_INTERRUPT);
 	intel_logical_ring_emit(ringbuf, MI_NOOP);
 	return intel_logical_ring_advance_and_submit(request);
@@ -1958,6 +1947,8 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *engine)
 	i915_cmd_parser_fini_ring(engine);
 	i915_gem_batch_pool_fini(&engine->batch_pool);
 
+	intel_engine_fini_breadcrumbs(engine);
+
 	if (engine->status_page.obj) {
 		i915_gem_object_unpin_map(engine->status_page.obj);
 		engine->status_page.obj = NULL;
@@ -1979,15 +1970,11 @@ logical_ring_default_vfuncs(struct intel_engine_cs *engine)
 	engine->init_hw = gen8_init_common_ring;
 	engine->emit_request = gen8_emit_request;
 	engine->emit_flush = gen8_emit_flush;
-	engine->irq_get = gen8_logical_ring_get_irq;
-	engine->irq_put = gen8_logical_ring_put_irq;
+	engine->irq_enable = gen8_logical_ring_enable_irq;
+	engine->irq_disable = gen8_logical_ring_disable_irq;
 	engine->emit_bb_start = gen8_emit_bb_start;
-	engine->get_seqno = gen8_get_seqno;
-	engine->set_seqno = gen8_set_seqno;
-	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1)) {
+	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1))
 		engine->irq_seqno_barrier = bxt_a_seqno_barrier;
-		engine->set_seqno = bxt_a_set_seqno;
-	}
 }
 
 static inline void
@@ -1995,7 +1982,6 @@ logical_ring_default_irqs(struct intel_engine_cs *engine, unsigned shift)
 {
 	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift;
 	engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift;
-	init_waitqueue_head(&engine->irq_queue);
 }
 
 static int
@@ -2016,12 +2002,94 @@ lrc_setup_hws(struct intel_engine_cs *engine,
 	return 0;
 }
 
+static int
+logical_ring_init(struct intel_engine_cs *engine)
+{
+	struct i915_gem_context *dctx = engine->i915->kernel_context;
+	int ret;
+
+	ret = intel_engine_init_breadcrumbs(engine);
+	if (ret)
+		goto error;
+
+	ret = i915_cmd_parser_init_ring(engine);
+	if (ret)
+		goto error;
+
+	ret = execlists_context_deferred_alloc(dctx, engine);
+	if (ret)
+		goto error;
+
+	/* As this is the default context, always pin it */
+	ret = intel_lr_context_pin(dctx, engine);
+	if (ret) {
+		DRM_ERROR("Failed to pin context for %s: %d\n",
+			  engine->name, ret);
+		goto error;
+	}
+
+	/* And setup the hardware status page. */
+	ret = lrc_setup_hws(engine, dctx->engine[engine->id].state);
+	if (ret) {
+		DRM_ERROR("Failed to set up hws %s: %d\n", engine->name, ret);
+		goto error;
+	}
+
+	return 0;
+
+error:
+	intel_logical_ring_cleanup(engine);
+	return ret;
+}
+
+static int logical_render_ring_init(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	if (HAS_L3_DPF(dev_priv))
+		engine->irq_keep_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+
+	/* Override some for render ring. */
+	if (INTEL_GEN(dev_priv) >= 9)
+		engine->init_hw = gen9_init_render_ring;
+	else
+		engine->init_hw = gen8_init_render_ring;
+	engine->init_context = gen8_init_rcs_context;
+	engine->cleanup = intel_fini_pipe_control;
+	engine->emit_flush = gen8_emit_flush_render;
+	engine->emit_request = gen8_emit_request_render;
+
+	ret = intel_init_pipe_control(engine, 4096);
+	if (ret)
+		return ret;
+
+	ret = intel_init_workaround_bb(engine);
+	if (ret) {
+		/*
+		 * We continue even if we fail to initialize WA batch
+		 * because we only expect rare glitches but nothing
+		 * critical to prevent us from using GPU
+		 */
+		DRM_ERROR("WA batch buffer initialization failed: %d\n",
+			  ret);
+	}
+
+	ret = logical_ring_init(engine);
+	if (ret) {
+		lrc_destroy_wa_ctx_obj(engine);
+	}
+
+	return ret;
+}
+
 static const struct logical_ring_info {
 	const char *name;
 	unsigned exec_id;
 	unsigned guc_id;
 	u32 mmio_base;
 	unsigned irq_shift;
+	int (*init)(struct intel_engine_cs *engine);
 } logical_rings[] = {
 	[RCS] = {
 		.name = "render ring",
@@ -2029,6 +2097,7 @@ static const struct logical_ring_info {
 		.guc_id = GUC_RENDER_ENGINE,
 		.mmio_base = RENDER_RING_BASE,
 		.irq_shift = GEN8_RCS_IRQ_SHIFT,
+		.init = logical_render_ring_init,
 	},
 	[BCS] = {
 		.name = "blitter ring",
@@ -2036,6 +2105,7 @@ static const struct logical_ring_info {
 		.guc_id = GUC_BLITTER_ENGINE,
 		.mmio_base = BLT_RING_BASE,
 		.irq_shift = GEN8_BCS_IRQ_SHIFT,
+		.init = logical_ring_init,
 	},
 	[VCS] = {
 		.name = "bsd ring",
@@ -2043,6 +2113,7 @@ static const struct logical_ring_info {
 		.guc_id = GUC_VIDEO_ENGINE,
 		.mmio_base = GEN6_BSD_RING_BASE,
 		.irq_shift = GEN8_VCS1_IRQ_SHIFT,
+		.init = logical_ring_init,
 	},
 	[VCS2] = {
 		.name = "bsd2 ring",
@@ -2050,6 +2121,7 @@ static const struct logical_ring_info {
 		.guc_id = GUC_VIDEO_ENGINE2,
 		.mmio_base = GEN8_BSD2_RING_BASE,
 		.irq_shift = GEN8_VCS2_IRQ_SHIFT,
+		.init = logical_ring_init,
 	},
 	[VECS] = {
 		.name = "video enhancement ring",
@@ -2057,14 +2129,14 @@ static const struct logical_ring_info {
 		.guc_id = GUC_VIDEOENHANCE_ENGINE,
 		.mmio_base = VEBOX_RING_BASE,
 		.irq_shift = GEN8_VECS_IRQ_SHIFT,
+		.init = logical_ring_init,
 	},
 };
 
 static struct intel_engine_cs *
-logical_ring_setup(struct drm_device *dev, enum intel_engine_id id)
+logical_ring_setup(struct drm_i915_private *dev_priv, enum intel_engine_id id)
 {
 	const struct logical_ring_info *info = &logical_rings[id];
-	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[id];
 	enum forcewake_domains fw_domains;
 
@@ -2107,169 +2179,62 @@ logical_ring_setup(struct drm_device *dev, enum intel_engine_id id)
 	logical_ring_default_irqs(engine, info->irq_shift);
 
 	intel_engine_init_hangcheck(engine);
-	i915_gem_batch_pool_init(dev, &engine->batch_pool);
+	i915_gem_batch_pool_init(&dev_priv->drm, &engine->batch_pool);
 
 	return engine;
 }
 
-static int
-logical_ring_init(struct intel_engine_cs *engine)
-{
-	struct i915_gem_context *dctx = engine->i915->kernel_context;
-	int ret;
-
-	ret = i915_cmd_parser_init_ring(engine);
-	if (ret)
-		goto error;
-
-	ret = execlists_context_deferred_alloc(dctx, engine);
-	if (ret)
-		goto error;
-
-	/* As this is the default context, always pin it */
-	ret = intel_lr_context_pin(dctx, engine);
-	if (ret) {
-		DRM_ERROR("Failed to pin context for %s: %d\n",
-			  engine->name, ret);
-		goto error;
-	}
-
-	/* And setup the hardware status page. */
-	ret = lrc_setup_hws(engine, dctx->engine[engine->id].state);
-	if (ret) {
-		DRM_ERROR("Failed to set up hws %s: %d\n", engine->name, ret);
-		goto error;
-	}
-
-	return 0;
-
-error:
-	intel_logical_ring_cleanup(engine);
-	return ret;
-}
-
-static int logical_render_ring_init(struct drm_device *dev)
-{
-	struct intel_engine_cs *engine = logical_ring_setup(dev, RCS);
-	int ret;
-
-	if (HAS_L3_DPF(dev))
-		engine->irq_keep_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
-
-	/* Override some for render ring. */
-	if (INTEL_INFO(dev)->gen >= 9)
-		engine->init_hw = gen9_init_render_ring;
-	else
-		engine->init_hw = gen8_init_render_ring;
-	engine->init_context = gen8_init_rcs_context;
-	engine->cleanup = intel_fini_pipe_control;
-	engine->emit_flush = gen8_emit_flush_render;
-	engine->emit_request = gen8_emit_request_render;
-
-	ret = intel_init_pipe_control(engine);
-	if (ret)
-		return ret;
-
-	ret = intel_init_workaround_bb(engine);
-	if (ret) {
-		/*
-		 * We continue even if we fail to initialize WA batch
-		 * because we only expect rare glitches but nothing
-		 * critical to prevent us from using GPU
-		 */
-		DRM_ERROR("WA batch buffer initialization failed: %d\n",
-			  ret);
-	}
-
-	ret = logical_ring_init(engine);
-	if (ret) {
-		lrc_destroy_wa_ctx_obj(engine);
-	}
-
-	return ret;
-}
-
-static int logical_bsd_ring_init(struct drm_device *dev)
-{
-	struct intel_engine_cs *engine = logical_ring_setup(dev, VCS);
-
-	return logical_ring_init(engine);
-}
-
-static int logical_bsd2_ring_init(struct drm_device *dev)
-{
-	struct intel_engine_cs *engine = logical_ring_setup(dev, VCS2);
-
-	return logical_ring_init(engine);
-}
-
-static int logical_blt_ring_init(struct drm_device *dev)
-{
-	struct intel_engine_cs *engine = logical_ring_setup(dev, BCS);
-
-	return logical_ring_init(engine);
-}
-
-static int logical_vebox_ring_init(struct drm_device *dev)
-{
-	struct intel_engine_cs *engine = logical_ring_setup(dev, VECS);
-
-	return logical_ring_init(engine);
-}
-
 /**
  * intel_logical_rings_init() - allocate, populate and init the Engine Command Streamers
  * @dev: DRM device.
  *
- * This function inits the engines for an Execlists submission style (the equivalent in the
- * legacy ringbuffer submission world would be i915_gem_init_engines). It does it only for
- * those engines that are present in the hardware.
+ * This function inits the engines for an Execlists submission style (the
+ * equivalent in the legacy ringbuffer submission world would be
+ * i915_gem_init_engines). It does it only for those engines that are present in
+ * the hardware.
  *
  * Return: non-zero if the initialization failed.
  */
 int intel_logical_rings_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	unsigned int mask = 0;
+	unsigned int i;
 	int ret;
 
-	ret = logical_render_ring_init(dev);
-	if (ret)
-		return ret;
+	WARN_ON(INTEL_INFO(dev_priv)->ring_mask &
+		GENMASK(sizeof(mask) * BITS_PER_BYTE - 1, I915_NUM_ENGINES));
 
-	if (HAS_BSD(dev)) {
-		ret = logical_bsd_ring_init(dev);
-		if (ret)
-			goto cleanup_render_ring;
-	}
+	for (i = 0; i < ARRAY_SIZE(logical_rings); i++) {
+		if (!HAS_ENGINE(dev_priv, i))
+			continue;
 
-	if (HAS_BLT(dev)) {
-		ret = logical_blt_ring_init(dev);
-		if (ret)
-			goto cleanup_bsd_ring;
-	}
+		if (!logical_rings[i].init)
+			continue;
 
-	if (HAS_VEBOX(dev)) {
-		ret = logical_vebox_ring_init(dev);
+		ret = logical_rings[i].init(logical_ring_setup(dev_priv, i));
 		if (ret)
-			goto cleanup_blt_ring;
+			goto cleanup;
+
+		mask |= ENGINE_MASK(i);
 	}
 
-	if (HAS_BSD2(dev)) {
-		ret = logical_bsd2_ring_init(dev);
-		if (ret)
-			goto cleanup_vebox_ring;
+	/*
+	 * Catch failures to update logical_rings table when the new engines
+	 * are added to the driver by a warning and disabling the forgotten
+	 * engines.
+	 */
+	if (WARN_ON(mask != INTEL_INFO(dev_priv)->ring_mask)) {
+		struct intel_device_info *info =
+			(struct intel_device_info *)&dev_priv->info;
+		info->ring_mask = mask;
 	}
 
 	return 0;
 
-cleanup_vebox_ring:
-	intel_logical_ring_cleanup(&dev_priv->engine[VECS]);
-cleanup_blt_ring:
-	intel_logical_ring_cleanup(&dev_priv->engine[BCS]);
-cleanup_bsd_ring:
-	intel_logical_ring_cleanup(&dev_priv->engine[VCS]);
-cleanup_render_ring:
-	intel_logical_ring_cleanup(&dev_priv->engine[RCS]);
+cleanup:
+	for (i = 0; i < I915_NUM_ENGINES; i++)
+		intel_logical_ring_cleanup(&dev_priv->engine[i]);
 
 	return ret;
 }
@@ -2546,7 +2511,7 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx,
 	/* One extra page as the sharing data between driver and GuC */
 	context_size += PAGE_SIZE * LRC_PPHWSP_PN;
 
-	ctx_obj = i915_gem_object_create(ctx->i915->dev, context_size);
+	ctx_obj = i915_gem_object_create(&ctx->i915->drm, context_size);
 	if (IS_ERR(ctx_obj)) {
 		DRM_DEBUG_DRIVER("Alloc LRC backing obj failed.\n");
 		return PTR_ERR(ctx_obj);
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index e9082185a375..49550470483e 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -72,7 +72,7 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 				    enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	enum intel_display_power_domain power_domain;
 	u32 tmp;
@@ -106,7 +106,7 @@ static void intel_lvds_get_config(struct intel_encoder *encoder,
 				  struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	u32 tmp, flags = 0;
 
@@ -140,7 +140,7 @@ static void intel_pre_enable_lvds(struct intel_encoder *encoder)
 {
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
 	int pipe = crtc->pipe;
@@ -184,8 +184,8 @@ static void intel_pre_enable_lvds(struct intel_encoder *encoder)
 	 * panels behave in the two modes. For now, let's just maintain the
 	 * value we got from the BIOS.
 	 */
-	 temp &= ~LVDS_A3_POWER_MASK;
-	 temp |= lvds_encoder->a3_power;
+	temp &= ~LVDS_A3_POWER_MASK;
+	temp |= lvds_encoder->a3_power;
 
 	/* Set the dithering flag on LVDS as needed, note that there is no
 	 * special lvds dither control bit on pch-split platforms, dithering is
@@ -216,7 +216,7 @@ static void intel_enable_lvds(struct intel_encoder *encoder)
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
 	struct intel_connector *intel_connector =
 		&lvds_encoder->attached_connector->base;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t ctl_reg, stat_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
@@ -231,7 +231,7 @@ static void intel_enable_lvds(struct intel_encoder *encoder)
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON);
 	POSTING_READ(lvds_encoder->reg);
-	if (wait_for((I915_READ(stat_reg) & PP_ON) != 0, 1000))
+	if (intel_wait_for_register(dev_priv, stat_reg, PP_ON, PP_ON, 1000))
 		DRM_ERROR("timed out waiting for panel to power on\n");
 
 	intel_panel_enable_backlight(intel_connector);
@@ -241,7 +241,7 @@ static void intel_disable_lvds(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	i915_reg_t ctl_reg, stat_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
@@ -253,7 +253,7 @@ static void intel_disable_lvds(struct intel_encoder *encoder)
 	}
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) & ~POWER_TARGET_ON);
-	if (wait_for((I915_READ(stat_reg) & PP_ON) == 0, 1000))
+	if (intel_wait_for_register(dev_priv, stat_reg, PP_ON, 0, 1000))
 		DRM_ERROR("timed out waiting for panel to power off\n");
 
 	I915_WRITE(lvds_encoder->reg, I915_READ(lvds_encoder->reg) & ~LVDS_PORT_EN);
@@ -442,7 +442,7 @@ static int intel_lid_notify(struct notifier_block *nb, unsigned long val,
 		container_of(nb, struct intel_lvds_connector, lid_notifier);
 	struct drm_connector *connector = &lvds_connector->base.base;
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (dev->switch_power_state != DRM_SWITCH_POWER_ON)
 		return NOTIFY_OK;
@@ -555,6 +555,7 @@ static const struct drm_connector_funcs intel_lvds_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_lvds_set_property,
 	.atomic_get_property = intel_connector_atomic_get_property,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_lvds_destroy,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
@@ -810,27 +811,29 @@ static const struct dmi_system_id intel_dual_link_lvds[] = {
 	{ }	/* terminating entry */
 };
 
-bool intel_is_dual_link_lvds(struct drm_device *dev)
+struct intel_encoder *intel_get_lvds_encoder(struct drm_device *dev)
 {
-	struct intel_encoder *encoder;
-	struct intel_lvds_encoder *lvds_encoder;
+	struct intel_encoder *intel_encoder;
 
-	for_each_intel_encoder(dev, encoder) {
-		if (encoder->type == INTEL_OUTPUT_LVDS) {
-			lvds_encoder = to_lvds_encoder(&encoder->base);
+	for_each_intel_encoder(dev, intel_encoder)
+		if (intel_encoder->type == INTEL_OUTPUT_LVDS)
+			return intel_encoder;
 
-			return lvds_encoder->is_dual_link;
-		}
-	}
+	return NULL;
+}
 
-	return false;
+bool intel_is_dual_link_lvds(struct drm_device *dev)
+{
+	struct intel_encoder *encoder = intel_get_lvds_encoder(dev);
+
+	return encoder && to_lvds_encoder(&encoder->base)->is_dual_link;
 }
 
 static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder)
 {
 	struct drm_device *dev = lvds_encoder->base.base.dev;
 	unsigned int val;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* use the module option value if specified */
 	if (i915.lvds_channel_mode > 0)
@@ -880,7 +883,7 @@ static bool intel_lvds_supported(struct drm_device *dev)
  */
 void intel_lvds_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_lvds_encoder *lvds_encoder;
 	struct intel_encoder *intel_encoder;
 	struct intel_lvds_connector *lvds_connector;
@@ -1118,6 +1121,7 @@ out:
 	mutex_unlock(&dev->mode_config.mutex);
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
+	intel_panel_setup_backlight(connector, INVALID_PIPE);
 
 	lvds_encoder->is_dual_link = compute_is_dual_link_lvds(lvds_encoder);
 	DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
@@ -1130,9 +1134,6 @@ out:
 		DRM_DEBUG_KMS("lid notifier registration failed\n");
 		lvds_connector->lid_notifier.notifier_call = NULL;
 	}
-	drm_connector_register(connector);
-
-	intel_panel_setup_backlight(connector, INVALID_PIPE);
 
 	return;
 
diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index 38a4c8ce7e63..f2584d0a01ab 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -82,7 +82,7 @@ void
 intel_attach_force_audio_property(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_property *prop;
 
 	prop = dev_priv->force_audio_property;
@@ -109,7 +109,7 @@ void
 intel_attach_broadcast_rgb_property(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_property *prop;
 
 	prop = dev_priv->broadcast_rgb_property;
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index f6d8a21d2c49..c27d5eb063d0 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -232,11 +232,28 @@ struct opregion_asle_ext {
 #define SWSCI_SBCB_POST_VBE_PM		SWSCI_FUNCTION_CODE(SWSCI_SBCB, 19)
 #define SWSCI_SBCB_ENABLE_DISABLE_AUDIO	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 21)
 
-#define ACPI_OTHER_OUTPUT (0<<8)
-#define ACPI_VGA_OUTPUT (1<<8)
-#define ACPI_TV_OUTPUT (2<<8)
-#define ACPI_DIGITAL_OUTPUT (3<<8)
-#define ACPI_LVDS_OUTPUT (4<<8)
+/*
+ * ACPI Specification, Revision 5.0, Appendix B.3.2 _DOD (Enumerate All Devices
+ * Attached to the Display Adapter).
+ */
+#define ACPI_DISPLAY_INDEX_SHIFT		0
+#define ACPI_DISPLAY_INDEX_MASK			(0xf << 0)
+#define ACPI_DISPLAY_PORT_ATTACHMENT_SHIFT	4
+#define ACPI_DISPLAY_PORT_ATTACHMENT_MASK	(0xf << 4)
+#define ACPI_DISPLAY_TYPE_SHIFT			8
+#define ACPI_DISPLAY_TYPE_MASK			(0xf << 8)
+#define ACPI_DISPLAY_TYPE_OTHER			(0 << 8)
+#define ACPI_DISPLAY_TYPE_VGA			(1 << 8)
+#define ACPI_DISPLAY_TYPE_TV			(2 << 8)
+#define ACPI_DISPLAY_TYPE_EXTERNAL_DIGITAL	(3 << 8)
+#define ACPI_DISPLAY_TYPE_INTERNAL_DIGITAL	(4 << 8)
+#define ACPI_VENDOR_SPECIFIC_SHIFT		12
+#define ACPI_VENDOR_SPECIFIC_MASK		(0xf << 12)
+#define ACPI_BIOS_CAN_DETECT			(1 << 16)
+#define ACPI_DEPENDS_ON_VGA			(1 << 17)
+#define ACPI_PIPE_ID_SHIFT			18
+#define ACPI_PIPE_ID_MASK			(7 << 18)
+#define ACPI_DEVICE_ID_SCHEME			(1 << 31)
 
 #define MAX_DSLP	1500
 
@@ -244,7 +261,7 @@ static int swsci(struct drm_i915_private *dev_priv,
 		 u32 function, u32 parm, u32 *parm_out)
 {
 	struct opregion_swsci *swsci = dev_priv->opregion.swsci;
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	u32 main_function, sub_function, scic;
 	u16 swsci_val;
 	u32 dslp;
@@ -366,7 +383,7 @@ int intel_opregion_notify_encoder(struct intel_encoder *intel_encoder,
 		type = DISPLAY_TYPE_CRT;
 		break;
 	case INTEL_OUTPUT_UNKNOWN:
-	case INTEL_OUTPUT_DISPLAYPORT:
+	case INTEL_OUTPUT_DP:
 	case INTEL_OUTPUT_HDMI:
 	case INTEL_OUTPUT_DP_MST:
 		type = DISPLAY_TYPE_EXTERNAL_FLAT_PANEL;
@@ -418,7 +435,7 @@ static u32 asle_set_backlight(struct drm_i915_private *dev_priv, u32 bclp)
 {
 	struct intel_connector *connector;
 	struct opregion_asle *asle = dev_priv->opregion.asle;
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	DRM_DEBUG_DRIVER("bclp = 0x%08x\n", bclp);
 
@@ -657,10 +674,51 @@ static void set_did(struct intel_opregion *opregion, int i, u32 val)
 	}
 }
 
+static u32 acpi_display_type(struct drm_connector *connector)
+{
+	u32 display_type;
+
+	switch (connector->connector_type) {
+	case DRM_MODE_CONNECTOR_VGA:
+	case DRM_MODE_CONNECTOR_DVIA:
+		display_type = ACPI_DISPLAY_TYPE_VGA;
+		break;
+	case DRM_MODE_CONNECTOR_Composite:
+	case DRM_MODE_CONNECTOR_SVIDEO:
+	case DRM_MODE_CONNECTOR_Component:
+	case DRM_MODE_CONNECTOR_9PinDIN:
+	case DRM_MODE_CONNECTOR_TV:
+		display_type = ACPI_DISPLAY_TYPE_TV;
+		break;
+	case DRM_MODE_CONNECTOR_DVII:
+	case DRM_MODE_CONNECTOR_DVID:
+	case DRM_MODE_CONNECTOR_DisplayPort:
+	case DRM_MODE_CONNECTOR_HDMIA:
+	case DRM_MODE_CONNECTOR_HDMIB:
+		display_type = ACPI_DISPLAY_TYPE_EXTERNAL_DIGITAL;
+		break;
+	case DRM_MODE_CONNECTOR_LVDS:
+	case DRM_MODE_CONNECTOR_eDP:
+	case DRM_MODE_CONNECTOR_DSI:
+		display_type = ACPI_DISPLAY_TYPE_INTERNAL_DIGITAL;
+		break;
+	case DRM_MODE_CONNECTOR_Unknown:
+	case DRM_MODE_CONNECTOR_VIRTUAL:
+		display_type = ACPI_DISPLAY_TYPE_OTHER;
+		break;
+	default:
+		MISSING_CASE(connector->connector_type);
+		display_type = ACPI_DISPLAY_TYPE_OTHER;
+		break;
+	}
+
+	return display_type;
+}
+
 static void intel_didl_outputs(struct drm_i915_private *dev_priv)
 {
 	struct intel_opregion *opregion = &dev_priv->opregion;
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	struct drm_connector *connector;
 	acpi_handle handle;
 	struct acpi_device *acpi_dev, *acpi_cdev, *acpi_video_bus = NULL;
@@ -724,37 +782,18 @@ end:
 
 blind_set:
 	i = 0;
-	list_for_each_entry(connector, &dev_priv->dev->mode_config.connector_list, head) {
-		int output_type = ACPI_OTHER_OUTPUT;
+	list_for_each_entry(connector,
+			    &dev_priv->drm.mode_config.connector_list, head) {
+		int display_type = acpi_display_type(connector);
+
 		if (i >= max_outputs) {
 			DRM_DEBUG_KMS("More than %u outputs in connector list\n",
 				      max_outputs);
 			return;
 		}
-		switch (connector->connector_type) {
-		case DRM_MODE_CONNECTOR_VGA:
-		case DRM_MODE_CONNECTOR_DVIA:
-			output_type = ACPI_VGA_OUTPUT;
-			break;
-		case DRM_MODE_CONNECTOR_Composite:
-		case DRM_MODE_CONNECTOR_SVIDEO:
-		case DRM_MODE_CONNECTOR_Component:
-		case DRM_MODE_CONNECTOR_9PinDIN:
-			output_type = ACPI_TV_OUTPUT;
-			break;
-		case DRM_MODE_CONNECTOR_DVII:
-		case DRM_MODE_CONNECTOR_DVID:
-		case DRM_MODE_CONNECTOR_DisplayPort:
-		case DRM_MODE_CONNECTOR_HDMIA:
-		case DRM_MODE_CONNECTOR_HDMIB:
-			output_type = ACPI_DIGITAL_OUTPUT;
-			break;
-		case DRM_MODE_CONNECTOR_LVDS:
-			output_type = ACPI_LVDS_OUTPUT;
-			break;
-		}
+
 		temp = get_did(opregion, i);
-		set_did(opregion, i, temp | (1 << 31) | output_type | i);
+		set_did(opregion, i, temp | (1 << 31) | display_type | i);
 		i++;
 	}
 	goto end;
@@ -916,7 +955,7 @@ static const struct dmi_system_id intel_no_opregion_vbt[] = {
 int intel_opregion_setup(struct drm_i915_private *dev_priv)
 {
 	struct intel_opregion *opregion = &dev_priv->opregion;
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	u32 asls, mboxes;
 	char buf[sizeof(OPREGION_SIGNATURE)];
 	int err = 0;
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index eb93f90bb74d..3212d8806b5a 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -409,7 +409,7 @@ static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
 	struct intel_engine_cs *engine = &dev_priv->engine[RCS];
 	int ret;
 
-	lockdep_assert_held(&dev_priv->dev->struct_mutex);
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
 
 	/* Only wait if there is actually an old frame to release to
 	 * guarantee forward progress.
@@ -741,8 +741,8 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	u32 swidth, swidthsw, sheight, ostride;
 	enum pipe pipe = overlay->crtc->pipe;
 
-	lockdep_assert_held(&dev_priv->dev->struct_mutex);
-	WARN_ON(!drm_modeset_is_locked(&dev_priv->dev->mode_config.connection_mutex));
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+	WARN_ON(!drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
 
 	ret = intel_overlay_release_old_vid(overlay);
 	if (ret != 0)
@@ -836,7 +836,8 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	overlay->old_vid_bo = overlay->vid_bo;
 	overlay->vid_bo = new_bo;
 
-	intel_frontbuffer_flip(dev_priv->dev, INTEL_FRONTBUFFER_OVERLAY(pipe));
+	intel_frontbuffer_flip(&dev_priv->drm,
+			       INTEL_FRONTBUFFER_OVERLAY(pipe));
 
 	return 0;
 
@@ -851,8 +852,8 @@ int intel_overlay_switch_off(struct intel_overlay *overlay)
 	struct overlay_registers __iomem *regs;
 	int ret;
 
-	lockdep_assert_held(&dev_priv->dev->struct_mutex);
-	WARN_ON(!drm_modeset_is_locked(&dev_priv->dev->mode_config.connection_mutex));
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+	WARN_ON(!drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
 
 	ret = intel_overlay_recover_from_interrupt(overlay);
 	if (ret != 0)
@@ -1084,7 +1085,7 @@ int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data,
 				  struct drm_file *file_priv)
 {
 	struct drm_intel_overlay_put_image *put_image_rec = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_overlay *overlay;
 	struct drm_crtc *drmmode_crtc;
 	struct intel_crtc *crtc;
@@ -1282,7 +1283,7 @@ int intel_overlay_attrs_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv)
 {
 	struct drm_intel_overlay_attrs *attrs = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_overlay *overlay;
 	struct overlay_registers __iomem *regs;
 	int ret;
@@ -1379,7 +1380,7 @@ void intel_setup_overlay(struct drm_i915_private *dev_priv)
 	if (!overlay)
 		return;
 
-	mutex_lock(&dev_priv->dev->struct_mutex);
+	mutex_lock(&dev_priv->drm.struct_mutex);
 	if (WARN_ON(dev_priv->overlay))
 		goto out_free;
 
@@ -1387,9 +1388,10 @@ void intel_setup_overlay(struct drm_i915_private *dev_priv)
 
 	reg_bo = NULL;
 	if (!OVERLAY_NEEDS_PHYSICAL(dev_priv))
-		reg_bo = i915_gem_object_create_stolen(dev_priv->dev, PAGE_SIZE);
+		reg_bo = i915_gem_object_create_stolen(&dev_priv->drm,
+						       PAGE_SIZE);
 	if (reg_bo == NULL)
-		reg_bo = i915_gem_object_create(dev_priv->dev, PAGE_SIZE);
+		reg_bo = i915_gem_object_create(&dev_priv->drm, PAGE_SIZE);
 	if (IS_ERR(reg_bo))
 		goto out_free;
 	overlay->reg_bo = reg_bo;
@@ -1434,7 +1436,7 @@ void intel_setup_overlay(struct drm_i915_private *dev_priv)
 	intel_overlay_unmap_regs(overlay, regs);
 
 	dev_priv->overlay = overlay;
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 	DRM_INFO("initialized overlay support\n");
 	return;
 
@@ -1444,7 +1446,7 @@ out_unpin_bo:
 out_free_bo:
 	drm_gem_object_unreference(&reg_bo->base);
 out_free:
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 	kfree(overlay);
 	return;
 }
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index bf721781c259..96c65d77e886 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -377,7 +377,7 @@ out:
 enum drm_connector_status
 intel_panel_detect(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Assume that the BIOS does not lie through the OpRegion... */
 	if (!i915.panel_ignore_lid && dev_priv->opregion.lid_state) {
@@ -504,7 +504,7 @@ static u32 i9xx_get_backlight(struct intel_connector *connector)
 	if (panel->backlight.combination_mode) {
 		u8 lbpc;
 
-		pci_read_config_byte(dev_priv->dev->pdev, LBPC, &lbpc);
+		pci_read_config_byte(dev_priv->drm.pdev, LBPC, &lbpc);
 		val *= lbpc;
 	}
 
@@ -592,7 +592,7 @@ static void i9xx_set_backlight(struct intel_connector *connector, u32 level)
 
 		lbpc = level * 0xfe / panel->backlight.max + 1;
 		level /= lbpc;
-		pci_write_config_byte(dev_priv->dev->pdev, LBPC, lbpc);
+		pci_write_config_byte(dev_priv->drm.pdev, LBPC, lbpc);
 	}
 
 	if (IS_GEN4(dev_priv)) {
@@ -822,7 +822,7 @@ void intel_panel_disable_backlight(struct intel_connector *connector)
 	 * backlight. This will leave the backlight on unnecessarily when
 	 * another client is not activated.
 	 */
-	if (dev_priv->dev->switch_power_state == DRM_SWITCH_POWER_CHANGING) {
+	if (dev_priv->drm.switch_power_state == DRM_SWITCH_POWER_CHANGING) {
 		DRM_DEBUG_DRIVER("Skipping backlight disable on vga switch\n");
 		return;
 	}
@@ -1142,7 +1142,7 @@ static int intel_backlight_device_get_brightness(struct backlight_device *bd)
 {
 	struct intel_connector *connector = bl_get_data(bd);
 	struct drm_device *dev = connector->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 hw_level;
 	int ret;
 
@@ -1163,7 +1163,7 @@ static const struct backlight_ops intel_backlight_device_ops = {
 	.get_brightness = intel_backlight_device_get_brightness,
 };
 
-static int intel_backlight_device_register(struct intel_connector *connector)
+int intel_backlight_device_register(struct intel_connector *connector)
 {
 	struct intel_panel *panel = &connector->panel;
 	struct backlight_properties props;
@@ -1225,11 +1225,6 @@ void intel_backlight_device_unregister(struct intel_connector *connector)
 		panel->backlight.device = NULL;
 	}
 }
-#else /* CONFIG_BACKLIGHT_CLASS_DEVICE */
-static int intel_backlight_device_register(struct intel_connector *connector)
-{
-	return 0;
-}
 #endif /* CONFIG_BACKLIGHT_CLASS_DEVICE */
 
 /*
@@ -1321,7 +1316,7 @@ static u32 i9xx_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
 static u32 i965_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
 {
 	struct drm_device *dev = connector->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int clock;
 
 	if (IS_G4X(dev_priv))
@@ -1736,7 +1731,8 @@ intel_panel_init_backlight_funcs(struct intel_panel *panel)
 		panel->backlight.set = bxt_set_backlight;
 		panel->backlight.get = bxt_get_backlight;
 		panel->backlight.hz_to_pwm = bxt_hz_to_pwm;
-	} else if (HAS_PCH_LPT(dev_priv) || HAS_PCH_SPT(dev_priv)) {
+	} else if (HAS_PCH_LPT(dev_priv) || HAS_PCH_SPT(dev_priv) ||
+		   HAS_PCH_KBP(dev_priv)) {
 		panel->backlight.setup = lpt_setup_backlight;
 		panel->backlight.enable = lpt_enable_backlight;
 		panel->backlight.disable = lpt_disable_backlight;
@@ -1809,11 +1805,3 @@ void intel_panel_fini(struct intel_panel *panel)
 		drm_mode_destroy(intel_connector->base.dev,
 				panel->downclock_mode);
 }
-
-void intel_backlight_register(struct drm_device *dev)
-{
-	struct intel_connector *connector;
-
-	for_each_intel_connector(dev, connector)
-		intel_backlight_device_register(connector);
-}
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 658a75659657..5a8ee0c76593 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -57,7 +57,7 @@
 
 static void gen9_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl */
 	I915_WRITE(CHICKEN_PAR1_1,
@@ -83,7 +83,7 @@ static void gen9_init_clock_gating(struct drm_device *dev)
 
 static void bxt_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	gen9_init_clock_gating(dev);
 
@@ -109,7 +109,7 @@ static void bxt_init_clock_gating(struct drm_device *dev)
 
 static void i915_pineview_get_mem_freq(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tmp;
 
 	tmp = I915_READ(CLKCFG);
@@ -148,7 +148,7 @@ static void i915_pineview_get_mem_freq(struct drm_device *dev)
 
 static void i915_ironlake_get_mem_freq(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u16 ddrpll, csipll;
 
 	ddrpll = I915_READ16(DDRMPLL1);
@@ -319,7 +319,7 @@ static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable)
 
 void intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	u32 val;
 
 	if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)) {
@@ -375,7 +375,7 @@ static const int pessimal_latency_ns = 5000;
 static int vlv_get_fifo_size(struct drm_device *dev,
 			      enum pipe pipe, int plane)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int sprite0_start, sprite1_start, size;
 
 	switch (pipe) {
@@ -426,7 +426,7 @@ static int vlv_get_fifo_size(struct drm_device *dev,
 
 static int i9xx_get_fifo_size(struct drm_device *dev, int plane)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dsparb = I915_READ(DSPARB);
 	int size;
 
@@ -442,7 +442,7 @@ static int i9xx_get_fifo_size(struct drm_device *dev, int plane)
 
 static int i830_get_fifo_size(struct drm_device *dev, int plane)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dsparb = I915_READ(DSPARB);
 	int size;
 
@@ -459,7 +459,7 @@ static int i830_get_fifo_size(struct drm_device *dev, int plane)
 
 static int i845_get_fifo_size(struct drm_device *dev, int plane)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dsparb = I915_READ(DSPARB);
 	int size;
 
@@ -637,7 +637,7 @@ static struct drm_crtc *single_enabled_crtc(struct drm_device *dev)
 static void pineview_update_wm(struct drm_crtc *unused_crtc)
 {
 	struct drm_device *dev = unused_crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	const struct cxsr_latency *latency;
 	u32 reg;
@@ -934,7 +934,7 @@ static unsigned int vlv_wm_method2(unsigned int pixel_rate,
 
 static void vlv_setup_wm_latency(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* all latencies in usec */
 	dev_priv->wm.pri_latency[VLV_WM_LEVEL_PM2] = 3;
@@ -1325,7 +1325,7 @@ static void vlv_merge_wm(struct drm_device *dev,
 static void vlv_update_wm(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	struct vlv_wm_values wm = {};
@@ -1381,7 +1381,7 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	static const int sr_latency_ns = 12000;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int planea_wm, planeb_wm, cursora_wm, cursorb_wm;
 	int plane_sr, cursor_sr;
 	unsigned int enabled = 0;
@@ -1438,7 +1438,7 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 static void i965_update_wm(struct drm_crtc *unused_crtc)
 {
 	struct drm_device *dev = unused_crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	int srwm = 1;
 	int cursor_sr = 16;
@@ -1512,7 +1512,7 @@ static void i965_update_wm(struct drm_crtc *unused_crtc)
 static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 {
 	struct drm_device *dev = unused_crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	const struct intel_watermark_params *wm_info;
 	uint32_t fwater_lo;
 	uint32_t fwater_hi;
@@ -1642,7 +1642,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 static void i845_update_wm(struct drm_crtc *unused_crtc)
 {
 	struct drm_device *dev = unused_crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	const struct drm_display_mode *adjusted_mode;
 	uint32_t fwater_lo;
@@ -2070,7 +2070,7 @@ hsw_compute_linetime_wm(const struct intel_crtc_state *cstate)
 
 static void intel_read_wm_latency(struct drm_device *dev, uint16_t wm[8])
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (IS_GEN9(dev)) {
 		uint32_t val;
@@ -2236,7 +2236,7 @@ static void intel_print_wm_latency(struct drm_device *dev,
 static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv,
 				    uint16_t wm[5], uint16_t min)
 {
-	int level, max_level = ilk_wm_max_level(dev_priv->dev);
+	int level, max_level = ilk_wm_max_level(&dev_priv->drm);
 
 	if (wm[0] >= min)
 		return false;
@@ -2250,7 +2250,7 @@ static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv,
 
 static void snb_wm_latency_quirk(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	bool changed;
 
 	/*
@@ -2272,7 +2272,7 @@ static void snb_wm_latency_quirk(struct drm_device *dev)
 
 static void ilk_setup_wm_latency(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_read_wm_latency(dev, dev_priv->wm.pri_latency);
 
@@ -2294,7 +2294,7 @@ static void ilk_setup_wm_latency(struct drm_device *dev)
 
 static void skl_setup_wm_latency(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_read_wm_latency(dev, dev_priv->wm.skl_latency);
 	intel_print_wm_latency(dev, "Gen9 Plane", dev_priv->wm.skl_latency);
@@ -2330,7 +2330,7 @@ static int ilk_compute_pipe_wm(struct intel_crtc_state *cstate)
 	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
 	struct intel_pipe_wm *pipe_wm;
 	struct drm_device *dev = state->dev;
-	const struct drm_i915_private *dev_priv = dev->dev_private;
+	const struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane;
 	struct intel_plane_state *pristate = NULL;
 	struct intel_plane_state *sprstate = NULL;
@@ -2505,7 +2505,7 @@ static void ilk_wm_merge(struct drm_device *dev,
 			 const struct ilk_wm_maximums *max,
 			 struct intel_pipe_wm *merged)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int level, max_level = ilk_wm_max_level(dev);
 	int last_enabled_level = max_level;
 
@@ -2565,7 +2565,7 @@ static int ilk_wm_lp_to_level(int wm_lp, const struct intel_pipe_wm *pipe_wm)
 /* The value we need to program into the WM_LPx latency field */
 static unsigned int ilk_wm_lp_latency(struct drm_device *dev, int level)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		return 2 * level;
@@ -2765,7 +2765,7 @@ static bool _ilk_disable_lp_wm(struct drm_i915_private *dev_priv,
 static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 				struct ilk_wm_values *results)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct ilk_wm_values *previous = &dev_priv->wm.hw;
 	unsigned int dirty;
 	uint32_t val;
@@ -2840,7 +2840,7 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 
 bool ilk_disable_lp_wm(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL);
 }
@@ -3498,7 +3498,6 @@ skl_compute_wm_level(const struct drm_i915_private *dev_priv,
 		     int level,
 		     struct skl_wm_level *result)
 {
-	struct drm_device *dev = dev_priv->dev;
 	struct drm_atomic_state *state = cstate->base.state;
 	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
 	struct drm_plane *plane;
@@ -3514,7 +3513,9 @@ skl_compute_wm_level(const struct drm_i915_private *dev_priv,
 	 */
 	memset(result, 0, sizeof(*result));
 
-	for_each_intel_plane_mask(dev, intel_plane, cstate->base.plane_mask) {
+	for_each_intel_plane_mask(&dev_priv->drm,
+				  intel_plane,
+				  cstate->base.plane_mask) {
 		int i = skl_wm_plane_id(intel_plane);
 
 		plane = &intel_plane->base;
@@ -3595,7 +3596,7 @@ static int skl_build_pipe_wm(struct intel_crtc_state *cstate,
 			     struct skl_pipe_wm *pipe_wm)
 {
 	struct drm_device *dev = cstate->base.crtc->dev;
-	const struct drm_i915_private *dev_priv = dev->dev_private;
+	const struct drm_i915_private *dev_priv = to_i915(dev);
 	int level, max_level = ilk_wm_max_level(dev);
 	int ret;
 
@@ -3682,7 +3683,7 @@ static void skl_ddb_entry_write(struct drm_i915_private *dev_priv,
 static void skl_write_wm_values(struct drm_i915_private *dev_priv,
 				const struct skl_wm_values *new)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_crtc *crtc;
 
 	for_each_intel_crtc(dev, crtc) {
@@ -3779,7 +3780,7 @@ skl_ddb_allocation_included(const struct skl_ddb_allocation *old,
 static void skl_flush_wm_values(struct drm_i915_private *dev_priv,
 				struct skl_wm_values *new_values)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct skl_ddb_allocation *cur_ddb, *new_ddb;
 	bool reallocated[I915_MAX_PIPES] = {};
 	struct intel_crtc *crtc;
@@ -3879,6 +3880,19 @@ static int skl_update_pipe_wm(struct drm_crtc_state *cstate,
 	return 0;
 }
 
+static uint32_t
+pipes_modified(struct drm_atomic_state *state)
+{
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *cstate;
+	uint32_t i, ret = 0;
+
+	for_each_crtc_in_state(state, crtc, cstate, i)
+		ret |= drm_crtc_mask(crtc);
+
+	return ret;
+}
+
 static int
 skl_compute_ddb(struct drm_atomic_state *state)
 {
@@ -3887,7 +3901,7 @@ skl_compute_ddb(struct drm_atomic_state *state)
 	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
 	struct intel_crtc *intel_crtc;
 	struct skl_ddb_allocation *ddb = &intel_state->wm_results.ddb;
-	unsigned realloc_pipes = dev_priv->active_crtcs;
+	uint32_t realloc_pipes = pipes_modified(state);
 	int ret;
 
 	/*
@@ -4002,7 +4016,7 @@ static void skl_update_wm(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct skl_wm_values *results = &dev_priv->wm.skl_results;
 	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
 	struct skl_pipe_wm *pipe_wm = &cstate->wm.skl.optimal;
@@ -4043,7 +4057,7 @@ static void ilk_compute_wm_config(struct drm_device *dev,
 
 static void ilk_program_watermarks(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
 	struct ilk_wm_maximums max;
 	struct intel_wm_config config = {};
@@ -4145,7 +4159,7 @@ static void skl_pipe_wm_active_state(uint32_t val,
 static void skl_pipe_wm_get_hw_state(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct skl_wm_values *hw = &dev_priv->wm.skl_hw;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
@@ -4199,7 +4213,7 @@ static void skl_pipe_wm_get_hw_state(struct drm_crtc *crtc)
 
 void skl_wm_get_hw_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct skl_ddb_allocation *ddb = &dev_priv->wm.skl_hw.ddb;
 	struct drm_crtc *crtc;
 
@@ -4219,7 +4233,7 @@ void skl_wm_get_hw_state(struct drm_device *dev)
 static void ilk_pipe_wm_get_hw_state(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct ilk_wm_values *hw = &dev_priv->wm.hw;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
@@ -4423,7 +4437,7 @@ void vlv_wm_get_hw_state(struct drm_device *dev)
 
 void ilk_wm_get_hw_state(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct ilk_wm_values *hw = &dev_priv->wm.hw;
 	struct drm_crtc *crtc;
 
@@ -4485,7 +4499,7 @@ void ilk_wm_get_hw_state(struct drm_device *dev)
  */
 void intel_update_watermarks(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 
 	if (dev_priv->display.update_wm)
 		dev_priv->display.update_wm(crtc);
@@ -4654,19 +4668,23 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val)
 	new_power = dev_priv->rps.power;
 	switch (dev_priv->rps.power) {
 	case LOW_POWER:
-		if (val > dev_priv->rps.efficient_freq + 1 && val > dev_priv->rps.cur_freq)
+		if (val > dev_priv->rps.efficient_freq + 1 &&
+		    val > dev_priv->rps.cur_freq)
 			new_power = BETWEEN;
 		break;
 
 	case BETWEEN:
-		if (val <= dev_priv->rps.efficient_freq && val < dev_priv->rps.cur_freq)
+		if (val <= dev_priv->rps.efficient_freq &&
+		    val < dev_priv->rps.cur_freq)
 			new_power = LOW_POWER;
-		else if (val >= dev_priv->rps.rp0_freq && val > dev_priv->rps.cur_freq)
+		else if (val >= dev_priv->rps.rp0_freq &&
+			 val > dev_priv->rps.cur_freq)
 			new_power = HIGH_POWER;
 		break;
 
 	case HIGH_POWER:
-		if (val < (dev_priv->rps.rp1_freq + dev_priv->rps.rp0_freq) >> 1 && val < dev_priv->rps.cur_freq)
+		if (val < (dev_priv->rps.rp1_freq + dev_priv->rps.rp0_freq) >> 1 &&
+		    val < dev_priv->rps.cur_freq)
 			new_power = BETWEEN;
 		break;
 	}
@@ -4712,22 +4730,24 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val)
 	}
 
 	I915_WRITE(GEN6_RP_UP_EI,
-		GT_INTERVAL_FROM_US(dev_priv, ei_up));
+		   GT_INTERVAL_FROM_US(dev_priv, ei_up));
 	I915_WRITE(GEN6_RP_UP_THRESHOLD,
-		GT_INTERVAL_FROM_US(dev_priv, (ei_up * threshold_up / 100)));
+		   GT_INTERVAL_FROM_US(dev_priv,
+				       ei_up * threshold_up / 100));
 
 	I915_WRITE(GEN6_RP_DOWN_EI,
-		GT_INTERVAL_FROM_US(dev_priv, ei_down));
+		   GT_INTERVAL_FROM_US(dev_priv, ei_down));
 	I915_WRITE(GEN6_RP_DOWN_THRESHOLD,
-		GT_INTERVAL_FROM_US(dev_priv, (ei_down * threshold_down / 100)));
+		   GT_INTERVAL_FROM_US(dev_priv,
+				       ei_down * threshold_down / 100));
 
-	 I915_WRITE(GEN6_RP_CONTROL,
-		    GEN6_RP_MEDIA_TURBO |
-		    GEN6_RP_MEDIA_HW_NORMAL_MODE |
-		    GEN6_RP_MEDIA_IS_GFX |
-		    GEN6_RP_ENABLE |
-		    GEN6_RP_UP_BUSY_AVG |
-		    GEN6_RP_DOWN_IDLE_AVG);
+	I915_WRITE(GEN6_RP_CONTROL,
+		   GEN6_RP_MEDIA_TURBO |
+		   GEN6_RP_MEDIA_HW_NORMAL_MODE |
+		   GEN6_RP_MEDIA_IS_GFX |
+		   GEN6_RP_ENABLE |
+		   GEN6_RP_UP_BUSY_AVG |
+		   GEN6_RP_DOWN_IDLE_AVG);
 
 	dev_priv->rps.power = new_power;
 	dev_priv->rps.up_threshold = threshold_up;
@@ -4844,12 +4864,27 @@ void gen6_rps_busy(struct drm_i915_private *dev_priv)
 			gen6_rps_reset_ei(dev_priv);
 		I915_WRITE(GEN6_PMINTRMSK,
 			   gen6_rps_pm_mask(dev_priv, dev_priv->rps.cur_freq));
+
+		gen6_enable_rps_interrupts(dev_priv);
+
+		/* Ensure we start at the user's desired frequency */
+		intel_set_rps(dev_priv,
+			      clamp(dev_priv->rps.cur_freq,
+				    dev_priv->rps.min_freq_softlimit,
+				    dev_priv->rps.max_freq_softlimit));
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
 void gen6_rps_idle(struct drm_i915_private *dev_priv)
 {
+	/* Flush our bottom-half so that it does not race with us
+	 * setting the idle frequency and so that it is bounded by
+	 * our rpm wakeref. And then disable the interrupts to stop any
+	 * futher RPS reclocking whilst we are asleep.
+	 */
+	gen6_disable_rps_interrupts(dev_priv);
+
 	mutex_lock(&dev_priv->rps.hw_lock);
 	if (dev_priv->rps.enabled) {
 		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
@@ -4874,7 +4909,7 @@ void gen6_rps_boost(struct drm_i915_private *dev_priv,
 	/* This is intentionally racy! We peek at the state here, then
 	 * validate inside the RPS worker.
 	 */
-	if (!(dev_priv->mm.busy &&
+	if (!(dev_priv->gt.awake &&
 	      dev_priv->rps.enabled &&
 	      dev_priv->rps.cur_freq < dev_priv->rps.max_freq_softlimit))
 		return;
@@ -4890,7 +4925,7 @@ void gen6_rps_boost(struct drm_i915_private *dev_priv,
 		spin_lock_irq(&dev_priv->irq_lock);
 		if (dev_priv->rps.interrupts_enabled) {
 			dev_priv->rps.client_boost = true;
-			queue_work(dev_priv->wq, &dev_priv->rps.work);
+			schedule_work(&dev_priv->rps.work);
 		}
 		spin_unlock_irq(&dev_priv->irq_lock);
 
@@ -4954,14 +4989,15 @@ static void intel_print_rc6_info(struct drm_i915_private *dev_priv, u32 mode)
 			mode = 0;
 	}
 	if (HAS_RC6p(dev_priv))
-		DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s RC6p %s RC6pp %s\n",
-			      onoff(mode & GEN6_RC_CTL_RC6_ENABLE),
-			      onoff(mode & GEN6_RC_CTL_RC6p_ENABLE),
-			      onoff(mode & GEN6_RC_CTL_RC6pp_ENABLE));
+		DRM_DEBUG_DRIVER("Enabling RC6 states: "
+				 "RC6 %s RC6p %s RC6pp %s\n",
+				 onoff(mode & GEN6_RC_CTL_RC6_ENABLE),
+				 onoff(mode & GEN6_RC_CTL_RC6p_ENABLE),
+				 onoff(mode & GEN6_RC_CTL_RC6pp_ENABLE));
 
 	else
-		DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s\n",
-			      onoff(mode & GEN6_RC_CTL_RC6_ENABLE));
+		DRM_DEBUG_DRIVER("Enabling RC6 states: RC6 %s\n",
+				 onoff(mode & GEN6_RC_CTL_RC6_ENABLE));
 }
 
 static bool bxt_check_bios_rc6_setup(struct drm_i915_private *dev_priv)
@@ -4969,9 +5005,20 @@ static bool bxt_check_bios_rc6_setup(struct drm_i915_private *dev_priv)
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	bool enable_rc6 = true;
 	unsigned long rc6_ctx_base;
+	u32 rc_ctl;
+	int rc_sw_target;
+
+	rc_ctl = I915_READ(GEN6_RC_CONTROL);
+	rc_sw_target = (I915_READ(GEN6_RC_STATE) & RC_SW_TARGET_STATE_MASK) >>
+		       RC_SW_TARGET_STATE_SHIFT;
+	DRM_DEBUG_DRIVER("BIOS enabled RC states: "
+			 "HW_CTRL %s HW_RC6 %s SW_TARGET_STATE %x\n",
+			 onoff(rc_ctl & GEN6_RC_CTL_HW_ENABLE),
+			 onoff(rc_ctl & GEN6_RC_CTL_RC6_ENABLE),
+			 rc_sw_target);
 
 	if (!(I915_READ(RC6_LOCATION) & RC6_CTX_IN_DRAM)) {
-		DRM_DEBUG_KMS("RC6 Base location not set properly.\n");
+		DRM_DEBUG_DRIVER("RC6 Base location not set properly.\n");
 		enable_rc6 = false;
 	}
 
@@ -4983,7 +5030,7 @@ static bool bxt_check_bios_rc6_setup(struct drm_i915_private *dev_priv)
 	if (!((rc6_ctx_base >= ggtt->stolen_reserved_base) &&
 	      (rc6_ctx_base + PAGE_SIZE <= ggtt->stolen_reserved_base +
 					ggtt->stolen_reserved_size))) {
-		DRM_DEBUG_KMS("RC6 Base address not as expected.\n");
+		DRM_DEBUG_DRIVER("RC6 Base address not as expected.\n");
 		enable_rc6 = false;
 	}
 
@@ -4991,15 +5038,24 @@ static bool bxt_check_bios_rc6_setup(struct drm_i915_private *dev_priv)
 	      ((I915_READ(PWRCTX_MAXCNT_VCSUNIT0) & IDLE_TIME_MASK) > 1) &&
 	      ((I915_READ(PWRCTX_MAXCNT_BCSUNIT) & IDLE_TIME_MASK) > 1) &&
 	      ((I915_READ(PWRCTX_MAXCNT_VECSUNIT) & IDLE_TIME_MASK) > 1))) {
-		DRM_DEBUG_KMS("Engine Idle wait time not set properly.\n");
+		DRM_DEBUG_DRIVER("Engine Idle wait time not set properly.\n");
 		enable_rc6 = false;
 	}
 
-	if (!(I915_READ(GEN6_RC_CONTROL) & (GEN6_RC_CTL_RC6_ENABLE |
-					    GEN6_RC_CTL_HW_ENABLE)) &&
-	    ((I915_READ(GEN6_RC_CONTROL) & GEN6_RC_CTL_HW_ENABLE) ||
-	     !(I915_READ(GEN6_RC_STATE) & RC6_STATE))) {
-		DRM_DEBUG_KMS("HW/SW RC6 is not enabled by BIOS.\n");
+	if (!I915_READ(GEN8_PUSHBUS_CONTROL) ||
+	    !I915_READ(GEN8_PUSHBUS_ENABLE) ||
+	    !I915_READ(GEN8_PUSHBUS_SHIFT)) {
+		DRM_DEBUG_DRIVER("Pushbus not setup properly.\n");
+		enable_rc6 = false;
+	}
+
+	if (!I915_READ(GEN6_GFXPAUSE)) {
+		DRM_DEBUG_DRIVER("GFX pause not setup properly.\n");
+		enable_rc6 = false;
+	}
+
+	if (!I915_READ(GEN8_MISC_CTRL0)) {
+		DRM_DEBUG_DRIVER("GPM control not setup properly.\n");
 		enable_rc6 = false;
 	}
 
@@ -5031,8 +5087,9 @@ int sanitize_rc6_option(struct drm_i915_private *dev_priv, int enable_rc6)
 			mask = INTEL_RC6_ENABLE;
 
 		if ((enable_rc6 & mask) != enable_rc6)
-			DRM_DEBUG_KMS("Adjusting RC6 mask to %d (requested %d, valid %d)\n",
-				      enable_rc6 & mask, enable_rc6, mask);
+			DRM_DEBUG_DRIVER("Adjusting RC6 mask to %d "
+					 "(requested %d, valid %d)\n",
+					 enable_rc6 & mask, enable_rc6, mask);
 
 		return enable_rc6 & mask;
 	}
@@ -5643,7 +5700,7 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv)
 	u32 pcbr;
 	int pctx_size = 24*1024;
 
-	mutex_lock(&dev_priv->dev->struct_mutex);
+	mutex_lock(&dev_priv->drm.struct_mutex);
 
 	pcbr = I915_READ(VLV_PCBR);
 	if (pcbr) {
@@ -5651,7 +5708,7 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv)
 		int pcbr_offset;
 
 		pcbr_offset = (pcbr & (~4095)) - dev_priv->mm.stolen_base;
-		pctx = i915_gem_object_create_stolen_for_preallocated(dev_priv->dev,
+		pctx = i915_gem_object_create_stolen_for_preallocated(&dev_priv->drm,
 								      pcbr_offset,
 								      I915_GTT_OFFSET_NONE,
 								      pctx_size);
@@ -5668,7 +5725,7 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv)
 	 * overlap with other ranges, such as the frame buffer, protected
 	 * memory, or any other relevant ranges.
 	 */
-	pctx = i915_gem_object_create_stolen(dev_priv->dev, pctx_size);
+	pctx = i915_gem_object_create_stolen(&dev_priv->drm, pctx_size);
 	if (!pctx) {
 		DRM_DEBUG("not enough stolen space for PCTX, disabling\n");
 		goto out;
@@ -5680,7 +5737,7 @@ static void valleyview_setup_pctx(struct drm_i915_private *dev_priv)
 out:
 	DRM_DEBUG_DRIVER("PCBR: 0x%08x\n", I915_READ(VLV_PCBR));
 	dev_priv->vlv_pctx = pctx;
-	mutex_unlock(&dev_priv->dev->struct_mutex);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 }
 
 static void valleyview_cleanup_pctx(struct drm_i915_private *dev_priv)
@@ -6624,9 +6681,9 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv)
 
 	if (IS_IRONLAKE_M(dev_priv)) {
 		ironlake_enable_drps(dev_priv);
-		mutex_lock(&dev_priv->dev->struct_mutex);
+		mutex_lock(&dev_priv->drm.struct_mutex);
 		intel_init_emon(dev_priv);
-		mutex_unlock(&dev_priv->dev->struct_mutex);
+		mutex_unlock(&dev_priv->drm.struct_mutex);
 	} else if (INTEL_INFO(dev_priv)->gen >= 6) {
 		/*
 		 * PCU communication is slow and this doesn't need to be
@@ -6657,7 +6714,7 @@ void intel_reset_gt_powersave(struct drm_i915_private *dev_priv)
 
 static void ibx_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/*
 	 * On Ibex Peak and Cougar Point, we need to disable clock
@@ -6669,7 +6726,7 @@ static void ibx_init_clock_gating(struct drm_device *dev)
 
 static void g4x_disable_trickle_feed(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 
 	for_each_pipe(dev_priv, pipe) {
@@ -6684,7 +6741,7 @@ static void g4x_disable_trickle_feed(struct drm_device *dev)
 
 static void ilk_init_lp_watermarks(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(WM3_LP_ILK, I915_READ(WM3_LP_ILK) & ~WM1_LP_SR_EN);
 	I915_WRITE(WM2_LP_ILK, I915_READ(WM2_LP_ILK) & ~WM1_LP_SR_EN);
@@ -6698,7 +6755,7 @@ static void ilk_init_lp_watermarks(struct drm_device *dev)
 
 static void ironlake_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
 
 	/*
@@ -6772,7 +6829,7 @@ static void ironlake_init_clock_gating(struct drm_device *dev)
 
 static void cpt_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	int pipe;
 	uint32_t val;
 
@@ -6809,7 +6866,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 
 static void gen6_check_mch_setup(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t tmp;
 
 	tmp = I915_READ(MCH_SSKPD);
@@ -6820,7 +6877,7 @@ static void gen6_check_mch_setup(struct drm_device *dev)
 
 static void gen6_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
 
 	I915_WRITE(ILK_DSPCLK_GATE_D, dspclk_gate);
@@ -6935,7 +6992,7 @@ static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
 
 static void lpt_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/*
 	 * TODO: this bit should only be enabled when really needed, then
@@ -6954,7 +7011,7 @@ static void lpt_init_clock_gating(struct drm_device *dev)
 
 static void lpt_suspend_hw(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (HAS_PCH_LPT_LP(dev)) {
 		uint32_t val = I915_READ(SOUTH_DSPCLK_GATE_D);
@@ -6989,7 +7046,7 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv,
 
 static void kabylake_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	gen9_init_clock_gating(dev);
 
@@ -7010,7 +7067,7 @@ static void kabylake_init_clock_gating(struct drm_device *dev)
 
 static void skylake_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	gen9_init_clock_gating(dev);
 
@@ -7025,7 +7082,7 @@ static void skylake_init_clock_gating(struct drm_device *dev)
 
 static void broadwell_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum pipe pipe;
 
 	ilk_init_lp_watermarks(dev);
@@ -7076,7 +7133,7 @@ static void broadwell_init_clock_gating(struct drm_device *dev)
 
 static void haswell_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	ilk_init_lp_watermarks(dev);
 
@@ -7132,7 +7189,7 @@ static void haswell_init_clock_gating(struct drm_device *dev)
 
 static void ivybridge_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t snpcr;
 
 	ilk_init_lp_watermarks(dev);
@@ -7230,7 +7287,7 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 
 static void valleyview_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* WaDisableEarlyCull:vlv */
 	I915_WRITE(_3D_CHICKEN3,
@@ -7312,7 +7369,7 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 
 static void cherryview_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* WaVSRefCountFullforceMissDisable:chv */
 	/* WaDSRefCountFullforceMissDisable:chv */
@@ -7348,7 +7405,7 @@ static void cherryview_init_clock_gating(struct drm_device *dev)
 
 static void g4x_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t dspclk_gate;
 
 	I915_WRITE(RENCLK_GATE_D1, 0);
@@ -7375,7 +7432,7 @@ static void g4x_init_clock_gating(struct drm_device *dev)
 
 static void crestline_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(RENCLK_GATE_D1, I965_RCC_CLOCK_GATE_DISABLE);
 	I915_WRITE(RENCLK_GATE_D2, 0);
@@ -7391,7 +7448,7 @@ static void crestline_init_clock_gating(struct drm_device *dev)
 
 static void broadwater_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(RENCLK_GATE_D1, I965_RCZ_CLOCK_GATE_DISABLE |
 		   I965_RCC_CLOCK_GATE_DISABLE |
@@ -7408,7 +7465,7 @@ static void broadwater_init_clock_gating(struct drm_device *dev)
 
 static void gen3_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 dstate = I915_READ(D_STATE);
 
 	dstate |= DSTATE_PLL_D3_OFF | DSTATE_GFX_CLOCK_GATING |
@@ -7433,7 +7490,7 @@ static void gen3_init_clock_gating(struct drm_device *dev)
 
 static void i85x_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(RENCLK_GATE_D1, SV_CLOCK_GATE_DISABLE);
 
@@ -7447,7 +7504,7 @@ static void i85x_init_clock_gating(struct drm_device *dev)
 
 static void i830_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(DSPCLK_GATE_D, OVRUNIT_CLOCK_GATE_DISABLE);
 
@@ -7458,7 +7515,7 @@ static void i830_init_clock_gating(struct drm_device *dev)
 
 void intel_init_clock_gating(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->display.init_clock_gating(dev);
 }
@@ -7526,7 +7583,7 @@ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
 /* Set up chip specific power management-related functions */
 void intel_init_pm(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_fbc_init(dev_priv);
 
@@ -7604,46 +7661,59 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val
 {
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
 
-	if (I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) {
+	/* GEN6_PCODE_* are outside of the forcewake domain, we can
+	 * use te fw I915_READ variants to reduce the amount of work
+	 * required when reading/writing.
+	 */
+
+	if (I915_READ_FW(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) {
 		DRM_DEBUG_DRIVER("warning: pcode (read) mailbox access failed\n");
 		return -EAGAIN;
 	}
 
-	I915_WRITE(GEN6_PCODE_DATA, *val);
-	I915_WRITE(GEN6_PCODE_DATA1, 0);
-	I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
+	I915_WRITE_FW(GEN6_PCODE_DATA, *val);
+	I915_WRITE_FW(GEN6_PCODE_DATA1, 0);
+	I915_WRITE_FW(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
 
-	if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) == 0,
-		     500)) {
+	if (intel_wait_for_register_fw(dev_priv,
+				       GEN6_PCODE_MAILBOX, GEN6_PCODE_READY, 0,
+				       500)) {
 		DRM_ERROR("timeout waiting for pcode read (%d) to finish\n", mbox);
 		return -ETIMEDOUT;
 	}
 
-	*val = I915_READ(GEN6_PCODE_DATA);
-	I915_WRITE(GEN6_PCODE_DATA, 0);
+	*val = I915_READ_FW(GEN6_PCODE_DATA);
+	I915_WRITE_FW(GEN6_PCODE_DATA, 0);
 
 	return 0;
 }
 
-int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val)
+int sandybridge_pcode_write(struct drm_i915_private *dev_priv,
+			       u32 mbox, u32 val)
 {
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
 
-	if (I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) {
+	/* GEN6_PCODE_* are outside of the forcewake domain, we can
+	 * use te fw I915_READ variants to reduce the amount of work
+	 * required when reading/writing.
+	 */
+
+	if (I915_READ_FW(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) {
 		DRM_DEBUG_DRIVER("warning: pcode (write) mailbox access failed\n");
 		return -EAGAIN;
 	}
 
-	I915_WRITE(GEN6_PCODE_DATA, val);
-	I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
+	I915_WRITE_FW(GEN6_PCODE_DATA, val);
+	I915_WRITE_FW(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
 
-	if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) == 0,
-		     500)) {
+	if (intel_wait_for_register_fw(dev_priv,
+				       GEN6_PCODE_MAILBOX, GEN6_PCODE_READY, 0,
+				       500)) {
 		DRM_ERROR("timeout waiting for pcode write (%d) to finish\n", mbox);
 		return -ETIMEDOUT;
 	}
 
-	I915_WRITE(GEN6_PCODE_DATA, 0);
+	I915_WRITE_FW(GEN6_PCODE_DATA, 0);
 
 	return 0;
 }
@@ -7713,7 +7783,7 @@ static void __intel_rps_boost_work(struct work_struct *work)
 	struct request_boost *boost = container_of(work, struct request_boost, work);
 	struct drm_i915_gem_request *req = boost->req;
 
-	if (!i915_gem_request_completed(req, true))
+	if (!i915_gem_request_completed(req))
 		gen6_rps_boost(req->i915, NULL, req->emitted_jiffies);
 
 	i915_gem_request_unreference(req);
@@ -7727,7 +7797,7 @@ void intel_queue_rps_boost_for_request(struct drm_i915_gem_request *req)
 	if (req == NULL || INTEL_GEN(req->i915) < 6)
 		return;
 
-	if (i915_gem_request_completed(req, true))
+	if (i915_gem_request_completed(req))
 		return;
 
 	boost = kmalloc(sizeof(*boost), GFP_ATOMIC);
@@ -7743,7 +7813,7 @@ void intel_queue_rps_boost_for_request(struct drm_i915_gem_request *req)
 
 void intel_pm_setup(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	mutex_init(&dev_priv->rps.hw_lock);
 	spin_lock_init(&dev_priv->rps.client_lock);
diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c
index 29a09bf6bd18..68bd0bb34817 100644
--- a/drivers/gpu/drm/i915/intel_psr.c
+++ b/drivers/gpu/drm/i915/intel_psr.c
@@ -63,7 +63,7 @@ static bool is_edp_psr(struct intel_dp *intel_dp)
 
 static bool vlv_is_psr_active_on_pipe(struct drm_device *dev, int pipe)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t val;
 
 	val = I915_READ(VLV_PSRSTAT(pipe)) &
@@ -77,7 +77,7 @@ static void intel_psr_write_vsc(struct intel_dp *intel_dp,
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc);
 	enum transcoder cpu_transcoder = crtc->config->cpu_transcoder;
 	i915_reg_t ctl_reg = HSW_TVIDEO_DIP_CTL(cpu_transcoder);
@@ -107,7 +107,7 @@ static void vlv_psr_setup_vsc(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = intel_dig_port->base.base.crtc;
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 	uint32_t val;
@@ -173,7 +173,7 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	uint32_t aux_clock_divider;
 	i915_reg_t aux_ctl_reg;
 	static const uint8_t aux_msg[] = {
@@ -220,7 +220,7 @@ static void vlv_psr_enable_source(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 
@@ -235,7 +235,7 @@ static void vlv_psr_activate(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 
@@ -252,7 +252,7 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	uint32_t max_sleep_time = 0x1f;
 	/* Lately it was identified that depending on panel idle frame count
@@ -324,7 +324,7 @@ static bool intel_psr_match_conditions(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
@@ -378,7 +378,7 @@ static void intel_psr_activate(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE);
 	WARN_ON(dev_priv->psr.active);
@@ -407,7 +407,7 @@ void intel_psr_enable(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(intel_dig_port->base.base.crtc);
 
 	if (!HAS_PSR(dev)) {
@@ -494,15 +494,18 @@ static void vlv_psr_disable(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(intel_dig_port->base.base.crtc);
 	uint32_t val;
 
 	if (dev_priv->psr.active) {
 		/* Put VLV PSR back to PSR_state 0 that is PSR Disabled. */
-		if (wait_for((I915_READ(VLV_PSRSTAT(intel_crtc->pipe)) &
-			      VLV_EDP_PSR_IN_TRANS) == 0, 1))
+		if (intel_wait_for_register(dev_priv,
+					    VLV_PSRSTAT(intel_crtc->pipe),
+					    VLV_EDP_PSR_IN_TRANS,
+					    0,
+					    1))
 			WARN(1, "PSR transition took longer than expected\n");
 
 		val = I915_READ(VLV_PSRCTL(intel_crtc->pipe));
@@ -521,16 +524,18 @@ static void hsw_psr_disable(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	if (dev_priv->psr.active) {
 		I915_WRITE(EDP_PSR_CTL,
 			   I915_READ(EDP_PSR_CTL) & ~EDP_PSR_ENABLE);
 
 		/* Wait till PSR is idle */
-		if (_wait_for((I915_READ(EDP_PSR_STATUS_CTL) &
-			       EDP_PSR_STATUS_STATE_MASK) == 0,
-			       2 * USEC_PER_SEC, 10 * USEC_PER_MSEC))
+		if (intel_wait_for_register(dev_priv,
+					    EDP_PSR_STATUS_CTL,
+					    EDP_PSR_STATUS_STATE_MASK,
+					    0,
+					    2000))
 			DRM_ERROR("Timed out waiting for PSR Idle State\n");
 
 		dev_priv->psr.active = false;
@@ -549,7 +554,7 @@ void intel_psr_disable(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	mutex_lock(&dev_priv->psr.lock);
 	if (!dev_priv->psr.enabled) {
@@ -586,14 +591,20 @@ static void intel_psr_work(struct work_struct *work)
 	 * and be ready for re-enable.
 	 */
 	if (HAS_DDI(dev_priv)) {
-		if (wait_for((I915_READ(EDP_PSR_STATUS_CTL) &
-			      EDP_PSR_STATUS_STATE_MASK) == 0, 50)) {
+		if (intel_wait_for_register(dev_priv,
+					    EDP_PSR_STATUS_CTL,
+					    EDP_PSR_STATUS_STATE_MASK,
+					    0,
+					    50)) {
 			DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n");
 			return;
 		}
 	} else {
-		if (wait_for((I915_READ(VLV_PSRSTAT(pipe)) &
-			      VLV_EDP_PSR_IN_TRANS) == 0, 1)) {
+		if (intel_wait_for_register(dev_priv,
+					    VLV_PSRSTAT(pipe),
+					    VLV_EDP_PSR_IN_TRANS,
+					    0,
+					    1)) {
 			DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n");
 			return;
 		}
@@ -619,7 +630,7 @@ unlock:
 
 static void intel_psr_exit(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_dp *intel_dp = dev_priv->psr.enabled;
 	struct drm_crtc *crtc = dp_to_dig_port(intel_dp)->base.base.crtc;
 	enum pipe pipe = to_intel_crtc(crtc)->pipe;
@@ -674,7 +685,7 @@ static void intel_psr_exit(struct drm_device *dev)
 void intel_psr_single_frame_update(struct drm_device *dev,
 				   unsigned frontbuffer_bits)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	enum pipe pipe;
 	u32 val;
@@ -722,7 +733,7 @@ void intel_psr_single_frame_update(struct drm_device *dev,
 void intel_psr_invalidate(struct drm_device *dev,
 			  unsigned frontbuffer_bits)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	enum pipe pipe;
 
@@ -760,7 +771,7 @@ void intel_psr_invalidate(struct drm_device *dev,
 void intel_psr_flush(struct drm_device *dev,
 		     unsigned frontbuffer_bits, enum fb_op_origin origin)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_crtc *crtc;
 	enum pipe pipe;
 
@@ -796,7 +807,7 @@ void intel_psr_flush(struct drm_device *dev,
  */
 void intel_psr_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	dev_priv->psr_mmio_base = IS_HASWELL(dev_priv) ?
 		HSW_EDP_PSR_BASE : BDW_EDP_PSR_BASE;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index fedd27049814..61e00bf9e87f 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -58,18 +58,10 @@ void intel_ring_update_space(struct intel_ringbuffer *ringbuf)
 					    ringbuf->tail, ringbuf->size);
 }
 
-bool intel_engine_stopped(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	return dev_priv->gpu_error.stop_rings & intel_engine_flag(engine);
-}
-
 static void __intel_ring_advance(struct intel_engine_cs *engine)
 {
 	struct intel_ringbuffer *ringbuf = engine->buffer;
 	ringbuf->tail &= ringbuf->size - 1;
-	if (intel_engine_stopped(engine))
-		return;
 	engine->write_tail(engine, ringbuf->tail);
 }
 
@@ -515,8 +507,9 @@ static void intel_ring_setup_status_page(struct intel_engine_cs *engine)
 		I915_WRITE(reg,
 			   _MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE |
 					      INSTPM_SYNC_FLUSH));
-		if (wait_for((I915_READ(reg) & INSTPM_SYNC_FLUSH) == 0,
-			     1000))
+		if (intel_wait_for_register(dev_priv,
+					    reg, INSTPM_SYNC_FLUSH, 0,
+					    1000))
 			DRM_ERROR("%s: wait for SyncFlush to complete for TLB invalidation timed out\n",
 				  engine->name);
 	}
@@ -528,7 +521,11 @@ static bool stop_ring(struct intel_engine_cs *engine)
 
 	if (!IS_GEN2(dev_priv)) {
 		I915_WRITE_MODE(engine, _MASKED_BIT_ENABLE(STOP_RING));
-		if (wait_for((I915_READ_MODE(engine) & MODE_IDLE) != 0, 1000)) {
+		if (intel_wait_for_register(dev_priv,
+					    RING_MI_MODE(engine->mmio_base),
+					    MODE_IDLE,
+					    MODE_IDLE,
+					    1000)) {
 			DRM_ERROR("%s : timed out trying to stop ring\n",
 				  engine->name);
 			/* Sometimes we observe that the idle flag is not
@@ -643,58 +640,42 @@ out:
 	return ret;
 }
 
-void
-intel_fini_pipe_control(struct intel_engine_cs *engine)
+void intel_fini_pipe_control(struct intel_engine_cs *engine)
 {
 	if (engine->scratch.obj == NULL)
 		return;
 
-	if (INTEL_GEN(engine->i915) >= 5) {
-		kunmap(sg_page(engine->scratch.obj->pages->sgl));
-		i915_gem_object_ggtt_unpin(engine->scratch.obj);
-	}
-
+	i915_gem_object_ggtt_unpin(engine->scratch.obj);
 	drm_gem_object_unreference(&engine->scratch.obj->base);
 	engine->scratch.obj = NULL;
 }
 
-int
-intel_init_pipe_control(struct intel_engine_cs *engine)
+int intel_init_pipe_control(struct intel_engine_cs *engine, int size)
 {
+	struct drm_i915_gem_object *obj;
 	int ret;
 
 	WARN_ON(engine->scratch.obj);
 
-	engine->scratch.obj = i915_gem_object_create(engine->i915->dev, 4096);
-	if (IS_ERR(engine->scratch.obj)) {
-		DRM_ERROR("Failed to allocate seqno page\n");
-		ret = PTR_ERR(engine->scratch.obj);
-		engine->scratch.obj = NULL;
+	obj = i915_gem_object_create_stolen(&engine->i915->drm, size);
+	if (!obj)
+		obj = i915_gem_object_create(&engine->i915->drm, size);
+	if (IS_ERR(obj)) {
+		DRM_ERROR("Failed to allocate scratch page\n");
+		ret = PTR_ERR(obj);
 		goto err;
 	}
 
-	ret = i915_gem_object_set_cache_level(engine->scratch.obj,
-					      I915_CACHE_LLC);
+	ret = i915_gem_obj_ggtt_pin(obj, 4096, PIN_HIGH);
 	if (ret)
 		goto err_unref;
 
-	ret = i915_gem_obj_ggtt_pin(engine->scratch.obj, 4096, 0);
-	if (ret)
-		goto err_unref;
-
-	engine->scratch.gtt_offset = i915_gem_obj_ggtt_offset(engine->scratch.obj);
-	engine->scratch.cpu_page = kmap(sg_page(engine->scratch.obj->pages->sgl));
-	if (engine->scratch.cpu_page == NULL) {
-		ret = -ENOMEM;
-		goto err_unpin;
-	}
-
+	engine->scratch.obj = obj;
+	engine->scratch.gtt_offset = i915_gem_obj_ggtt_offset(obj);
 	DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n",
 			 engine->name, engine->scratch.gtt_offset);
 	return 0;
 
-err_unpin:
-	i915_gem_object_ggtt_unpin(engine->scratch.obj);
 err_unref:
 	drm_gem_object_unreference(&engine->scratch.obj->base);
 err:
@@ -1324,8 +1305,7 @@ static int init_render_ring(struct intel_engine_cs *engine)
 	if (IS_GEN(dev_priv, 6, 7))
 		I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
 
-	if (HAS_L3_DPF(dev_priv))
-		I915_WRITE_IMR(engine, ~GT_PARITY_ERROR(dev_priv));
+	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
 
 	return init_workarounds_ring(engine);
 }
@@ -1362,19 +1342,17 @@ static int gen8_rcs_signal(struct drm_i915_gem_request *signaller_req,
 		return ret;
 
 	for_each_engine_id(waiter, dev_priv, id) {
-		u32 seqno;
 		u64 gtt_offset = signaller->semaphore.signal_ggtt[id];
 		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
 			continue;
 
-		seqno = i915_gem_request_get_seqno(signaller_req);
 		intel_ring_emit(signaller, GFX_OP_PIPE_CONTROL(6));
 		intel_ring_emit(signaller, PIPE_CONTROL_GLOBAL_GTT_IVB |
 					   PIPE_CONTROL_QW_WRITE |
 					   PIPE_CONTROL_CS_STALL);
 		intel_ring_emit(signaller, lower_32_bits(gtt_offset));
 		intel_ring_emit(signaller, upper_32_bits(gtt_offset));
-		intel_ring_emit(signaller, seqno);
+		intel_ring_emit(signaller, signaller_req->seqno);
 		intel_ring_emit(signaller, 0);
 		intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL |
 					   MI_SEMAPHORE_TARGET(waiter->hw_id));
@@ -1403,18 +1381,16 @@ static int gen8_xcs_signal(struct drm_i915_gem_request *signaller_req,
 		return ret;
 
 	for_each_engine_id(waiter, dev_priv, id) {
-		u32 seqno;
 		u64 gtt_offset = signaller->semaphore.signal_ggtt[id];
 		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
 			continue;
 
-		seqno = i915_gem_request_get_seqno(signaller_req);
 		intel_ring_emit(signaller, (MI_FLUSH_DW + 1) |
 					   MI_FLUSH_DW_OP_STOREDW);
 		intel_ring_emit(signaller, lower_32_bits(gtt_offset) |
 					   MI_FLUSH_DW_USE_GTT);
 		intel_ring_emit(signaller, upper_32_bits(gtt_offset));
-		intel_ring_emit(signaller, seqno);
+		intel_ring_emit(signaller, signaller_req->seqno);
 		intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL |
 					   MI_SEMAPHORE_TARGET(waiter->hw_id));
 		intel_ring_emit(signaller, 0);
@@ -1445,11 +1421,9 @@ static int gen6_signal(struct drm_i915_gem_request *signaller_req,
 		i915_reg_t mbox_reg = signaller->semaphore.mbox.signal[id];
 
 		if (i915_mmio_reg_valid(mbox_reg)) {
-			u32 seqno = i915_gem_request_get_seqno(signaller_req);
-
 			intel_ring_emit(signaller, MI_LOAD_REGISTER_IMM(1));
 			intel_ring_emit_reg(signaller, mbox_reg);
-			intel_ring_emit(signaller, seqno);
+			intel_ring_emit(signaller, signaller_req->seqno);
 		}
 	}
 
@@ -1485,7 +1459,7 @@ gen6_add_request(struct drm_i915_gem_request *req)
 	intel_ring_emit(engine, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(engine,
 			I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-	intel_ring_emit(engine, i915_gem_request_get_seqno(req));
+	intel_ring_emit(engine, req->seqno);
 	intel_ring_emit(engine, MI_USER_INTERRUPT);
 	__intel_ring_advance(engine);
 
@@ -1542,6 +1516,7 @@ gen8_ring_sync(struct drm_i915_gem_request *waiter_req,
 {
 	struct intel_engine_cs *waiter = waiter_req->engine;
 	struct drm_i915_private *dev_priv = waiter_req->i915;
+	u64 offset = GEN8_WAIT_OFFSET(waiter, signaller->id);
 	struct i915_hw_ppgtt *ppgtt;
 	int ret;
 
@@ -1553,10 +1528,8 @@ gen8_ring_sync(struct drm_i915_gem_request *waiter_req,
 				MI_SEMAPHORE_GLOBAL_GTT |
 				MI_SEMAPHORE_SAD_GTE_SDD);
 	intel_ring_emit(waiter, seqno);
-	intel_ring_emit(waiter,
-			lower_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id)));
-	intel_ring_emit(waiter,
-			upper_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id)));
+	intel_ring_emit(waiter, lower_32_bits(offset));
+	intel_ring_emit(waiter, upper_32_bits(offset));
 	intel_ring_advance(waiter);
 
 	/* When the !RCS engines idle waiting upon a semaphore, they lose their
@@ -1611,66 +1584,22 @@ gen6_ring_sync(struct drm_i915_gem_request *waiter_req,
 	return 0;
 }
 
-#define PIPE_CONTROL_FLUSH(ring__, addr__)					\
-do {									\
-	intel_ring_emit(ring__, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |		\
-		 PIPE_CONTROL_DEPTH_STALL);				\
-	intel_ring_emit(ring__, (addr__) | PIPE_CONTROL_GLOBAL_GTT);			\
-	intel_ring_emit(ring__, 0);							\
-	intel_ring_emit(ring__, 0);							\
-} while (0)
-
-static int
-pc_render_add_request(struct drm_i915_gem_request *req)
+static void
+gen5_seqno_barrier(struct intel_engine_cs *ring)
 {
-	struct intel_engine_cs *engine = req->engine;
-	u32 scratch_addr = engine->scratch.gtt_offset + 2 * CACHELINE_BYTES;
-	int ret;
-
-	/* For Ironlake, MI_USER_INTERRUPT was deprecated and apparently
-	 * incoherent with writes to memory, i.e. completely fubar,
-	 * so we need to use PIPE_NOTIFY instead.
+	/* MI_STORE are internally buffered by the GPU and not flushed
+	 * either by MI_FLUSH or SyncFlush or any other combination of
+	 * MI commands.
 	 *
-	 * However, we also need to workaround the qword write
-	 * incoherence by flushing the 6 PIPE_NOTIFY buffers out to
-	 * memory before requesting an interrupt.
+	 * "Only the submission of the store operation is guaranteed.
+	 * The write result will be complete (coherent) some time later
+	 * (this is practically a finite period but there is no guaranteed
+	 * latency)."
+	 *
+	 * Empirically, we observe that we need a delay of at least 75us to
+	 * be sure that the seqno write is visible by the CPU.
 	 */
-	ret = intel_ring_begin(req, 32);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(engine,
-			GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |
-			PIPE_CONTROL_WRITE_FLUSH |
-			PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE);
-	intel_ring_emit(engine,
-			engine->scratch.gtt_offset | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(engine, i915_gem_request_get_seqno(req));
-	intel_ring_emit(engine, 0);
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-	scratch_addr += 2 * CACHELINE_BYTES; /* write to separate cachelines */
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-	scratch_addr += 2 * CACHELINE_BYTES;
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-	scratch_addr += 2 * CACHELINE_BYTES;
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-	scratch_addr += 2 * CACHELINE_BYTES;
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-	scratch_addr += 2 * CACHELINE_BYTES;
-	PIPE_CONTROL_FLUSH(engine, scratch_addr);
-
-	intel_ring_emit(engine,
-			GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |
-			PIPE_CONTROL_WRITE_FLUSH |
-			PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE |
-			PIPE_CONTROL_NOTIFY);
-	intel_ring_emit(engine,
-			engine->scratch.gtt_offset | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(engine, i915_gem_request_get_seqno(req));
-	intel_ring_emit(engine, 0);
-	__intel_ring_advance(engine);
-
-	return 0;
+	usleep_range(125, 250);
 }
 
 static void
@@ -1698,127 +1627,54 @@ gen6_seqno_barrier(struct intel_engine_cs *engine)
 	spin_unlock_irq(&dev_priv->uncore.lock);
 }
 
-static u32
-ring_get_seqno(struct intel_engine_cs *engine)
-{
-	return intel_read_status_page(engine, I915_GEM_HWS_INDEX);
-}
-
 static void
-ring_set_seqno(struct intel_engine_cs *engine, u32 seqno)
-{
-	intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
-}
-
-static u32
-pc_render_get_seqno(struct intel_engine_cs *engine)
+gen5_irq_enable(struct intel_engine_cs *engine)
 {
-	return engine->scratch.cpu_page[0];
+	gen5_enable_gt_irq(engine->i915, engine->irq_enable_mask);
 }
 
 static void
-pc_render_set_seqno(struct intel_engine_cs *engine, u32 seqno)
+gen5_irq_disable(struct intel_engine_cs *engine)
 {
-	engine->scratch.cpu_page[0] = seqno;
-}
-
-static bool
-gen5_ring_get_irq(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return false;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0)
-		gen5_enable_gt_irq(dev_priv, engine->irq_enable_mask);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	gen5_disable_gt_irq(engine->i915, engine->irq_enable_mask);
 }
 
 static void
-gen5_ring_put_irq(struct intel_engine_cs *engine)
+i9xx_irq_enable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0)
-		gen5_disable_gt_irq(dev_priv, engine->irq_enable_mask);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-}
-
-static bool
-i9xx_ring_get_irq(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	if (!intel_irqs_enabled(dev_priv))
-		return false;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		dev_priv->irq_mask &= ~engine->irq_enable_mask;
-		I915_WRITE(IMR, dev_priv->irq_mask);
-		POSTING_READ(IMR);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	dev_priv->irq_mask &= ~engine->irq_enable_mask;
+	I915_WRITE(IMR, dev_priv->irq_mask);
+	POSTING_READ_FW(RING_IMR(engine->mmio_base));
 }
 
 static void
-i9xx_ring_put_irq(struct intel_engine_cs *engine)
+i9xx_irq_disable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		dev_priv->irq_mask |= engine->irq_enable_mask;
-		I915_WRITE(IMR, dev_priv->irq_mask);
-		POSTING_READ(IMR);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	dev_priv->irq_mask |= engine->irq_enable_mask;
+	I915_WRITE(IMR, dev_priv->irq_mask);
 }
 
-static bool
-i8xx_ring_get_irq(struct intel_engine_cs *engine)
+static void
+i8xx_irq_enable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	if (!intel_irqs_enabled(dev_priv))
-		return false;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		dev_priv->irq_mask &= ~engine->irq_enable_mask;
-		I915_WRITE16(IMR, dev_priv->irq_mask);
-		POSTING_READ16(IMR);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	dev_priv->irq_mask &= ~engine->irq_enable_mask;
+	I915_WRITE16(IMR, dev_priv->irq_mask);
+	POSTING_READ16(RING_IMR(engine->mmio_base));
 }
 
 static void
-i8xx_ring_put_irq(struct intel_engine_cs *engine)
+i8xx_irq_disable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		dev_priv->irq_mask |= engine->irq_enable_mask;
-		I915_WRITE16(IMR, dev_priv->irq_mask);
-		POSTING_READ16(IMR);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	dev_priv->irq_mask |= engine->irq_enable_mask;
+	I915_WRITE16(IMR, dev_priv->irq_mask);
 }
 
 static int
@@ -1852,129 +1708,68 @@ i9xx_add_request(struct drm_i915_gem_request *req)
 	intel_ring_emit(engine, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(engine,
 			I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-	intel_ring_emit(engine, i915_gem_request_get_seqno(req));
+	intel_ring_emit(engine, req->seqno);
 	intel_ring_emit(engine, MI_USER_INTERRUPT);
 	__intel_ring_advance(engine);
 
 	return 0;
 }
 
-static bool
-gen6_ring_get_irq(struct intel_engine_cs *engine)
+static void
+gen6_irq_enable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
-
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return false;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		if (HAS_L3_DPF(dev_priv) && engine->id == RCS)
-			I915_WRITE_IMR(engine,
-				       ~(engine->irq_enable_mask |
-					 GT_PARITY_ERROR(dev_priv)));
-		else
-			I915_WRITE_IMR(engine, ~engine->irq_enable_mask);
-		gen5_enable_gt_irq(dev_priv, engine->irq_enable_mask);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
-	return true;
+	I915_WRITE_IMR(engine,
+		       ~(engine->irq_enable_mask |
+			 engine->irq_keep_mask));
+	gen5_enable_gt_irq(dev_priv, engine->irq_enable_mask);
 }
 
 static void
-gen6_ring_put_irq(struct intel_engine_cs *engine)
+gen6_irq_disable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		if (HAS_L3_DPF(dev_priv) && engine->id == RCS)
-			I915_WRITE_IMR(engine, ~GT_PARITY_ERROR(dev_priv));
-		else
-			I915_WRITE_IMR(engine, ~0);
-		gen5_disable_gt_irq(dev_priv, engine->irq_enable_mask);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
+	gen5_disable_gt_irq(dev_priv, engine->irq_enable_mask);
 }
 
-static bool
-hsw_vebox_get_irq(struct intel_engine_cs *engine)
+static void
+hsw_vebox_irq_enable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return false;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		I915_WRITE_IMR(engine, ~engine->irq_enable_mask);
-		gen6_enable_pm_irq(dev_priv, engine->irq_enable_mask);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	I915_WRITE_IMR(engine, ~engine->irq_enable_mask);
+	gen6_enable_pm_irq(dev_priv, engine->irq_enable_mask);
 }
 
 static void
-hsw_vebox_put_irq(struct intel_engine_cs *engine)
+hsw_vebox_irq_disable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		I915_WRITE_IMR(engine, ~0);
-		gen6_disable_pm_irq(dev_priv, engine->irq_enable_mask);
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	I915_WRITE_IMR(engine, ~0);
+	gen6_disable_pm_irq(dev_priv, engine->irq_enable_mask);
 }
 
-static bool
-gen8_ring_get_irq(struct intel_engine_cs *engine)
+static void
+gen8_irq_enable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return false;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (engine->irq_refcount++ == 0) {
-		if (HAS_L3_DPF(dev_priv) && engine->id == RCS) {
-			I915_WRITE_IMR(engine,
-				       ~(engine->irq_enable_mask |
-					 GT_RENDER_L3_PARITY_ERROR_INTERRUPT));
-		} else {
-			I915_WRITE_IMR(engine, ~engine->irq_enable_mask);
-		}
-		POSTING_READ(RING_IMR(engine->mmio_base));
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return true;
+	I915_WRITE_IMR(engine,
+		       ~(engine->irq_enable_mask |
+			 engine->irq_keep_mask));
+	POSTING_READ_FW(RING_IMR(engine->mmio_base));
 }
 
 static void
-gen8_ring_put_irq(struct intel_engine_cs *engine)
+gen8_irq_disable(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--engine->irq_refcount == 0) {
-		if (HAS_L3_DPF(dev_priv) && engine->id == RCS) {
-			I915_WRITE_IMR(engine,
-				       ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT);
-		} else {
-			I915_WRITE_IMR(engine, ~0);
-		}
-		POSTING_READ(RING_IMR(engine->mmio_base));
-	}
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
 }
 
 static int
@@ -2093,7 +1888,7 @@ static void cleanup_phys_status_page(struct intel_engine_cs *engine)
 	if (!dev_priv->status_page_dmah)
 		return;
 
-	drm_pci_free(dev_priv->dev, dev_priv->status_page_dmah);
+	drm_pci_free(&dev_priv->drm, dev_priv->status_page_dmah);
 	engine->status_page.page_addr = NULL;
 }
 
@@ -2119,7 +1914,7 @@ static int init_status_page(struct intel_engine_cs *engine)
 		unsigned flags;
 		int ret;
 
-		obj = i915_gem_object_create(engine->i915->dev, 4096);
+		obj = i915_gem_object_create(&engine->i915->drm, 4096);
 		if (IS_ERR(obj)) {
 			DRM_ERROR("Failed to allocate status page\n");
 			return PTR_ERR(obj);
@@ -2168,7 +1963,7 @@ static int init_phys_status_page(struct intel_engine_cs *engine)
 
 	if (!dev_priv->status_page_dmah) {
 		dev_priv->status_page_dmah =
-			drm_pci_alloc(dev_priv->dev, PAGE_SIZE, PAGE_SIZE);
+			drm_pci_alloc(&dev_priv->drm, PAGE_SIZE, PAGE_SIZE);
 		if (!dev_priv->status_page_dmah)
 			return -ENOMEM;
 	}
@@ -2301,7 +2096,7 @@ intel_engine_create_ringbuffer(struct intel_engine_cs *engine, int size)
 	ring->last_retired_head = -1;
 	intel_ring_update_space(ring);
 
-	ret = intel_alloc_ringbuffer_obj(engine->i915->dev, ring);
+	ret = intel_alloc_ringbuffer_obj(&engine->i915->drm, ring);
 	if (ret) {
 		DRM_DEBUG_DRIVER("Failed to allocate ringbuffer %s: %d\n",
 				 engine->name, ret);
@@ -2321,6 +2116,57 @@ intel_ringbuffer_free(struct intel_ringbuffer *ring)
 	kfree(ring);
 }
 
+static int intel_ring_context_pin(struct i915_gem_context *ctx,
+				  struct intel_engine_cs *engine)
+{
+	struct intel_context *ce = &ctx->engine[engine->id];
+	int ret;
+
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
+
+	if (ce->pin_count++)
+		return 0;
+
+	if (ce->state) {
+		ret = i915_gem_obj_ggtt_pin(ce->state, ctx->ggtt_alignment, 0);
+		if (ret)
+			goto error;
+	}
+
+	/* The kernel context is only used as a placeholder for flushing the
+	 * active context. It is never used for submitting user rendering and
+	 * as such never requires the golden render context, and so we can skip
+	 * emitting it when we switch to the kernel context. This is required
+	 * as during eviction we cannot allocate and pin the renderstate in
+	 * order to initialise the context.
+	 */
+	if (ctx == ctx->i915->kernel_context)
+		ce->initialised = true;
+
+	i915_gem_context_reference(ctx);
+	return 0;
+
+error:
+	ce->pin_count = 0;
+	return ret;
+}
+
+static void intel_ring_context_unpin(struct i915_gem_context *ctx,
+				     struct intel_engine_cs *engine)
+{
+	struct intel_context *ce = &ctx->engine[engine->id];
+
+	lockdep_assert_held(&ctx->i915->drm.struct_mutex);
+
+	if (--ce->pin_count)
+		return;
+
+	if (ce->state)
+		i915_gem_object_ggtt_unpin(ce->state);
+
+	i915_gem_context_unreference(ctx);
+}
+
 static int intel_init_ring_buffer(struct drm_device *dev,
 				  struct intel_engine_cs *engine)
 {
@@ -2339,7 +2185,20 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 	memset(engine->semaphore.sync_seqno, 0,
 	       sizeof(engine->semaphore.sync_seqno));
 
-	init_waitqueue_head(&engine->irq_queue);
+	ret = intel_engine_init_breadcrumbs(engine);
+	if (ret)
+		goto error;
+
+	/* We may need to do things with the shrinker which
+	 * require us to immediately switch back to the default
+	 * context. This can cause a problem as pinning the
+	 * default context also requires GTT space which may not
+	 * be available. To avoid this we always pin the default
+	 * context.
+	 */
+	ret = intel_ring_context_pin(dev_priv->kernel_context, engine);
+	if (ret)
+		goto error;
 
 	ringbuf = intel_engine_create_ringbuffer(engine, 32 * PAGE_SIZE);
 	if (IS_ERR(ringbuf)) {
@@ -2408,6 +2267,10 @@ void intel_cleanup_engine(struct intel_engine_cs *engine)
 
 	i915_cmd_parser_fini_ring(engine);
 	i915_gem_batch_pool_fini(&engine->batch_pool);
+	intel_engine_fini_breadcrumbs(engine);
+
+	intel_ring_context_unpin(dev_priv->kernel_context, engine);
+
 	engine->i915 = NULL;
 }
 
@@ -2603,10 +2466,19 @@ void intel_ring_init_seqno(struct intel_engine_cs *engine, u32 seqno)
 	memset(engine->semaphore.sync_seqno, 0,
 	       sizeof(engine->semaphore.sync_seqno));
 
-	engine->set_seqno(engine, seqno);
+	intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
+	if (engine->irq_seqno_barrier)
+		engine->irq_seqno_barrier(engine);
 	engine->last_submitted_seqno = seqno;
 
 	engine->hangcheck.seqno = seqno;
+
+	/* After manually advancing the seqno, fake the interrupt in case
+	 * there are any waiters for that seqno.
+	 */
+	rcu_read_lock();
+	intel_engine_wakeup(engine);
+	rcu_read_unlock();
 }
 
 static void gen6_bsd_ring_write_tail(struct intel_engine_cs *engine,
@@ -2614,32 +2486,38 @@ static void gen6_bsd_ring_write_tail(struct intel_engine_cs *engine,
 {
 	struct drm_i915_private *dev_priv = engine->i915;
 
+	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
+
        /* Every tail move must follow the sequence below */
 
 	/* Disable notification that the ring is IDLE. The GT
 	 * will then assume that it is busy and bring it out of rc6.
 	 */
-	I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
-		   _MASKED_BIT_ENABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
+	I915_WRITE_FW(GEN6_BSD_SLEEP_PSMI_CONTROL,
+		      _MASKED_BIT_ENABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
 
 	/* Clear the context id. Here be magic! */
-	I915_WRITE64(GEN6_BSD_RNCID, 0x0);
+	I915_WRITE64_FW(GEN6_BSD_RNCID, 0x0);
 
 	/* Wait for the ring not to be idle, i.e. for it to wake up. */
-	if (wait_for((I915_READ(GEN6_BSD_SLEEP_PSMI_CONTROL) &
-		      GEN6_BSD_SLEEP_INDICATOR) == 0,
-		     50))
+	if (intel_wait_for_register_fw(dev_priv,
+				       GEN6_BSD_SLEEP_PSMI_CONTROL,
+				       GEN6_BSD_SLEEP_INDICATOR,
+				       0,
+				       50))
 		DRM_ERROR("timed out waiting for the BSD ring to wake up\n");
 
 	/* Now that the ring is fully powered up, update the tail */
-	I915_WRITE_TAIL(engine, value);
-	POSTING_READ(RING_TAIL(engine->mmio_base));
+	I915_WRITE_FW(RING_TAIL(engine->mmio_base), value);
+	POSTING_READ_FW(RING_TAIL(engine->mmio_base));
 
 	/* Let the ring send IDLE messages to the GT again,
 	 * and so let it sleep to conserve power when idle.
 	 */
-	I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
-		   _MASKED_BIT_DISABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
+	I915_WRITE_FW(GEN6_BSD_SLEEP_PSMI_CONTROL,
+		      _MASKED_BIT_DISABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
+
+	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
 }
 
 static int gen6_bsd_ring_flush(struct drm_i915_gem_request *req,
@@ -2808,11 +2686,159 @@ static int gen6_ring_flush(struct drm_i915_gem_request *req,
 	return 0;
 }
 
+static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv,
+				       struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	int ret, i;
+
+	if (!i915_semaphore_is_enabled(dev_priv))
+		return;
+
+	if (INTEL_GEN(dev_priv) >= 8 && !dev_priv->semaphore_obj) {
+		obj = i915_gem_object_create(&dev_priv->drm, 4096);
+		if (IS_ERR(obj)) {
+			DRM_ERROR("Failed to allocate semaphore bo. Disabling semaphores\n");
+			i915.semaphores = 0;
+		} else {
+			i915_gem_object_set_cache_level(obj, I915_CACHE_LLC);
+			ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_NONBLOCK);
+			if (ret != 0) {
+				drm_gem_object_unreference(&obj->base);
+				DRM_ERROR("Failed to pin semaphore bo. Disabling semaphores\n");
+				i915.semaphores = 0;
+			} else {
+				dev_priv->semaphore_obj = obj;
+			}
+		}
+	}
+
+	if (!i915_semaphore_is_enabled(dev_priv))
+		return;
+
+	if (INTEL_GEN(dev_priv) >= 8) {
+		u64 offset = i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj);
+
+		engine->semaphore.sync_to = gen8_ring_sync;
+		engine->semaphore.signal = gen8_xcs_signal;
+
+		for (i = 0; i < I915_NUM_ENGINES; i++) {
+			u64 ring_offset;
+
+			if (i != engine->id)
+				ring_offset = offset + GEN8_SEMAPHORE_OFFSET(engine->id, i);
+			else
+				ring_offset = MI_SEMAPHORE_SYNC_INVALID;
+
+			engine->semaphore.signal_ggtt[i] = ring_offset;
+		}
+	} else if (INTEL_GEN(dev_priv) >= 6) {
+		engine->semaphore.sync_to = gen6_ring_sync;
+		engine->semaphore.signal = gen6_signal;
+
+		/*
+		 * The current semaphore is only applied on pre-gen8
+		 * platform.  And there is no VCS2 ring on the pre-gen8
+		 * platform. So the semaphore between RCS and VCS2 is
+		 * initialized as INVALID.  Gen8 will initialize the
+		 * sema between VCS2 and RCS later.
+		 */
+		for (i = 0; i < I915_NUM_ENGINES; i++) {
+			static const struct {
+				u32 wait_mbox;
+				i915_reg_t mbox_reg;
+			} sem_data[I915_NUM_ENGINES][I915_NUM_ENGINES] = {
+				[RCS] = {
+					[VCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_RV,  .mbox_reg = GEN6_VRSYNC },
+					[BCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_RB,  .mbox_reg = GEN6_BRSYNC },
+					[VECS] = { .wait_mbox = MI_SEMAPHORE_SYNC_RVE, .mbox_reg = GEN6_VERSYNC },
+				},
+				[VCS] = {
+					[RCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VR,  .mbox_reg = GEN6_RVSYNC },
+					[BCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VB,  .mbox_reg = GEN6_BVSYNC },
+					[VECS] = { .wait_mbox = MI_SEMAPHORE_SYNC_VVE, .mbox_reg = GEN6_VEVSYNC },
+				},
+				[BCS] = {
+					[RCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_BR,  .mbox_reg = GEN6_RBSYNC },
+					[VCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_BV,  .mbox_reg = GEN6_VBSYNC },
+					[VECS] = { .wait_mbox = MI_SEMAPHORE_SYNC_BVE, .mbox_reg = GEN6_VEBSYNC },
+				},
+				[VECS] = {
+					[RCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VER, .mbox_reg = GEN6_RVESYNC },
+					[VCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VEV, .mbox_reg = GEN6_VVESYNC },
+					[BCS] =  { .wait_mbox = MI_SEMAPHORE_SYNC_VEB, .mbox_reg = GEN6_BVESYNC },
+				},
+			};
+			u32 wait_mbox;
+			i915_reg_t mbox_reg;
+
+			if (i == engine->id || i == VCS2) {
+				wait_mbox = MI_SEMAPHORE_SYNC_INVALID;
+				mbox_reg = GEN6_NOSYNC;
+			} else {
+				wait_mbox = sem_data[engine->id][i].wait_mbox;
+				mbox_reg = sem_data[engine->id][i].mbox_reg;
+			}
+
+			engine->semaphore.mbox.wait[i] = wait_mbox;
+			engine->semaphore.mbox.signal[i] = mbox_reg;
+		}
+	}
+}
+
+static void intel_ring_init_irq(struct drm_i915_private *dev_priv,
+				struct intel_engine_cs *engine)
+{
+	if (INTEL_GEN(dev_priv) >= 8) {
+		engine->irq_enable = gen8_irq_enable;
+		engine->irq_disable = gen8_irq_disable;
+		engine->irq_seqno_barrier = gen6_seqno_barrier;
+	} else if (INTEL_GEN(dev_priv) >= 6) {
+		engine->irq_enable = gen6_irq_enable;
+		engine->irq_disable = gen6_irq_disable;
+		engine->irq_seqno_barrier = gen6_seqno_barrier;
+	} else if (INTEL_GEN(dev_priv) >= 5) {
+		engine->irq_enable = gen5_irq_enable;
+		engine->irq_disable = gen5_irq_disable;
+		engine->irq_seqno_barrier = gen5_seqno_barrier;
+	} else if (INTEL_GEN(dev_priv) >= 3) {
+		engine->irq_enable = i9xx_irq_enable;
+		engine->irq_disable = i9xx_irq_disable;
+	} else {
+		engine->irq_enable = i8xx_irq_enable;
+		engine->irq_disable = i8xx_irq_disable;
+	}
+}
+
+static void intel_ring_default_vfuncs(struct drm_i915_private *dev_priv,
+				      struct intel_engine_cs *engine)
+{
+	engine->init_hw = init_ring_common;
+	engine->write_tail = ring_write_tail;
+
+	engine->add_request = i9xx_add_request;
+	if (INTEL_GEN(dev_priv) >= 6)
+		engine->add_request = gen6_add_request;
+
+	if (INTEL_GEN(dev_priv) >= 8)
+		engine->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
+	else if (INTEL_GEN(dev_priv) >= 6)
+		engine->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
+	else if (INTEL_GEN(dev_priv) >= 4)
+		engine->dispatch_execbuffer = i965_dispatch_execbuffer;
+	else if (IS_I830(dev_priv) || IS_845G(dev_priv))
+		engine->dispatch_execbuffer = i830_dispatch_execbuffer;
+	else
+		engine->dispatch_execbuffer = i915_dispatch_execbuffer;
+
+	intel_ring_init_irq(dev_priv, engine);
+	intel_ring_init_semaphores(dev_priv, engine);
+}
+
 int intel_init_render_ring_buffer(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[RCS];
-	struct drm_i915_gem_object *obj;
 	int ret;
 
 	engine->name = "render ring";
@@ -2821,139 +2847,49 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 	engine->hw_id = 0;
 	engine->mmio_base = RENDER_RING_BASE;
 
-	if (INTEL_GEN(dev_priv) >= 8) {
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			obj = i915_gem_object_create(dev, 4096);
-			if (IS_ERR(obj)) {
-				DRM_ERROR("Failed to allocate semaphore bo. Disabling semaphores\n");
-				i915.semaphores = 0;
-			} else {
-				i915_gem_object_set_cache_level(obj, I915_CACHE_LLC);
-				ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_NONBLOCK);
-				if (ret != 0) {
-					drm_gem_object_unreference(&obj->base);
-					DRM_ERROR("Failed to pin semaphore bo. Disabling semaphores\n");
-					i915.semaphores = 0;
-				} else
-					dev_priv->semaphore_obj = obj;
-			}
-		}
+	intel_ring_default_vfuncs(dev_priv, engine);
+
+	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
+	if (HAS_L3_DPF(dev_priv))
+		engine->irq_keep_mask = GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 
+	if (INTEL_GEN(dev_priv) >= 8) {
 		engine->init_context = intel_rcs_ctx_init;
 		engine->add_request = gen8_render_add_request;
 		engine->flush = gen8_render_ring_flush;
-		engine->irq_get = gen8_ring_get_irq;
-		engine->irq_put = gen8_ring_put_irq;
-		engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
-		engine->get_seqno = ring_get_seqno;
-		engine->set_seqno = ring_set_seqno;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			WARN_ON(!dev_priv->semaphore_obj);
-			engine->semaphore.sync_to = gen8_ring_sync;
+		if (i915_semaphore_is_enabled(dev_priv))
 			engine->semaphore.signal = gen8_rcs_signal;
-			GEN8_RING_SEMAPHORE_INIT(engine);
-		}
 	} else if (INTEL_GEN(dev_priv) >= 6) {
 		engine->init_context = intel_rcs_ctx_init;
-		engine->add_request = gen6_add_request;
 		engine->flush = gen7_render_ring_flush;
 		if (IS_GEN6(dev_priv))
 			engine->flush = gen6_render_ring_flush;
-		engine->irq_get = gen6_ring_get_irq;
-		engine->irq_put = gen6_ring_put_irq;
-		engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
-		engine->irq_seqno_barrier = gen6_seqno_barrier;
-		engine->get_seqno = ring_get_seqno;
-		engine->set_seqno = ring_set_seqno;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			engine->semaphore.sync_to = gen6_ring_sync;
-			engine->semaphore.signal = gen6_signal;
-			/*
-			 * The current semaphore is only applied on pre-gen8
-			 * platform.  And there is no VCS2 ring on the pre-gen8
-			 * platform. So the semaphore between RCS and VCS2 is
-			 * initialized as INVALID.  Gen8 will initialize the
-			 * sema between VCS2 and RCS later.
-			 */
-			engine->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_RV;
-			engine->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_RB;
-			engine->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_RVE;
-			engine->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.signal[RCS] = GEN6_NOSYNC;
-			engine->semaphore.mbox.signal[VCS] = GEN6_VRSYNC;
-			engine->semaphore.mbox.signal[BCS] = GEN6_BRSYNC;
-			engine->semaphore.mbox.signal[VECS] = GEN6_VERSYNC;
-			engine->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
-		}
 	} else if (IS_GEN5(dev_priv)) {
-		engine->add_request = pc_render_add_request;
 		engine->flush = gen4_render_ring_flush;
-		engine->get_seqno = pc_render_get_seqno;
-		engine->set_seqno = pc_render_set_seqno;
-		engine->irq_get = gen5_ring_get_irq;
-		engine->irq_put = gen5_ring_put_irq;
-		engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT |
-					GT_RENDER_PIPECTL_NOTIFY_INTERRUPT;
 	} else {
-		engine->add_request = i9xx_add_request;
 		if (INTEL_GEN(dev_priv) < 4)
 			engine->flush = gen2_render_ring_flush;
 		else
 			engine->flush = gen4_render_ring_flush;
-		engine->get_seqno = ring_get_seqno;
-		engine->set_seqno = ring_set_seqno;
-		if (IS_GEN2(dev_priv)) {
-			engine->irq_get = i8xx_ring_get_irq;
-			engine->irq_put = i8xx_ring_put_irq;
-		} else {
-			engine->irq_get = i9xx_ring_get_irq;
-			engine->irq_put = i9xx_ring_put_irq;
-		}
 		engine->irq_enable_mask = I915_USER_INTERRUPT;
 	}
-	engine->write_tail = ring_write_tail;
 
 	if (IS_HASWELL(dev_priv))
 		engine->dispatch_execbuffer = hsw_ring_dispatch_execbuffer;
-	else if (IS_GEN8(dev_priv))
-		engine->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
-	else if (INTEL_GEN(dev_priv) >= 6)
-		engine->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
-	else if (INTEL_GEN(dev_priv) >= 4)
-		engine->dispatch_execbuffer = i965_dispatch_execbuffer;
-	else if (IS_I830(dev_priv) || IS_845G(dev_priv))
-		engine->dispatch_execbuffer = i830_dispatch_execbuffer;
-	else
-		engine->dispatch_execbuffer = i915_dispatch_execbuffer;
+
 	engine->init_hw = init_render_ring;
 	engine->cleanup = render_ring_cleanup;
 
-	/* Workaround batchbuffer to combat CS tlb bug. */
-	if (HAS_BROKEN_CS_TLB(dev_priv)) {
-		obj = i915_gem_object_create(dev, I830_WA_SIZE);
-		if (IS_ERR(obj)) {
-			DRM_ERROR("Failed to allocate batch bo\n");
-			return PTR_ERR(obj);
-		}
-
-		ret = i915_gem_obj_ggtt_pin(obj, 0, 0);
-		if (ret != 0) {
-			drm_gem_object_unreference(&obj->base);
-			DRM_ERROR("Failed to ping batch bo\n");
-			return ret;
-		}
-
-		engine->scratch.obj = obj;
-		engine->scratch.gtt_offset = i915_gem_obj_ggtt_offset(obj);
-	}
-
 	ret = intel_init_ring_buffer(dev, engine);
 	if (ret)
 		return ret;
 
-	if (INTEL_GEN(dev_priv) >= 5) {
-		ret = intel_init_pipe_control(engine);
+	if (INTEL_GEN(dev_priv) >= 6) {
+		ret = intel_init_pipe_control(engine, 4096);
+		if (ret)
+			return ret;
+	} else if (HAS_BROKEN_CS_TLB(dev_priv)) {
+		ret = intel_init_pipe_control(engine, I830_WA_SIZE);
 		if (ret)
 			return ret;
 	}
@@ -2963,7 +2899,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 
 int intel_init_bsd_ring_buffer(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[VCS];
 
 	engine->name = "bsd ring";
@@ -2971,68 +2907,27 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 	engine->exec_id = I915_EXEC_BSD;
 	engine->hw_id = 1;
 
-	engine->write_tail = ring_write_tail;
+	intel_ring_default_vfuncs(dev_priv, engine);
+
 	if (INTEL_GEN(dev_priv) >= 6) {
 		engine->mmio_base = GEN6_BSD_RING_BASE;
 		/* gen6 bsd needs a special wa for tail updates */
 		if (IS_GEN6(dev_priv))
 			engine->write_tail = gen6_bsd_ring_write_tail;
 		engine->flush = gen6_bsd_ring_flush;
-		engine->add_request = gen6_add_request;
-		engine->irq_seqno_barrier = gen6_seqno_barrier;
-		engine->get_seqno = ring_get_seqno;
-		engine->set_seqno = ring_set_seqno;
-		if (INTEL_GEN(dev_priv) >= 8) {
+		if (INTEL_GEN(dev_priv) >= 8)
 			engine->irq_enable_mask =
 				GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT;
-			engine->irq_get = gen8_ring_get_irq;
-			engine->irq_put = gen8_ring_put_irq;
-			engine->dispatch_execbuffer =
-				gen8_ring_dispatch_execbuffer;
-			if (i915_semaphore_is_enabled(dev_priv)) {
-				engine->semaphore.sync_to = gen8_ring_sync;
-				engine->semaphore.signal = gen8_xcs_signal;
-				GEN8_RING_SEMAPHORE_INIT(engine);
-			}
-		} else {
+		else
 			engine->irq_enable_mask = GT_BSD_USER_INTERRUPT;
-			engine->irq_get = gen6_ring_get_irq;
-			engine->irq_put = gen6_ring_put_irq;
-			engine->dispatch_execbuffer =
-				gen6_ring_dispatch_execbuffer;
-			if (i915_semaphore_is_enabled(dev_priv)) {
-				engine->semaphore.sync_to = gen6_ring_sync;
-				engine->semaphore.signal = gen6_signal;
-				engine->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VR;
-				engine->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID;
-				engine->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VB;
-				engine->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_VVE;
-				engine->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-				engine->semaphore.mbox.signal[RCS] = GEN6_RVSYNC;
-				engine->semaphore.mbox.signal[VCS] = GEN6_NOSYNC;
-				engine->semaphore.mbox.signal[BCS] = GEN6_BVSYNC;
-				engine->semaphore.mbox.signal[VECS] = GEN6_VEVSYNC;
-				engine->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
-			}
-		}
 	} else {
 		engine->mmio_base = BSD_RING_BASE;
 		engine->flush = bsd_ring_flush;
-		engine->add_request = i9xx_add_request;
-		engine->get_seqno = ring_get_seqno;
-		engine->set_seqno = ring_set_seqno;
-		if (IS_GEN5(dev_priv)) {
+		if (IS_GEN5(dev_priv))
 			engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
-			engine->irq_get = gen5_ring_get_irq;
-			engine->irq_put = gen5_ring_put_irq;
-		} else {
+		else
 			engine->irq_enable_mask = I915_BSD_USER_INTERRUPT;
-			engine->irq_get = i9xx_ring_get_irq;
-			engine->irq_put = i9xx_ring_put_irq;
-		}
-		engine->dispatch_execbuffer = i965_dispatch_execbuffer;
 	}
-	engine->init_hw = init_ring_common;
 
 	return intel_init_ring_buffer(dev, engine);
 }
@@ -3042,147 +2937,70 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
  */
 int intel_init_bsd2_ring_buffer(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[VCS2];
 
 	engine->name = "bsd2 ring";
 	engine->id = VCS2;
 	engine->exec_id = I915_EXEC_BSD;
 	engine->hw_id = 4;
-
-	engine->write_tail = ring_write_tail;
 	engine->mmio_base = GEN8_BSD2_RING_BASE;
+
+	intel_ring_default_vfuncs(dev_priv, engine);
+
 	engine->flush = gen6_bsd_ring_flush;
-	engine->add_request = gen6_add_request;
-	engine->irq_seqno_barrier = gen6_seqno_barrier;
-	engine->get_seqno = ring_get_seqno;
-	engine->set_seqno = ring_set_seqno;
 	engine->irq_enable_mask =
 			GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT;
-	engine->irq_get = gen8_ring_get_irq;
-	engine->irq_put = gen8_ring_put_irq;
-	engine->dispatch_execbuffer =
-			gen8_ring_dispatch_execbuffer;
-	if (i915_semaphore_is_enabled(dev_priv)) {
-		engine->semaphore.sync_to = gen8_ring_sync;
-		engine->semaphore.signal = gen8_xcs_signal;
-		GEN8_RING_SEMAPHORE_INIT(engine);
-	}
-	engine->init_hw = init_ring_common;
 
 	return intel_init_ring_buffer(dev, engine);
 }
 
 int intel_init_blt_ring_buffer(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[BCS];
 
 	engine->name = "blitter ring";
 	engine->id = BCS;
 	engine->exec_id = I915_EXEC_BLT;
 	engine->hw_id = 2;
-
 	engine->mmio_base = BLT_RING_BASE;
-	engine->write_tail = ring_write_tail;
+
+	intel_ring_default_vfuncs(dev_priv, engine);
+
 	engine->flush = gen6_ring_flush;
-	engine->add_request = gen6_add_request;
-	engine->irq_seqno_barrier = gen6_seqno_barrier;
-	engine->get_seqno = ring_get_seqno;
-	engine->set_seqno = ring_set_seqno;
-	if (INTEL_GEN(dev_priv) >= 8) {
+	if (INTEL_GEN(dev_priv) >= 8)
 		engine->irq_enable_mask =
 			GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT;
-		engine->irq_get = gen8_ring_get_irq;
-		engine->irq_put = gen8_ring_put_irq;
-		engine->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			engine->semaphore.sync_to = gen8_ring_sync;
-			engine->semaphore.signal = gen8_xcs_signal;
-			GEN8_RING_SEMAPHORE_INIT(engine);
-		}
-	} else {
+	else
 		engine->irq_enable_mask = GT_BLT_USER_INTERRUPT;
-		engine->irq_get = gen6_ring_get_irq;
-		engine->irq_put = gen6_ring_put_irq;
-		engine->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			engine->semaphore.signal = gen6_signal;
-			engine->semaphore.sync_to = gen6_ring_sync;
-			/*
-			 * The current semaphore is only applied on pre-gen8
-			 * platform.  And there is no VCS2 ring on the pre-gen8
-			 * platform. So the semaphore between BCS and VCS2 is
-			 * initialized as INVALID.  Gen8 will initialize the
-			 * sema between BCS and VCS2 later.
-			 */
-			engine->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_BR;
-			engine->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_BV;
-			engine->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_BVE;
-			engine->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.signal[RCS] = GEN6_RBSYNC;
-			engine->semaphore.mbox.signal[VCS] = GEN6_VBSYNC;
-			engine->semaphore.mbox.signal[BCS] = GEN6_NOSYNC;
-			engine->semaphore.mbox.signal[VECS] = GEN6_VEBSYNC;
-			engine->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
-		}
-	}
-	engine->init_hw = init_ring_common;
 
 	return intel_init_ring_buffer(dev, engine);
 }
 
 int intel_init_vebox_ring_buffer(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_engine_cs *engine = &dev_priv->engine[VECS];
 
 	engine->name = "video enhancement ring";
 	engine->id = VECS;
 	engine->exec_id = I915_EXEC_VEBOX;
 	engine->hw_id = 3;
-
 	engine->mmio_base = VEBOX_RING_BASE;
-	engine->write_tail = ring_write_tail;
+
+	intel_ring_default_vfuncs(dev_priv, engine);
+
 	engine->flush = gen6_ring_flush;
-	engine->add_request = gen6_add_request;
-	engine->irq_seqno_barrier = gen6_seqno_barrier;
-	engine->get_seqno = ring_get_seqno;
-	engine->set_seqno = ring_set_seqno;
 
 	if (INTEL_GEN(dev_priv) >= 8) {
 		engine->irq_enable_mask =
 			GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT;
-		engine->irq_get = gen8_ring_get_irq;
-		engine->irq_put = gen8_ring_put_irq;
-		engine->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			engine->semaphore.sync_to = gen8_ring_sync;
-			engine->semaphore.signal = gen8_xcs_signal;
-			GEN8_RING_SEMAPHORE_INIT(engine);
-		}
 	} else {
 		engine->irq_enable_mask = PM_VEBOX_USER_INTERRUPT;
-		engine->irq_get = hsw_vebox_get_irq;
-		engine->irq_put = hsw_vebox_put_irq;
-		engine->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
-		if (i915_semaphore_is_enabled(dev_priv)) {
-			engine->semaphore.sync_to = gen6_ring_sync;
-			engine->semaphore.signal = gen6_signal;
-			engine->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VER;
-			engine->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_VEV;
-			engine->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VEB;
-			engine->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-			engine->semaphore.mbox.signal[RCS] = GEN6_RVESYNC;
-			engine->semaphore.mbox.signal[VCS] = GEN6_VVESYNC;
-			engine->semaphore.mbox.signal[BCS] = GEN6_BVESYNC;
-			engine->semaphore.mbox.signal[VECS] = GEN6_NOSYNC;
-			engine->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
-		}
+		engine->irq_enable = hsw_vebox_irq_enable;
+		engine->irq_disable = hsw_vebox_irq_disable;
 	}
-	engine->init_hw = init_ring_common;
 
 	return intel_init_ring_buffer(dev, engine);
 }
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index b33c876fed20..12cb7ed90014 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -62,18 +62,6 @@ struct  intel_hw_status_page {
 	(i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj) + \
 	 GEN8_SEMAPHORE_OFFSET(from, (__ring)->id))
 
-#define GEN8_RING_SEMAPHORE_INIT(e) do { \
-	if (!dev_priv->semaphore_obj) { \
-		break; \
-	} \
-	(e)->semaphore.signal_ggtt[RCS] = GEN8_SIGNAL_OFFSET((e), RCS); \
-	(e)->semaphore.signal_ggtt[VCS] = GEN8_SIGNAL_OFFSET((e), VCS); \
-	(e)->semaphore.signal_ggtt[BCS] = GEN8_SIGNAL_OFFSET((e), BCS); \
-	(e)->semaphore.signal_ggtt[VECS] = GEN8_SIGNAL_OFFSET((e), VECS); \
-	(e)->semaphore.signal_ggtt[VCS2] = GEN8_SIGNAL_OFFSET((e), VCS2); \
-	(e)->semaphore.signal_ggtt[(e)->id] = MI_SEMAPHORE_SYNC_INVALID; \
-	} while(0)
-
 enum intel_ring_hangcheck_action {
 	HANGCHECK_IDLE = 0,
 	HANGCHECK_WAIT,
@@ -86,8 +74,8 @@ enum intel_ring_hangcheck_action {
 
 struct intel_ring_hangcheck {
 	u64 acthd;
+	unsigned long user_interrupts;
 	u32 seqno;
-	unsigned user_interrupts;
 	int score;
 	enum intel_ring_hangcheck_action action;
 	int deadlock;
@@ -141,6 +129,8 @@ struct  i915_ctx_workarounds {
 	struct drm_i915_gem_object *obj;
 };
 
+struct drm_i915_gem_request;
+
 struct intel_engine_cs {
 	struct drm_i915_private *i915;
 	const char	*name;
@@ -160,6 +150,39 @@ struct intel_engine_cs {
 	struct intel_ringbuffer *buffer;
 	struct list_head buffers;
 
+	/* Rather than have every client wait upon all user interrupts,
+	 * with the herd waking after every interrupt and each doing the
+	 * heavyweight seqno dance, we delegate the task (of being the
+	 * bottom-half of the user interrupt) to the first client. After
+	 * every interrupt, we wake up one client, who does the heavyweight
+	 * coherent seqno read and either goes back to sleep (if incomplete),
+	 * or wakes up all the completed clients in parallel, before then
+	 * transferring the bottom-half status to the next client in the queue.
+	 *
+	 * Compared to walking the entire list of waiters in a single dedicated
+	 * bottom-half, we reduce the latency of the first waiter by avoiding
+	 * a context switch, but incur additional coherent seqno reads when
+	 * following the chain of request breadcrumbs. Since it is most likely
+	 * that we have a single client waiting on each seqno, then reducing
+	 * the overhead of waking that client is much preferred.
+	 */
+	struct intel_breadcrumbs {
+		struct task_struct *irq_seqno_bh; /* bh for user interrupts */
+		unsigned long irq_wakeups;
+		bool irq_posted;
+
+		spinlock_t lock; /* protects the lists of requests */
+		struct rb_root waiters; /* sorted by retirement, priority */
+		struct rb_root signals; /* sorted by retirement */
+		struct intel_wait *first_wait; /* oldest waiter by retirement */
+		struct task_struct *signaler; /* used for fence signalling */
+		struct drm_i915_gem_request *first_signal;
+		struct timer_list fake_irq; /* used after a missed interrupt */
+
+		bool irq_enabled : 1;
+		bool rpm_wakelock : 1;
+	} breadcrumbs;
+
 	/*
 	 * A pool of objects to use as shadow copies of client batch buffers
 	 * when the command parser is enabled. Prevents the client from
@@ -170,11 +193,10 @@ struct intel_engine_cs {
 	struct intel_hw_status_page status_page;
 	struct i915_ctx_workarounds wa_ctx;
 
-	unsigned irq_refcount; /* protected by dev_priv->irq_lock */
-	u32		irq_enable_mask;	/* bitmask to enable ring interrupt */
-	struct drm_i915_gem_request *trace_irq_req;
-	bool __must_check (*irq_get)(struct intel_engine_cs *ring);
-	void		(*irq_put)(struct intel_engine_cs *ring);
+	u32             irq_keep_mask; /* always keep these interrupts */
+	u32		irq_enable_mask; /* bitmask to enable ring interrupt */
+	void		(*irq_enable)(struct intel_engine_cs *ring);
+	void		(*irq_disable)(struct intel_engine_cs *ring);
 
 	int		(*init_hw)(struct intel_engine_cs *ring);
 
@@ -193,9 +215,6 @@ struct intel_engine_cs {
 	 * monotonic, even if not coherent.
 	 */
 	void		(*irq_seqno_barrier)(struct intel_engine_cs *ring);
-	u32		(*get_seqno)(struct intel_engine_cs *ring);
-	void		(*set_seqno)(struct intel_engine_cs *ring,
-				     u32 seqno);
 	int		(*dispatch_execbuffer)(struct drm_i915_gem_request *req,
 					       u64 offset, u32 length,
 					       unsigned dispatch_flags);
@@ -272,7 +291,6 @@ struct intel_engine_cs {
 	unsigned int idle_lite_restore_wa;
 	bool disable_lite_restore_wa;
 	u32 ctx_desc_template;
-	u32             irq_keep_mask; /* bitmask for interrupts that should not be masked */
 	int		(*emit_request)(struct drm_i915_gem_request *request);
 	int		(*emit_flush)(struct drm_i915_gem_request *request,
 				      u32 invalidate_domains,
@@ -304,12 +322,9 @@ struct intel_engine_cs {
 	 * inspecting request list.
 	 */
 	u32 last_submitted_seqno;
-	unsigned user_interrupts;
 
 	bool gpu_caches_dirty;
 
-	wait_queue_head_t irq_queue;
-
 	struct i915_gem_context *last_context;
 
 	struct intel_ring_hangcheck hangcheck;
@@ -317,7 +332,6 @@ struct intel_engine_cs {
 	struct {
 		struct drm_i915_gem_object *obj;
 		u32 gtt_offset;
-		volatile u32 *cpu_page;
 	} scratch;
 
 	bool needs_cmd_parser;
@@ -348,13 +362,13 @@ struct intel_engine_cs {
 };
 
 static inline bool
-intel_engine_initialized(struct intel_engine_cs *engine)
+intel_engine_initialized(const struct intel_engine_cs *engine)
 {
 	return engine->i915 != NULL;
 }
 
 static inline unsigned
-intel_engine_flag(struct intel_engine_cs *engine)
+intel_engine_flag(const struct intel_engine_cs *engine)
 {
 	return 1 << engine->id;
 }
@@ -456,15 +470,14 @@ static inline void intel_ring_advance(struct intel_engine_cs *engine)
 }
 int __intel_ring_space(int head, int tail, int size);
 void intel_ring_update_space(struct intel_ringbuffer *ringbuf);
-bool intel_engine_stopped(struct intel_engine_cs *engine);
 
 int __must_check intel_engine_idle(struct intel_engine_cs *engine);
 void intel_ring_init_seqno(struct intel_engine_cs *engine, u32 seqno);
 int intel_ring_flush_all_caches(struct drm_i915_gem_request *req);
 int intel_ring_invalidate_all_caches(struct drm_i915_gem_request *req);
 
+int intel_init_pipe_control(struct intel_engine_cs *engine, int size);
 void intel_fini_pipe_control(struct intel_engine_cs *engine);
-int intel_init_pipe_control(struct intel_engine_cs *engine);
 
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
@@ -473,6 +486,10 @@ int intel_init_blt_ring_buffer(struct drm_device *dev);
 int intel_init_vebox_ring_buffer(struct drm_device *dev);
 
 u64 intel_ring_get_active_head(struct intel_engine_cs *engine);
+static inline u32 intel_engine_get_seqno(struct intel_engine_cs *engine)
+{
+	return intel_read_status_page(engine, I915_GEM_HWS_INDEX);
+}
 
 int init_workarounds_ring(struct intel_engine_cs *engine);
 
@@ -495,4 +512,62 @@ static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine)
 	return engine->status_page.gfx_addr + I915_GEM_HWS_INDEX_ADDR;
 }
 
+/* intel_breadcrumbs.c -- user interrupt bottom-half for waiters */
+struct intel_wait {
+	struct rb_node node;
+	struct task_struct *tsk;
+	u32 seqno;
+};
+
+struct intel_signal_node {
+	struct rb_node node;
+	struct intel_wait wait;
+};
+
+int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine);
+
+static inline void intel_wait_init(struct intel_wait *wait, u32 seqno)
+{
+	wait->tsk = current;
+	wait->seqno = seqno;
+}
+
+static inline bool intel_wait_complete(const struct intel_wait *wait)
+{
+	return RB_EMPTY_NODE(&wait->node);
+}
+
+bool intel_engine_add_wait(struct intel_engine_cs *engine,
+			   struct intel_wait *wait);
+void intel_engine_remove_wait(struct intel_engine_cs *engine,
+			      struct intel_wait *wait);
+void intel_engine_enable_signaling(struct drm_i915_gem_request *request);
+
+static inline bool intel_engine_has_waiter(struct intel_engine_cs *engine)
+{
+	return READ_ONCE(engine->breadcrumbs.irq_seqno_bh);
+}
+
+static inline bool intel_engine_wakeup(struct intel_engine_cs *engine)
+{
+	bool wakeup = false;
+	struct task_struct *tsk = READ_ONCE(engine->breadcrumbs.irq_seqno_bh);
+	/* Note that for this not to dangerously chase a dangling pointer,
+	 * the caller is responsible for ensure that the task remain valid for
+	 * wake_up_process() i.e. that the RCU grace period cannot expire.
+	 *
+	 * Also note that tsk is likely to be in !TASK_RUNNING state so an
+	 * early test for tsk->state != TASK_RUNNING before wake_up_process()
+	 * is unlikely to be beneficial.
+	 */
+	if (tsk)
+		wakeup = wake_up_process(tsk);
+	return wakeup;
+}
+
+void intel_engine_enable_fake_irq(struct intel_engine_cs *engine);
+void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);
+unsigned int intel_kick_waiters(struct drm_i915_private *i915);
+unsigned int intel_kick_signalers(struct drm_i915_private *i915);
+
 #endif /* _INTEL_RINGBUFFER_H_ */
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index e856d49d6dc3..6b78295f53db 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -287,7 +287,7 @@ void intel_display_set_init_power(struct drm_i915_private *dev_priv,
  */
 static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	/*
 	 * After we re-enable the power well, if we touch VGA register 0x3d5
@@ -318,7 +318,7 @@ static void hsw_power_well_pre_disable(struct drm_i915_private *dev_priv)
 static void skl_power_well_post_enable(struct drm_i915_private *dev_priv,
 				       struct i915_power_well *power_well)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 
 	/*
 	 * After we re-enable the power well, if we touch VGA register 0x3d5
@@ -365,8 +365,11 @@ static void hsw_set_power_well(struct drm_i915_private *dev_priv,
 
 		if (!is_enabled) {
 			DRM_DEBUG_KMS("Enabling power well\n");
-			if (wait_for((I915_READ(HSW_PWR_WELL_DRIVER) &
-				      HSW_PWR_WELL_STATE_ENABLED), 20))
+			if (intel_wait_for_register(dev_priv,
+						    HSW_PWR_WELL_DRIVER,
+						    HSW_PWR_WELL_STATE_ENABLED,
+						    HSW_PWR_WELL_STATE_ENABLED,
+						    20))
 				DRM_ERROR("Timeout enabling power well\n");
 			hsw_power_well_post_enable(dev_priv);
 		}
@@ -578,6 +581,7 @@ void bxt_enable_dc9(struct drm_i915_private *dev_priv)
 
 	DRM_DEBUG_KMS("Enabling DC9\n");
 
+	intel_power_sequencer_reset(dev_priv);
 	gen9_set_dc_state(dev_priv, DC_STATE_EN_DC9);
 }
 
@@ -699,8 +703,11 @@ static void skl_set_power_well(struct drm_i915_private *dev_priv,
 
 	switch (power_well->data) {
 	case SKL_DISP_PW_1:
-		if (wait_for((I915_READ(SKL_FUSE_STATUS) &
-			SKL_FUSE_PG0_DIST_STATUS), 1)) {
+		if (intel_wait_for_register(dev_priv,
+					    SKL_FUSE_STATUS,
+					    SKL_FUSE_PG0_DIST_STATUS,
+					    SKL_FUSE_PG0_DIST_STATUS,
+					    1)) {
 			DRM_ERROR("PG0 not enabled\n");
 			return;
 		}
@@ -761,12 +768,18 @@ static void skl_set_power_well(struct drm_i915_private *dev_priv,
 
 	if (check_fuse_status) {
 		if (power_well->data == SKL_DISP_PW_1) {
-			if (wait_for((I915_READ(SKL_FUSE_STATUS) &
-				SKL_FUSE_PG1_DIST_STATUS), 1))
+			if (intel_wait_for_register(dev_priv,
+						    SKL_FUSE_STATUS,
+						    SKL_FUSE_PG1_DIST_STATUS,
+						    SKL_FUSE_PG1_DIST_STATUS,
+						    1))
 				DRM_ERROR("PG1 distributing status timeout\n");
 		} else if (power_well->data == SKL_DISP_PW_2) {
-			if (wait_for((I915_READ(SKL_FUSE_STATUS) &
-				SKL_FUSE_PG2_DIST_STATUS), 1))
+			if (intel_wait_for_register(dev_priv,
+						    SKL_FUSE_STATUS,
+						    SKL_FUSE_PG2_DIST_STATUS,
+						    SKL_FUSE_PG2_DIST_STATUS,
+						    1))
 				DRM_ERROR("PG2 distributing status timeout\n");
 		}
 	}
@@ -917,7 +930,7 @@ static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv,
 	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
 
 	WARN_ON(dev_priv->cdclk_freq !=
-		dev_priv->display.get_display_clock_speed(dev_priv->dev));
+		dev_priv->display.get_display_clock_speed(&dev_priv->drm));
 
 	gen9_assert_dbuf_enabled(dev_priv);
 
@@ -1075,7 +1088,7 @@ static void vlv_display_power_well_init(struct drm_i915_private *dev_priv)
 	 *
 	 * CHV DPLL B/C have some issues if VGA mode is enabled.
 	 */
-	for_each_pipe(dev_priv->dev, pipe) {
+	for_each_pipe(&dev_priv->drm, pipe) {
 		u32 val = I915_READ(DPLL(pipe));
 
 		val |= DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
@@ -1100,7 +1113,7 @@ static void vlv_display_power_well_init(struct drm_i915_private *dev_priv)
 
 	intel_hpd_init(dev_priv);
 
-	i915_redisable_vga_power_on(dev_priv->dev);
+	i915_redisable_vga_power_on(&dev_priv->drm);
 }
 
 static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv)
@@ -1110,9 +1123,9 @@ static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv)
 	spin_unlock_irq(&dev_priv->irq_lock);
 
 	/* make sure we're done processing display irqs */
-	synchronize_irq(dev_priv->dev->irq);
+	synchronize_irq(dev_priv->drm.irq);
 
-	vlv_power_sequencer_reset(dev_priv);
+	intel_power_sequencer_reset(dev_priv);
 }
 
 static void vlv_display_power_well_enable(struct drm_i915_private *dev_priv,
@@ -1205,7 +1218,6 @@ static void assert_chv_phy_status(struct drm_i915_private *dev_priv)
 	u32 phy_control = dev_priv->chv_phy_control;
 	u32 phy_status = 0;
 	u32 phy_status_mask = 0xffffffff;
-	u32 tmp;
 
 	/*
 	 * The BIOS can leave the PHY is some weird state
@@ -1293,10 +1305,14 @@ static void assert_chv_phy_status(struct drm_i915_private *dev_priv)
 	 * The PHY may be busy with some initial calibration and whatnot,
 	 * so the power state can take a while to actually change.
 	 */
-	if (wait_for((tmp = I915_READ(DISPLAY_PHY_STATUS) & phy_status_mask) == phy_status, 10))
-		WARN(phy_status != tmp,
-		     "Unexpected PHY_STATUS 0x%08x, expected 0x%08x (PHY_CONTROL=0x%08x)\n",
-		     tmp, phy_status, dev_priv->chv_phy_control);
+	if (intel_wait_for_register(dev_priv,
+				    DISPLAY_PHY_STATUS,
+				    phy_status_mask,
+				    phy_status,
+				    10))
+		DRM_ERROR("Unexpected PHY_STATUS 0x%08x, expected 0x%08x (PHY_CONTROL=0x%08x)\n",
+			  I915_READ(DISPLAY_PHY_STATUS) & phy_status_mask,
+			   phy_status, dev_priv->chv_phy_control);
 }
 
 #undef BITS_SET
@@ -1324,7 +1340,11 @@ static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
 	vlv_set_power_well(dev_priv, power_well, true);
 
 	/* Poll for phypwrgood signal */
-	if (wait_for(I915_READ(DISPLAY_PHY_STATUS) & PHY_POWERGOOD(phy), 1))
+	if (intel_wait_for_register(dev_priv,
+				    DISPLAY_PHY_STATUS,
+				    PHY_POWERGOOD(phy),
+				    PHY_POWERGOOD(phy),
+				    1))
 		DRM_ERROR("Display PHY %d is not power up\n", phy);
 
 	mutex_lock(&dev_priv->sb_lock);
@@ -2255,7 +2275,7 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
  */
 void intel_power_domains_fini(struct drm_i915_private *dev_priv)
 {
-	struct device *device = &dev_priv->dev->pdev->dev;
+	struct device *device = &dev_priv->drm.pdev->dev;
 
 	/*
 	 * The i915.ko module is still not prepared to be loaded when
@@ -2556,7 +2576,7 @@ static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
  */
 void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct i915_power_domains *power_domains = &dev_priv->power_domains;
 
 	power_domains->initializing = true;
@@ -2618,7 +2638,7 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv)
  */
 void intel_runtime_pm_get(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct device *device = &dev->pdev->dev;
 
 	pm_runtime_get_sync(device);
@@ -2639,7 +2659,7 @@ void intel_runtime_pm_get(struct drm_i915_private *dev_priv)
  */
 bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct device *device = &dev->pdev->dev;
 
 	if (IS_ENABLED(CONFIG_PM)) {
@@ -2681,7 +2701,7 @@ bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv)
  */
 void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct device *device = &dev->pdev->dev;
 
 	assert_rpm_wakelock_held(dev_priv);
@@ -2700,7 +2720,7 @@ void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv)
  */
 void intel_runtime_pm_put(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct device *device = &dev->pdev->dev;
 
 	assert_rpm_wakelock_held(dev_priv);
@@ -2723,7 +2743,7 @@ void intel_runtime_pm_put(struct drm_i915_private *dev_priv)
  */
 void intel_runtime_pm_enable(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_device *dev = &dev_priv->drm;
 	struct device *device = &dev->pdev->dev;
 
 	pm_runtime_set_autosuspend_delay(device, 10000); /* 10s */
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 02b4a6695528..e378f35365a2 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -240,7 +240,7 @@ intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo,
 static void intel_sdvo_write_sdvox(struct intel_sdvo *intel_sdvo, u32 val)
 {
 	struct drm_device *dev = intel_sdvo->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 bval = val, cval = val;
 	int i;
 
@@ -1195,7 +1195,7 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 static void intel_sdvo_pre_enable(struct intel_encoder *intel_encoder)
 {
 	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *crtc = to_intel_crtc(intel_encoder->base.crtc);
 	const struct drm_display_mode *adjusted_mode = &crtc->config->base.adjusted_mode;
 	struct drm_display_mode *mode = &crtc->config->base.mode;
@@ -1330,7 +1330,7 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 				    enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
 	u16 active_outputs = 0;
 	u32 tmp;
@@ -1353,7 +1353,7 @@ static void intel_sdvo_get_config(struct intel_encoder *encoder,
 				  struct intel_crtc_state *pipe_config)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
 	struct intel_sdvo_dtd dtd;
 	int encoder_pixel_multiplier = 0;
@@ -1436,7 +1436,7 @@ static void intel_sdvo_get_config(struct intel_encoder *encoder,
 
 static void intel_disable_sdvo(struct intel_encoder *encoder)
 {
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
 	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	u32 temp;
@@ -1471,7 +1471,7 @@ static void intel_disable_sdvo(struct intel_encoder *encoder)
 		temp &= ~SDVO_ENABLE;
 		intel_sdvo_write_sdvox(intel_sdvo, temp);
 
-		intel_wait_for_vblank_if_active(dev_priv->dev, PIPE_A);
+		intel_wait_for_vblank_if_active(&dev_priv->drm, PIPE_A);
 		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
 	}
@@ -1489,7 +1489,7 @@ static void pch_post_disable_sdvo(struct intel_encoder *encoder)
 static void intel_enable_sdvo(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	u32 temp;
@@ -1633,7 +1633,7 @@ intel_sdvo_get_edid(struct drm_connector *connector)
 static struct edid *
 intel_sdvo_get_analog_edid(struct drm_connector *connector)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 
 	return drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
@@ -1916,7 +1916,7 @@ static void intel_sdvo_get_tv_modes(struct drm_connector *connector)
 static void intel_sdvo_get_lvds_modes(struct drm_connector *connector)
 {
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct drm_display_mode *newmode;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
@@ -2001,7 +2001,7 @@ intel_sdvo_set_property(struct drm_connector *connector,
 {
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector);
 	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	uint16_t temp_value;
 	uint8_t cmd;
 	int ret;
@@ -2177,6 +2177,21 @@ done:
 #undef CHECK_PROPERTY
 }
 
+static int
+intel_sdvo_connector_register(struct drm_connector *connector)
+{
+	struct intel_sdvo *sdvo = intel_attached_sdvo(connector);
+	int ret;
+
+	ret = intel_connector_register(connector);
+	if (ret)
+		return ret;
+
+	return sysfs_create_link(&connector->kdev->kobj,
+				 &sdvo->ddc.dev.kobj,
+				 sdvo->ddc.dev.kobj.name);
+}
+
 static void
 intel_sdvo_connector_unregister(struct drm_connector *connector)
 {
@@ -2193,6 +2208,7 @@ static const struct drm_connector_funcs intel_sdvo_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_sdvo_set_property,
 	.atomic_get_property = intel_connector_atomic_get_property,
+	.late_register = intel_sdvo_connector_register,
 	.early_unregister = intel_sdvo_connector_unregister,
 	.destroy = intel_sdvo_destroy,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
@@ -2322,7 +2338,7 @@ intel_sdvo_is_hdmi_connector(struct intel_sdvo *intel_sdvo, int device)
 static u8
 intel_sdvo_get_slave_addr(struct drm_device *dev, struct intel_sdvo *sdvo)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct sdvo_device_mapping *my_mapping, *other_mapping;
 
 	if (sdvo->port == PORT_B) {
@@ -2380,24 +2396,8 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 	connector->base.get_hw_state = intel_sdvo_connector_get_hw_state;
 
 	intel_connector_attach_encoder(&connector->base, &encoder->base);
-	ret = drm_connector_register(drm_connector);
-	if (ret < 0)
-		goto err1;
-
-	ret = sysfs_create_link(&drm_connector->kdev->kobj,
-				&encoder->ddc.dev.kobj,
-				encoder->ddc.dev.kobj.name);
-	if (ret < 0)
-		goto err2;
 
 	return 0;
-
-err2:
-	drm_connector_unregister(drm_connector);
-err1:
-	drm_connector_cleanup(drm_connector);
-
-	return ret;
 }
 
 static void
@@ -2524,7 +2524,6 @@ intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type)
 	return true;
 
 err:
-	drm_connector_unregister(connector);
 	intel_sdvo_destroy(connector);
 	return false;
 }
@@ -2603,7 +2602,6 @@ intel_sdvo_lvds_init(struct intel_sdvo *intel_sdvo, int device)
 	return true;
 
 err:
-	drm_connector_unregister(connector);
 	intel_sdvo_destroy(connector);
 	return false;
 }
@@ -2954,7 +2952,7 @@ static void assert_sdvo_port_valid(const struct drm_i915_private *dev_priv,
 bool intel_sdvo_init(struct drm_device *dev,
 		     i915_reg_t sdvo_reg, enum port port)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_encoder *intel_encoder;
 	struct intel_sdvo *intel_sdvo;
 	int i;
diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c
index c3998188cf35..1a840bf92eea 100644
--- a/drivers/gpu/drm/i915/intel_sideband.c
+++ b/drivers/gpu/drm/i915/intel_sideband.c
@@ -51,7 +51,9 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
 
 	WARN_ON(!mutex_is_locked(&dev_priv->sb_lock));
 
-	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+	if (intel_wait_for_register(dev_priv,
+				    VLV_IOSF_DOORBELL_REQ, IOSF_SB_BUSY, 0,
+				    5)) {
 		DRM_DEBUG_DRIVER("IOSF sideband idle wait (%s) timed out\n",
 				 is_read ? "read" : "write");
 		return -EAGAIN;
@@ -62,7 +64,9 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
 		I915_WRITE(VLV_IOSF_DATA, *val);
 	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
 
-	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+	if (intel_wait_for_register(dev_priv,
+				    VLV_IOSF_DOORBELL_REQ, IOSF_SB_BUSY, 0,
+				    5)) {
 		DRM_DEBUG_DRIVER("IOSF sideband finish wait (%s) timed out\n",
 				 is_read ? "read" : "write");
 		return -ETIMEDOUT;
@@ -202,8 +206,9 @@ u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
 	u32 value = 0;
 	WARN_ON(!mutex_is_locked(&dev_priv->sb_lock));
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
+	if (intel_wait_for_register(dev_priv,
+				    SBI_CTL_STAT, SBI_BUSY, 0,
+				    100)) {
 		DRM_ERROR("timeout waiting for SBI to become ready\n");
 		return 0;
 	}
@@ -216,8 +221,11 @@ u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
 		value = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
 	I915_WRITE(SBI_CTL_STAT, value | SBI_BUSY);
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
+	if (intel_wait_for_register(dev_priv,
+				    SBI_CTL_STAT,
+				    SBI_BUSY | SBI_RESPONSE_FAIL,
+				    0,
+				    100)) {
 		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
 		return 0;
 	}
@@ -232,8 +240,9 @@ void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
 
 	WARN_ON(!mutex_is_locked(&dev_priv->sb_lock));
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
+	if (intel_wait_for_register(dev_priv,
+				    SBI_CTL_STAT, SBI_BUSY, 0,
+				    100)) {
 		DRM_ERROR("timeout waiting for SBI to become ready\n");
 		return;
 	}
@@ -247,8 +256,11 @@ void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
 		tmp = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IOWR;
 	I915_WRITE(SBI_CTL_STAT, SBI_BUSY | tmp);
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
+	if (intel_wait_for_register(dev_priv,
+				    SBI_CTL_STAT,
+				    SBI_BUSY | SBI_RESPONSE_FAIL,
+				    0,
+				    100)) {
 		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
 		return;
 	}
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index fc654173c491..0de935ad01c2 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -199,7 +199,7 @@ skl_update_plane(struct drm_plane *drm_plane,
 		 const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = drm_plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(drm_plane);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -303,7 +303,7 @@ static void
 skl_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc)
 {
 	struct drm_device *dev = dplane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(dplane);
 	const int pipe = intel_plane->pipe;
 	const int plane = intel_plane->plane + 1;
@@ -317,7 +317,7 @@ skl_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc)
 static void
 chv_update_csc(struct intel_plane *intel_plane, uint32_t format)
 {
-	struct drm_i915_private *dev_priv = intel_plane->base.dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(intel_plane->base.dev);
 	int plane = intel_plane->plane;
 
 	/* Seems RGB data bypasses the CSC always */
@@ -359,7 +359,7 @@ vlv_update_plane(struct drm_plane *dplane,
 		 const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = dplane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(dplane);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -485,7 +485,7 @@ static void
 vlv_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc)
 {
 	struct drm_device *dev = dplane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(dplane);
 	int pipe = intel_plane->pipe;
 	int plane = intel_plane->plane;
@@ -502,7 +502,7 @@ ivb_update_plane(struct drm_plane *plane,
 		 const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -624,7 +624,7 @@ static void
 ivb_disable_plane(struct drm_plane *plane, struct drm_crtc *crtc)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	int pipe = intel_plane->pipe;
 
@@ -643,7 +643,7 @@ ilk_update_plane(struct drm_plane *plane,
 		 const struct intel_plane_state *plane_state)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
@@ -753,7 +753,7 @@ static void
 ilk_disable_plane(struct drm_plane *plane, struct drm_crtc *crtc)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	int pipe = intel_plane->pipe;
 
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index 4ce70a9f9df2..49136ad5473e 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -826,7 +826,7 @@ static bool
 intel_tv_get_hw_state(struct intel_encoder *encoder, enum pipe *pipe)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tmp = I915_READ(TV_CTL);
 
 	if (!(tmp & TV_ENC_ENABLE))
@@ -841,7 +841,7 @@ static void
 intel_enable_tv(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Prevents vblank waits from timing out in intel_tv_detect_type() */
 	intel_wait_for_vblank(encoder->base.dev,
@@ -854,7 +854,7 @@ static void
 intel_disable_tv(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	I915_WRITE(TV_CTL, I915_READ(TV_CTL) & ~TV_ENC_ENABLE);
 }
@@ -1013,7 +1013,7 @@ static void set_color_conversion(struct drm_i915_private *dev_priv,
 static void intel_tv_pre_enable(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_tv *intel_tv = enc_to_tv(encoder);
 	const struct tv_mode *tv_mode = intel_tv_mode_find(intel_tv);
@@ -1173,7 +1173,7 @@ intel_tv_detect_type(struct intel_tv *intel_tv,
 	struct drm_crtc *crtc = connector->state->crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	u32 tv_ctl, save_tv_ctl;
 	u32 tv_dac, save_tv_dac;
 	int type;
@@ -1501,6 +1501,7 @@ out:
 static const struct drm_connector_funcs intel_tv_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.detect = intel_tv_detect,
+	.late_register = intel_connector_register,
 	.early_unregister = intel_connector_unregister,
 	.destroy = intel_tv_destroy,
 	.set_property = intel_tv_set_property,
@@ -1522,7 +1523,7 @@ static const struct drm_encoder_funcs intel_tv_enc_funcs = {
 void
 intel_tv_init(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_connector *connector;
 	struct intel_tv *intel_tv;
 	struct intel_encoder *intel_encoder;
@@ -1641,5 +1642,4 @@ intel_tv_init(struct drm_device *dev)
 	drm_object_attach_property(&connector->base,
 				   dev->mode_config.tv_bottom_margin_property,
 				   intel_tv->margin[TV_MARGIN_BOTTOM]);
-	drm_connector_register(connector);
 }
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index c1ca458d688e..ff80a81b1a84 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -1299,9 +1299,11 @@ static void intel_uncore_fw_domains_init(struct drm_i915_private *dev_priv)
 		fw_domain_init(dev_priv, FW_DOMAIN_ID_RENDER,
 			       FORCEWAKE_MT, FORCEWAKE_MT_ACK);
 
+		spin_lock_irq(&dev_priv->uncore.lock);
 		fw_domains_get_with_thread_status(dev_priv, FORCEWAKE_ALL);
 		ecobus = __raw_i915_read32(dev_priv, ECOBUS);
 		fw_domains_put_with_fifo(dev_priv, FORCEWAKE_ALL);
+		spin_unlock_irq(&dev_priv->uncore.lock);
 
 		if (!(ecobus & FORCEWAKE_MT_ENABLE)) {
 			DRM_INFO("No MT forcewake available on Ivybridge, this can result in issues\n");
@@ -1407,7 +1409,7 @@ static const struct register_whitelist {
 int i915_reg_read_ioctl(struct drm_device *dev,
 			void *data, struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_i915_reg_read *reg = data;
 	struct register_whitelist const *entry = whitelist;
 	unsigned size;
@@ -1469,7 +1471,7 @@ static int i915_reset_complete(struct pci_dev *pdev)
 
 static int i915_do_reset(struct drm_i915_private *dev_priv, unsigned engine_mask)
 {
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 
 	/* assert reset for at least 20 usec */
 	pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
@@ -1488,14 +1490,14 @@ static int g4x_reset_complete(struct pci_dev *pdev)
 
 static int g33_do_reset(struct drm_i915_private *dev_priv, unsigned engine_mask)
 {
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
 	return wait_for(g4x_reset_complete(pdev), 500);
 }
 
 static int g4x_do_reset(struct drm_i915_private *dev_priv, unsigned engine_mask)
 {
-	struct pci_dev *pdev = dev_priv->dev->pdev;
+	struct pci_dev *pdev = dev_priv->drm.pdev;
 	int ret;
 
 	pci_write_config_byte(pdev, I915_GDRST,
@@ -1530,15 +1532,17 @@ static int ironlake_do_reset(struct drm_i915_private *dev_priv,
 
 	I915_WRITE(ILK_GDSR,
 		   ILK_GRDOM_RENDER | ILK_GRDOM_RESET_ENABLE);
-	ret = wait_for((I915_READ(ILK_GDSR) &
-			ILK_GRDOM_RESET_ENABLE) == 0, 500);
+	ret = intel_wait_for_register(dev_priv,
+				      ILK_GDSR, ILK_GRDOM_RESET_ENABLE, 0,
+				      500);
 	if (ret)
 		return ret;
 
 	I915_WRITE(ILK_GDSR,
 		   ILK_GRDOM_MEDIA | ILK_GRDOM_RESET_ENABLE);
-	ret = wait_for((I915_READ(ILK_GDSR) &
-			ILK_GRDOM_RESET_ENABLE) == 0, 500);
+	ret = intel_wait_for_register(dev_priv,
+				      ILK_GDSR, ILK_GRDOM_RESET_ENABLE, 0,
+				      500);
 	if (ret)
 		return ret;
 
@@ -1551,20 +1555,16 @@ static int ironlake_do_reset(struct drm_i915_private *dev_priv,
 static int gen6_hw_domain_reset(struct drm_i915_private *dev_priv,
 				u32 hw_domain_mask)
 {
-	int ret;
-
 	/* GEN6_GDRST is not in the gt power well, no need to check
 	 * for fifo space for the write or forcewake the chip for
 	 * the read
 	 */
 	__raw_i915_write32(dev_priv, GEN6_GDRST, hw_domain_mask);
 
-#define ACKED ((__raw_i915_read32(dev_priv, GEN6_GDRST) & hw_domain_mask) == 0)
 	/* Spin waiting for the device to ack the reset requests */
-	ret = wait_for(ACKED, 500);
-#undef ACKED
-
-	return ret;
+	return intel_wait_for_register_fw(dev_priv,
+					  GEN6_GDRST, hw_domain_mask, 0,
+					  500);
 }
 
 /**
@@ -1609,13 +1609,74 @@ static int gen6_reset_engines(struct drm_i915_private *dev_priv,
 	return ret;
 }
 
-static int wait_for_register_fw(struct drm_i915_private *dev_priv,
-				i915_reg_t reg,
-				const u32 mask,
-				const u32 value,
-				const unsigned long timeout_ms)
+/**
+ * intel_wait_for_register_fw - wait until register matches expected state
+ * @dev_priv: the i915 device
+ * @reg: the register to read
+ * @mask: mask to apply to register value
+ * @value: expected value
+ * @timeout_ms: timeout in millisecond
+ *
+ * This routine waits until the target register @reg contains the expected
+ * @value after applying the @mask, i.e. it waits until
+ *   (I915_READ_FW(@reg) & @mask) == @value
+ * Otherwise, the wait will timeout after @timeout_ms milliseconds.
+ *
+ * Note that this routine assumes the caller holds forcewake asserted, it is
+ * not suitable for very long waits. See intel_wait_for_register() if you
+ * wish to wait without holding forcewake for the duration (i.e. you expect
+ * the wait to be slow).
+ *
+ * Returns 0 if the register matches the desired condition, or -ETIMEOUT.
+ */
+int intel_wait_for_register_fw(struct drm_i915_private *dev_priv,
+			       i915_reg_t reg,
+			       const u32 mask,
+			       const u32 value,
+			       const unsigned long timeout_ms)
+{
+#define done ((I915_READ_FW(reg) & mask) == value)
+	int ret = wait_for_us(done, 2);
+	if (ret)
+		ret = wait_for(done, timeout_ms);
+	return ret;
+#undef done
+}
+
+/**
+ * intel_wait_for_register - wait until register matches expected state
+ * @dev_priv: the i915 device
+ * @reg: the register to read
+ * @mask: mask to apply to register value
+ * @value: expected value
+ * @timeout_ms: timeout in millisecond
+ *
+ * This routine waits until the target register @reg contains the expected
+ * @value after applying the @mask, i.e. it waits until
+ *   (I915_READ(@reg) & @mask) == @value
+ * Otherwise, the wait will timeout after @timeout_ms milliseconds.
+ *
+ * Returns 0 if the register matches the desired condition, or -ETIMEOUT.
+ */
+int intel_wait_for_register(struct drm_i915_private *dev_priv,
+			    i915_reg_t reg,
+			    const u32 mask,
+			    const u32 value,
+			    const unsigned long timeout_ms)
 {
-	return wait_for((I915_READ_FW(reg) & mask) == value, timeout_ms);
+
+	unsigned fw =
+		intel_uncore_forcewake_for_reg(dev_priv, reg, FW_REG_READ);
+	int ret;
+
+	intel_uncore_forcewake_get(dev_priv, fw);
+	ret = wait_for_us((I915_READ_FW(reg) & mask) == value, 2);
+	intel_uncore_forcewake_put(dev_priv, fw);
+	if (ret)
+		ret = wait_for((I915_READ_NOTRACE(reg) & mask) == value,
+			       timeout_ms);
+
+	return ret;
 }
 
 static int gen8_request_engine_reset(struct intel_engine_cs *engine)
@@ -1626,11 +1687,11 @@ static int gen8_request_engine_reset(struct intel_engine_cs *engine)
 	I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base),
 		      _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET));
 
-	ret = wait_for_register_fw(dev_priv,
-				   RING_RESET_CTL(engine->mmio_base),
-				   RESET_CTL_READY_TO_RESET,
-				   RESET_CTL_READY_TO_RESET,
-				   700);
+	ret = intel_wait_for_register_fw(dev_priv,
+					 RING_RESET_CTL(engine->mmio_base),
+					 RESET_CTL_READY_TO_RESET,
+					 RESET_CTL_READY_TO_RESET,
+					 700);
 	if (ret)
 		DRM_ERROR("%s: reset request timeout\n", engine->name);
 
diff --git a/drivers/gpu/drm/imx/dw_hdmi-imx.c b/drivers/gpu/drm/imx/dw_hdmi-imx.c
index a24631fdf4ad..359cd2765552 100644
--- a/drivers/gpu/drm/imx/dw_hdmi-imx.c
+++ b/drivers/gpu/drm/imx/dw_hdmi-imx.c
@@ -28,6 +28,11 @@ struct imx_hdmi {
 	struct regmap *regmap;
 };
 
+static inline struct imx_hdmi *enc_to_imx_hdmi(struct drm_encoder *e)
+{
+	return container_of(e, struct imx_hdmi, encoder);
+}
+
 static const struct dw_hdmi_mpll_config imx_mpll_cfg[] = {
 	{
 		45250000, {
@@ -109,15 +114,9 @@ static void dw_hdmi_imx_encoder_disable(struct drm_encoder *encoder)
 {
 }
 
-static void dw_hdmi_imx_encoder_mode_set(struct drm_encoder *encoder,
-					 struct drm_display_mode *mode,
-					 struct drm_display_mode *adj_mode)
+static void dw_hdmi_imx_encoder_enable(struct drm_encoder *encoder)
 {
-}
-
-static void dw_hdmi_imx_encoder_commit(struct drm_encoder *encoder)
-{
-	struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder);
+	struct imx_hdmi *hdmi = enc_to_imx_hdmi(encoder);
 	int mux = drm_of_encoder_active_port_id(hdmi->dev->of_node, encoder);
 
 	regmap_update_bits(hdmi->regmap, IOMUXC_GPR3,
@@ -125,16 +124,23 @@ static void dw_hdmi_imx_encoder_commit(struct drm_encoder *encoder)
 			   mux << IMX6Q_GPR3_HDMI_MUX_CTL_SHIFT);
 }
 
-static void dw_hdmi_imx_encoder_prepare(struct drm_encoder *encoder)
+static int dw_hdmi_imx_atomic_check(struct drm_encoder *encoder,
+				    struct drm_crtc_state *crtc_state,
+				    struct drm_connector_state *conn_state)
 {
-	imx_drm_set_bus_format(encoder, MEDIA_BUS_FMT_RGB888_1X24);
+	struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state);
+
+	imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB888_1X24;
+	imx_crtc_state->di_hsync_pin = 2;
+	imx_crtc_state->di_vsync_pin = 3;
+
+	return 0;
 }
 
 static const struct drm_encoder_helper_funcs dw_hdmi_imx_encoder_helper_funcs = {
-	.mode_set   = dw_hdmi_imx_encoder_mode_set,
-	.prepare    = dw_hdmi_imx_encoder_prepare,
-	.commit     = dw_hdmi_imx_encoder_commit,
+	.enable     = dw_hdmi_imx_encoder_enable,
 	.disable    = dw_hdmi_imx_encoder_disable,
+	.atomic_check = dw_hdmi_imx_atomic_check,
 };
 
 static const struct drm_encoder_funcs dw_hdmi_imx_encoder_funcs = {
diff --git a/drivers/gpu/drm/imx/imx-drm-core.c b/drivers/gpu/drm/imx/imx-drm-core.c
index 7746418a4c08..9f7dafce3a4c 100644
--- a/drivers/gpu/drm/imx/imx-drm-core.c
+++ b/drivers/gpu/drm/imx/imx-drm-core.c
@@ -15,10 +15,14 @@
  */
 #include <linux/component.h>
 #include <linux/device.h>
+#include <linux/dma-buf.h>
 #include <linux/fb.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
+#include <linux/reservation.h>
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_gem_cma_helper.h>
@@ -41,6 +45,7 @@ struct imx_drm_device {
 	struct imx_drm_crtc			*crtc[MAX_CRTC];
 	unsigned int				pipes;
 	struct drm_fbdev_cma			*fbhelper;
+	struct drm_atomic_state			*state;
 };
 
 struct imx_drm_crtc {
@@ -85,45 +90,6 @@ static int imx_drm_driver_unload(struct drm_device *drm)
 	return 0;
 }
 
-static struct imx_drm_crtc *imx_drm_find_crtc(struct drm_crtc *crtc)
-{
-	struct imx_drm_device *imxdrm = crtc->dev->dev_private;
-	unsigned i;
-
-	for (i = 0; i < MAX_CRTC; i++)
-		if (imxdrm->crtc[i] && imxdrm->crtc[i]->crtc == crtc)
-			return imxdrm->crtc[i];
-
-	return NULL;
-}
-
-int imx_drm_set_bus_config(struct drm_encoder *encoder, u32 bus_format,
-		int hsync_pin, int vsync_pin, u32 bus_flags)
-{
-	struct imx_drm_crtc_helper_funcs *helper;
-	struct imx_drm_crtc *imx_crtc;
-
-	imx_crtc = imx_drm_find_crtc(encoder->crtc);
-	if (!imx_crtc)
-		return -EINVAL;
-
-	helper = &imx_crtc->imx_drm_helper_funcs;
-	if (helper->set_interface_pix_fmt)
-		return helper->set_interface_pix_fmt(encoder->crtc,
-					bus_format, hsync_pin, vsync_pin,
-					bus_flags);
-	return 0;
-}
-EXPORT_SYMBOL_GPL(imx_drm_set_bus_config);
-
-int imx_drm_set_bus_format(struct drm_encoder *encoder, u32 bus_format)
-{
-	return imx_drm_set_bus_config(encoder, bus_format, 2, 3,
-				      DRM_BUS_FLAG_DE_HIGH |
-				      DRM_BUS_FLAG_PIXDATA_NEGEDGE);
-}
-EXPORT_SYMBOL_GPL(imx_drm_set_bus_format);
-
 int imx_drm_crtc_vblank_get(struct imx_drm_crtc *imx_drm_crtc)
 {
 	return drm_crtc_vblank_get(imx_drm_crtc->crtc);
@@ -208,6 +174,63 @@ static void imx_drm_output_poll_changed(struct drm_device *drm)
 static const struct drm_mode_config_funcs imx_drm_mode_config_funcs = {
 	.fb_create = drm_fb_cma_create,
 	.output_poll_changed = imx_drm_output_poll_changed,
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = drm_atomic_helper_commit,
+};
+
+static void imx_drm_atomic_commit_tail(struct drm_atomic_state *state)
+{
+	struct drm_device *dev = state->dev;
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *crtc_state;
+	struct drm_plane_state *plane_state;
+	struct drm_gem_cma_object *cma_obj;
+	struct fence *excl;
+	unsigned shared_count;
+	struct fence **shared;
+	unsigned int i, j;
+	int ret;
+
+	/* Wait for fences. */
+	for_each_crtc_in_state(state, crtc, crtc_state, i) {
+		plane_state = crtc->primary->state;
+		if (plane_state->fb) {
+			cma_obj = drm_fb_cma_get_gem_obj(plane_state->fb, 0);
+			if (cma_obj->base.dma_buf) {
+				ret = reservation_object_get_fences_rcu(
+					cma_obj->base.dma_buf->resv, &excl,
+					&shared_count, &shared);
+				if (unlikely(ret))
+					DRM_ERROR("failed to get fences "
+						  "for buffer\n");
+
+				if (excl) {
+					fence_wait(excl, false);
+					fence_put(excl);
+				}
+				for (j = 0; j < shared_count; i++) {
+					fence_wait(shared[j], false);
+					fence_put(shared[j]);
+				}
+			}
+		}
+	}
+
+	drm_atomic_helper_commit_modeset_disables(dev, state);
+
+	drm_atomic_helper_commit_planes(dev, state, true);
+
+	drm_atomic_helper_commit_modeset_enables(dev, state);
+
+	drm_atomic_helper_commit_hw_done(state);
+
+	drm_atomic_helper_wait_for_vblanks(dev, state);
+
+	drm_atomic_helper_cleanup_planes(dev, state);
+}
+
+static struct drm_mode_config_helper_funcs imx_drm_mode_config_helpers = {
+	.atomic_commit_tail = imx_drm_atomic_commit_tail,
 };
 
 /*
@@ -249,6 +272,7 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags)
 	drm->mode_config.max_width = 4096;
 	drm->mode_config.max_height = 4096;
 	drm->mode_config.funcs = &imx_drm_mode_config_funcs;
+	drm->mode_config.helper_private = &imx_drm_mode_config_helpers;
 
 	drm_mode_config_init(drm);
 
@@ -279,6 +303,8 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags)
 		}
 	}
 
+	drm_mode_config_reset(drm);
+
 	/*
 	 * All components are now initialised, so setup the fb helper.
 	 * The fb helper takes copies of key hardware information, so the
@@ -289,7 +315,6 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags)
 		dev_warn(drm->dev, "Invalid legacyfb_depth.  Defaulting to 16bpp\n");
 		legacyfb_depth = 16;
 	}
-	drm_helper_disable_unused_functions(drm);
 	imxdrm->fbhelper = drm_fbdev_cma_init(drm, legacyfb_depth,
 				drm->mode_config.num_crtc, MAX_CRTC);
 	if (IS_ERR(imxdrm->fbhelper)) {
@@ -403,7 +428,8 @@ static const struct drm_ioctl_desc imx_drm_ioctls[] = {
 };
 
 static struct drm_driver imx_drm_driver = {
-	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
+	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME |
+				  DRIVER_ATOMIC,
 	.load			= imx_drm_driver_load,
 	.unload			= imx_drm_driver_unload,
 	.lastclose		= imx_drm_driver_lastclose,
@@ -491,6 +517,7 @@ static int imx_drm_platform_remove(struct platform_device *pdev)
 static int imx_drm_suspend(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct imx_drm_device *imxdrm;
 
 	/* The drm_dev is NULL before .load hook is called */
 	if (drm_dev == NULL)
@@ -498,17 +525,26 @@ static int imx_drm_suspend(struct device *dev)
 
 	drm_kms_helper_poll_disable(drm_dev);
 
+	imxdrm = drm_dev->dev_private;
+	imxdrm->state = drm_atomic_helper_suspend(drm_dev);
+	if (IS_ERR(imxdrm->state)) {
+		drm_kms_helper_poll_enable(drm_dev);
+		return PTR_ERR(imxdrm->state);
+	}
+
 	return 0;
 }
 
 static int imx_drm_resume(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
+	struct imx_drm_device *imx_drm;
 
 	if (drm_dev == NULL)
 		return 0;
 
-	drm_helper_resume_force_mode(drm_dev);
+	imx_drm = drm_dev->dev_private;
+	drm_atomic_helper_resume(drm_dev, imx_drm->state);
 	drm_kms_helper_poll_enable(drm_dev);
 
 	return 0;
diff --git a/drivers/gpu/drm/imx/imx-drm.h b/drivers/gpu/drm/imx/imx-drm.h
index 74320a1723b7..07d33e45f90f 100644
--- a/drivers/gpu/drm/imx/imx-drm.h
+++ b/drivers/gpu/drm/imx/imx-drm.h
@@ -15,12 +15,22 @@ struct platform_device;
 
 unsigned int imx_drm_crtc_id(struct imx_drm_crtc *crtc);
 
+struct imx_crtc_state {
+	struct drm_crtc_state			base;
+	u32					bus_format;
+	u32					bus_flags;
+	int					di_hsync_pin;
+	int					di_vsync_pin;
+};
+
+static inline struct imx_crtc_state *to_imx_crtc_state(struct drm_crtc_state *s)
+{
+	return container_of(s, struct imx_crtc_state, base);
+}
+
 struct imx_drm_crtc_helper_funcs {
 	int (*enable_vblank)(struct drm_crtc *crtc);
 	void (*disable_vblank)(struct drm_crtc *crtc);
-	int (*set_interface_pix_fmt)(struct drm_crtc *crtc,
-			u32 bus_format, int hsync_pin, int vsync_pin,
-			u32 bus_flags);
 	const struct drm_crtc_helper_funcs *crtc_helper_funcs;
 	const struct drm_crtc_funcs *crtc_funcs;
 };
@@ -42,11 +52,6 @@ void imx_drm_mode_config_init(struct drm_device *drm);
 
 struct drm_gem_cma_object *imx_drm_fb_get_obj(struct drm_framebuffer *fb);
 
-int imx_drm_set_bus_config(struct drm_encoder *encoder, u32 bus_format,
-		int hsync_pin, int vsync_pin, u32 bus_flags);
-int imx_drm_set_bus_format(struct drm_encoder *encoder,
-		u32 bus_format);
-
 int imx_drm_encoder_parse_of(struct drm_device *drm,
 	struct drm_encoder *encoder, struct device_node *np);
 
diff --git a/drivers/gpu/drm/imx/imx-ldb.c b/drivers/gpu/drm/imx/imx-ldb.c
index beff793bb717..5d2831dfb8b9 100644
--- a/drivers/gpu/drm/imx/imx-ldb.c
+++ b/drivers/gpu/drm/imx/imx-ldb.c
@@ -17,6 +17,8 @@
 #include <linux/clk.h>
 #include <linux/component.h>
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_of.h>
@@ -49,9 +51,6 @@
 #define LDB_DI1_VS_POL_ACT_LOW		(1 << 10)
 #define LDB_BGREF_RMODE_INT		(1 << 15)
 
-#define con_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, connector)
-#define enc_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, encoder)
-
 struct imx_ldb;
 
 struct imx_ldb_channel {
@@ -66,9 +65,19 @@ struct imx_ldb_channel {
 	int edid_len;
 	struct drm_display_mode mode;
 	int mode_valid;
-	int bus_format;
+	u32 bus_format;
 };
 
+static inline struct imx_ldb_channel *con_to_imx_ldb_ch(struct drm_connector *c)
+{
+	return container_of(c, struct imx_ldb_channel, connector);
+}
+
+static inline struct imx_ldb_channel *enc_to_imx_ldb_ch(struct drm_encoder *e)
+{
+	return container_of(e, struct imx_ldb_channel, encoder);
+}
+
 struct bus_mux {
 	int reg;
 	int shift;
@@ -93,6 +102,32 @@ static enum drm_connector_status imx_ldb_connector_detect(
 	return connector_status_connected;
 }
 
+static void imx_ldb_ch_set_bus_format(struct imx_ldb_channel *imx_ldb_ch,
+				      u32 bus_format)
+{
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN;
+
+	switch (bus_format) {
+	case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG:
+		break;
+	case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG:
+		if (imx_ldb_ch->chno == 0 || dual)
+			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24;
+		if (imx_ldb_ch->chno == 1 || dual)
+			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24;
+		break;
+	case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA:
+		if (imx_ldb_ch->chno == 0 || dual)
+			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24 |
+					 LDB_BIT_MAP_CH0_JEIDA;
+		if (imx_ldb_ch->chno == 1 || dual)
+			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24 |
+					 LDB_BIT_MAP_CH1_JEIDA;
+		break;
+	}
+}
+
 static int imx_ldb_connector_get_modes(struct drm_connector *connector)
 {
 	struct imx_ldb_channel *imx_ldb_ch = con_to_imx_ldb_ch(connector);
@@ -100,11 +135,7 @@ static int imx_ldb_connector_get_modes(struct drm_connector *connector)
 
 	if (imx_ldb_ch->panel && imx_ldb_ch->panel->funcs &&
 	    imx_ldb_ch->panel->funcs->get_modes) {
-		struct drm_display_info *di = &connector->display_info;
-
 		num_modes = imx_ldb_ch->panel->funcs->get_modes(imx_ldb_ch->panel);
-		if (!imx_ldb_ch->bus_format && di->num_bus_formats)
-			imx_ldb_ch->bus_format = di->bus_formats[0];
 		if (num_modes > 0)
 			return num_modes;
 	}
@@ -141,10 +172,6 @@ static struct drm_encoder *imx_ldb_connector_best_encoder(
 	return &imx_ldb_ch->encoder;
 }
 
-static void imx_ldb_encoder_dpms(struct drm_encoder *encoder, int mode)
-{
-}
-
 static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno,
 		unsigned long serial_clk, unsigned long di_clk)
 {
@@ -173,43 +200,7 @@ static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno,
 			chno);
 }
 
-static void imx_ldb_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
-	struct imx_ldb *ldb = imx_ldb_ch->ldb;
-	int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN;
-	u32 bus_format;
-
-	switch (imx_ldb_ch->bus_format) {
-	default:
-		dev_warn(ldb->dev,
-			 "could not determine data mapping, default to 18-bit \"spwg\"\n");
-		/* fallthrough */
-	case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG:
-		bus_format = MEDIA_BUS_FMT_RGB666_1X18;
-		break;
-	case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG:
-		bus_format = MEDIA_BUS_FMT_RGB888_1X24;
-		if (imx_ldb_ch->chno == 0 || dual)
-			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24;
-		if (imx_ldb_ch->chno == 1 || dual)
-			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24;
-		break;
-	case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA:
-		bus_format = MEDIA_BUS_FMT_RGB888_1X24;
-		if (imx_ldb_ch->chno == 0 || dual)
-			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24 |
-					 LDB_BIT_MAP_CH0_JEIDA;
-		if (imx_ldb_ch->chno == 1 || dual)
-			ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24 |
-					 LDB_BIT_MAP_CH1_JEIDA;
-		break;
-	}
-
-	imx_drm_set_bus_format(encoder, bus_format);
-}
-
-static void imx_ldb_encoder_commit(struct drm_encoder *encoder)
+static void imx_ldb_encoder_enable(struct drm_encoder *encoder)
 {
 	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
 	struct imx_ldb *ldb = imx_ldb_ch->ldb;
@@ -219,8 +210,13 @@ static void imx_ldb_encoder_commit(struct drm_encoder *encoder)
 	drm_panel_prepare(imx_ldb_ch->panel);
 
 	if (dual) {
+		clk_set_parent(ldb->clk_sel[mux], ldb->clk[0]);
+		clk_set_parent(ldb->clk_sel[mux], ldb->clk[1]);
+
 		clk_prepare_enable(ldb->clk[0]);
 		clk_prepare_enable(ldb->clk[1]);
+	} else {
+		clk_set_parent(ldb->clk_sel[mux], ldb->clk[imx_ldb_ch->chno]);
 	}
 
 	if (imx_ldb_ch == &ldb->channel[0] || dual) {
@@ -265,6 +261,7 @@ static void imx_ldb_encoder_mode_set(struct drm_encoder *encoder,
 	unsigned long serial_clk;
 	unsigned long di_clk = mode->clock * 1000;
 	int mux = drm_of_encoder_active_port_id(imx_ldb_ch->child, encoder);
+	u32 bus_format = imx_ldb_ch->bus_format;
 
 	if (mode->clock > 170000) {
 		dev_warn(ldb->dev,
@@ -286,18 +283,36 @@ static void imx_ldb_encoder_mode_set(struct drm_encoder *encoder,
 	}
 
 	/* FIXME - assumes straight connections DI0 --> CH0, DI1 --> CH1 */
-	if (imx_ldb_ch == &ldb->channel[0]) {
+	if (imx_ldb_ch == &ldb->channel[0] || dual) {
 		if (mode->flags & DRM_MODE_FLAG_NVSYNC)
 			ldb->ldb_ctrl |= LDB_DI0_VS_POL_ACT_LOW;
 		else if (mode->flags & DRM_MODE_FLAG_PVSYNC)
 			ldb->ldb_ctrl &= ~LDB_DI0_VS_POL_ACT_LOW;
 	}
-	if (imx_ldb_ch == &ldb->channel[1]) {
+	if (imx_ldb_ch == &ldb->channel[1] || dual) {
 		if (mode->flags & DRM_MODE_FLAG_NVSYNC)
 			ldb->ldb_ctrl |= LDB_DI1_VS_POL_ACT_LOW;
 		else if (mode->flags & DRM_MODE_FLAG_PVSYNC)
 			ldb->ldb_ctrl &= ~LDB_DI1_VS_POL_ACT_LOW;
 	}
+
+	if (!bus_format) {
+		struct drm_connector_state *conn_state;
+		struct drm_connector *connector;
+		int i;
+
+		for_each_connector_in_state(encoder->crtc->state->state,
+					    connector, conn_state, i) {
+			struct drm_display_info *di = &connector->display_info;
+
+			if (conn_state->crtc == encoder->crtc &&
+			    di->num_bus_formats) {
+				bus_format = di->bus_formats[0];
+				break;
+			}
+		}
+	}
+	imx_ldb_ch_set_bus_format(imx_ldb_ch, bus_format);
 }
 
 static void imx_ldb_encoder_disable(struct drm_encoder *encoder)
@@ -357,11 +372,45 @@ static void imx_ldb_encoder_disable(struct drm_encoder *encoder)
 	drm_panel_unprepare(imx_ldb_ch->panel);
 }
 
+static int imx_ldb_encoder_atomic_check(struct drm_encoder *encoder,
+					struct drm_crtc_state *crtc_state,
+					struct drm_connector_state *conn_state)
+{
+	struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state);
+	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
+	struct drm_display_info *di = &conn_state->connector->display_info;
+	u32 bus_format = imx_ldb_ch->bus_format;
+
+	/* Bus format description in DT overrides connector display info. */
+	if (!bus_format && di->num_bus_formats)
+		bus_format = di->bus_formats[0];
+	switch (bus_format) {
+	case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG:
+		imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB666_1X18;
+		break;
+	case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG:
+	case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA:
+		imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB888_1X24;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	imx_crtc_state->di_hsync_pin = 2;
+	imx_crtc_state->di_vsync_pin = 3;
+
+	return 0;
+}
+
+
 static const struct drm_connector_funcs imx_ldb_connector_funcs = {
-	.dpms = drm_helper_connector_dpms,
+	.dpms = drm_atomic_helper_connector_dpms,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.detect = imx_ldb_connector_detect,
 	.destroy = imx_drm_connector_destroy,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
 };
 
 static const struct drm_connector_helper_funcs imx_ldb_connector_helper_funcs = {
@@ -374,11 +423,10 @@ static const struct drm_encoder_funcs imx_ldb_encoder_funcs = {
 };
 
 static const struct drm_encoder_helper_funcs imx_ldb_encoder_helper_funcs = {
-	.dpms = imx_ldb_encoder_dpms,
-	.prepare = imx_ldb_encoder_prepare,
-	.commit = imx_ldb_encoder_commit,
 	.mode_set = imx_ldb_encoder_mode_set,
+	.enable = imx_ldb_encoder_enable,
 	.disable = imx_ldb_encoder_disable,
+	.atomic_check = imx_ldb_encoder_atomic_check,
 };
 
 static int imx_ldb_get_clk(struct imx_ldb *ldb, int chno)
@@ -400,10 +448,10 @@ static int imx_ldb_register(struct drm_device *drm,
 	struct imx_ldb_channel *imx_ldb_ch)
 {
 	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	struct drm_encoder *encoder = &imx_ldb_ch->encoder;
 	int ret;
 
-	ret = imx_drm_encoder_parse_of(drm, &imx_ldb_ch->encoder,
-				       imx_ldb_ch->child);
+	ret = imx_drm_encoder_parse_of(drm, encoder, imx_ldb_ch->child);
 	if (ret)
 		return ret;
 
@@ -417,9 +465,8 @@ static int imx_ldb_register(struct drm_device *drm,
 			return ret;
 	}
 
-	drm_encoder_helper_add(&imx_ldb_ch->encoder,
-			&imx_ldb_encoder_helper_funcs);
-	drm_encoder_init(drm, &imx_ldb_ch->encoder, &imx_ldb_encoder_funcs,
+	drm_encoder_helper_add(encoder, &imx_ldb_encoder_helper_funcs);
+	drm_encoder_init(drm, encoder, &imx_ldb_encoder_funcs,
 			 DRM_MODE_ENCODER_LVDS, NULL);
 
 	drm_connector_helper_add(&imx_ldb_ch->connector,
@@ -427,11 +474,14 @@ static int imx_ldb_register(struct drm_device *drm,
 	drm_connector_init(drm, &imx_ldb_ch->connector,
 			   &imx_ldb_connector_funcs, DRM_MODE_CONNECTOR_LVDS);
 
-	if (imx_ldb_ch->panel)
-		drm_panel_attach(imx_ldb_ch->panel, &imx_ldb_ch->connector);
+	if (imx_ldb_ch->panel) {
+		ret = drm_panel_attach(imx_ldb_ch->panel,
+				       &imx_ldb_ch->connector);
+		if (ret)
+			return ret;
+	}
 
-	drm_mode_connector_attach_encoder(&imx_ldb_ch->connector,
-			&imx_ldb_ch->encoder);
+	drm_mode_connector_attach_encoder(&imx_ldb_ch->connector, encoder);
 
 	return 0;
 }
@@ -560,6 +610,7 @@ static int imx_ldb_bind(struct device *dev, struct device *master, void *data)
 		struct imx_ldb_channel *channel;
 		struct device_node *ddc_node;
 		struct device_node *ep;
+		int bus_format;
 
 		ret = of_property_read_u32(child, "reg", &i);
 		if (ret || i < 0 || i > 1)
@@ -632,21 +683,22 @@ static int imx_ldb_bind(struct device *dev, struct device *master, void *data)
 			}
 		}
 
-		channel->bus_format = of_get_bus_format(dev, child);
-		if (channel->bus_format == -EINVAL) {
+		bus_format = of_get_bus_format(dev, child);
+		if (bus_format == -EINVAL) {
 			/*
 			 * If no bus format was specified in the device tree,
 			 * we can still get it from the connected panel later.
 			 */
 			if (channel->panel && channel->panel->funcs &&
 			    channel->panel->funcs->get_modes)
-				channel->bus_format = 0;
+				bus_format = 0;
 		}
-		if (channel->bus_format < 0) {
+		if (bus_format < 0) {
 			dev_err(dev, "could not determine data mapping: %d\n",
-				channel->bus_format);
-			return channel->bus_format;
+				bus_format);
+			return bus_format;
 		}
+		channel->bus_format = bus_format;
 
 		ret = imx_ldb_register(drm, channel);
 		if (ret)
diff --git a/drivers/gpu/drm/imx/imx-tve.c b/drivers/gpu/drm/imx/imx-tve.c
index baf788121287..5e875944ffa2 100644
--- a/drivers/gpu/drm/imx/imx-tve.c
+++ b/drivers/gpu/drm/imx/imx-tve.c
@@ -23,6 +23,7 @@
 #include <linux/spinlock.h>
 #include <linux/videodev2.h>
 #include <drm/drmP.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <video/imx-ipu-v3.h>
@@ -97,9 +98,6 @@
 /* TVE_TST_MODE_REG */
 #define TVE_TVDAC_TEST_MODE_MASK	(0x7 << 0)
 
-#define con_to_tve(x) container_of(x, struct imx_tve, connector)
-#define enc_to_tve(x) container_of(x, struct imx_tve, encoder)
-
 enum {
 	TVE_MODE_TVOUT,
 	TVE_MODE_VGA,
@@ -112,6 +110,8 @@ struct imx_tve {
 	spinlock_t lock;	/* register lock */
 	bool enabled;
 	int mode;
+	int di_hsync_pin;
+	int di_vsync_pin;
 
 	struct regmap *regmap;
 	struct regulator *dac_reg;
@@ -120,10 +120,18 @@ struct imx_tve {
 	struct clk *di_sel_clk;
 	struct clk_hw clk_hw_di;
 	struct clk *di_clk;
-	int vsync_pin;
-	int hsync_pin;
 };
 
+static inline struct imx_tve *con_to_tve(struct drm_connector *c)
+{
+	return container_of(c, struct imx_tve, connector);
+}
+
+static inline struct imx_tve *enc_to_tve(struct drm_encoder *e)
+{
+	return container_of(e, struct imx_tve, encoder);
+}
+
 static void tve_lock(void *__tve)
 __acquires(&tve->lock)
 {
@@ -148,8 +156,7 @@ static void tve_enable(struct imx_tve *tve)
 		tve->enabled = true;
 		clk_prepare_enable(tve->clk);
 		ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
-					 TVE_IPU_CLK_EN | TVE_EN,
-					 TVE_IPU_CLK_EN | TVE_EN);
+					 TVE_EN, TVE_EN);
 	}
 
 	/* clear interrupt status register */
@@ -172,7 +179,7 @@ static void tve_disable(struct imx_tve *tve)
 	if (tve->enabled) {
 		tve->enabled = false;
 		ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
-					 TVE_IPU_CLK_EN | TVE_EN, 0);
+					 TVE_EN, 0);
 		clk_disable_unprepare(tve->clk);
 	}
 }
@@ -275,36 +282,6 @@ static struct drm_encoder *imx_tve_connector_best_encoder(
 	return &tve->encoder;
 }
 
-static void imx_tve_encoder_dpms(struct drm_encoder *encoder, int mode)
-{
-	struct imx_tve *tve = enc_to_tve(encoder);
-	int ret;
-
-	ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
-				 TVE_TV_OUT_MODE_MASK, TVE_TV_OUT_DISABLE);
-	if (ret < 0)
-		dev_err(tve->dev, "failed to disable TVOUT: %d\n", ret);
-}
-
-static void imx_tve_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct imx_tve *tve = enc_to_tve(encoder);
-
-	tve_disable(tve);
-
-	switch (tve->mode) {
-	case TVE_MODE_VGA:
-		imx_drm_set_bus_config(encoder, MEDIA_BUS_FMT_GBR888_1X24,
-				       tve->hsync_pin, tve->vsync_pin,
-				       DRM_BUS_FLAG_DE_HIGH |
-				       DRM_BUS_FLAG_PIXDATA_NEGEDGE);
-		break;
-	case TVE_MODE_TVOUT:
-		imx_drm_set_bus_format(encoder, MEDIA_BUS_FMT_YUV8_1X24);
-		break;
-	}
-}
-
 static void imx_tve_encoder_mode_set(struct drm_encoder *encoder,
 				     struct drm_display_mode *orig_mode,
 				     struct drm_display_mode *mode)
@@ -333,6 +310,9 @@ static void imx_tve_encoder_mode_set(struct drm_encoder *encoder,
 			ret);
 	}
 
+	regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
+			   TVE_IPU_CLK_EN, TVE_IPU_CLK_EN);
+
 	if (tve->mode == TVE_MODE_VGA)
 		ret = tve_setup_vga(tve);
 	else
@@ -341,7 +321,7 @@ static void imx_tve_encoder_mode_set(struct drm_encoder *encoder,
 		dev_err(tve->dev, "failed to set configuration: %d\n", ret);
 }
 
-static void imx_tve_encoder_commit(struct drm_encoder *encoder)
+static void imx_tve_encoder_enable(struct drm_encoder *encoder)
 {
 	struct imx_tve *tve = enc_to_tve(encoder);
 
@@ -355,11 +335,28 @@ static void imx_tve_encoder_disable(struct drm_encoder *encoder)
 	tve_disable(tve);
 }
 
+static int imx_tve_atomic_check(struct drm_encoder *encoder,
+				struct drm_crtc_state *crtc_state,
+				struct drm_connector_state *conn_state)
+{
+	struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state);
+	struct imx_tve *tve = enc_to_tve(encoder);
+
+	imx_crtc_state->bus_format = MEDIA_BUS_FMT_GBR888_1X24;
+	imx_crtc_state->di_hsync_pin = tve->di_hsync_pin;
+	imx_crtc_state->di_vsync_pin = tve->di_vsync_pin;
+
+	return 0;
+}
+
 static const struct drm_connector_funcs imx_tve_connector_funcs = {
-	.dpms = drm_helper_connector_dpms,
+	.dpms = drm_atomic_helper_connector_dpms,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.detect = imx_tve_connector_detect,
 	.destroy = imx_drm_connector_destroy,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
 };
 
 static const struct drm_connector_helper_funcs imx_tve_connector_helper_funcs = {
@@ -373,11 +370,10 @@ static const struct drm_encoder_funcs imx_tve_encoder_funcs = {
 };
 
 static const struct drm_encoder_helper_funcs imx_tve_encoder_helper_funcs = {
-	.dpms = imx_tve_encoder_dpms,
-	.prepare = imx_tve_encoder_prepare,
 	.mode_set = imx_tve_encoder_mode_set,
-	.commit = imx_tve_encoder_commit,
+	.enable = imx_tve_encoder_enable,
 	.disable = imx_tve_encoder_disable,
+	.atomic_check = imx_tve_atomic_check,
 };
 
 static irqreturn_t imx_tve_irq_handler(int irq, void *data)
@@ -495,8 +491,7 @@ static int imx_tve_register(struct drm_device *drm, struct imx_tve *tve)
 	encoder_type = tve->mode == TVE_MODE_VGA ?
 				DRM_MODE_ENCODER_DAC : DRM_MODE_ENCODER_TVDAC;
 
-	ret = imx_drm_encoder_parse_of(drm, &tve->encoder,
-				       tve->dev->of_node);
+	ret = imx_drm_encoder_parse_of(drm, &tve->encoder, tve->dev->of_node);
 	if (ret)
 		return ret;
 
@@ -587,15 +582,15 @@ static int imx_tve_bind(struct device *dev, struct device *master, void *data)
 
 	if (tve->mode == TVE_MODE_VGA) {
 		ret = of_property_read_u32(np, "fsl,hsync-pin",
-					   &tve->hsync_pin);
+					   &tve->di_hsync_pin);
 
 		if (ret < 0) {
-			dev_err(dev, "failed to get vsync pin\n");
+			dev_err(dev, "failed to get hsync pin\n");
 			return ret;
 		}
 
-		ret |= of_property_read_u32(np, "fsl,vsync-pin",
-					    &tve->vsync_pin);
+		ret = of_property_read_u32(np, "fsl,vsync-pin",
+					   &tve->di_vsync_pin);
 
 		if (ret < 0) {
 			dev_err(dev, "failed to get vsync pin\n");
@@ -633,7 +628,9 @@ static int imx_tve_bind(struct device *dev, struct device *master, void *data)
 
 	tve->dac_reg = devm_regulator_get(dev, "dac");
 	if (!IS_ERR(tve->dac_reg)) {
-		regulator_set_voltage(tve->dac_reg, 2750000, 2750000);
+		ret = regulator_set_voltage(tve->dac_reg, 2750000, 2750000);
+		if (ret)
+			return ret;
 		ret = regulator_enable(tve->dac_reg);
 		if (ret)
 			return ret;
diff --git a/drivers/gpu/drm/imx/ipuv3-crtc.c b/drivers/gpu/drm/imx/ipuv3-crtc.c
index fc040417e1e8..08e188bc10fc 100644
--- a/drivers/gpu/drm/imx/ipuv3-crtc.c
+++ b/drivers/gpu/drm/imx/ipuv3-crtc.c
@@ -18,12 +18,12 @@
 #include <linux/device.h>
 #include <linux/platform_device.h>
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <linux/fb.h>
 #include <linux/clk.h>
 #include <linux/errno.h>
-#include <linux/reservation.h>
-#include <linux/dma-buf.h>
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 
@@ -33,23 +33,6 @@
 
 #define DRIVER_DESC		"i.MX IPUv3 Graphics"
 
-enum ipu_flip_status {
-	IPU_FLIP_NONE,
-	IPU_FLIP_PENDING,
-	IPU_FLIP_SUBMITTED,
-};
-
-struct ipu_flip_work {
-	struct work_struct		unref_work;
-	struct drm_gem_object		*bo;
-	struct drm_pending_vblank_event *page_flip_event;
-	struct work_struct		fence_work;
-	struct ipu_crtc			*crtc;
-	struct fence			*excl;
-	unsigned			shared_count;
-	struct fence			**shared;
-};
-
 struct ipu_crtc {
 	struct device		*dev;
 	struct drm_crtc		base;
@@ -60,201 +43,166 @@ struct ipu_crtc {
 
 	struct ipu_dc		*dc;
 	struct ipu_di		*di;
-	int			enabled;
-	enum ipu_flip_status	flip_state;
-	struct workqueue_struct *flip_queue;
-	struct ipu_flip_work	*flip_work;
 	int			irq;
-	u32			bus_format;
-	u32			bus_flags;
-	int			di_hsync_pin;
-	int			di_vsync_pin;
 };
 
-#define to_ipu_crtc(x) container_of(x, struct ipu_crtc, base)
+static inline struct ipu_crtc *to_ipu_crtc(struct drm_crtc *crtc)
+{
+	return container_of(crtc, struct ipu_crtc, base);
+}
 
-static void ipu_fb_enable(struct ipu_crtc *ipu_crtc)
+static void ipu_crtc_enable(struct drm_crtc *crtc)
 {
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
 	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
 
-	if (ipu_crtc->enabled)
-		return;
-
 	ipu_dc_enable(ipu);
-	ipu_plane_enable(ipu_crtc->plane[0]);
-	/* Start DC channel and DI after IDMAC */
 	ipu_dc_enable_channel(ipu_crtc->dc);
 	ipu_di_enable(ipu_crtc->di);
-	drm_crtc_vblank_on(&ipu_crtc->base);
-
-	ipu_crtc->enabled = 1;
 }
 
-static void ipu_fb_disable(struct ipu_crtc *ipu_crtc)
+static void ipu_crtc_disable(struct drm_crtc *crtc)
 {
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
 	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
 
-	if (!ipu_crtc->enabled)
-		return;
-
-	/* Stop DC channel and DI before IDMAC */
 	ipu_dc_disable_channel(ipu_crtc->dc);
 	ipu_di_disable(ipu_crtc->di);
-	ipu_plane_disable(ipu_crtc->plane[0]);
 	ipu_dc_disable(ipu);
-	drm_crtc_vblank_off(&ipu_crtc->base);
 
-	ipu_crtc->enabled = 0;
+	spin_lock_irq(&crtc->dev->event_lock);
+	if (crtc->state->event) {
+		drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+	spin_unlock_irq(&crtc->dev->event_lock);
 }
 
-static void ipu_crtc_dpms(struct drm_crtc *crtc, int mode)
+static void imx_drm_crtc_reset(struct drm_crtc *crtc)
 {
-	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+	struct imx_crtc_state *state;
 
-	dev_dbg(ipu_crtc->dev, "%s mode: %d\n", __func__, mode);
-
-	switch (mode) {
-	case DRM_MODE_DPMS_ON:
-		ipu_fb_enable(ipu_crtc);
-		break;
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-	case DRM_MODE_DPMS_OFF:
-		ipu_fb_disable(ipu_crtc);
-		break;
+	if (crtc->state) {
+		if (crtc->state->mode_blob)
+			drm_property_unreference_blob(crtc->state->mode_blob);
+
+		state = to_imx_crtc_state(crtc->state);
+		memset(state, 0, sizeof(*state));
+	} else {
+		state = kzalloc(sizeof(*state), GFP_KERNEL);
+		if (!state)
+			return;
+		crtc->state = &state->base;
 	}
+
+	state->base.crtc = crtc;
 }
 
-static void ipu_flip_unref_work_func(struct work_struct *__work)
+static struct drm_crtc_state *imx_drm_crtc_duplicate_state(struct drm_crtc *crtc)
 {
-	struct ipu_flip_work *work =
-			container_of(__work, struct ipu_flip_work, unref_work);
+	struct imx_crtc_state *state;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
 
-	drm_gem_object_unreference_unlocked(work->bo);
-	kfree(work);
+	__drm_atomic_helper_crtc_duplicate_state(crtc, &state->base);
+
+	WARN_ON(state->base.crtc != crtc);
+	state->base.crtc = crtc;
+
+	return &state->base;
 }
 
-static void ipu_flip_fence_work_func(struct work_struct *__work)
+static void imx_drm_crtc_destroy_state(struct drm_crtc *crtc,
+				       struct drm_crtc_state *state)
 {
-	struct ipu_flip_work *work =
-			container_of(__work, struct ipu_flip_work, fence_work);
-	int i;
-
-	/* wait for all fences attached to the FB obj to signal */
-	if (work->excl) {
-		fence_wait(work->excl, false);
-		fence_put(work->excl);
-	}
-	for (i = 0; i < work->shared_count; i++) {
-		fence_wait(work->shared[i], false);
-		fence_put(work->shared[i]);
-	}
+	__drm_atomic_helper_crtc_destroy_state(state);
+	kfree(to_imx_crtc_state(state));
+}
+
+static const struct drm_crtc_funcs ipu_crtc_funcs = {
+	.set_config = drm_atomic_helper_set_config,
+	.destroy = drm_crtc_cleanup,
+	.page_flip = drm_atomic_helper_page_flip,
+	.reset = imx_drm_crtc_reset,
+	.atomic_duplicate_state = imx_drm_crtc_duplicate_state,
+	.atomic_destroy_state = imx_drm_crtc_destroy_state,
+};
 
-	work->crtc->flip_state = IPU_FLIP_SUBMITTED;
+static irqreturn_t ipu_irq_handler(int irq, void *dev_id)
+{
+	struct ipu_crtc *ipu_crtc = dev_id;
+
+	imx_drm_handle_vblank(ipu_crtc->imx_crtc);
+
+	return IRQ_HANDLED;
 }
 
-static int ipu_page_flip(struct drm_crtc *crtc,
-		struct drm_framebuffer *fb,
-		struct drm_pending_vblank_event *event,
-		uint32_t page_flip_flags)
+static bool ipu_crtc_mode_fixup(struct drm_crtc *crtc,
+				  const struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
 {
-	struct drm_gem_cma_object *cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
 	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
-	struct ipu_flip_work *flip_work;
+	struct videomode vm;
 	int ret;
 
-	if (ipu_crtc->flip_state != IPU_FLIP_NONE)
-		return -EBUSY;
-
-	ret = imx_drm_crtc_vblank_get(ipu_crtc->imx_crtc);
-	if (ret) {
-		dev_dbg(ipu_crtc->dev, "failed to acquire vblank counter\n");
-		list_del(&event->base.link);
-
-		return ret;
-	}
+	drm_display_mode_to_videomode(adjusted_mode, &vm);
 
-	flip_work = kzalloc(sizeof *flip_work, GFP_KERNEL);
-	if (!flip_work) {
-		ret = -ENOMEM;
-		goto put_vblank;
-	}
-	INIT_WORK(&flip_work->unref_work, ipu_flip_unref_work_func);
-	flip_work->page_flip_event = event;
+	ret = ipu_di_adjust_videomode(ipu_crtc->di, &vm);
+	if (ret)
+		return false;
 
-	/* get BO backing the old framebuffer and take a reference */
-	flip_work->bo = &drm_fb_cma_get_gem_obj(crtc->primary->fb, 0)->base;
-	drm_gem_object_reference(flip_work->bo);
+	if ((vm.vsync_len == 0) || (vm.hsync_len == 0))
+		return false;
 
-	ipu_crtc->flip_work = flip_work;
-	/*
-	 * If the object has a DMABUF attached, we need to wait on its fences
-	 * if there are any.
-	 */
-	if (cma_obj->base.dma_buf) {
-		INIT_WORK(&flip_work->fence_work, ipu_flip_fence_work_func);
-		flip_work->crtc = ipu_crtc;
+	drm_display_mode_from_videomode(&vm, adjusted_mode);
 
-		ret = reservation_object_get_fences_rcu(
-				cma_obj->base.dma_buf->resv, &flip_work->excl,
-				&flip_work->shared_count, &flip_work->shared);
+	return true;
+}
 
-		if (unlikely(ret)) {
-			DRM_ERROR("failed to get fences for buffer\n");
-			goto free_flip_work;
-		}
+static int ipu_crtc_atomic_check(struct drm_crtc *crtc,
+				 struct drm_crtc_state *state)
+{
+	u32 primary_plane_mask = 1 << drm_plane_index(crtc->primary);
 
-		/* No need to queue the worker if the are no fences */
-		if (!flip_work->excl && !flip_work->shared_count) {
-			ipu_crtc->flip_state = IPU_FLIP_SUBMITTED;
-		} else {
-			ipu_crtc->flip_state = IPU_FLIP_PENDING;
-			queue_work(ipu_crtc->flip_queue,
-				   &flip_work->fence_work);
-		}
-	} else {
-		ipu_crtc->flip_state = IPU_FLIP_SUBMITTED;
-	}
+	if (state->active && (primary_plane_mask & state->plane_mask) == 0)
+		return -EINVAL;
 
 	return 0;
-
-free_flip_work:
-	drm_gem_object_unreference_unlocked(flip_work->bo);
-	kfree(flip_work);
-	ipu_crtc->flip_work = NULL;
-put_vblank:
-	imx_drm_crtc_vblank_put(ipu_crtc->imx_crtc);
-
-	return ret;
 }
 
-static const struct drm_crtc_funcs ipu_crtc_funcs = {
-	.set_config = drm_crtc_helper_set_config,
-	.destroy = drm_crtc_cleanup,
-	.page_flip = ipu_page_flip,
-};
+static void ipu_crtc_atomic_begin(struct drm_crtc *crtc,
+				  struct drm_crtc_state *old_crtc_state)
+{
+	spin_lock_irq(&crtc->dev->event_lock);
+	if (crtc->state->event) {
+		WARN_ON(drm_crtc_vblank_get(crtc));
+		drm_crtc_arm_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+	spin_unlock_irq(&crtc->dev->event_lock);
+}
 
-static int ipu_crtc_mode_set(struct drm_crtc *crtc,
-			       struct drm_display_mode *orig_mode,
-			       struct drm_display_mode *mode,
-			       int x, int y,
-			       struct drm_framebuffer *old_fb)
+static void ipu_crtc_mode_set_nofb(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_encoder *encoder;
 	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc->state);
 	struct ipu_di_signal_cfg sig_cfg = {};
 	unsigned long encoder_types = 0;
-	int ret;
 
 	dev_dbg(ipu_crtc->dev, "%s: mode->hdisplay: %d\n", __func__,
 			mode->hdisplay);
 	dev_dbg(ipu_crtc->dev, "%s: mode->vdisplay: %d\n", __func__,
 			mode->vdisplay);
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head)
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
 		if (encoder->crtc == crtc)
 			encoder_types |= BIT(encoder->encoder_type);
+	}
 
 	dev_dbg(ipu_crtc->dev, "%s: attached to encoder types 0x%lx\n",
 		__func__, encoder_types);
@@ -272,114 +220,30 @@ static int ipu_crtc_mode_set(struct drm_crtc *crtc,
 	else
 		sig_cfg.clkflags = 0;
 
-	sig_cfg.enable_pol = !(ipu_crtc->bus_flags & DRM_BUS_FLAG_DE_LOW);
+	sig_cfg.enable_pol = !(imx_crtc_state->bus_flags & DRM_BUS_FLAG_DE_LOW);
 	/* Default to driving pixel data on negative clock edges */
-	sig_cfg.clk_pol = !!(ipu_crtc->bus_flags &
+	sig_cfg.clk_pol = !!(imx_crtc_state->bus_flags &
 			     DRM_BUS_FLAG_PIXDATA_POSEDGE);
-	sig_cfg.bus_format = ipu_crtc->bus_format;
+	sig_cfg.bus_format = imx_crtc_state->bus_format;
 	sig_cfg.v_to_h_sync = 0;
-	sig_cfg.hsync_pin = ipu_crtc->di_hsync_pin;
-	sig_cfg.vsync_pin = ipu_crtc->di_vsync_pin;
+	sig_cfg.hsync_pin = imx_crtc_state->di_hsync_pin;
+	sig_cfg.vsync_pin = imx_crtc_state->di_vsync_pin;
 
 	drm_display_mode_to_videomode(mode, &sig_cfg.mode);
 
-	ret = ipu_dc_init_sync(ipu_crtc->dc, ipu_crtc->di,
-			       mode->flags & DRM_MODE_FLAG_INTERLACE,
-			       ipu_crtc->bus_format, mode->hdisplay);
-	if (ret) {
-		dev_err(ipu_crtc->dev,
-				"initializing display controller failed with %d\n",
-				ret);
-		return ret;
-	}
-
-	ret = ipu_di_init_sync_panel(ipu_crtc->di, &sig_cfg);
-	if (ret) {
-		dev_err(ipu_crtc->dev,
-				"initializing panel failed with %d\n", ret);
-		return ret;
-	}
-
-	return ipu_plane_mode_set(ipu_crtc->plane[0], crtc, mode,
-				  crtc->primary->fb,
-				  0, 0, mode->hdisplay, mode->vdisplay,
-				  x, y, mode->hdisplay, mode->vdisplay,
-				  mode->flags & DRM_MODE_FLAG_INTERLACE);
-}
-
-static void ipu_crtc_handle_pageflip(struct ipu_crtc *ipu_crtc)
-{
-	unsigned long flags;
-	struct drm_device *drm = ipu_crtc->base.dev;
-	struct ipu_flip_work *work = ipu_crtc->flip_work;
-
-	spin_lock_irqsave(&drm->event_lock, flags);
-	if (work->page_flip_event)
-		drm_crtc_send_vblank_event(&ipu_crtc->base,
-					   work->page_flip_event);
-	imx_drm_crtc_vblank_put(ipu_crtc->imx_crtc);
-	spin_unlock_irqrestore(&drm->event_lock, flags);
-}
-
-static irqreturn_t ipu_irq_handler(int irq, void *dev_id)
-{
-	struct ipu_crtc *ipu_crtc = dev_id;
-
-	imx_drm_handle_vblank(ipu_crtc->imx_crtc);
-
-	if (ipu_crtc->flip_state == IPU_FLIP_SUBMITTED) {
-		struct ipu_plane *plane = ipu_crtc->plane[0];
-
-		ipu_plane_set_base(plane, ipu_crtc->base.primary->fb,
-				   plane->x, plane->y);
-		ipu_crtc_handle_pageflip(ipu_crtc);
-		queue_work(ipu_crtc->flip_queue,
-			   &ipu_crtc->flip_work->unref_work);
-		ipu_crtc->flip_state = IPU_FLIP_NONE;
-	}
-
-	return IRQ_HANDLED;
-}
-
-static bool ipu_crtc_mode_fixup(struct drm_crtc *crtc,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode)
-{
-	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
-	struct videomode vm;
-	int ret;
-
-	drm_display_mode_to_videomode(adjusted_mode, &vm);
-
-	ret = ipu_di_adjust_videomode(ipu_crtc->di, &vm);
-	if (ret)
-		return false;
-
-	drm_display_mode_from_videomode(&vm, adjusted_mode);
-
-	return true;
-}
-
-static void ipu_crtc_prepare(struct drm_crtc *crtc)
-{
-	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
-
-	ipu_fb_disable(ipu_crtc);
-}
-
-static void ipu_crtc_commit(struct drm_crtc *crtc)
-{
-	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
-
-	ipu_fb_enable(ipu_crtc);
+	ipu_dc_init_sync(ipu_crtc->dc, ipu_crtc->di,
+			 mode->flags & DRM_MODE_FLAG_INTERLACE,
+			 imx_crtc_state->bus_format, mode->hdisplay);
+	ipu_di_init_sync_panel(ipu_crtc->di, &sig_cfg);
 }
 
 static const struct drm_crtc_helper_funcs ipu_helper_funcs = {
-	.dpms = ipu_crtc_dpms,
 	.mode_fixup = ipu_crtc_mode_fixup,
-	.mode_set = ipu_crtc_mode_set,
-	.prepare = ipu_crtc_prepare,
-	.commit = ipu_crtc_commit,
+	.mode_set_nofb = ipu_crtc_mode_set_nofb,
+	.atomic_check = ipu_crtc_atomic_check,
+	.atomic_begin = ipu_crtc_atomic_begin,
+	.disable = ipu_crtc_disable,
+	.enable = ipu_crtc_enable,
 };
 
 static int ipu_enable_vblank(struct drm_crtc *crtc)
@@ -398,23 +262,9 @@ static void ipu_disable_vblank(struct drm_crtc *crtc)
 	disable_irq_nosync(ipu_crtc->irq);
 }
 
-static int ipu_set_interface_pix_fmt(struct drm_crtc *crtc,
-		u32 bus_format, int hsync_pin, int vsync_pin, u32 bus_flags)
-{
-	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
-
-	ipu_crtc->bus_format = bus_format;
-	ipu_crtc->bus_flags = bus_flags;
-	ipu_crtc->di_hsync_pin = hsync_pin;
-	ipu_crtc->di_vsync_pin = vsync_pin;
-
-	return 0;
-}
-
 static const struct imx_drm_crtc_helper_funcs ipu_crtc_helper_funcs = {
 	.enable_vblank = ipu_enable_vblank,
 	.disable_vblank = ipu_disable_vblank,
-	.set_interface_pix_fmt = ipu_set_interface_pix_fmt,
 	.crtc_funcs = &ipu_crtc_funcs,
 	.crtc_helper_funcs = &ipu_helper_funcs,
 };
@@ -496,8 +346,16 @@ static int ipu_crtc_init(struct ipu_crtc *ipu_crtc,
 						IPU_DP_FLOW_SYNC_FG,
 						drm_crtc_mask(&ipu_crtc->base),
 						DRM_PLANE_TYPE_OVERLAY);
-		if (IS_ERR(ipu_crtc->plane[1]))
+		if (IS_ERR(ipu_crtc->plane[1])) {
 			ipu_crtc->plane[1] = NULL;
+		} else {
+			ret = ipu_plane_get_resources(ipu_crtc->plane[1]);
+			if (ret) {
+				dev_err(ipu_crtc->dev, "getting plane 1 "
+					"resources failed with %d.\n", ret);
+				goto err_put_plane0_res;
+			}
+		}
 	}
 
 	ipu_crtc->irq = ipu_plane_irq(ipu_crtc->plane[0]);
@@ -505,16 +363,17 @@ static int ipu_crtc_init(struct ipu_crtc *ipu_crtc,
 			"imx_drm", ipu_crtc);
 	if (ret < 0) {
 		dev_err(ipu_crtc->dev, "irq request failed with %d.\n", ret);
-		goto err_put_plane_res;
+		goto err_put_plane1_res;
 	}
 	/* Only enable IRQ when we actually need it to trigger work. */
 	disable_irq(ipu_crtc->irq);
 
-	ipu_crtc->flip_queue = create_singlethread_workqueue("ipu-crtc-flip");
-
 	return 0;
 
-err_put_plane_res:
+err_put_plane1_res:
+	if (ipu_crtc->plane[1])
+		ipu_plane_put_resources(ipu_crtc->plane[1]);
+err_put_plane0_res:
 	ipu_plane_put_resources(ipu_crtc->plane[0]);
 err_remove_crtc:
 	imx_drm_remove_crtc(ipu_crtc->imx_crtc);
@@ -553,9 +412,10 @@ static void ipu_drm_unbind(struct device *dev, struct device *master,
 
 	imx_drm_remove_crtc(ipu_crtc->imx_crtc);
 
-	destroy_workqueue(ipu_crtc->flip_queue);
-	ipu_plane_put_resources(ipu_crtc->plane[0]);
 	ipu_put_resources(ipu_crtc);
+	if (ipu_crtc->plane[1])
+		ipu_plane_put_resources(ipu_crtc->plane[1]);
+	ipu_plane_put_resources(ipu_crtc->plane[0]);
 }
 
 static const struct component_ops ipu_crtc_ops = {
diff --git a/drivers/gpu/drm/imx/ipuv3-plane.c b/drivers/gpu/drm/imx/ipuv3-plane.c
index a4bb44118d33..4ad67d015ec7 100644
--- a/drivers/gpu/drm/imx/ipuv3-plane.c
+++ b/drivers/gpu/drm/imx/ipuv3-plane.c
@@ -14,13 +14,19 @@
  */
 
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include "video/imx-ipu-v3.h"
 #include "ipuv3-plane.h"
 
-#define to_ipu_plane(x)	container_of(x, struct ipu_plane, base)
+static inline struct ipu_plane *to_ipu_plane(struct drm_plane *p)
+{
+	return container_of(p, struct ipu_plane, base);
+}
 
 static const uint32_t ipu_plane_formats[] = {
 	DRM_FORMAT_ARGB1555,
@@ -53,62 +59,67 @@ int ipu_plane_irq(struct ipu_plane *ipu_plane)
 				     IPU_IRQ_EOF);
 }
 
-static int calc_vref(struct drm_display_mode *mode)
+static inline unsigned long
+drm_plane_state_to_eba(struct drm_plane_state *state)
 {
-	unsigned long htotal, vtotal;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_gem_cma_object *cma_obj;
 
-	htotal = mode->htotal;
-	vtotal = mode->vtotal;
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
+	BUG_ON(!cma_obj);
 
-	if (!htotal || !vtotal)
-		return 60;
-
-	return DIV_ROUND_UP(mode->clock * 1000, vtotal * htotal);
+	return cma_obj->paddr + fb->offsets[0] +
+	       fb->pitches[0] * (state->src_y >> 16) +
+	       (fb->bits_per_pixel >> 3) * (state->src_x >> 16);
 }
 
-static inline int calc_bandwidth(int width, int height, unsigned int vref)
+static inline unsigned long
+drm_plane_state_to_ubo(struct drm_plane_state *state)
 {
-	return width * height * vref;
-}
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_gem_cma_object *cma_obj;
+	unsigned long eba = drm_plane_state_to_eba(state);
 
-int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb,
-		       int x, int y)
-{
-	struct drm_gem_cma_object *cma_obj[3];
-	unsigned long eba, ubo, vbo;
-	int active, i;
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 1);
+	BUG_ON(!cma_obj);
 
-	for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) {
-		cma_obj[i] = drm_fb_cma_get_gem_obj(fb, i);
-		if (!cma_obj[i]) {
-			DRM_DEBUG_KMS("plane %d entry is null.\n", i);
-			return -EFAULT;
-		}
-	}
+	return cma_obj->paddr + fb->offsets[1] +
+	       fb->pitches[1] * (state->src_y >> 16) / 2 +
+	       (state->src_x >> 16) / 2 - eba;
+}
 
-	eba = cma_obj[0]->paddr + fb->offsets[0] +
-	      fb->pitches[0] * y + (fb->bits_per_pixel >> 3) * x;
+static inline unsigned long
+drm_plane_state_to_vbo(struct drm_plane_state *state)
+{
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_gem_cma_object *cma_obj;
+	unsigned long eba = drm_plane_state_to_eba(state);
 
-	if (eba & 0x7) {
-		DRM_DEBUG_KMS("base address must be a multiple of 8.\n");
-		return -EINVAL;
-	}
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 2);
+	BUG_ON(!cma_obj);
 
-	if (fb->pitches[0] < 1 || fb->pitches[0] > 16384) {
-		DRM_DEBUG_KMS("pitches out of range.\n");
-		return -EINVAL;
-	}
+	return cma_obj->paddr + fb->offsets[2] +
+	       fb->pitches[2] * (state->src_y >> 16) / 2 +
+	       (state->src_x >> 16) / 2 - eba;
+}
 
-	if (ipu_plane->enabled && fb->pitches[0] != ipu_plane->stride[0]) {
-		DRM_DEBUG_KMS("pitches must not change while plane is enabled.\n");
-		return -EINVAL;
-	}
+static void ipu_plane_atomic_set_base(struct ipu_plane *ipu_plane,
+				      struct drm_plane_state *old_state)
+{
+	struct drm_plane *plane = &ipu_plane->base;
+	struct drm_plane_state *state = plane->state;
+	struct drm_framebuffer *fb = state->fb;
+	unsigned long eba, ubo, vbo;
+	int active;
 
-	ipu_plane->stride[0] = fb->pitches[0];
+	eba = drm_plane_state_to_eba(state);
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_YUV420:
 	case DRM_FORMAT_YVU420:
+		if (old_state->fb)
+			break;
+
 		/*
 		 * Multiplanar formats have to meet the following restrictions:
 		 * - The (up to) three plane addresses are EBA, EBA+UBO, EBA+VBO
@@ -117,59 +128,28 @@ int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb,
 		 * - Only EBA may be changed while scanout is active
 		 * - The strides of U and V planes must be identical.
 		 */
-		ubo = cma_obj[1]->paddr + fb->offsets[1] +
-		      fb->pitches[1] * y / 2 + x / 2 - eba;
-		vbo = cma_obj[2]->paddr + fb->offsets[2] +
-		      fb->pitches[2] * y / 2 + x / 2 - eba;
+		ubo = drm_plane_state_to_ubo(state);
+		vbo = drm_plane_state_to_vbo(state);
 
-		if ((ubo & 0x7) || (vbo & 0x7)) {
-			DRM_DEBUG_KMS("U/V buffer offsets must be a multiple of 8.\n");
-			return -EINVAL;
-		}
-
-		if ((ubo > 0xfffff8) || (vbo > 0xfffff8)) {
-			DRM_DEBUG_KMS("U/V buffer offsets must be positive and not larger than 0xfffff8.\n");
-			return -EINVAL;
-		}
-
-		if (ipu_plane->enabled && ((ipu_plane->u_offset != ubo) ||
-					   (ipu_plane->v_offset != vbo))) {
-			DRM_DEBUG_KMS("U/V buffer offsets must not change while plane is enabled.\n");
-			return -EINVAL;
-		}
-
-		if (fb->pitches[1] != fb->pitches[2]) {
-			DRM_DEBUG_KMS("U/V pitches must be identical.\n");
-			return -EINVAL;
-		}
-
-		if (fb->pitches[1] < 1 || fb->pitches[1] > 16384) {
-			DRM_DEBUG_KMS("U/V pitches out of range.\n");
-			return -EINVAL;
-		}
-
-		if (ipu_plane->enabled &&
-		    (ipu_plane->stride[1] != fb->pitches[1])) {
-			DRM_DEBUG_KMS("U/V pitches must not change while plane is enabled.\n");
-			return -EINVAL;
-		}
-
-		ipu_plane->u_offset = ubo;
-		ipu_plane->v_offset = vbo;
-		ipu_plane->stride[1] = fb->pitches[1];
+		if (fb->pixel_format == DRM_FORMAT_YUV420)
+			ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch,
+						      fb->pitches[1], ubo, vbo);
+		else
+			ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch,
+						      fb->pitches[1], vbo, ubo);
 
 		dev_dbg(ipu_plane->base.dev->dev,
-			"phys = %pad %pad %pad, x = %d, y = %d",
-			&cma_obj[0]->paddr, &cma_obj[1]->paddr,
-			&cma_obj[2]->paddr, x, y);
+			"phy = %lu %lu %lu, x = %d, y = %d", eba, ubo, vbo,
+			state->src_x >> 16, state->src_y >> 16);
 		break;
 	default:
-		dev_dbg(ipu_plane->base.dev->dev, "phys = %pad, x = %d, y = %d",
-			&cma_obj[0]->paddr, x, y);
+		dev_dbg(ipu_plane->base.dev->dev, "phys = %lu, x = %d, y = %d",
+			eba, state->src_x >> 16, state->src_y >> 16);
+
 		break;
 	}
 
-	if (ipu_plane->enabled) {
+	if (old_state->fb) {
 		active = ipu_idmac_get_current_buffer(ipu_plane->ipu_ch);
 		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, !active, eba);
 		ipu_idmac_select_buffer(ipu_plane->ipu_ch, !active);
@@ -177,155 +157,6 @@ int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb,
 		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 0, eba);
 		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 1, eba);
 	}
-
-	/* cache offsets for subsequent pageflips */
-	ipu_plane->x = x;
-	ipu_plane->y = y;
-
-	return 0;
-}
-
-int ipu_plane_mode_set(struct ipu_plane *ipu_plane, struct drm_crtc *crtc,
-		       struct drm_display_mode *mode,
-		       struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-		       unsigned int crtc_w, unsigned int crtc_h,
-		       uint32_t src_x, uint32_t src_y,
-		       uint32_t src_w, uint32_t src_h, bool interlaced)
-{
-	struct device *dev = ipu_plane->base.dev->dev;
-	int ret;
-
-	/* no scaling */
-	if (src_w != crtc_w || src_h != crtc_h)
-		return -EINVAL;
-
-	/* clip to crtc bounds */
-	if (crtc_x < 0) {
-		if (-crtc_x > crtc_w)
-			return -EINVAL;
-		src_x += -crtc_x;
-		src_w -= -crtc_x;
-		crtc_w -= -crtc_x;
-		crtc_x = 0;
-	}
-	if (crtc_y < 0) {
-		if (-crtc_y > crtc_h)
-			return -EINVAL;
-		src_y += -crtc_y;
-		src_h -= -crtc_y;
-		crtc_h -= -crtc_y;
-		crtc_y = 0;
-	}
-	if (crtc_x + crtc_w > mode->hdisplay) {
-		if (crtc_x > mode->hdisplay)
-			return -EINVAL;
-		crtc_w = mode->hdisplay - crtc_x;
-		src_w = crtc_w;
-	}
-	if (crtc_y + crtc_h > mode->vdisplay) {
-		if (crtc_y > mode->vdisplay)
-			return -EINVAL;
-		crtc_h = mode->vdisplay - crtc_y;
-		src_h = crtc_h;
-	}
-	/* full plane minimum width is 13 pixels */
-	if (crtc_w < 13 && (ipu_plane->dp_flow != IPU_DP_FLOW_SYNC_FG))
-		return -EINVAL;
-	if (crtc_h < 2)
-		return -EINVAL;
-
-	/*
-	 * since we cannot touch active IDMAC channels, we do not support
-	 * resizing the enabled plane or changing its format
-	 */
-	if (ipu_plane->enabled) {
-		if (src_w != ipu_plane->w || src_h != ipu_plane->h ||
-		    fb->pixel_format != ipu_plane->base.fb->pixel_format)
-			return -EINVAL;
-
-		return ipu_plane_set_base(ipu_plane, fb, src_x, src_y);
-	}
-
-	switch (ipu_plane->dp_flow) {
-	case IPU_DP_FLOW_SYNC_BG:
-		ret = ipu_dp_setup_channel(ipu_plane->dp,
-				IPUV3_COLORSPACE_RGB,
-				IPUV3_COLORSPACE_RGB);
-		if (ret) {
-			dev_err(dev,
-				"initializing display processor failed with %d\n",
-				ret);
-			return ret;
-		}
-		ipu_dp_set_global_alpha(ipu_plane->dp, true, 0, true);
-		break;
-	case IPU_DP_FLOW_SYNC_FG:
-		ipu_dp_setup_channel(ipu_plane->dp,
-				ipu_drm_fourcc_to_colorspace(fb->pixel_format),
-				IPUV3_COLORSPACE_UNKNOWN);
-		ipu_dp_set_window_pos(ipu_plane->dp, crtc_x, crtc_y);
-		/* Enable local alpha on partial plane */
-		switch (fb->pixel_format) {
-		case DRM_FORMAT_ARGB1555:
-		case DRM_FORMAT_ABGR1555:
-		case DRM_FORMAT_RGBA5551:
-		case DRM_FORMAT_BGRA5551:
-		case DRM_FORMAT_ARGB4444:
-		case DRM_FORMAT_ARGB8888:
-		case DRM_FORMAT_ABGR8888:
-		case DRM_FORMAT_RGBA8888:
-		case DRM_FORMAT_BGRA8888:
-			ipu_dp_set_global_alpha(ipu_plane->dp, false, 0, false);
-			break;
-		default:
-			break;
-		}
-	}
-
-	ret = ipu_dmfc_alloc_bandwidth(ipu_plane->dmfc,
-			calc_bandwidth(crtc_w, crtc_h,
-				       calc_vref(mode)), 64);
-	if (ret) {
-		dev_err(dev, "allocating dmfc bandwidth failed with %d\n", ret);
-		return ret;
-	}
-
-	ipu_dmfc_config_wait4eot(ipu_plane->dmfc, crtc_w);
-
-	ipu_cpmem_zero(ipu_plane->ipu_ch);
-	ipu_cpmem_set_resolution(ipu_plane->ipu_ch, src_w, src_h);
-	ret = ipu_cpmem_set_fmt(ipu_plane->ipu_ch, fb->pixel_format);
-	if (ret < 0) {
-		dev_err(dev, "unsupported pixel format 0x%08x\n",
-			fb->pixel_format);
-		return ret;
-	}
-	ipu_cpmem_set_high_priority(ipu_plane->ipu_ch);
-	ipu_idmac_set_double_buffer(ipu_plane->ipu_ch, 1);
-	ipu_cpmem_set_stride(ipu_plane->ipu_ch, fb->pitches[0]);
-
-	ret = ipu_plane_set_base(ipu_plane, fb, src_x, src_y);
-	if (ret < 0)
-		return ret;
-	if (interlaced)
-		ipu_cpmem_interlaced_scan(ipu_plane->ipu_ch, fb->pitches[0]);
-
-	if (fb->pixel_format == DRM_FORMAT_YUV420) {
-		ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch,
-					      ipu_plane->stride[1],
-					      ipu_plane->u_offset,
-					      ipu_plane->v_offset);
-	} else if (fb->pixel_format == DRM_FORMAT_YVU420) {
-		ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch,
-					      ipu_plane->stride[1],
-					      ipu_plane->v_offset,
-					      ipu_plane->u_offset);
-	}
-
-	ipu_plane->w = src_w;
-	ipu_plane->h = src_h;
-
-	return 0;
 }
 
 void ipu_plane_put_resources(struct ipu_plane *ipu_plane)
@@ -372,7 +203,7 @@ err_out:
 	return ret;
 }
 
-void ipu_plane_enable(struct ipu_plane *ipu_plane)
+static void ipu_plane_enable(struct ipu_plane *ipu_plane)
 {
 	if (ipu_plane->dp)
 		ipu_dp_enable(ipu_plane->ipu);
@@ -380,14 +211,10 @@ void ipu_plane_enable(struct ipu_plane *ipu_plane)
 	ipu_idmac_enable_channel(ipu_plane->ipu_ch);
 	if (ipu_plane->dp)
 		ipu_dp_enable_channel(ipu_plane->dp);
-
-	ipu_plane->enabled = true;
 }
 
-void ipu_plane_disable(struct ipu_plane *ipu_plane)
+static void ipu_plane_disable(struct ipu_plane *ipu_plane)
 {
-	ipu_plane->enabled = false;
-
 	ipu_idmac_wait_busy(ipu_plane->ipu_ch, 50);
 
 	if (ipu_plane->dp)
@@ -398,74 +225,225 @@ void ipu_plane_disable(struct ipu_plane *ipu_plane)
 		ipu_dp_disable(ipu_plane->ipu);
 }
 
-/*
- * drm_plane API
- */
-
-static int ipu_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
-			    struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-			    unsigned int crtc_w, unsigned int crtc_h,
-			    uint32_t src_x, uint32_t src_y,
-			    uint32_t src_w, uint32_t src_h)
+static int ipu_disable_plane(struct drm_plane *plane)
 {
 	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
-	int ret = 0;
-
-	DRM_DEBUG_KMS("plane - %p\n", plane);
-
-	if (!ipu_plane->enabled)
-		ret = ipu_plane_get_resources(ipu_plane);
-	if (ret < 0)
-		return ret;
-
-	ret = ipu_plane_mode_set(ipu_plane, crtc, &crtc->hwmode, fb,
-			crtc_x, crtc_y, crtc_w, crtc_h,
-			src_x >> 16, src_y >> 16, src_w >> 16, src_h >> 16,
-			false);
-	if (ret < 0) {
-		ipu_plane_put_resources(ipu_plane);
-		return ret;
-	}
 
-	if (crtc != plane->crtc)
-		dev_dbg(plane->dev->dev, "crtc change: %p -> %p\n",
-				plane->crtc, crtc);
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-	if (!ipu_plane->enabled)
-		ipu_plane_enable(ipu_plane);
+	ipu_plane_disable(ipu_plane);
 
 	return 0;
 }
 
-static int ipu_disable_plane(struct drm_plane *plane)
+static void ipu_plane_destroy(struct drm_plane *plane)
 {
 	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
 
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-	if (ipu_plane->enabled)
-		ipu_plane_disable(ipu_plane);
+	ipu_disable_plane(plane);
+	drm_plane_cleanup(plane);
+	kfree(ipu_plane);
+}
 
-	ipu_plane_put_resources(ipu_plane);
+static const struct drm_plane_funcs ipu_plane_funcs = {
+	.update_plane	= drm_atomic_helper_update_plane,
+	.disable_plane	= drm_atomic_helper_disable_plane,
+	.destroy	= ipu_plane_destroy,
+	.reset		= drm_atomic_helper_plane_reset,
+	.atomic_duplicate_state	= drm_atomic_helper_plane_duplicate_state,
+	.atomic_destroy_state	= drm_atomic_helper_plane_destroy_state,
+};
+
+static int ipu_plane_atomic_check(struct drm_plane *plane,
+				  struct drm_plane_state *state)
+{
+	struct drm_plane_state *old_state = plane->state;
+	struct drm_crtc_state *crtc_state;
+	struct device *dev = plane->dev->dev;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_framebuffer *old_fb = old_state->fb;
+	unsigned long eba, ubo, vbo, old_ubo, old_vbo;
+
+	/* Ok to disable */
+	if (!fb)
+		return 0;
+
+	if (!state->crtc)
+		return -EINVAL;
+
+	crtc_state =
+		drm_atomic_get_existing_crtc_state(state->state, state->crtc);
+	if (WARN_ON(!crtc_state))
+		return -EINVAL;
+
+	/* CRTC should be enabled */
+	if (!crtc_state->enable)
+		return -EINVAL;
+
+	/* no scaling */
+	if (state->src_w >> 16 != state->crtc_w ||
+	    state->src_h >> 16 != state->crtc_h)
+		return -EINVAL;
+
+	switch (plane->type) {
+	case DRM_PLANE_TYPE_PRIMARY:
+		/* full plane doesn't support partial off screen */
+		if (state->crtc_x || state->crtc_y ||
+		    state->crtc_w != crtc_state->adjusted_mode.hdisplay ||
+		    state->crtc_h != crtc_state->adjusted_mode.vdisplay)
+			return -EINVAL;
+
+		/* full plane minimum width is 13 pixels */
+		if (state->crtc_w < 13)
+			return -EINVAL;
+		break;
+	case DRM_PLANE_TYPE_OVERLAY:
+		if (state->crtc_x < 0 || state->crtc_y < 0)
+			return -EINVAL;
+
+		if (state->crtc_x + state->crtc_w >
+		    crtc_state->adjusted_mode.hdisplay)
+			return -EINVAL;
+		if (state->crtc_y + state->crtc_h >
+		    crtc_state->adjusted_mode.vdisplay)
+			return -EINVAL;
+		break;
+	default:
+		dev_warn(dev, "Unsupported plane type\n");
+		return -EINVAL;
+	}
+
+	if (state->crtc_h < 2)
+		return -EINVAL;
+
+	/*
+	 * since we cannot touch active IDMAC channels, we do not support
+	 * resizing the enabled plane or changing its format
+	 */
+	if (old_fb && (state->src_w != old_state->src_w ||
+			      state->src_h != old_state->src_h ||
+			      fb->pixel_format != old_fb->pixel_format))
+		return -EINVAL;
+
+	eba = drm_plane_state_to_eba(state);
+
+	if (eba & 0x7)
+		return -EINVAL;
+
+	if (fb->pitches[0] < 1 || fb->pitches[0] > 16384)
+		return -EINVAL;
+
+	if (old_fb && fb->pitches[0] != old_fb->pitches[0])
+		return -EINVAL;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_YUV420:
+	case DRM_FORMAT_YVU420:
+		/*
+		 * Multiplanar formats have to meet the following restrictions:
+		 * - The (up to) three plane addresses are EBA, EBA+UBO, EBA+VBO
+		 * - EBA, UBO and VBO are a multiple of 8
+		 * - UBO and VBO are unsigned and not larger than 0xfffff8
+		 * - Only EBA may be changed while scanout is active
+		 * - The strides of U and V planes must be identical.
+		 */
+		ubo = drm_plane_state_to_ubo(state);
+		vbo = drm_plane_state_to_vbo(state);
+
+		if ((ubo & 0x7) || (vbo & 0x7))
+			return -EINVAL;
+
+		if ((ubo > 0xfffff8) || (vbo > 0xfffff8))
+			return -EINVAL;
+
+		if (old_fb) {
+			old_ubo = drm_plane_state_to_ubo(old_state);
+			old_vbo = drm_plane_state_to_vbo(old_state);
+			if (ubo != old_ubo || vbo != old_vbo)
+				return -EINVAL;
+		}
+
+		if (fb->pitches[1] != fb->pitches[2])
+			return -EINVAL;
+
+		if (fb->pitches[1] < 1 || fb->pitches[1] > 16384)
+			return -EINVAL;
+
+		if (old_fb && old_fb->pitches[1] != fb->pitches[1])
+			return -EINVAL;
+	}
 
 	return 0;
 }
 
-static void ipu_plane_destroy(struct drm_plane *plane)
+static void ipu_plane_atomic_disable(struct drm_plane *plane,
+				     struct drm_plane_state *old_state)
+{
+	ipu_disable_plane(plane);
+}
+
+static void ipu_plane_atomic_update(struct drm_plane *plane,
+				    struct drm_plane_state *old_state)
 {
 	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
+	struct drm_plane_state *state = plane->state;
+	enum ipu_color_space ics;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+	if (old_state->fb) {
+		ipu_plane_atomic_set_base(ipu_plane, old_state);
+		return;
+	}
 
-	ipu_disable_plane(plane);
-	drm_plane_cleanup(plane);
-	kfree(ipu_plane);
+	switch (ipu_plane->dp_flow) {
+	case IPU_DP_FLOW_SYNC_BG:
+		ipu_dp_setup_channel(ipu_plane->dp,
+					IPUV3_COLORSPACE_RGB,
+					IPUV3_COLORSPACE_RGB);
+		ipu_dp_set_global_alpha(ipu_plane->dp, true, 0, true);
+		break;
+	case IPU_DP_FLOW_SYNC_FG:
+		ics = ipu_drm_fourcc_to_colorspace(state->fb->pixel_format);
+		ipu_dp_setup_channel(ipu_plane->dp, ics,
+					IPUV3_COLORSPACE_UNKNOWN);
+		ipu_dp_set_window_pos(ipu_plane->dp, state->crtc_x,
+					state->crtc_y);
+		/* Enable local alpha on partial plane */
+		switch (state->fb->pixel_format) {
+		case DRM_FORMAT_ARGB1555:
+		case DRM_FORMAT_ABGR1555:
+		case DRM_FORMAT_RGBA5551:
+		case DRM_FORMAT_BGRA5551:
+		case DRM_FORMAT_ARGB4444:
+		case DRM_FORMAT_ARGB8888:
+		case DRM_FORMAT_ABGR8888:
+		case DRM_FORMAT_RGBA8888:
+		case DRM_FORMAT_BGRA8888:
+			ipu_dp_set_global_alpha(ipu_plane->dp, false, 0, false);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ipu_dmfc_config_wait4eot(ipu_plane->dmfc, state->crtc_w);
+
+	ipu_cpmem_zero(ipu_plane->ipu_ch);
+	ipu_cpmem_set_resolution(ipu_plane->ipu_ch, state->src_w >> 16,
+					state->src_h >> 16);
+	ipu_cpmem_set_fmt(ipu_plane->ipu_ch, state->fb->pixel_format);
+	ipu_cpmem_set_high_priority(ipu_plane->ipu_ch);
+	ipu_idmac_set_double_buffer(ipu_plane->ipu_ch, 1);
+	ipu_cpmem_set_stride(ipu_plane->ipu_ch, state->fb->pitches[0]);
+	ipu_plane_atomic_set_base(ipu_plane, old_state);
+	ipu_plane_enable(ipu_plane);
 }
 
-static const struct drm_plane_funcs ipu_plane_funcs = {
-	.update_plane	= ipu_update_plane,
-	.disable_plane	= ipu_disable_plane,
-	.destroy	= ipu_plane_destroy,
+static const struct drm_plane_helper_funcs ipu_plane_helper_funcs = {
+	.atomic_check = ipu_plane_atomic_check,
+	.atomic_disable = ipu_plane_atomic_disable,
+	.atomic_update = ipu_plane_atomic_update,
 };
 
 struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu,
@@ -498,5 +476,7 @@ struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu,
 		return ERR_PTR(ret);
 	}
 
+	drm_plane_helper_add(&ipu_plane->base, &ipu_plane_helper_funcs);
+
 	return ipu_plane;
 }
diff --git a/drivers/gpu/drm/imx/ipuv3-plane.h b/drivers/gpu/drm/imx/ipuv3-plane.h
index 4448fd4ad4eb..338b88a74eb6 100644
--- a/drivers/gpu/drm/imx/ipuv3-plane.h
+++ b/drivers/gpu/drm/imx/ipuv3-plane.h
@@ -23,17 +23,6 @@ struct ipu_plane {
 
 	int			dma;
 	int			dp_flow;
-
-	int			x;
-	int			y;
-	int			w;
-	int			h;
-
-	unsigned int		u_offset;
-	unsigned int		v_offset;
-	unsigned int		stride[2];
-
-	bool			enabled;
 };
 
 struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu,
@@ -48,11 +37,6 @@ int ipu_plane_mode_set(struct ipu_plane *plane, struct drm_crtc *crtc,
 		       uint32_t src_x, uint32_t src_y, uint32_t src_w,
 		       uint32_t src_h, bool interlaced);
 
-void ipu_plane_enable(struct ipu_plane *plane);
-void ipu_plane_disable(struct ipu_plane *plane);
-int ipu_plane_set_base(struct ipu_plane *plane, struct drm_framebuffer *fb,
-		       int x, int y);
-
 int ipu_plane_get_resources(struct ipu_plane *plane);
 void ipu_plane_put_resources(struct ipu_plane *plane);
 
diff --git a/drivers/gpu/drm/imx/parallel-display.c b/drivers/gpu/drm/imx/parallel-display.c
index 2d1fd02cd3d6..1dad297b01fd 100644
--- a/drivers/gpu/drm/imx/parallel-display.c
+++ b/drivers/gpu/drm/imx/parallel-display.c
@@ -16,6 +16,7 @@
 #include <linux/component.h>
 #include <linux/module.h>
 #include <drm/drmP.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_panel.h>
@@ -25,9 +26,6 @@
 
 #include "imx-drm.h"
 
-#define con_to_imxpd(x) container_of(x, struct imx_parallel_display, connector)
-#define enc_to_imxpd(x) container_of(x, struct imx_parallel_display, encoder)
-
 struct imx_parallel_display {
 	struct drm_connector connector;
 	struct drm_encoder encoder;
@@ -37,8 +35,19 @@ struct imx_parallel_display {
 	u32 bus_format;
 	struct drm_display_mode mode;
 	struct drm_panel *panel;
+	struct drm_bridge *bridge;
 };
 
+static inline struct imx_parallel_display *con_to_imxpd(struct drm_connector *c)
+{
+	return container_of(c, struct imx_parallel_display, connector);
+}
+
+static inline struct imx_parallel_display *enc_to_imxpd(struct drm_encoder *e)
+{
+	return container_of(e, struct imx_parallel_display, encoder);
+}
+
 static enum drm_connector_status imx_pd_connector_detect(
 		struct drm_connector *connector, bool force)
 {
@@ -53,11 +62,7 @@ static int imx_pd_connector_get_modes(struct drm_connector *connector)
 
 	if (imxpd->panel && imxpd->panel->funcs &&
 	    imxpd->panel->funcs->get_modes) {
-		struct drm_display_info *di = &connector->display_info;
-
 		num_modes = imxpd->panel->funcs->get_modes(imxpd->panel);
-		if (!imxpd->bus_format && di->num_bus_formats)
-			imxpd->bus_format = di->bus_formats[0];
 		if (num_modes > 0)
 			return num_modes;
 	}
@@ -69,10 +74,16 @@ static int imx_pd_connector_get_modes(struct drm_connector *connector)
 
 	if (np) {
 		struct drm_display_mode *mode = drm_mode_create(connector->dev);
+		int ret;
 
 		if (!mode)
 			return -EINVAL;
-		of_get_drm_display_mode(np, &imxpd->mode, OF_USE_NATIVE_MODE);
+
+		ret = of_get_drm_display_mode(np, &imxpd->mode,
+					      OF_USE_NATIVE_MODE);
+		if (ret)
+			return ret;
+
 		drm_mode_copy(mode, &imxpd->mode);
 		mode->type |= DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
 		drm_mode_probed_add(connector, mode);
@@ -90,24 +101,7 @@ static struct drm_encoder *imx_pd_connector_best_encoder(
 	return &imxpd->encoder;
 }
 
-static void imx_pd_encoder_dpms(struct drm_encoder *encoder, int mode)
-{
-	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
-
-	if (mode != DRM_MODE_DPMS_ON)
-		drm_panel_disable(imxpd->panel);
-	else
-		drm_panel_enable(imxpd->panel);
-}
-
-static void imx_pd_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
-	imx_drm_set_bus_config(encoder, imxpd->bus_format, 2, 3,
-			       imxpd->connector.display_info.bus_flags);
-}
-
-static void imx_pd_encoder_commit(struct drm_encoder *encoder)
+static void imx_pd_encoder_enable(struct drm_encoder *encoder)
 {
 	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
 
@@ -115,12 +109,6 @@ static void imx_pd_encoder_commit(struct drm_encoder *encoder)
 	drm_panel_enable(imxpd->panel);
 }
 
-static void imx_pd_encoder_mode_set(struct drm_encoder *encoder,
-			 struct drm_display_mode *orig_mode,
-			 struct drm_display_mode *mode)
-{
-}
-
 static void imx_pd_encoder_disable(struct drm_encoder *encoder)
 {
 	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
@@ -129,11 +117,33 @@ static void imx_pd_encoder_disable(struct drm_encoder *encoder)
 	drm_panel_unprepare(imxpd->panel);
 }
 
+static int imx_pd_encoder_atomic_check(struct drm_encoder *encoder,
+				       struct drm_crtc_state *crtc_state,
+				       struct drm_connector_state *conn_state)
+{
+	struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state);
+	struct drm_display_info *di = &conn_state->connector->display_info;
+	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
+
+	imx_crtc_state->bus_flags = di->bus_flags;
+	if (!imxpd->bus_format && di->num_bus_formats)
+		imx_crtc_state->bus_format = di->bus_formats[0];
+	else
+		imx_crtc_state->bus_format = imxpd->bus_format;
+	imx_crtc_state->di_hsync_pin = 2;
+	imx_crtc_state->di_vsync_pin = 3;
+
+	return 0;
+}
+
 static const struct drm_connector_funcs imx_pd_connector_funcs = {
-	.dpms = drm_helper_connector_dpms,
+	.dpms = drm_atomic_helper_connector_dpms,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.detect = imx_pd_connector_detect,
 	.destroy = imx_drm_connector_destroy,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
 };
 
 static const struct drm_connector_helper_funcs imx_pd_connector_helper_funcs = {
@@ -146,20 +156,18 @@ static const struct drm_encoder_funcs imx_pd_encoder_funcs = {
 };
 
 static const struct drm_encoder_helper_funcs imx_pd_encoder_helper_funcs = {
-	.dpms = imx_pd_encoder_dpms,
-	.prepare = imx_pd_encoder_prepare,
-	.commit = imx_pd_encoder_commit,
-	.mode_set = imx_pd_encoder_mode_set,
+	.enable = imx_pd_encoder_enable,
 	.disable = imx_pd_encoder_disable,
+	.atomic_check = imx_pd_encoder_atomic_check,
 };
 
 static int imx_pd_register(struct drm_device *drm,
 	struct imx_parallel_display *imxpd)
 {
+	struct drm_encoder *encoder = &imxpd->encoder;
 	int ret;
 
-	ret = imx_drm_encoder_parse_of(drm, &imxpd->encoder,
-				       imxpd->dev->of_node);
+	ret = imx_drm_encoder_parse_of(drm, encoder, imxpd->dev->of_node);
 	if (ret)
 		return ret;
 
@@ -170,19 +178,33 @@ static int imx_pd_register(struct drm_device *drm,
 	 */
 	imxpd->connector.dpms = DRM_MODE_DPMS_OFF;
 
-	drm_encoder_helper_add(&imxpd->encoder, &imx_pd_encoder_helper_funcs);
-	drm_encoder_init(drm, &imxpd->encoder, &imx_pd_encoder_funcs,
+	drm_encoder_helper_add(encoder, &imx_pd_encoder_helper_funcs);
+	drm_encoder_init(drm, encoder, &imx_pd_encoder_funcs,
 			 DRM_MODE_ENCODER_NONE, NULL);
 
-	drm_connector_helper_add(&imxpd->connector,
-			&imx_pd_connector_helper_funcs);
-	drm_connector_init(drm, &imxpd->connector, &imx_pd_connector_funcs,
-			   DRM_MODE_CONNECTOR_VGA);
+	if (!imxpd->bridge) {
+		drm_connector_helper_add(&imxpd->connector,
+				&imx_pd_connector_helper_funcs);
+		drm_connector_init(drm, &imxpd->connector,
+				   &imx_pd_connector_funcs,
+				   DRM_MODE_CONNECTOR_VGA);
+	}
 
 	if (imxpd->panel)
 		drm_panel_attach(imxpd->panel, &imxpd->connector);
 
-	drm_mode_connector_attach_encoder(&imxpd->connector, &imxpd->encoder);
+	if (imxpd->bridge) {
+		imxpd->bridge->encoder = encoder;
+		encoder->bridge = imxpd->bridge;
+		ret = drm_bridge_attach(drm, imxpd->bridge);
+		if (ret < 0) {
+			dev_err(imxpd->dev, "failed to attach bridge: %d\n",
+				ret);
+			return ret;
+		}
+	} else {
+		drm_mode_connector_attach_encoder(&imxpd->connector, encoder);
+	}
 
 	return 0;
 }
@@ -195,6 +217,7 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data)
 	const u8 *edidp;
 	struct imx_parallel_display *imxpd;
 	int ret;
+	u32 bus_format = 0;
 	const char *fmt;
 
 	imxpd = devm_kzalloc(dev, sizeof(*imxpd), GFP_KERNEL);
@@ -208,14 +231,15 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data)
 	ret = of_property_read_string(np, "interface-pix-fmt", &fmt);
 	if (!ret) {
 		if (!strcmp(fmt, "rgb24"))
-			imxpd->bus_format = MEDIA_BUS_FMT_RGB888_1X24;
+			bus_format = MEDIA_BUS_FMT_RGB888_1X24;
 		else if (!strcmp(fmt, "rgb565"))
-			imxpd->bus_format = MEDIA_BUS_FMT_RGB565_1X16;
+			bus_format = MEDIA_BUS_FMT_RGB565_1X16;
 		else if (!strcmp(fmt, "bgr666"))
-			imxpd->bus_format = MEDIA_BUS_FMT_RGB666_1X18;
+			bus_format = MEDIA_BUS_FMT_RGB666_1X18;
 		else if (!strcmp(fmt, "lvds666"))
-			imxpd->bus_format = MEDIA_BUS_FMT_RGB666_1X24_CPADHI;
+			bus_format = MEDIA_BUS_FMT_RGB666_1X24_CPADHI;
 	}
+	imxpd->bus_format = bus_format;
 
 	/* port@1 is the output port */
 	ep = of_graph_get_endpoint_by_regs(np, 1, -1);
@@ -223,13 +247,30 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data)
 		struct device_node *remote;
 
 		remote = of_graph_get_remote_port_parent(ep);
+		if (!remote) {
+			dev_warn(dev, "endpoint %s not connected\n",
+				 ep->full_name);
+			of_node_put(ep);
+			return -ENODEV;
+		}
 		of_node_put(ep);
-		if (remote) {
-			imxpd->panel = of_drm_find_panel(remote);
-			of_node_put(remote);
+
+		imxpd->panel = of_drm_find_panel(remote);
+		if (imxpd->panel) {
+			dev_dbg(dev, "found panel %s\n", remote->full_name);
+		} else {
+			imxpd->bridge = of_drm_find_bridge(remote);
+			if (imxpd->bridge)
+				dev_dbg(dev, "found bridge %s\n",
+					remote->full_name);
 		}
-		if (!imxpd->panel)
+		if (!imxpd->panel && !imxpd->bridge) {
+			dev_dbg(dev, "waiting for panel or bridge %s\n",
+				remote->full_name);
+			of_node_put(remote);
 			return -EPROBE_DEFER;
+		}
+		of_node_put(remote);
 	}
 
 	imxpd->dev = dev;
diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c
index ba812ef2c9d1..334562d06731 100644
--- a/drivers/gpu/drm/mediatek/mtk_hdmi.c
+++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c
@@ -1535,7 +1535,7 @@ static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi,
  * HDMI audio codec callbacks
  */
 
-static int mtk_hdmi_audio_hw_params(struct device *dev,
+static int mtk_hdmi_audio_hw_params(struct device *dev, void *data,
 				    struct hdmi_codec_daifmt *daifmt,
 				    struct hdmi_codec_params *params)
 {
@@ -1604,7 +1604,7 @@ static int mtk_hdmi_audio_hw_params(struct device *dev,
 	return 0;
 }
 
-static int mtk_hdmi_audio_startup(struct device *dev)
+static int mtk_hdmi_audio_startup(struct device *dev, void *data)
 {
 	struct mtk_hdmi *hdmi = dev_get_drvdata(dev);
 
@@ -1615,7 +1615,7 @@ static int mtk_hdmi_audio_startup(struct device *dev)
 	return 0;
 }
 
-static void mtk_hdmi_audio_shutdown(struct device *dev)
+static void mtk_hdmi_audio_shutdown(struct device *dev, void *data)
 {
 	struct mtk_hdmi *hdmi = dev_get_drvdata(dev);
 
@@ -1624,7 +1624,7 @@ static void mtk_hdmi_audio_shutdown(struct device *dev)
 	mtk_hdmi_audio_disable(hdmi);
 }
 
-int mtk_hdmi_audio_digital_mute(struct device *dev, bool enable)
+int mtk_hdmi_audio_digital_mute(struct device *dev, void *data, bool enable)
 {
 	struct mtk_hdmi *hdmi = dev_get_drvdata(dev);
 
@@ -1638,7 +1638,7 @@ int mtk_hdmi_audio_digital_mute(struct device *dev, bool enable)
 	return 0;
 }
 
-static int mtk_hdmi_audio_get_eld(struct device *dev, uint8_t *buf, size_t len)
+static int mtk_hdmi_audio_get_eld(struct device *dev, void *data, uint8_t *buf, size_t len)
 {
 	struct mtk_hdmi *hdmi = dev_get_drvdata(dev);
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 9d5083d0f1ee..68268e55d595 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -186,17 +186,6 @@ static void mgag200_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_r
 {
 }
 
-static int mgag200_bo_move(struct ttm_buffer_object *bo,
-		       bool evict, bool interruptible,
-		       bool no_wait_gpu,
-		       struct ttm_mem_reg *new_mem)
-{
-	int r;
-	r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
-	return r;
-}
-
-
 static void mgag200_ttm_backend_destroy(struct ttm_tt *tt)
 {
 	ttm_tt_fini(tt);
@@ -241,7 +230,7 @@ struct ttm_bo_driver mgag200_bo_driver = {
 	.ttm_tt_unpopulate = mgag200_ttm_tt_unpopulate,
 	.init_mem_type = mgag200_bo_init_mem_type,
 	.evict_flags = mgag200_bo_evict_flags,
-	.move = mgag200_bo_move,
+	.move = NULL,
 	.verify_access = mgag200_bo_verify_access,
 	.io_mem_reserve = &mgag200_ttm_io_mem_reserve,
 	.io_mem_free = &mgag200_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
index fbe304ee6c80..2aec27dbb5bb 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
@@ -408,7 +408,7 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 	}
 
 	adreno_gpu->memptrs = msm_gem_vaddr(adreno_gpu->memptrs_bo);
-	if (!adreno_gpu->memptrs) {
+	if (IS_ERR(adreno_gpu->memptrs)) {
 		dev_err(drm->dev, "could not vmap memptrs\n");
 		return -ENOMEM;
 	}
diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c
index 1a061e3e8b9e..a9223bea871b 100644
--- a/drivers/gpu/drm/msm/msm_fbdev.c
+++ b/drivers/gpu/drm/msm/msm_fbdev.c
@@ -159,6 +159,10 @@ static int msm_fbdev_create(struct drm_fb_helper *helper,
 	dev->mode_config.fb_base = paddr;
 
 	fbi->screen_base = msm_gem_vaddr_locked(fbdev->bo);
+	if (IS_ERR(fbi->screen_base)) {
+		ret = PTR_ERR(fbi->screen_base);
+		goto fail_unlock;
+	}
 	fbi->screen_size = fbdev->bo->size;
 	fbi->fix.smem_start = paddr;
 	fbi->fix.smem_len = fbdev->bo->size;
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index 7daf4054dd2b..69836f5685b1 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -398,6 +398,8 @@ void *msm_gem_vaddr_locked(struct drm_gem_object *obj)
 			return ERR_CAST(pages);
 		msm_obj->vaddr = vmap(pages, obj->size >> PAGE_SHIFT,
 				VM_MAP, pgprot_writecombine(PAGE_KERNEL));
+		if (msm_obj->vaddr == NULL)
+			return ERR_PTR(-ENOMEM);
 	}
 	return msm_obj->vaddr;
 }
diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c
index b89ca5174863..eb4bb8b2f3a5 100644
--- a/drivers/gpu/drm/msm/msm_gem_submit.c
+++ b/drivers/gpu/drm/msm/msm_gem_submit.c
@@ -40,12 +40,14 @@ static struct msm_gem_submit *submit_create(struct drm_device *dev,
 
 	submit->dev = dev;
 	submit->gpu = gpu;
+	submit->fence = NULL;
 	submit->pid = get_pid(task_pid(current));
 
 	/* initially, until copy_from_user() and bo lookup succeeds: */
 	submit->nr_bos = 0;
 	submit->nr_cmds = 0;
 
+	INIT_LIST_HEAD(&submit->node);
 	INIT_LIST_HEAD(&submit->bo_list);
 	ww_acquire_init(&submit->ticket, &reservation_ww_class);
 
@@ -75,6 +77,11 @@ static int submit_lookup_objects(struct msm_gem_submit *submit,
 		void __user *userptr =
 			u64_to_user_ptr(args->bos + (i * sizeof(submit_bo)));
 
+		/* make sure we don't have garbage flags, in case we hit
+		 * error path before flags is initialized:
+		 */
+		submit->bos[i].flags = 0;
+
 		ret = copy_from_user(&submit_bo, userptr, sizeof(submit_bo));
 		if (ret) {
 			ret = -EFAULT;
diff --git a/drivers/gpu/drm/msm/msm_rd.c b/drivers/gpu/drm/msm/msm_rd.c
index b48f73ac6389..0857710c2ff2 100644
--- a/drivers/gpu/drm/msm/msm_rd.c
+++ b/drivers/gpu/drm/msm/msm_rd.c
@@ -312,6 +312,9 @@ void msm_rd_dump_submit(struct msm_gem_submit *submit)
 		struct msm_gem_object *obj = submit->bos[idx].obj;
 		const char *buf = msm_gem_vaddr_locked(&obj->base);
 
+		if (IS_ERR(buf))
+			continue;
+
 		buf += iova - submit->bos[idx].iova;
 
 		rd_write_section(rd, RD_GPUADDR,
diff --git a/drivers/gpu/drm/msm/msm_ringbuffer.c b/drivers/gpu/drm/msm/msm_ringbuffer.c
index 1f14b908b221..42f5359cf988 100644
--- a/drivers/gpu/drm/msm/msm_ringbuffer.c
+++ b/drivers/gpu/drm/msm/msm_ringbuffer.c
@@ -40,6 +40,10 @@ struct msm_ringbuffer *msm_ringbuffer_new(struct msm_gpu *gpu, int size)
 	}
 
 	ring->start = msm_gem_vaddr_locked(ring->bo);
+	if (IS_ERR(ring->start)) {
+		ret = PTR_ERR(ring->start);
+		goto fail;
+	}
 	ring->end   = ring->start + (size / 4);
 	ring->cur   = ring->start;
 
diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
index 331620a52afa..287a7d6fa480 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h
@@ -29,6 +29,7 @@ struct nv_device_info_v0 {
 #define NV_DEVICE_INFO_V0_FERMI                                            0x07
 #define NV_DEVICE_INFO_V0_KEPLER                                           0x08
 #define NV_DEVICE_INFO_V0_MAXWELL                                          0x09
+#define NV_DEVICE_INFO_V0_PASCAL                                           0x0a
 	__u8  family;
 	__u8  pad06[2];
 	__u64 ram_size;
diff --git a/drivers/gpu/drm/nouveau/include/nvif/class.h b/drivers/gpu/drm/nouveau/include/nvif/class.h
index 982aad8fa645..e6e9537537cf 100644
--- a/drivers/gpu/drm/nouveau/include/nvif/class.h
+++ b/drivers/gpu/drm/nouveau/include/nvif/class.h
@@ -39,6 +39,7 @@
 #define KEPLER_CHANNEL_GPFIFO_A                       /* cla06f.h */ 0x0000a06f
 #define KEPLER_CHANNEL_GPFIFO_B                       /* cla06f.h */ 0x0000a16f
 #define MAXWELL_CHANNEL_GPFIFO_A                      /* cla06f.h */ 0x0000b06f
+#define PASCAL_CHANNEL_GPFIFO_A                       /* cla06f.h */ 0x0000c06f
 
 #define NV50_DISP                                     /* cl5070.h */ 0x00005070
 #define G82_DISP                                      /* cl5070.h */ 0x00008270
@@ -50,6 +51,8 @@
 #define GK110_DISP                                    /* cl5070.h */ 0x00009270
 #define GM107_DISP                                    /* cl5070.h */ 0x00009470
 #define GM200_DISP                                    /* cl5070.h */ 0x00009570
+#define GP100_DISP                                    /* cl5070.h */ 0x00009770
+#define GP104_DISP                                    /* cl5070.h */ 0x00009870
 
 #define NV31_MPEG                                                    0x00003174
 #define G82_MPEG                                                     0x00008274
@@ -86,6 +89,8 @@
 #define GK110_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000927d
 #define GM107_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000947d
 #define GM200_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000957d
+#define GP100_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000977d
+#define GP104_DISP_CORE_CHANNEL_DMA                   /* cl507d.h */ 0x0000987d
 
 #define NV50_DISP_OVERLAY_CHANNEL_DMA                 /* cl507e.h */ 0x0000507e
 #define G82_DISP_OVERLAY_CHANNEL_DMA                  /* cl507e.h */ 0x0000827e
@@ -105,6 +110,8 @@
 #define MAXWELL_A                                     /* cl9097.h */ 0x0000b097
 #define MAXWELL_B                                     /* cl9097.h */ 0x0000b197
 
+#define PASCAL_A                                      /* cl9097.h */ 0x0000c097
+
 #define NV74_BSP                                                     0x000074b0
 
 #define GT212_MSVLD                                                  0x000085b1
@@ -128,6 +135,8 @@
 #define FERMI_DMA                                                    0x000090b5
 #define KEPLER_DMA_COPY_A                                            0x0000a0b5
 #define MAXWELL_DMA_COPY_A                                           0x0000b0b5
+#define PASCAL_DMA_COPY_A                                            0x0000c0b5
+#define PASCAL_DMA_COPY_B                                            0x0000c1b5
 
 #define FERMI_DECOMPRESS                                             0x000090b8
 
@@ -137,6 +146,7 @@
 #define KEPLER_COMPUTE_B                                             0x0000a1c0
 #define MAXWELL_COMPUTE_A                                            0x0000b0c0
 #define MAXWELL_COMPUTE_B                                            0x0000b1c0
+#define PASCAL_COMPUTE_A                                             0x0000c0c0
 
 #define NV74_CIPHER                                                  0x000074c1
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
index c612dc1f1eb4..7ea8aa7ca408 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h
@@ -16,9 +16,9 @@ enum nvkm_devidx {
 	NVKM_SUBDEV_MC,
 	NVKM_SUBDEV_BUS,
 	NVKM_SUBDEV_TIMER,
+	NVKM_SUBDEV_INSTMEM,
 	NVKM_SUBDEV_FB,
 	NVKM_SUBDEV_LTC,
-	NVKM_SUBDEV_INSTMEM,
 	NVKM_SUBDEV_MMU,
 	NVKM_SUBDEV_BAR,
 	NVKM_SUBDEV_PMU,
@@ -33,7 +33,10 @@ enum nvkm_devidx {
 	NVKM_ENGINE_CE0,
 	NVKM_ENGINE_CE1,
 	NVKM_ENGINE_CE2,
-	NVKM_ENGINE_CE_LAST = NVKM_ENGINE_CE2,
+	NVKM_ENGINE_CE3,
+	NVKM_ENGINE_CE4,
+	NVKM_ENGINE_CE5,
+	NVKM_ENGINE_CE_LAST = NVKM_ENGINE_CE5,
 
 	NVKM_ENGINE_CIPHER,
 	NVKM_ENGINE_DISP,
@@ -50,7 +53,8 @@ enum nvkm_devidx {
 
 	NVKM_ENGINE_NVENC0,
 	NVKM_ENGINE_NVENC1,
-	NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC1,
+	NVKM_ENGINE_NVENC2,
+	NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC2,
 
 	NVKM_ENGINE_NVDEC,
 	NVKM_ENGINE_PM,
@@ -102,6 +106,7 @@ struct nvkm_device {
 		NV_C0    = 0xc0,
 		NV_E0    = 0xe0,
 		GM100    = 0x110,
+		GP100    = 0x130,
 	} card_type;
 	u32 chipset;
 	u8  chiprev;
@@ -136,7 +141,7 @@ struct nvkm_device {
 	struct nvkm_volt *volt;
 
 	struct nvkm_engine *bsp;
-	struct nvkm_engine *ce[3];
+	struct nvkm_engine *ce[6];
 	struct nvkm_engine *cipher;
 	struct nvkm_disp *disp;
 	struct nvkm_dma *dma;
@@ -149,7 +154,7 @@ struct nvkm_device {
 	struct nvkm_engine *mspdec;
 	struct nvkm_engine *msppp;
 	struct nvkm_engine *msvld;
-	struct nvkm_engine *nvenc[2];
+	struct nvkm_engine *nvenc[3];
 	struct nvkm_engine *nvdec;
 	struct nvkm_pm *pm;
 	struct nvkm_engine *sec;
@@ -170,7 +175,6 @@ struct nvkm_device_func {
 	void (*fini)(struct nvkm_device *, bool suspend);
 	resource_size_t (*resource_addr)(struct nvkm_device *, unsigned bar);
 	resource_size_t (*resource_size)(struct nvkm_device *, unsigned bar);
-	bool cpu_coherent;
 };
 
 struct nvkm_device_quirk {
@@ -206,7 +210,7 @@ struct nvkm_device_chip {
 	int (*volt    )(struct nvkm_device *, int idx, struct nvkm_volt **);
 
 	int (*bsp     )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*ce[3]   )(struct nvkm_device *, int idx, struct nvkm_engine **);
+	int (*ce[6]   )(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*cipher  )(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*disp    )(struct nvkm_device *, int idx, struct nvkm_disp **);
 	int (*dma     )(struct nvkm_device *, int idx, struct nvkm_dma **);
@@ -219,7 +223,7 @@ struct nvkm_device_chip {
 	int (*mspdec  )(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*msppp   )(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*msvld   )(struct nvkm_device *, int idx, struct nvkm_engine **);
-	int (*nvenc[2])(struct nvkm_device *, int idx, struct nvkm_engine **);
+	int (*nvenc[3])(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*nvdec   )(struct nvkm_device *, int idx, struct nvkm_engine **);
 	int (*pm      )(struct nvkm_device *, int idx, struct nvkm_pm **);
 	int (*sec     )(struct nvkm_device *, int idx, struct nvkm_engine **);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h b/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h
index b5370cb56e3c..e5c9b6268dcc 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h
@@ -28,6 +28,7 @@ struct nvkm_device_tegra {
 	} iommu;
 
 	int gpu_speedo;
+	int gpu_speedo_id;
 };
 
 struct nvkm_device_tegra_func {
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
index 594d719ba41e..d3d26a1e215d 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h
@@ -7,4 +7,6 @@ int gf100_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
 int gk104_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
 int gm107_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
 int gm200_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
+int gp100_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
+int gp104_ce_new(struct nvkm_device *, int, struct nvkm_engine **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
index d4fdce27b297..e82049667ce4 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h
@@ -32,4 +32,6 @@ int gk104_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
 int gk110_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
 int gm107_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
 int gm200_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
+int gp100_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
+int gp104_disp_new(struct nvkm_device *, int, struct nvkm_disp **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
index 15ddfcf5e8db..ed92fec5292c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h
@@ -66,4 +66,5 @@ int gk20a_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
 int gm107_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
 int gm200_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
 int gm20b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
+int gp100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
index 6515f5810a26..89cf99307828 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h
@@ -42,4 +42,5 @@ int gk20a_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
 int gm107_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
 int gm200_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
 int gm20b_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
+int gp100_gr_new(struct nvkm_device *, int, struct nvkm_gr **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
index e39a1fea930b..a72f3290528a 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h
@@ -7,6 +7,9 @@ struct nvkm_bios {
 	u32 size;
 	u8 *data;
 
+	u32 image0_size;
+	u32 imaged_addr;
+
 	u32 bmp_offset;
 	u32 bit_offset;
 
@@ -22,10 +25,9 @@ struct nvkm_bios {
 u8  nvbios_checksum(const u8 *data, int size);
 u16 nvbios_findstr(const u8 *data, int size, const char *str, int len);
 int nvbios_memcmp(struct nvkm_bios *, u32 addr, const char *, u32 len);
-
-#define nvbios_rd08(b,o) (b)->data[(o)]
-#define nvbios_rd16(b,o) get_unaligned_le16(&(b)->data[(o)])
-#define nvbios_rd32(b,o) get_unaligned_le32(&(b)->data[(o)])
+u8  nvbios_rd08(struct nvkm_bios *, u32 addr);
+u16 nvbios_rd16(struct nvkm_bios *, u32 addr);
+u32 nvbios_rd32(struct nvkm_bios *, u32 addr);
 
 int nvkm_bios_new(struct nvkm_device *, int, struct nvkm_bios **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h
index db10c11f0595..c5a6ebd5a478 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h
@@ -25,7 +25,8 @@ u16 nvbios_outp_match(struct nvkm_bios *, u16 type, u16 mask,
 		      u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_outp *);
 
 struct nvbios_ocfg {
-	u16 match;
+	u8  proto;
+	u8  flags;
 	u16 clkcmp[2];
 };
 
@@ -33,7 +34,7 @@ u16 nvbios_ocfg_entry(struct nvkm_bios *, u16 outp, u8 idx,
 		      u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
 u16 nvbios_ocfg_parse(struct nvkm_bios *, u16 outp, u8 idx,
 		      u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *);
-u16 nvbios_ocfg_match(struct nvkm_bios *, u16 outp, u16 type,
+u16 nvbios_ocfg_match(struct nvkm_bios *, u16 outp, u8 proto, u8 flags,
 		      u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *);
 u16 nvbios_oclk_match(struct nvkm_bios *, u16 cmp, u32 khz);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
index 0a734fd06acf..3a410275fa71 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h
@@ -56,6 +56,8 @@ struct nvkm_fb {
 		int regions;
 	} tile;
 
+	u8 page;
+
 	struct nvkm_memory *mmu_rd;
 	struct nvkm_memory *mmu_wr;
 };
@@ -91,6 +93,8 @@ int gk104_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
 int gk20a_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
 int gm107_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
 int gm200_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
+int gp100_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
+int gp104_fb_new(struct nvkm_device *, int, struct nvkm_fb **);
 
 #include <subdev/bios.h>
 #include <subdev/bios/ramcfg.h>
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
index c6b90b6543b3..cd755baf9cab 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h
@@ -38,4 +38,5 @@ int gk104_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
 int gk20a_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
 int gm107_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
 int gm200_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
+int gp100_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
index 2e80682b2da1..27d25b18d85c 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h
@@ -7,11 +7,14 @@ struct nvkm_mc {
 	struct nvkm_subdev subdev;
 };
 
-void nvkm_mc_intr(struct nvkm_mc *, bool *handled);
-void nvkm_mc_intr_unarm(struct nvkm_mc *);
-void nvkm_mc_intr_rearm(struct nvkm_mc *);
-void nvkm_mc_reset(struct nvkm_mc *, enum nvkm_devidx);
-void nvkm_mc_unk260(struct nvkm_mc *, u32 data);
+void nvkm_mc_enable(struct nvkm_device *, enum nvkm_devidx);
+void nvkm_mc_disable(struct nvkm_device *, enum nvkm_devidx);
+void nvkm_mc_reset(struct nvkm_device *, enum nvkm_devidx);
+void nvkm_mc_intr(struct nvkm_device *, bool *handled);
+void nvkm_mc_intr_unarm(struct nvkm_device *);
+void nvkm_mc_intr_rearm(struct nvkm_device *);
+void nvkm_mc_intr_mask(struct nvkm_device *, enum nvkm_devidx, bool enable);
+void nvkm_mc_unk260(struct nvkm_device *, u32 data);
 
 int nv04_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
 int nv11_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
@@ -24,4 +27,5 @@ int gt215_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
 int gf100_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
 int gk104_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
 int gk20a_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
+int gp100_mc_new(struct nvkm_device *, int, struct nvkm_mc **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
index ddb913889d7e..e6523e2cea9f 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h
@@ -47,6 +47,7 @@ int g94_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
 int gf100_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
 int gf106_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
 int gk104_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
+int gp100_pci_new(struct nvkm_device *, int, struct nvkm_pci **);
 
 /* pcie functions */
 int nvkm_pcie_set_link(struct nvkm_pci *, enum nvkm_pcie_speed, u8 width);
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h
index c6edd95a5b69..b04c38c07761 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h
@@ -43,9 +43,8 @@ struct nvkm_secboot {
 	const struct nvkm_secboot_func *func;
 	struct nvkm_subdev subdev;
 
+	enum nvkm_devidx devidx;
 	u32 base;
-	u32 irq_mask;
-	u32 enable_mask;
 };
 #define nvkm_secboot(p) container_of((p), struct nvkm_secboot, subdev)
 
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
index 8fb575a92c48..71ebbfd4484f 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h
@@ -8,10 +8,11 @@ struct nvkm_top {
 	struct list_head device;
 };
 
-u32 nvkm_top_reset(struct nvkm_top *, enum nvkm_devidx);
-u32 nvkm_top_intr(struct nvkm_top *, u32 intr, u64 *subdevs);
-enum nvkm_devidx nvkm_top_fault(struct nvkm_top *, int fault);
-enum nvkm_devidx nvkm_top_engine(struct nvkm_top *, int, int *runl, int *engn);
+u32 nvkm_top_reset(struct nvkm_device *, enum nvkm_devidx);
+u32 nvkm_top_intr(struct nvkm_device *, u32 intr, u64 *subdevs);
+u32 nvkm_top_intr_mask(struct nvkm_device *, enum nvkm_devidx);
+enum nvkm_devidx nvkm_top_fault(struct nvkm_device *, int fault);
+enum nvkm_devidx nvkm_top_engine(struct nvkm_device *, int, int *runl, int *engn);
 
 int gk104_top_new(struct nvkm_device *, int, struct nvkm_top **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
index feff55cff05b..b765f4ffcde6 100644
--- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
+++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h
@@ -12,6 +12,9 @@ struct nvkm_volt {
 		u32 uv;
 		u8 vid;
 	} vid[256];
+
+	u32 max_uv;
+	u32 min_uv;
 };
 
 int nvkm_volt_get(struct nvkm_volt *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index eb7de487a2b3..7bd4683216d0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -100,6 +100,7 @@ nouveau_abi16_swclass(struct nouveau_drm *drm)
 	case NV_DEVICE_INFO_V0_FERMI:
 	case NV_DEVICE_INFO_V0_KEPLER:
 	case NV_DEVICE_INFO_V0_MAXWELL:
+	case NV_DEVICE_INFO_V0_PASCAL:
 		return NVIF_CLASS_SW_GF100;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 5e3f3e826476..528bdeffb339 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -209,8 +209,7 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	nvbo->tile_flags = tile_flags;
 	nvbo->bo.bdev = &drm->ttm.bdev;
 
-	if (!nvxx_device(&drm->device)->func->cpu_coherent)
-		nvbo->force_coherent = flags & TTM_PL_FLAG_UNCACHED;
+	nvbo->force_coherent = flags & TTM_PL_FLAG_UNCACHED;
 
 	nvbo->page_shift = 12;
 	if (drm->client.vm) {
@@ -424,13 +423,7 @@ nouveau_bo_map(struct nouveau_bo *nvbo)
 	if (ret)
 		return ret;
 
-	/*
-	 * TTM buffers allocated using the DMA API already have a mapping, let's
-	 * use it instead.
-	 */
-	if (!nvbo->force_coherent)
-		ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages,
-				  &nvbo->kmap);
+	ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages, &nvbo->kmap);
 
 	ttm_bo_unreserve(&nvbo->bo);
 	return ret;
@@ -442,12 +435,7 @@ nouveau_bo_unmap(struct nouveau_bo *nvbo)
 	if (!nvbo)
 		return;
 
-	/*
-	 * TTM buffers allocated using the DMA API already had a coherent
-	 * mapping which we used, no need to unmap.
-	 */
-	if (!nvbo->force_coherent)
-		ttm_bo_kunmap(&nvbo->kmap);
+	ttm_bo_kunmap(&nvbo->kmap);
 }
 
 void
@@ -506,35 +494,13 @@ nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
 	return 0;
 }
 
-static inline void *
-_nouveau_bo_mem_index(struct nouveau_bo *nvbo, unsigned index, void *mem, u8 sz)
-{
-	struct ttm_dma_tt *dma_tt;
-	u8 *m = mem;
-
-	index *= sz;
-
-	if (m) {
-		/* kmap'd address, return the corresponding offset */
-		m += index;
-	} else {
-		/* DMA-API mapping, lookup the right address */
-		dma_tt = (struct ttm_dma_tt *)nvbo->bo.ttm;
-		m = dma_tt->cpu_address[index / PAGE_SIZE];
-		m += index % PAGE_SIZE;
-	}
-
-	return m;
-}
-#define nouveau_bo_mem_index(o, i, m) _nouveau_bo_mem_index(o, i, m, sizeof(*m))
-
 void
 nouveau_bo_wr16(struct nouveau_bo *nvbo, unsigned index, u16 val)
 {
 	bool is_iomem;
 	u16 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
 
-	mem = nouveau_bo_mem_index(nvbo, index, mem);
+	mem += index;
 
 	if (is_iomem)
 		iowrite16_native(val, (void __force __iomem *)mem);
@@ -548,7 +514,7 @@ nouveau_bo_rd32(struct nouveau_bo *nvbo, unsigned index)
 	bool is_iomem;
 	u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
 
-	mem = nouveau_bo_mem_index(nvbo, index, mem);
+	mem += index;
 
 	if (is_iomem)
 		return ioread32_native((void __force __iomem *)mem);
@@ -562,7 +528,7 @@ nouveau_bo_wr32(struct nouveau_bo *nvbo, unsigned index, u32 val)
 	bool is_iomem;
 	u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
 
-	mem = nouveau_bo_mem_index(nvbo, index, mem);
+	mem += index;
 
 	if (is_iomem)
 		iowrite32_native(val, (void __force __iomem *)mem);
@@ -1082,7 +1048,6 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 				ret = ttm_bo_move_accel_cleanup(bo,
 								&fence->base,
 								evict,
-								no_wait_gpu,
 								new_mem);
 				nouveau_fence_unref(&fence);
 			}
@@ -1104,6 +1069,10 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 			    struct ttm_mem_reg *, struct ttm_mem_reg *);
 		int (*init)(struct nouveau_channel *, u32 handle);
 	} _methods[] = {
+		{  "COPY", 4, 0xc1b5, nve0_bo_move_copy, nve0_bo_move_init },
+		{  "GRCE", 0, 0xc1b5, nve0_bo_move_copy, nvc0_bo_move_init },
+		{  "COPY", 4, 0xc0b5, nve0_bo_move_copy, nve0_bo_move_init },
+		{  "GRCE", 0, 0xc0b5, nve0_bo_move_copy, nvc0_bo_move_init },
 		{  "COPY", 4, 0xb0b5, nve0_bo_move_copy, nve0_bo_move_init },
 		{  "GRCE", 0, 0xb0b5, nve0_bo_move_copy, nvc0_bo_move_init },
 		{  "COPY", 4, 0xa0b5, nve0_bo_move_copy, nve0_bo_move_init },
@@ -1289,6 +1258,10 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	struct nouveau_drm_tile *new_tile = NULL;
 	int ret = 0;
 
+	ret = ttm_bo_wait(bo, intr, no_wait_gpu);
+	if (ret)
+		return ret;
+
 	if (nvbo->pin_refcnt)
 		NV_WARN(drm, "Moving pinned object %p!\n", nvbo);
 
@@ -1324,7 +1297,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	/* Fallback to software copy. */
 	ret = ttm_bo_wait(bo, intr, no_wait_gpu);
 	if (ret == 0)
-		ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+		ret = ttm_bo_move_memcpy(bo, evict, intr, no_wait_gpu, new_mem);
 
 out:
 	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
@@ -1488,14 +1461,6 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 	dev = drm->dev;
 	pdev = device->dev;
 
-	/*
-	 * Objects matching this condition have been marked as force_coherent,
-	 * so use the DMA API for them.
-	 */
-	if (!nvxx_device(&drm->device)->func->cpu_coherent &&
-	    ttm->caching_state == tt_uncached)
-		return ttm_dma_populate(ttm_dma, dev->dev);
-
 #if IS_ENABLED(CONFIG_AGP)
 	if (drm->agp.bridge) {
 		return ttm_agp_tt_populate(ttm);
@@ -1553,16 +1518,6 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	dev = drm->dev;
 	pdev = device->dev;
 
-	/*
-	 * Objects matching this condition have been marked as force_coherent,
-	 * so use the DMA API for them.
-	 */
-	if (!nvxx_device(&drm->device)->func->cpu_coherent &&
-	    ttm->caching_state == tt_uncached) {
-		ttm_dma_unpopulate(ttm_dma, dev->dev);
-		return;
-	}
-
 #if IS_ENABLED(CONFIG_AGP)
 	if (drm->agp.bridge) {
 		ttm_agp_tt_unpopulate(ttm);
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index b1d2527c5625..f9b3c811187e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -191,7 +191,8 @@ static int
 nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device,
 		    u32 engine, struct nouveau_channel **pchan)
 {
-	static const u16 oclasses[] = { MAXWELL_CHANNEL_GPFIFO_A,
+	static const u16 oclasses[] = { PASCAL_CHANNEL_GPFIFO_A,
+					MAXWELL_CHANNEL_GPFIFO_A,
 					KEPLER_CHANNEL_GPFIFO_B,
 					KEPLER_CHANNEL_GPFIFO_A,
 					FERMI_CHANNEL_GPFIFO,
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 6e97862d31a1..afbf557b23d4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -495,6 +495,8 @@ nouveau_display_create(struct drm_device *dev)
 
 	if (nouveau_modeset != 2 && drm->vbios.dcb.entries) {
 		static const u16 oclass[] = {
+			GP104_DISP,
+			GP100_DISP,
 			GM200_DISP,
 			GM107_DISP,
 			GK110_DISP,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 1161f8b452f2..66c1280c0f1f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -198,6 +198,7 @@ nouveau_accel_init(struct nouveau_drm *drm)
 		case KEPLER_CHANNEL_GPFIFO_A:
 		case KEPLER_CHANNEL_GPFIFO_B:
 		case MAXWELL_CHANNEL_GPFIFO_A:
+		case PASCAL_CHANNEL_GPFIFO_A:
 			ret = nvc0_fence_create(drm);
 			break;
 		default:
@@ -316,7 +317,16 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 	if (vga_switcheroo_client_probe_defer(pdev))
 		return -EPROBE_DEFER;
 
-	/* remove conflicting drivers (vesafb, efifb etc) */
+	/* We need to check that the chipset is supported before booting
+	 * fbdev off the hardware, as there's no way to put it back.
+	 */
+	ret = nvkm_device_pci_new(pdev, NULL, "error", true, false, 0, &device);
+	if (ret)
+		return ret;
+
+	nvkm_device_del(&device);
+
+	/* Remove conflicting drivers (vesafb, efifb etc). */
 	aper = alloc_apertures(3);
 	if (!aper)
 		return -ENOMEM;
@@ -430,6 +440,11 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 	nouveau_vga_init(drm);
 
 	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		if (!nvxx_device(&drm->device)->mmu) {
+			ret = -ENOSYS;
+			goto fail_device;
+		}
+
 		ret = nvkm_vm_new(nvxx_device(&drm->device), 0, (1ULL << 40),
 				  0x1000, NULL, &drm->client.vm);
 		if (ret)
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 57aaf98a26f9..d1f248fd3506 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -552,6 +552,8 @@ nouveau_fbcon_init(struct drm_device *dev)
 	if (ret)
 		goto fini;
 
+	if (fbcon->helper.fbdev)
+		fbcon->helper.fbdev->pixmap.buf_align = 4;
 	return 0;
 
 fini:
diff --git a/drivers/gpu/drm/nouveau/nouveau_hwmon.c b/drivers/gpu/drm/nouveau/nouveau_hwmon.c
index 1ff4166af26e..71f764bf4cc6 100644
--- a/drivers/gpu/drm/nouveau/nouveau_hwmon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_hwmon.c
@@ -535,6 +535,40 @@ static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO,
 			  nouveau_hwmon_get_in0_input, NULL, 0);
 
 static ssize_t
+nouveau_hwmon_get_in0_min(struct device *d,
+			    struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nvkm_volt *volt = nvxx_volt(&drm->device);
+
+	if (!volt || !volt->min_uv)
+		return -ENODEV;
+
+	return sprintf(buf, "%i\n", volt->min_uv / 1000);
+}
+
+static SENSOR_DEVICE_ATTR(in0_min, S_IRUGO,
+			  nouveau_hwmon_get_in0_min, NULL, 0);
+
+static ssize_t
+nouveau_hwmon_get_in0_max(struct device *d,
+			    struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nvkm_volt *volt = nvxx_volt(&drm->device);
+
+	if (!volt || !volt->max_uv)
+		return -ENODEV;
+
+	return sprintf(buf, "%i\n", volt->max_uv / 1000);
+}
+
+static SENSOR_DEVICE_ATTR(in0_max, S_IRUGO,
+			  nouveau_hwmon_get_in0_max, NULL, 0);
+
+static ssize_t
 nouveau_hwmon_get_in0_label(struct device *d,
 			    struct device_attribute *a, char *buf)
 {
@@ -594,6 +628,8 @@ static struct attribute *hwmon_pwm_fan_attributes[] = {
 
 static struct attribute *hwmon_in0_attributes[] = {
 	&sensor_dev_attr_in0_input.dev_attr.attr,
+	&sensor_dev_attr_in0_min.dev_attr.attr,
+	&sensor_dev_attr_in0_max.dev_attr.attr,
 	&sensor_dev_attr_in0_label.dev_attr.attr,
 	NULL
 };
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index bcee91497eb9..1825dbc33192 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -164,6 +164,7 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 	case NV_DEVICE_INFO_V0_FERMI:
 	case NV_DEVICE_INFO_V0_KEPLER:
 	case NV_DEVICE_INFO_V0_MAXWELL:
+	case NV_DEVICE_INFO_V0_PASCAL:
 		node->memtype = (nvbo->tile_flags & 0xff00) >> 8;
 		break;
 	default:
diff --git a/drivers/gpu/drm/nouveau/nv04_fbcon.c b/drivers/gpu/drm/nouveau/nv04_fbcon.c
index 0f3e4bb411cc..7d9248b8c664 100644
--- a/drivers/gpu/drm/nouveau/nv04_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv04_fbcon.c
@@ -82,7 +82,6 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	uint32_t fg;
 	uint32_t bg;
 	uint32_t dsize;
-	uint32_t width;
 	uint32_t *data = (uint32_t *)image->data;
 	int ret;
 
@@ -93,9 +92,6 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	if (ret)
 		return ret;
 
-	width = ALIGN(image->width, 8);
-	dsize = ALIGN(width * image->height, 32) >> 5;
-
 	if (info->fix.visual == FB_VISUAL_TRUECOLOR ||
 	    info->fix.visual == FB_VISUAL_DIRECTCOLOR) {
 		fg = ((uint32_t *) info->pseudo_palette)[image->fg_color];
@@ -111,10 +107,11 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 			 ((image->dx + image->width) & 0xffff));
 	OUT_RING(chan, bg);
 	OUT_RING(chan, fg);
-	OUT_RING(chan, (image->height << 16) | width);
+	OUT_RING(chan, (image->height << 16) | image->width);
 	OUT_RING(chan, (image->height << 16) | image->width);
 	OUT_RING(chan, (image->dy << 16) | (image->dx & 0xffff));
 
+	dsize = ALIGN(image->width * image->height, 32) >> 5;
 	while (dsize) {
 		int iter_len = dsize > 128 ? 128 : dsize;
 
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 7a7788212df7..7d0edcbcfca7 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -297,6 +297,8 @@ nv50_core_create(struct nvif_device *device, struct nvif_object *disp,
 		.pushbuf = 0xb0007d00,
 	};
 	static const s32 oclass[] = {
+		GP104_DISP_CORE_CHANNEL_DMA,
+		GP100_DISP_CORE_CHANNEL_DMA,
 		GM200_DISP_CORE_CHANNEL_DMA,
 		GM107_DISP_CORE_CHANNEL_DMA,
 		GK110_DISP_CORE_CHANNEL_DMA,
diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c
index 33d9ee0fac40..1aeb698e9707 100644
--- a/drivers/gpu/drm/nouveau/nv50_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c
@@ -95,7 +95,7 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	struct nouveau_fbdev *nfbdev = info->par;
 	struct nouveau_drm *drm = nouveau_drm(nfbdev->dev);
 	struct nouveau_channel *chan = drm->channel;
-	uint32_t width, dwords, *data = (uint32_t *)image->data;
+	uint32_t dwords, *data = (uint32_t *)image->data;
 	uint32_t mask = ~(~0 >> (32 - info->var.bits_per_pixel));
 	uint32_t *palette = info->pseudo_palette;
 	int ret;
@@ -107,9 +107,6 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	if (ret)
 		return ret;
 
-	width = ALIGN(image->width, 32);
-	dwords = (width * image->height) >> 5;
-
 	BEGIN_NV04(chan, NvSub2D, 0x0814, 2);
 	if (info->fix.visual == FB_VISUAL_TRUECOLOR ||
 	    info->fix.visual == FB_VISUAL_DIRECTCOLOR) {
@@ -128,6 +125,7 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	OUT_RING(chan, 0);
 	OUT_RING(chan, image->dy);
 
+	dwords = ALIGN(image->width * image->height, 32) >> 5;
 	while (dwords) {
 		int push = dwords > 2047 ? 2047 : dwords;
 
diff --git a/drivers/gpu/drm/nouveau/nvc0_fbcon.c b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
index a0913359ac05..839f4c8c1805 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
@@ -95,7 +95,7 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	struct nouveau_fbdev *nfbdev = info->par;
 	struct nouveau_drm *drm = nouveau_drm(nfbdev->dev);
 	struct nouveau_channel *chan = drm->channel;
-	uint32_t width, dwords, *data = (uint32_t *)image->data;
+	uint32_t dwords, *data = (uint32_t *)image->data;
 	uint32_t mask = ~(~0 >> (32 - info->var.bits_per_pixel));
 	uint32_t *palette = info->pseudo_palette;
 	int ret;
@@ -107,9 +107,6 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	if (ret)
 		return ret;
 
-	width = ALIGN(image->width, 32);
-	dwords = (width * image->height) >> 5;
-
 	BEGIN_NVC0(chan, NvSub2D, 0x0814, 2);
 	if (info->fix.visual == FB_VISUAL_TRUECOLOR ||
 	    info->fix.visual == FB_VISUAL_DIRECTCOLOR) {
@@ -128,6 +125,7 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	OUT_RING  (chan, 0);
 	OUT_RING  (chan, image->dy);
 
+	dwords = ALIGN(image->width * image->height, 32) >> 5;
 	while (dwords) {
 		int push = dwords > 2047 ? 2047 : dwords;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
index b18557858f19..19044aba265e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
+++ b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c
@@ -57,6 +57,9 @@ nvkm_subdev_name[NVKM_SUBDEV_NR] = {
 	[NVKM_ENGINE_CE0     ] = "ce0",
 	[NVKM_ENGINE_CE1     ] = "ce1",
 	[NVKM_ENGINE_CE2     ] = "ce2",
+	[NVKM_ENGINE_CE3     ] = "ce3",
+	[NVKM_ENGINE_CE4     ] = "ce4",
+	[NVKM_ENGINE_CE5     ] = "ce5",
 	[NVKM_ENGINE_CIPHER  ] = "cipher",
 	[NVKM_ENGINE_DISP    ] = "disp",
 	[NVKM_ENGINE_DMAOBJ  ] = "dma",
@@ -71,6 +74,7 @@ nvkm_subdev_name[NVKM_SUBDEV_NR] = {
 	[NVKM_ENGINE_MSVLD   ] = "msvld",
 	[NVKM_ENGINE_NVENC0  ] = "nvenc0",
 	[NVKM_ENGINE_NVENC1  ] = "nvenc1",
+	[NVKM_ENGINE_NVENC2  ] = "nvenc2",
 	[NVKM_ENGINE_NVDEC   ] = "nvdec",
 	[NVKM_ENGINE_PM      ] = "pm",
 	[NVKM_ENGINE_SEC     ] = "sec",
@@ -105,7 +109,7 @@ nvkm_subdev_fini(struct nvkm_subdev *subdev, bool suspend)
 		}
 	}
 
-	nvkm_mc_reset(device->mc, subdev->index);
+	nvkm_mc_reset(device, subdev->index);
 
 	time = ktime_to_us(ktime_get()) - time;
 	nvkm_trace(subdev, "%s completed in %lldus\n", action, time);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild
index 9c19d59b47df..a4458a8eb30a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild
@@ -3,3 +3,5 @@ nvkm-y += nvkm/engine/ce/gf100.o
 nvkm-y += nvkm/engine/ce/gk104.o
 nvkm-y += nvkm/engine/ce/gm107.o
 nvkm-y += nvkm/engine/ce/gm200.o
+nvkm-y += nvkm/engine/ce/gp100.o
+nvkm-y += nvkm/engine/ce/gp104.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c
new file mode 100644
index 000000000000..c7710456bc30
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c
@@ -0,0 +1,102 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "priv.h"
+#include <core/enum.h>
+
+#include <nvif/class.h>
+
+static const struct nvkm_enum
+gp100_ce_launcherr_report[] = {
+	{ 0x0, "NO_ERR" },
+	{ 0x1, "2D_LAYER_EXCEEDS_DEPTH" },
+	{ 0x2, "INVALID_ALIGNMENT" },
+	{ 0x3, "MEM2MEM_RECT_OUT_OF_BOUNDS" },
+	{ 0x4, "SRC_LINE_EXCEEDS_PITCH" },
+	{ 0x5, "SRC_LINE_EXCEEDS_NEG_PITCH" },
+	{ 0x6, "DST_LINE_EXCEEDS_PITCH" },
+	{ 0x7, "DST_LINE_EXCEEDS_NEG_PITCH" },
+	{ 0x8, "BAD_SRC_PIXEL_COMP_REF" },
+	{ 0x9, "INVALID_VALUE" },
+	{ 0xa, "UNUSED_FIELD" },
+	{ 0xb, "INVALID_OPERATION" },
+	{ 0xc, "NO_RESOURCES" },
+	{ 0xd, "INVALID_CONFIG" },
+	{}
+};
+
+static void
+gp100_ce_intr_launcherr(struct nvkm_engine *ce, const u32 base)
+{
+	struct nvkm_subdev *subdev = &ce->subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 stat = nvkm_rd32(device, 0x104418 + base);
+	const struct nvkm_enum *en =
+		nvkm_enum_find(gp100_ce_launcherr_report, stat & 0x0000000f);
+	nvkm_warn(subdev, "LAUNCHERR %08x [%s]\n", stat, en ? en->name : "");
+}
+
+void
+gp100_ce_intr(struct nvkm_engine *ce)
+{
+	const u32 base = (ce->subdev.index - NVKM_ENGINE_CE0) * 0x80;
+	struct nvkm_subdev *subdev = &ce->subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 mask = nvkm_rd32(device, 0x10440c + base);
+	u32 intr = nvkm_rd32(device, 0x104410 + base) & mask;
+	if (intr & 0x00000001) { //XXX: guess
+		nvkm_warn(subdev, "BLOCKPIPE\n");
+		nvkm_wr32(device, 0x104410 + base, 0x00000001);
+		intr &= ~0x00000001;
+	}
+	if (intr & 0x00000002) { //XXX: guess
+		nvkm_warn(subdev, "NONBLOCKPIPE\n");
+		nvkm_wr32(device, 0x104410 + base, 0x00000002);
+		intr &= ~0x00000002;
+	}
+	if (intr & 0x00000004) {
+		gp100_ce_intr_launcherr(ce, base);
+		nvkm_wr32(device, 0x104410 + base, 0x00000004);
+		intr &= ~0x00000004;
+	}
+	if (intr) {
+		nvkm_warn(subdev, "intr %08x\n", intr);
+		nvkm_wr32(device, 0x104410 + base, intr);
+	}
+}
+
+static const struct nvkm_engine_func
+gp100_ce = {
+	.intr = gp100_ce_intr,
+	.sclass = {
+		{ -1, -1, PASCAL_DMA_COPY_A },
+		{}
+	}
+};
+
+int
+gp100_ce_new(struct nvkm_device *device, int index,
+	     struct nvkm_engine **pengine)
+{
+	return nvkm_engine_new_(&gp100_ce, device, index, true, pengine);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c
new file mode 100644
index 000000000000..20e019788a53
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "priv.h"
+#include <core/enum.h>
+
+#include <nvif/class.h>
+
+static const struct nvkm_engine_func
+gp104_ce = {
+	.intr = gp100_ce_intr,
+	.sclass = {
+		{ -1, -1, PASCAL_DMA_COPY_B },
+		{ -1, -1, PASCAL_DMA_COPY_A },
+		{}
+	}
+};
+
+int
+gp104_ce_new(struct nvkm_device *device, int index,
+	     struct nvkm_engine **pengine)
+{
+	return nvkm_engine_new_(&gp104_ce, device, index, true, pengine);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h
index e2fa8b161943..2dce405976ad 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h
@@ -4,4 +4,5 @@
 
 void gt215_ce_intr(struct nvkm_falcon *, struct nvkm_fifo_chan *);
 void gk104_ce_intr(struct nvkm_engine *);
+void gp100_ce_intr(struct nvkm_engine *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
index 4572debcb0c9..7218a067a6c5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c
@@ -2148,6 +2148,67 @@ nv12b_chipset = {
 	.sw = gf100_sw_new,
 };
 
+static const struct nvkm_device_chip
+nv130_chipset = {
+	.name = "GP100",
+	.bar = gf100_bar_new,
+	.bios = nvkm_bios_new,
+	.bus = gf100_bus_new,
+	.devinit = gm200_devinit_new,
+	.fb = gp100_fb_new,
+	.fuse = gm107_fuse_new,
+	.gpio = gk104_gpio_new,
+	.i2c = gm200_i2c_new,
+	.ibus = gm200_ibus_new,
+	.imem = nv50_instmem_new,
+	.ltc = gp100_ltc_new,
+	.mc = gp100_mc_new,
+	.mmu = gf100_mmu_new,
+	.secboot = gm200_secboot_new,
+	.pci = gp100_pci_new,
+	.timer = gk20a_timer_new,
+	.top = gk104_top_new,
+	.ce[0] = gp100_ce_new,
+	.ce[1] = gp100_ce_new,
+	.ce[2] = gp100_ce_new,
+	.ce[3] = gp100_ce_new,
+	.ce[4] = gp100_ce_new,
+	.ce[5] = gp100_ce_new,
+	.dma = gf119_dma_new,
+	.disp = gp100_disp_new,
+	.fifo = gp100_fifo_new,
+	.gr = gp100_gr_new,
+	.sw = gf100_sw_new,
+};
+
+static const struct nvkm_device_chip
+nv134_chipset = {
+	.name = "GP104",
+	.bar = gf100_bar_new,
+	.bios = nvkm_bios_new,
+	.bus = gf100_bus_new,
+	.devinit = gm200_devinit_new,
+	.fb = gp104_fb_new,
+	.fuse = gm107_fuse_new,
+	.gpio = gk104_gpio_new,
+	.i2c = gm200_i2c_new,
+	.ibus = gm200_ibus_new,
+	.imem = nv50_instmem_new,
+	.ltc = gp100_ltc_new,
+	.mc = gp100_mc_new,
+	.mmu = gf100_mmu_new,
+	.pci = gp100_pci_new,
+	.timer = gk20a_timer_new,
+	.top = gk104_top_new,
+	.ce[0] = gp104_ce_new,
+	.ce[1] = gp104_ce_new,
+	.ce[2] = gp104_ce_new,
+	.ce[3] = gp104_ce_new,
+	.disp = gp104_disp_new,
+	.dma = gf119_dma_new,
+	.fifo = gp100_fifo_new,
+};
+
 static int
 nvkm_device_event_ctor(struct nvkm_object *object, void *data, u32 size,
 		       struct nvkm_notify *notify)
@@ -2221,6 +2282,9 @@ nvkm_device_engine(struct nvkm_device *device, int index)
 	_(CE0    , device->ce[0]   ,  device->ce[0]);
 	_(CE1    , device->ce[1]   ,  device->ce[1]);
 	_(CE2    , device->ce[2]   ,  device->ce[2]);
+	_(CE3    , device->ce[3]   ,  device->ce[3]);
+	_(CE4    , device->ce[4]   ,  device->ce[4]);
+	_(CE5    , device->ce[5]   ,  device->ce[5]);
 	_(CIPHER , device->cipher  ,  device->cipher);
 	_(DISP   , device->disp    , &device->disp->engine);
 	_(DMAOBJ , device->dma     , &device->dma->engine);
@@ -2235,6 +2299,7 @@ nvkm_device_engine(struct nvkm_device *device, int index)
 	_(MSVLD  , device->msvld   ,  device->msvld);
 	_(NVENC0 , device->nvenc[0],  device->nvenc[0]);
 	_(NVENC1 , device->nvenc[1],  device->nvenc[1]);
+	_(NVENC2 , device->nvenc[2],  device->nvenc[2]);
 	_(NVDEC  , device->nvdec   ,  device->nvdec);
 	_(PM     , device->pm      , &device->pm->engine);
 	_(SEC    , device->sec     ,  device->sec);
@@ -2492,6 +2557,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 			case 0x100: device->card_type = NV_E0; break;
 			case 0x110:
 			case 0x120: device->card_type = GM100; break;
+			case 0x130: device->card_type = GP100; break;
 			default:
 				break;
 			}
@@ -2576,6 +2642,8 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 		case 0x124: device->chip = &nv124_chipset; break;
 		case 0x126: device->chip = &nv126_chipset; break;
 		case 0x12b: device->chip = &nv12b_chipset; break;
+		case 0x130: device->chip = &nv130_chipset; break;
+		case 0x134: device->chip = &nv134_chipset; break;
 		default:
 			nvdev_error(device, "unknown chipset (%08x)\n", boot0);
 			goto done;
@@ -2659,6 +2727,9 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 		_(NVKM_ENGINE_CE0     ,    ce[0]);
 		_(NVKM_ENGINE_CE1     ,    ce[1]);
 		_(NVKM_ENGINE_CE2     ,    ce[2]);
+		_(NVKM_ENGINE_CE3     ,    ce[3]);
+		_(NVKM_ENGINE_CE4     ,    ce[4]);
+		_(NVKM_ENGINE_CE5     ,    ce[5]);
 		_(NVKM_ENGINE_CIPHER  ,   cipher);
 		_(NVKM_ENGINE_DISP    ,     disp);
 		_(NVKM_ENGINE_DMAOBJ  ,      dma);
@@ -2673,6 +2744,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 		_(NVKM_ENGINE_MSVLD   ,    msvld);
 		_(NVKM_ENGINE_NVENC0  , nvenc[0]);
 		_(NVKM_ENGINE_NVENC1  , nvenc[1]);
+		_(NVKM_ENGINE_NVENC2  , nvenc[2]);
 		_(NVKM_ENGINE_NVDEC   ,    nvdec);
 		_(NVKM_ENGINE_PM      ,       pm);
 		_(NVKM_ENGINE_SEC     ,      sec);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c
index 18fab3973ce5..b1b693219db3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c
@@ -1614,7 +1614,6 @@ nvkm_device_pci_func = {
 	.fini = nvkm_device_pci_fini,
 	.resource_addr = nvkm_device_pci_resource_addr,
 	.resource_size = nvkm_device_pci_resource_size,
-	.cpu_coherent = !IS_ENABLED(CONFIG_ARM) && !IS_ENABLED(CONFIG_ARM64),
 };
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c
index ec12efb4689a..939682f18788 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c
@@ -191,13 +191,11 @@ static irqreturn_t
 nvkm_device_tegra_intr(int irq, void *arg)
 {
 	struct nvkm_device_tegra *tdev = arg;
-	struct nvkm_mc *mc = tdev->device.mc;
+	struct nvkm_device *device = &tdev->device;
 	bool handled = false;
-	if (likely(mc)) {
-		nvkm_mc_intr_unarm(mc);
-		nvkm_mc_intr(mc, &handled);
-		nvkm_mc_intr_rearm(mc);
-	}
+	nvkm_mc_intr_unarm(device);
+	nvkm_mc_intr(device, &handled);
+	nvkm_mc_intr_rearm(device);
 	return handled ? IRQ_HANDLED : IRQ_NONE;
 }
 
@@ -247,7 +245,6 @@ nvkm_device_tegra_func = {
 	.fini = nvkm_device_tegra_fini,
 	.resource_addr = nvkm_device_tegra_resource_addr,
 	.resource_size = nvkm_device_tegra_resource_size,
-	.cpu_coherent = false,
 };
 
 int
@@ -313,6 +310,7 @@ nvkm_device_tegra_new(const struct nvkm_device_tegra_func *func,
 		goto remove;
 
 	tdev->gpu_speedo = tegra_sku_info.gpu_speedo_value;
+	tdev->gpu_speedo_id = tegra_sku_info.gpu_speedo_id;
 	ret = nvkm_device_ctor(&nvkm_device_tegra_func, NULL, &pdev->dev,
 			       NVKM_DEVICE_TEGRA, pdev->id, NULL,
 			       cfg, dbg, detect, mmio, subdev_mask,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
index 137066426ed7..79a8f71cf788 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c
@@ -102,6 +102,7 @@ nvkm_udevice_info(struct nvkm_udevice *udev, void *data, u32 size)
 	case NV_C0: args->v0.family = NV_DEVICE_INFO_V0_FERMI; break;
 	case NV_E0: args->v0.family = NV_DEVICE_INFO_V0_KEPLER; break;
 	case GM100: args->v0.family = NV_DEVICE_INFO_V0_MAXWELL; break;
+	case GP100: args->v0.family = NV_DEVICE_INFO_V0_PASCAL; break;
 	default:
 		args->v0.family = 0;
 		break;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
index a74c5dd27dc0..77a52b54a31e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild
@@ -10,6 +10,8 @@ nvkm-y += nvkm/engine/disp/gk104.o
 nvkm-y += nvkm/engine/disp/gk110.o
 nvkm-y += nvkm/engine/disp/gm107.o
 nvkm-y += nvkm/engine/disp/gm200.o
+nvkm-y += nvkm/engine/disp/gp100.o
+nvkm-y += nvkm/engine/disp/gp104.o
 
 nvkm-y += nvkm/engine/disp/outp.o
 nvkm-y += nvkm/engine/disp/outpdp.o
@@ -18,6 +20,7 @@ nvkm-y += nvkm/engine/disp/piornv50.o
 nvkm-y += nvkm/engine/disp/sornv50.o
 nvkm-y += nvkm/engine/disp/sorg94.o
 nvkm-y += nvkm/engine/disp/sorgf119.o
+nvkm-y += nvkm/engine/disp/sorgm107.o
 nvkm-y += nvkm/engine/disp/sorgm200.o
 nvkm-y += nvkm/engine/disp/dport.o
 
@@ -44,12 +47,15 @@ nvkm-y += nvkm/engine/disp/rootgk104.o
 nvkm-y += nvkm/engine/disp/rootgk110.o
 nvkm-y += nvkm/engine/disp/rootgm107.o
 nvkm-y += nvkm/engine/disp/rootgm200.o
+nvkm-y += nvkm/engine/disp/rootgp100.o
+nvkm-y += nvkm/engine/disp/rootgp104.o
 
 nvkm-y += nvkm/engine/disp/channv50.o
 nvkm-y += nvkm/engine/disp/changf119.o
 
 nvkm-y += nvkm/engine/disp/dmacnv50.o
 nvkm-y += nvkm/engine/disp/dmacgf119.o
+nvkm-y += nvkm/engine/disp/dmacgp104.o
 
 nvkm-y += nvkm/engine/disp/basenv50.o
 nvkm-y += nvkm/engine/disp/baseg84.o
@@ -58,6 +64,7 @@ nvkm-y += nvkm/engine/disp/basegt215.o
 nvkm-y += nvkm/engine/disp/basegf119.o
 nvkm-y += nvkm/engine/disp/basegk104.o
 nvkm-y += nvkm/engine/disp/basegk110.o
+nvkm-y += nvkm/engine/disp/basegp104.o
 
 nvkm-y += nvkm/engine/disp/corenv50.o
 nvkm-y += nvkm/engine/disp/coreg84.o
@@ -69,6 +76,8 @@ nvkm-y += nvkm/engine/disp/coregk104.o
 nvkm-y += nvkm/engine/disp/coregk110.o
 nvkm-y += nvkm/engine/disp/coregm107.o
 nvkm-y += nvkm/engine/disp/coregm200.o
+nvkm-y += nvkm/engine/disp/coregp100.o
+nvkm-y += nvkm/engine/disp/coregp104.o
 
 nvkm-y += nvkm/engine/disp/ovlynv50.o
 nvkm-y += nvkm/engine/disp/ovlyg84.o
@@ -76,6 +85,7 @@ nvkm-y += nvkm/engine/disp/ovlygt200.o
 nvkm-y += nvkm/engine/disp/ovlygt215.o
 nvkm-y += nvkm/engine/disp/ovlygf119.o
 nvkm-y += nvkm/engine/disp/ovlygk104.o
+nvkm-y += nvkm/engine/disp/ovlygp104.o
 
 nvkm-y += nvkm/engine/disp/piocnv50.o
 nvkm-y += nvkm/engine/disp/piocgf119.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c
new file mode 100644
index 000000000000..51688e37c54e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c
@@ -0,0 +1,38 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "dmacnv50.h"
+#include "rootnv50.h"
+
+#include <nvif/class.h>
+
+const struct nv50_disp_dmac_oclass
+gp104_disp_base_oclass = {
+	.base.oclass = GK110_DISP_BASE_CHANNEL_DMA,
+	.base.minver = 0,
+	.base.maxver = 0,
+	.ctor = nv50_disp_base_new,
+	.func = &gp104_disp_dmac_func,
+	.mthd = &gf119_disp_base_chan_mthd,
+	.chid = 1,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h
index aee374884c96..f5f683d9fd20 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h
@@ -85,6 +85,7 @@ extern const struct nv50_disp_mthd_list gf119_disp_core_mthd_pior;
 extern const struct nv50_disp_chan_mthd gf119_disp_base_chan_mthd;
 
 extern const struct nv50_disp_chan_mthd gk104_disp_core_chan_mthd;
+extern const struct nv50_disp_chan_mthd gk104_disp_ovly_chan_mthd;
 
 struct nv50_disp_pioc_oclass {
 	int (*ctor)(const struct nv50_disp_chan_func *,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c
index 6b1dc703dac7..21fbf89b6319 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c
@@ -171,7 +171,7 @@ gf119_disp_core_chan_mthd = {
 	}
 };
 
-static void
+void
 gf119_disp_core_fini(struct nv50_disp_dmac *chan)
 {
 	struct nv50_disp *disp = chan->base.root->disp;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c
new file mode 100644
index 000000000000..d5dff6619d4d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c
@@ -0,0 +1,38 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "dmacnv50.h"
+#include "rootnv50.h"
+
+#include <nvif/class.h>
+
+const struct nv50_disp_dmac_oclass
+gp100_disp_core_oclass = {
+	.base.oclass = GP100_DISP_CORE_CHANNEL_DMA,
+	.base.minver = 0,
+	.base.maxver = 0,
+	.ctor = nv50_disp_core_new,
+	.func = &gf119_disp_core_func,
+	.mthd = &gk104_disp_core_chan_mthd,
+	.chid = 0,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c
new file mode 100644
index 000000000000..6922f4007b61
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "dmacnv50.h"
+#include "rootnv50.h"
+
+#include <subdev/timer.h>
+
+#include <nvif/class.h>
+
+static int
+gp104_disp_core_init(struct nv50_disp_dmac *chan)
+{
+	struct nv50_disp *disp = chan->base.root->disp;
+	struct nvkm_subdev *subdev = &disp->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+
+	/* enable error reporting */
+	nvkm_mask(device, 0x6100a0, 0x00000001, 0x00000001);
+
+	/* initialise channel for dma command submission */
+	nvkm_wr32(device, 0x611494, chan->push);
+	nvkm_wr32(device, 0x611498, 0x00010000);
+	nvkm_wr32(device, 0x61149c, 0x00000001);
+	nvkm_mask(device, 0x610490, 0x00000010, 0x00000010);
+	nvkm_wr32(device, 0x640000, 0x00000000);
+	nvkm_wr32(device, 0x610490, 0x01000013);
+
+	/* wait for it to go inactive */
+	if (nvkm_msec(device, 2000,
+		if (!(nvkm_rd32(device, 0x610490) & 0x80000000))
+			break;
+	) < 0) {
+		nvkm_error(subdev, "core init: %08x\n",
+			   nvkm_rd32(device, 0x610490));
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+const struct nv50_disp_dmac_func
+gp104_disp_core_func = {
+	.init = gp104_disp_core_init,
+	.fini = gf119_disp_core_fini,
+	.bind = gf119_disp_dmac_bind,
+};
+
+const struct nv50_disp_dmac_oclass
+gp104_disp_core_oclass = {
+	.base.oclass = GP104_DISP_CORE_CHANNEL_DMA,
+	.base.minver = 0,
+	.base.maxver = 0,
+	.ctor = nv50_disp_core_new,
+	.func = &gp104_disp_core_func,
+	.mthd = &gk104_disp_core_chan_mthd,
+	.chid = 0,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c
index 876b14549a58..a57f7cef307a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c
@@ -36,7 +36,7 @@ gf119_disp_dmac_bind(struct nv50_disp_dmac *chan,
 				 chan->base.chid << 27 | 0x00000001);
 }
 
-static void
+void
 gf119_disp_dmac_fini(struct nv50_disp_dmac *chan)
 {
 	struct nv50_disp *disp = chan->base.root->disp;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c
new file mode 100644
index 000000000000..ad24c2c57696
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c
@@ -0,0 +1,66 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "dmacnv50.h"
+#include "rootnv50.h"
+
+#include <subdev/timer.h>
+
+static int
+gp104_disp_dmac_init(struct nv50_disp_dmac *chan)
+{
+	struct nv50_disp *disp = chan->base.root->disp;
+	struct nvkm_subdev *subdev = &disp->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	int chid = chan->base.chid;
+
+	/* enable error reporting */
+	nvkm_mask(device, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid);
+
+	/* initialise channel for dma command submission */
+	nvkm_wr32(device, 0x611494 + (chid * 0x0010), chan->push);
+	nvkm_wr32(device, 0x611498 + (chid * 0x0010), 0x00010000);
+	nvkm_wr32(device, 0x61149c + (chid * 0x0010), 0x00000001);
+	nvkm_mask(device, 0x610490 + (chid * 0x0010), 0x00000010, 0x00000010);
+	nvkm_wr32(device, 0x640000 + (chid * 0x1000), 0x00000000);
+	nvkm_wr32(device, 0x610490 + (chid * 0x0010), 0x00000013);
+
+	/* wait for it to go inactive */
+	if (nvkm_msec(device, 2000,
+		if (!(nvkm_rd32(device, 0x610490 + (chid * 0x10)) & 0x80000000))
+			break;
+	) < 0) {
+		nvkm_error(subdev, "ch %d init: %08x\n", chid,
+			   nvkm_rd32(device, 0x610490 + (chid * 0x10)));
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+const struct nv50_disp_dmac_func
+gp104_disp_dmac_func = {
+	.init = gp104_disp_dmac_init,
+	.fini = gf119_disp_dmac_fini,
+	.bind = gf119_disp_dmac_bind,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h
index fc84eb8b5c45..43ac05857853 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h
@@ -25,8 +25,12 @@ int nv50_disp_dmac_bind(struct nv50_disp_dmac *, struct nvkm_object *, u32);
 extern const struct nv50_disp_dmac_func nv50_disp_core_func;
 
 extern const struct nv50_disp_dmac_func gf119_disp_dmac_func;
+void gf119_disp_dmac_fini(struct nv50_disp_dmac *);
 int gf119_disp_dmac_bind(struct nv50_disp_dmac *, struct nvkm_object *, u32);
 extern const struct nv50_disp_dmac_func gf119_disp_core_func;
+void gf119_disp_core_fini(struct nv50_disp_dmac *);
+
+extern const struct nv50_disp_dmac_func gp104_disp_dmac_func;
 
 struct nv50_disp_dmac_oclass {
 	int (*ctor)(const struct nv50_disp_dmac_func *,
@@ -88,4 +92,10 @@ extern const struct nv50_disp_dmac_oclass gk110_disp_base_oclass;
 extern const struct nv50_disp_dmac_oclass gm107_disp_core_oclass;
 
 extern const struct nv50_disp_dmac_oclass gm200_disp_core_oclass;
+
+extern const struct nv50_disp_dmac_oclass gp100_disp_core_oclass;
+
+extern const struct nv50_disp_dmac_oclass gp104_disp_core_oclass;
+extern const struct nv50_disp_dmac_oclass gp104_disp_base_oclass;
+extern const struct nv50_disp_dmac_oclass gp104_disp_ovly_oclass;
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
index f0314664349c..29e84b241cca 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c
@@ -79,8 +79,7 @@ exec_lookup(struct nv50_disp *disp, int head, int or, u32 ctrl,
 	list_for_each_entry(outp, &disp->base.outp, head) {
 		if ((outp->info.hasht & 0xff) == type &&
 		    (outp->info.hashm & mask) == mask) {
-			*data = nvbios_outp_match(bios, outp->info.hasht,
-							outp->info.hashm,
+			*data = nvbios_outp_match(bios, outp->info.hasht, mask,
 						  ver, hdr, cnt, len, info);
 			if (!*data)
 				return NULL;
@@ -155,25 +154,21 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf)
 	if (!outp)
 		return NULL;
 
+	*conf = (ctrl & 0x00000f00) >> 8;
 	switch (outp->info.type) {
 	case DCB_OUTPUT_TMDS:
-		*conf = (ctrl & 0x00000f00) >> 8;
 		if (*conf == 5)
 			*conf |= 0x0100;
 		break;
 	case DCB_OUTPUT_LVDS:
-		*conf = disp->sor.lvdsconf;
+		*conf |= disp->sor.lvdsconf;
 		break;
-	case DCB_OUTPUT_DP:
-		*conf = (ctrl & 0x00000f00) >> 8;
-		break;
-	case DCB_OUTPUT_ANALOG:
 	default:
-		*conf = 0x00ff;
 		break;
 	}
 
-	data = nvbios_ocfg_match(bios, data, *conf, &ver, &hdr, &cnt, &len, &info2);
+	data = nvbios_ocfg_match(bios, data, *conf & 0xff, *conf >> 8,
+				 &ver, &hdr, &cnt, &len, &info2);
 	if (data && id < 0xff) {
 		data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk);
 		if (data) {
@@ -418,7 +413,7 @@ gf119_disp_intr_supervisor(struct work_struct *work)
 	nvkm_wr32(device, 0x6101d0, 0x80000000);
 }
 
-static void
+void
 gf119_disp_intr_error(struct nv50_disp *disp, int chid)
 {
 	struct nvkm_subdev *subdev = &disp->base.engine.subdev;
@@ -466,7 +461,7 @@ gf119_disp_intr(struct nv50_disp *disp)
 		u32 stat = nvkm_rd32(device, 0x61009c);
 		int chid = ffs(stat) - 1;
 		if (chid >= 0)
-			gf119_disp_intr_error(disp, chid);
+			disp->func->intr_error(disp, chid);
 		intr &= ~0x00000002;
 	}
 
@@ -510,6 +505,7 @@ gf119_disp_new_(const struct nv50_disp_func *func, struct nvkm_device *device,
 static const struct nv50_disp_func
 gf119_disp = {
 	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
 	.uevent = &gf119_disp_chan_uevent,
 	.super = gf119_disp_intr_supervisor,
 	.root = &gf119_disp_root_oclass,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
index a86384b8e388..37f145cf30d7 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c
@@ -27,6 +27,7 @@
 static const struct nv50_disp_func
 gk104_disp = {
 	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
 	.uevent = &gf119_disp_chan_uevent,
 	.super = gf119_disp_intr_supervisor,
 	.root = &gk104_disp_root_oclass,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
index 0d574c7e594a..e14ac946608c 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c
@@ -27,6 +27,7 @@
 static const struct nv50_disp_func
 gk110_disp = {
 	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
 	.uevent = &gf119_disp_chan_uevent,
 	.super = gf119_disp_intr_supervisor,
 	.root = &gk110_disp_root_oclass,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
index b6944142d616..2f2437cc5891 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c
@@ -27,6 +27,7 @@
 static const struct nv50_disp_func
 gm107_disp = {
 	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
 	.uevent = &gf119_disp_chan_uevent,
 	.super = gf119_disp_intr_supervisor,
 	.root = &gm107_disp_root_oclass,
@@ -36,7 +37,7 @@ gm107_disp = {
 	.outp.internal.crt = nv50_dac_output_new,
 	.outp.internal.tmds = nv50_sor_output_new,
 	.outp.internal.lvds = nv50_sor_output_new,
-	.outp.internal.dp = gf119_sor_dp_new,
+	.outp.internal.dp = gm107_sor_dp_new,
 	.dac.nr = 3,
 	.dac.power = nv50_dac_power,
 	.dac.sense = nv50_dac_sense,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
index 67eec8620719..9f368d4ee61e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c
@@ -27,6 +27,7 @@
 static const struct nv50_disp_func
 gm200_disp = {
 	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
 	.uevent = &gf119_disp_chan_uevent,
 	.super = gf119_disp_intr_supervisor,
 	.root = &gm200_disp_root_oclass,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
new file mode 100644
index 000000000000..4f81bf31435e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c
@@ -0,0 +1,55 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "nv50.h"
+#include "rootnv50.h"
+
+static const struct nv50_disp_func
+gp100_disp = {
+	.intr = gf119_disp_intr,
+	.intr_error = gf119_disp_intr_error,
+	.uevent = &gf119_disp_chan_uevent,
+	.super = gf119_disp_intr_supervisor,
+	.root = &gp100_disp_root_oclass,
+	.head.vblank_init = gf119_disp_vblank_init,
+	.head.vblank_fini = gf119_disp_vblank_fini,
+	.head.scanoutpos = gf119_disp_root_scanoutpos,
+	.outp.internal.crt = nv50_dac_output_new,
+	.outp.internal.tmds = nv50_sor_output_new,
+	.outp.internal.lvds = nv50_sor_output_new,
+	.outp.internal.dp = gm200_sor_dp_new,
+	.dac.nr = 3,
+	.dac.power = nv50_dac_power,
+	.dac.sense = nv50_dac_sense,
+	.sor.nr = 4,
+	.sor.power = nv50_sor_power,
+	.sor.hda_eld = gf119_hda_eld,
+	.sor.hdmi = gk104_hdmi_ctrl,
+	.sor.magic = gm200_sor_magic,
+};
+
+int
+gp100_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+{
+	return gf119_disp_new_(&gp100_disp, device, index, pdisp);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c
new file mode 100644
index 000000000000..3bf3380336e4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c
@@ -0,0 +1,81 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "nv50.h"
+#include "rootnv50.h"
+
+static void
+gp104_disp_intr_error(struct nv50_disp *disp, int chid)
+{
+	struct nvkm_subdev *subdev = &disp->base.engine.subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 mthd = nvkm_rd32(device, 0x6111f0 + (chid * 12));
+	u32 data = nvkm_rd32(device, 0x6111f4 + (chid * 12));
+	u32 unkn = nvkm_rd32(device, 0x6111f8 + (chid * 12));
+
+	nvkm_error(subdev, "chid %d mthd %04x data %08x %08x %08x\n",
+		   chid, (mthd & 0x0000ffc), data, mthd, unkn);
+
+	if (chid < ARRAY_SIZE(disp->chan)) {
+		switch (mthd & 0xffc) {
+		case 0x0080:
+			nv50_disp_chan_mthd(disp->chan[chid], NV_DBG_ERROR);
+			break;
+		default:
+			break;
+		}
+	}
+
+	nvkm_wr32(device, 0x61009c, (1 << chid));
+	nvkm_wr32(device, 0x6111f0 + (chid * 12), 0x90000000);
+}
+
+static const struct nv50_disp_func
+gp104_disp = {
+	.intr = gf119_disp_intr,
+	.intr_error = gp104_disp_intr_error,
+	.uevent = &gf119_disp_chan_uevent,
+	.super = gf119_disp_intr_supervisor,
+	.root = &gp104_disp_root_oclass,
+	.head.vblank_init = gf119_disp_vblank_init,
+	.head.vblank_fini = gf119_disp_vblank_fini,
+	.head.scanoutpos = gf119_disp_root_scanoutpos,
+	.outp.internal.crt = nv50_dac_output_new,
+	.outp.internal.tmds = nv50_sor_output_new,
+	.outp.internal.lvds = nv50_sor_output_new,
+	.outp.internal.dp = gm200_sor_dp_new,
+	.dac.nr = 3,
+	.dac.power = nv50_dac_power,
+	.dac.sense = nv50_dac_sense,
+	.sor.nr = 4,
+	.sor.power = nv50_sor_power,
+	.sor.hda_eld = gf119_hda_eld,
+	.sor.hdmi = gk104_hdmi_ctrl,
+	.sor.magic = gm200_sor_magic,
+};
+
+int
+gp104_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp)
+{
+	return gf119_disp_new_(&gp104_disp, device, index, pdisp);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
index 4226d2153b9c..fbb8c7dc18fd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c
@@ -32,6 +32,7 @@
 #include <subdev/bios/init.h>
 #include <subdev/bios/pll.h>
 #include <subdev/devinit.h>
+#include <subdev/timer.h>
 
 static const struct nvkm_disp_oclass *
 nv50_disp_root_(struct nvkm_disp *base)
@@ -269,8 +270,7 @@ exec_lookup(struct nv50_disp *disp, int head, int or, u32 ctrl,
 	list_for_each_entry(outp, &disp->base.outp, head) {
 		if ((outp->info.hasht & 0xff) == type &&
 		    (outp->info.hashm & mask) == mask) {
-			*data = nvbios_outp_match(bios, outp->info.hasht,
-							outp->info.hashm,
+			*data = nvbios_outp_match(bios, outp->info.hasht, mask,
 						  ver, hdr, cnt, len, info);
 			if (!*data)
 				return NULL;
@@ -387,22 +387,17 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf)
 	if (!outp)
 		return NULL;
 
+	*conf = (ctrl & 0x00000f00) >> 8;
 	if (outp->info.location == 0) {
 		switch (outp->info.type) {
 		case DCB_OUTPUT_TMDS:
-			*conf = (ctrl & 0x00000f00) >> 8;
 			if (*conf == 5)
 				*conf |= 0x0100;
 			break;
 		case DCB_OUTPUT_LVDS:
-			*conf = disp->sor.lvdsconf;
+			*conf |= disp->sor.lvdsconf;
 			break;
-		case DCB_OUTPUT_DP:
-			*conf = (ctrl & 0x00000f00) >> 8;
-			break;
-		case DCB_OUTPUT_ANALOG:
 		default:
-			*conf = 0x00ff;
 			break;
 		}
 	} else {
@@ -410,7 +405,8 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf)
 		pclk = pclk / 2;
 	}
 
-	data = nvbios_ocfg_match(bios, data, *conf, &ver, &hdr, &cnt, &len, &info2);
+	data = nvbios_ocfg_match(bios, data, *conf & 0xff, *conf >> 8,
+				 &ver, &hdr, &cnt, &len, &info2);
 	if (data && id < 0xff) {
 		data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk);
 		if (data) {
@@ -430,6 +426,134 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf)
 	return outp;
 }
 
+static bool
+nv50_disp_dptmds_war(struct nvkm_device *device)
+{
+	switch (device->chipset) {
+	case 0x94:
+	case 0x96:
+	case 0x98:
+	case 0xaa:
+	case 0xac:
+		return true;
+	default:
+		break;
+	}
+	return false;
+}
+
+static bool
+nv50_disp_dptmds_war_needed(struct nv50_disp *disp, struct dcb_output *outp)
+{
+	struct nvkm_device *device = disp->base.engine.subdev.device;
+	const u32 soff = __ffs(outp->or) * 0x800;
+	if (nv50_disp_dptmds_war(device) && outp->type == DCB_OUTPUT_TMDS) {
+		switch (nvkm_rd32(device, 0x614300 + soff) & 0x00030000) {
+		case 0x00000000:
+		case 0x00030000:
+			return true;
+		default:
+			break;
+		}
+	}
+	return false;
+
+}
+
+static void
+nv50_disp_dptmds_war_2(struct nv50_disp *disp, struct dcb_output *outp)
+{
+	struct nvkm_device *device = disp->base.engine.subdev.device;
+	const u32 soff = __ffs(outp->or) * 0x800;
+
+	if (!nv50_disp_dptmds_war_needed(disp, outp))
+		return;
+
+	nvkm_mask(device, 0x00e840, 0x80000000, 0x80000000);
+	nvkm_mask(device, 0x614300 + soff, 0x03000000, 0x03000000);
+	nvkm_mask(device, 0x61c10c + soff, 0x00000001, 0x00000001);
+
+	nvkm_mask(device, 0x61c00c + soff, 0x0f000000, 0x00000000);
+	nvkm_mask(device, 0x61c008 + soff, 0xff000000, 0x14000000);
+	nvkm_usec(device, 400, NVKM_DELAY);
+	nvkm_mask(device, 0x61c008 + soff, 0xff000000, 0x00000000);
+	nvkm_mask(device, 0x61c00c + soff, 0x0f000000, 0x01000000);
+
+	if (nvkm_rd32(device, 0x61c004 + soff) & 0x00000001) {
+		u32 seqctl = nvkm_rd32(device, 0x61c030 + soff);
+		u32  pu_pc = seqctl & 0x0000000f;
+		nvkm_wr32(device, 0x61c040 + soff + pu_pc * 4, 0x1f008000);
+	}
+}
+
+static void
+nv50_disp_dptmds_war_3(struct nv50_disp *disp, struct dcb_output *outp)
+{
+	struct nvkm_device *device = disp->base.engine.subdev.device;
+	const u32 soff = __ffs(outp->or) * 0x800;
+	u32 sorpwr;
+
+	if (!nv50_disp_dptmds_war_needed(disp, outp))
+		return;
+
+	sorpwr = nvkm_rd32(device, 0x61c004 + soff);
+	if (sorpwr & 0x00000001) {
+		u32 seqctl = nvkm_rd32(device, 0x61c030 + soff);
+		u32  pd_pc = (seqctl & 0x00000f00) >> 8;
+		u32  pu_pc =  seqctl & 0x0000000f;
+
+		nvkm_wr32(device, 0x61c040 + soff + pd_pc * 4, 0x1f008000);
+
+		nvkm_msec(device, 2000,
+			if (!(nvkm_rd32(device, 0x61c030 + soff) & 0x10000000))
+				break;
+		);
+		nvkm_mask(device, 0x61c004 + soff, 0x80000001, 0x80000000);
+		nvkm_msec(device, 2000,
+			if (!(nvkm_rd32(device, 0x61c030 + soff) & 0x10000000))
+				break;
+		);
+
+		nvkm_wr32(device, 0x61c040 + soff + pd_pc * 4, 0x00002000);
+		nvkm_wr32(device, 0x61c040 + soff + pu_pc * 4, 0x1f000000);
+	}
+
+	nvkm_mask(device, 0x61c10c + soff, 0x00000001, 0x00000000);
+	nvkm_mask(device, 0x614300 + soff, 0x03000000, 0x00000000);
+
+	if (sorpwr & 0x00000001) {
+		nvkm_mask(device, 0x61c004 + soff, 0x80000001, 0x80000001);
+	}
+}
+
+static void
+nv50_disp_update_sppll1(struct nv50_disp *disp)
+{
+	struct nvkm_device *device = disp->base.engine.subdev.device;
+	bool used = false;
+	int sor;
+
+	if (!nv50_disp_dptmds_war(device))
+		return;
+
+	for (sor = 0; sor < disp->func->sor.nr; sor++) {
+		u32 clksor = nvkm_rd32(device, 0x614300 + (sor * 0x800));
+		switch (clksor & 0x03000000) {
+		case 0x02000000:
+		case 0x03000000:
+			used = true;
+			break;
+		default:
+			break;
+		}
+	}
+
+	if (used)
+		return;
+
+	nvkm_mask(device, 0x00e840, 0x80000000, 0x00000000);
+}
+
 static void
 nv50_disp_intr_unk10_0(struct nv50_disp *disp, int head)
 {
@@ -683,6 +807,8 @@ nv50_disp_intr_unk20_2(struct nv50_disp *disp, int head)
 
 	nvkm_mask(device, hreg, 0x0000000f, hval);
 	nvkm_mask(device, oreg, mask, oval);
+
+	nv50_disp_dptmds_war_2(disp, &outp->info);
 }
 
 /* If programming a TMDS output on a SOR that can also be configured for
@@ -724,6 +850,7 @@ nv50_disp_intr_unk40_0(struct nv50_disp *disp, int head)
 
 	if (outp->info.location == 0 && outp->info.type == DCB_OUTPUT_TMDS)
 		nv50_disp_intr_unk40_0_tmds(disp, &outp->info);
+	nv50_disp_dptmds_war_3(disp, &outp->info);
 }
 
 void
@@ -771,6 +898,7 @@ nv50_disp_intr_supervisor(struct work_struct *work)
 				continue;
 			nv50_disp_intr_unk40_0(disp, head);
 		}
+		nv50_disp_update_sppll1(disp);
 	}
 
 	nvkm_wr32(device, 0x610030, 0x80000000);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
index aecebd8717e5..1e1de6bfe85a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h
@@ -68,6 +68,7 @@ struct nv50_disp_func_outp {
 
 struct nv50_disp_func {
 	void (*intr)(struct nv50_disp *);
+	void (*intr_error)(struct nv50_disp *, int chid);
 
 	const struct nvkm_event_func *uevent;
 	void (*super)(struct work_struct *);
@@ -114,4 +115,5 @@ void gf119_disp_vblank_init(struct nv50_disp *, int);
 void gf119_disp_vblank_fini(struct nv50_disp *, int);
 void gf119_disp_intr(struct nv50_disp *);
 void gf119_disp_intr_supervisor(struct work_struct *);
+void gf119_disp_intr_error(struct nv50_disp *, int);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h
index e9067ba4e179..4e983f6d7032 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h
@@ -62,7 +62,12 @@ int g94_sor_dp_lnk_pwr(struct nvkm_output_dp *, int);
 int gf119_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *,
 		     struct nvkm_output **);
 int gf119_sor_dp_lnk_ctl(struct nvkm_output_dp *, int, int, bool);
+int gf119_sor_dp_drv_ctl(struct nvkm_output_dp *, int, int, int, int);
 
-int  gm200_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *,
-		      struct nvkm_output **);
+int gm107_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *,
+		     struct nvkm_output **);
+int gm107_sor_dp_pattern(struct nvkm_output_dp *, int);
+
+int gm200_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *,
+		     struct nvkm_output **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c
index 2e2dc0641ef2..2f0220b39f34 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c
@@ -80,7 +80,7 @@ gk104_disp_ovly_mthd_base = {
 	}
 };
 
-static const struct nv50_disp_chan_mthd
+const struct nv50_disp_chan_mthd
 gk104_disp_ovly_chan_mthd = {
 	.name = "Overlay",
 	.addr = 0x001000,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c
new file mode 100644
index 000000000000..97e2dd2d908e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c
@@ -0,0 +1,38 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "dmacnv50.h"
+#include "rootnv50.h"
+
+#include <nvif/class.h>
+
+const struct nv50_disp_dmac_oclass
+gp104_disp_ovly_oclass = {
+	.base.oclass = GK104_DISP_OVERLAY_CONTROL_DMA,
+	.base.minver = 0,
+	.base.maxver = 0,
+	.ctor = nv50_disp_ovly_new,
+	.func = &gp104_disp_dmac_func,
+	.mthd = &gk104_disp_ovly_chan_mthd,
+	.chid = 5,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c
new file mode 100644
index 000000000000..ac8fdd728ec6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "rootnv50.h"
+#include "dmacnv50.h"
+
+#include <nvif/class.h>
+
+static const struct nv50_disp_root_func
+gp100_disp_root = {
+	.init = gf119_disp_root_init,
+	.fini = gf119_disp_root_fini,
+	.dmac = {
+		&gp100_disp_core_oclass,
+		&gk110_disp_base_oclass,
+		&gk104_disp_ovly_oclass,
+	},
+	.pioc = {
+		&gk104_disp_oimm_oclass,
+		&gk104_disp_curs_oclass,
+	},
+};
+
+static int
+gp100_disp_root_new(struct nvkm_disp *disp, const struct nvkm_oclass *oclass,
+		    void *data, u32 size, struct nvkm_object **pobject)
+{
+	return nv50_disp_root_new_(&gp100_disp_root, disp, oclass,
+				   data, size, pobject);
+}
+
+const struct nvkm_disp_oclass
+gp100_disp_root_oclass = {
+	.base.oclass = GP100_DISP,
+	.base.minver = -1,
+	.base.maxver = -1,
+	.ctor = gp100_disp_root_new,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c
new file mode 100644
index 000000000000..8443e04dc626
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "rootnv50.h"
+#include "dmacnv50.h"
+
+#include <nvif/class.h>
+
+static const struct nv50_disp_root_func
+gp104_disp_root = {
+	.init = gf119_disp_root_init,
+	.fini = gf119_disp_root_fini,
+	.dmac = {
+		&gp104_disp_core_oclass,
+		&gp104_disp_base_oclass,
+		&gp104_disp_ovly_oclass,
+	},
+	.pioc = {
+		&gk104_disp_oimm_oclass,
+		&gk104_disp_curs_oclass,
+	},
+};
+
+static int
+gp104_disp_root_new(struct nvkm_disp *disp, const struct nvkm_oclass *oclass,
+		    void *data, u32 size, struct nvkm_object **pobject)
+{
+	return nv50_disp_root_new_(&gp104_disp_root, disp, oclass,
+				   data, size, pobject);
+}
+
+const struct nvkm_disp_oclass
+gp104_disp_root_oclass = {
+	.base.oclass = GP104_DISP,
+	.base.minver = -1,
+	.base.maxver = -1,
+	.ctor = gp104_disp_root_new,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
index cb449ed8d92c..ad00f1724b72 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h
@@ -40,4 +40,6 @@ extern const struct nvkm_disp_oclass gk104_disp_root_oclass;
 extern const struct nvkm_disp_oclass gk110_disp_root_oclass;
 extern const struct nvkm_disp_oclass gm107_disp_root_oclass;
 extern const struct nvkm_disp_oclass gm200_disp_root_oclass;
+extern const struct nvkm_disp_oclass gp100_disp_root_oclass;
+extern const struct nvkm_disp_oclass gp104_disp_root_oclass;
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
index b4b41b135643..22706c0a54b5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c
@@ -40,8 +40,7 @@ static int
 gf119_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern)
 {
 	struct nvkm_device *device = outp->base.disp->engine.subdev.device;
-	const u32 loff = gf119_sor_loff(outp);
-	nvkm_mask(device, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern);
+	nvkm_mask(device, 0x61c110, 0x0f0f0f0f, 0x01010101 * pattern);
 	return 0;
 }
 
@@ -64,7 +63,7 @@ gf119_sor_dp_lnk_ctl(struct nvkm_output_dp *outp, int nr, int bw, bool ef)
 	return 0;
 }
 
-static int
+int
 gf119_sor_dp_drv_ctl(struct nvkm_output_dp *outp,
 		     int ln, int vs, int pe, int pc)
 {
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c
new file mode 100644
index 000000000000..37790b2617c5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c
@@ -0,0 +1,53 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "nv50.h"
+#include "outpdp.h"
+
+int
+gm107_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern)
+{
+	struct nvkm_device *device = outp->base.disp->engine.subdev.device;
+	const u32 soff = outp->base.or * 0x800;
+	const u32 data = 0x01010101 * pattern;
+	if (outp->base.info.sorconf.link & 1)
+		nvkm_mask(device, 0x61c110 + soff, 0x0f0f0f0f, data);
+	else
+		nvkm_mask(device, 0x61c12c + soff, 0x0f0f0f0f, data);
+	return 0;
+}
+
+static const struct nvkm_output_dp_func
+gm107_sor_dp_func = {
+	.pattern = gm107_sor_dp_pattern,
+	.lnk_pwr = g94_sor_dp_lnk_pwr,
+	.lnk_ctl = gf119_sor_dp_lnk_ctl,
+	.drv_ctl = gf119_sor_dp_drv_ctl,
+};
+
+int
+gm107_sor_dp_new(struct nvkm_disp *disp, int index,
+		 struct dcb_output *dcbE, struct nvkm_output **poutp)
+{
+	return nvkm_output_dp_new_(&gm107_sor_dp_func, disp, index, dcbE, poutp);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
index 2cfbef9c344f..c44fa7ea672a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c
@@ -57,19 +57,6 @@ gm200_sor_dp_lane_map(struct nvkm_device *device, u8 lane)
 }
 
 static int
-gm200_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern)
-{
-	struct nvkm_device *device = outp->base.disp->engine.subdev.device;
-	const u32 soff = gm200_sor_soff(outp);
-	const u32 data = 0x01010101 * pattern;
-	if (outp->base.info.sorconf.link & 1)
-		nvkm_mask(device, 0x61c110 + soff, 0x0f0f0f0f, data);
-	else
-		nvkm_mask(device, 0x61c12c + soff, 0x0f0f0f0f, data);
-	return 0;
-}
-
-static int
 gm200_sor_dp_lnk_pwr(struct nvkm_output_dp *outp, int nr)
 {
 	struct nvkm_device *device = outp->base.disp->engine.subdev.device;
@@ -129,7 +116,7 @@ gm200_sor_dp_drv_ctl(struct nvkm_output_dp *outp,
 
 static const struct nvkm_output_dp_func
 gm200_sor_dp_func = {
-	.pattern = gm200_sor_dp_pattern,
+	.pattern = gm107_sor_dp_pattern,
 	.lnk_pwr = gm200_sor_dp_lnk_pwr,
 	.lnk_ctl = gf119_sor_dp_lnk_ctl,
 	.drv_ctl = gm200_sor_dp_drv_ctl,
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild
index 65e5d291ecda..98651a43bc12 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild
@@ -13,6 +13,7 @@ nvkm-y += nvkm/engine/fifo/gk20a.o
 nvkm-y += nvkm/engine/fifo/gm107.o
 nvkm-y += nvkm/engine/fifo/gm200.o
 nvkm-y += nvkm/engine/fifo/gm20b.o
+nvkm-y += nvkm/engine/fifo/gp100.o
 
 nvkm-y += nvkm/engine/fifo/chan.o
 nvkm-y += nvkm/engine/fifo/channv50.o
@@ -31,3 +32,4 @@ nvkm-y += nvkm/engine/fifo/gpfifogf100.o
 nvkm-y += nvkm/engine/fifo/gpfifogk104.o
 nvkm-y += nvkm/engine/fifo/gpfifogk110.o
 nvkm-y += nvkm/engine/fifo/gpfifogm200.o
+nvkm-y += nvkm/engine/fifo/gpfifogp100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
index e06f4d46f802..230f64e5f731 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h
@@ -27,4 +27,5 @@ int gk104_fifo_gpfifo_new(struct nvkm_fifo *, const struct nvkm_oclass *,
 extern const struct nvkm_fifo_chan_oclass gk104_fifo_gpfifo_oclass;
 extern const struct nvkm_fifo_chan_oclass gk110_fifo_gpfifo_oclass;
 extern const struct nvkm_fifo_chan_oclass gm200_fifo_gpfifo_oclass;
+extern const struct nvkm_fifo_chan_oclass gp100_fifo_gpfifo_oclass;
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
index 743f3a189f28..103c0afaaa6d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c
@@ -329,7 +329,7 @@ gk104_fifo_intr_fault(struct gk104_fifo *fifo, int unit)
 	}
 
 	if (eu == NULL) {
-		enum nvkm_devidx engidx = nvkm_top_fault(device->top, unit);
+		enum nvkm_devidx engidx = nvkm_top_fault(device, unit);
 		if (engidx < NVKM_SUBDEV_NR) {
 			const char *src = nvkm_subdev_name[engidx];
 			char *dst = en;
@@ -589,7 +589,6 @@ gk104_fifo_oneinit(struct nvkm_fifo *base)
 	struct gk104_fifo *fifo = gk104_fifo(base);
 	struct nvkm_subdev *subdev = &fifo->base.engine.subdev;
 	struct nvkm_device *device = subdev->device;
-	struct nvkm_top *top = device->top;
 	int engn, runl, pbid, ret, i, j;
 	enum nvkm_devidx engidx;
 	u32 *map;
@@ -608,7 +607,7 @@ gk104_fifo_oneinit(struct nvkm_fifo *base)
 
 	/* Determine runlist configuration from topology device info. */
 	i = 0;
-	while ((int)(engidx = nvkm_top_engine(top, i++, &runl, &engn)) >= 0) {
+	while ((int)(engidx = nvkm_top_engine(device, i++, &runl, &engn)) >= 0) {
 		/* Determine which PBDMA handles requests for this engine. */
 		for (j = 0, pbid = -1; j < fifo->pbdma_nr; j++) {
 			if (map[j] & (1 << runl)) {
@@ -617,8 +616,8 @@ gk104_fifo_oneinit(struct nvkm_fifo *base)
 			}
 		}
 
-		nvkm_debug(subdev, "engine %2d: runlist %2d pbdma %2d\n",
-			   engn, runl, pbid);
+		nvkm_debug(subdev, "engine %2d: runlist %2d pbdma %2d (%s)\n",
+			   engn, runl, pbid, nvkm_subdev_name[engidx]);
 
 		fifo->engine[engn].engine = nvkm_device_engine(device, engidx);
 		fifo->engine[engn].runl = runl;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c
new file mode 100644
index 000000000000..eff83f7fb705
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c
@@ -0,0 +1,67 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "gk104.h"
+#include "changk104.h"
+
+static const struct nvkm_enum
+gp100_fifo_fault_engine[] = {
+	{ 0x01, "DISPLAY" },
+	{ 0x03, "IFB", NULL, NVKM_ENGINE_IFB },
+	{ 0x04, "BAR1", NULL, NVKM_SUBDEV_BAR },
+	{ 0x05, "BAR2", NULL, NVKM_SUBDEV_INSTMEM },
+	{ 0x06, "HOST0" },
+	{ 0x07, "HOST1" },
+	{ 0x08, "HOST2" },
+	{ 0x09, "HOST3" },
+	{ 0x0a, "HOST4" },
+	{ 0x0b, "HOST5" },
+	{ 0x0c, "HOST6" },
+	{ 0x0d, "HOST7" },
+	{ 0x0e, "HOST8" },
+	{ 0x0f, "HOST9" },
+	{ 0x10, "HOST10" },
+	{ 0x13, "PERF" },
+	{ 0x17, "PMU" },
+	{ 0x18, "PTP" },
+	{ 0x1f, "PHYSICAL" },
+	{}
+};
+
+static const struct gk104_fifo_func
+gp100_fifo = {
+	.fault.engine = gp100_fifo_fault_engine,
+	.fault.reason = gk104_fifo_fault_reason,
+	.fault.hubclient = gk104_fifo_fault_hubclient,
+	.fault.gpcclient = gk104_fifo_fault_gpcclient,
+	.chan = {
+		&gp100_fifo_gpfifo_oclass,
+		NULL
+	},
+};
+
+int
+gp100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo)
+{
+	return gk104_fifo_new_(&gp100_fifo, device, index, 4096, pfifo);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c
new file mode 100644
index 000000000000..1530a9217aea
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c
@@ -0,0 +1,34 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "changk104.h"
+
+#include <nvif/class.h>
+
+const struct nvkm_fifo_chan_oclass
+gp100_fifo_gpfifo_oclass = {
+	.base.oclass = PASCAL_CHANNEL_GPFIFO_A,
+	.base.minver = 0,
+	.base.maxver = 0,
+	.ctor = gk104_fifo_gpfifo_new,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild
index 290ed0db8047..f1c494182248 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild
@@ -31,6 +31,7 @@ nvkm-y += nvkm/engine/gr/gk20a.o
 nvkm-y += nvkm/engine/gr/gm107.o
 nvkm-y += nvkm/engine/gr/gm200.o
 nvkm-y += nvkm/engine/gr/gm20b.o
+nvkm-y += nvkm/engine/gr/gp100.o
 
 nvkm-y += nvkm/engine/gr/ctxnv40.o
 nvkm-y += nvkm/engine/gr/ctxnv50.o
@@ -48,3 +49,4 @@ nvkm-y += nvkm/engine/gr/ctxgk20a.o
 nvkm-y += nvkm/engine/gr/ctxgm107.o
 nvkm-y += nvkm/engine/gr/ctxgm200.o
 nvkm-y += nvkm/engine/gr/ctxgm20b.o
+nvkm-y += nvkm/engine/gr/ctxgp100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c
index b02d8f50ea6a..bc77eea351a5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c
@@ -1240,7 +1240,7 @@ gf100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	const struct gf100_grctx_func *grctx = gr->func->grctx;
 	u32 idle_timeout;
 
-	nvkm_mc_unk260(device->mc, 0);
+	nvkm_mc_unk260(device, 0);
 
 	gf100_gr_mmio(gr, grctx->hub);
 	gf100_gr_mmio(gr, grctx->gpc);
@@ -1264,7 +1264,7 @@ gf100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	gf100_gr_icmd(gr, grctx->icmd);
 	nvkm_wr32(device, 0x404154, idle_timeout);
 	gf100_gr_mthd(gr, grctx->mthd);
-	nvkm_mc_unk260(device->mc, 1);
+	nvkm_mc_unk260(device, 1);
 }
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h
index ac895edce164..52048b5a5274 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h
@@ -101,6 +101,8 @@ void gm200_grctx_generate_405b60(struct gf100_gr *);
 
 extern const struct gf100_grctx_func gm20b_grctx;
 
+extern const struct gf100_grctx_func gp100_grctx;
+
 /* context init value lists */
 
 extern const struct gf100_gr_pack gf100_grctx_pack_icmd[];
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c
index f521de11a299..c925ade5880e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c
@@ -226,7 +226,7 @@ gf117_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	u32 idle_timeout;
 	int i;
 
-	nvkm_mc_unk260(device->mc, 0);
+	nvkm_mc_unk260(device, 0);
 
 	gf100_gr_mmio(gr, grctx->hub);
 	gf100_gr_mmio(gr, grctx->gpc);
@@ -253,7 +253,7 @@ gf117_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	gf100_gr_icmd(gr, grctx->icmd);
 	nvkm_wr32(device, 0x404154, idle_timeout);
 	gf100_gr_mthd(gr, grctx->mthd);
-	nvkm_mc_unk260(device->mc, 1);
+	nvkm_mc_unk260(device, 1);
 }
 
 const struct gf100_grctx_func
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c
index 9ba337778ef5..c46b3fdf7203 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c
@@ -950,7 +950,7 @@ gk104_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	u32 idle_timeout;
 	int i;
 
-	nvkm_mc_unk260(device->mc, 0);
+	nvkm_mc_unk260(device, 0);
 
 	gf100_gr_mmio(gr, grctx->hub);
 	gf100_gr_mmio(gr, grctx->gpc);
@@ -979,7 +979,7 @@ gk104_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
 	gf100_gr_icmd(gr, grctx->icmd);
 	nvkm_wr32(device, 0x404154, idle_timeout);
 	gf100_gr_mthd(gr, grctx->mthd);
-	nvkm_mc_unk260(device->mc, 1);
+	nvkm_mc_unk260(device, 1);
 
 	nvkm_mask(device, 0x418800, 0x00200000, 0x00200000);
 	nvkm_mask(device, 0x41be10, 0x00800000, 0x00800000);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c
new file mode 100644
index 000000000000..3d1ae7ddf7dd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c
@@ -0,0 +1,179 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "ctxgf100.h"
+
+#include <subdev/fb.h>
+
+/*******************************************************************************
+ * PGRAPH context implementation
+ ******************************************************************************/
+
+static void
+gp100_grctx_generate_pagepool(struct gf100_grctx *info)
+{
+	const struct gf100_grctx_func *grctx = info->gr->func->grctx;
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, grctx->pagepool_size, (1 << s), access);
+	mmio_refn(info, 0x40800c, 0x00000000, s, b);
+	mmio_wr32(info, 0x408010, 0x80000000);
+	mmio_refn(info, 0x419004, 0x00000000, s, b);
+	mmio_wr32(info, 0x419008, 0x00000000);
+}
+
+static void
+gp100_grctx_generate_attrib(struct gf100_grctx *info)
+{
+	struct gf100_gr *gr = info->gr;
+	const struct gf100_grctx_func *grctx = gr->func->grctx;
+	const u32  alpha = grctx->alpha_nr;
+	const u32 attrib = grctx->attrib_nr;
+	const u32 pertpc = 0x20 * (grctx->attrib_nr_max + grctx->alpha_nr_max);
+	const u32   size = roundup(gr->tpc_total * pertpc, 0x80);
+	const u32 access = NV_MEM_ACCESS_RW;
+	const int s = 12;
+	const int b = mmio_vram(info, size, (1 << s), access);
+	const int max_batches = 0xffff;
+	u32 ao = 0;
+	u32 bo = ao + grctx->alpha_nr_max * gr->tpc_total;
+	int gpc, ppc, n = 0;
+
+	mmio_refn(info, 0x418810, 0x80000000, s, b);
+	mmio_refn(info, 0x419848, 0x10000000, s, b);
+	mmio_refn(info, 0x419c2c, 0x10000000, s, b);
+	mmio_refn(info, 0x419b00, 0x00000000, s, b);
+	mmio_wr32(info, 0x419b04, 0x80000000 | size >> 7);
+	mmio_wr32(info, 0x405830, attrib);
+	mmio_wr32(info, 0x40585c, alpha);
+	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
+
+	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
+		for (ppc = 0; ppc < gr->ppc_nr[gpc]; ppc++, n++) {
+			const u32 as =  alpha * gr->ppc_tpc_nr[gpc][ppc];
+			const u32 bs = attrib * gr->ppc_tpc_nr[gpc][ppc];
+			const u32 u = 0x418ea0 + (n * 0x04);
+			const u32 o = PPC_UNIT(gpc, ppc, 0);
+			if (!(gr->ppc_mask[gpc] & (1 << ppc)))
+				continue;
+			mmio_wr32(info, o + 0xc0, bs);
+			mmio_wr32(info, o + 0xf4, bo);
+			mmio_wr32(info, o + 0xf0, bs);
+			bo += grctx->attrib_nr_max * gr->ppc_tpc_nr[gpc][ppc];
+			mmio_wr32(info, o + 0xe4, as);
+			mmio_wr32(info, o + 0xf8, ao);
+			ao += grctx->alpha_nr_max * gr->ppc_tpc_nr[gpc][ppc];
+			mmio_wr32(info, u, bs);
+		}
+	}
+
+	mmio_wr32(info, 0x418eec, 0x00000000);
+	mmio_wr32(info, 0x41befc, 0x00000000);
+}
+
+static void
+gp100_grctx_generate_405b60(struct gf100_gr *gr)
+{
+	struct nvkm_device *device = gr->base.engine.subdev.device;
+	const u32 dist_nr = DIV_ROUND_UP(gr->tpc_total, 4);
+	u32 dist[TPC_MAX / 4] = {};
+	u32 gpcs[GPC_MAX * 2] = {};
+	u8  tpcnr[GPC_MAX];
+	int tpc, gpc, i;
+
+	memcpy(tpcnr, gr->tpc_nr, sizeof(gr->tpc_nr));
+
+	/* won't result in the same distribution as the binary driver where
+	 * some of the gpcs have more tpcs than others, but this shall do
+	 * for the moment.  the code for earlier gpus has this issue too.
+	 */
+	for (gpc = -1, i = 0; i < gr->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % gr->gpc_nr;
+		} while(!tpcnr[gpc]);
+		tpc = gr->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		dist[i / 4] |= ((gpc << 4) | tpc) << ((i % 4) * 8);
+		gpcs[gpc + (gr->gpc_nr * (tpc / 4))] |= i << (tpc * 8);
+	}
+
+	for (i = 0; i < dist_nr; i++)
+		nvkm_wr32(device, 0x405b60 + (i * 4), dist[i]);
+	for (i = 0; i < gr->gpc_nr * 2; i++)
+		nvkm_wr32(device, 0x405ba0 + (i * 4), gpcs[i]);
+}
+
+static void
+gp100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info)
+{
+	struct nvkm_device *device = gr->base.engine.subdev.device;
+	const struct gf100_grctx_func *grctx = gr->func->grctx;
+	u32 idle_timeout, tmp;
+	int i;
+
+	gf100_gr_mmio(gr, gr->fuc_sw_ctx);
+
+	idle_timeout = nvkm_mask(device, 0x404154, 0xffffffff, 0x00000000);
+
+	grctx->pagepool(info);
+	grctx->bundle(info);
+	grctx->attrib(info);
+	grctx->unkn(gr);
+
+	gm200_grctx_generate_tpcid(gr);
+	gf100_grctx_generate_r406028(gr);
+	gk104_grctx_generate_r418bb8(gr);
+
+	for (i = 0; i < 8; i++)
+		nvkm_wr32(device, 0x4064d0 + (i * 0x04), 0x00000000);
+	nvkm_wr32(device, 0x406500, 0x00000000);
+
+	nvkm_wr32(device, 0x405b00, (gr->tpc_total << 8) | gr->gpc_nr);
+
+	for (tmp = 0, i = 0; i < gr->gpc_nr; i++)
+		tmp |= ((1 << gr->tpc_nr[i]) - 1) << (i * 5);
+	nvkm_wr32(device, 0x4041c4, tmp);
+
+	gp100_grctx_generate_405b60(gr);
+
+	gf100_gr_icmd(gr, gr->fuc_bundle);
+	nvkm_wr32(device, 0x404154, idle_timeout);
+	gf100_gr_mthd(gr, gr->fuc_method);
+}
+
+const struct gf100_grctx_func
+gp100_grctx = {
+	.main  = gp100_grctx_generate_main,
+	.unkn  = gk104_grctx_generate_unkn,
+	.bundle = gm107_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0x180,
+	.bundle_token_limit = 0x1080,
+	.pagepool = gp100_grctx_generate_pagepool,
+	.pagepool_size = 0x20000,
+	.attrib = gp100_grctx_generate_attrib,
+	.attrib_nr_max = 0x660,
+	.attrib_nr = 0x440,
+	.alpha_nr_max = 0xc00,
+	.alpha_nr = 0x800,
+};
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
index 9513badb8220..157919c788e6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c
@@ -949,22 +949,41 @@ gf100_gr_trap_gpc_rop(struct gf100_gr *gr, int gpc)
 }
 
 static const struct nvkm_enum gf100_mp_warp_error[] = {
-	{ 0x00, "NO_ERROR" },
-	{ 0x01, "STACK_MISMATCH" },
+	{ 0x01, "STACK_ERROR" },
+	{ 0x02, "API_STACK_ERROR" },
+	{ 0x03, "RET_EMPTY_STACK_ERROR" },
+	{ 0x04, "PC_WRAP" },
 	{ 0x05, "MISALIGNED_PC" },
-	{ 0x08, "MISALIGNED_GPR" },
-	{ 0x09, "INVALID_OPCODE" },
-	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
-	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
-	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
+	{ 0x06, "PC_OVERFLOW" },
+	{ 0x07, "MISALIGNED_IMMC_ADDR" },
+	{ 0x08, "MISALIGNED_REG" },
+	{ 0x09, "ILLEGAL_INSTR_ENCODING" },
+	{ 0x0a, "ILLEGAL_SPH_INSTR_COMBO" },
+	{ 0x0b, "ILLEGAL_INSTR_PARAM" },
+	{ 0x0c, "INVALID_CONST_ADDR" },
+	{ 0x0d, "OOR_REG" },
+	{ 0x0e, "OOR_ADDR" },
+	{ 0x0f, "MISALIGNED_ADDR" },
 	{ 0x10, "INVALID_ADDR_SPACE" },
-	{ 0x11, "INVALID_PARAM" },
+	{ 0x11, "ILLEGAL_INSTR_PARAM2" },
+	{ 0x12, "INVALID_CONST_ADDR_LDC" },
+	{ 0x13, "GEOMETRY_SM_ERROR" },
+	{ 0x14, "DIVERGENT" },
+	{ 0x15, "WARP_EXIT" },
 	{}
 };
 
 static const struct nvkm_bitfield gf100_mp_global_error[] = {
+	{ 0x00000001, "SM_TO_SM_FAULT" },
+	{ 0x00000002, "L1_ERROR" },
 	{ 0x00000004, "MULTIPLE_WARP_ERRORS" },
-	{ 0x00000008, "OUT_OF_STACK_SPACE" },
+	{ 0x00000008, "PHYSICAL_STACK_OVERFLOW" },
+	{ 0x00000010, "BPT_INT" },
+	{ 0x00000020, "BPT_PAUSE" },
+	{ 0x00000040, "SINGLE_STEP_COMPLETE" },
+	{ 0x20000000, "ECC_SEC_ERROR" },
+	{ 0x40000000, "ECC_DED_ERROR" },
+	{ 0x80000000, "TIMEOUT" },
 	{}
 };
 
@@ -1438,24 +1457,30 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr)
 	struct nvkm_device *device = subdev->device;
 	struct nvkm_secboot *sb = device->secboot;
 	int i;
+	int ret = 0;
 
 	if (gr->firmware) {
 		/* load fuc microcode */
-		nvkm_mc_unk260(device->mc, 0);
+		nvkm_mc_unk260(device, 0);
 
 		/* securely-managed falcons must be reset using secure boot */
 		if (nvkm_secboot_is_managed(sb, NVKM_SECBOOT_FALCON_FECS))
-			nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_FECS);
+			ret = nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_FECS);
 		else
 			gf100_gr_init_fw(gr, 0x409000, &gr->fuc409c,
 					 &gr->fuc409d);
+		if (ret)
+			return ret;
+
 		if (nvkm_secboot_is_managed(sb, NVKM_SECBOOT_FALCON_GPCCS))
-			nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_GPCCS);
+			ret = nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_GPCCS);
 		else
 			gf100_gr_init_fw(gr, 0x41a000, &gr->fuc41ac,
 					 &gr->fuc41ad);
+		if (ret)
+			return ret;
 
-		nvkm_mc_unk260(device->mc, 1);
+		nvkm_mc_unk260(device, 1);
 
 		/* start both of them running */
 		nvkm_wr32(device, 0x409840, 0xffffffff);
@@ -1557,7 +1582,7 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr)
 	}
 
 	/* load HUB microcode */
-	nvkm_mc_unk260(device->mc, 0);
+	nvkm_mc_unk260(device, 0);
 	nvkm_wr32(device, 0x4091c0, 0x01000000);
 	for (i = 0; i < gr->func->fecs.ucode->data.size / 4; i++)
 		nvkm_wr32(device, 0x4091c4, gr->func->fecs.ucode->data.data[i]);
@@ -1580,7 +1605,7 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr)
 			nvkm_wr32(device, 0x41a188, i >> 6);
 		nvkm_wr32(device, 0x41a184, gr->func->gpccs.ucode->code.data[i]);
 	}
-	nvkm_mc_unk260(device->mc, 1);
+	nvkm_mc_unk260(device, 1);
 
 	/* load register lists */
 	gf100_gr_init_csdata(gr, grctx->hub, 0x409000, 0x000, 0x000000);
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
index 2b98abdb9270..268b8d60ff73 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h
@@ -292,4 +292,6 @@ extern const struct gf100_gr_init gm107_gr_init_l1c_0[];
 extern const struct gf100_gr_init gm107_gr_init_wwdx_0[];
 extern const struct gf100_gr_init gm107_gr_init_cbm_0[];
 void gm107_gr_init_bios(struct gf100_gr *);
+
+void gm200_gr_init_gpc_mmu(struct gf100_gr *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
index 4ca8ed15191c..de8b806b88fd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c
@@ -361,6 +361,5 @@ gk20a_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
 	if (ret)
 		return ret;
 
-
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
index 4dfa4513bb6c..6435f1257572 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c
@@ -38,7 +38,7 @@ gm200_gr_rops(struct gf100_gr *gr)
 	return nvkm_rd32(gr->base.engine.subdev.device, 0x12006c);
 }
 
-static void
+void
 gm200_gr_init_gpc_mmu(struct gf100_gr *gr)
 {
 	struct nvkm_device *device = gr->base.engine.subdev.device;
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c
new file mode 100644
index 000000000000..26ad79def0ff
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c
@@ -0,0 +1,171 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "gf100.h"
+#include "ctxgf100.h"
+
+#include <nvif/class.h>
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static void
+gp100_gr_init_rop_active_fbps(struct gf100_gr *gr)
+{
+	struct nvkm_device *device = gr->base.engine.subdev.device;
+	/*XXX: otherwise identical to gm200 aside from mask.. do everywhere? */
+	const u32 fbp_count = nvkm_rd32(device, 0x12006c) & 0x0000000f;
+	nvkm_mask(device, 0x408850, 0x0000000f, fbp_count); /* zrop */
+	nvkm_mask(device, 0x408958, 0x0000000f, fbp_count); /* crop */
+}
+
+static int
+gp100_gr_init(struct gf100_gr *gr)
+{
+	struct nvkm_device *device = gr->base.engine.subdev.device;
+	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, gr->tpc_total);
+	u32 data[TPC_MAX / 8] = {};
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc, rop;
+	int i;
+
+	gr->func->init_gpc_mmu(gr);
+
+	gf100_gr_mmio(gr, gr->fuc_sw_nonctx);
+
+	nvkm_wr32(device, GPC_UNIT(0, 0x3018), 0x00000001);
+
+	memset(data, 0x00, sizeof(data));
+	memcpy(tpcnr, gr->tpc_nr, sizeof(gr->tpc_nr));
+	for (i = 0, gpc = -1; i < gr->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % gr->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpc = gr->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		data[i / 8] |= tpc << ((i % 8) * 4);
+	}
+
+	nvkm_wr32(device, GPC_BCAST(0x0980), data[0]);
+	nvkm_wr32(device, GPC_BCAST(0x0984), data[1]);
+	nvkm_wr32(device, GPC_BCAST(0x0988), data[2]);
+	nvkm_wr32(device, GPC_BCAST(0x098c), data[3]);
+
+	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0914),
+			  gr->screen_tile_row_offset << 8 | gr->tpc_nr[gpc]);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0910), 0x00040000 |
+							 gr->tpc_total);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0918), magicgpc918);
+	}
+
+	nvkm_wr32(device, GPC_BCAST(0x3fd4), magicgpc918);
+	nvkm_wr32(device, GPC_BCAST(0x08ac), nvkm_rd32(device, 0x100800));
+	nvkm_wr32(device, GPC_BCAST(0x033c), nvkm_rd32(device, 0x100804));
+
+	gr->func->init_rop_active_fbps(gr);
+
+	nvkm_wr32(device, 0x400500, 0x00010001);
+	nvkm_wr32(device, 0x400100, 0xffffffff);
+	nvkm_wr32(device, 0x40013c, 0xffffffff);
+	nvkm_wr32(device, 0x400124, 0x00000002);
+	nvkm_wr32(device, 0x409c24, 0x000f0002);
+	nvkm_wr32(device, 0x405848, 0xc0000000);
+	nvkm_mask(device, 0x40584c, 0x00000000, 0x00000001);
+	nvkm_wr32(device, 0x404000, 0xc0000000);
+	nvkm_wr32(device, 0x404600, 0xc0000000);
+	nvkm_wr32(device, 0x408030, 0xc0000000);
+	nvkm_wr32(device, 0x404490, 0xc0000000);
+	nvkm_wr32(device, 0x406018, 0xc0000000);
+	nvkm_wr32(device, 0x407020, 0x40000000);
+	nvkm_wr32(device, 0x405840, 0xc0000000);
+	nvkm_wr32(device, 0x405844, 0x00ffffff);
+	nvkm_mask(device, 0x419cc0, 0x00000008, 0x00000008);
+
+	nvkm_mask(device, 0x419c9c, 0x00010000, 0x00010000);
+	nvkm_mask(device, 0x419c9c, 0x00020000, 0x00020000);
+
+	gr->func->init_ppc_exceptions(gr);
+
+	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+		for (tpc = 0; tpc < gr->tpc_nr[gpc]; tpc++) {
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x430), 0xc0000000);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x644), 0x00dffffe);
+			nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x64c), 0x00000105);
+		}
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
+		nvkm_wr32(device, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
+	}
+
+	for (rop = 0; rop < gr->rop_nr; rop++) {
+		nvkm_wr32(device, ROP_UNIT(rop, 0x144), 0x40000000);
+		nvkm_wr32(device, ROP_UNIT(rop, 0x070), 0x40000000);
+		nvkm_wr32(device, ROP_UNIT(rop, 0x204), 0xffffffff);
+		nvkm_wr32(device, ROP_UNIT(rop, 0x208), 0xffffffff);
+	}
+
+	nvkm_wr32(device, 0x400108, 0xffffffff);
+	nvkm_wr32(device, 0x400138, 0xffffffff);
+	nvkm_wr32(device, 0x400118, 0xffffffff);
+	nvkm_wr32(device, 0x400130, 0xffffffff);
+	nvkm_wr32(device, 0x40011c, 0xffffffff);
+	nvkm_wr32(device, 0x400134, 0xffffffff);
+
+	gf100_gr_zbc_init(gr);
+
+	return gf100_gr_init_ctxctl(gr);
+}
+
+static const struct gf100_gr_func
+gp100_gr = {
+	.init = gp100_gr_init,
+	.init_gpc_mmu = gm200_gr_init_gpc_mmu,
+	.init_rop_active_fbps = gp100_gr_init_rop_active_fbps,
+	.init_ppc_exceptions = gk104_gr_init_ppc_exceptions,
+	.rops = gm200_gr_rops,
+	.ppc_nr = 2,
+	.grctx = &gp100_grctx,
+	.sclass = {
+		{ -1, -1, FERMI_TWOD_A },
+		{ -1, -1, KEPLER_INLINE_TO_MEMORY_B },
+		{ -1, -1, PASCAL_A, &gf100_fermi },
+		{ -1, -1, PASCAL_COMPUTE_A },
+		{}
+	}
+};
+
+int
+gp100_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr)
+{
+	return gm200_gr_new_(&gp100_gr, device, index, pgr);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
index e15b9627b07e..f3c30b2a788e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c
@@ -26,6 +26,49 @@
 #include <subdev/bios.h>
 #include <subdev/bios/bmp.h>
 #include <subdev/bios/bit.h>
+#include <subdev/bios/image.h>
+
+static bool
+nvbios_addr(struct nvkm_bios *bios, u32 *addr, u8 size)
+{
+	u32 p = *addr;
+
+	if (*addr > bios->image0_size && bios->imaged_addr) {
+		*addr -= bios->image0_size;
+		*addr += bios->imaged_addr;
+	}
+
+	if (unlikely(*addr + size >= bios->size)) {
+		nvkm_error(&bios->subdev, "OOB %d %08x %08x\n", size, p, *addr);
+		return false;
+	}
+
+	return true;
+}
+
+u8
+nvbios_rd08(struct nvkm_bios *bios, u32 addr)
+{
+	if (likely(nvbios_addr(bios, &addr, 1)))
+		return bios->data[addr];
+	return 0x00;
+}
+
+u16
+nvbios_rd16(struct nvkm_bios *bios, u32 addr)
+{
+	if (likely(nvbios_addr(bios, &addr, 2)))
+		return get_unaligned_le16(&bios->data[addr]);
+	return 0x0000;
+}
+
+u32
+nvbios_rd32(struct nvkm_bios *bios, u32 addr)
+{
+	if (likely(nvbios_addr(bios, &addr, 4)))
+		return get_unaligned_le32(&bios->data[addr]);
+	return 0x00000000;
+}
 
 u8
 nvbios_checksum(const u8 *data, int size)
@@ -100,8 +143,9 @@ int
 nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios)
 {
 	struct nvkm_bios *bios;
+	struct nvbios_image image;
 	struct bit_entry bit_i;
-	int ret;
+	int ret, idx = 0;
 
 	if (!(bios = *pbios = kzalloc(sizeof(*bios), GFP_KERNEL)))
 		return -ENOMEM;
@@ -111,6 +155,19 @@ nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios)
 	if (ret)
 		return ret;
 
+	/* Some tables have weird pointers that need adjustment before
+	 * they're dereferenced.  I'm not entirely sure why...
+	 */
+	if (nvbios_image(bios, idx++, &image)) {
+		bios->image0_size = image.size;
+		while (nvbios_image(bios, idx++, &image)) {
+			if (image.type == 0xe0) {
+				bios->imaged_addr = image.base;
+				break;
+			}
+		}
+	}
+
 	/* detect type of vbios we're dealing with */
 	bios->bmp_offset = nvbios_findstr(bios->data, bios->size,
 					  "\xff\x7f""NV\0", 5);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c
index a5e92135cd77..9efb1b48cd54 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c
@@ -141,7 +141,8 @@ nvbios_ocfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx,
 {
 	u16 data = nvbios_ocfg_entry(bios, outp, idx, ver, hdr, cnt, len);
 	if (data) {
-		info->match     = nvbios_rd16(bios, data + 0x00);
+		info->proto     = nvbios_rd08(bios, data + 0x00);
+		info->flags     = nvbios_rd16(bios, data + 0x01);
 		info->clkcmp[0] = nvbios_rd16(bios, data + 0x02);
 		info->clkcmp[1] = nvbios_rd16(bios, data + 0x04);
 	}
@@ -149,12 +150,13 @@ nvbios_ocfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx,
 }
 
 u16
-nvbios_ocfg_match(struct nvkm_bios *bios, u16 outp, u16 type,
+nvbios_ocfg_match(struct nvkm_bios *bios, u16 outp, u8 proto, u8 flags,
 		  u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *info)
 {
 	u16 data, idx = 0;
 	while ((data = nvbios_ocfg_parse(bios, outp, idx++, ver, hdr, cnt, len, info))) {
-		if (info->match == type)
+		if ((info->proto == proto || info->proto == 0xff) &&
+		    (info->flags == flags))
 			break;
 	}
 	return data;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c
index 05332476354a..d89e78c4e689 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c
@@ -40,6 +40,7 @@ nvbios_dp_table(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 				case 0x30:
 				case 0x40:
 				case 0x41:
+				case 0x42:
 					*hdr = nvbios_rd08(bios, data + 0x01);
 					*len = nvbios_rd08(bios, data + 0x02);
 					*cnt = nvbios_rd08(bios, data + 0x03);
@@ -70,6 +71,7 @@ nvbios_dpout_entry(struct nvkm_bios *bios, u8 idx,
 			break;
 		case 0x40:
 		case 0x41:
+		case 0x42:
 			*hdr = nvbios_rd08(bios, data + 0x04);
 			*cnt = 0;
 			*len = 0;
@@ -109,6 +111,7 @@ nvbios_dpout_parse(struct nvkm_bios *bios, u8 idx,
 			break;
 		case 0x40:
 		case 0x41:
+		case 0x42:
 			info->flags     = nvbios_rd08(bios, data + 0x04);
 			info->script[0] = nvbios_rd16(bios, data + 0x05);
 			info->script[1] = nvbios_rd16(bios, data + 0x07);
@@ -180,6 +183,11 @@ nvbios_dpcfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx,
 			info->pe    = nvbios_rd08(bios, data + 0x02);
 			info->tx_pu = nvbios_rd08(bios, data + 0x03);
 			break;
+		case 0x42:
+			info->dc    = nvbios_rd08(bios, data + 0x00);
+			info->pe    = nvbios_rd08(bios, data + 0x01);
+			info->tx_pu = nvbios_rd08(bios, data + 0x02);
+			break;
 		default:
 			data = 0x0000;
 			break;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c
index 74b14cf09308..1dbff7aeafec 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c
@@ -68,11 +68,16 @@ nvbios_imagen(struct nvkm_bios *bios, struct nvbios_image *image)
 bool
 nvbios_image(struct nvkm_bios *bios, int idx, struct nvbios_image *image)
 {
+	u32 imaged_addr = bios->imaged_addr;
 	memset(image, 0x00, sizeof(*image));
+	bios->imaged_addr = 0;
 	do {
 		image->base += image->size;
-		if (image->last || !nvbios_imagen(bios, image))
+		if (image->last || !nvbios_imagen(bios, image)) {
+			bios->imaged_addr = imaged_addr;
 			return false;
+		}
 	} while(idx--);
+	bios->imaged_addr = imaged_addr;
 	return true;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c
index 91a7dc56e406..2ca23a9157ab 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c
@@ -77,15 +77,17 @@ g84_pll_mapping[] = {
 	{}
 };
 
-static u16
+static u32
 pll_limits_table(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 {
 	struct bit_entry bit_C;
-	u16 data = 0x0000;
+	u32 data = 0x0000;
 
 	if (!bit_entry(bios, 'C', &bit_C)) {
 		if (bit_C.version == 1 && bit_C.length >= 10)
 			data = nvbios_rd16(bios, bit_C.offset + 8);
+		if (bit_C.version == 2 && bit_C.length >= 4)
+			data = nvbios_rd32(bios, bit_C.offset + 0);
 		if (data) {
 			*ver = nvbios_rd08(bios, data + 0);
 			*hdr = nvbios_rd08(bios, data + 1);
@@ -137,12 +139,12 @@ pll_map(struct nvkm_bios *bios)
 	}
 }
 
-static u16
+static u32
 pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len)
 {
 	struct pll_mapping *map;
 	u8  hdr, cnt;
-	u16 data;
+	u32 data;
 
 	data = pll_limits_table(bios, ver, &hdr, &cnt, len);
 	if (data && *ver >= 0x30) {
@@ -160,7 +162,7 @@ pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len)
 	map = pll_map(bios);
 	while (map && map->reg) {
 		if (map->reg == reg && *ver >= 0x20) {
-			u16 addr = (data += hdr);
+			u32 addr = (data += hdr);
 			*type = map->type;
 			while (cnt--) {
 				if (nvbios_rd32(bios, data) == map->reg)
@@ -179,12 +181,12 @@ pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len)
 	return 0x0000;
 }
 
-static u16
+static u32
 pll_map_type(struct nvkm_bios *bios, u8 type, u32 *reg, u8 *ver, u8 *len)
 {
 	struct pll_mapping *map;
 	u8  hdr, cnt;
-	u16 data;
+	u32 data;
 
 	data = pll_limits_table(bios, ver, &hdr, &cnt, len);
 	if (data && *ver >= 0x30) {
@@ -202,7 +204,7 @@ pll_map_type(struct nvkm_bios *bios, u8 type, u32 *reg, u8 *ver, u8 *len)
 	map = pll_map(bios);
 	while (map && map->reg) {
 		if (map->type == type && *ver >= 0x20) {
-			u16 addr = (data += hdr);
+			u32 addr = (data += hdr);
 			*reg = map->reg;
 			while (cnt--) {
 				if (nvbios_rd32(bios, data) == map->reg)
@@ -228,7 +230,7 @@ nvbios_pll_parse(struct nvkm_bios *bios, u32 type, struct nvbios_pll *info)
 	struct nvkm_device *device = subdev->device;
 	u8  ver, len;
 	u32 reg = type;
-	u16 data;
+	u32 data;
 
 	if (type > PLL_MAX) {
 		reg  = type;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c
index c268e5afe852..b4a308f3cf7b 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c
@@ -26,21 +26,6 @@
 #include <subdev/bios/image.h>
 #include <subdev/bios/pmu.h>
 
-static u32
-weirdo_pointer(struct nvkm_bios *bios, u32 data)
-{
-	struct nvbios_image image;
-	int idx = 0;
-	if (nvbios_image(bios, idx++, &image)) {
-		data -= image.size;
-		while (nvbios_image(bios, idx++, &image)) {
-			if (image.type == 0xe0)
-				return image.base + data;
-		}
-	}
-	return 0;
-}
-
 u32
 nvbios_pmuTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 {
@@ -50,7 +35,7 @@ nvbios_pmuTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 	if (!bit_entry(bios, 'p', &bit_p)) {
 		if (bit_p.version == 2 && bit_p.length >= 4)
 			data = nvbios_rd32(bios, bit_p.offset + 0x00);
-		if ((data = weirdo_pointer(bios, data))) {
+		if (data) {
 			*ver = nvbios_rd08(bios, data + 0x00); /* maybe? */
 			*hdr = nvbios_rd08(bios, data + 0x01);
 			*len = nvbios_rd08(bios, data + 0x02);
@@ -97,8 +82,7 @@ nvbios_pmuRm(struct nvkm_bios *bios, u8 type, struct nvbios_pmuR *info)
 	u32 data;
 	memset(info, 0x00, sizeof(*info));
 	while ((data = nvbios_pmuEp(bios, idx++, &ver, &hdr, &pmuE))) {
-		if ( pmuE.type == type &&
-		    (data = weirdo_pointer(bios, pmuE.data))) {
+		if (pmuE.type == type && (data = pmuE.data)) {
 			info->init_addr_pmu = nvbios_rd32(bios, data + 0x08);
 			info->args_addr_pmu = nvbios_rd32(bios, data + 0x0c);
 			info->boot_addr     = data + 0x30;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c
index d0ae7454764e..b57c370c725d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c
@@ -30,11 +30,11 @@ nvbios_rammapTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr,
 		u8 *cnt, u8 *len, u8 *snr, u8 *ssz)
 {
 	struct bit_entry bit_P;
-	u16 rammap = 0x0000;
+	u32 rammap = 0x0000;
 
 	if (!bit_entry(bios, 'P', &bit_P)) {
 		if (bit_P.version == 2)
-			rammap = nvbios_rd16(bios, bit_P.offset + 4);
+			rammap = nvbios_rd32(bios, bit_P.offset + 4);
 
 		if (rammap) {
 			*ver = nvbios_rd08(bios, rammap + 0);
@@ -61,7 +61,7 @@ nvbios_rammapEe(struct nvkm_bios *bios, int idx,
 		u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 {
 	u8  snr, ssz;
-	u16 rammap = nvbios_rammapTe(bios, ver, hdr, cnt, len, &snr, &ssz);
+	u32 rammap = nvbios_rammapTe(bios, ver, hdr, cnt, len, &snr, &ssz);
 	if (rammap && idx < *cnt) {
 		rammap = rammap + *hdr + (idx * (*len + (snr * ssz)));
 		*hdr = *len;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
index 78c449b417b7..89d5543118cf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c
@@ -99,7 +99,7 @@ read_div(struct gf100_clk *clk, int doff, u32 dsrc, u32 dctl)
 {
 	struct nvkm_device *device = clk->base.subdev.device;
 	u32 ssrc = nvkm_rd32(device, dsrc + (doff * 4));
-	u32 sctl = nvkm_rd32(device, dctl + (doff * 4));
+	u32 sclk, sctl, sdiv = 2;
 
 	switch (ssrc & 0x00000003) {
 	case 0:
@@ -109,13 +109,21 @@ read_div(struct gf100_clk *clk, int doff, u32 dsrc, u32 dctl)
 	case 2:
 		return 100000;
 	case 3:
-		if (sctl & 0x80000000) {
-			u32 sclk = read_vco(clk, dsrc + (doff * 4));
-			u32 sdiv = (sctl & 0x0000003f) + 2;
-			return (sclk * 2) / sdiv;
+		sclk = read_vco(clk, dsrc + (doff * 4));
+
+		/* Memclk has doff of 0 despite its alt. location */
+		if (doff <= 2) {
+			sctl = nvkm_rd32(device, dctl + (doff * 4));
+
+			if (sctl & 0x80000000) {
+				if (ssrc & 0x100)
+					sctl >>= 8;
+
+				sdiv = (sctl & 0x3f) + 2;
+			}
 		}
 
-		return read_vco(clk, dsrc + (doff * 4));
+		return (sclk * 2) / sdiv;
 	default:
 		return 0;
 	}
@@ -366,11 +374,17 @@ gf100_clk_prog_2(struct gf100_clk *clk, int idx)
 		if (info->coef) {
 			nvkm_wr32(device, addr + 0x04, info->coef);
 			nvkm_mask(device, addr + 0x00, 0x00000001, 0x00000001);
+
+			/* Test PLL lock */
+			nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000000);
 			nvkm_msec(device, 2000,
 				if (nvkm_rd32(device, addr + 0x00) & 0x00020000)
 					break;
 			);
-			nvkm_mask(device, addr + 0x00, 0x00020004, 0x00000004);
+			nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000010);
+
+			/* Enable sync mode */
+			nvkm_mask(device, addr + 0x00, 0x00000004, 0x00000004);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
index 975c401bccab..06bc0d2d6ae1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c
@@ -393,11 +393,17 @@ gk104_clk_prog_2(struct gk104_clk *clk, int idx)
 	if (info->coef) {
 		nvkm_wr32(device, addr + 0x04, info->coef);
 		nvkm_mask(device, addr + 0x00, 0x00000001, 0x00000001);
+
+		/* Test PLL lock */
+		nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000000);
 		nvkm_msec(device, 2000,
 			if (nvkm_rd32(device, addr + 0x00) & 0x00020000)
 				break;
 		);
-		nvkm_mask(device, addr + 0x00, 0x00020004, 0x00000004);
+		nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000010);
+
+		/* Enable sync mode */
+		nvkm_mask(device, addr + 0x00, 0x00000004, 0x00000004);
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
index 5f0ee24e31b8..218893e3e5f9 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
@@ -28,69 +28,6 @@
 #include <core/tegra.h>
 #include <subdev/timer.h>
 
-#define KHZ (1000)
-#define MHZ (KHZ * 1000)
-
-#define MASK(w)	((1 << w) - 1)
-
-#define GPCPLL_CFG		(SYS_GPCPLL_CFG_BASE + 0)
-#define GPCPLL_CFG_ENABLE	BIT(0)
-#define GPCPLL_CFG_IDDQ		BIT(1)
-#define GPCPLL_CFG_LOCK_DET_OFF	BIT(4)
-#define GPCPLL_CFG_LOCK		BIT(17)
-
-#define GPCPLL_COEFF		(SYS_GPCPLL_CFG_BASE + 4)
-#define GPCPLL_COEFF_M_SHIFT	0
-#define GPCPLL_COEFF_M_WIDTH	8
-#define GPCPLL_COEFF_N_SHIFT	8
-#define GPCPLL_COEFF_N_WIDTH	8
-#define GPCPLL_COEFF_P_SHIFT	16
-#define GPCPLL_COEFF_P_WIDTH	6
-
-#define GPCPLL_CFG2			(SYS_GPCPLL_CFG_BASE + 0xc)
-#define GPCPLL_CFG2_SETUP2_SHIFT	16
-#define GPCPLL_CFG2_PLL_STEPA_SHIFT	24
-
-#define GPCPLL_CFG3			(SYS_GPCPLL_CFG_BASE + 0x18)
-#define GPCPLL_CFG3_PLL_STEPB_SHIFT	16
-
-#define GPC_BCASE_GPCPLL_CFG_BASE		0x00132800
-#define GPCPLL_NDIV_SLOWDOWN			(SYS_GPCPLL_CFG_BASE + 0x1c)
-#define GPCPLL_NDIV_SLOWDOWN_NDIV_LO_SHIFT	0
-#define GPCPLL_NDIV_SLOWDOWN_NDIV_MID_SHIFT	8
-#define GPCPLL_NDIV_SLOWDOWN_STEP_SIZE_LO2MID_SHIFT	16
-#define GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT	22
-#define GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT	31
-
-#define SEL_VCO				(SYS_GPCPLL_CFG_BASE + 0x100)
-#define SEL_VCO_GPC2CLK_OUT_SHIFT	0
-
-#define GPC2CLK_OUT			(SYS_GPCPLL_CFG_BASE + 0x250)
-#define GPC2CLK_OUT_SDIV14_INDIV4_WIDTH	1
-#define GPC2CLK_OUT_SDIV14_INDIV4_SHIFT	31
-#define GPC2CLK_OUT_SDIV14_INDIV4_MODE	1
-#define GPC2CLK_OUT_VCODIV_WIDTH	6
-#define GPC2CLK_OUT_VCODIV_SHIFT	8
-#define GPC2CLK_OUT_VCODIV1		0
-#define GPC2CLK_OUT_VCODIV_MASK		(MASK(GPC2CLK_OUT_VCODIV_WIDTH) << \
-					GPC2CLK_OUT_VCODIV_SHIFT)
-#define GPC2CLK_OUT_BYPDIV_WIDTH	6
-#define GPC2CLK_OUT_BYPDIV_SHIFT	0
-#define GPC2CLK_OUT_BYPDIV31		0x3c
-#define GPC2CLK_OUT_INIT_MASK	((MASK(GPC2CLK_OUT_SDIV14_INDIV4_WIDTH) << \
-		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT)\
-		| (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << GPC2CLK_OUT_VCODIV_SHIFT)\
-		| (MASK(GPC2CLK_OUT_BYPDIV_WIDTH) << GPC2CLK_OUT_BYPDIV_SHIFT))
-#define GPC2CLK_OUT_INIT_VAL	((GPC2CLK_OUT_SDIV14_INDIV4_MODE << \
-		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT) \
-		| (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT) \
-		| (GPC2CLK_OUT_BYPDIV31 << GPC2CLK_OUT_BYPDIV_SHIFT))
-
-#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG	(GPC_BCASE_GPCPLL_CFG_BASE + 0xa0)
-#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT	24
-#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK \
-	    (0x1 << GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT)
-
 static const u8 _pl_to_div[] = {
 /* PL:   0, 1, 2, 3, 4, 5, 6,  7,  8,  9, 10, 11, 12, 13, 14 */
 /* p: */ 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 12, 16, 20, 24, 32,
@@ -124,7 +61,7 @@ static const struct gk20a_clk_pllg_params gk20a_pllg_params = {
 	.min_pl = 1, .max_pl = 32,
 };
 
-static void
+void
 gk20a_pllg_read_mnp(struct gk20a_clk *clk, struct gk20a_pll *pll)
 {
 	struct nvkm_device *device = clk->base.subdev.device;
@@ -136,20 +73,33 @@ gk20a_pllg_read_mnp(struct gk20a_clk *clk, struct gk20a_pll *pll)
 	pll->pl = (val >> GPCPLL_COEFF_P_SHIFT) & MASK(GPCPLL_COEFF_P_WIDTH);
 }
 
-static u32
-gk20a_pllg_calc_rate(struct gk20a_clk *clk)
+void
+gk20a_pllg_write_mnp(struct gk20a_clk *clk, const struct gk20a_pll *pll)
+{
+	struct nvkm_device *device = clk->base.subdev.device;
+	u32 val;
+
+	val = (pll->m & MASK(GPCPLL_COEFF_M_WIDTH)) << GPCPLL_COEFF_M_SHIFT;
+	val |= (pll->n & MASK(GPCPLL_COEFF_N_WIDTH)) << GPCPLL_COEFF_N_SHIFT;
+	val |= (pll->pl & MASK(GPCPLL_COEFF_P_WIDTH)) << GPCPLL_COEFF_P_SHIFT;
+	nvkm_wr32(device, GPCPLL_COEFF, val);
+}
+
+u32
+gk20a_pllg_calc_rate(struct gk20a_clk *clk, struct gk20a_pll *pll)
 {
 	u32 rate;
 	u32 divider;
 
-	rate = clk->parent_rate * clk->pll.n;
-	divider = clk->pll.m * clk->pl_to_div(clk->pll.pl);
+	rate = clk->parent_rate * pll->n;
+	divider = pll->m * clk->pl_to_div(pll->pl);
 
 	return rate / divider / 2;
 }
 
-static int
-gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate)
+int
+gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate,
+		    struct gk20a_pll *pll)
 {
 	struct nvkm_subdev *subdev = &clk->base.subdev;
 	u32 target_clk_f, ref_clk_f, target_freq;
@@ -163,16 +113,13 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate)
 	target_clk_f = rate * 2 / KHZ;
 	ref_clk_f = clk->parent_rate / KHZ;
 
-	max_vco_f = clk->params->max_vco;
+	target_vco_f = target_clk_f + target_clk_f / 50;
+	max_vco_f = max(clk->params->max_vco, target_vco_f);
 	min_vco_f = clk->params->min_vco;
 	best_m = clk->params->max_m;
 	best_n = clk->params->min_n;
 	best_pl = clk->params->min_pl;
 
-	target_vco_f = target_clk_f + target_clk_f / 50;
-	if (max_vco_f < target_vco_f)
-		max_vco_f = target_vco_f;
-
 	/* min_pl <= high_pl <= max_pl */
 	high_pl = (max_vco_f + target_vco_f - 1) / target_vco_f;
 	high_pl = min(high_pl, clk->params->max_pl);
@@ -195,9 +142,7 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate)
 		target_vco_f = target_clk_f * clk->pl_to_div(pl);
 
 		for (m = clk->params->min_m; m <= clk->params->max_m; m++) {
-			u32 u_f, vco_f;
-
-			u_f = ref_clk_f / m;
+			u32 u_f = ref_clk_f / m;
 
 			if (u_f < clk->params->min_u)
 				break;
@@ -211,6 +156,8 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate)
 				break;
 
 			for (; n <= n2; n++) {
+				u32 vco_f;
+
 				if (n < clk->params->min_n)
 					continue;
 				if (n > clk->params->max_n)
@@ -247,16 +194,16 @@ found_match:
 			   "no best match for target @ %dMHz on gpc_pll",
 			   target_clk_f / KHZ);
 
-	clk->pll.m = best_m;
-	clk->pll.n = best_n;
-	clk->pll.pl = best_pl;
+	pll->m = best_m;
+	pll->n = best_n;
+	pll->pl = best_pl;
 
-	target_freq = gk20a_pllg_calc_rate(clk);
+	target_freq = gk20a_pllg_calc_rate(clk, pll);
 
 	nvkm_debug(subdev,
-		   "actual target freq %d MHz, M %d, N %d, PL %d(div%d)\n",
-		   target_freq / MHZ, clk->pll.m, clk->pll.n, clk->pll.pl,
-		   clk->pl_to_div(clk->pll.pl));
+		   "actual target freq %d KHz, M %d, N %d, PL %d(div%d)\n",
+		   target_freq / KHZ, pll->m, pll->n, pll->pl,
+		   clk->pl_to_div(pll->pl));
 	return 0;
 }
 
@@ -265,45 +212,36 @@ gk20a_pllg_slide(struct gk20a_clk *clk, u32 n)
 {
 	struct nvkm_subdev *subdev = &clk->base.subdev;
 	struct nvkm_device *device = subdev->device;
-	u32 val;
-	int ramp_timeout;
+	struct gk20a_pll pll;
+	int ret = 0;
 
 	/* get old coefficients */
-	val = nvkm_rd32(device, GPCPLL_COEFF);
+	gk20a_pllg_read_mnp(clk, &pll);
 	/* do nothing if NDIV is the same */
-	if (n == ((val >> GPCPLL_COEFF_N_SHIFT) & MASK(GPCPLL_COEFF_N_WIDTH)))
+	if (n == pll.n)
 		return 0;
 
-	/* setup */
-	nvkm_mask(device, GPCPLL_CFG2, 0xff << GPCPLL_CFG2_PLL_STEPA_SHIFT,
-		0x2b << GPCPLL_CFG2_PLL_STEPA_SHIFT);
-	nvkm_mask(device, GPCPLL_CFG3, 0xff << GPCPLL_CFG3_PLL_STEPB_SHIFT,
-		0xb << GPCPLL_CFG3_PLL_STEPB_SHIFT);
-
 	/* pll slowdown mode */
 	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
 		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT),
 		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT));
 
 	/* new ndiv ready for ramp */
-	val = nvkm_rd32(device, GPCPLL_COEFF);
-	val &= ~(MASK(GPCPLL_COEFF_N_WIDTH) << GPCPLL_COEFF_N_SHIFT);
-	val |= (n & MASK(GPCPLL_COEFF_N_WIDTH)) << GPCPLL_COEFF_N_SHIFT;
+	pll.n = n;
 	udelay(1);
-	nvkm_wr32(device, GPCPLL_COEFF, val);
+	gk20a_pllg_write_mnp(clk, &pll);
 
 	/* dynamic ramp to new ndiv */
-	val = nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN);
-	val |= 0x1 << GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT;
 	udelay(1);
-	nvkm_wr32(device, GPCPLL_NDIV_SLOWDOWN, val);
+	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
+		  BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT),
+		  BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT));
 
-	for (ramp_timeout = 500; ramp_timeout > 0; ramp_timeout--) {
-		udelay(1);
-		val = nvkm_rd32(device, GPC_BCAST_NDIV_SLOWDOWN_DEBUG);
-		if (val & GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK)
-			break;
-	}
+	/* wait for ramping to complete */
+	if (nvkm_wait_usec(device, 500, GPC_BCAST_NDIV_SLOWDOWN_DEBUG,
+		GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK,
+		GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK) < 0)
+		ret = -ETIMEDOUT;
 
 	/* exit slowdown mode */
 	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
@@ -311,21 +249,35 @@ gk20a_pllg_slide(struct gk20a_clk *clk, u32 n)
 		BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), 0);
 	nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN);
 
-	if (ramp_timeout <= 0) {
-		nvkm_error(subdev, "gpcpll dynamic ramp timeout\n");
-		return -ETIMEDOUT;
-	}
-
-	return 0;
+	return ret;
 }
 
-static void
+static int
 gk20a_pllg_enable(struct gk20a_clk *clk)
 {
 	struct nvkm_device *device = clk->base.subdev.device;
+	u32 val;
 
 	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, GPCPLL_CFG_ENABLE);
 	nvkm_rd32(device, GPCPLL_CFG);
+
+	/* enable lock detection */
+	val = nvkm_rd32(device, GPCPLL_CFG);
+	if (val & GPCPLL_CFG_LOCK_DET_OFF) {
+		val &= ~GPCPLL_CFG_LOCK_DET_OFF;
+		nvkm_wr32(device, GPCPLL_CFG, val);
+	}
+
+	/* wait for lock */
+	if (nvkm_wait_usec(device, 300, GPCPLL_CFG, GPCPLL_CFG_LOCK,
+			   GPCPLL_CFG_LOCK) < 0)
+		return -ETIMEDOUT;
+
+	/* switch to VCO mode */
+	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT),
+		BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+
+	return 0;
 }
 
 static void
@@ -333,117 +285,81 @@ gk20a_pllg_disable(struct gk20a_clk *clk)
 {
 	struct nvkm_device *device = clk->base.subdev.device;
 
+	/* put PLL in bypass before disabling it */
+	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0);
+
 	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, 0);
 	nvkm_rd32(device, GPCPLL_CFG);
 }
 
 static int
-_gk20a_pllg_program_mnp(struct gk20a_clk *clk, bool allow_slide)
+gk20a_pllg_program_mnp(struct gk20a_clk *clk, const struct gk20a_pll *pll)
 {
 	struct nvkm_subdev *subdev = &clk->base.subdev;
 	struct nvkm_device *device = subdev->device;
-	u32 val, cfg;
-	struct gk20a_pll old_pll;
-	u32 n_lo;
-
-	/* get old coefficients */
-	gk20a_pllg_read_mnp(clk, &old_pll);
-
-	/* do NDIV slide if there is no change in M and PL */
-	cfg = nvkm_rd32(device, GPCPLL_CFG);
-	if (allow_slide && clk->pll.m == old_pll.m &&
-	    clk->pll.pl == old_pll.pl && (cfg & GPCPLL_CFG_ENABLE)) {
-		return gk20a_pllg_slide(clk, clk->pll.n);
-	}
-
-	/* slide down to NDIV_LO */
-	if (allow_slide && (cfg & GPCPLL_CFG_ENABLE)) {
-		int ret;
-
-		n_lo = DIV_ROUND_UP(old_pll.m * clk->params->min_vco,
-				    clk->parent_rate / KHZ);
-		ret = gk20a_pllg_slide(clk, n_lo);
+	struct gk20a_pll cur_pll;
+	int ret;
 
-		if (ret)
-			return ret;
-	}
+	gk20a_pllg_read_mnp(clk, &cur_pll);
 
-	/* split FO-to-bypass jump in halfs by setting out divider 1:2 */
+	/* split VCO-to-bypass jump in half by setting out divider 1:2 */
 	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
-		0x2 << GPC2CLK_OUT_VCODIV_SHIFT);
-
-	/* put PLL in bypass before programming it */
-	val = nvkm_rd32(device, SEL_VCO);
-	val &= ~(BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+		  GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT);
+	/* Intentional 2nd write to assure linear divider operation */
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT);
+	nvkm_rd32(device, GPC2CLK_OUT);
 	udelay(2);
-	nvkm_wr32(device, SEL_VCO, val);
-
-	/* get out from IDDQ */
-	val = nvkm_rd32(device, GPCPLL_CFG);
-	if (val & GPCPLL_CFG_IDDQ) {
-		val &= ~GPCPLL_CFG_IDDQ;
-		nvkm_wr32(device, GPCPLL_CFG, val);
-		nvkm_rd32(device, GPCPLL_CFG);
-		udelay(2);
-	}
 
 	gk20a_pllg_disable(clk);
 
-	nvkm_debug(subdev, "%s: m=%d n=%d pl=%d\n", __func__,
-		   clk->pll.m, clk->pll.n, clk->pll.pl);
-
-	n_lo = DIV_ROUND_UP(clk->pll.m * clk->params->min_vco,
-			    clk->parent_rate / KHZ);
-	val = clk->pll.m << GPCPLL_COEFF_M_SHIFT;
-	val |= (allow_slide ? n_lo : clk->pll.n) << GPCPLL_COEFF_N_SHIFT;
-	val |= clk->pll.pl << GPCPLL_COEFF_P_SHIFT;
-	nvkm_wr32(device, GPCPLL_COEFF, val);
+	gk20a_pllg_write_mnp(clk, pll);
 
-	gk20a_pllg_enable(clk);
-
-	val = nvkm_rd32(device, GPCPLL_CFG);
-	if (val & GPCPLL_CFG_LOCK_DET_OFF) {
-		val &= ~GPCPLL_CFG_LOCK_DET_OFF;
-		nvkm_wr32(device, GPCPLL_CFG, val);
-	}
-
-	if (nvkm_usec(device, 300,
-		if (nvkm_rd32(device, GPCPLL_CFG) & GPCPLL_CFG_LOCK)
-			break;
-	) < 0)
-		return -ETIMEDOUT;
-
-	/* switch to VCO mode */
-	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT),
-		  BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+	ret = gk20a_pllg_enable(clk);
+	if (ret)
+		return ret;
 
 	/* restore out divider 1:1 */
-	val = nvkm_rd32(device, GPC2CLK_OUT);
-	if ((val & GPC2CLK_OUT_VCODIV_MASK) !=
-	    (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT)) {
-		val &= ~GPC2CLK_OUT_VCODIV_MASK;
-		val |= GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT;
-		udelay(2);
-		nvkm_wr32(device, GPC2CLK_OUT, val);
-		/* Intentional 2nd write to assure linear divider operation */
-		nvkm_wr32(device, GPC2CLK_OUT, val);
-		nvkm_rd32(device, GPC2CLK_OUT);
-	}
+	udelay(2);
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT);
+	/* Intentional 2nd write to assure linear divider operation */
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT);
+	nvkm_rd32(device, GPC2CLK_OUT);
 
-	/* slide up to new NDIV */
-	return allow_slide ? gk20a_pllg_slide(clk, clk->pll.n) : 0;
+	return 0;
 }
 
 static int
-gk20a_pllg_program_mnp(struct gk20a_clk *clk)
+gk20a_pllg_program_mnp_slide(struct gk20a_clk *clk, const struct gk20a_pll *pll)
 {
-	int err;
+	struct gk20a_pll cur_pll;
+	int ret;
 
-	err = _gk20a_pllg_program_mnp(clk, true);
-	if (err)
-		err = _gk20a_pllg_program_mnp(clk, false);
+	if (gk20a_pllg_is_enabled(clk)) {
+		gk20a_pllg_read_mnp(clk, &cur_pll);
+
+		/* just do NDIV slide if there is no change to M and PL */
+		if (pll->m == cur_pll.m && pll->pl == cur_pll.pl)
+			return gk20a_pllg_slide(clk, pll->n);
+
+		/* slide down to current NDIV_LO */
+		cur_pll.n = gk20a_pllg_n_lo(clk, &cur_pll);
+		ret = gk20a_pllg_slide(clk, cur_pll.n);
+		if (ret)
+			return ret;
+	}
+
+	/* program MNP with the new clock parameters and new NDIV_LO */
+	cur_pll = *pll;
+	cur_pll.n = gk20a_pllg_n_lo(clk, &cur_pll);
+	ret = gk20a_pllg_program_mnp(clk, &cur_pll);
+	if (ret)
+		return ret;
 
-	return err;
+	/* slide up to new NDIV */
+	return gk20a_pllg_slide(clk, pll->n);
 }
 
 static struct nvkm_pstate
@@ -546,13 +462,14 @@ gk20a_clk_read(struct nvkm_clk *base, enum nv_clk_src src)
 	struct gk20a_clk *clk = gk20a_clk(base);
 	struct nvkm_subdev *subdev = &clk->base.subdev;
 	struct nvkm_device *device = subdev->device;
+	struct gk20a_pll pll;
 
 	switch (src) {
 	case nv_clk_src_crystal:
 		return device->crystal;
 	case nv_clk_src_gpc:
-		gk20a_pllg_read_mnp(clk, &clk->pll);
-		return gk20a_pllg_calc_rate(clk) / GK20A_CLK_GPC_MDIV;
+		gk20a_pllg_read_mnp(clk, &pll);
+		return gk20a_pllg_calc_rate(clk, &pll) / GK20A_CLK_GPC_MDIV;
 	default:
 		nvkm_error(subdev, "invalid clock source %d\n", src);
 		return -EINVAL;
@@ -565,15 +482,20 @@ gk20a_clk_calc(struct nvkm_clk *base, struct nvkm_cstate *cstate)
 	struct gk20a_clk *clk = gk20a_clk(base);
 
 	return gk20a_pllg_calc_mnp(clk, cstate->domain[nv_clk_src_gpc] *
-					 GK20A_CLK_GPC_MDIV);
+					 GK20A_CLK_GPC_MDIV, &clk->pll);
 }
 
 int
 gk20a_clk_prog(struct nvkm_clk *base)
 {
 	struct gk20a_clk *clk = gk20a_clk(base);
+	int ret;
+
+	ret = gk20a_pllg_program_mnp_slide(clk, &clk->pll);
+	if (ret)
+		ret = gk20a_pllg_program_mnp(clk, &clk->pll);
 
-	return gk20a_pllg_program_mnp(clk);
+	return ret;
 }
 
 void
@@ -581,29 +503,62 @@ gk20a_clk_tidy(struct nvkm_clk *base)
 {
 }
 
+int
+gk20a_clk_setup_slide(struct gk20a_clk *clk)
+{
+	struct nvkm_subdev *subdev = &clk->base.subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 step_a, step_b;
+
+	switch (clk->parent_rate) {
+	case 12000000:
+	case 12800000:
+	case 13000000:
+		step_a = 0x2b;
+		step_b = 0x0b;
+		break;
+	case 19200000:
+		step_a = 0x12;
+		step_b = 0x08;
+		break;
+	case 38400000:
+		step_a = 0x04;
+		step_b = 0x05;
+		break;
+	default:
+		nvkm_error(subdev, "invalid parent clock rate %u KHz",
+			   clk->parent_rate / KHZ);
+		return -EINVAL;
+	}
+
+	nvkm_mask(device, GPCPLL_CFG2, 0xff << GPCPLL_CFG2_PLL_STEPA_SHIFT,
+		step_a << GPCPLL_CFG2_PLL_STEPA_SHIFT);
+	nvkm_mask(device, GPCPLL_CFG3, 0xff << GPCPLL_CFG3_PLL_STEPB_SHIFT,
+		step_b << GPCPLL_CFG3_PLL_STEPB_SHIFT);
+
+	return 0;
+}
+
 void
 gk20a_clk_fini(struct nvkm_clk *base)
 {
 	struct nvkm_device *device = base->subdev.device;
 	struct gk20a_clk *clk = gk20a_clk(base);
-	u32 val;
 
 	/* slide to VCO min */
-	val = nvkm_rd32(device, GPCPLL_CFG);
-	if (val & GPCPLL_CFG_ENABLE) {
+	if (gk20a_pllg_is_enabled(clk)) {
 		struct gk20a_pll pll;
 		u32 n_lo;
 
 		gk20a_pllg_read_mnp(clk, &pll);
-		n_lo = DIV_ROUND_UP(pll.m * clk->params->min_vco,
-				    clk->parent_rate / KHZ);
+		n_lo = gk20a_pllg_n_lo(clk, &pll);
 		gk20a_pllg_slide(clk, n_lo);
 	}
 
-	/* put PLL in bypass before disabling it */
-	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0);
-
 	gk20a_pllg_disable(clk);
+
+	/* set IDDQ */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 1);
 }
 
 static int
@@ -614,9 +569,18 @@ gk20a_clk_init(struct nvkm_clk *base)
 	struct nvkm_device *device = subdev->device;
 	int ret;
 
+	/* get out from IDDQ */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 0);
+	nvkm_rd32(device, GPCPLL_CFG);
+	udelay(5);
+
 	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_INIT_MASK,
 		  GPC2CLK_OUT_INIT_VAL);
 
+	ret = gk20a_clk_setup_slide(clk);
+	if (ret)
+		return ret;
+
 	/* Start with lowest frequency */
 	base->func->calc(base, &base->func->pstates[0].base);
 	ret = base->func->prog(&clk->base);
@@ -646,7 +610,7 @@ gk20a_clk = {
 };
 
 int
-_gk20a_clk_ctor(struct nvkm_device *device, int index,
+gk20a_clk_ctor(struct nvkm_device *device, int index,
 		const struct nvkm_clk_func *func,
 		const struct gk20a_clk_pllg_params *params,
 		struct gk20a_clk *clk)
@@ -685,7 +649,7 @@ gk20a_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	ret = _gk20a_clk_ctor(device, index, &gk20a_clk, &gk20a_pllg_params,
+	ret = gk20a_clk_ctor(device, index, &gk20a_clk, &gk20a_pllg_params,
 			      clk);
 
 	clk->pl_to_div = pl_to_div;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
index 13c46740197d..0d1450972162 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h
@@ -24,9 +24,79 @@
 #ifndef __NVKM_CLK_GK20A_H__
 #define __NVKM_CLK_GK20A_H__
 
+#define KHZ (1000)
+#define MHZ (KHZ * 1000)
+
+#define MASK(w)	((1 << (w)) - 1)
+
 #define GK20A_CLK_GPC_MDIV 1000
 
 #define SYS_GPCPLL_CFG_BASE	0x00137000
+#define GPCPLL_CFG		(SYS_GPCPLL_CFG_BASE + 0)
+#define GPCPLL_CFG_ENABLE	BIT(0)
+#define GPCPLL_CFG_IDDQ		BIT(1)
+#define GPCPLL_CFG_LOCK_DET_OFF	BIT(4)
+#define GPCPLL_CFG_LOCK		BIT(17)
+
+#define GPCPLL_CFG2		(SYS_GPCPLL_CFG_BASE + 0xc)
+#define GPCPLL_CFG2_SETUP2_SHIFT	16
+#define GPCPLL_CFG2_PLL_STEPA_SHIFT	24
+
+#define GPCPLL_CFG3			(SYS_GPCPLL_CFG_BASE + 0x18)
+#define GPCPLL_CFG3_VCO_CTRL_SHIFT		0
+#define GPCPLL_CFG3_VCO_CTRL_WIDTH		9
+#define GPCPLL_CFG3_VCO_CTRL_MASK		\
+	(MASK(GPCPLL_CFG3_VCO_CTRL_WIDTH) << GPCPLL_CFG3_VCO_CTRL_SHIFT)
+#define GPCPLL_CFG3_PLL_STEPB_SHIFT		16
+#define GPCPLL_CFG3_PLL_STEPB_WIDTH		8
+
+#define GPCPLL_COEFF		(SYS_GPCPLL_CFG_BASE + 4)
+#define GPCPLL_COEFF_M_SHIFT	0
+#define GPCPLL_COEFF_M_WIDTH	8
+#define GPCPLL_COEFF_N_SHIFT	8
+#define GPCPLL_COEFF_N_WIDTH	8
+#define GPCPLL_COEFF_N_MASK	\
+	(MASK(GPCPLL_COEFF_N_WIDTH) << GPCPLL_COEFF_N_SHIFT)
+#define GPCPLL_COEFF_P_SHIFT	16
+#define GPCPLL_COEFF_P_WIDTH	6
+
+#define GPCPLL_NDIV_SLOWDOWN			(SYS_GPCPLL_CFG_BASE + 0x1c)
+#define GPCPLL_NDIV_SLOWDOWN_NDIV_LO_SHIFT	0
+#define GPCPLL_NDIV_SLOWDOWN_NDIV_MID_SHIFT	8
+#define GPCPLL_NDIV_SLOWDOWN_STEP_SIZE_LO2MID_SHIFT	16
+#define GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT	22
+#define GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT	31
+
+#define GPC_BCAST_GPCPLL_CFG_BASE		0x00132800
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG	(GPC_BCAST_GPCPLL_CFG_BASE + 0xa0)
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT	24
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK \
+	(0x1 << GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT)
+
+#define SEL_VCO				(SYS_GPCPLL_CFG_BASE + 0x100)
+#define SEL_VCO_GPC2CLK_OUT_SHIFT	0
+
+#define GPC2CLK_OUT			(SYS_GPCPLL_CFG_BASE + 0x250)
+#define GPC2CLK_OUT_SDIV14_INDIV4_WIDTH	1
+#define GPC2CLK_OUT_SDIV14_INDIV4_SHIFT	31
+#define GPC2CLK_OUT_SDIV14_INDIV4_MODE	1
+#define GPC2CLK_OUT_VCODIV_WIDTH	6
+#define GPC2CLK_OUT_VCODIV_SHIFT	8
+#define GPC2CLK_OUT_VCODIV1		0
+#define GPC2CLK_OUT_VCODIV2		2
+#define GPC2CLK_OUT_VCODIV_MASK		(MASK(GPC2CLK_OUT_VCODIV_WIDTH) << \
+					GPC2CLK_OUT_VCODIV_SHIFT)
+#define GPC2CLK_OUT_BYPDIV_WIDTH	6
+#define GPC2CLK_OUT_BYPDIV_SHIFT	0
+#define GPC2CLK_OUT_BYPDIV31		0x3c
+#define GPC2CLK_OUT_INIT_MASK	((MASK(GPC2CLK_OUT_SDIV14_INDIV4_WIDTH) << \
+		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT)\
+		| (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << GPC2CLK_OUT_VCODIV_SHIFT)\
+		| (MASK(GPC2CLK_OUT_BYPDIV_WIDTH) << GPC2CLK_OUT_BYPDIV_SHIFT))
+#define GPC2CLK_OUT_INIT_VAL	((GPC2CLK_OUT_SDIV14_INDIV4_MODE << \
+		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT) \
+		| (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT) \
+		| (GPC2CLK_OUT_BYPDIV31 << GPC2CLK_OUT_BYPDIV_SHIFT))
 
 /* All frequencies in Khz */
 struct gk20a_clk_pllg_params {
@@ -54,7 +124,29 @@ struct gk20a_clk {
 };
 #define gk20a_clk(p) container_of((p), struct gk20a_clk, base)
 
-int _gk20a_clk_ctor(struct nvkm_device *, int, const struct nvkm_clk_func *,
+u32 gk20a_pllg_calc_rate(struct gk20a_clk *, struct gk20a_pll *);
+int gk20a_pllg_calc_mnp(struct gk20a_clk *, unsigned long, struct gk20a_pll *);
+void gk20a_pllg_read_mnp(struct gk20a_clk *, struct gk20a_pll *);
+void gk20a_pllg_write_mnp(struct gk20a_clk *, const struct gk20a_pll *);
+
+static inline bool
+gk20a_pllg_is_enabled(struct gk20a_clk *clk)
+{
+	struct nvkm_device *device = clk->base.subdev.device;
+	u32 val;
+
+	val = nvkm_rd32(device, GPCPLL_CFG);
+	return val & GPCPLL_CFG_ENABLE;
+}
+
+static inline u32
+gk20a_pllg_n_lo(struct gk20a_clk *clk, struct gk20a_pll *pll)
+{
+	return DIV_ROUND_UP(pll->m * clk->params->min_vco,
+			    clk->parent_rate / KHZ);
+}
+
+int gk20a_clk_ctor(struct nvkm_device *, int, const struct nvkm_clk_func *,
 		    const struct gk20a_clk_pllg_params *, struct gk20a_clk *);
 void gk20a_clk_fini(struct nvkm_clk *);
 int gk20a_clk_read(struct nvkm_clk *, enum nv_clk_src);
@@ -62,4 +154,6 @@ int gk20a_clk_calc(struct nvkm_clk *, struct nvkm_cstate *);
 int gk20a_clk_prog(struct nvkm_clk *);
 void gk20a_clk_tidy(struct nvkm_clk *);
 
+int gk20a_clk_setup_slide(struct gk20a_clk *);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
index 71b2bbb61973..b284e949f732 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c
@@ -21,20 +21,123 @@
  */
 
 #include <subdev/clk.h>
+#include <subdev/volt.h>
+#include <subdev/timer.h>
 #include <core/device.h>
+#include <core/tegra.h>
 
 #include "priv.h"
 #include "gk20a.h"
 
-#define KHZ (1000)
-#define MHZ (KHZ * 1000)
-
-#define MASK(w)	((1 << w) - 1)
+#define GPCPLL_CFG_SYNC_MODE	BIT(2)
 
 #define BYPASSCTRL_SYS	(SYS_GPCPLL_CFG_BASE + 0x340)
 #define BYPASSCTRL_SYS_GPCPLL_SHIFT	0
 #define BYPASSCTRL_SYS_GPCPLL_WIDTH	1
 
+#define GPCPLL_CFG2_SDM_DIN_SHIFT	0
+#define GPCPLL_CFG2_SDM_DIN_WIDTH	8
+#define GPCPLL_CFG2_SDM_DIN_MASK	\
+	(MASK(GPCPLL_CFG2_SDM_DIN_WIDTH) << GPCPLL_CFG2_SDM_DIN_SHIFT)
+#define GPCPLL_CFG2_SDM_DIN_NEW_SHIFT	8
+#define GPCPLL_CFG2_SDM_DIN_NEW_WIDTH	15
+#define GPCPLL_CFG2_SDM_DIN_NEW_MASK	\
+	(MASK(GPCPLL_CFG2_SDM_DIN_NEW_WIDTH) << GPCPLL_CFG2_SDM_DIN_NEW_SHIFT)
+#define GPCPLL_CFG2_SETUP2_SHIFT	16
+#define GPCPLL_CFG2_PLL_STEPA_SHIFT	24
+
+#define GPCPLL_DVFS0	(SYS_GPCPLL_CFG_BASE + 0x10)
+#define GPCPLL_DVFS0_DFS_COEFF_SHIFT	0
+#define GPCPLL_DVFS0_DFS_COEFF_WIDTH	7
+#define GPCPLL_DVFS0_DFS_COEFF_MASK	\
+	(MASK(GPCPLL_DVFS0_DFS_COEFF_WIDTH) << GPCPLL_DVFS0_DFS_COEFF_SHIFT)
+#define GPCPLL_DVFS0_DFS_DET_MAX_SHIFT	8
+#define GPCPLL_DVFS0_DFS_DET_MAX_WIDTH	7
+#define GPCPLL_DVFS0_DFS_DET_MAX_MASK	\
+	(MASK(GPCPLL_DVFS0_DFS_DET_MAX_WIDTH) << GPCPLL_DVFS0_DFS_DET_MAX_SHIFT)
+
+#define GPCPLL_DVFS1		(SYS_GPCPLL_CFG_BASE + 0x14)
+#define GPCPLL_DVFS1_DFS_EXT_DET_SHIFT		0
+#define GPCPLL_DVFS1_DFS_EXT_DET_WIDTH		7
+#define GPCPLL_DVFS1_DFS_EXT_STRB_SHIFT		7
+#define GPCPLL_DVFS1_DFS_EXT_STRB_WIDTH		1
+#define GPCPLL_DVFS1_DFS_EXT_CAL_SHIFT		8
+#define GPCPLL_DVFS1_DFS_EXT_CAL_WIDTH		7
+#define GPCPLL_DVFS1_DFS_EXT_SEL_SHIFT		15
+#define GPCPLL_DVFS1_DFS_EXT_SEL_WIDTH		1
+#define GPCPLL_DVFS1_DFS_CTRL_SHIFT		16
+#define GPCPLL_DVFS1_DFS_CTRL_WIDTH		12
+#define GPCPLL_DVFS1_EN_SDM_SHIFT		28
+#define GPCPLL_DVFS1_EN_SDM_WIDTH		1
+#define GPCPLL_DVFS1_EN_SDM_BIT			BIT(28)
+#define GPCPLL_DVFS1_EN_DFS_SHIFT		29
+#define GPCPLL_DVFS1_EN_DFS_WIDTH		1
+#define GPCPLL_DVFS1_EN_DFS_BIT			BIT(29)
+#define GPCPLL_DVFS1_EN_DFS_CAL_SHIFT		30
+#define GPCPLL_DVFS1_EN_DFS_CAL_WIDTH		1
+#define GPCPLL_DVFS1_EN_DFS_CAL_BIT		BIT(30)
+#define GPCPLL_DVFS1_DFS_CAL_DONE_SHIFT		31
+#define GPCPLL_DVFS1_DFS_CAL_DONE_WIDTH		1
+#define GPCPLL_DVFS1_DFS_CAL_DONE_BIT		BIT(31)
+
+#define GPC_BCAST_GPCPLL_DVFS2	(GPC_BCAST_GPCPLL_CFG_BASE + 0x20)
+#define GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT	BIT(16)
+
+#define GPCPLL_CFG3_PLL_DFS_TESTOUT_SHIFT	24
+#define GPCPLL_CFG3_PLL_DFS_TESTOUT_WIDTH	7
+
+#define DFS_DET_RANGE	6	/* -2^6 ... 2^6-1 */
+#define SDM_DIN_RANGE	12	/* -2^12 ... 2^12-1 */
+
+struct gm20b_clk_dvfs_params {
+	s32 coeff_slope;
+	s32 coeff_offs;
+	u32 vco_ctrl;
+};
+
+static const struct gm20b_clk_dvfs_params gm20b_dvfs_params = {
+	.coeff_slope = -165230,
+	.coeff_offs = 214007,
+	.vco_ctrl = 0x7 << 3,
+};
+
+/*
+ * base.n is now the *integer* part of the N factor.
+ * sdm_din contains n's decimal part.
+ */
+struct gm20b_pll {
+	struct gk20a_pll base;
+	u32 sdm_din;
+};
+
+struct gm20b_clk_dvfs {
+	u32 dfs_coeff;
+	s32 dfs_det_max;
+	s32 dfs_ext_cal;
+};
+
+struct gm20b_clk {
+	/* currently applied parameters */
+	struct gk20a_clk base;
+	struct gm20b_clk_dvfs dvfs;
+	u32 uv;
+
+	/* new parameters to apply */
+	struct gk20a_pll new_pll;
+	struct gm20b_clk_dvfs new_dvfs;
+	u32 new_uv;
+
+	const struct gm20b_clk_dvfs_params *dvfs_params;
+
+	/* fused parameters */
+	s32 uvdet_slope;
+	s32 uvdet_offs;
+
+	/* safe frequency we can use at minimum voltage */
+	u32 safe_fmax_vmin;
+};
+#define gm20b_clk(p) container_of((gk20a_clk(p)), struct gm20b_clk, base)
+
 static u32 pl_to_div(u32 pl)
 {
 	return pl;
@@ -53,6 +156,484 @@ static const struct gk20a_clk_pllg_params gm20b_pllg_params = {
 	.min_pl = 1, .max_pl = 31,
 };
 
+static void
+gm20b_pllg_read_mnp(struct gm20b_clk *clk, struct gm20b_pll *pll)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	struct nvkm_device *device = subdev->device;
+	u32 val;
+
+	gk20a_pllg_read_mnp(&clk->base, &pll->base);
+	val = nvkm_rd32(device, GPCPLL_CFG2);
+	pll->sdm_din = (val >> GPCPLL_CFG2_SDM_DIN_SHIFT) &
+		       MASK(GPCPLL_CFG2_SDM_DIN_WIDTH);
+}
+
+static void
+gm20b_pllg_write_mnp(struct gm20b_clk *clk, const struct gm20b_pll *pll)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+
+	nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_MASK,
+		  pll->sdm_din << GPCPLL_CFG2_SDM_DIN_SHIFT);
+	gk20a_pllg_write_mnp(&clk->base, &pll->base);
+}
+
+/*
+ * Determine DFS_COEFF for the requested voltage. Always select external
+ * calibration override equal to the voltage, and set maximum detection
+ * limit "0" (to make sure that PLL output remains under F/V curve when
+ * voltage increases).
+ */
+static void
+gm20b_dvfs_calc_det_coeff(struct gm20b_clk *clk, s32 uv,
+			  struct gm20b_clk_dvfs *dvfs)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	const struct gm20b_clk_dvfs_params *p = clk->dvfs_params;
+	u32 coeff;
+	/* Work with mv as uv would likely trigger an overflow */
+	s32 mv = DIV_ROUND_CLOSEST(uv, 1000);
+
+	/* coeff = slope * voltage + offset */
+	coeff = DIV_ROUND_CLOSEST(mv * p->coeff_slope, 1000) + p->coeff_offs;
+	coeff = DIV_ROUND_CLOSEST(coeff, 1000);
+	dvfs->dfs_coeff = min_t(u32, coeff, MASK(GPCPLL_DVFS0_DFS_COEFF_WIDTH));
+
+	dvfs->dfs_ext_cal = DIV_ROUND_CLOSEST(uv - clk->uvdet_offs,
+					     clk->uvdet_slope);
+	/* should never happen */
+	if (abs(dvfs->dfs_ext_cal) >= BIT(DFS_DET_RANGE))
+		nvkm_error(subdev, "dfs_ext_cal overflow!\n");
+
+	dvfs->dfs_det_max = 0;
+
+	nvkm_debug(subdev, "%s uv: %d coeff: %x, ext_cal: %d, det_max: %d\n",
+		   __func__, uv, dvfs->dfs_coeff, dvfs->dfs_ext_cal,
+		   dvfs->dfs_det_max);
+}
+
+/*
+ * Solve equation for integer and fractional part of the effective NDIV:
+ *
+ * n_eff = n_int + 1/2 + (SDM_DIN / 2^(SDM_DIN_RANGE + 1)) +
+ *         (DVFS_COEFF * DVFS_DET_DELTA) / 2^DFS_DET_RANGE
+ *
+ * The SDM_DIN LSB is finally shifted out, since it is not accessible by sw.
+ */
+static void
+gm20b_dvfs_calc_ndiv(struct gm20b_clk *clk, u32 n_eff, u32 *n_int, u32 *sdm_din)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	const struct gk20a_clk_pllg_params *p = clk->base.params;
+	u32 n;
+	s32 det_delta;
+	u32 rem, rem_range;
+
+	/* calculate current ext_cal and subtract previous one */
+	det_delta = DIV_ROUND_CLOSEST(((s32)clk->uv) - clk->uvdet_offs,
+				      clk->uvdet_slope);
+	det_delta -= clk->dvfs.dfs_ext_cal;
+	det_delta = min(det_delta, clk->dvfs.dfs_det_max);
+	det_delta *= clk->dvfs.dfs_coeff;
+
+	/* integer part of n */
+	n = (n_eff << DFS_DET_RANGE) - det_delta;
+	/* should never happen! */
+	if (n <= 0) {
+		nvkm_error(subdev, "ndiv <= 0 - setting to 1...\n");
+		n = 1 << DFS_DET_RANGE;
+	}
+	if (n >> DFS_DET_RANGE > p->max_n) {
+		nvkm_error(subdev, "ndiv > max_n - setting to max_n...\n");
+		n = p->max_n << DFS_DET_RANGE;
+	}
+	*n_int = n >> DFS_DET_RANGE;
+
+	/* fractional part of n */
+	rem = ((u32)n) & MASK(DFS_DET_RANGE);
+	rem_range = SDM_DIN_RANGE + 1 - DFS_DET_RANGE;
+	/* subtract 2^SDM_DIN_RANGE to account for the 1/2 of the equation */
+	rem = (rem << rem_range) - BIT(SDM_DIN_RANGE);
+	/* lose 8 LSB and clip - sdm_din only keeps the most significant byte */
+	*sdm_din = (rem >> BITS_PER_BYTE) & MASK(GPCPLL_CFG2_SDM_DIN_WIDTH);
+
+	nvkm_debug(subdev, "%s n_eff: %d, n_int: %d, sdm_din: %d\n", __func__,
+		   n_eff, *n_int, *sdm_din);
+}
+
+static int
+gm20b_pllg_slide(struct gm20b_clk *clk, u32 n)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	struct nvkm_device *device = subdev->device;
+	struct gm20b_pll pll;
+	u32 n_int, sdm_din;
+	int ret = 0;
+
+	/* calculate the new n_int/sdm_din for this n/uv */
+	gm20b_dvfs_calc_ndiv(clk, n, &n_int, &sdm_din);
+
+	/* get old coefficients */
+	gm20b_pllg_read_mnp(clk, &pll);
+	/* do nothing if NDIV is the same */
+	if (n_int == pll.base.n && sdm_din == pll.sdm_din)
+		return 0;
+
+	/* pll slowdown mode */
+	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT),
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT));
+
+	/* new ndiv ready for ramp */
+	/* in DVFS mode SDM is updated via "new" field */
+	nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_NEW_MASK,
+		  sdm_din << GPCPLL_CFG2_SDM_DIN_NEW_SHIFT);
+	pll.base.n = n_int;
+	udelay(1);
+	gk20a_pllg_write_mnp(&clk->base, &pll.base);
+
+	/* dynamic ramp to new ndiv */
+	udelay(1);
+	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
+		  BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT),
+		  BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT));
+
+	/* wait for ramping to complete */
+	if (nvkm_wait_usec(device, 500, GPC_BCAST_NDIV_SLOWDOWN_DEBUG,
+		GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK,
+		GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK) < 0)
+		ret = -ETIMEDOUT;
+
+	/* in DVFS mode complete SDM update */
+	nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_MASK,
+		  sdm_din << GPCPLL_CFG2_SDM_DIN_SHIFT);
+
+	/* exit slowdown mode */
+	nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN,
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT) |
+		BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), 0);
+	nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN);
+
+	return ret;
+}
+
+static int
+gm20b_pllg_enable(struct gm20b_clk *clk)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, GPCPLL_CFG_ENABLE);
+	nvkm_rd32(device, GPCPLL_CFG);
+
+	/* In DVFS mode lock cannot be used - so just delay */
+	udelay(40);
+
+	/* set SYNC_MODE for glitchless switch out of bypass */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_SYNC_MODE,
+		       GPCPLL_CFG_SYNC_MODE);
+	nvkm_rd32(device, GPCPLL_CFG);
+
+	/* switch to VCO mode */
+	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT),
+		  BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+
+	return 0;
+}
+
+static void
+gm20b_pllg_disable(struct gm20b_clk *clk)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+
+	/* put PLL in bypass before disabling it */
+	nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0);
+
+	/* clear SYNC_MODE before disabling PLL */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_SYNC_MODE, 0);
+
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, 0);
+	nvkm_rd32(device, GPCPLL_CFG);
+}
+
+static int
+gm20b_pllg_program_mnp(struct gm20b_clk *clk, const struct gk20a_pll *pll)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	struct nvkm_device *device = subdev->device;
+	struct gm20b_pll cur_pll;
+	u32 n_int, sdm_din;
+	/* if we only change pdiv, we can do a glitchless transition */
+	bool pdiv_only;
+	int ret;
+
+	gm20b_dvfs_calc_ndiv(clk, pll->n, &n_int, &sdm_din);
+	gm20b_pllg_read_mnp(clk, &cur_pll);
+	pdiv_only = cur_pll.base.n == n_int && cur_pll.sdm_din == sdm_din &&
+		    cur_pll.base.m == pll->m;
+
+	/* need full sequence if clock not enabled yet */
+	if (!gk20a_pllg_is_enabled(&clk->base))
+		pdiv_only = false;
+
+	/* split VCO-to-bypass jump in half by setting out divider 1:2 */
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT);
+	/* Intentional 2nd write to assure linear divider operation */
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT);
+	nvkm_rd32(device, GPC2CLK_OUT);
+	udelay(2);
+
+	if (pdiv_only) {
+		u32 old = cur_pll.base.pl;
+		u32 new = pll->pl;
+
+		/*
+		 * we can do a glitchless transition only if the old and new PL
+		 * parameters share at least one bit set to 1. If this is not
+		 * the case, calculate and program an interim PL that will allow
+		 * us to respect that rule.
+		 */
+		if ((old & new) == 0) {
+			cur_pll.base.pl = min(old | BIT(ffs(new) - 1),
+					      new | BIT(ffs(old) - 1));
+			gk20a_pllg_write_mnp(&clk->base, &cur_pll.base);
+		}
+
+		cur_pll.base.pl = new;
+		gk20a_pllg_write_mnp(&clk->base, &cur_pll.base);
+	} else {
+		/* disable before programming if more than pdiv changes */
+		gm20b_pllg_disable(clk);
+
+		cur_pll.base = *pll;
+		cur_pll.base.n = n_int;
+		cur_pll.sdm_din = sdm_din;
+		gm20b_pllg_write_mnp(clk, &cur_pll);
+
+		ret = gm20b_pllg_enable(clk);
+		if (ret)
+			return ret;
+	}
+
+	/* restore out divider 1:1 */
+	udelay(2);
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT);
+	/* Intentional 2nd write to assure linear divider operation */
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		  GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT);
+	nvkm_rd32(device, GPC2CLK_OUT);
+
+	return 0;
+}
+
+static int
+gm20b_pllg_program_mnp_slide(struct gm20b_clk *clk, const struct gk20a_pll *pll)
+{
+	struct gk20a_pll cur_pll;
+	int ret;
+
+	if (gk20a_pllg_is_enabled(&clk->base)) {
+		gk20a_pllg_read_mnp(&clk->base, &cur_pll);
+
+		/* just do NDIV slide if there is no change to M and PL */
+		if (pll->m == cur_pll.m && pll->pl == cur_pll.pl)
+			return gm20b_pllg_slide(clk, pll->n);
+
+		/* slide down to current NDIV_LO */
+		cur_pll.n = gk20a_pllg_n_lo(&clk->base, &cur_pll);
+		ret = gm20b_pllg_slide(clk, cur_pll.n);
+		if (ret)
+			return ret;
+	}
+
+	/* program MNP with the new clock parameters and new NDIV_LO */
+	cur_pll = *pll;
+	cur_pll.n = gk20a_pllg_n_lo(&clk->base, &cur_pll);
+	ret = gm20b_pllg_program_mnp(clk, &cur_pll);
+	if (ret)
+		return ret;
+
+	/* slide up to new NDIV */
+	return gm20b_pllg_slide(clk, pll->n);
+}
+
+static int
+gm20b_clk_calc(struct nvkm_clk *base, struct nvkm_cstate *cstate)
+{
+	struct gm20b_clk *clk = gm20b_clk(base);
+	struct nvkm_subdev *subdev = &base->subdev;
+	struct nvkm_volt *volt = base->subdev.device->volt;
+	int ret;
+
+	ret = gk20a_pllg_calc_mnp(&clk->base, cstate->domain[nv_clk_src_gpc] *
+					     GK20A_CLK_GPC_MDIV, &clk->new_pll);
+	if (ret)
+		return ret;
+
+	clk->new_uv = volt->vid[cstate->voltage].uv;
+	gm20b_dvfs_calc_det_coeff(clk, clk->new_uv, &clk->new_dvfs);
+
+	nvkm_debug(subdev, "%s uv: %d uv\n", __func__, clk->new_uv);
+
+	return 0;
+}
+
+/*
+ * Compute PLL parameters that are always safe for the current voltage
+ */
+static void
+gm20b_dvfs_calc_safe_pll(struct gm20b_clk *clk, struct gk20a_pll *pll)
+{
+	u32 rate = gk20a_pllg_calc_rate(&clk->base, pll) / KHZ;
+	u32 parent_rate = clk->base.parent_rate / KHZ;
+	u32 nmin, nsafe;
+
+	/* remove a safe margin of 10% */
+	if (rate > clk->safe_fmax_vmin)
+		rate = rate * (100 - 10) / 100;
+
+	/* gpc2clk */
+	rate *= 2;
+
+	nmin = DIV_ROUND_UP(pll->m * clk->base.params->min_vco, parent_rate);
+	nsafe = pll->m * rate / (clk->base.parent_rate);
+
+	if (nsafe < nmin) {
+		pll->pl = DIV_ROUND_UP(nmin * parent_rate, pll->m * rate);
+		nsafe = nmin;
+	}
+
+	pll->n = nsafe;
+}
+
+static void
+gm20b_dvfs_program_coeff(struct gm20b_clk *clk, u32 coeff)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+
+	/* strobe to read external DFS coefficient */
+	nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT);
+
+	nvkm_mask(device, GPCPLL_DVFS0, GPCPLL_DVFS0_DFS_COEFF_MASK,
+		  coeff << GPCPLL_DVFS0_DFS_COEFF_SHIFT);
+
+	udelay(1);
+	nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, 0);
+}
+
+static void
+gm20b_dvfs_program_ext_cal(struct gm20b_clk *clk, u32 dfs_det_cal)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+	u32 val;
+
+	nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, MASK(DFS_DET_RANGE + 1),
+		  dfs_det_cal);
+	udelay(1);
+
+	val = nvkm_rd32(device, GPCPLL_DVFS1);
+	if (!(val & BIT(25))) {
+		/* Use external value to overwrite calibration value */
+		val |= BIT(25) | BIT(16);
+		nvkm_wr32(device, GPCPLL_DVFS1, val);
+	}
+}
+
+static void
+gm20b_dvfs_program_dfs_detection(struct gm20b_clk *clk,
+				 struct gm20b_clk_dvfs *dvfs)
+{
+	struct nvkm_device *device = clk->base.base.subdev.device;
+
+	/* strobe to read external DFS coefficient */
+	nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT);
+
+	nvkm_mask(device, GPCPLL_DVFS0,
+		  GPCPLL_DVFS0_DFS_COEFF_MASK | GPCPLL_DVFS0_DFS_DET_MAX_MASK,
+		  dvfs->dfs_coeff << GPCPLL_DVFS0_DFS_COEFF_SHIFT |
+		  dvfs->dfs_det_max << GPCPLL_DVFS0_DFS_DET_MAX_SHIFT);
+
+	udelay(1);
+	nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2,
+		  GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, 0);
+
+	gm20b_dvfs_program_ext_cal(clk, dvfs->dfs_ext_cal);
+}
+
+static int
+gm20b_clk_prog(struct nvkm_clk *base)
+{
+	struct gm20b_clk *clk = gm20b_clk(base);
+	u32 cur_freq;
+	int ret;
+
+	/* No change in DVFS settings? */
+	if (clk->uv == clk->new_uv)
+		goto prog;
+
+	/*
+	 * Interim step for changing DVFS detection settings: low enough
+	 * frequency to be safe at at DVFS coeff = 0.
+	 *
+	 * 1. If voltage is increasing:
+	 * - safe frequency target matches the lowest - old - frequency
+	 * - DVFS settings are still old
+	 * - Voltage already increased to new level by volt, but maximum
+	 *   detection limit assures PLL output remains under F/V curve
+	 *
+	 * 2. If voltage is decreasing:
+	 * - safe frequency target matches the lowest - new - frequency
+	 * - DVFS settings are still old
+	 * - Voltage is also old, it will be lowered by volt afterwards
+	 *
+	 * Interim step can be skipped if old frequency is below safe minimum,
+	 * i.e., it is low enough to be safe at any voltage in operating range
+	 * with zero DVFS coefficient.
+	 */
+	cur_freq = nvkm_clk_read(&clk->base.base, nv_clk_src_gpc);
+	if (cur_freq > clk->safe_fmax_vmin) {
+		struct gk20a_pll pll_safe;
+
+		if (clk->uv < clk->new_uv)
+			/* voltage will raise: safe frequency is current one */
+			pll_safe = clk->base.pll;
+		else
+			/* voltage will drop: safe frequency is new one */
+			pll_safe = clk->new_pll;
+
+		gm20b_dvfs_calc_safe_pll(clk, &pll_safe);
+		ret = gm20b_pllg_program_mnp_slide(clk, &pll_safe);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * DVFS detection settings transition:
+	 * - Set DVFS coefficient zero
+	 * - Set calibration level to new voltage
+	 * - Set DVFS coefficient to match new voltage
+	 */
+	gm20b_dvfs_program_coeff(clk, 0);
+	gm20b_dvfs_program_ext_cal(clk, clk->new_dvfs.dfs_ext_cal);
+	gm20b_dvfs_program_coeff(clk, clk->new_dvfs.dfs_coeff);
+	gm20b_dvfs_program_dfs_detection(clk, &clk->new_dvfs);
+
+prog:
+	clk->uv = clk->new_uv;
+	clk->dvfs = clk->new_dvfs;
+	clk->base.pll = clk->new_pll;
+
+	return gm20b_pllg_program_mnp_slide(clk, &clk->base.pll);
+}
+
 static struct nvkm_pstate
 gm20b_pstates[] = {
 	{
@@ -133,9 +714,99 @@ gm20b_pstates[] = {
 			.voltage = 12,
 		},
 	},
-
 };
 
+static void
+gm20b_clk_fini(struct nvkm_clk *base)
+{
+	struct nvkm_device *device = base->subdev.device;
+	struct gm20b_clk *clk = gm20b_clk(base);
+
+	/* slide to VCO min */
+	if (gk20a_pllg_is_enabled(&clk->base)) {
+		struct gk20a_pll pll;
+		u32 n_lo;
+
+		gk20a_pllg_read_mnp(&clk->base, &pll);
+		n_lo = gk20a_pllg_n_lo(&clk->base, &pll);
+		gm20b_pllg_slide(clk, n_lo);
+	}
+
+	gm20b_pllg_disable(clk);
+
+	/* set IDDQ */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 1);
+}
+
+static int
+gm20b_clk_init_dvfs(struct gm20b_clk *clk)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	struct nvkm_device *device = subdev->device;
+	bool fused = clk->uvdet_offs && clk->uvdet_slope;
+	static const s32 ADC_SLOPE_UV = 10000; /* default ADC detection slope */
+	u32 data;
+	int ret;
+
+	/* Enable NA DVFS */
+	nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_BIT,
+		  GPCPLL_DVFS1_EN_DFS_BIT);
+
+	/* Set VCO_CTRL */
+	if (clk->dvfs_params->vco_ctrl)
+		nvkm_mask(device, GPCPLL_CFG3, GPCPLL_CFG3_VCO_CTRL_MASK,
+		      clk->dvfs_params->vco_ctrl << GPCPLL_CFG3_VCO_CTRL_SHIFT);
+
+	if (fused) {
+		/* Start internal calibration, but ignore results */
+		nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_CAL_BIT,
+			  GPCPLL_DVFS1_EN_DFS_CAL_BIT);
+
+		/* got uvdev parameters from fuse, skip calibration */
+		goto calibrated;
+	}
+
+	/*
+	 * If calibration parameters are not fused, start internal calibration,
+	 * wait for completion, and use results along with default slope to
+	 * calculate ADC offset during boot.
+	 */
+	nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_CAL_BIT,
+			  GPCPLL_DVFS1_EN_DFS_CAL_BIT);
+
+	/* Wait for internal calibration done (spec < 2us). */
+	ret = nvkm_wait_usec(device, 10, GPCPLL_DVFS1,
+			     GPCPLL_DVFS1_DFS_CAL_DONE_BIT,
+			     GPCPLL_DVFS1_DFS_CAL_DONE_BIT);
+	if (ret < 0) {
+		nvkm_error(subdev, "GPCPLL calibration timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	data = nvkm_rd32(device, GPCPLL_CFG3) >>
+			 GPCPLL_CFG3_PLL_DFS_TESTOUT_SHIFT;
+	data &= MASK(GPCPLL_CFG3_PLL_DFS_TESTOUT_WIDTH);
+
+	clk->uvdet_slope = ADC_SLOPE_UV;
+	clk->uvdet_offs = ((s32)clk->uv) - data * ADC_SLOPE_UV;
+
+	nvkm_debug(subdev, "calibrated DVFS parameters: offs %d, slope %d\n",
+		   clk->uvdet_offs, clk->uvdet_slope);
+
+calibrated:
+	/* Compute and apply initial DVFS parameters */
+	gm20b_dvfs_calc_det_coeff(clk, clk->uv, &clk->dvfs);
+	gm20b_dvfs_program_coeff(clk, 0);
+	gm20b_dvfs_program_ext_cal(clk, clk->dvfs.dfs_ext_cal);
+	gm20b_dvfs_program_coeff(clk, clk->dvfs.dfs_coeff);
+	gm20b_dvfs_program_dfs_detection(clk, &clk->new_dvfs);
+
+	return 0;
+}
+
+/* Forward declaration to detect speedo >=1 in gm20b_clk_init() */
+static const struct nvkm_clk_func gm20b_clk;
+
 static int
 gm20b_clk_init(struct nvkm_clk *base)
 {
@@ -143,15 +814,56 @@ gm20b_clk_init(struct nvkm_clk *base)
 	struct nvkm_subdev *subdev = &clk->base.subdev;
 	struct nvkm_device *device = subdev->device;
 	int ret;
+	u32 data;
+
+	/* get out from IDDQ */
+	nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 0);
+	nvkm_rd32(device, GPCPLL_CFG);
+	udelay(5);
+
+	nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_INIT_MASK,
+		  GPC2CLK_OUT_INIT_VAL);
 
 	/* Set the global bypass control to VCO */
 	nvkm_mask(device, BYPASSCTRL_SYS,
 	       MASK(BYPASSCTRL_SYS_GPCPLL_WIDTH) << BYPASSCTRL_SYS_GPCPLL_SHIFT,
 	       0);
 
+	ret = gk20a_clk_setup_slide(clk);
+	if (ret)
+		return ret;
+
+	/* If not fused, set RAM SVOP PDP data 0x2, and enable fuse override */
+	data = nvkm_rd32(device, 0x021944);
+	if (!(data & 0x3)) {
+		data |= 0x2;
+		nvkm_wr32(device, 0x021944, data);
+
+		data = nvkm_rd32(device, 0x021948);
+		data |=  0x1;
+		nvkm_wr32(device, 0x021948, data);
+	}
+
+	/* Disable idle slow down  */
+	nvkm_mask(device, 0x20160, 0x003f0000, 0x0);
+
+	/* speedo >= 1? */
+	if (clk->base.func == &gm20b_clk) {
+		struct gm20b_clk *_clk = gm20b_clk(base);
+		struct nvkm_volt *volt = device->volt;
+
+		/* Get current voltage */
+		_clk->uv = nvkm_volt_get(volt);
+
+		/* Initialize DVFS */
+		ret = gm20b_clk_init_dvfs(_clk);
+		if (ret)
+			return ret;
+	}
+
 	/* Start with lowest frequency */
 	base->func->calc(base, &base->func->pstates[0].base);
-	ret = base->func->prog(&clk->base);
+	ret = base->func->prog(base);
 	if (ret) {
 		nvkm_error(subdev, "cannot initialize clock\n");
 		return ret;
@@ -169,6 +881,7 @@ gm20b_clk_speedo0 = {
 	.prog = gk20a_clk_prog,
 	.tidy = gk20a_clk_tidy,
 	.pstates = gm20b_pstates,
+	/* Speedo 0 only supports 12 voltages */
 	.nr_pstates = ARRAY_SIZE(gm20b_pstates) - 1,
 	.domains = {
 		{ nv_clk_src_crystal, 0xff },
@@ -177,8 +890,26 @@ gm20b_clk_speedo0 = {
 	},
 };
 
-int
-gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+static const struct nvkm_clk_func
+gm20b_clk = {
+	.init = gm20b_clk_init,
+	.fini = gm20b_clk_fini,
+	.read = gk20a_clk_read,
+	.calc = gm20b_clk_calc,
+	.prog = gm20b_clk_prog,
+	.tidy = gk20a_clk_tidy,
+	.pstates = gm20b_pstates,
+	.nr_pstates = ARRAY_SIZE(gm20b_pstates),
+	.domains = {
+		{ nv_clk_src_crystal, 0xff },
+		{ nv_clk_src_gpc, 0xff, 0, "core", GK20A_CLK_GPC_MDIV },
+		{ nv_clk_src_max },
+	},
+};
+
+static int
+gm20b_clk_new_speedo0(struct nvkm_device *device, int index,
+		      struct nvkm_clk **pclk)
 {
 	struct gk20a_clk *clk;
 	int ret;
@@ -188,11 +919,156 @@ gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
 		return -ENOMEM;
 	*pclk = &clk->base;
 
-	ret = _gk20a_clk_ctor(device, index, &gm20b_clk_speedo0,
-			      &gm20b_pllg_params, clk);
+	ret = gk20a_clk_ctor(device, index, &gm20b_clk_speedo0,
+			     &gm20b_pllg_params, clk);
 
 	clk->pl_to_div = pl_to_div;
 	clk->div_to_pl = div_to_pl;
 
 	return ret;
 }
+
+/* FUSE register */
+#define FUSE_RESERVED_CALIB0	0x204
+#define FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_SHIFT	0
+#define FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_WIDTH	4
+#define FUSE_RESERVED_CALIB0_INTERCEPT_INT_SHIFT	4
+#define FUSE_RESERVED_CALIB0_INTERCEPT_INT_WIDTH	10
+#define FUSE_RESERVED_CALIB0_SLOPE_FRAC_SHIFT		14
+#define FUSE_RESERVED_CALIB0_SLOPE_FRAC_WIDTH		10
+#define FUSE_RESERVED_CALIB0_SLOPE_INT_SHIFT		24
+#define FUSE_RESERVED_CALIB0_SLOPE_INT_WIDTH		6
+#define FUSE_RESERVED_CALIB0_FUSE_REV_SHIFT		30
+#define FUSE_RESERVED_CALIB0_FUSE_REV_WIDTH		2
+
+static int
+gm20b_clk_init_fused_params(struct gm20b_clk *clk)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	u32 val = 0;
+	u32 rev = 0;
+
+#if IS_ENABLED(CONFIG_ARCH_TEGRA)
+	tegra_fuse_readl(FUSE_RESERVED_CALIB0, &val);
+	rev = (val >> FUSE_RESERVED_CALIB0_FUSE_REV_SHIFT) &
+	      MASK(FUSE_RESERVED_CALIB0_FUSE_REV_WIDTH);
+#endif
+
+	/* No fused parameters, we will calibrate later */
+	if (rev == 0)
+		return -EINVAL;
+
+	/* Integer part in mV + fractional part in uV */
+	clk->uvdet_slope = ((val >> FUSE_RESERVED_CALIB0_SLOPE_INT_SHIFT) &
+			MASK(FUSE_RESERVED_CALIB0_SLOPE_INT_WIDTH)) * 1000 +
+			((val >> FUSE_RESERVED_CALIB0_SLOPE_FRAC_SHIFT) &
+			MASK(FUSE_RESERVED_CALIB0_SLOPE_FRAC_WIDTH));
+
+	/* Integer part in mV + fractional part in 100uV */
+	clk->uvdet_offs = ((val >> FUSE_RESERVED_CALIB0_INTERCEPT_INT_SHIFT) &
+			MASK(FUSE_RESERVED_CALIB0_INTERCEPT_INT_WIDTH)) * 1000 +
+			((val >> FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_SHIFT) &
+			 MASK(FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_WIDTH)) * 100;
+
+	nvkm_debug(subdev, "fused calibration data: slope %d, offs %d\n",
+		   clk->uvdet_slope, clk->uvdet_offs);
+	return 0;
+}
+
+static int
+gm20b_clk_init_safe_fmax(struct gm20b_clk *clk)
+{
+	struct nvkm_subdev *subdev = &clk->base.base.subdev;
+	struct nvkm_volt *volt = subdev->device->volt;
+	struct nvkm_pstate *pstates = clk->base.base.func->pstates;
+	int nr_pstates = clk->base.base.func->nr_pstates;
+	int vmin, id = 0;
+	u32 fmax = 0;
+	int i;
+
+	/* find lowest voltage we can use */
+	vmin = volt->vid[0].uv;
+	for (i = 1; i < volt->vid_nr; i++) {
+		if (volt->vid[i].uv <= vmin) {
+			vmin = volt->vid[i].uv;
+			id = volt->vid[i].vid;
+		}
+	}
+
+	/* find max frequency at this voltage */
+	for (i = 0; i < nr_pstates; i++)
+		if (pstates[i].base.voltage == id)
+			fmax = max(fmax,
+				   pstates[i].base.domain[nv_clk_src_gpc]);
+
+	if (!fmax) {
+		nvkm_error(subdev, "failed to evaluate safe fmax\n");
+		return -EINVAL;
+	}
+
+	/* we are safe at 90% of the max frequency */
+	clk->safe_fmax_vmin = fmax * (100 - 10) / 100;
+	nvkm_debug(subdev, "safe fmax @ vmin = %u Khz\n", clk->safe_fmax_vmin);
+
+	return 0;
+}
+
+int
+gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk)
+{
+	struct nvkm_device_tegra *tdev = device->func->tegra(device);
+	struct gm20b_clk *clk;
+	struct nvkm_subdev *subdev;
+	struct gk20a_clk_pllg_params *clk_params;
+	int ret;
+
+	/* Speedo 0 GPUs cannot use noise-aware PLL */
+	if (tdev->gpu_speedo_id == 0)
+		return gm20b_clk_new_speedo0(device, index, pclk);
+
+	/* Speedo >= 1, use NAPLL */
+	clk = kzalloc(sizeof(*clk) + sizeof(*clk_params), GFP_KERNEL);
+	if (!clk)
+		return -ENOMEM;
+	*pclk = &clk->base.base;
+	subdev = &clk->base.base.subdev;
+
+	/* duplicate the clock parameters since we will patch them below */
+	clk_params = (void *) (clk + 1);
+	*clk_params = gm20b_pllg_params;
+	ret = gk20a_clk_ctor(device, index, &gm20b_clk, clk_params,
+			     &clk->base);
+	if (ret)
+		return ret;
+
+	/*
+	 * NAPLL can only work with max_u, clamp the m range so
+	 * gk20a_pllg_calc_mnp always uses it
+	 */
+	clk_params->max_m = clk_params->min_m = DIV_ROUND_UP(clk_params->max_u,
+						(clk->base.parent_rate / KHZ));
+	if (clk_params->max_m == 0) {
+		nvkm_warn(subdev, "cannot use NAPLL, using legacy clock...\n");
+		kfree(clk);
+		return gm20b_clk_new_speedo0(device, index, pclk);
+	}
+
+	clk->base.pl_to_div = pl_to_div;
+	clk->base.div_to_pl = div_to_pl;
+
+	clk->dvfs_params = &gm20b_dvfs_params;
+
+	ret = gm20b_clk_init_fused_params(clk);
+	/*
+	 * we will calibrate during init - should never happen on
+	 * prod parts
+	 */
+	if (ret)
+		nvkm_warn(subdev, "no fused calibration parameters\n");
+
+	ret = gm20b_clk_init_safe_fmax(clk);
+	if (ret)
+		return ret;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
index 842d5de96d73..edcc157e6ac8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild
@@ -24,6 +24,8 @@ nvkm-y += nvkm/subdev/fb/gk104.o
 nvkm-y += nvkm/subdev/fb/gk20a.o
 nvkm-y += nvkm/subdev/fb/gm107.o
 nvkm-y += nvkm/subdev/fb/gm200.o
+nvkm-y += nvkm/subdev/fb/gp100.o
+nvkm-y += nvkm/subdev/fb/gp104.o
 
 nvkm-y += nvkm/subdev/fb/ram.o
 nvkm-y += nvkm/subdev/fb/ramnv04.o
@@ -41,6 +43,7 @@ nvkm-y += nvkm/subdev/fb/rammcp77.o
 nvkm-y += nvkm/subdev/fb/ramgf100.o
 nvkm-y += nvkm/subdev/fb/ramgk104.o
 nvkm-y += nvkm/subdev/fb/ramgm107.o
+nvkm-y += nvkm/subdev/fb/ramgp100.o
 nvkm-y += nvkm/subdev/fb/sddr2.o
 nvkm-y += nvkm/subdev/fb/sddr3.o
 nvkm-y += nvkm/subdev/fb/gddr3.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
index ce90242b8cce..a7049c041594 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c
@@ -25,6 +25,7 @@
 #include "ram.h"
 
 #include <core/memory.h>
+#include <core/option.h>
 #include <subdev/bios.h>
 #include <subdev/bios/M0203.h>
 #include <engine/gr.h>
@@ -134,6 +135,10 @@ nvkm_fb_init(struct nvkm_subdev *subdev)
 
 	if (fb->func->init)
 		fb->func->init(fb);
+	if (fb->func->init_page)
+		fb->func->init_page(fb);
+	if (fb->func->init_unkn)
+		fb->func->init_unkn(fb);
 	return 0;
 }
 
@@ -171,6 +176,7 @@ nvkm_fb_ctor(const struct nvkm_fb_func *func, struct nvkm_device *device,
 	nvkm_subdev_ctor(&nvkm_fb, device, index, &fb->subdev);
 	fb->func = func;
 	fb->tile.regions = fb->func->tile.regions;
+	fb->page = nvkm_longopt(device->cfgopt, "NvFbBigPage", 0);
 }
 
 int
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
index e649ead5ccfc..76433cc66fff 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c
@@ -72,6 +72,22 @@ gf100_fb_oneinit(struct nvkm_fb *fb)
 }
 
 void
+gf100_fb_init_page(struct nvkm_fb *fb)
+{
+	struct nvkm_device *device = fb->subdev.device;
+	switch (fb->page) {
+	case 16:
+		nvkm_mask(device, 0x100c80, 0x00000001, 0x00000001);
+		break;
+	case 17:
+	default:
+		nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000);
+		fb->page = 17;
+		break;
+	}
+}
+
+void
 gf100_fb_init(struct nvkm_fb *base)
 {
 	struct gf100_fb *fb = gf100_fb(base);
@@ -79,8 +95,6 @@ gf100_fb_init(struct nvkm_fb *base)
 
 	if (fb->r100c10_page)
 		nvkm_wr32(device, 0x100c10, fb->r100c10 >> 8);
-
-	nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */
 }
 
 void *
@@ -125,6 +139,7 @@ gf100_fb = {
 	.dtor = gf100_fb_dtor,
 	.oneinit = gf100_fb_oneinit,
 	.init = gf100_fb_init,
+	.init_page = gf100_fb_init_page,
 	.intr = gf100_fb_intr,
 	.ram_new = gf100_ram_new,
 	.memtype_valid = gf100_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
index 2160e5a39c9a..449f431644b3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h
@@ -14,4 +14,6 @@ int gf100_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *,
 void *gf100_fb_dtor(struct nvkm_fb *);
 void gf100_fb_init(struct nvkm_fb *);
 void gf100_fb_intr(struct nvkm_fb *);
+
+void gp100_fb_init(struct nvkm_fb *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
index b41f0f70038c..4245e2e6e604 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c
@@ -29,6 +29,7 @@ gk104_fb = {
 	.dtor = gf100_fb_dtor,
 	.oneinit = gf100_fb_oneinit,
 	.init = gf100_fb_init,
+	.init_page = gf100_fb_init_page,
 	.intr = gf100_fb_intr,
 	.ram_new = gk104_ram_new,
 	.memtype_valid = gf100_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
index 7306f7dfc3b9..f815fe2bbf08 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c
@@ -27,7 +27,6 @@ static void
 gk20a_fb_init(struct nvkm_fb *fb)
 {
 	struct nvkm_device *device = fb->subdev.device;
-	nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */
 	nvkm_wr32(device, 0x100cc8, nvkm_memory_addr(fb->mmu_wr) >> 8);
 	nvkm_wr32(device, 0x100ccc, nvkm_memory_addr(fb->mmu_rd) >> 8);
 }
@@ -36,6 +35,7 @@ static const struct nvkm_fb_func
 gk20a_fb = {
 	.oneinit = gf100_fb_oneinit,
 	.init = gk20a_fb_init,
+	.init_page = gf100_fb_init_page,
 	.memtype_valid = gf100_fb_memtype_valid,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
index 4869fdb753c9..db699025f546 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c
@@ -29,6 +29,7 @@ gm107_fb = {
 	.dtor = gf100_fb_dtor,
 	.oneinit = gf100_fb_oneinit,
 	.init = gf100_fb_init,
+	.init_page = gf100_fb_init_page,
 	.intr = gf100_fb_intr,
 	.ram_new = gm107_ram_new,
 	.memtype_valid = gf100_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
index 44f5716f64d8..62f653240be3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c
@@ -26,6 +26,24 @@
 
 #include <core/memory.h>
 
+void
+gm200_fb_init_page(struct nvkm_fb *fb)
+{
+	struct nvkm_device *device = fb->subdev.device;
+	switch (fb->page) {
+	case 16:
+		nvkm_mask(device, 0x100c80, 0x00000801, 0x00000001);
+		break;
+	case 17:
+		nvkm_mask(device, 0x100c80, 0x00000801, 0x00000000);
+		break;
+	default:
+		nvkm_mask(device, 0x100c80, 0x00000800, 0x00000800);
+		fb->page = 0;
+		break;
+	}
+}
+
 static void
 gm200_fb_init(struct nvkm_fb *base)
 {
@@ -48,6 +66,7 @@ gm200_fb = {
 	.dtor = gf100_fb_dtor,
 	.oneinit = gf100_fb_oneinit,
 	.init = gm200_fb_init,
+	.init_page = gm200_fb_init_page,
 	.intr = gf100_fb_intr,
 	.ram_new = gm107_ram_new,
 	.memtype_valid = gf100_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c
new file mode 100644
index 000000000000..98474aec1921
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "gf100.h"
+#include "ram.h"
+
+#include <core/memory.h>
+
+static void
+gp100_fb_init_unkn(struct nvkm_fb *base)
+{
+	struct nvkm_device *device = gf100_fb(base)->base.subdev.device;
+	nvkm_wr32(device, 0x1fac80, nvkm_rd32(device, 0x100c80));
+	nvkm_wr32(device, 0x1facc4, nvkm_rd32(device, 0x100cc4));
+	nvkm_wr32(device, 0x1facc8, nvkm_rd32(device, 0x100cc8));
+	nvkm_wr32(device, 0x1faccc, nvkm_rd32(device, 0x100ccc));
+}
+
+void
+gp100_fb_init(struct nvkm_fb *base)
+{
+	struct gf100_fb *fb = gf100_fb(base);
+	struct nvkm_device *device = fb->base.subdev.device;
+
+	if (fb->r100c10_page)
+		nvkm_wr32(device, 0x100c10, fb->r100c10 >> 8);
+
+	nvkm_wr32(device, 0x100cc8, nvkm_memory_addr(fb->base.mmu_wr) >> 8);
+	nvkm_wr32(device, 0x100ccc, nvkm_memory_addr(fb->base.mmu_rd) >> 8);
+	nvkm_mask(device, 0x100cc4, 0x00060000,
+		  max(nvkm_memory_size(fb->base.mmu_rd) >> 16, (u64)2) << 17);
+}
+
+static const struct nvkm_fb_func
+gp100_fb = {
+	.dtor = gf100_fb_dtor,
+	.oneinit = gf100_fb_oneinit,
+	.init = gp100_fb_init,
+	.init_page = gm200_fb_init_page,
+	.init_unkn = gp100_fb_init_unkn,
+	.ram_new = gp100_ram_new,
+	.memtype_valid = gf100_fb_memtype_valid,
+};
+
+int
+gp100_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+{
+	return gf100_fb_new_(&gp100_fb, device, index, pfb);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c
new file mode 100644
index 000000000000..92cb71861bec
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c
@@ -0,0 +1,43 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "gf100.h"
+#include "ram.h"
+
+#include <core/memory.h>
+
+static const struct nvkm_fb_func
+gp104_fb = {
+	.dtor = gf100_fb_dtor,
+	.oneinit = gf100_fb_oneinit,
+	.init = gp100_fb_init,
+	.init_page = gm200_fb_init_page,
+	.ram_new = gp100_ram_new,
+	.memtype_valid = gf100_fb_memtype_valid,
+};
+
+int
+gp104_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb)
+{
+	return gf100_fb_new_(&gp104_fb, device, index, pfb);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
index d97d640e60a0..e905d44fa1d5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h
@@ -8,6 +8,8 @@ struct nvkm_fb_func {
 	void *(*dtor)(struct nvkm_fb *);
 	int (*oneinit)(struct nvkm_fb *);
 	void (*init)(struct nvkm_fb *);
+	void (*init_page)(struct nvkm_fb *);
+	void (*init_unkn)(struct nvkm_fb *);
 	void (*intr)(struct nvkm_fb *);
 
 	struct {
@@ -60,5 +62,8 @@ void nv46_fb_tile_init(struct nvkm_fb *, int i, u32 addr, u32 size,
 		       u32 pitch, u32 flags, struct nvkm_fb_tile *);
 
 int gf100_fb_oneinit(struct nvkm_fb *);
+void gf100_fb_init_page(struct nvkm_fb *);
 bool gf100_fb_memtype_valid(struct nvkm_fb *, u32);
+
+void gm200_fb_init_page(struct nvkm_fb *);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
index f816cbf2ced3..b9ec0ae6723a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h
@@ -47,4 +47,5 @@ int mcp77_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gf100_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gk104_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 int gm107_ram_new(struct nvkm_fb *, struct nvkm_ram **);
+int gp100_ram_new(struct nvkm_fb *, struct nvkm_ram **);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c
new file mode 100644
index 000000000000..f3be408b5e5e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c
@@ -0,0 +1,146 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "ram.h"
+
+#include <subdev/bios.h>
+#include <subdev/bios/init.h>
+#include <subdev/bios/rammap.h>
+
+static int
+gp100_ram_init(struct nvkm_ram *ram)
+{
+	struct nvkm_subdev *subdev = &ram->fb->subdev;
+	struct nvkm_device *device = subdev->device;
+	struct nvkm_bios *bios = device->bios;
+	u8  ver, hdr, cnt, len, snr, ssz;
+	u32 data;
+	int i;
+
+	/* run a bunch of tables from rammap table.  there's actually
+	 * individual pointers for each rammap entry too, but, nvidia
+	 * seem to just run the last two entries' scripts early on in
+	 * their init, and never again.. we'll just run 'em all once
+	 * for now.
+	 *
+	 * i strongly suspect that each script is for a separate mode
+	 * (likely selected by 0x9a065c's lower bits?), and the
+	 * binary driver skips the one that's already been setup by
+	 * the init tables.
+	 */
+	data = nvbios_rammapTe(bios, &ver, &hdr, &cnt, &len, &snr, &ssz);
+	if (!data || hdr < 0x15)
+		return -EINVAL;
+
+	cnt  = nvbios_rd08(bios, data + 0x14); /* guess at count */
+	data = nvbios_rd32(bios, data + 0x10); /* guess u32... */
+	if (cnt) {
+		u32 save = nvkm_rd32(device, 0x9a065c) & 0x000000f0;
+		for (i = 0; i < cnt; i++, data += 4) {
+			if (i != save >> 4) {
+				nvkm_mask(device, 0x9a065c, 0x000000f0, i << 4);
+				nvbios_exec(&(struct nvbios_init) {
+						.subdev = subdev,
+						.bios = bios,
+						.offset = nvbios_rd32(bios, data),
+						.execute = 1,
+					    });
+			}
+		}
+		nvkm_mask(device, 0x9a065c, 0x000000f0, save);
+	}
+
+	nvkm_mask(device, 0x9a0584, 0x11000000, 0x00000000);
+	nvkm_wr32(device, 0x10ecc0, 0xffffffff);
+	nvkm_mask(device, 0x9a0160, 0x00000010, 0x00000010);
+	return 0;
+}
+
+static const struct nvkm_ram_func
+gp100_ram_func = {
+	.init = gp100_ram_init,
+	.get = gf100_ram_get,
+	.put = gf100_ram_put,
+};
+
+int
+gp100_ram_new(struct nvkm_fb *fb, struct nvkm_ram **pram)
+{
+	struct nvkm_ram *ram;
+	struct nvkm_subdev *subdev = &fb->subdev;
+	struct nvkm_device *device = subdev->device;
+	enum nvkm_ram_type type = nvkm_fb_bios_memtype(device->bios);
+	const u32 rsvd_head = ( 256 * 1024); /* vga memory */
+	const u32 rsvd_tail = (1024 * 1024); /* vbios etc */
+	u32 fbpa_num = nvkm_rd32(device, 0x022438), fbpa;
+	u32 fbio_opt = nvkm_rd32(device, 0x021c14);
+	u64 part, size = 0, comm = ~0ULL;
+	bool mixed = false;
+	int ret;
+
+	nvkm_debug(subdev, "022438: %08x\n", fbpa_num);
+	nvkm_debug(subdev, "021c14: %08x\n", fbio_opt);
+	for (fbpa = 0; fbpa < fbpa_num; fbpa++) {
+		if (!(fbio_opt & (1 << fbpa))) {
+			part = nvkm_rd32(device, 0x90020c + (fbpa * 0x4000));
+			nvkm_debug(subdev, "fbpa %02x: %lld MiB\n", fbpa, part);
+			part = part << 20;
+			if (part != comm) {
+				if (comm != ~0ULL)
+					mixed = true;
+				comm = min(comm, part);
+			}
+			size = size + part;
+		}
+	}
+
+	ret = nvkm_ram_new_(&gp100_ram_func, fb, type, size, 0, &ram);
+	*pram = ram;
+	if (ret)
+		return ret;
+
+	nvkm_mm_fini(&ram->vram);
+
+	if (mixed) {
+		ret = nvkm_mm_init(&ram->vram, rsvd_head >> NVKM_RAM_MM_SHIFT,
+				   ((comm * fbpa_num) - rsvd_head) >>
+				   NVKM_RAM_MM_SHIFT, 1);
+		if (ret)
+			return ret;
+
+		ret = nvkm_mm_init(&ram->vram, (0x1000000000ULL + comm) >>
+				   NVKM_RAM_MM_SHIFT,
+				   (size - (comm * fbpa_num) - rsvd_tail) >>
+				   NVKM_RAM_MM_SHIFT, 1);
+		if (ret)
+			return ret;
+	} else {
+		ret = nvkm_mm_init(&ram->vram, rsvd_head >> NVKM_RAM_MM_SHIFT,
+				   (size - rsvd_head - rsvd_tail) >>
+				   NVKM_RAM_MM_SHIFT, 1);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
index 323c79abe468..41bd5d0f7692 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c
@@ -276,6 +276,8 @@ nvkm_iccsense_oneinit(struct nvkm_subdev *subdev)
 		struct pwr_rail_t *r = &stbl.rail[i];
 		struct nvkm_iccsense_rail *rail;
 		struct nvkm_iccsense_sensor *sensor;
+		int (*read)(struct nvkm_iccsense *,
+			    struct nvkm_iccsense_rail *);
 
 		if (!r->mode || r->resistor_mohm == 0)
 			continue;
@@ -284,31 +286,31 @@ nvkm_iccsense_oneinit(struct nvkm_subdev *subdev)
 		if (!sensor)
 			continue;
 
-		rail = kmalloc(sizeof(*rail), GFP_KERNEL);
-		if (!rail)
-			return -ENOMEM;
-
 		switch (sensor->type) {
 		case NVBIOS_EXTDEV_INA209:
 			if (r->rail != 0)
 				continue;
-			rail->read = nvkm_iccsense_ina209_read;
+			read = nvkm_iccsense_ina209_read;
 			break;
 		case NVBIOS_EXTDEV_INA219:
 			if (r->rail != 0)
 				continue;
-			rail->read = nvkm_iccsense_ina219_read;
+			read = nvkm_iccsense_ina219_read;
 			break;
 		case NVBIOS_EXTDEV_INA3221:
 			if (r->rail >= 3)
 				continue;
-			rail->read = nvkm_iccsense_ina3221_read;
+			read = nvkm_iccsense_ina3221_read;
 			break;
 		default:
 			continue;
 		}
 
+		rail = kmalloc(sizeof(*rail), GFP_KERNEL);
+		if (!rail)
+			return -ENOMEM;
 		sensor->rail_mask |= 1 << r->rail;
+		rail->read = read;
 		rail->sensor = sensor;
 		rail->idx = r->rail;
 		rail->mohm = r->resistor_mohm;
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild
index 932b366598aa..12d6f4f102cb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild
@@ -3,3 +3,4 @@ nvkm-y += nvkm/subdev/ltc/gf100.o
 nvkm-y += nvkm/subdev/ltc/gk104.o
 nvkm-y += nvkm/subdev/ltc/gm107.o
 nvkm-y += nvkm/subdev/ltc/gm200.o
+nvkm-y += nvkm/subdev/ltc/gp100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
index c9eb677967a8..4a0fa0a9b802 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c
@@ -23,7 +23,6 @@
  */
 #include "priv.h"
 
-#include <core/enum.h>
 #include <subdev/fb.h>
 #include <subdev/timer.h>
 
@@ -71,7 +70,7 @@ gf100_ltc_zbc_clear_depth(struct nvkm_ltc *ltc, int i, const u32 depth)
 	nvkm_wr32(device, 0x17ea58, depth);
 }
 
-static const struct nvkm_bitfield
+const struct nvkm_bitfield
 gf100_ltc_lts_intr_name[] = {
 	{ 0x00000001, "IDLE_ERROR_IQ" },
 	{ 0x00000002, "IDLE_ERROR_CBC" },
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
index e292f5679418..ec0a3844b2d1 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c
@@ -68,18 +68,22 @@ gm107_ltc_zbc_clear_depth(struct nvkm_ltc *ltc, int i, const u32 depth)
 	nvkm_wr32(device, 0x17e34c, depth);
 }
 
-static void
-gm107_ltc_lts_isr(struct nvkm_ltc *ltc, int c, int s)
+void
+gm107_ltc_intr_lts(struct nvkm_ltc *ltc, int c, int s)
 {
 	struct nvkm_subdev *subdev = &ltc->subdev;
 	struct nvkm_device *device = subdev->device;
-	u32 base = 0x140000 + (c * 0x2000) + (s * 0x200);
-	u32 stat = nvkm_rd32(device, base + 0x00c);
+	u32 base = 0x140400 + (c * 0x2000) + (s * 0x200);
+	u32 intr = nvkm_rd32(device, base + 0x00c);
+	u16 stat = intr & 0x0000ffff;
+	char msg[128];
 
 	if (stat) {
-		nvkm_error(subdev, "LTC%d_LTS%d: %08x\n", c, s, stat);
-		nvkm_wr32(device, base + 0x00c, stat);
+		nvkm_snprintbf(msg, sizeof(msg), gf100_ltc_lts_intr_name, stat);
+		nvkm_error(subdev, "LTC%d_LTS%d: %08x [%s]\n", c, s, intr, msg);
 	}
+
+	nvkm_wr32(device, base + 0x00c, intr);
 }
 
 void
@@ -92,7 +96,7 @@ gm107_ltc_intr(struct nvkm_ltc *ltc)
 	while (mask) {
 		u32 s, c = __ffs(mask);
 		for (s = 0; s < ltc->lts_nr; s++)
-			gm107_ltc_lts_isr(ltc, c, s);
+			gm107_ltc_intr_lts(ltc, c, s);
 		mask &= ~(1 << c);
 	}
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
index 2a29bfd5125a..e18e0dc19ec8 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c
@@ -46,7 +46,7 @@ static const struct nvkm_ltc_func
 gm200_ltc = {
 	.oneinit = gm200_ltc_oneinit,
 	.init = gm200_ltc_init,
-	.intr = gm107_ltc_intr, /*XXX: not validated */
+	.intr = gm107_ltc_intr,
 	.cbc_clear = gm107_ltc_cbc_clear,
 	.cbc_wait = gm107_ltc_cbc_wait,
 	.zbc = 16,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c
new file mode 100644
index 000000000000..0bdfb2f40266
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c
@@ -0,0 +1,75 @@
+/*
+ * Copyright 2016 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#include "priv.h"
+
+static void
+gp100_ltc_intr(struct nvkm_ltc *ltc)
+{
+	struct nvkm_device *device = ltc->subdev.device;
+	u32 mask;
+
+	mask = nvkm_rd32(device, 0x0001c0);
+	while (mask) {
+		u32 s, c = __ffs(mask);
+		for (s = 0; s < ltc->lts_nr; s++)
+			gm107_ltc_intr_lts(ltc, c, s);
+		mask &= ~(1 << c);
+	}
+}
+
+static int
+gp100_ltc_oneinit(struct nvkm_ltc *ltc)
+{
+	struct nvkm_device *device = ltc->subdev.device;
+	ltc->ltc_nr = nvkm_rd32(device, 0x12006c);
+	ltc->lts_nr = nvkm_rd32(device, 0x17e280) >> 28;
+	/*XXX: tagram allocation - TBD */
+	return nvkm_mm_init(&ltc->tags, 0, 0, 1);
+}
+
+static void
+gp100_ltc_init(struct nvkm_ltc *ltc)
+{
+	/*XXX: PMU LS call to setup tagram address */
+}
+
+static const struct nvkm_ltc_func
+gp100_ltc = {
+	.oneinit = gp100_ltc_oneinit,
+	.init = gp100_ltc_init,
+	.intr = gp100_ltc_intr,
+	.cbc_clear = gm107_ltc_cbc_clear,
+	.cbc_wait = gm107_ltc_cbc_wait,
+	.zbc = 16,
+	.zbc_clear_color = gm107_ltc_zbc_clear_color,
+	.zbc_clear_depth = gm107_ltc_zbc_clear_depth,
+	.invalidate = gf100_ltc_invalidate,
+	.flush = gf100_ltc_flush,
+};
+
+int
+gp100_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc)
+{
+	return nvkm_ltc_new_(&gp100_ltc, device, index, pltc);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
index 6d81c695ed0d..8b95f96e3ffa 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h
@@ -2,6 +2,7 @@
 #define __NVKM_LTC_PRIV_H__
 #define nvkm_ltc(p) container_of((p), struct nvkm_ltc, subdev)
 #include <subdev/ltc.h>
+#include <core/enum.h>
 
 int nvkm_ltc_new_(const struct nvkm_ltc_func *, struct nvkm_device *,
 		  int index, struct nvkm_ltc **);
@@ -31,8 +32,10 @@ void gf100_ltc_zbc_clear_color(struct nvkm_ltc *, int, const u32[4]);
 void gf100_ltc_zbc_clear_depth(struct nvkm_ltc *, int, const u32);
 void gf100_ltc_invalidate(struct nvkm_ltc *);
 void gf100_ltc_flush(struct nvkm_ltc *);
+extern const struct nvkm_bitfield gf100_ltc_lts_intr_name[];
 
 void gm107_ltc_intr(struct nvkm_ltc *);
+void gm107_ltc_intr_lts(struct nvkm_ltc *, int ltc, int lts);
 void gm107_ltc_cbc_clear(struct nvkm_ltc *, u32, u32);
 void gm107_ltc_cbc_wait(struct nvkm_ltc *);
 void gm107_ltc_zbc_clear_color(struct nvkm_ltc *, int, const u32[4]);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
index 49695ac7be2e..12943f92c206 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild
@@ -10,3 +10,4 @@ nvkm-y += nvkm/subdev/mc/gt215.o
 nvkm-y += nvkm/subdev/mc/gf100.o
 nvkm-y += nvkm/subdev/mc/gk104.o
 nvkm-y += nvkm/subdev/mc/gk20a.o
+nvkm-y += nvkm/subdev/mc/gp100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
index 350a8caa84c8..6b25e25f9eba 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c
@@ -27,43 +27,67 @@
 #include <subdev/top.h>
 
 void
-nvkm_mc_unk260(struct nvkm_mc *mc, u32 data)
+nvkm_mc_unk260(struct nvkm_device *device, u32 data)
 {
-	if (mc->func->unk260)
+	struct nvkm_mc *mc = device->mc;
+	if (likely(mc) && mc->func->unk260)
 		mc->func->unk260(mc, data);
 }
 
 void
-nvkm_mc_intr_unarm(struct nvkm_mc *mc)
+nvkm_mc_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx, bool en)
 {
-	return mc->func->intr_unarm(mc);
+	struct nvkm_mc *mc = device->mc;
+	const struct nvkm_mc_map *map;
+	if (likely(mc) && mc->func->intr_mask) {
+		u32 mask = nvkm_top_intr_mask(device, devidx);
+		for (map = mc->func->intr; !mask && map->stat; map++) {
+			if (map->unit == devidx)
+				mask = map->stat;
+		}
+		mc->func->intr_mask(mc, mask, en ? mask : 0);
+	}
+}
+
+void
+nvkm_mc_intr_unarm(struct nvkm_device *device)
+{
+	struct nvkm_mc *mc = device->mc;
+	if (likely(mc))
+		mc->func->intr_unarm(mc);
 }
 
 void
-nvkm_mc_intr_rearm(struct nvkm_mc *mc)
+nvkm_mc_intr_rearm(struct nvkm_device *device)
 {
-	return mc->func->intr_rearm(mc);
+	struct nvkm_mc *mc = device->mc;
+	if (likely(mc))
+		mc->func->intr_rearm(mc);
 }
 
 static u32
-nvkm_mc_intr_mask(struct nvkm_mc *mc)
+nvkm_mc_intr_stat(struct nvkm_mc *mc)
 {
-	u32 intr = mc->func->intr_mask(mc);
+	u32 intr = mc->func->intr_stat(mc);
 	if (WARN_ON_ONCE(intr == 0xffffffff))
 		intr = 0; /* likely fallen off the bus */
 	return intr;
 }
 
 void
-nvkm_mc_intr(struct nvkm_mc *mc, bool *handled)
+nvkm_mc_intr(struct nvkm_device *device, bool *handled)
 {
-	struct nvkm_device *device = mc->subdev.device;
+	struct nvkm_mc *mc = device->mc;
 	struct nvkm_subdev *subdev;
-	const struct nvkm_mc_map *map = mc->func->intr;
-	u32 stat, intr = nvkm_mc_intr_mask(mc);
+	const struct nvkm_mc_map *map;
+	u32 stat, intr;
 	u64 subdevs;
 
-	stat = nvkm_top_intr(device->top, intr, &subdevs);
+	if (unlikely(!mc))
+		return;
+
+	intr = nvkm_mc_intr_stat(mc);
+	stat = nvkm_top_intr(device, intr, &subdevs);
 	while (subdevs) {
 		enum nvkm_devidx subidx = __ffs64(subdevs);
 		subdev = nvkm_device_subdev(device, subidx);
@@ -72,14 +96,13 @@ nvkm_mc_intr(struct nvkm_mc *mc, bool *handled)
 		subdevs &= ~BIT_ULL(subidx);
 	}
 
-	while (map->stat) {
+	for (map = mc->func->intr; map->stat; map++) {
 		if (intr & map->stat) {
 			subdev = nvkm_device_subdev(device, map->unit);
 			if (subdev)
 				nvkm_subdev_intr(subdev);
 			stat &= ~map->stat;
 		}
-		map++;
 	}
 
 	if (stat)
@@ -87,22 +110,32 @@ nvkm_mc_intr(struct nvkm_mc *mc, bool *handled)
 	*handled = intr != 0;
 }
 
-static void
-nvkm_mc_reset_(struct nvkm_mc *mc, enum nvkm_devidx devidx)
+static u32
+nvkm_mc_reset_mask(struct nvkm_device *device, bool isauto,
+		   enum nvkm_devidx devidx)
 {
-	struct nvkm_device *device = mc->subdev.device;
+	struct nvkm_mc *mc = device->mc;
 	const struct nvkm_mc_map *map;
-	u64 pmc_enable;
-
-	if (!(pmc_enable = nvkm_top_reset(device->top, devidx))) {
-		for (map = mc->func->reset; map && map->stat; map++) {
-			if (map->unit == devidx) {
-				pmc_enable = map->stat;
-				break;
+	u64 pmc_enable = 0;
+	if (likely(mc)) {
+		if (!(pmc_enable = nvkm_top_reset(device, devidx))) {
+			for (map = mc->func->reset; map && map->stat; map++) {
+				if (!isauto || !map->noauto) {
+					if (map->unit == devidx) {
+						pmc_enable = map->stat;
+						break;
+					}
+				}
 			}
 		}
 	}
+	return pmc_enable;
+}
 
+void
+nvkm_mc_reset(struct nvkm_device *device, enum nvkm_devidx devidx)
+{
+	u64 pmc_enable = nvkm_mc_reset_mask(device, true, devidx);
 	if (pmc_enable) {
 		nvkm_mask(device, 0x000200, pmc_enable, 0x00000000);
 		nvkm_mask(device, 0x000200, pmc_enable, pmc_enable);
@@ -111,17 +144,27 @@ nvkm_mc_reset_(struct nvkm_mc *mc, enum nvkm_devidx devidx)
 }
 
 void
-nvkm_mc_reset(struct nvkm_mc *mc, enum nvkm_devidx devidx)
+nvkm_mc_disable(struct nvkm_device *device, enum nvkm_devidx devidx)
 {
-	if (likely(mc))
-		nvkm_mc_reset_(mc, devidx);
+	u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx);
+	if (pmc_enable)
+		nvkm_mask(device, 0x000200, pmc_enable, 0x00000000);
+}
+
+void
+nvkm_mc_enable(struct nvkm_device *device, enum nvkm_devidx devidx)
+{
+	u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx);
+	if (pmc_enable) {
+		nvkm_mask(device, 0x000200, pmc_enable, pmc_enable);
+		nvkm_rd32(device, 0x000200);
+	}
 }
 
 static int
 nvkm_mc_fini(struct nvkm_subdev *subdev, bool suspend)
 {
-	struct nvkm_mc *mc = nvkm_mc(subdev);
-	nvkm_mc_intr_unarm(mc);
+	nvkm_mc_intr_unarm(subdev->device);
 	return 0;
 }
 
@@ -131,7 +174,7 @@ nvkm_mc_init(struct nvkm_subdev *subdev)
 	struct nvkm_mc *mc = nvkm_mc(subdev);
 	if (mc->func->init)
 		mc->func->init(mc);
-	nvkm_mc_intr_rearm(mc);
+	nvkm_mc_intr_rearm(subdev->device);
 	return 0;
 }
 
@@ -148,16 +191,21 @@ nvkm_mc = {
 	.fini = nvkm_mc_fini,
 };
 
+void
+nvkm_mc_ctor(const struct nvkm_mc_func *func, struct nvkm_device *device,
+	     int index, struct nvkm_mc *mc)
+{
+	nvkm_subdev_ctor(&nvkm_mc, device, index, &mc->subdev);
+	mc->func = func;
+}
+
 int
 nvkm_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device,
 	     int index, struct nvkm_mc **pmc)
 {
 	struct nvkm_mc *mc;
-
 	if (!(mc = *pmc = kzalloc(sizeof(*mc), GFP_KERNEL)))
 		return -ENOMEM;
-
-	nvkm_subdev_ctor(&nvkm_mc, device, index, &mc->subdev);
-	mc->func = func;
+	nvkm_mc_ctor(func, device, index, *pmc);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
index 5c85b47f071d..c3d66ef5dc12 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c
@@ -57,7 +57,7 @@ g84_mc = {
 	.intr = g84_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = g84_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
index 0280b43cc10c..93ad4982ce5f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c
@@ -57,7 +57,7 @@ g98_mc = {
 	.intr = g98_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = g98_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
index 8397e223bd43..d2c4d6033abb 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c
@@ -76,7 +76,7 @@ gf100_mc_intr_rearm(struct nvkm_mc *mc)
 }
 
 u32
-gf100_mc_intr_mask(struct nvkm_mc *mc)
+gf100_mc_intr_stat(struct nvkm_mc *mc)
 {
 	struct nvkm_device *device = mc->subdev.device;
 	u32 intr0 = nvkm_rd32(device, 0x000100);
@@ -85,6 +85,14 @@ gf100_mc_intr_mask(struct nvkm_mc *mc)
 }
 
 void
+gf100_mc_intr_mask(struct nvkm_mc *mc, u32 mask, u32 stat)
+{
+	struct nvkm_device *device = mc->subdev.device;
+	nvkm_mask(device, 0x000640, mask, stat);
+	nvkm_mask(device, 0x000644, mask, stat);
+}
+
+void
 gf100_mc_unk260(struct nvkm_mc *mc, u32 data)
 {
 	nvkm_wr32(mc->subdev.device, 0x000260, data);
@@ -97,6 +105,7 @@ gf100_mc = {
 	.intr_unarm = gf100_mc_intr_unarm,
 	.intr_rearm = gf100_mc_intr_rearm,
 	.intr_mask = gf100_mc_intr_mask,
+	.intr_stat = gf100_mc_intr_stat,
 	.reset = gf100_mc_reset,
 	.unk260 = gf100_mc_unk260,
 };
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
index 317464212c7d..7b8c6ecad1a5 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c
@@ -26,6 +26,7 @@
 const struct nvkm_mc_map
 gk104_mc_reset[] = {
 	{ 0x00000100, NVKM_ENGINE_FIFO },
+	{ 0x00002000, NVKM_SUBDEV_PMU, true },
 	{}
 };
 
@@ -53,6 +54,7 @@ gk104_mc = {
 	.intr_unarm = gf100_mc_intr_unarm,
 	.intr_rearm = gf100_mc_intr_rearm,
 	.intr_mask = gf100_mc_intr_mask,
+	.intr_stat = gf100_mc_intr_stat,
 	.reset = gk104_mc_reset,
 	.unk260 = gf100_mc_unk260,
 };
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
index 60b044f517ed..ca1bf3279dbe 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c
@@ -30,6 +30,7 @@ gk20a_mc = {
 	.intr_unarm = gf100_mc_intr_unarm,
 	.intr_rearm = gf100_mc_intr_rearm,
 	.intr_mask = gf100_mc_intr_mask,
+	.intr_stat = gf100_mc_intr_stat,
 	.reset = gk104_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c
new file mode 100644
index 000000000000..4d22f4abd6de
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c
@@ -0,0 +1,103 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+#define gp100_mc(p) container_of((p), struct gp100_mc, base)
+#include "priv.h"
+
+struct gp100_mc {
+	struct nvkm_mc base;
+	spinlock_t lock;
+	bool intr;
+	u32 mask;
+};
+
+static void
+gp100_mc_intr_update(struct gp100_mc *mc)
+{
+	struct nvkm_device *device = mc->base.subdev.device;
+	u32 mask = mc->intr ? mc->mask : 0, i;
+	for (i = 0; i < 2; i++) {
+		nvkm_wr32(device, 0x000180 + (i * 0x04), ~mask);
+		nvkm_wr32(device, 0x000160 + (i * 0x04),  mask);
+	}
+}
+
+static void
+gp100_mc_intr_unarm(struct nvkm_mc *base)
+{
+	struct gp100_mc *mc = gp100_mc(base);
+	unsigned long flags;
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->intr = false;
+	gp100_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+static void
+gp100_mc_intr_rearm(struct nvkm_mc *base)
+{
+	struct gp100_mc *mc = gp100_mc(base);
+	unsigned long flags;
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->intr = true;
+	gp100_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+static void
+gp100_mc_intr_mask(struct nvkm_mc *base, u32 mask, u32 intr)
+{
+	struct gp100_mc *mc = gp100_mc(base);
+	unsigned long flags;
+	spin_lock_irqsave(&mc->lock, flags);
+	mc->mask = (mc->mask & ~mask) | intr;
+	gp100_mc_intr_update(mc);
+	spin_unlock_irqrestore(&mc->lock, flags);
+}
+
+static const struct nvkm_mc_func
+gp100_mc = {
+	.init = nv50_mc_init,
+	.intr = gk104_mc_intr,
+	.intr_unarm = gp100_mc_intr_unarm,
+	.intr_rearm = gp100_mc_intr_rearm,
+	.intr_mask = gp100_mc_intr_mask,
+	.intr_stat = gf100_mc_intr_stat,
+	.reset = gk104_mc_reset,
+};
+
+int
+gp100_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc)
+{
+	struct gp100_mc *mc;
+
+	if (!(mc = kzalloc(sizeof(*mc), GFP_KERNEL)))
+		return -ENOMEM;
+	nvkm_mc_ctor(&gp100_mc, device, index, &mc->base);
+	*pmc = &mc->base;
+
+	spin_lock_init(&mc->lock);
+	mc->intr = false;
+	mc->mask = 0x7fffffff;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
index aad0ba95bf18..99d50a3d956f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c
@@ -53,13 +53,20 @@ gt215_mc_intr[] = {
 	{},
 };
 
+static void
+gt215_mc_intr_mask(struct nvkm_mc *mc, u32 mask, u32 stat)
+{
+	nvkm_mask(mc->subdev.device, 0x000640, mask, stat);
+}
+
 static const struct nvkm_mc_func
 gt215_mc = {
 	.init = nv50_mc_init,
 	.intr = gt215_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_mask = gt215_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = gt215_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
index a062624e906b..6509defd1460 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c
@@ -56,7 +56,7 @@ nv04_mc_intr_rearm(struct nvkm_mc *mc)
 }
 
 u32
-nv04_mc_intr_mask(struct nvkm_mc *mc)
+nv04_mc_intr_stat(struct nvkm_mc *mc)
 {
 	return nvkm_rd32(mc->subdev.device, 0x000100);
 }
@@ -75,7 +75,7 @@ nv04_mc = {
 	.intr = nv04_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = nv04_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
index 55f0b9166b52..9213107901e6 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c
@@ -39,7 +39,7 @@ nv11_mc = {
 	.intr = nv11_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = nv04_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
index c40fa67f79a5..64bf5bbf8146 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c
@@ -48,7 +48,7 @@ nv17_mc = {
 	.intr = nv17_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = nv17_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
index cc56271db564..65fa44a64b98 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c
@@ -43,7 +43,7 @@ nv44_mc = {
 	.intr = nv17_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = nv17_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
index 343b6078580d..fe93b4fd7100 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c
@@ -50,7 +50,7 @@ nv50_mc = {
 	.intr = nv50_mc_intr,
 	.intr_unarm = nv04_mc_intr_unarm,
 	.intr_rearm = nv04_mc_intr_rearm,
-	.intr_mask = nv04_mc_intr_mask,
+	.intr_stat = nv04_mc_intr_stat,
 	.reset = nv17_mc_reset,
 };
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
index a12038118512..4f0576a06d24 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h
@@ -3,12 +3,15 @@
 #define nvkm_mc(p) container_of((p), struct nvkm_mc, subdev)
 #include <subdev/mc.h>
 
+void nvkm_mc_ctor(const struct nvkm_mc_func *, struct nvkm_device *,
+		  int index, struct nvkm_mc *);
 int nvkm_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *,
 		 int index, struct nvkm_mc **);
 
 struct nvkm_mc_map {
 	u32 stat;
 	u32 unit;
+	bool noauto;
 };
 
 struct nvkm_mc_func {
@@ -18,8 +21,10 @@ struct nvkm_mc_func {
 	void (*intr_unarm)(struct nvkm_mc *);
 	/* enable reporting of interrupts to host */
 	void (*intr_rearm)(struct nvkm_mc *);
+	/* (un)mask delivery of specific interrupts */
+	void (*intr_mask)(struct nvkm_mc *, u32 mask, u32 stat);
 	/* retrieve pending interrupt mask (NV_PMC_INTR) */
-	u32 (*intr_mask)(struct nvkm_mc *);
+	u32 (*intr_stat)(struct nvkm_mc *);
 	const struct nvkm_mc_map *reset;
 	void (*unk260)(struct nvkm_mc *, u32);
 };
@@ -27,7 +32,7 @@ struct nvkm_mc_func {
 void nv04_mc_init(struct nvkm_mc *);
 void nv04_mc_intr_unarm(struct nvkm_mc *);
 void nv04_mc_intr_rearm(struct nvkm_mc *);
-u32 nv04_mc_intr_mask(struct nvkm_mc *);
+u32 nv04_mc_intr_stat(struct nvkm_mc *);
 extern const struct nvkm_mc_map nv04_mc_reset[];
 
 extern const struct nvkm_mc_map nv17_mc_intr[];
@@ -39,7 +44,8 @@ void nv50_mc_init(struct nvkm_mc *);
 
 void gf100_mc_intr_unarm(struct nvkm_mc *);
 void gf100_mc_intr_rearm(struct nvkm_mc *);
-u32 gf100_mc_intr_mask(struct nvkm_mc *);
+void gf100_mc_intr_mask(struct nvkm_mc *, u32, u32);
+u32 gf100_mc_intr_stat(struct nvkm_mc *);
 void gf100_mc_unk260(struct nvkm_mc *, u32);
 
 extern const struct nvkm_mc_map gk104_mc_intr[];
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild
index 3c2519fdeb81..2a31b7d66a6d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild
@@ -10,3 +10,4 @@ nvkm-y += nvkm/subdev/pci/g94.o
 nvkm-y += nvkm/subdev/pci/gf100.o
 nvkm-y += nvkm/subdev/pci/gf106.o
 nvkm-y += nvkm/subdev/pci/gk104.o
+nvkm-y += nvkm/subdev/pci/gp100.o
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
index 6b0328bd7eed..eb9b278198b2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
@@ -69,15 +69,13 @@ static irqreturn_t
 nvkm_pci_intr(int irq, void *arg)
 {
 	struct nvkm_pci *pci = arg;
-	struct nvkm_mc *mc = pci->subdev.device->mc;
+	struct nvkm_device *device = pci->subdev.device;
 	bool handled = false;
-	if (likely(mc)) {
-		nvkm_mc_intr_unarm(mc);
-		if (pci->msi)
-			pci->func->msi_rearm(pci);
-		nvkm_mc_intr(mc, &handled);
-		nvkm_mc_intr_rearm(mc);
-	}
+	nvkm_mc_intr_unarm(device);
+	if (pci->msi)
+		pci->func->msi_rearm(pci);
+	nvkm_mc_intr(device, &handled);
+	nvkm_mc_intr_rearm(device);
 	return handled ? IRQ_HANDLED : IRQ_NONE;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c
new file mode 100644
index 000000000000..82c5234a06ff
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2015 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+#include "priv.h"
+
+static void
+gp100_pci_msi_rearm(struct nvkm_pci *pci)
+{
+	nvkm_pci_wr32(pci, 0x0704, 0x00000000);
+}
+
+static const struct nvkm_pci_func
+gp100_pci_func = {
+	.rd32 = nv40_pci_rd32,
+	.wr08 = nv40_pci_wr08,
+	.wr32 = nv40_pci_wr32,
+	.msi_rearm = gp100_pci_msi_rearm,
+};
+
+int
+gp100_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci)
+{
+	return nvkm_pci_new_(&gp100_pci_func, device, index, ppci);
+}
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c
index 213fdba6cfa0..314be2192b7d 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c
@@ -19,8 +19,9 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
-
 #include "priv.h"
+
+#include <subdev/mc.h>
 #include <subdev/timer.h>
 
 static const char *
@@ -70,12 +71,11 @@ nvkm_secboot_falcon_enable(struct nvkm_secboot *sb)
 	int ret;
 
 	/* enable engine */
-	nvkm_mask(device, 0x200, sb->enable_mask, sb->enable_mask);
-	nvkm_rd32(device, 0x200);
+	nvkm_mc_enable(device, sb->devidx);
 	ret = nvkm_wait_msec(device, 10, sb->base + 0x10c, 0x6, 0x0);
 	if (ret < 0) {
-		nvkm_mask(device, 0x200, sb->enable_mask, 0x0);
 		nvkm_error(&sb->subdev, "Falcon mem scrubbing timeout\n");
+		nvkm_mc_disable(device, sb->devidx);
 		return ret;
 	}
 
@@ -85,8 +85,7 @@ nvkm_secboot_falcon_enable(struct nvkm_secboot *sb)
 
 	/* enable IRQs */
 	nvkm_wr32(device, sb->base + 0x010, 0xff);
-	nvkm_mask(device, 0x640, sb->irq_mask, sb->irq_mask);
-	nvkm_mask(device, 0x644, sb->irq_mask, sb->irq_mask);
+	nvkm_mc_intr_mask(device, sb->devidx, true);
 
 	return 0;
 }
@@ -97,14 +96,13 @@ nvkm_secboot_falcon_disable(struct nvkm_secboot *sb)
 	struct nvkm_device *device = sb->subdev.device;
 
 	/* disable IRQs and wait for any previous code to complete */
-	nvkm_mask(device, 0x644, sb->irq_mask, 0x0);
-	nvkm_mask(device, 0x640, sb->irq_mask, 0x0);
+	nvkm_mc_intr_mask(device, sb->devidx, false);
 	nvkm_wr32(device, sb->base + 0x014, 0xff);
 
 	falcon_wait_idle(device, sb->base);
 
 	/* disable engine */
-	nvkm_mask(device, 0x200, sb->enable_mask, 0x0);
+	nvkm_mc_disable(device, sb->devidx);
 
 	return 0;
 }
@@ -216,14 +214,7 @@ nvkm_secboot_oneinit(struct nvkm_subdev *subdev)
 		return ret;
 	}
 
-	/*
-	 * Build all blobs - the same blobs can be used to perform secure boot
-	 * multiple times
-	 */
-	if (sb->func->prepare_blobs)
-		ret = sb->func->prepare_blobs(sb);
-
-	return ret;
+	return 0;
 }
 
 static int
@@ -270,9 +261,8 @@ nvkm_secboot_ctor(const struct nvkm_secboot_func *func,
 	/* setup the performing falcon's base address and masks */
 	switch (func->boot_falcon) {
 	case NVKM_SECBOOT_FALCON_PMU:
+		sb->devidx = NVKM_SUBDEV_PMU;
 		sb->base = 0x10a000;
-		sb->irq_mask = 0x1000000;
-		sb->enable_mask = 0x2000;
 		break;
 	default:
 		nvkm_error(&sb->subdev, "invalid secure boot falcon\n");
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c
index cc100dc940ea..f1e2dc914366 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c
@@ -860,6 +860,8 @@ gm200_secboot_prepare_ls_blob(struct gm200_secboot *gsb)
 
 	/* Write LS blob */
 	ret = ls_ucode_mgr_write_wpr(gsb, &mgr, gsb->ls_blob);
+	if (ret)
+		nvkm_gpuobj_del(&gsb->ls_blob);
 
 cleanup:
 	ls_ucode_mgr_cleanup(&mgr);
@@ -1023,29 +1025,34 @@ gm20x_secboot_prepare_blobs(struct gm200_secboot *gsb)
 	int ret;
 
 	/* Load and prepare the managed falcon's firmwares */
-	ret = gm200_secboot_prepare_ls_blob(gsb);
-	if (ret)
-		return ret;
+	if (!gsb->ls_blob) {
+		ret = gm200_secboot_prepare_ls_blob(gsb);
+		if (ret)
+			return ret;
+	}
 
 	/* Load the HS firmware that will load the LS firmwares */
-	ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_load",
-					    &gsb->acr_load_blob,
-					    &gsb->acr_load_bl_desc, true);
-	if (ret)
-		return ret;
+	if (!gsb->acr_load_blob) {
+		ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_load",
+						&gsb->acr_load_blob,
+						&gsb->acr_load_bl_desc, true);
+		if (ret)
+			return ret;
+	}
 
 	/* Load the HS firmware bootloader */
-	ret = gm200_secboot_prepare_hsbl_blob(gsb);
-	if (ret)
-		return ret;
+	if (!gsb->hsbl_blob) {
+		ret = gm200_secboot_prepare_hsbl_blob(gsb);
+		if (ret)
+			return ret;
+	}
 
 	return 0;
 }
 
 static int
-gm200_secboot_prepare_blobs(struct nvkm_secboot *sb)
+gm200_secboot_prepare_blobs(struct gm200_secboot *gsb)
 {
-	struct gm200_secboot *gsb = gm200_secboot(sb);
 	int ret;
 
 	ret = gm20x_secboot_prepare_blobs(gsb);
@@ -1053,15 +1060,37 @@ gm200_secboot_prepare_blobs(struct nvkm_secboot *sb)
 		return ret;
 
 	/* dGPU only: load the HS firmware that unprotects the WPR region */
-	ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_unload",
-					    &gsb->acr_unload_blob,
-					    &gsb->acr_unload_bl_desc, false);
-	if (ret)
-		return ret;
+	if (!gsb->acr_unload_blob) {
+		ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_unload",
+					       &gsb->acr_unload_blob,
+					       &gsb->acr_unload_bl_desc, false);
+		if (ret)
+			return ret;
+	}
 
 	return 0;
 }
 
+static int
+gm200_secboot_blobs_ready(struct gm200_secboot *gsb)
+{
+	struct nvkm_subdev *subdev = &gsb->base.subdev;
+	int ret;
+
+	/* firmware already loaded, nothing to do... */
+	if (gsb->firmware_ok)
+		return 0;
+
+	ret = gsb->func->prepare_blobs(gsb);
+	if (ret) {
+		nvkm_error(subdev, "failed to load secure firmware\n");
+		return ret;
+	}
+
+	gsb->firmware_ok = true;
+
+	return 0;
+}
 
 
 /*
@@ -1234,6 +1263,11 @@ gm200_secboot_reset(struct nvkm_secboot *sb, enum nvkm_secboot_falcon falcon)
 	struct gm200_secboot *gsb = gm200_secboot(sb);
 	int ret;
 
+	/* Make sure all blobs are ready */
+	ret = gm200_secboot_blobs_ready(gsb);
+	if (ret)
+		return ret;
+
 	/*
 	 * Dummy GM200 implementation: perform secure boot each time we are
 	 * called on FECS. Since only FECS and GPCCS are managed and started
@@ -1373,7 +1407,6 @@ gm200_secboot = {
 	.dtor = gm200_secboot_dtor,
 	.init = gm200_secboot_init,
 	.fini = gm200_secboot_fini,
-	.prepare_blobs = gm200_secboot_prepare_blobs,
 	.reset = gm200_secboot_reset,
 	.start = gm200_secboot_start,
 	.managed_falcons = BIT(NVKM_SECBOOT_FALCON_FECS) |
@@ -1415,6 +1448,7 @@ gm200_secboot_func = {
 	.bl_desc_size = sizeof(struct gm200_flcn_bl_desc),
 	.fixup_bl_desc = gm200_secboot_fixup_bl_desc,
 	.fixup_hs_desc = gm200_secboot_fixup_hs_desc,
+	.prepare_blobs = gm200_secboot_prepare_blobs,
 };
 
 int
@@ -1487,3 +1521,19 @@ MODULE_FIRMWARE("nvidia/gm206/gr/sw_ctx.bin");
 MODULE_FIRMWARE("nvidia/gm206/gr/sw_nonctx.bin");
 MODULE_FIRMWARE("nvidia/gm206/gr/sw_bundle_init.bin");
 MODULE_FIRMWARE("nvidia/gm206/gr/sw_method_init.bin");
+
+MODULE_FIRMWARE("nvidia/gp100/acr/bl.bin");
+MODULE_FIRMWARE("nvidia/gp100/acr/ucode_load.bin");
+MODULE_FIRMWARE("nvidia/gp100/acr/ucode_unload.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/fecs_bl.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/fecs_inst.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/fecs_data.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/fecs_sig.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_bl.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_inst.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_data.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_sig.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/sw_ctx.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/sw_nonctx.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/sw_bundle_init.bin");
+MODULE_FIRMWARE("nvidia/gp100/gr/sw_method_init.bin");
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c
index 684320484b70..d5395ebfe8d3 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c
@@ -42,6 +42,32 @@ struct gm20b_flcn_bl_desc {
 	u32 data_size;
 };
 
+static int
+gm20b_secboot_prepare_blobs(struct gm200_secboot *gsb)
+{
+	struct nvkm_subdev *subdev = &gsb->base.subdev;
+	int acr_size;
+	int ret;
+
+	ret = gm20x_secboot_prepare_blobs(gsb);
+	if (ret)
+		return ret;
+
+	acr_size = gsb->acr_load_blob->size;
+	/*
+	 * On Tegra the WPR region is set by the bootloader. It is illegal for
+	 * the HS blob to be larger than this region.
+	 */
+	if (acr_size > gsb->wpr_size) {
+		nvkm_error(subdev, "WPR region too small for FW blob!\n");
+		nvkm_error(subdev, "required: %dB\n", acr_size);
+		nvkm_error(subdev, "WPR size: %dB\n", gsb->wpr_size);
+		return -ENOSPC;
+	}
+
+	return 0;
+}
+
 /**
  * gm20b_secboot_fixup_bl_desc - adapt BL descriptor to format used by GM20B FW
  *
@@ -88,6 +114,7 @@ gm20b_secboot_func = {
 	.bl_desc_size = sizeof(struct gm20b_flcn_bl_desc),
 	.fixup_bl_desc = gm20b_secboot_fixup_bl_desc,
 	.fixup_hs_desc = gm20b_secboot_fixup_hs_desc,
+	.prepare_blobs = gm20b_secboot_prepare_blobs,
 };
 
 
@@ -147,32 +174,6 @@ gm20b_tegra_read_wpr(struct gm200_secboot *gsb)
 #endif
 
 static int
-gm20b_secboot_prepare_blobs(struct nvkm_secboot *sb)
-{
-	struct gm200_secboot *gsb = gm200_secboot(sb);
-	int acr_size;
-	int ret;
-
-	ret = gm20x_secboot_prepare_blobs(gsb);
-	if (ret)
-		return ret;
-
-	acr_size = gsb->acr_load_blob->size;
-	/*
-	 * On Tegra the WPR region is set by the bootloader. It is illegal for
-	 * the HS blob to be larger than this region.
-	 */
-	if (acr_size > gsb->wpr_size) {
-		nvkm_error(&sb->subdev, "WPR region too small for FW blob!\n");
-		nvkm_error(&sb->subdev, "required: %dB\n", acr_size);
-		nvkm_error(&sb->subdev, "WPR size: %dB\n", gsb->wpr_size);
-		return -ENOSPC;
-	}
-
-	return 0;
-}
-
-static int
 gm20b_secboot_init(struct nvkm_secboot *sb)
 {
 	struct gm200_secboot *gsb = gm200_secboot(sb);
@@ -189,7 +190,6 @@ static const struct nvkm_secboot_func
 gm20b_secboot = {
 	.dtor = gm200_secboot_dtor,
 	.init = gm20b_secboot_init,
-	.prepare_blobs = gm20b_secboot_prepare_blobs,
 	.reset = gm200_secboot_reset,
 	.start = gm200_secboot_start,
 	.managed_falcons = BIT(NVKM_SECBOOT_FALCON_FECS),
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h
index f2b09dee7c5d..a9a8a0e1017e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h
@@ -30,7 +30,6 @@ struct nvkm_secboot_func {
 	int (*init)(struct nvkm_secboot *);
 	int (*fini)(struct nvkm_secboot *, bool suspend);
 	void *(*dtor)(struct nvkm_secboot *);
-	int (*prepare_blobs)(struct nvkm_secboot *);
 	int (*reset)(struct nvkm_secboot *, enum nvkm_secboot_falcon);
 	int (*start)(struct nvkm_secboot *, enum nvkm_secboot_falcon);
 
@@ -147,10 +146,8 @@ struct hsflcn_acr_desc {
  * @inst:		instance block for HS falcon
  * @pgd:		page directory for the HS falcon
  * @vm:			address space used by the HS falcon
- * @bl_desc_size:	size of the BL descriptor used by this chip.
- * @fixup_bl_desc:	hook that generates the proper BL descriptor format from
- *			the generic GM200 format into a data array of size
- *			bl_desc_size
+ * @falcon_state:	current state of the managed falcons
+ * @firmware_ok:	whether the firmware blobs have been created
  */
 struct gm200_secboot {
 	struct nvkm_secboot base;
@@ -196,9 +193,19 @@ struct gm200_secboot {
 		RUNNING,
 	} falcon_state[NVKM_SECBOOT_FALCON_END];
 
+	bool firmware_ok;
 };
 #define gm200_secboot(sb) container_of(sb, struct gm200_secboot, base)
 
+/**
+ * Contains functions we wish to abstract between GM200-like implementations
+ * @bl_desc_size:	size of the BL descriptor used by this chip.
+ * @fixup_bl_desc:	hook that generates the proper BL descriptor format from
+ *			the generic GM200 format into a data array of size
+ *			bl_desc_size
+ * @fixup_hs_desc:	hook that twiddles the HS descriptor before it is used
+ * @prepare_blobs:	prepares the various blobs needed for secure booting
+ */
 struct gm200_secboot_func {
 	/*
 	 * Size of the bootloader descriptor for this chip. A block of this
@@ -214,6 +221,7 @@ struct gm200_secboot_func {
 	 * we want the HS FW to set up.
 	 */
 	void (*fixup_hs_desc)(struct gm200_secboot *, struct hsflcn_acr_desc *);
+	int (*prepare_blobs)(struct gm200_secboot *);
 };
 
 int gm200_secboot_init(struct nvkm_secboot *);
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
index a1b264664aad..fe063d5728e2 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c
@@ -41,8 +41,9 @@ nvkm_top_device_new(struct nvkm_top *top)
 }
 
 u32
-nvkm_top_reset(struct nvkm_top *top, enum nvkm_devidx index)
+nvkm_top_reset(struct nvkm_device *device, enum nvkm_devidx index)
 {
+	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	if (top) {
@@ -56,8 +57,25 @@ nvkm_top_reset(struct nvkm_top *top, enum nvkm_devidx index)
 }
 
 u32
-nvkm_top_intr(struct nvkm_top *top, u32 intr, u64 *psubdevs)
+nvkm_top_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx)
 {
+	struct nvkm_top *top = device->top;
+	struct nvkm_top_device *info;
+
+	if (top) {
+		list_for_each_entry(info, &top->device, head) {
+			if (info->index == devidx && info->intr >= 0)
+				return BIT(info->intr);
+		}
+	}
+
+	return 0;
+}
+
+u32
+nvkm_top_intr(struct nvkm_device *device, u32 intr, u64 *psubdevs)
+{
+	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 	u64 subdevs = 0;
 	u32 handled = 0;
@@ -78,8 +96,9 @@ nvkm_top_intr(struct nvkm_top *top, u32 intr, u64 *psubdevs)
 }
 
 enum nvkm_devidx
-nvkm_top_fault(struct nvkm_top *top, int fault)
+nvkm_top_fault(struct nvkm_device *device, int fault)
 {
+	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 
 	list_for_each_entry(info, &top->device, head) {
@@ -91,8 +110,9 @@ nvkm_top_fault(struct nvkm_top *top, int fault)
 }
 
 enum nvkm_devidx
-nvkm_top_engine(struct nvkm_top *top, int index, int *runl, int *engn)
+nvkm_top_engine(struct nvkm_device *device, int index, int *runl, int *engn)
 {
+	struct nvkm_top *top = device->top;
 	struct nvkm_top_device *info;
 	int n = 0;
 
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
index e06acc340e99..efac3402f9dd 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c
@@ -29,7 +29,7 @@ gk104_top_oneinit(struct nvkm_top *top)
 	struct nvkm_subdev *subdev = &top->subdev;
 	struct nvkm_device *device = subdev->device;
 	struct nvkm_top_device *info = NULL;
-	u32 data, type;
+	u32 data, type, inst;
 	int i;
 
 	for (i = 0; i < 64; i++) {
@@ -37,6 +37,7 @@ gk104_top_oneinit(struct nvkm_top *top)
 			if (!(info = nvkm_top_device_new(top)))
 				return -ENOMEM;
 			type = ~0;
+			inst = 0;
 		}
 
 		data = nvkm_rd32(device, 0x022700 + (i * 0x04));
@@ -45,6 +46,7 @@ gk104_top_oneinit(struct nvkm_top *top)
 		case 0x00000000: /* NOT_VALID */
 			continue;
 		case 0x00000001: /* DATA */
+			inst        = (data & 0x3c000000) >> 26;
 			info->addr  = (data & 0x00fff000);
 			info->fault = (data & 0x000000f8) >> 3;
 			break;
@@ -67,27 +69,32 @@ gk104_top_oneinit(struct nvkm_top *top)
 			continue;
 
 		/* Translate engine type to NVKM engine identifier. */
+#define A_(A) if (inst == 0) info->index = NVKM_ENGINE_##A
+#define B_(A) if (inst + NVKM_ENGINE_##A##0 < NVKM_ENGINE_##A##_LAST + 1)      \
+		info->index = NVKM_ENGINE_##A##0 + inst
 		switch (type) {
-		case 0x00000000: info->index = NVKM_ENGINE_GR; break;
-		case 0x00000001: info->index = NVKM_ENGINE_CE0; break;
-		case 0x00000002: info->index = NVKM_ENGINE_CE1; break;
-		case 0x00000003: info->index = NVKM_ENGINE_CE2; break;
-		case 0x00000008: info->index = NVKM_ENGINE_MSPDEC; break;
-		case 0x00000009: info->index = NVKM_ENGINE_MSPPP; break;
-		case 0x0000000a: info->index = NVKM_ENGINE_MSVLD; break;
-		case 0x0000000b: info->index = NVKM_ENGINE_MSENC; break;
-		case 0x0000000c: info->index = NVKM_ENGINE_VIC; break;
-		case 0x0000000d: info->index = NVKM_ENGINE_SEC; break;
-		case 0x0000000e: info->index = NVKM_ENGINE_NVENC0; break;
-		case 0x0000000f: info->index = NVKM_ENGINE_NVENC1; break;
-		case 0x00000010: info->index = NVKM_ENGINE_NVDEC; break;
+		case 0x00000000: A_(GR    ); break;
+		case 0x00000001: A_(CE0   ); break;
+		case 0x00000002: A_(CE1   ); break;
+		case 0x00000003: A_(CE2   ); break;
+		case 0x00000008: A_(MSPDEC); break;
+		case 0x00000009: A_(MSPPP ); break;
+		case 0x0000000a: A_(MSVLD ); break;
+		case 0x0000000b: A_(MSENC ); break;
+		case 0x0000000c: A_(VIC   ); break;
+		case 0x0000000d: A_(SEC   ); break;
+		case 0x0000000e: B_(NVENC ); break;
+		case 0x0000000f: A_(NVENC1); break;
+		case 0x00000010: A_(NVDEC ); break;
+		case 0x00000013: B_(CE    ); break;
 			break;
 		default:
 			break;
 		}
 
-		nvkm_debug(subdev, "%02x (%8s): addr %06x fault %2d engine %2d "
-				   "runlist %2d intr %2d reset %2d\n", type,
+		nvkm_debug(subdev, "%02x.%d (%8s): addr %06x fault %2d "
+				   "engine %2d runlist %2d intr %2d "
+				   "reset %2d\n", type, inst,
 			   info->index == NVKM_SUBDEV_NR ? NULL :
 					  nvkm_subdev_name[info->index],
 			   info->addr, info->fault, info->engine, info->runlist,
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
index 6b2d7531a7ff..1c3d23b0e84a 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c
@@ -120,6 +120,8 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt)
 
 	data = nvbios_volt_parse(bios, &ver, &hdr, &cnt, &len, &info);
 	if (data && info.vidmask && info.base && info.step) {
+		volt->min_uv = info.min;
+		volt->max_uv = info.max;
 		for (i = 0; i < info.vidmask + 1; i++) {
 			if (info.base >= info.min &&
 				info.base <= info.max) {
@@ -131,6 +133,8 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt)
 		}
 		volt->vid_mask = info.vidmask;
 	} else if (data && info.vidmask) {
+		volt->min_uv = 0xffffffff;
+		volt->max_uv = 0;
 		for (i = 0; i < cnt; i++) {
 			data = nvbios_volt_entry_parse(bios, i, &ver, &hdr,
 						       &ivid);
@@ -138,9 +142,14 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt)
 				volt->vid[volt->vid_nr].uv = ivid.voltage;
 				volt->vid[volt->vid_nr].vid = ivid.vid;
 				volt->vid_nr++;
+				volt->min_uv = min(volt->min_uv, ivid.voltage);
+				volt->max_uv = max(volt->max_uv, ivid.voltage);
 			}
 		}
 		volt->vid_mask = info.vidmask;
+	} else if (data && info.type == NVBIOS_VOLT_PWM) {
+		volt->min_uv = info.base;
+		volt->max_uv = info.base + info.pwm_range;
 	}
 }
 
@@ -181,8 +190,11 @@ nvkm_volt_ctor(const struct nvkm_volt_func *func, struct nvkm_device *device,
 	volt->func = func;
 
 	/* Assuming the non-bios device should build the voltage table later */
-	if (bios)
+	if (bios) {
 		nvkm_volt_parse_bios(bios, volt);
+		nvkm_debug(&volt->subdev, "min: %iuv max: %iuv\n",
+			   volt->min_uv, volt->max_uv);
+	}
 
 	if (volt->vid_nr) {
 		for (i = 0; i < volt->vid_nr; i++) {
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
index d554455326da..ce5d83cdc7cf 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c
@@ -77,18 +77,19 @@ gk20a_volt_get_cvb_t_voltage(int speedo, int temp, int s_scale, int t_scale,
 	return mv;
 }
 
-int
+static int
 gk20a_volt_calc_voltage(const struct cvb_coef *coef, int speedo)
 {
+	static const int v_scale = 1000;
 	int mv;
 
 	mv = gk20a_volt_get_cvb_t_voltage(speedo, -10, 100, 10, coef);
-	mv = DIV_ROUND_UP(mv, 1000);
+	mv = DIV_ROUND_UP(mv, v_scale);
 
 	return mv * 1000;
 }
 
-int
+static int
 gk20a_volt_vid_get(struct nvkm_volt *base)
 {
 	struct gk20a_volt *volt = gk20a_volt(base);
@@ -103,7 +104,7 @@ gk20a_volt_vid_get(struct nvkm_volt *base)
 	return -EINVAL;
 }
 
-int
+static int
 gk20a_volt_vid_set(struct nvkm_volt *base, u8 vid)
 {
 	struct gk20a_volt *volt = gk20a_volt(base);
@@ -113,7 +114,7 @@ gk20a_volt_vid_set(struct nvkm_volt *base, u8 vid)
 	return regulator_set_voltage(volt->vdd, volt->base.vid[vid].uv, 1200000);
 }
 
-int
+static int
 gk20a_volt_set_id(struct nvkm_volt *base, u8 id, int condition)
 {
 	struct gk20a_volt *volt = gk20a_volt(base);
@@ -143,9 +144,9 @@ gk20a_volt = {
 };
 
 int
-_gk20a_volt_ctor(struct nvkm_device *device, int index,
-		 const struct cvb_coef *coefs, int nb_coefs,
-		 struct gk20a_volt *volt)
+gk20a_volt_ctor(struct nvkm_device *device, int index,
+		const struct cvb_coef *coefs, int nb_coefs,
+		int vmin, struct gk20a_volt *volt)
 {
 	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	int i, uv;
@@ -160,9 +161,9 @@ _gk20a_volt_ctor(struct nvkm_device *device, int index,
 	volt->base.vid_nr = nb_coefs;
 	for (i = 0; i < volt->base.vid_nr; i++) {
 		volt->base.vid[i].vid = i;
-		volt->base.vid[i].uv =
-			gk20a_volt_calc_voltage(&coefs[i],
-						tdev->gpu_speedo);
+		volt->base.vid[i].uv = max(
+			gk20a_volt_calc_voltage(&coefs[i], tdev->gpu_speedo),
+			vmin);
 		nvkm_debug(&volt->base.subdev, "%2d: vid=%d, uv=%d\n", i,
 			   volt->base.vid[i].vid, volt->base.vid[i].uv);
 	}
@@ -180,6 +181,6 @@ gk20a_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
 		return -ENOMEM;
 	*pvolt = &volt->base;
 
-	return _gk20a_volt_ctor(device, index, gk20a_cvb_coef,
-				ARRAY_SIZE(gk20a_cvb_coef), volt);
+	return gk20a_volt_ctor(device, index, gk20a_cvb_coef,
+			       ARRAY_SIZE(gk20a_cvb_coef), 0, volt);
 }
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
index 0fa3b502bcf8..6a6c97f9684e 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h
@@ -37,13 +37,8 @@ struct gk20a_volt {
 	struct regulator *vdd;
 };
 
-int _gk20a_volt_ctor(struct nvkm_device *device, int index,
-		     const struct cvb_coef *coefs, int nb_coefs,
-		     struct gk20a_volt *volt);
-
-int gk20a_volt_calc_voltage(const struct cvb_coef *coef, int speedo);
-int gk20a_volt_vid_get(struct nvkm_volt *volt);
-int gk20a_volt_vid_set(struct nvkm_volt *volt, u8 vid);
-int gk20a_volt_set_id(struct nvkm_volt *volt, u8 id, int condition);
+int gk20a_volt_ctor(struct nvkm_device *device, int index,
+		    const struct cvb_coef *coefs, int nb_coefs,
+		    int vmin, struct gk20a_volt *volt);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
index 49b5ecb701e4..74db4d28930f 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c
@@ -41,16 +41,52 @@ const struct cvb_coef gm20b_cvb_coef[] = {
 	/* 921600 */ { 2647676, -106455, 1632 },
 };
 
+static const struct cvb_coef gm20b_na_cvb_coef[] = {
+	/* KHz,         c0,     c1,   c2,    c3,     c4,   c5 */
+	/*  76800 */ {  814294, 8144, -940, 808, -21583, 226 },
+	/* 153600 */ {  856185, 8144, -940, 808, -21583, 226 },
+	/* 230400 */ {  898077, 8144, -940, 808, -21583, 226 },
+	/* 307200 */ {  939968, 8144, -940, 808, -21583, 226 },
+	/* 384000 */ {  981860, 8144, -940, 808, -21583, 226 },
+	/* 460800 */ { 1023751, 8144, -940, 808, -21583, 226 },
+	/* 537600 */ { 1065642, 8144, -940, 808, -21583, 226 },
+	/* 614400 */ { 1107534, 8144, -940, 808, -21583, 226 },
+	/* 691200 */ { 1149425, 8144, -940, 808, -21583, 226 },
+	/* 768000 */ { 1191317, 8144, -940, 808, -21583, 226 },
+	/* 844800 */ { 1233208, 8144, -940, 808, -21583, 226 },
+	/* 921600 */ { 1275100, 8144, -940, 808, -21583, 226 },
+	/* 998400 */ { 1316991, 8144, -940, 808, -21583, 226 },
+};
+
+const u32 speedo_to_vmin[] = {
+	/*   0,      1,      2,      3,      4, */
+	950000, 840000, 818750, 840000, 810000,
+};
+
 int
 gm20b_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt)
 {
+	struct nvkm_device_tegra *tdev = device->func->tegra(device);
 	struct gk20a_volt *volt;
+	u32 vmin;
+
+	if (tdev->gpu_speedo_id >= ARRAY_SIZE(speedo_to_vmin)) {
+		nvdev_error(device, "unsupported speedo %d\n",
+			    tdev->gpu_speedo_id);
+		return -EINVAL;
+	}
 
 	volt = kzalloc(sizeof(*volt), GFP_KERNEL);
 	if (!volt)
 		return -ENOMEM;
 	*pvolt = &volt->base;
 
-	return _gk20a_volt_ctor(device, index, gm20b_cvb_coef,
-				ARRAY_SIZE(gm20b_cvb_coef), volt);
+	vmin = speedo_to_vmin[tdev->gpu_speedo_id];
+
+	if (tdev->gpu_speedo_id >= 1)
+		return gk20a_volt_ctor(device, index, gm20b_na_cvb_coef,
+				     ARRAY_SIZE(gm20b_na_cvb_coef), vmin, volt);
+	else
+		return gk20a_volt_ctor(device, index, gm20b_cvb_coef,
+					ARRAY_SIZE(gm20b_cvb_coef), vmin, volt);
 }
diff --git a/drivers/gpu/drm/omapdrm/dss/dsi.c b/drivers/gpu/drm/omapdrm/dss/dsi.c
index 6f45e9d00b41..e1be5e795cd8 100644
--- a/drivers/gpu/drm/omapdrm/dss/dsi.c
+++ b/drivers/gpu/drm/omapdrm/dss/dsi.c
@@ -1167,7 +1167,6 @@ static int dsi_regulator_init(struct platform_device *dsidev)
 {
 	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
 	struct regulator *vdds_dsi;
-	int r;
 
 	if (dsi->vdds_dsi_reg != NULL)
 		return 0;
diff --git a/drivers/gpu/drm/omapdrm/dss/hdmi5.c b/drivers/gpu/drm/omapdrm/dss/hdmi5.c
index 061f9bab4c9b..0c0a5139a301 100644
--- a/drivers/gpu/drm/omapdrm/dss/hdmi5.c
+++ b/drivers/gpu/drm/omapdrm/dss/hdmi5.c
@@ -120,7 +120,6 @@ static irqreturn_t hdmi_irq_handler(int irq, void *data)
 
 static int hdmi_init_regulator(void)
 {
-	int r;
 	struct regulator *reg;
 
 	if (hdmi.vdda_reg != NULL)
diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c
index 3a7bdf1c842b..85143d1b9b31 100644
--- a/drivers/gpu/drm/panel/panel-simple.c
+++ b/drivers/gpu/drm/panel/panel-simple.c
@@ -168,6 +168,7 @@ static int panel_simple_disable(struct drm_panel *panel)
 
 	if (p->backlight) {
 		p->backlight->props.power = FB_BLANK_POWERDOWN;
+		p->backlight->props.state |= BL_CORE_FBBLANK;
 		backlight_update_status(p->backlight);
 	}
 
@@ -235,6 +236,7 @@ static int panel_simple_enable(struct drm_panel *panel)
 		msleep(p->desc->delay.enable);
 
 	if (p->backlight) {
+		p->backlight->props.state &= ~BL_CORE_FBBLANK;
 		p->backlight->props.power = FB_BLANK_UNBLANK;
 		backlight_update_status(p->backlight);
 	}
@@ -964,8 +966,8 @@ static const struct panel_desc innolux_zj070na_01p = {
 	.num_modes = 1,
 	.bpc = 6,
 	.size = {
-		.width = 1024,
-		.height = 600,
+		.width = 154,
+		.height = 90,
 	},
 };
 
@@ -1017,6 +1019,51 @@ static const struct panel_desc lg_lb070wv8 = {
 	.bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_SPWG,
 };
 
+static const struct drm_display_mode lg_lp079qx1_sp0v_mode = {
+	.clock = 200000,
+	.hdisplay = 1536,
+	.hsync_start = 1536 + 12,
+	.hsync_end = 1536 + 12 + 16,
+	.htotal = 1536 + 12 + 16 + 48,
+	.vdisplay = 2048,
+	.vsync_start = 2048 + 8,
+	.vsync_end = 2048 + 8 + 4,
+	.vtotal = 2048 + 8 + 4 + 8,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc lg_lp079qx1_sp0v = {
+	.modes = &lg_lp079qx1_sp0v_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 129,
+		.height = 171,
+	},
+};
+
+static const struct drm_display_mode lg_lp097qx1_spa1_mode = {
+	.clock = 205210,
+	.hdisplay = 2048,
+	.hsync_start = 2048 + 150,
+	.hsync_end = 2048 + 150 + 5,
+	.htotal = 2048 + 150 + 5 + 5,
+	.vdisplay = 1536,
+	.vsync_start = 1536 + 3,
+	.vsync_end = 1536 + 3 + 1,
+	.vtotal = 1536 + 3 + 1 + 9,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc lg_lp097qx1_spa1 = {
+	.modes = &lg_lp097qx1_spa1_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 208,
+		.height = 147,
+	},
+};
+
 static const struct drm_display_mode lg_lp120up1_mode = {
 	.clock = 162300,
 	.hdisplay = 1920,
@@ -1224,6 +1271,28 @@ static const struct panel_desc qd43003c0_40 = {
 	.bus_format = MEDIA_BUS_FMT_RGB888_1X24,
 };
 
+static const struct drm_display_mode samsung_lsn122dl01_c01_mode = {
+	.clock = 271560,
+	.hdisplay = 2560,
+	.hsync_start = 2560 + 48,
+	.hsync_end = 2560 + 48 + 32,
+	.htotal = 2560 + 48 + 32 + 80,
+	.vdisplay = 1600,
+	.vsync_start = 1600 + 2,
+	.vsync_end = 1600 + 2 + 5,
+	.vtotal = 1600 + 2 + 5 + 57,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc samsung_lsn122dl01_c01 = {
+	.modes = &samsung_lsn122dl01_c01_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 263,
+		.height = 164,
+	},
+};
+
 static const struct drm_display_mode samsung_ltn101nt05_mode = {
 	.clock = 54030,
 	.hdisplay = 1024,
@@ -1242,8 +1311,8 @@ static const struct panel_desc samsung_ltn101nt05 = {
 	.num_modes = 1,
 	.bpc = 6,
 	.size = {
-		.width = 1024,
-		.height = 600,
+		.width = 223,
+		.height = 125,
 	},
 };
 
@@ -1270,6 +1339,53 @@ static const struct panel_desc samsung_ltn140at29_301 = {
 	},
 };
 
+static const struct display_timing sharp_lq101k1ly04_timing = {
+	.pixelclock = { 60000000, 65000000, 80000000 },
+	.hactive = { 1280, 1280, 1280 },
+	.hfront_porch = { 20, 20, 20 },
+	.hback_porch = { 20, 20, 20 },
+	.hsync_len = { 10, 10, 10 },
+	.vactive = { 800, 800, 800 },
+	.vfront_porch = { 4, 4, 4 },
+	.vback_porch = { 4, 4, 4 },
+	.vsync_len = { 4, 4, 4 },
+	.flags = DISPLAY_FLAGS_PIXDATA_POSEDGE,
+};
+
+static const struct panel_desc sharp_lq101k1ly04 = {
+	.timings = &sharp_lq101k1ly04_timing,
+	.num_timings = 1,
+	.bpc = 8,
+	.size = {
+		.width = 217,
+		.height = 136,
+	},
+	.bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA,
+};
+
+static const struct drm_display_mode sharp_lq123p1jx31_mode = {
+	.clock = 252750,
+	.hdisplay = 2400,
+	.hsync_start = 2400 + 48,
+	.hsync_end = 2400 + 48 + 32,
+	.htotal = 2400 + 48 + 32 + 80,
+	.vdisplay = 1600,
+	.vsync_start = 1600 + 3,
+	.vsync_end = 1600 + 3 + 10,
+	.vtotal = 1600 + 3 + 10 + 33,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc sharp_lq123p1jx31 = {
+	.modes = &sharp_lq123p1jx31_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 259,
+		.height = 173,
+	},
+};
+
 static const struct drm_display_mode shelly_sca07010_bfn_lnn_mode = {
 	.clock = 33300,
 	.hdisplay = 800,
@@ -1293,6 +1409,29 @@ static const struct panel_desc shelly_sca07010_bfn_lnn = {
 	.bus_format = MEDIA_BUS_FMT_RGB666_1X18,
 };
 
+static const struct drm_display_mode starry_kr122ea0sra_mode = {
+	.clock = 147000,
+	.hdisplay = 1920,
+	.hsync_start = 1920 + 16,
+	.hsync_end = 1920 + 16 + 16,
+	.htotal = 1920 + 16 + 16 + 32,
+	.vdisplay = 1200,
+	.vsync_start = 1200 + 15,
+	.vsync_end = 1200 + 15 + 2,
+	.vtotal = 1200 + 15 + 2 + 18,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc starry_kr122ea0sra = {
+	.modes = &starry_kr122ea0sra_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 263,
+		.height = 164,
+	},
+};
+
 static const struct drm_display_mode tpk_f07a_0102_mode = {
 	.clock = 33260,
 	.hdisplay = 800,
@@ -1457,6 +1596,12 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "lg,lb070wv8",
 		.data = &lg_lb070wv8,
 	}, {
+		.compatible = "lg,lp079qx1-sp0v",
+		.data = &lg_lp079qx1_sp0v,
+	}, {
+		.compatible = "lg,lp097qx1-spa1",
+		.data = &lg_lp097qx1_spa1,
+	}, {
 		.compatible = "lg,lp120up1",
 		.data = &lg_lp120up1,
 	}, {
@@ -1481,15 +1626,27 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "qiaodian,qd43003c0-40",
 		.data = &qd43003c0_40,
 	}, {
+		.compatible = "samsung,lsn122dl01-c01",
+		.data = &samsung_lsn122dl01_c01,
+	}, {
 		.compatible = "samsung,ltn101nt05",
 		.data = &samsung_ltn101nt05,
 	}, {
 		.compatible = "samsung,ltn140at29-301",
 		.data = &samsung_ltn140at29_301,
 	}, {
+		.compatible = "sharp,lq101k1ly04",
+		.data = &sharp_lq101k1ly04,
+	}, {
+		.compatible = "sharp,lq123p1jx31",
+		.data = &sharp_lq123p1jx31,
+	}, {
 		.compatible = "shelly,sca07010-bfn-lnn",
 		.data = &shelly_sca07010_bfn_lnn,
 	}, {
+		.compatible = "starry,kr122ea0sra",
+		.data = &starry_kr122ea0sra,
+	}, {
 		.compatible = "tpk,f07a-0102",
 		.data = &tpk_f07a_0102,
 	}, {
@@ -1701,7 +1858,6 @@ static const struct panel_desc_dsi panasonic_vvx10f004b00 = {
 	.lanes = 4,
 };
 
-
 static const struct of_device_id dsi_of_match[] = {
 	{
 		.compatible = "auo,b080uan01",
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 0738d74c8d04..d50c9679e631 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -350,11 +350,19 @@ static int qxl_bo_move(struct ttm_buffer_object *bo,
 		       struct ttm_mem_reg *new_mem)
 {
 	struct ttm_mem_reg *old_mem = &bo->mem;
+	int ret;
+
+	ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+	if (ret)
+		return ret;
+
+
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
 		qxl_move_null(bo, new_mem);
 		return 0;
 	}
-	return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+	return ttm_bo_move_memcpy(bo, evict, interruptible,
+				  no_wait_gpu, new_mem);
 }
 
 static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index e91763d5d800..a97abc8af657 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -589,7 +589,8 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 		/* use frac fb div on RS780/RS880 */
-		if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+		if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+		    && !radeon_crtc->ss_enabled)
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 		if (ASIC_IS_DCE32(rdev) && mode->clock > 165000)
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
@@ -626,7 +627,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 			if (radeon_crtc->ss.refdiv) {
 				radeon_crtc->pll_flags |= RADEON_PLL_USE_REF_DIV;
 				radeon_crtc->pll_reference_div = radeon_crtc->ss.refdiv;
-				if (ASIC_IS_AVIVO(rdev))
+				if (rdev->family >= CHIP_RV770)
 					radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 			}
 		}
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index ba192a35c607..0c1b9ff433af 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -53,6 +53,7 @@ MODULE_FIRMWARE("radeon/bonaire_mc.bin");
 MODULE_FIRMWARE("radeon/bonaire_rlc.bin");
 MODULE_FIRMWARE("radeon/bonaire_sdma.bin");
 MODULE_FIRMWARE("radeon/bonaire_smc.bin");
+MODULE_FIRMWARE("radeon/bonaire_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/HAWAII_pfp.bin");
 MODULE_FIRMWARE("radeon/HAWAII_me.bin");
@@ -72,6 +73,7 @@ MODULE_FIRMWARE("radeon/hawaii_mc.bin");
 MODULE_FIRMWARE("radeon/hawaii_rlc.bin");
 MODULE_FIRMWARE("radeon/hawaii_sdma.bin");
 MODULE_FIRMWARE("radeon/hawaii_smc.bin");
+MODULE_FIRMWARE("radeon/hawaii_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/KAVERI_pfp.bin");
 MODULE_FIRMWARE("radeon/KAVERI_me.bin");
@@ -1990,12 +1992,17 @@ static int cik_init_microcode(struct radeon_device *rdev)
 	int new_fw = 0;
 	int err;
 	int num_fw;
+	bool new_smc = false;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
 		chip_name = "BONAIRE";
+		if ((rdev->pdev->revision == 0x80) ||
+		    (rdev->pdev->revision == 0x81) ||
+		    (rdev->pdev->device == 0x665f))
+			new_smc = true;
 		new_chip_name = "bonaire";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
@@ -2010,6 +2017,8 @@ static int cik_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_HAWAII:
 		chip_name = "HAWAII";
+		if (rdev->pdev->revision == 0x80)
+			new_smc = true;
 		new_chip_name = "hawaii";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
@@ -2259,7 +2268,10 @@ static int cik_init_microcode(struct radeon_device *rdev)
 			}
 		}
 
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
+		if (new_smc)
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name);
+		else
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 		if (err) {
 			snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
@@ -8354,7 +8366,8 @@ static int cik_startup(struct radeon_device *rdev)
 		}
 	}
 	rdev->rlc.cs_data = ci_cs_data;
-	rdev->rlc.cp_table_size = CP_ME_TABLE_SIZE * 5 * 4;
+	rdev->rlc.cp_table_size = ALIGN(CP_ME_TABLE_SIZE * 5 * 4, 2048); /* CP JT */
+	rdev->rlc.cp_table_size += 64 * 1024; /* GDS */
 	r = sumo_rlc_init(rdev);
 	if (r) {
 		DRM_ERROR("Failed to init rlc BOs!\n");
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 0d3f744de35a..d960d3915408 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -2209,6 +2209,12 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		break;
 	}
+	case PACKET3_PFP_SYNC_ME:
+		if (pkt->count) {
+			DRM_ERROR("bad PFP_SYNC_ME\n");
+			return -EINVAL;
+		}
+		break;
 	case PACKET3_SURFACE_SYNC:
 		if (pkt->count != 3) {
 			DRM_ERROR("bad SURFACE_SYNC\n");
@@ -3381,6 +3387,7 @@ static int evergreen_vm_packet3_check(struct radeon_device *rdev,
 	case PACKET3_MPEG_INDEX:
 	case PACKET3_WAIT_REG_MEM:
 	case PACKET3_MEM_WRITE:
+	case PACKET3_PFP_SYNC_ME:
 	case PACKET3_SURFACE_SYNC:
 	case PACKET3_EVENT_WRITE:
 	case PACKET3_EVENT_WRITE_EOP:
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 0b174e14e9a6..c8e3d394cde7 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -1624,6 +1624,7 @@
 		 */
 #              define PACKET3_CP_DMA_CMD_SAIC      (1 << 28)
 #              define PACKET3_CP_DMA_CMD_DAIC      (1 << 29)
+#define	PACKET3_PFP_SYNC_ME				0x42
 #define	PACKET3_SURFACE_SYNC				0x43
 #              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
 #              define PACKET3_CB1_DEST_BASE_ENA    (1 << 7)
diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index 95f4fea89302..86dcdf38b732 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -10,6 +10,7 @@
 #include <linux/slab.h>
 #include <linux/acpi.h>
 #include <linux/pci.h>
+#include <linux/delay.h>
 
 #include "radeon_acpi.h"
 
@@ -27,6 +28,7 @@ struct radeon_atpx_functions {
 struct radeon_atpx {
 	acpi_handle handle;
 	struct radeon_atpx_functions functions;
+	bool is_hybrid;
 };
 
 static struct radeon_atpx_priv {
@@ -62,6 +64,14 @@ bool radeon_has_atpx(void) {
 	return radeon_atpx_priv.atpx_detected;
 }
 
+bool radeon_has_atpx_dgpu_power_cntl(void) {
+	return radeon_atpx_priv.atpx.functions.power_cntl;
+}
+
+bool radeon_is_atpx_hybrid(void) {
+	return radeon_atpx_priv.atpx.is_hybrid;
+}
+
 /**
  * radeon_atpx_call - call an ATPX method
  *
@@ -141,18 +151,12 @@ static void radeon_atpx_parse_functions(struct radeon_atpx_functions *f, u32 mas
  */
 static int radeon_atpx_validate(struct radeon_atpx *atpx)
 {
-	/* make sure required functions are enabled */
-	/* dGPU power control is required */
-	if (atpx->functions.power_cntl == false) {
-		printk("ATPX dGPU power cntl not present, forcing\n");
-		atpx->functions.power_cntl = true;
-	}
+	u32 valid_bits = 0;
 
 	if (atpx->functions.px_params) {
 		union acpi_object *info;
 		struct atpx_px_params output;
 		size_t size;
-		u32 valid_bits;
 
 		info = radeon_atpx_call(atpx->handle, ATPX_FUNCTION_GET_PX_PARAMETERS, NULL);
 		if (!info)
@@ -171,19 +175,42 @@ static int radeon_atpx_validate(struct radeon_atpx *atpx)
 		memcpy(&output, info->buffer.pointer, size);
 
 		valid_bits = output.flags & output.valid_flags;
-		/* if separate mux flag is set, mux controls are required */
-		if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) {
-			atpx->functions.i2c_mux_cntl = true;
-			atpx->functions.disp_mux_cntl = true;
-		}
-		/* if any outputs are muxed, mux controls are required */
-		if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED |
-				  ATPX_TV_SIGNAL_MUXED |
-				  ATPX_DFP_SIGNAL_MUXED))
-			atpx->functions.disp_mux_cntl = true;
 
 		kfree(info);
 	}
+
+	/* if separate mux flag is set, mux controls are required */
+	if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) {
+		atpx->functions.i2c_mux_cntl = true;
+		atpx->functions.disp_mux_cntl = true;
+	}
+	/* if any outputs are muxed, mux controls are required */
+	if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED |
+			  ATPX_TV_SIGNAL_MUXED |
+			  ATPX_DFP_SIGNAL_MUXED))
+		atpx->functions.disp_mux_cntl = true;
+
+	/* some bioses set these bits rather than flagging power_cntl as supported */
+	if (valid_bits & (ATPX_DYNAMIC_PX_SUPPORTED |
+			  ATPX_DYNAMIC_DGPU_POWER_OFF_SUPPORTED))
+		atpx->functions.power_cntl = true;
+
+	atpx->is_hybrid = false;
+	if (valid_bits & ATPX_MS_HYBRID_GFX_SUPPORTED) {
+		printk("ATPX Hybrid Graphics\n");
+#if 1
+		/* This is a temporary hack until the D3 cold support
+		 * makes it upstream.  The ATPX power_control method seems
+		 * to still work on even if the system should be using
+		 * the new standardized hybrid D3 cold ACPI interface.
+		 */
+		atpx->functions.power_cntl = true;
+#else
+		atpx->functions.power_cntl = false;
+#endif
+		atpx->is_hybrid = true;
+	}
+
 	return 0;
 }
 
@@ -258,6 +285,10 @@ static int radeon_atpx_set_discrete_state(struct radeon_atpx *atpx, u8 state)
 		if (!info)
 			return -EIO;
 		kfree(info);
+
+		/* 200ms delay is required after off */
+		if (state == 0)
+			msleep(200);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 81a63d7f5cd9..b79f3b002471 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -2064,7 +2064,6 @@ radeon_add_atom_connector(struct drm_device *dev,
 							   RADEON_OUTPUT_CSC_BYPASS);
 			/* no HPD on analog connectors */
 			radeon_connector->hpd.hpd = RADEON_HPD_NONE;
-			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 			connector->interlace_allowed = true;
 			connector->doublescan_allowed = true;
 			break;
@@ -2314,8 +2313,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 	}
 
 	if (radeon_connector->hpd.hpd == RADEON_HPD_NONE) {
-		if (i2c_bus->valid)
-			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+		if (i2c_bus->valid) {
+			connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+			                    DRM_CONNECTOR_POLL_DISCONNECT;
+		}
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
@@ -2391,7 +2392,6 @@ radeon_add_legacy_connector(struct drm_device *dev,
 					      1);
 		/* no HPD on analog connectors */
 		radeon_connector->hpd.hpd = RADEON_HPD_NONE;
-		connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 		connector->interlace_allowed = true;
 		connector->doublescan_allowed = true;
 		break;
@@ -2476,10 +2476,13 @@ radeon_add_legacy_connector(struct drm_device *dev,
 	}
 
 	if (radeon_connector->hpd.hpd == RADEON_HPD_NONE) {
-		if (i2c_bus->valid)
-			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+		if (i2c_bus->valid) {
+			connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+			                    DRM_CONNECTOR_POLL_DISCONNECT;
+		}
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
+
 	connector->display_info.subpixel_order = subpixel_order;
 	drm_connector_register(connector);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index e0bf77843d79..a00dd2f74527 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -631,6 +631,23 @@ void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc)
 /*
  * GPU helpers function.
  */
+
+/**
+ * radeon_device_is_virtual - check if we are running is a virtual environment
+ *
+ * Check if the asic has been passed through to a VM (all asics).
+ * Used at driver startup.
+ * Returns true if virtual or false if not.
+ */
+static bool radeon_device_is_virtual(void)
+{
+#ifdef CONFIG_X86
+	return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#else
+	return false;
+#endif
+}
+
 /**
  * radeon_card_posted - check if the hw has already been initialized
  *
@@ -644,6 +661,10 @@ bool radeon_card_posted(struct radeon_device *rdev)
 {
 	uint32_t reg;
 
+	/* for pass through, always force asic_init */
+	if (radeon_device_is_virtual())
+		return false;
+
 	/* required for EFI mode on macbook2,1 which uses an r5xx asic */
 	if (efi_enabled(EFI_BOOT) &&
 	    (rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE) &&
@@ -1635,7 +1656,7 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend,
 	radeon_agp_suspend(rdev);
 
 	pci_save_state(dev->pdev);
-	if (freeze && rdev->family >= CHIP_R600) {
+	if (freeze && rdev->family >= CHIP_CEDAR) {
 		rdev->asic->asic_reset(rdev, true);
 		pci_restore_state(dev->pdev);
 	} else if (suspend) {
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index a455dc7d4aa1..c01a7c6abb49 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -93,9 +93,10 @@
  *   2.43.0 - RADEON_INFO_GPU_RESET_COUNTER
  *   2.44.0 - SET_APPEND_CNT packet3 support
  *   2.45.0 - Allow setting shader registers using DMA/COPY packet3 on SI
+ *   2.46.0 - Add PFP_SYNC_ME support on evergreen
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	45
+#define KMS_DRIVER_MINOR	46
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -162,9 +163,13 @@ void radeon_debugfs_cleanup(struct drm_minor *minor);
 #if defined(CONFIG_VGA_SWITCHEROO)
 void radeon_register_atpx_handler(void);
 void radeon_unregister_atpx_handler(void);
+bool radeon_has_atpx_dgpu_power_cntl(void);
+bool radeon_is_atpx_hybrid(void);
 #else
 static inline void radeon_register_atpx_handler(void) {}
 static inline void radeon_unregister_atpx_handler(void) {}
+static inline bool radeon_has_atpx_dgpu_power_cntl(void) { return false; }
+static inline bool radeon_is_atpx_hybrid(void) { return false; }
 #endif
 
 int radeon_no_wb;
@@ -404,7 +409,10 @@ static int radeon_pmops_runtime_suspend(struct device *dev)
 	pci_save_state(pdev);
 	pci_disable_device(pdev);
 	pci_ignore_hotplug(pdev);
-	pci_set_power_state(pdev, PCI_D3cold);
+	if (radeon_is_atpx_hybrid())
+		pci_set_power_state(pdev, PCI_D3cold);
+	else if (!radeon_has_atpx_dgpu_power_cntl())
+		pci_set_power_state(pdev, PCI_D3hot);
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
 
 	return 0;
@@ -421,7 +429,9 @@ static int radeon_pmops_runtime_resume(struct device *dev)
 
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
-	pci_set_power_state(pdev, PCI_D0);
+	if (radeon_is_atpx_hybrid() ||
+	    !radeon_has_atpx_dgpu_power_cntl())
+		pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 	ret = pci_enable_device(pdev);
 	if (ret)
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 590b0377fbe2..ffdad81ef964 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -300,8 +300,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	if (IS_ERR(fence))
 		return PTR_ERR(fence);
 
-	r = ttm_bo_move_accel_cleanup(bo, &fence->base,
-				      evict, no_wait_gpu, new_mem);
+	r = ttm_bo_move_accel_cleanup(bo, &fence->base, evict, new_mem);
 	radeon_fence_unref(&fence);
 	return r;
 }
@@ -403,6 +402,10 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	int r;
 
+	r = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+	if (r)
+		return r;
+
 	/* Can't move a pinned BO */
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	if (WARN_ON_ONCE(rbo->pin_count > 0))
@@ -441,7 +444,8 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
 
 	if (r) {
 memcpy:
-		r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+		r = ttm_bo_move_memcpy(bo, evict, interruptible,
+				       no_wait_gpu, new_mem);
 		if (r) {
 			return r;
 		}
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index b30e719dd56d..2523ca96c6c7 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -50,6 +50,7 @@ MODULE_FIRMWARE("radeon/tahiti_ce.bin");
 MODULE_FIRMWARE("radeon/tahiti_mc.bin");
 MODULE_FIRMWARE("radeon/tahiti_rlc.bin");
 MODULE_FIRMWARE("radeon/tahiti_smc.bin");
+MODULE_FIRMWARE("radeon/tahiti_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
@@ -65,6 +66,7 @@ MODULE_FIRMWARE("radeon/pitcairn_ce.bin");
 MODULE_FIRMWARE("radeon/pitcairn_mc.bin");
 MODULE_FIRMWARE("radeon/pitcairn_rlc.bin");
 MODULE_FIRMWARE("radeon/pitcairn_smc.bin");
+MODULE_FIRMWARE("radeon/pitcairn_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
 MODULE_FIRMWARE("radeon/VERDE_me.bin");
@@ -80,6 +82,7 @@ MODULE_FIRMWARE("radeon/verde_ce.bin");
 MODULE_FIRMWARE("radeon/verde_mc.bin");
 MODULE_FIRMWARE("radeon/verde_rlc.bin");
 MODULE_FIRMWARE("radeon/verde_smc.bin");
+MODULE_FIRMWARE("radeon/verde_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
 MODULE_FIRMWARE("radeon/OLAND_me.bin");
@@ -95,6 +98,7 @@ MODULE_FIRMWARE("radeon/oland_ce.bin");
 MODULE_FIRMWARE("radeon/oland_mc.bin");
 MODULE_FIRMWARE("radeon/oland_rlc.bin");
 MODULE_FIRMWARE("radeon/oland_smc.bin");
+MODULE_FIRMWARE("radeon/oland_k_smc.bin");
 
 MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
 MODULE_FIRMWARE("radeon/HAINAN_me.bin");
@@ -110,6 +114,7 @@ MODULE_FIRMWARE("radeon/hainan_ce.bin");
 MODULE_FIRMWARE("radeon/hainan_mc.bin");
 MODULE_FIRMWARE("radeon/hainan_rlc.bin");
 MODULE_FIRMWARE("radeon/hainan_smc.bin");
+MODULE_FIRMWARE("radeon/hainan_k_smc.bin");
 
 static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh);
 static void si_pcie_gen3_enable(struct radeon_device *rdev);
@@ -1653,12 +1658,16 @@ static int si_init_microcode(struct radeon_device *rdev)
 	char fw_name[30];
 	int err;
 	int new_fw = 0;
+	bool new_smc = false;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_TAHITI:
 		chip_name = "TAHITI";
+		/* XXX: figure out which Tahitis need the new ucode */
+		if (0)
+			new_smc = true;
 		new_chip_name = "tahiti";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
@@ -1670,6 +1679,13 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_PITCAIRN:
 		chip_name = "PITCAIRN";
+		if ((rdev->pdev->revision == 0x81) ||
+		    (rdev->pdev->device == 0x6810) ||
+		    (rdev->pdev->device == 0x6811) ||
+		    (rdev->pdev->device == 0x6816) ||
+		    (rdev->pdev->device == 0x6817) ||
+		    (rdev->pdev->device == 0x6806))
+			new_smc = true;
 		new_chip_name = "pitcairn";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
@@ -1681,6 +1697,16 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_VERDE:
 		chip_name = "VERDE";
+		if ((rdev->pdev->revision == 0x81) ||
+		    (rdev->pdev->revision == 0x83) ||
+		    (rdev->pdev->revision == 0x87) ||
+		    (rdev->pdev->device == 0x6820) ||
+		    (rdev->pdev->device == 0x6821) ||
+		    (rdev->pdev->device == 0x6822) ||
+		    (rdev->pdev->device == 0x6823) ||
+		    (rdev->pdev->device == 0x682A) ||
+		    (rdev->pdev->device == 0x682B))
+			new_smc = true;
 		new_chip_name = "verde";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
@@ -1692,6 +1718,13 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_OLAND:
 		chip_name = "OLAND";
+		if ((rdev->pdev->revision == 0xC7) ||
+		    (rdev->pdev->revision == 0x80) ||
+		    (rdev->pdev->revision == 0x81) ||
+		    (rdev->pdev->revision == 0x83) ||
+		    (rdev->pdev->device == 0x6604) ||
+		    (rdev->pdev->device == 0x6605))
+			new_smc = true;
 		new_chip_name = "oland";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
@@ -1702,6 +1735,13 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_HAINAN:
 		chip_name = "HAINAN";
+		if ((rdev->pdev->revision == 0x81) ||
+		    (rdev->pdev->revision == 0x83) ||
+		    (rdev->pdev->revision == 0xC3) ||
+		    (rdev->pdev->device == 0x6664) ||
+		    (rdev->pdev->device == 0x6665) ||
+		    (rdev->pdev->device == 0x6667))
+			new_smc = true;
 		new_chip_name = "hainan";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
@@ -1847,7 +1887,10 @@ static int si_init_microcode(struct radeon_device *rdev)
 		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
+	if (new_smc)
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name);
+	else
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 	err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 	if (err) {
 		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
diff --git a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
index c120172add5c..e81e19a660ad 100644
--- a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
+++ b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
@@ -14,6 +14,7 @@
 
 #include <linux/component.h>
 #include <linux/mfd/syscon.h>
+#include <linux/of_device.h>
 #include <linux/of_graph.h>
 #include <linux/regmap.h>
 #include <linux/reset.h>
@@ -33,13 +34,28 @@
 #include "rockchip_drm_drv.h"
 #include "rockchip_drm_vop.h"
 
+#define RK3288_GRF_SOC_CON6		0x25c
+#define RK3288_EDP_LCDC_SEL		BIT(5)
+#define RK3399_GRF_SOC_CON20		0x6250
+#define RK3399_EDP_LCDC_SEL		BIT(5)
+
+#define HIWORD_UPDATE(val, mask)	(val | (mask) << 16)
+
 #define to_dp(nm)	container_of(nm, struct rockchip_dp_device, nm)
 
-/* dp grf register offset */
-#define GRF_SOC_CON6                            0x025c
-#define GRF_EDP_LCD_SEL_MASK                    BIT(5)
-#define GRF_EDP_SEL_VOP_LIT                     BIT(5)
-#define GRF_EDP_SEL_VOP_BIG                     0
+/**
+ * struct rockchip_dp_chip_data - splite the grf setting of kind of chips
+ * @lcdsel_grf_reg: grf register offset of lcdc select
+ * @lcdsel_big: reg value of selecting vop big for eDP
+ * @lcdsel_lit: reg value of selecting vop little for eDP
+ * @chip_type: specific chip type
+ */
+struct rockchip_dp_chip_data {
+	u32	lcdsel_grf_reg;
+	u32	lcdsel_big;
+	u32	lcdsel_lit;
+	u32	chip_type;
+};
 
 struct rockchip_dp_device {
 	struct drm_device        *drm_dev;
@@ -48,9 +64,12 @@ struct rockchip_dp_device {
 	struct drm_display_mode  mode;
 
 	struct clk               *pclk;
+	struct clk               *grfclk;
 	struct regmap            *grf;
 	struct reset_control     *rst;
 
+	const struct rockchip_dp_chip_data *data;
+
 	struct analogix_dp_plat_data plat_data;
 };
 
@@ -92,6 +111,23 @@ static int rockchip_dp_powerdown(struct analogix_dp_plat_data *plat_data)
 	return 0;
 }
 
+static int rockchip_dp_get_modes(struct analogix_dp_plat_data *plat_data,
+				 struct drm_connector *connector)
+{
+	struct drm_display_info *di = &connector->display_info;
+	/* VOP couldn't output YUV video format for eDP rightly */
+	u32 mask = DRM_COLOR_FORMAT_YCRCB444 | DRM_COLOR_FORMAT_YCRCB422;
+
+	if ((di->color_formats & mask)) {
+		DRM_DEBUG_KMS("Swapping display color format from YUV to RGB\n");
+		di->color_formats &= ~mask;
+		di->color_formats |= DRM_COLOR_FORMAT_RGB444;
+		di->bpc = 8;
+	}
+
+	return 0;
+}
+
 static bool
 rockchip_dp_drm_encoder_mode_fixup(struct drm_encoder *encoder,
 				   const struct drm_display_mode *mode,
@@ -119,17 +155,23 @@ static void rockchip_dp_drm_encoder_enable(struct drm_encoder *encoder)
 		return;
 
 	if (ret)
-		val = GRF_EDP_SEL_VOP_LIT | (GRF_EDP_LCD_SEL_MASK << 16);
+		val = dp->data->lcdsel_lit;
 	else
-		val = GRF_EDP_SEL_VOP_BIG | (GRF_EDP_LCD_SEL_MASK << 16);
+		val = dp->data->lcdsel_big;
 
 	dev_dbg(dp->dev, "vop %s output to dp\n", (ret) ? "LIT" : "BIG");
 
-	ret = regmap_write(dp->grf, GRF_SOC_CON6, val);
-	if (ret != 0) {
-		dev_err(dp->dev, "Could not write to GRF: %d\n", ret);
+	ret = clk_prepare_enable(dp->grfclk);
+	if (ret < 0) {
+		dev_err(dp->dev, "failed to enable grfclk %d\n", ret);
 		return;
 	}
+
+	ret = regmap_write(dp->grf, dp->data->lcdsel_grf_reg, val);
+	if (ret != 0)
+		dev_err(dp->dev, "Could not write to GRF: %d\n", ret);
+
+	clk_disable_unprepare(dp->grfclk);
 }
 
 static void rockchip_dp_drm_encoder_nop(struct drm_encoder *encoder)
@@ -143,22 +185,29 @@ rockchip_dp_drm_encoder_atomic_check(struct drm_encoder *encoder,
 				      struct drm_connector_state *conn_state)
 {
 	struct rockchip_crtc_state *s = to_rockchip_crtc_state(crtc_state);
+	struct rockchip_dp_device *dp = to_dp(encoder);
+	int ret;
 
 	/*
-	 * FIXME(Yakir): driver should configure the CRTC output video
-	 * mode with the display information which indicated the monitor
-	 * support colorimetry.
-	 *
-	 * But don't know why the CRTC driver seems could only output the
-	 * RGBaaa rightly. For example, if connect the "innolux,n116bge"
-	 * eDP screen, EDID would indicated that screen only accepted the
-	 * 6bpc mode. But if I configure CRTC to RGB666 output, then eDP
-	 * screen would show a blue picture (RGB888 show a green picture).
-	 * But if I configure CTRC to RGBaaa, and eDP driver still keep
-	 * RGB666 input video mode, then screen would works prefect.
+	 * The hardware IC designed that VOP must output the RGB10 video
+	 * format to eDP controller, and if eDP panel only support RGB8,
+	 * then eDP controller should cut down the video data, not via VOP
+	 * controller, that's why we need to hardcode the VOP output mode
+	 * to RGA10 here.
 	 */
+
 	s->output_mode = ROCKCHIP_OUT_MODE_AAAA;
 	s->output_type = DRM_MODE_CONNECTOR_eDP;
+	if (dp->data->chip_type == RK3399_EDP) {
+		/*
+		 * For RK3399, VOP Lit must code the out mode to RGB888,
+		 * VOP Big must code the out mode to RGB10.
+		 */
+		ret = drm_of_encoder_active_endpoint_id(dp->dev->of_node,
+							encoder);
+		if (ret > 0)
+			s->output_mode = ROCKCHIP_OUT_MODE_P888;
+	}
 
 	return 0;
 }
@@ -192,6 +241,16 @@ static int rockchip_dp_init(struct rockchip_dp_device *dp)
 		return PTR_ERR(dp->grf);
 	}
 
+	dp->grfclk = devm_clk_get(dev, "grf");
+	if (PTR_ERR(dp->grfclk) == -ENOENT) {
+		dp->grfclk = NULL;
+	} else if (PTR_ERR(dp->grfclk) == -EPROBE_DEFER) {
+		return -EPROBE_DEFER;
+	} else if (IS_ERR(dp->grfclk)) {
+		dev_err(dev, "failed to get grf clock\n");
+		return PTR_ERR(dp->grfclk);
+	}
+
 	dp->pclk = devm_clk_get(dev, "pclk");
 	if (IS_ERR(dp->pclk)) {
 		dev_err(dev, "failed to get pclk property\n");
@@ -246,6 +305,7 @@ static int rockchip_dp_bind(struct device *dev, struct device *master,
 			    void *data)
 {
 	struct rockchip_dp_device *dp = dev_get_drvdata(dev);
+	const struct rockchip_dp_chip_data *dp_data;
 	struct drm_device *drm_dev = data;
 	int ret;
 
@@ -256,10 +316,15 @@ static int rockchip_dp_bind(struct device *dev, struct device *master,
 	 */
 	dev_set_drvdata(dev, NULL);
 
+	dp_data = of_device_get_match_data(dev);
+	if (!dp_data)
+		return -ENODEV;
+
 	ret = rockchip_dp_init(dp);
 	if (ret < 0)
 		return ret;
 
+	dp->data = dp_data;
 	dp->drm_dev = drm_dev;
 
 	ret = rockchip_dp_drm_create_encoder(dp);
@@ -270,9 +335,10 @@ static int rockchip_dp_bind(struct device *dev, struct device *master,
 
 	dp->plat_data.encoder = &dp->encoder;
 
-	dp->plat_data.dev_type = RK3288_DP;
+	dp->plat_data.dev_type = dp->data->chip_type;
 	dp->plat_data.power_on = rockchip_dp_poweron;
 	dp->plat_data.power_off = rockchip_dp_powerdown;
+	dp->plat_data.get_modes = rockchip_dp_get_modes;
 
 	return analogix_dp_bind(dev, dp->drm_dev, &dp->plat_data);
 }
@@ -292,38 +358,33 @@ static int rockchip_dp_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *panel_node, *port, *endpoint;
+	struct drm_panel *panel = NULL;
 	struct rockchip_dp_device *dp;
-	struct drm_panel *panel;
 
 	port = of_graph_get_port_by_id(dev->of_node, 1);
-	if (!port) {
-		dev_err(dev, "can't find output port\n");
-		return -EINVAL;
-	}
-
-	endpoint = of_get_child_by_name(port, "endpoint");
-	of_node_put(port);
-	if (!endpoint) {
-		dev_err(dev, "no output endpoint found\n");
-		return -EINVAL;
-	}
-
-	panel_node = of_graph_get_remote_port_parent(endpoint);
-	of_node_put(endpoint);
-	if (!panel_node) {
-		dev_err(dev, "no output node found\n");
-		return -EINVAL;
-	}
-
-	panel = of_drm_find_panel(panel_node);
-	if (!panel) {
-		DRM_ERROR("failed to find panel\n");
+	if (port) {
+		endpoint = of_get_child_by_name(port, "endpoint");
+		of_node_put(port);
+		if (!endpoint) {
+			dev_err(dev, "no output endpoint found\n");
+			return -EINVAL;
+		}
+
+		panel_node = of_graph_get_remote_port_parent(endpoint);
+		of_node_put(endpoint);
+		if (!panel_node) {
+			dev_err(dev, "no output node found\n");
+			return -EINVAL;
+		}
+
+		panel = of_drm_find_panel(panel_node);
 		of_node_put(panel_node);
-		return -EPROBE_DEFER;
+		if (!panel) {
+			DRM_ERROR("failed to find panel\n");
+			return -EPROBE_DEFER;
+		}
 	}
 
-	of_node_put(panel_node);
-
 	dp = devm_kzalloc(dev, sizeof(*dp), GFP_KERNEL);
 	if (!dp)
 		return -ENOMEM;
@@ -356,8 +417,23 @@ static const struct dev_pm_ops rockchip_dp_pm_ops = {
 #endif
 };
 
+static const struct rockchip_dp_chip_data rk3399_edp = {
+	.lcdsel_grf_reg = RK3399_GRF_SOC_CON20,
+	.lcdsel_big = HIWORD_UPDATE(0, RK3399_EDP_LCDC_SEL),
+	.lcdsel_lit = HIWORD_UPDATE(RK3399_EDP_LCDC_SEL, RK3399_EDP_LCDC_SEL),
+	.chip_type = RK3399_EDP,
+};
+
+static const struct rockchip_dp_chip_data rk3288_dp = {
+	.lcdsel_grf_reg = RK3288_GRF_SOC_CON6,
+	.lcdsel_big = HIWORD_UPDATE(0, RK3288_EDP_LCDC_SEL),
+	.lcdsel_lit = HIWORD_UPDATE(RK3288_EDP_LCDC_SEL, RK3288_EDP_LCDC_SEL),
+	.chip_type = RK3288_DP,
+};
+
 static const struct of_device_id rockchip_dp_dt_ids[] = {
-	{.compatible = "rockchip,rk3288-dp",},
+	{.compatible = "rockchip,rk3288-dp", .data = &rk3288_dp },
+	{.compatible = "rockchip,rk3399-edp", .data = &rk3399_edp },
 	{}
 };
 MODULE_DEVICE_TABLE(of, rockchip_dp_dt_ids);
diff --git a/drivers/gpu/drm/sti/Kconfig b/drivers/gpu/drm/sti/Kconfig
index 5ad43a1bb260..494ab257f77c 100644
--- a/drivers/gpu/drm/sti/Kconfig
+++ b/drivers/gpu/drm/sti/Kconfig
@@ -7,5 +7,6 @@ config DRM_STI
 	select DRM_KMS_CMA_HELPER
 	select DRM_PANEL
 	select FW_LOADER
+	select SND_SOC_HDMI_CODEC if SND_SOC
 	help
 	  Choose this option to enable DRM on STM stiH41x chipset
diff --git a/drivers/gpu/drm/sti/sti_awg_utils.c b/drivers/gpu/drm/sti/sti_awg_utils.c
index a516eb869f6f..2da7d6866d5d 100644
--- a/drivers/gpu/drm/sti/sti_awg_utils.c
+++ b/drivers/gpu/drm/sti/sti_awg_utils.c
@@ -6,6 +6,8 @@
 
 #include "sti_awg_utils.h"
 
+#define AWG_DELAY (-5)
+
 #define AWG_OPCODE_OFFSET 10
 #define AWG_MAX_ARG       0x3ff
 
@@ -125,7 +127,7 @@ static int awg_generate_line_signal(
 		val = timing->blanking_level;
 		ret |= awg_generate_instr(RPLSET, val, 0, 0, fwparams);
 
-		val = timing->trailing_pixels - 1;
+		val = timing->trailing_pixels - 1 + AWG_DELAY;
 		ret |= awg_generate_instr(SKIP, val, 0, 0, fwparams);
 	}
 
diff --git a/drivers/gpu/drm/sti/sti_crtc.c b/drivers/gpu/drm/sti/sti_crtc.c
index 7fab3af7473b..c7d734dc3cf4 100644
--- a/drivers/gpu/drm/sti/sti_crtc.c
+++ b/drivers/gpu/drm/sti/sti_crtc.c
@@ -23,22 +23,11 @@
 static void sti_crtc_enable(struct drm_crtc *crtc)
 {
 	struct sti_mixer *mixer = to_sti_mixer(crtc);
-	struct device *dev = mixer->dev;
-	struct sti_compositor *compo = dev_get_drvdata(dev);
 
 	DRM_DEBUG_DRIVER("\n");
 
 	mixer->status = STI_MIXER_READY;
 
-	/* Prepare and enable the compo IP clock */
-	if (mixer->id == STI_MIXER_MAIN) {
-		if (clk_prepare_enable(compo->clk_compo_main))
-			DRM_INFO("Failed to prepare/enable compo_main clk\n");
-	} else {
-		if (clk_prepare_enable(compo->clk_compo_aux))
-			DRM_INFO("Failed to prepare/enable compo_aux clk\n");
-	}
-
 	drm_crtc_vblank_on(crtc);
 }
 
@@ -57,9 +46,8 @@ sti_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode)
 	struct sti_mixer *mixer = to_sti_mixer(crtc);
 	struct device *dev = mixer->dev;
 	struct sti_compositor *compo = dev_get_drvdata(dev);
-	struct clk *clk;
+	struct clk *compo_clk, *pix_clk;
 	int rate = mode->clock * 1000;
-	int res;
 
 	DRM_DEBUG_KMS("CRTC:%d (%s) mode:%d (%s)\n",
 		      crtc->base.id, sti_mixer_to_str(mixer),
@@ -74,32 +62,46 @@ sti_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode)
 		      mode->vsync_start, mode->vsync_end,
 		      mode->vtotal, mode->type, mode->flags);
 
-	/* Set rate and prepare/enable pixel clock */
-	if (mixer->id == STI_MIXER_MAIN)
-		clk = compo->clk_pix_main;
-	else
-		clk = compo->clk_pix_aux;
+	if (mixer->id == STI_MIXER_MAIN) {
+		compo_clk = compo->clk_compo_main;
+		pix_clk = compo->clk_pix_main;
+	} else {
+		compo_clk = compo->clk_compo_aux;
+		pix_clk = compo->clk_pix_aux;
+	}
+
+	/* Prepare and enable the compo IP clock */
+	if (clk_prepare_enable(compo_clk)) {
+		DRM_INFO("Failed to prepare/enable compositor clk\n");
+		goto compo_error;
+	}
 
-	res = clk_set_rate(clk, rate);
-	if (res < 0) {
+	/* Set rate and prepare/enable pixel clock */
+	if (clk_set_rate(pix_clk, rate) < 0) {
 		DRM_ERROR("Cannot set rate (%dHz) for pix clk\n", rate);
-		return -EINVAL;
+		goto pix_error;
 	}
-	if (clk_prepare_enable(clk)) {
+	if (clk_prepare_enable(pix_clk)) {
 		DRM_ERROR("Failed to prepare/enable pix clk\n");
-		return -EINVAL;
+		goto pix_error;
 	}
 
 	sti_vtg_set_config(mixer->id == STI_MIXER_MAIN ?
 			compo->vtg_main : compo->vtg_aux, &crtc->mode);
 
-	res = sti_mixer_active_video_area(mixer, &crtc->mode);
-	if (res) {
+	if (sti_mixer_active_video_area(mixer, &crtc->mode)) {
 		DRM_ERROR("Can't set active video area\n");
-		return -EINVAL;
+		goto mixer_error;
 	}
 
-	return res;
+	return 0;
+
+mixer_error:
+	clk_disable_unprepare(pix_clk);
+pix_error:
+	clk_disable_unprepare(compo_clk);
+compo_error:
+	return -EINVAL;
 }
 
 static void sti_crtc_disable(struct drm_crtc *crtc)
@@ -130,7 +132,6 @@ static void sti_crtc_disable(struct drm_crtc *crtc)
 static void
 sti_crtc_mode_set_nofb(struct drm_crtc *crtc)
 {
-	sti_crtc_enable(crtc);
 	sti_crtc_mode_set(crtc, &crtc->state->adjusted_mode);
 }
 
@@ -221,9 +222,7 @@ static void sti_crtc_atomic_flush(struct drm_crtc *crtc,
 static const struct drm_crtc_helper_funcs sti_crtc_helper_funcs = {
 	.enable = sti_crtc_enable,
 	.disable = sti_crtc_disabling,
-	.mode_set = drm_helper_crtc_mode_set,
 	.mode_set_nofb = sti_crtc_mode_set_nofb,
-	.mode_set_base = drm_helper_crtc_mode_set_base,
 	.atomic_begin = sti_crtc_atomic_begin,
 	.atomic_flush = sti_crtc_atomic_flush,
 };
diff --git a/drivers/gpu/drm/sti/sti_hdmi.c b/drivers/gpu/drm/sti/sti_hdmi.c
index 8d1402b245bf..fedc17f98d9b 100644
--- a/drivers/gpu/drm/sti/sti_hdmi.c
+++ b/drivers/gpu/drm/sti/sti_hdmi.c
@@ -18,6 +18,8 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_edid.h>
 
+#include <sound/hdmi-codec.h>
+
 #include "sti_hdmi.h"
 #include "sti_hdmi_tx3g4c28phy.h"
 #include "sti_hdmi_tx3g0c55phy.h"
@@ -35,6 +37,8 @@
 #define HDMI_DFLT_CHL0_DAT              0x0110
 #define HDMI_DFLT_CHL1_DAT              0x0114
 #define HDMI_DFLT_CHL2_DAT              0x0118
+#define HDMI_AUDIO_CFG                  0x0200
+#define HDMI_SPDIF_FIFO_STATUS          0x0204
 #define HDMI_SW_DI_1_HEAD_WORD          0x0210
 #define HDMI_SW_DI_1_PKT_WORD0          0x0214
 #define HDMI_SW_DI_1_PKT_WORD1          0x0218
@@ -44,6 +48,9 @@
 #define HDMI_SW_DI_1_PKT_WORD5          0x0228
 #define HDMI_SW_DI_1_PKT_WORD6          0x022C
 #define HDMI_SW_DI_CFG                  0x0230
+#define HDMI_SAMPLE_FLAT_MASK           0x0244
+#define HDMI_AUDN                       0x0400
+#define HDMI_AUD_CTS                    0x0404
 #define HDMI_SW_DI_2_HEAD_WORD          0x0600
 #define HDMI_SW_DI_2_PKT_WORD0          0x0604
 #define HDMI_SW_DI_2_PKT_WORD1          0x0608
@@ -103,6 +110,7 @@
 #define HDMI_INT_DLL_LCK                BIT(5)
 #define HDMI_INT_NEW_FRAME              BIT(6)
 #define HDMI_INT_GENCTRL_PKT            BIT(7)
+#define HDMI_INT_AUDIO_FIFO_XRUN        BIT(8)
 #define HDMI_INT_SINK_TERM_PRESENT      BIT(11)
 
 #define HDMI_DEFAULT_INT (HDMI_INT_SINK_TERM_PRESENT \
@@ -111,6 +119,7 @@
 			| HDMI_INT_GLOBAL)
 
 #define HDMI_WORKING_INT (HDMI_INT_SINK_TERM_PRESENT \
+			| HDMI_INT_AUDIO_FIFO_XRUN \
 			| HDMI_INT_GENCTRL_PKT \
 			| HDMI_INT_NEW_FRAME \
 			| HDMI_INT_DLL_LCK \
@@ -121,6 +130,27 @@
 
 #define HDMI_STA_SW_RST                 BIT(1)
 
+#define HDMI_AUD_CFG_8CH		BIT(0)
+#define HDMI_AUD_CFG_SPDIF_DIV_2	BIT(1)
+#define HDMI_AUD_CFG_SPDIF_DIV_3	BIT(2)
+#define HDMI_AUD_CFG_SPDIF_CLK_DIV_4	(BIT(1) | BIT(2))
+#define HDMI_AUD_CFG_CTS_CLK_256FS	BIT(12)
+#define HDMI_AUD_CFG_DTS_INVALID	BIT(16)
+#define HDMI_AUD_CFG_ONE_BIT_INVALID	(BIT(18) | BIT(19) | BIT(20) |  BIT(21))
+#define HDMI_AUD_CFG_CH12_VALID	BIT(28)
+#define HDMI_AUD_CFG_CH34_VALID	BIT(29)
+#define HDMI_AUD_CFG_CH56_VALID	BIT(30)
+#define HDMI_AUD_CFG_CH78_VALID	BIT(31)
+
+/* sample flat mask */
+#define HDMI_SAMPLE_FLAT_NO	 0
+#define HDMI_SAMPLE_FLAT_SP0 BIT(0)
+#define HDMI_SAMPLE_FLAT_SP1 BIT(1)
+#define HDMI_SAMPLE_FLAT_SP2 BIT(2)
+#define HDMI_SAMPLE_FLAT_SP3 BIT(3)
+#define HDMI_SAMPLE_FLAT_ALL (HDMI_SAMPLE_FLAT_SP0 | HDMI_SAMPLE_FLAT_SP1 |\
+			      HDMI_SAMPLE_FLAT_SP2 | HDMI_SAMPLE_FLAT_SP3)
+
 #define HDMI_INFOFRAME_HEADER_TYPE(x)    (((x) & 0xff) <<  0)
 #define HDMI_INFOFRAME_HEADER_VERSION(x) (((x) & 0xff) <<  8)
 #define HDMI_INFOFRAME_HEADER_LEN(x)     (((x) & 0x0f) << 16)
@@ -171,6 +201,10 @@ static irqreturn_t hdmi_irq_thread(int irq, void *arg)
 		wake_up_interruptible(&hdmi->wait_event);
 	}
 
+	/* Audio FIFO underrun IRQ */
+	if (hdmi->irq_status & HDMI_INT_AUDIO_FIFO_XRUN)
+		DRM_INFO("Warning: audio FIFO underrun occurs!");
+
 	return IRQ_HANDLED;
 }
 
@@ -441,26 +475,29 @@ static int hdmi_avi_infoframe_config(struct sti_hdmi *hdmi)
  */
 static int hdmi_audio_infoframe_config(struct sti_hdmi *hdmi)
 {
-	struct hdmi_audio_infoframe infofame;
+	struct hdmi_audio_params *audio = &hdmi->audio;
 	u8 buffer[HDMI_INFOFRAME_SIZE(AUDIO)];
-	int ret;
-
-	ret = hdmi_audio_infoframe_init(&infofame);
-	if (ret < 0) {
-		DRM_ERROR("failed to setup audio infoframe: %d\n", ret);
-		return ret;
-	}
-
-	infofame.channels = 2;
-
-	ret = hdmi_audio_infoframe_pack(&infofame, buffer, sizeof(buffer));
-	if (ret < 0) {
-		DRM_ERROR("failed to pack audio infoframe: %d\n", ret);
-		return ret;
+	int ret, val;
+
+	DRM_DEBUG_DRIVER("enter %s, AIF %s\n", __func__,
+			 audio->enabled ? "enable" : "disable");
+	if (audio->enabled) {
+		/* set audio parameters stored*/
+		ret = hdmi_audio_infoframe_pack(&audio->cea, buffer,
+						sizeof(buffer));
+		if (ret < 0) {
+			DRM_ERROR("failed to pack audio infoframe: %d\n", ret);
+			return ret;
+		}
+		hdmi_infoframe_write_infopack(hdmi, buffer, ret);
+	} else {
+		/*disable audio info frame transmission */
+		val = hdmi_read(hdmi, HDMI_SW_DI_CFG);
+		val &= ~HDMI_IFRAME_CFG_DI_N(HDMI_IFRAME_MASK,
+					     HDMI_IFRAME_SLOT_AUDIO);
+		hdmi_write(hdmi, val, HDMI_SW_DI_CFG);
 	}
 
-	hdmi_infoframe_write_infopack(hdmi, buffer, ret);
-
 	return 0;
 }
 
@@ -650,6 +687,10 @@ static int hdmi_dbg_show(struct seq_file *s, void *data)
 	DBGFS_DUMP("", HDMI_SW_DI_CFG);
 	hdmi_dbg_sw_di_cfg(s, hdmi_read(hdmi, HDMI_SW_DI_CFG));
 
+	DBGFS_DUMP("\n", HDMI_AUDIO_CFG);
+	DBGFS_DUMP("\n", HDMI_SPDIF_FIFO_STATUS);
+	DBGFS_DUMP("\n", HDMI_AUDN);
+
 	seq_printf(s, "\n AVI Infoframe (Data Island slot N=%d):",
 		   HDMI_IFRAME_SLOT_AVI);
 	DBGFS_DUMP_DI(HDMI_SW_DI_N_HEAD_WORD, HDMI_IFRAME_SLOT_AVI);
@@ -854,6 +895,7 @@ static int sti_hdmi_connector_get_modes(struct drm_connector *connector)
 
 	count = drm_add_edid_modes(connector, edid);
 	drm_mode_connector_update_edid_property(connector, edid);
+	drm_edid_to_eld(connector, edid);
 
 	kfree(edid);
 	return count;
@@ -1036,6 +1078,207 @@ static struct drm_encoder *sti_hdmi_find_encoder(struct drm_device *dev)
 	return NULL;
 }
 
+/**
+ * sti_hdmi_audio_get_non_coherent_n() - get N parameter for non-coherent
+ * clocks. None-coherent clocks means that audio and TMDS clocks have not the
+ * same source (drifts between clocks). In this case assumption is that CTS is
+ * automatically calculated by hardware.
+ *
+ * @audio_fs: audio frame clock frequency in Hz
+ *
+ * Values computed are based on table described in HDMI specification 1.4b
+ *
+ * Returns n value.
+ */
+static int sti_hdmi_audio_get_non_coherent_n(unsigned int audio_fs)
+{
+	unsigned int n;
+
+	switch (audio_fs) {
+	case 32000:
+		n = 4096;
+		break;
+	case 44100:
+		n = 6272;
+		break;
+	case 48000:
+		n = 6144;
+		break;
+	case 88200:
+		n = 6272 * 2;
+		break;
+	case 96000:
+		n = 6144 * 2;
+		break;
+	case 176400:
+		n = 6272 * 4;
+		break;
+	case 192000:
+		n = 6144 * 4;
+		break;
+	default:
+		/* Not pre-defined, recommended value: 128 * fs / 1000 */
+		n = (audio_fs * 128) / 1000;
+	}
+
+	return n;
+}
+
+static int hdmi_audio_configure(struct sti_hdmi *hdmi,
+				struct hdmi_audio_params *params)
+{
+	int audio_cfg, n;
+	struct hdmi_audio_infoframe *info = &params->cea;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (!hdmi->enabled)
+		return 0;
+
+	/* update N parameter */
+	n = sti_hdmi_audio_get_non_coherent_n(params->sample_rate);
+
+	DRM_DEBUG_DRIVER("Audio rate = %d Hz, TMDS clock = %d Hz, n = %d\n",
+			 params->sample_rate, hdmi->mode.clock * 1000, n);
+	hdmi_write(hdmi, n, HDMI_AUDN);
+
+	/* update HDMI registers according to configuration */
+	audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID |
+		    HDMI_AUD_CFG_ONE_BIT_INVALID;
+
+	switch (info->channels) {
+	case 8:
+		audio_cfg |= HDMI_AUD_CFG_CH78_VALID;
+	case 6:
+		audio_cfg |= HDMI_AUD_CFG_CH56_VALID;
+	case 4:
+		audio_cfg |= HDMI_AUD_CFG_CH34_VALID | HDMI_AUD_CFG_8CH;
+	case 2:
+		audio_cfg |= HDMI_AUD_CFG_CH12_VALID;
+		break;
+	default:
+		DRM_ERROR("ERROR: Unsupported number of channels (%d)!\n",
+			  info->channels);
+		return -EINVAL;
+	}
+
+	hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG);
+
+	hdmi->audio = *params;
+
+	return hdmi_audio_infoframe_config(hdmi);
+}
+
+static void hdmi_audio_shutdown(struct device *dev, void *data)
+{
+	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
+	int audio_cfg;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* disable audio */
+	audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID |
+		    HDMI_AUD_CFG_ONE_BIT_INVALID;
+	hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG);
+
+	hdmi->audio.enabled = 0;
+	hdmi_audio_infoframe_config(hdmi);
+}
+
+static int hdmi_audio_hw_params(struct device *dev,
+				void *data,
+				struct hdmi_codec_daifmt *daifmt,
+				struct hdmi_codec_params *params)
+{
+	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
+	int ret;
+	struct hdmi_audio_params audio = {
+		.sample_width = params->sample_width,
+		.sample_rate = params->sample_rate,
+		.cea = params->cea,
+	};
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (!hdmi->enabled)
+		return 0;
+
+	if ((daifmt->fmt != HDMI_I2S) || daifmt->bit_clk_inv ||
+	    daifmt->frame_clk_inv || daifmt->bit_clk_master ||
+	    daifmt->frame_clk_master) {
+		dev_err(dev, "%s: Bad flags %d %d %d %d\n", __func__,
+			daifmt->bit_clk_inv, daifmt->frame_clk_inv,
+			daifmt->bit_clk_master,
+			daifmt->frame_clk_master);
+		return -EINVAL;
+	}
+
+	audio.enabled = 1;
+
+	ret = hdmi_audio_configure(hdmi, &audio);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static int hdmi_audio_digital_mute(struct device *dev, void *data, bool enable)
+{
+	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
+
+	DRM_DEBUG_DRIVER("%s\n", enable ? "enable" : "disable");
+
+	if (enable)
+		hdmi_write(hdmi, HDMI_SAMPLE_FLAT_ALL, HDMI_SAMPLE_FLAT_MASK);
+	else
+		hdmi_write(hdmi, HDMI_SAMPLE_FLAT_NO, HDMI_SAMPLE_FLAT_MASK);
+
+	return 0;
+}
+
+static int hdmi_audio_get_eld(struct device *dev, void *data, uint8_t *buf, size_t len)
+{
+	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
+	struct drm_connector *connector = hdmi->drm_connector;
+
+	DRM_DEBUG_DRIVER("\n");
+	memcpy(buf, connector->eld, min(sizeof(connector->eld), len));
+
+	return 0;
+}
+
+static const struct hdmi_codec_ops audio_codec_ops = {
+	.hw_params = hdmi_audio_hw_params,
+	.audio_shutdown = hdmi_audio_shutdown,
+	.digital_mute = hdmi_audio_digital_mute,
+	.get_eld = hdmi_audio_get_eld,
+};
+
+static int sti_hdmi_register_audio_driver(struct device *dev,
+					  struct sti_hdmi *hdmi)
+{
+	struct hdmi_codec_pdata codec_data = {
+		.ops = &audio_codec_ops,
+		.max_i2s_channels = 8,
+		.i2s = 1,
+	};
+
+	DRM_DEBUG_DRIVER("\n");
+
+	hdmi->audio.enabled = 0;
+
+	hdmi->audio_pdev = platform_device_register_data(
+		dev, HDMI_CODEC_DRV_NAME, PLATFORM_DEVID_AUTO,
+		&codec_data, sizeof(codec_data));
+
+	if (IS_ERR(hdmi->audio_pdev))
+		return PTR_ERR(hdmi->audio_pdev);
+
+	DRM_INFO("%s Driver bound %s\n", HDMI_CODEC_DRV_NAME, dev_name(dev));
+
+	return 0;
+}
+
 static int sti_hdmi_bind(struct device *dev, struct device *master, void *data)
 {
 	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
@@ -1082,12 +1325,27 @@ static int sti_hdmi_bind(struct device *dev, struct device *master, void *data)
 	/* initialise property */
 	sti_hdmi_connector_init_property(drm_dev, drm_connector);
 
+	hdmi->drm_connector = drm_connector;
+
 	err = drm_mode_connector_attach_encoder(drm_connector, encoder);
 	if (err) {
 		DRM_ERROR("Failed to attach a connector to a encoder\n");
 		goto err_sysfs;
 	}
 
+	err = sti_hdmi_register_audio_driver(dev, hdmi);
+	if (err) {
+		DRM_ERROR("Failed to attach an audio codec\n");
+		goto err_sysfs;
+	}
+
+	/* Initialize audio infoframe */
+	err = hdmi_audio_infoframe_init(&hdmi->audio.cea);
+	if (err) {
+		DRM_ERROR("Failed to init audio infoframe\n");
+		goto err_sysfs;
+	}
+
 	/* Enable default interrupts */
 	hdmi_write(hdmi, HDMI_DEFAULT_INT, HDMI_INT_EN);
 
@@ -1095,6 +1353,7 @@ static int sti_hdmi_bind(struct device *dev, struct device *master, void *data)
 
 err_sysfs:
 	drm_bridge_remove(bridge);
+	hdmi->drm_connector = NULL;
 	return -EINVAL;
 }
 
@@ -1244,6 +1503,8 @@ static int sti_hdmi_remove(struct platform_device *pdev)
 	struct sti_hdmi *hdmi = dev_get_drvdata(&pdev->dev);
 
 	i2c_put_adapter(hdmi->ddc_adapt);
+	if (hdmi->audio_pdev)
+		platform_device_unregister(hdmi->audio_pdev);
 	component_del(&pdev->dev, &sti_hdmi_ops);
 
 	return 0;
diff --git a/drivers/gpu/drm/sti/sti_hdmi.h b/drivers/gpu/drm/sti/sti_hdmi.h
index ef3a94583bbd..119bc3582ac7 100644
--- a/drivers/gpu/drm/sti/sti_hdmi.h
+++ b/drivers/gpu/drm/sti/sti_hdmi.h
@@ -23,6 +23,13 @@ struct hdmi_phy_ops {
 	void (*stop)(struct sti_hdmi *hdmi);
 };
 
+struct hdmi_audio_params {
+	bool enabled;
+	unsigned int sample_width;
+	unsigned int sample_rate;
+	struct hdmi_audio_infoframe cea;
+};
+
 /* values for the framing mode property */
 enum sti_hdmi_modes {
 	HDMI_MODE_HDMI,
@@ -67,6 +74,9 @@ static const struct drm_prop_enum_list colorspace_mode_names[] = {
  * @ddc_adapt: i2c ddc adapter
  * @colorspace: current colorspace selected
  * @hdmi_mode: select framing for HDMI or DVI
+ * @audio_pdev: ASoC hdmi-codec platform device
+ * @audio: hdmi audio parameters.
+ * @drm_connector: hdmi connector
  */
 struct sti_hdmi {
 	struct device dev;
@@ -89,6 +99,9 @@ struct sti_hdmi {
 	struct i2c_adapter *ddc_adapt;
 	enum hdmi_colorspace colorspace;
 	enum sti_hdmi_modes hdmi_mode;
+	struct platform_device *audio_pdev;
+	struct hdmi_audio_params audio;
+	struct drm_connector *drm_connector;
 };
 
 u32 hdmi_read(struct sti_hdmi *hdmi, int offset);
diff --git a/drivers/gpu/drm/sti/sti_plane.c b/drivers/gpu/drm/sti/sti_plane.c
index 85cee9098439..0cf3335ef37c 100644
--- a/drivers/gpu/drm/sti/sti_plane.c
+++ b/drivers/gpu/drm/sti/sti_plane.c
@@ -45,25 +45,15 @@ const char *sti_plane_to_str(struct sti_plane *plane)
 
 #define STI_FPS_INTERVAL_MS     3000
 
-static int sti_plane_timespec_ms_diff(struct timespec lhs, struct timespec rhs)
-{
-	struct timespec tmp_ts = timespec_sub(lhs, rhs);
-	u64 tmp_ns = (u64)timespec_to_ns(&tmp_ts);
-
-	do_div(tmp_ns, NSEC_PER_MSEC);
-
-	return (u32)tmp_ns;
-}
-
 void sti_plane_update_fps(struct sti_plane *plane,
 			  bool new_frame,
 			  bool new_field)
 {
-	struct timespec now;
+	ktime_t now;
 	struct sti_fps_info *fps;
 	int fpks, fipks, ms_since_last, num_frames, num_fields;
 
-	getrawmonotonic(&now);
+	now = ktime_get();
 
 	/* Compute number of frame updates */
 	fps = &plane->fps_info;
@@ -76,7 +66,7 @@ void sti_plane_update_fps(struct sti_plane *plane,
 		return;
 
 	fps->curr_frame_counter++;
-	ms_since_last = sti_plane_timespec_ms_diff(now, fps->last_timestamp);
+	ms_since_last = ktime_to_ms(ktime_sub(now, fps->last_timestamp));
 	num_frames = fps->curr_frame_counter - fps->last_frame_counter;
 
 	if (num_frames <= 0  || ms_since_last < STI_FPS_INTERVAL_MS)
diff --git a/drivers/gpu/drm/sti/sti_plane.h b/drivers/gpu/drm/sti/sti_plane.h
index 39d39f5b7dd9..e0ea1dd3bb88 100644
--- a/drivers/gpu/drm/sti/sti_plane.h
+++ b/drivers/gpu/drm/sti/sti_plane.h
@@ -55,7 +55,7 @@ struct sti_fps_info {
 	unsigned int last_frame_counter;
 	unsigned int curr_field_counter;
 	unsigned int last_field_counter;
-	struct timespec last_timestamp;
+	ktime_t	     last_timestamp;
 	char fps_str[FPS_LENGTH];
 	char fips_str[FPS_LENGTH];
 };
diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c
index 0100c7c668b7..0bdc385eec17 100644
--- a/drivers/gpu/drm/sti/sti_vtg.c
+++ b/drivers/gpu/drm/sti/sti_vtg.c
@@ -65,7 +65,7 @@
 #define HDMI_DELAY          (5)
 
 /* Delay introduced by the DVO in nb of pixel */
-#define DVO_DELAY           (2)
+#define DVO_DELAY           (7)
 
 /* delay introduced by the Arbitrary Waveform Generator in nb of pixels */
 #define AWG_DELAY_HD        (-9)
diff --git a/drivers/gpu/drm/sun4i/Kconfig b/drivers/gpu/drm/sun4i/Kconfig
index 99510e64e91a..a4b357db8856 100644
--- a/drivers/gpu/drm/sun4i/Kconfig
+++ b/drivers/gpu/drm/sun4i/Kconfig
@@ -1,6 +1,6 @@
 config DRM_SUN4I
 	tristate "DRM Support for Allwinner A10 Display Engine"
-	depends on DRM && ARM
+	depends on DRM && ARM && COMMON_CLK
 	depends on ARCH_SUNXI || COMPILE_TEST
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_HELPER
diff --git a/drivers/gpu/drm/sun4i/sun4i_backend.c b/drivers/gpu/drm/sun4i/sun4i_backend.c
index f7a15c1a93bf..3ab560450a82 100644
--- a/drivers/gpu/drm/sun4i/sun4i_backend.c
+++ b/drivers/gpu/drm/sun4i/sun4i_backend.c
@@ -190,7 +190,7 @@ int sun4i_backend_update_layer_buffer(struct sun4i_backend *backend,
 	/* Get the physical address of the buffer in memory */
 	gem = drm_fb_cma_get_gem_obj(fb, 0);
 
-	DRM_DEBUG_DRIVER("Using GEM @ 0x%x\n", gem->paddr);
+	DRM_DEBUG_DRIVER("Using GEM @ %pad\n", &gem->paddr);
 
 	/* Compute the start of the displayed memory */
 	bpp = drm_format_plane_cpp(fb->pixel_format, 0);
@@ -198,7 +198,7 @@ int sun4i_backend_update_layer_buffer(struct sun4i_backend *backend,
 	paddr += (state->src_x >> 16) * bpp;
 	paddr += (state->src_y >> 16) * fb->pitches[0];
 
-	DRM_DEBUG_DRIVER("Setting buffer address to 0x%x\n", paddr);
+	DRM_DEBUG_DRIVER("Setting buffer address to %pad\n", &paddr);
 
 	/* Write the 32 lower bits of the address (in bits) */
 	lo_paddr = paddr << 3;
diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c
index 3ff668cb463c..5b3463197c48 100644
--- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c
+++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c
@@ -72,14 +72,40 @@ static unsigned long sun4i_dclk_recalc_rate(struct clk_hw *hw,
 static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate,
 				  unsigned long *parent_rate)
 {
-	return *parent_rate / DIV_ROUND_CLOSEST(*parent_rate, rate);
+	unsigned long best_parent = 0;
+	u8 best_div = 1;
+	int i;
+
+	for (i = 6; i < 127; i++) {
+		unsigned long ideal = rate * i;
+		unsigned long rounded;
+
+		rounded = clk_hw_round_rate(clk_hw_get_parent(hw),
+					    ideal);
+
+		if (rounded == ideal) {
+			best_parent = rounded;
+			best_div = i;
+			goto out;
+		}
+
+		if ((rounded < ideal) && (rounded > best_parent)) {
+			best_parent = rounded;
+			best_div = i;
+		}
+	}
+
+out:
+	*parent_rate = best_parent;
+
+	return best_parent / best_div;
 }
 
 static int sun4i_dclk_set_rate(struct clk_hw *hw, unsigned long rate,
 			       unsigned long parent_rate)
 {
 	struct sun4i_dclk *dclk = hw_to_dclk(hw);
-	int div = DIV_ROUND_CLOSEST(parent_rate, rate);
+	u8 div = parent_rate / rate;
 
 	return regmap_update_bits(dclk->regmap, SUN4I_TCON0_DCLK_REG,
 				  GENMASK(6, 0), div);
@@ -127,10 +153,14 @@ int sun4i_dclk_create(struct device *dev, struct sun4i_tcon *tcon)
 	const char *clk_name, *parent_name;
 	struct clk_init_data init;
 	struct sun4i_dclk *dclk;
+	int ret;
 
 	parent_name = __clk_get_name(tcon->sclk0);
-	of_property_read_string_index(dev->of_node, "clock-output-names", 0,
-				      &clk_name);
+	ret = of_property_read_string_index(dev->of_node,
+					    "clock-output-names", 0,
+					    &clk_name);
+	if (ret)
+		return ret;
 
 	dclk = devm_kzalloc(dev, sizeof(*dclk), GFP_KERNEL);
 	if (!dclk)
@@ -140,6 +170,7 @@ int sun4i_dclk_create(struct device *dev, struct sun4i_tcon *tcon)
 	init.ops = &sun4i_dclk_ops;
 	init.parent_names = &parent_name;
 	init.num_parents = 1;
+	init.flags = CLK_SET_RATE_PARENT;
 
 	dclk->regmap = tcon->regs;
 	dclk->hw.init = &init;
diff --git a/drivers/gpu/drm/sun4i/sun4i_drv.c b/drivers/gpu/drm/sun4i/sun4i_drv.c
index 9a67f927a53e..5b89940edcb1 100644
--- a/drivers/gpu/drm/sun4i/sun4i_drv.c
+++ b/drivers/gpu/drm/sun4i/sun4i_drv.c
@@ -97,6 +97,22 @@ static struct drm_driver sun4i_drv_driver = {
 	.disable_vblank		= sun4i_drv_disable_vblank,
 };
 
+static void sun4i_remove_framebuffers(void)
+{
+	struct apertures_struct *ap;
+
+	ap = alloc_apertures(1);
+	if (!ap)
+		return;
+
+	/* The framebuffer can be located anywhere in RAM */
+	ap->ranges[0].base = 0;
+	ap->ranges[0].size = ~0;
+
+	remove_conflicting_framebuffers(ap, "sun4i-drm-fb", false);
+	kfree(ap);
+}
+
 static int sun4i_drv_bind(struct device *dev)
 {
 	struct drm_device *drm;
@@ -140,6 +156,9 @@ static int sun4i_drv_bind(struct device *dev)
 	}
 	drm->irq_enabled = true;
 
+	/* Remove early framebuffers (ie. simplefb) */
+	sun4i_remove_framebuffers();
+
 	/* Create our framebuffer */
 	drv->fbdev = sun4i_framebuffer_init(drm);
 	if (IS_ERR(drv->fbdev)) {
@@ -166,6 +185,7 @@ static void sun4i_drv_unbind(struct device *dev)
 {
 	struct drm_device *drm = dev_get_drvdata(dev);
 
+	drm_connector_unregister_all(drm);
 	drm_dev_unregister(drm);
 	drm_kms_helper_poll_fini(drm);
 	sun4i_framebuffer_free(drm);
diff --git a/drivers/gpu/drm/sun4i/sun4i_rgb.c b/drivers/gpu/drm/sun4i/sun4i_rgb.c
index 442cfe271688..f5bbac6efb4c 100644
--- a/drivers/gpu/drm/sun4i/sun4i_rgb.c
+++ b/drivers/gpu/drm/sun4i/sun4i_rgb.c
@@ -54,8 +54,13 @@ static int sun4i_rgb_get_modes(struct drm_connector *connector)
 static int sun4i_rgb_mode_valid(struct drm_connector *connector,
 				struct drm_display_mode *mode)
 {
+	struct sun4i_rgb *rgb = drm_connector_to_sun4i_rgb(connector);
+	struct sun4i_drv *drv = rgb->drv;
+	struct sun4i_tcon *tcon = drv->tcon;
 	u32 hsync = mode->hsync_end - mode->hsync_start;
 	u32 vsync = mode->vsync_end - mode->vsync_start;
+	unsigned long rate = mode->clock * 1000;
+	long rounded_rate;
 
 	DRM_DEBUG_DRIVER("Validating modes...\n");
 
@@ -87,6 +92,15 @@ static int sun4i_rgb_mode_valid(struct drm_connector *connector,
 
 	DRM_DEBUG_DRIVER("Vertical parameters OK\n");
 
+	rounded_rate = clk_round_rate(tcon->dclk, rate);
+	if (rounded_rate < rate)
+		return MODE_CLOCK_LOW;
+
+	if (rounded_rate > rate)
+		return MODE_CLOCK_HIGH;
+
+	DRM_DEBUG_DRIVER("Clock rate OK\n");
+
 	return MODE_OK;
 }
 
@@ -193,7 +207,7 @@ int sun4i_rgb_init(struct drm_device *drm)
 	int ret;
 
 	/* If we don't have a panel, there's no point in going on */
-	if (!tcon->panel)
+	if (IS_ERR(tcon->panel))
 		return -ENODEV;
 
 	rgb = devm_kzalloc(drm->dev, sizeof(*rgb), GFP_KERNEL);
diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.c b/drivers/gpu/drm/sun4i/sun4i_tcon.c
index 9f19b0e08560..652385f09735 100644
--- a/drivers/gpu/drm/sun4i/sun4i_tcon.c
+++ b/drivers/gpu/drm/sun4i/sun4i_tcon.c
@@ -425,11 +425,11 @@ static struct drm_panel *sun4i_tcon_find_panel(struct device_node *node)
 
 	remote = of_graph_get_remote_port_parent(end_node);
 	if (!remote) {
-		DRM_DEBUG_DRIVER("Enable to parse remote node\n");
+		DRM_DEBUG_DRIVER("Unable to parse remote node\n");
 		return ERR_PTR(-EINVAL);
 	}
 
-	return of_drm_find_panel(remote);
+	return of_drm_find_panel(remote) ?: ERR_PTR(-EPROBE_DEFER);
 }
 
 static int sun4i_tcon_bind(struct device *dev, struct device *master,
@@ -490,7 +490,11 @@ static int sun4i_tcon_bind(struct device *dev, struct device *master,
 		return 0;
 	}
 
-	return sun4i_rgb_init(drm);
+	ret = sun4i_rgb_init(drm);
+	if (ret < 0)
+		goto err_free_clocks;
+
+	return 0;
 
 err_free_clocks:
 	sun4i_tcon_free_clocks(tcon);
@@ -522,12 +526,13 @@ static int sun4i_tcon_probe(struct platform_device *pdev)
 	 * Defer the probe.
 	 */
 	panel = sun4i_tcon_find_panel(node);
-	if (IS_ERR(panel)) {
-		/*
-		 * If we don't have a panel endpoint, just go on
-		 */
-		if (PTR_ERR(panel) != -ENODEV)
-			return -EPROBE_DEFER;
+
+	/*
+	 * If we don't have a panel endpoint, just go on
+	 */
+	if (PTR_ERR(panel) == -EPROBE_DEFER) {
+		DRM_DEBUG_DRIVER("Still waiting for our panel. Deferring...\n");
+		return -EPROBE_DEFER;
 	}
 
 	return component_add(&pdev->dev, &sun4i_tcon_ops);
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 39940f5b7c91..8495bd01b544 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -10,6 +10,7 @@
 #include <linux/clk.h>
 #include <linux/debugfs.h>
 #include <linux/iommu.h>
+#include <linux/pm_runtime.h>
 #include <linux/reset.h>
 
 #include <soc/tegra/pmc.h>
@@ -1216,6 +1217,8 @@ static void tegra_crtc_disable(struct drm_crtc *crtc)
 
 	tegra_dc_stats_reset(&dc->stats);
 	drm_crtc_vblank_off(crtc);
+
+	pm_runtime_put_sync(dc->dev);
 }
 
 static void tegra_crtc_enable(struct drm_crtc *crtc)
@@ -1225,6 +1228,48 @@ static void tegra_crtc_enable(struct drm_crtc *crtc)
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	u32 value;
 
+	pm_runtime_get_sync(dc->dev);
+
+	/* initialize display controller */
+	if (dc->syncpt) {
+		u32 syncpt = host1x_syncpt_id(dc->syncpt);
+
+		value = SYNCPT_CNTRL_NO_STALL;
+		tegra_dc_writel(dc, value, DC_CMD_GENERAL_INCR_SYNCPT_CNTRL);
+
+		value = SYNCPT_VSYNC_ENABLE | syncpt;
+		tegra_dc_writel(dc, value, DC_CMD_CONT_SYNCPT_VSYNC);
+	}
+
+	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_TYPE);
+
+	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY);
+
+	/* initialize timer */
+	value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) |
+		WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20);
+	tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY);
+
+	value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) |
+		WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1);
+	tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER);
+
+	value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
+
+	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+
+	if (dc->soc->supports_border_color)
+		tegra_dc_writel(dc, 0, DC_DISP_BORDER_COLOR);
+
+	/* apply PLL and pixel clock changes */
 	tegra_dc_commit_state(dc, state);
 
 	/* program display mode */
@@ -1685,7 +1730,6 @@ static int tegra_dc_init(struct host1x_client *client)
 	struct tegra_drm *tegra = drm->dev_private;
 	struct drm_plane *primary = NULL;
 	struct drm_plane *cursor = NULL;
-	u32 value;
 	int err;
 
 	dc->syncpt = host1x_syncpt_request(dc->dev, flags);
@@ -1755,47 +1799,6 @@ static int tegra_dc_init(struct host1x_client *client)
 		goto cleanup;
 	}
 
-	/* initialize display controller */
-	if (dc->syncpt) {
-		u32 syncpt = host1x_syncpt_id(dc->syncpt);
-
-		value = SYNCPT_CNTRL_NO_STALL;
-		tegra_dc_writel(dc, value, DC_CMD_GENERAL_INCR_SYNCPT_CNTRL);
-
-		value = SYNCPT_VSYNC_ENABLE | syncpt;
-		tegra_dc_writel(dc, value, DC_CMD_CONT_SYNCPT_VSYNC);
-	}
-
-	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
-		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_TYPE);
-
-	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
-		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY);
-
-	/* initialize timer */
-	value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) |
-		WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20);
-	tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY);
-
-	value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) |
-		WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1);
-	tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER);
-
-	value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
-		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
-
-	value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
-		WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
-	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
-
-	if (dc->soc->supports_border_color)
-		tegra_dc_writel(dc, 0, DC_DISP_BORDER_COLOR);
-
-	tegra_dc_stats_reset(&dc->stats);
-
 	return 0;
 
 cleanup:
@@ -1987,33 +1990,15 @@ static int tegra_dc_probe(struct platform_device *pdev)
 		return PTR_ERR(dc->rst);
 	}
 
+	reset_control_assert(dc->rst);
+
 	if (dc->soc->has_powergate) {
 		if (dc->pipe == 0)
 			dc->powergate = TEGRA_POWERGATE_DIS;
 		else
 			dc->powergate = TEGRA_POWERGATE_DISB;
 
-		err = tegra_powergate_sequence_power_up(dc->powergate, dc->clk,
-							dc->rst);
-		if (err < 0) {
-			dev_err(&pdev->dev, "failed to power partition: %d\n",
-				err);
-			return err;
-		}
-	} else {
-		err = clk_prepare_enable(dc->clk);
-		if (err < 0) {
-			dev_err(&pdev->dev, "failed to enable clock: %d\n",
-				err);
-			return err;
-		}
-
-		err = reset_control_deassert(dc->rst);
-		if (err < 0) {
-			dev_err(&pdev->dev, "failed to deassert reset: %d\n",
-				err);
-			return err;
-		}
+		tegra_powergate_power_off(dc->powergate);
 	}
 
 	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -2027,16 +2012,19 @@ static int tegra_dc_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	INIT_LIST_HEAD(&dc->client.list);
-	dc->client.ops = &dc_client_ops;
-	dc->client.dev = &pdev->dev;
-
 	err = tegra_dc_rgb_probe(dc);
 	if (err < 0 && err != -ENODEV) {
 		dev_err(&pdev->dev, "failed to probe RGB output: %d\n", err);
 		return err;
 	}
 
+	platform_set_drvdata(pdev, dc);
+	pm_runtime_enable(&pdev->dev);
+
+	INIT_LIST_HEAD(&dc->client.list);
+	dc->client.ops = &dc_client_ops;
+	dc->client.dev = &pdev->dev;
+
 	err = host1x_client_register(&dc->client);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
@@ -2044,8 +2032,6 @@ static int tegra_dc_probe(struct platform_device *pdev)
 		return err;
 	}
 
-	platform_set_drvdata(pdev, dc);
-
 	return 0;
 }
 
@@ -2067,7 +2053,22 @@ static int tegra_dc_remove(struct platform_device *pdev)
 		return err;
 	}
 
-	reset_control_assert(dc->rst);
+	pm_runtime_disable(&pdev->dev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int tegra_dc_suspend(struct device *dev)
+{
+	struct tegra_dc *dc = dev_get_drvdata(dev);
+	int err;
+
+	err = reset_control_assert(dc->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to assert reset: %d\n", err);
+		return err;
+	}
 
 	if (dc->soc->has_powergate)
 		tegra_powergate_power_off(dc->powergate);
@@ -2077,10 +2078,45 @@ static int tegra_dc_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static int tegra_dc_resume(struct device *dev)
+{
+	struct tegra_dc *dc = dev_get_drvdata(dev);
+	int err;
+
+	if (dc->soc->has_powergate) {
+		err = tegra_powergate_sequence_power_up(dc->powergate, dc->clk,
+							dc->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to power partition: %d\n", err);
+			return err;
+		}
+	} else {
+		err = clk_prepare_enable(dc->clk);
+		if (err < 0) {
+			dev_err(dev, "failed to enable clock: %d\n", err);
+			return err;
+		}
+
+		err = reset_control_deassert(dc->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to deassert reset: %d\n", err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops tegra_dc_pm_ops = {
+	SET_RUNTIME_PM_OPS(tegra_dc_suspend, tegra_dc_resume, NULL)
+};
+
 struct platform_driver tegra_dc_driver = {
 	.driver = {
 		.name = "tegra-dc",
 		.of_match_table = tegra_dc_of_match,
+		.pm = &tegra_dc_pm_ops,
 	},
 	.probe = tegra_dc_probe,
 	.remove = tegra_dc_remove,
diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c
index b24a0f14821a..059f409556d5 100644
--- a/drivers/gpu/drm/tegra/dpaux.c
+++ b/drivers/gpu/drm/tegra/dpaux.c
@@ -12,6 +12,9 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/of_gpio.h>
+#include <linux/pinctrl/pinconf-generic.h>
+#include <linux/pinctrl/pinctrl.h>
+#include <linux/pinctrl/pinmux.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
 #include <linux/regulator/consumer.h>
@@ -44,6 +47,11 @@ struct tegra_dpaux {
 	struct completion complete;
 	struct work_struct work;
 	struct list_head list;
+
+#ifdef CONFIG_GENERIC_PINCONF
+	struct pinctrl_dev *pinctrl;
+	struct pinctrl_desc desc;
+#endif
 };
 
 static inline struct tegra_dpaux *to_dpaux(struct drm_dp_aux *aux)
@@ -267,6 +275,148 @@ static irqreturn_t tegra_dpaux_irq(int irq, void *data)
 	return ret;
 }
 
+enum tegra_dpaux_functions {
+	DPAUX_PADCTL_FUNC_AUX,
+	DPAUX_PADCTL_FUNC_I2C,
+	DPAUX_PADCTL_FUNC_OFF,
+};
+
+static void tegra_dpaux_pad_power_down(struct tegra_dpaux *dpaux)
+{
+	u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
+
+	value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+}
+
+static void tegra_dpaux_pad_power_up(struct tegra_dpaux *dpaux)
+{
+	u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
+
+	value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+}
+
+static int tegra_dpaux_pad_config(struct tegra_dpaux *dpaux, unsigned function)
+{
+	u32 value;
+
+	switch (function) {
+	case DPAUX_PADCTL_FUNC_AUX:
+		value = DPAUX_HYBRID_PADCTL_AUX_CMH(2) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVZ(4) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVI(0x18) |
+			DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_MODE_AUX;
+		break;
+
+	case DPAUX_PADCTL_FUNC_I2C:
+		value = DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_MODE_I2C;
+		break;
+
+	case DPAUX_PADCTL_FUNC_OFF:
+		tegra_dpaux_pad_power_down(dpaux);
+		return 0;
+
+	default:
+		return -ENOTSUPP;
+	}
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL);
+	tegra_dpaux_pad_power_up(dpaux);
+
+	return 0;
+}
+
+#ifdef CONFIG_GENERIC_PINCONF
+static const struct pinctrl_pin_desc tegra_dpaux_pins[] = {
+	PINCTRL_PIN(0, "DP_AUX_CHx_P"),
+	PINCTRL_PIN(1, "DP_AUX_CHx_N"),
+};
+
+static const unsigned tegra_dpaux_pin_numbers[] = { 0, 1 };
+
+static const char * const tegra_dpaux_groups[] = {
+	"dpaux-io",
+};
+
+static const char * const tegra_dpaux_functions[] = {
+	"aux",
+	"i2c",
+	"off",
+};
+
+static int tegra_dpaux_get_groups_count(struct pinctrl_dev *pinctrl)
+{
+	return ARRAY_SIZE(tegra_dpaux_groups);
+}
+
+static const char *tegra_dpaux_get_group_name(struct pinctrl_dev *pinctrl,
+					      unsigned int group)
+{
+	return tegra_dpaux_groups[group];
+}
+
+static int tegra_dpaux_get_group_pins(struct pinctrl_dev *pinctrl,
+				      unsigned group, const unsigned **pins,
+				      unsigned *num_pins)
+{
+	*pins = tegra_dpaux_pin_numbers;
+	*num_pins = ARRAY_SIZE(tegra_dpaux_pin_numbers);
+
+	return 0;
+}
+
+static const struct pinctrl_ops tegra_dpaux_pinctrl_ops = {
+	.get_groups_count = tegra_dpaux_get_groups_count,
+	.get_group_name = tegra_dpaux_get_group_name,
+	.get_group_pins = tegra_dpaux_get_group_pins,
+	.dt_node_to_map = pinconf_generic_dt_node_to_map_group,
+	.dt_free_map = pinconf_generic_dt_free_map,
+};
+
+static int tegra_dpaux_get_functions_count(struct pinctrl_dev *pinctrl)
+{
+	return ARRAY_SIZE(tegra_dpaux_functions);
+}
+
+static const char *tegra_dpaux_get_function_name(struct pinctrl_dev *pinctrl,
+						 unsigned int function)
+{
+	return tegra_dpaux_functions[function];
+}
+
+static int tegra_dpaux_get_function_groups(struct pinctrl_dev *pinctrl,
+					   unsigned int function,
+					   const char * const **groups,
+					   unsigned * const num_groups)
+{
+	*num_groups = ARRAY_SIZE(tegra_dpaux_groups);
+	*groups = tegra_dpaux_groups;
+
+	return 0;
+}
+
+static int tegra_dpaux_set_mux(struct pinctrl_dev *pinctrl,
+			       unsigned int function, unsigned int group)
+{
+	struct tegra_dpaux *dpaux = pinctrl_dev_get_drvdata(pinctrl);
+
+	return tegra_dpaux_pad_config(dpaux, function);
+}
+
+static const struct pinmux_ops tegra_dpaux_pinmux_ops = {
+	.get_functions_count = tegra_dpaux_get_functions_count,
+	.get_function_name = tegra_dpaux_get_function_name,
+	.get_function_groups = tegra_dpaux_get_function_groups,
+	.set_mux = tegra_dpaux_set_mux,
+};
+#endif
+
 static int tegra_dpaux_probe(struct platform_device *pdev)
 {
 	struct tegra_dpaux *dpaux;
@@ -294,11 +444,14 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	dpaux->rst = devm_reset_control_get(&pdev->dev, "dpaux");
-	if (IS_ERR(dpaux->rst)) {
-		dev_err(&pdev->dev, "failed to get reset control: %ld\n",
-			PTR_ERR(dpaux->rst));
-		return PTR_ERR(dpaux->rst);
+	if (!pdev->dev.pm_domain) {
+		dpaux->rst = devm_reset_control_get(&pdev->dev, "dpaux");
+		if (IS_ERR(dpaux->rst)) {
+			dev_err(&pdev->dev,
+				"failed to get reset control: %ld\n",
+				PTR_ERR(dpaux->rst));
+			return PTR_ERR(dpaux->rst);
+		}
 	}
 
 	dpaux->clk = devm_clk_get(&pdev->dev, NULL);
@@ -315,34 +468,37 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 		return err;
 	}
 
-	reset_control_deassert(dpaux->rst);
+	if (dpaux->rst)
+		reset_control_deassert(dpaux->rst);
 
 	dpaux->clk_parent = devm_clk_get(&pdev->dev, "parent");
 	if (IS_ERR(dpaux->clk_parent)) {
 		dev_err(&pdev->dev, "failed to get parent clock: %ld\n",
 			PTR_ERR(dpaux->clk_parent));
-		return PTR_ERR(dpaux->clk_parent);
+		err = PTR_ERR(dpaux->clk_parent);
+		goto assert_reset;
 	}
 
 	err = clk_prepare_enable(dpaux->clk_parent);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to enable parent clock: %d\n",
 			err);
-		return err;
+		goto assert_reset;
 	}
 
 	err = clk_set_rate(dpaux->clk_parent, 270000000);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to set clock to 270 MHz: %d\n",
 			err);
-		return err;
+		goto disable_parent_clk;
 	}
 
 	dpaux->vdd = devm_regulator_get(&pdev->dev, "vdd");
 	if (IS_ERR(dpaux->vdd)) {
 		dev_err(&pdev->dev, "failed to get VDD supply: %ld\n",
 			PTR_ERR(dpaux->vdd));
-		return PTR_ERR(dpaux->vdd);
+		err = PTR_ERR(dpaux->vdd);
+		goto disable_parent_clk;
 	}
 
 	err = devm_request_irq(dpaux->dev, dpaux->irq, tegra_dpaux_irq, 0,
@@ -350,7 +506,7 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 	if (err < 0) {
 		dev_err(dpaux->dev, "failed to request IRQ#%u: %d\n",
 			dpaux->irq, err);
-		return err;
+		goto disable_parent_clk;
 	}
 
 	disable_irq(dpaux->irq);
@@ -360,7 +516,7 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 
 	err = drm_dp_aux_register(&dpaux->aux);
 	if (err < 0)
-		return err;
+		goto disable_parent_clk;
 
 	/*
 	 * Assume that by default the DPAUX/I2C pads will be used for HDMI,
@@ -370,16 +526,24 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 	 * is no possibility to perform the I2C mode configuration in the
 	 * HDMI path.
 	 */
-	value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
-	value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
-
-	value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_PADCTL);
-	value = DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV |
-		DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV |
-		DPAUX_HYBRID_PADCTL_MODE_I2C;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL);
+	err = tegra_dpaux_pad_config(dpaux, DPAUX_HYBRID_PADCTL_MODE_I2C);
+	if (err < 0)
+		return err;
 
+#ifdef CONFIG_GENERIC_PINCONF
+	dpaux->desc.name = dev_name(&pdev->dev);
+	dpaux->desc.pins = tegra_dpaux_pins;
+	dpaux->desc.npins = ARRAY_SIZE(tegra_dpaux_pins);
+	dpaux->desc.pctlops = &tegra_dpaux_pinctrl_ops;
+	dpaux->desc.pmxops = &tegra_dpaux_pinmux_ops;
+	dpaux->desc.owner = THIS_MODULE;
+
+	dpaux->pinctrl = devm_pinctrl_register(&pdev->dev, &dpaux->desc, dpaux);
+	if (!dpaux->pinctrl) {
+		dev_err(&pdev->dev, "failed to register pincontrol\n");
+		return -ENODEV;
+	}
+#endif
 	/* enable and clear all interrupts */
 	value = DPAUX_INTR_AUX_DONE | DPAUX_INTR_IRQ_EVENT |
 		DPAUX_INTR_UNPLUG_EVENT | DPAUX_INTR_PLUG_EVENT;
@@ -393,17 +557,24 @@ static int tegra_dpaux_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, dpaux);
 
 	return 0;
+
+disable_parent_clk:
+	clk_disable_unprepare(dpaux->clk_parent);
+assert_reset:
+	if (dpaux->rst)
+		reset_control_assert(dpaux->rst);
+
+	clk_disable_unprepare(dpaux->clk);
+
+	return err;
 }
 
 static int tegra_dpaux_remove(struct platform_device *pdev)
 {
 	struct tegra_dpaux *dpaux = platform_get_drvdata(pdev);
-	u32 value;
 
 	/* make sure pads are powered down when not in use */
-	value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
-	value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+	tegra_dpaux_pad_power_down(dpaux);
 
 	drm_dp_aux_unregister(&dpaux->aux);
 
@@ -414,7 +585,10 @@ static int tegra_dpaux_remove(struct platform_device *pdev)
 	cancel_work_sync(&dpaux->work);
 
 	clk_disable_unprepare(dpaux->clk_parent);
-	reset_control_assert(dpaux->rst);
+
+	if (dpaux->rst)
+		reset_control_assert(dpaux->rst);
+
 	clk_disable_unprepare(dpaux->clk);
 
 	return 0;
@@ -528,30 +702,15 @@ enum drm_connector_status drm_dp_aux_detect(struct drm_dp_aux *aux)
 int drm_dp_aux_enable(struct drm_dp_aux *aux)
 {
 	struct tegra_dpaux *dpaux = to_dpaux(aux);
-	u32 value;
-
-	value = DPAUX_HYBRID_PADCTL_AUX_CMH(2) |
-		DPAUX_HYBRID_PADCTL_AUX_DRVZ(4) |
-		DPAUX_HYBRID_PADCTL_AUX_DRVI(0x18) |
-		DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV |
-		DPAUX_HYBRID_PADCTL_MODE_AUX;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL);
-
-	value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
-	value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
 
-	return 0;
+	return tegra_dpaux_pad_config(dpaux, DPAUX_PADCTL_FUNC_AUX);
 }
 
 int drm_dp_aux_disable(struct drm_dp_aux *aux)
 {
 	struct tegra_dpaux *dpaux = to_dpaux(aux);
-	u32 value;
 
-	value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
-	value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
-	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+	tegra_dpaux_pad_power_down(dpaux);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index a177a42a9849..755264d9db22 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -56,8 +56,8 @@ static void tegra_atomic_complete(struct tegra_drm *tegra,
 	 */
 
 	drm_atomic_helper_commit_modeset_disables(drm, state);
-	drm_atomic_helper_commit_planes(drm, state, false);
 	drm_atomic_helper_commit_modeset_enables(drm, state);
+	drm_atomic_helper_commit_planes(drm, state, true);
 
 	drm_atomic_helper_wait_for_vblanks(drm, state);
 
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
index 099cccb2fbcb..3d228ad90e0f 100644
--- a/drivers/gpu/drm/tegra/dsi.c
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -13,6 +13,7 @@
 #include <linux/of.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
 #include <linux/reset.h>
 
 #include <linux/regulator/consumer.h>
@@ -677,6 +678,45 @@ static void tegra_dsi_ganged_disable(struct tegra_dsi *dsi)
 	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_CONTROL);
 }
 
+static int tegra_dsi_pad_enable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = DSI_PAD_CONTROL_VS1_PULLDN(0) | DSI_PAD_CONTROL_VS1_PDIO(0);
+	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0);
+
+	return 0;
+}
+
+static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	/*
+	 * XXX Is this still needed? The module reset is deasserted right
+	 * before this function is called.
+	 */
+	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0);
+	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1);
+	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_2);
+	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_3);
+	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_4);
+
+	/* start calibration */
+	tegra_dsi_pad_enable(dsi);
+
+	value = DSI_PAD_SLEW_UP(0x7) | DSI_PAD_SLEW_DN(0x7) |
+		DSI_PAD_LP_UP(0x1) | DSI_PAD_LP_DN(0x1) |
+		DSI_PAD_OUT_CLK(0x0);
+	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_2);
+
+	value = DSI_PAD_PREEMP_PD_CLK(0x3) | DSI_PAD_PREEMP_PU_CLK(0x3) |
+		DSI_PAD_PREEMP_PD(0x03) | DSI_PAD_PREEMP_PU(0x3);
+	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_3);
+
+	return tegra_mipi_calibrate(dsi->mipi);
+}
+
 static void tegra_dsi_set_timeout(struct tegra_dsi *dsi, unsigned long bclk,
 				  unsigned int vrefresh)
 {
@@ -836,7 +876,7 @@ static void tegra_dsi_encoder_disable(struct drm_encoder *encoder)
 
 	tegra_dsi_disable(dsi);
 
-	return;
+	pm_runtime_put(dsi->dev);
 }
 
 static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)
@@ -847,6 +887,13 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)
 	struct tegra_dsi *dsi = to_dsi(output);
 	struct tegra_dsi_state *state;
 	u32 value;
+	int err;
+
+	pm_runtime_get_sync(dsi->dev);
+
+	err = tegra_dsi_pad_calibrate(dsi);
+	if (err < 0)
+		dev_err(dsi->dev, "MIPI calibration failed: %d\n", err);
 
 	state = tegra_dsi_get_state(dsi);
 
@@ -875,8 +922,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)
 
 	if (output->panel)
 		drm_panel_enable(output->panel);
-
-	return;
 }
 
 static int
@@ -966,55 +1011,12 @@ static const struct drm_encoder_helper_funcs tegra_dsi_encoder_helper_funcs = {
 	.atomic_check = tegra_dsi_encoder_atomic_check,
 };
 
-static int tegra_dsi_pad_enable(struct tegra_dsi *dsi)
-{
-	u32 value;
-
-	value = DSI_PAD_CONTROL_VS1_PULLDN(0) | DSI_PAD_CONTROL_VS1_PDIO(0);
-	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0);
-
-	return 0;
-}
-
-static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi)
-{
-	u32 value;
-
-	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0);
-	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1);
-	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_2);
-	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_3);
-	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_4);
-
-	/* start calibration */
-	tegra_dsi_pad_enable(dsi);
-
-	value = DSI_PAD_SLEW_UP(0x7) | DSI_PAD_SLEW_DN(0x7) |
-		DSI_PAD_LP_UP(0x1) | DSI_PAD_LP_DN(0x1) |
-		DSI_PAD_OUT_CLK(0x0);
-	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_2);
-
-	value = DSI_PAD_PREEMP_PD_CLK(0x3) | DSI_PAD_PREEMP_PU_CLK(0x3) |
-		DSI_PAD_PREEMP_PD(0x03) | DSI_PAD_PREEMP_PU(0x3);
-	tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_3);
-
-	return tegra_mipi_calibrate(dsi->mipi);
-}
-
 static int tegra_dsi_init(struct host1x_client *client)
 {
 	struct drm_device *drm = dev_get_drvdata(client->parent);
 	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
 	int err;
 
-	reset_control_deassert(dsi->rst);
-
-	err = tegra_dsi_pad_calibrate(dsi);
-	if (err < 0) {
-		dev_err(dsi->dev, "MIPI calibration failed: %d\n", err);
-		goto reset;
-	}
-
 	/* Gangsters must not register their own outputs. */
 	if (!dsi->master) {
 		dsi->output.dev = client->dev;
@@ -1037,12 +1039,9 @@ static int tegra_dsi_init(struct host1x_client *client)
 		drm_connector_register(&dsi->output.connector);
 
 		err = tegra_output_init(drm, &dsi->output);
-		if (err < 0) {
-			dev_err(client->dev,
-				"failed to initialize output: %d\n",
+		if (err < 0)
+			dev_err(dsi->dev, "failed to initialize output: %d\n",
 				err);
-			goto reset;
-		}
 
 		dsi->output.encoder.possible_crtcs = 0x3;
 	}
@@ -1054,10 +1053,6 @@ static int tegra_dsi_init(struct host1x_client *client)
 	}
 
 	return 0;
-
-reset:
-	reset_control_assert(dsi->rst);
-	return err;
 }
 
 static int tegra_dsi_exit(struct host1x_client *client)
@@ -1069,7 +1064,7 @@ static int tegra_dsi_exit(struct host1x_client *client)
 	if (IS_ENABLED(CONFIG_DEBUG_FS))
 		tegra_dsi_debugfs_exit(dsi);
 
-	reset_control_assert(dsi->rst);
+	regulator_disable(dsi->vdd);
 
 	return 0;
 }
@@ -1493,74 +1488,50 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 	dsi->format = MIPI_DSI_FMT_RGB888;
 	dsi->lanes = 4;
 
-	dsi->rst = devm_reset_control_get(&pdev->dev, "dsi");
-	if (IS_ERR(dsi->rst))
-		return PTR_ERR(dsi->rst);
+	if (!pdev->dev.pm_domain) {
+		dsi->rst = devm_reset_control_get(&pdev->dev, "dsi");
+		if (IS_ERR(dsi->rst))
+			return PTR_ERR(dsi->rst);
+	}
 
 	dsi->clk = devm_clk_get(&pdev->dev, NULL);
 	if (IS_ERR(dsi->clk)) {
 		dev_err(&pdev->dev, "cannot get DSI clock\n");
-		err = PTR_ERR(dsi->clk);
-		goto reset;
-	}
-
-	err = clk_prepare_enable(dsi->clk);
-	if (err < 0) {
-		dev_err(&pdev->dev, "cannot enable DSI clock\n");
-		goto reset;
+		return PTR_ERR(dsi->clk);
 	}
 
 	dsi->clk_lp = devm_clk_get(&pdev->dev, "lp");
 	if (IS_ERR(dsi->clk_lp)) {
 		dev_err(&pdev->dev, "cannot get low-power clock\n");
-		err = PTR_ERR(dsi->clk_lp);
-		goto disable_clk;
-	}
-
-	err = clk_prepare_enable(dsi->clk_lp);
-	if (err < 0) {
-		dev_err(&pdev->dev, "cannot enable low-power clock\n");
-		goto disable_clk;
+		return PTR_ERR(dsi->clk_lp);
 	}
 
 	dsi->clk_parent = devm_clk_get(&pdev->dev, "parent");
 	if (IS_ERR(dsi->clk_parent)) {
 		dev_err(&pdev->dev, "cannot get parent clock\n");
-		err = PTR_ERR(dsi->clk_parent);
-		goto disable_clk_lp;
+		return PTR_ERR(dsi->clk_parent);
 	}
 
 	dsi->vdd = devm_regulator_get(&pdev->dev, "avdd-dsi-csi");
 	if (IS_ERR(dsi->vdd)) {
 		dev_err(&pdev->dev, "cannot get VDD supply\n");
-		err = PTR_ERR(dsi->vdd);
-		goto disable_clk_lp;
-	}
-
-	err = regulator_enable(dsi->vdd);
-	if (err < 0) {
-		dev_err(&pdev->dev, "cannot enable VDD supply\n");
-		goto disable_clk_lp;
+		return PTR_ERR(dsi->vdd);
 	}
 
 	err = tegra_dsi_setup_clocks(dsi);
 	if (err < 0) {
 		dev_err(&pdev->dev, "cannot setup clocks\n");
-		goto disable_vdd;
+		return err;
 	}
 
 	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	dsi->regs = devm_ioremap_resource(&pdev->dev, regs);
-	if (IS_ERR(dsi->regs)) {
-		err = PTR_ERR(dsi->regs);
-		goto disable_vdd;
-	}
+	if (IS_ERR(dsi->regs))
+		return PTR_ERR(dsi->regs);
 
 	dsi->mipi = tegra_mipi_request(&pdev->dev);
-	if (IS_ERR(dsi->mipi)) {
-		err = PTR_ERR(dsi->mipi);
-		goto disable_vdd;
-	}
+	if (IS_ERR(dsi->mipi))
+		return PTR_ERR(dsi->mipi);
 
 	dsi->host.ops = &tegra_dsi_host_ops;
 	dsi->host.dev = &pdev->dev;
@@ -1571,6 +1542,9 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 		goto mipi_free;
 	}
 
+	platform_set_drvdata(pdev, dsi);
+	pm_runtime_enable(&pdev->dev);
+
 	INIT_LIST_HEAD(&dsi->client.list);
 	dsi->client.ops = &dsi_client_ops;
 	dsi->client.dev = &pdev->dev;
@@ -1582,22 +1556,12 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 		goto unregister;
 	}
 
-	platform_set_drvdata(pdev, dsi);
-
 	return 0;
 
 unregister:
 	mipi_dsi_host_unregister(&dsi->host);
 mipi_free:
 	tegra_mipi_free(dsi->mipi);
-disable_vdd:
-	regulator_disable(dsi->vdd);
-disable_clk_lp:
-	clk_disable_unprepare(dsi->clk_lp);
-disable_clk:
-	clk_disable_unprepare(dsi->clk);
-reset:
-	reset_control_assert(dsi->rst);
 	return err;
 }
 
@@ -1606,6 +1570,8 @@ static int tegra_dsi_remove(struct platform_device *pdev)
 	struct tegra_dsi *dsi = platform_get_drvdata(pdev);
 	int err;
 
+	pm_runtime_disable(&pdev->dev);
+
 	err = host1x_client_unregister(&dsi->client);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to unregister host1x client: %d\n",
@@ -1618,14 +1584,82 @@ static int tegra_dsi_remove(struct platform_device *pdev)
 	mipi_dsi_host_unregister(&dsi->host);
 	tegra_mipi_free(dsi->mipi);
 
-	regulator_disable(dsi->vdd);
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int tegra_dsi_suspend(struct device *dev)
+{
+	struct tegra_dsi *dsi = dev_get_drvdata(dev);
+	int err;
+
+	if (dsi->rst) {
+		err = reset_control_assert(dsi->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to assert reset: %d\n", err);
+			return err;
+		}
+	}
+
+	usleep_range(1000, 2000);
+
 	clk_disable_unprepare(dsi->clk_lp);
 	clk_disable_unprepare(dsi->clk);
-	reset_control_assert(dsi->rst);
+
+	regulator_disable(dsi->vdd);
 
 	return 0;
 }
 
+static int tegra_dsi_resume(struct device *dev)
+{
+	struct tegra_dsi *dsi = dev_get_drvdata(dev);
+	int err;
+
+	err = regulator_enable(dsi->vdd);
+	if (err < 0) {
+		dev_err(dsi->dev, "failed to enable VDD supply: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(dsi->clk);
+	if (err < 0) {
+		dev_err(dev, "cannot enable DSI clock: %d\n", err);
+		goto disable_vdd;
+	}
+
+	err = clk_prepare_enable(dsi->clk_lp);
+	if (err < 0) {
+		dev_err(dev, "cannot enable low-power clock: %d\n", err);
+		goto disable_clk;
+	}
+
+	usleep_range(1000, 2000);
+
+	if (dsi->rst) {
+		err = reset_control_deassert(dsi->rst);
+		if (err < 0) {
+			dev_err(dev, "cannot assert reset: %d\n", err);
+			goto disable_clk_lp;
+		}
+	}
+
+	return 0;
+
+disable_clk_lp:
+	clk_disable_unprepare(dsi->clk_lp);
+disable_clk:
+	clk_disable_unprepare(dsi->clk);
+disable_vdd:
+	regulator_disable(dsi->vdd);
+	return err;
+}
+#endif
+
+static const struct dev_pm_ops tegra_dsi_pm_ops = {
+	SET_RUNTIME_PM_OPS(tegra_dsi_suspend, tegra_dsi_resume, NULL)
+};
+
 static const struct of_device_id tegra_dsi_of_match[] = {
 	{ .compatible = "nvidia,tegra210-dsi", },
 	{ .compatible = "nvidia,tegra132-dsi", },
@@ -1639,6 +1673,7 @@ struct platform_driver tegra_dsi_driver = {
 	.driver = {
 		.name = "tegra-dsi",
 		.of_match_table = tegra_dsi_of_match,
+		.pm = &tegra_dsi_pm_ops,
 	},
 	.probe = tegra_dsi_probe,
 	.remove = tegra_dsi_remove,
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
index 2fdb8796443e..cda0491ed6bf 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -11,6 +11,7 @@
 #include <linux/debugfs.h>
 #include <linux/gpio.h>
 #include <linux/hdmi.h>
+#include <linux/pm_runtime.h>
 #include <linux/regulator/consumer.h>
 #include <linux/reset.h>
 
@@ -18,10 +19,14 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
 
+#include <sound/hda_verbs.h>
+
 #include "hdmi.h"
 #include "drm.h"
 #include "dc.h"
 
+#define HDMI_ELD_BUFFER_SIZE 96
+
 struct tmds_config {
 	unsigned int pclk;
 	u32 pll0;
@@ -39,6 +44,8 @@ struct tegra_hdmi_config {
 	u32 fuse_override_value;
 
 	bool has_sor_io_peak_current;
+	bool has_hda;
+	bool has_hbr;
 };
 
 struct tegra_hdmi {
@@ -60,7 +67,10 @@ struct tegra_hdmi {
 	const struct tegra_hdmi_config *config;
 
 	unsigned int audio_source;
-	unsigned int audio_freq;
+	unsigned int audio_sample_rate;
+	unsigned int audio_channels;
+
+	unsigned int pixel_clock;
 	bool stereo;
 	bool dvi;
 
@@ -402,11 +412,11 @@ static const struct tmds_config tegra124_tmds_config[] = {
 };
 
 static const struct tegra_hdmi_audio_config *
-tegra_hdmi_get_audio_config(unsigned int audio_freq, unsigned int pclk)
+tegra_hdmi_get_audio_config(unsigned int sample_rate, unsigned int pclk)
 {
 	const struct tegra_hdmi_audio_config *table;
 
-	switch (audio_freq) {
+	switch (sample_rate) {
 	case 32000:
 		table = tegra_hdmi_audio_32k;
 		break;
@@ -476,44 +486,114 @@ static void tegra_hdmi_setup_audio_fs_tables(struct tegra_hdmi *hdmi)
 	}
 }
 
-static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk)
+static void tegra_hdmi_write_aval(struct tegra_hdmi *hdmi, u32 value)
+{
+	static const struct {
+		unsigned int sample_rate;
+		unsigned int offset;
+	} regs[] = {
+		{  32000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320 },
+		{  44100, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441 },
+		{  48000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480 },
+		{  88200, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882 },
+		{  96000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960 },
+		{ 176400, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764 },
+		{ 192000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920 },
+	};
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs); i++) {
+		if (regs[i].sample_rate == hdmi->audio_sample_rate) {
+			tegra_hdmi_writel(hdmi, value, regs[i].offset);
+			break;
+		}
+	}
+}
+
+static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi)
 {
-	struct device_node *node = hdmi->dev->of_node;
 	const struct tegra_hdmi_audio_config *config;
-	unsigned int offset = 0;
-	u32 value;
+	u32 source, value;
 
 	switch (hdmi->audio_source) {
 	case HDA:
-		value = AUDIO_CNTRL0_SOURCE_SELECT_HDAL;
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL;
+		else
+			return -EINVAL;
+
 		break;
 
 	case SPDIF:
-		value = AUDIO_CNTRL0_SOURCE_SELECT_SPDIF;
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF;
+		else
+			source = AUDIO_CNTRL0_SOURCE_SELECT_SPDIF;
 		break;
 
 	default:
-		value = AUDIO_CNTRL0_SOURCE_SELECT_AUTO;
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO;
+		else
+			source = AUDIO_CNTRL0_SOURCE_SELECT_AUTO;
 		break;
 	}
 
-	if (of_device_is_compatible(node, "nvidia,tegra30-hdmi")) {
-		value |= AUDIO_CNTRL0_ERROR_TOLERANCE(6) |
-			 AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0);
-		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0);
-	} else {
-		value |= AUDIO_CNTRL0_INJECT_NULLSMPL;
+	/*
+	 * Tegra30 and later use a slightly modified version of the register
+	 * layout to accomodate for changes related to supporting HDA as the
+	 * audio input source for HDMI. The source select field has moved to
+	 * the SOR_AUDIO_CNTRL0 register, but the error tolerance and frames
+	 * per block fields remain in the AUDIO_CNTRL0 register.
+	 */
+	if (hdmi->config->has_hda) {
+		/*
+		 * Inject null samples into the audio FIFO for every frame in
+		 * which the codec did not receive any samples. This applies
+		 * to stereo LPCM only.
+		 *
+		 * XXX: This seems to be a remnant of MCP days when this was
+		 * used to work around issues with monitors not being able to
+		 * play back system startup sounds early. It is possibly not
+		 * needed on Linux at all.
+		 */
+		if (hdmi->audio_channels == 2)
+			value = SOR_AUDIO_CNTRL0_INJECT_NULLSMPL;
+		else
+			value = 0;
+
+		value |= source;
+
 		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_CNTRL0);
+	}
 
-		value = AUDIO_CNTRL0_ERROR_TOLERANCE(6) |
-			AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0);
-		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0);
+	/*
+	 * On Tegra20, HDA is not a supported audio source and the source
+	 * select field is part of the AUDIO_CNTRL0 register.
+	 */
+	value = AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0) |
+		AUDIO_CNTRL0_ERROR_TOLERANCE(6);
+
+	if (!hdmi->config->has_hda)
+		value |= source;
+
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0);
+
+	/*
+	 * Advertise support for High Bit-Rate on Tegra114 and later.
+	 */
+	if (hdmi->config->has_hbr) {
+		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_SPARE0);
+		value |= SOR_AUDIO_SPARE0_HBR_ENABLE;
+		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_SPARE0);
 	}
 
-	config = tegra_hdmi_get_audio_config(hdmi->audio_freq, pclk);
+	config = tegra_hdmi_get_audio_config(hdmi->audio_sample_rate,
+					     hdmi->pixel_clock);
 	if (!config) {
-		dev_err(hdmi->dev, "cannot set audio to %u at %u pclk\n",
-			hdmi->audio_freq, pclk);
+		dev_err(hdmi->dev,
+			"cannot set audio to %u Hz at %u Hz pixel clock\n",
+			hdmi->audio_sample_rate, hdmi->pixel_clock);
 		return -EINVAL;
 	}
 
@@ -526,8 +606,8 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk)
 	tegra_hdmi_writel(hdmi, ACR_SUBPACK_N(config->n) | ACR_ENABLE,
 			  HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH);
 
-	value = ACR_SUBPACK_CTS(config->cts);
-	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW);
+	tegra_hdmi_writel(hdmi, ACR_SUBPACK_CTS(config->cts),
+			  HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW);
 
 	value = SPARE_HW_CTS | SPARE_FORCE_SW_CTS | SPARE_CTS_RESET_VAL(1);
 	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_SPARE);
@@ -536,43 +616,53 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk)
 	value &= ~AUDIO_N_RESETF;
 	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N);
 
-	if (of_device_is_compatible(node, "nvidia,tegra30-hdmi")) {
-		switch (hdmi->audio_freq) {
-		case 32000:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320;
-			break;
+	if (hdmi->config->has_hda)
+		tegra_hdmi_write_aval(hdmi, config->aval);
 
-		case 44100:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441;
-			break;
+	tegra_hdmi_setup_audio_fs_tables(hdmi);
 
-		case 48000:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480;
-			break;
+	return 0;
+}
 
-		case 88200:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882;
-			break;
+static void tegra_hdmi_disable_audio(struct tegra_hdmi *hdmi)
+{
+	u32 value;
 
-		case 96000:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960;
-			break;
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value &= ~GENERIC_CTRL_AUDIO;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
 
-		case 176400:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764;
-			break;
+static void tegra_hdmi_enable_audio(struct tegra_hdmi *hdmi)
+{
+	u32 value;
 
-		case 192000:
-			offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920;
-			break;
-		}
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value |= GENERIC_CTRL_AUDIO;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
 
-		tegra_hdmi_writel(hdmi, config->aval, offset);
-	}
+static void tegra_hdmi_write_eld(struct tegra_hdmi *hdmi)
+{
+	size_t length = drm_eld_size(hdmi->output.connector.eld), i;
+	u32 value;
 
-	tegra_hdmi_setup_audio_fs_tables(hdmi);
+	for (i = 0; i < length; i++)
+		tegra_hdmi_writel(hdmi, i << 8 | hdmi->output.connector.eld[i],
+				  HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR);
 
-	return 0;
+	/*
+	 * The HDA codec will always report an ELD buffer size of 96 bytes and
+	 * the HDA codec driver will check that each byte read from the buffer
+	 * is valid. Therefore every byte must be written, even if no 96 bytes
+	 * were parsed from EDID.
+	 */
+	for (i = length; i < HDMI_ELD_BUFFER_SIZE; i++)
+		tegra_hdmi_writel(hdmi, i << 8 | 0,
+				  HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR);
+
+	value = SOR_AUDIO_HDA_PRESENSE_VALID | SOR_AUDIO_HDA_PRESENSE_PRESENT;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE);
 }
 
 static inline u32 tegra_hdmi_subpack(const u8 *ptr, size_t size)
@@ -644,12 +734,6 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
 	u8 buffer[17];
 	ssize_t err;
 
-	if (hdmi->dvi) {
-		tegra_hdmi_writel(hdmi, 0,
-				  HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
-		return;
-	}
-
 	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode);
 	if (err < 0) {
 		dev_err(hdmi->dev, "failed to setup AVI infoframe: %zd\n", err);
@@ -663,9 +747,24 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
 	}
 
 	tegra_hdmi_write_infopack(hdmi, buffer, err);
+}
+
+static void tegra_hdmi_disable_avi_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
 
-	tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE,
-			  HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_enable_avi_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
 }
 
 static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
@@ -674,12 +773,6 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 	u8 buffer[14];
 	ssize_t err;
 
-	if (hdmi->dvi) {
-		tegra_hdmi_writel(hdmi, 0,
-				  HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
-		return;
-	}
-
 	err = hdmi_audio_infoframe_init(&frame);
 	if (err < 0) {
 		dev_err(hdmi->dev, "failed to setup audio infoframe: %zd\n",
@@ -687,7 +780,7 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 		return;
 	}
 
-	frame.channels = 2;
+	frame.channels = hdmi->audio_channels;
 
 	err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer));
 	if (err < 0) {
@@ -703,9 +796,24 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
 	 * bytes can be programmed.
 	 */
 	tegra_hdmi_write_infopack(hdmi, buffer, min_t(size_t, 10, err));
+}
 
-	tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE,
-			  HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+static void tegra_hdmi_disable_audio_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_enable_audio_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
 }
 
 static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
@@ -713,14 +821,6 @@ static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
 	struct hdmi_vendor_infoframe frame;
 	u8 buffer[10];
 	ssize_t err;
-	u32 value;
-
-	if (!hdmi->stereo) {
-		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
-		value &= ~GENERIC_CTRL_ENABLE;
-		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
-		return;
-	}
 
 	hdmi_vendor_infoframe_init(&frame);
 	frame.s3d_struct = HDMI_3D_STRUCTURE_FRAME_PACKING;
@@ -733,6 +833,20 @@ static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
 	}
 
 	tegra_hdmi_write_infopack(hdmi, buffer, err);
+}
+
+static void tegra_hdmi_disable_stereo_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value &= ~GENERIC_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
+
+static void tegra_hdmi_enable_stereo_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
 
 	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
 	value |= GENERIC_CTRL_ENABLE;
@@ -772,10 +886,25 @@ static bool tegra_output_is_hdmi(struct tegra_output *output)
 	return drm_detect_hdmi_monitor(edid);
 }
 
+static enum drm_connector_status
+tegra_hdmi_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	enum drm_connector_status status;
+
+	status = tegra_output_connector_detect(connector, force);
+	if (status == connector_status_connected)
+		return status;
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE);
+	return status;
+}
+
 static const struct drm_connector_funcs tegra_hdmi_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
 	.reset = drm_atomic_helper_connector_reset,
-	.detect = tegra_output_connector_detect,
+	.detect = tegra_hdmi_connector_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.destroy = tegra_output_connector_destroy,
 	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
@@ -814,7 +943,9 @@ static const struct drm_encoder_funcs tegra_hdmi_encoder_funcs = {
 
 static void tegra_hdmi_encoder_disable(struct drm_encoder *encoder)
 {
+	struct tegra_output *output = encoder_to_output(encoder);
 	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
 	u32 value;
 
 	/*
@@ -828,6 +959,20 @@ static void tegra_hdmi_encoder_disable(struct drm_encoder *encoder)
 
 		tegra_dc_commit(dc);
 	}
+
+	if (!hdmi->dvi) {
+		if (hdmi->stereo)
+			tegra_hdmi_disable_stereo_infoframe(hdmi);
+
+		tegra_hdmi_disable_audio_infoframe(hdmi);
+		tegra_hdmi_disable_avi_infoframe(hdmi);
+		tegra_hdmi_disable_audio(hdmi);
+	}
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_ENABLE);
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_MASK);
+
+	pm_runtime_put(hdmi->dev);
 }
 
 static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
@@ -836,21 +981,28 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
 	unsigned int h_sync_width, h_front_porch, h_back_porch, i, rekey;
 	struct tegra_output *output = encoder_to_output(encoder);
 	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
-	struct device_node *node = output->dev->of_node;
 	struct tegra_hdmi *hdmi = to_hdmi(output);
-	unsigned int pulse_start, div82, pclk;
+	unsigned int pulse_start, div82;
 	int retries = 1000;
 	u32 value;
 	int err;
 
-	hdmi->dvi = !tegra_output_is_hdmi(output);
+	pm_runtime_get_sync(hdmi->dev);
 
-	pclk = mode->clock * 1000;
+	/*
+	 * Enable and unmask the HDA codec SCRATCH0 register interrupt. This
+	 * is used for interoperability between the HDA codec driver and the
+	 * HDMI driver.
+	 */
+	tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_ENABLE);
+	tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_MASK);
+
+	hdmi->pixel_clock = mode->clock * 1000;
 	h_sync_width = mode->hsync_end - mode->hsync_start;
 	h_back_porch = mode->htotal - mode->hsync_end;
 	h_front_porch = mode->hsync_start - mode->hdisplay;
 
-	err = clk_set_rate(hdmi->clk, pclk);
+	err = clk_set_rate(hdmi->clk, hdmi->pixel_clock);
 	if (err < 0) {
 		dev_err(hdmi->dev, "failed to set HDMI clock frequency: %d\n",
 			err);
@@ -909,17 +1061,15 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
 	value = SOR_REFCLK_DIV_INT(div82 >> 2) | SOR_REFCLK_DIV_FRAC(div82);
 	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_REFCLK);
 
+	hdmi->dvi = !tegra_output_is_hdmi(output);
 	if (!hdmi->dvi) {
-		err = tegra_hdmi_setup_audio(hdmi, pclk);
+		err = tegra_hdmi_setup_audio(hdmi);
 		if (err < 0)
 			hdmi->dvi = true;
 	}
 
-	if (of_device_is_compatible(node, "nvidia,tegra20-hdmi")) {
-		/*
-		 * TODO: add ELD support
-		 */
-	}
+	if (hdmi->config->has_hda)
+		tegra_hdmi_write_eld(hdmi);
 
 	rekey = HDMI_REKEY_DEFAULT;
 	value = HDMI_CTRL_REKEY(rekey);
@@ -931,20 +1081,17 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
 
 	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_CTRL);
 
-	if (hdmi->dvi)
-		tegra_hdmi_writel(hdmi, 0x0,
-				  HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
-	else
-		tegra_hdmi_writel(hdmi, GENERIC_CTRL_AUDIO,
-				  HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	if (!hdmi->dvi) {
+		tegra_hdmi_setup_avi_infoframe(hdmi, mode);
+		tegra_hdmi_setup_audio_infoframe(hdmi);
 
-	tegra_hdmi_setup_avi_infoframe(hdmi, mode);
-	tegra_hdmi_setup_audio_infoframe(hdmi);
-	tegra_hdmi_setup_stereo_infoframe(hdmi);
+		if (hdmi->stereo)
+			tegra_hdmi_setup_stereo_infoframe(hdmi);
+	}
 
 	/* TMDS CONFIG */
 	for (i = 0; i < hdmi->config->num_tmds; i++) {
-		if (pclk <= hdmi->config->tmds[i].pclk) {
+		if (hdmi->pixel_clock <= hdmi->config->tmds[i].pclk) {
 			tegra_hdmi_setup_tmds(hdmi, &hdmi->config->tmds[i]);
 			break;
 		}
@@ -1031,6 +1178,15 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
 
 	tegra_dc_commit(dc);
 
+	if (!hdmi->dvi) {
+		tegra_hdmi_enable_avi_infoframe(hdmi);
+		tegra_hdmi_enable_audio_infoframe(hdmi);
+		tegra_hdmi_enable_audio(hdmi);
+
+		if (hdmi->stereo)
+			tegra_hdmi_enable_stereo_infoframe(hdmi);
+	}
+
 	/* TODO: add HDCP support */
 }
 
@@ -1235,8 +1391,14 @@ static int tegra_hdmi_show_regs(struct seq_file *s, void *data)
 	DUMP_REG(HDMI_NV_PDISP_KEY_HDCP_KEY_TRIG);
 	DUMP_REG(HDMI_NV_PDISP_KEY_SKEY_INDEX);
 	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_CNTRL0);
+	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_SPARE0);
+	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0);
+	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1);
 	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR);
 	DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE);
+	DUMP_REG(HDMI_NV_PDISP_INT_STATUS);
+	DUMP_REG(HDMI_NV_PDISP_INT_MASK);
+	DUMP_REG(HDMI_NV_PDISP_INT_ENABLE);
 	DUMP_REG(HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT);
 
 #undef DUMP_REG
@@ -1360,14 +1522,6 @@ static int tegra_hdmi_init(struct host1x_client *client)
 		return err;
 	}
 
-	err = clk_prepare_enable(hdmi->clk);
-	if (err < 0) {
-		dev_err(hdmi->dev, "failed to enable clock: %d\n", err);
-		return err;
-	}
-
-	reset_control_deassert(hdmi->rst);
-
 	return 0;
 }
 
@@ -1377,9 +1531,6 @@ static int tegra_hdmi_exit(struct host1x_client *client)
 
 	tegra_output_exit(&hdmi->output);
 
-	reset_control_assert(hdmi->rst);
-	clk_disable_unprepare(hdmi->clk);
-
 	regulator_disable(hdmi->vdd);
 	regulator_disable(hdmi->pll);
 	regulator_disable(hdmi->hdmi);
@@ -1401,6 +1552,8 @@ static const struct tegra_hdmi_config tegra20_hdmi_config = {
 	.fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT,
 	.fuse_override_value = 1 << 31,
 	.has_sor_io_peak_current = false,
+	.has_hda = false,
+	.has_hbr = false,
 };
 
 static const struct tegra_hdmi_config tegra30_hdmi_config = {
@@ -1409,6 +1562,8 @@ static const struct tegra_hdmi_config tegra30_hdmi_config = {
 	.fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT,
 	.fuse_override_value = 1 << 31,
 	.has_sor_io_peak_current = false,
+	.has_hda = true,
+	.has_hbr = false,
 };
 
 static const struct tegra_hdmi_config tegra114_hdmi_config = {
@@ -1417,6 +1572,8 @@ static const struct tegra_hdmi_config tegra114_hdmi_config = {
 	.fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0,
 	.fuse_override_value = 1 << 31,
 	.has_sor_io_peak_current = true,
+	.has_hda = true,
+	.has_hbr = true,
 };
 
 static const struct tegra_hdmi_config tegra124_hdmi_config = {
@@ -1425,6 +1582,8 @@ static const struct tegra_hdmi_config tegra124_hdmi_config = {
 	.fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0,
 	.fuse_override_value = 1 << 31,
 	.has_sor_io_peak_current = true,
+	.has_hda = true,
+	.has_hbr = true,
 };
 
 static const struct of_device_id tegra_hdmi_of_match[] = {
@@ -1436,6 +1595,67 @@ static const struct of_device_id tegra_hdmi_of_match[] = {
 };
 MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match);
 
+static void hda_format_parse(unsigned int format, unsigned int *rate,
+			     unsigned int *channels)
+{
+	unsigned int mul, div;
+
+	if (format & AC_FMT_BASE_44K)
+		*rate = 44100;
+	else
+		*rate = 48000;
+
+	mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT;
+	div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT;
+
+	*rate = *rate * (mul + 1) / (div + 1);
+
+	*channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT;
+}
+
+static irqreturn_t tegra_hdmi_irq(int irq, void *data)
+{
+	struct tegra_hdmi *hdmi = data;
+	u32 value;
+	int err;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_INT_STATUS);
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_INT_STATUS);
+
+	if (value & INT_CODEC_SCRATCH0) {
+		unsigned int format;
+		u32 value;
+
+		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0);
+
+		if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) {
+			unsigned int sample_rate, channels;
+
+			format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK;
+
+			hda_format_parse(format, &sample_rate, &channels);
+
+			hdmi->audio_sample_rate = sample_rate;
+			hdmi->audio_channels = channels;
+
+			err = tegra_hdmi_setup_audio(hdmi);
+			if (err < 0) {
+				tegra_hdmi_disable_audio_infoframe(hdmi);
+				tegra_hdmi_disable_audio(hdmi);
+			} else {
+				tegra_hdmi_setup_audio_infoframe(hdmi);
+				tegra_hdmi_enable_audio_infoframe(hdmi);
+				tegra_hdmi_enable_audio(hdmi);
+			}
+		} else {
+			tegra_hdmi_disable_audio_infoframe(hdmi);
+			tegra_hdmi_disable_audio(hdmi);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
 static int tegra_hdmi_probe(struct platform_device *pdev)
 {
 	const struct of_device_id *match;
@@ -1453,8 +1673,10 @@ static int tegra_hdmi_probe(struct platform_device *pdev)
 
 	hdmi->config = match->data;
 	hdmi->dev = &pdev->dev;
+
 	hdmi->audio_source = AUTO;
-	hdmi->audio_freq = 44100;
+	hdmi->audio_sample_rate = 48000;
+	hdmi->audio_channels = 2;
 	hdmi->stereo = false;
 	hdmi->dvi = false;
 
@@ -1515,6 +1737,17 @@ static int tegra_hdmi_probe(struct platform_device *pdev)
 
 	hdmi->irq = err;
 
+	err = devm_request_irq(hdmi->dev, hdmi->irq, tegra_hdmi_irq, 0,
+			       dev_name(hdmi->dev), hdmi);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to request IRQ#%u: %d\n",
+			hdmi->irq, err);
+		return err;
+	}
+
+	platform_set_drvdata(pdev, hdmi);
+	pm_runtime_enable(&pdev->dev);
+
 	INIT_LIST_HEAD(&hdmi->client.list);
 	hdmi->client.ops = &hdmi_client_ops;
 	hdmi->client.dev = &pdev->dev;
@@ -1526,8 +1759,6 @@ static int tegra_hdmi_probe(struct platform_device *pdev)
 		return err;
 	}
 
-	platform_set_drvdata(pdev, hdmi);
-
 	return 0;
 }
 
@@ -1536,6 +1767,8 @@ static int tegra_hdmi_remove(struct platform_device *pdev)
 	struct tegra_hdmi *hdmi = platform_get_drvdata(pdev);
 	int err;
 
+	pm_runtime_disable(&pdev->dev);
+
 	err = host1x_client_unregister(&hdmi->client);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to unregister host1x client: %d\n",
@@ -1545,17 +1778,61 @@ static int tegra_hdmi_remove(struct platform_device *pdev)
 
 	tegra_output_remove(&hdmi->output);
 
-	clk_disable_unprepare(hdmi->clk_parent);
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int tegra_hdmi_suspend(struct device *dev)
+{
+	struct tegra_hdmi *hdmi = dev_get_drvdata(dev);
+	int err;
+
+	err = reset_control_assert(hdmi->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to assert reset: %d\n", err);
+		return err;
+	}
+
+	usleep_range(1000, 2000);
+
 	clk_disable_unprepare(hdmi->clk);
 
 	return 0;
 }
 
+static int tegra_hdmi_resume(struct device *dev)
+{
+	struct tegra_hdmi *hdmi = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_prepare_enable(hdmi->clk);
+	if (err < 0) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		return err;
+	}
+
+	usleep_range(1000, 2000);
+
+	err = reset_control_deassert(hdmi->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to deassert reset: %d\n", err);
+		clk_disable_unprepare(hdmi->clk);
+		return err;
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops tegra_hdmi_pm_ops = {
+	SET_RUNTIME_PM_OPS(tegra_hdmi_suspend, tegra_hdmi_resume, NULL)
+};
+
 struct platform_driver tegra_hdmi_driver = {
 	.driver = {
 		.name = "tegra-hdmi",
-		.owner = THIS_MODULE,
 		.of_match_table = tegra_hdmi_of_match,
+		.pm = &tegra_hdmi_pm_ops,
 	},
 	.probe = tegra_hdmi_probe,
 	.remove = tegra_hdmi_remove,
diff --git a/drivers/gpu/drm/tegra/hdmi.h b/drivers/gpu/drm/tegra/hdmi.h
index a882514389cd..2339f134a09a 100644
--- a/drivers/gpu/drm/tegra/hdmi.h
+++ b/drivers/gpu/drm/tegra/hdmi.h
@@ -468,9 +468,20 @@
 #define HDMI_NV_PDISP_KEY_SKEY_INDEX				0xa3
 
 #define HDMI_NV_PDISP_SOR_AUDIO_CNTRL0				0xac
-#define AUDIO_CNTRL0_INJECT_NULLSMPL (1 << 29)
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO	(0 << 20)
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF	(1 << 20)
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL	(2 << 20)
+#define  SOR_AUDIO_CNTRL0_INJECT_NULLSMPL	(1 << 29)
+#define HDMI_NV_PDISP_SOR_AUDIO_SPARE0				0xae
+#define  SOR_AUDIO_SPARE0_HBR_ENABLE		(1 << 27)
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0		0xba
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID	(1 << 30)
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK	0xffff
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1		0xbb
 #define HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR			0xbc
 #define HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE			0xbd
+#define  SOR_AUDIO_HDA_PRESENSE_VALID		(1 << 1)
+#define  SOR_AUDIO_HDA_PRESENSE_PRESENT		(1 << 0)
 
 #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320    0xbf
 #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441    0xc0
@@ -481,6 +492,14 @@
 #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920    0xc5
 #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_DEFAULT 0xc5
 
+#define HDMI_NV_PDISP_INT_STATUS			0xcc
+#define  INT_SCRATCH		(1 << 3)
+#define  INT_CP_REQUEST		(1 << 2)
+#define  INT_CODEC_SCRATCH1	(1 << 1)
+#define  INT_CODEC_SCRATCH0	(1 << 0)
+#define HDMI_NV_PDISP_INT_MASK				0xcd
+#define HDMI_NV_PDISP_INT_ENABLE			0xce
+
 #define HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT		0xd1
 #define PEAK_CURRENT_LANE0(x) (((x) & 0x7f) <<  0)
 #define PEAK_CURRENT_LANE1(x) (((x) & 0x7f) <<  8)
diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c
index 1480f6aaffe4..595d1ec3e02e 100644
--- a/drivers/gpu/drm/tegra/output.c
+++ b/drivers/gpu/drm/tegra/output.c
@@ -36,6 +36,7 @@ int tegra_output_connector_get_modes(struct drm_connector *connector)
 
 	if (edid) {
 		err = drm_add_edid_modes(connector, edid);
+		drm_edid_to_eld(connector, edid);
 		kfree(edid);
 	}
 
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
index 34958d71284b..74d0540b8d4c 100644
--- a/drivers/gpu/drm/tegra/sor.c
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -7,11 +7,13 @@
  */
 
 #include <linux/clk.h>
+#include <linux/clk-provider.h>
 #include <linux/debugfs.h>
 #include <linux/gpio.h>
 #include <linux/io.h>
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
 #include <linux/regulator/consumer.h>
 #include <linux/reset.h>
 
@@ -149,6 +151,8 @@ struct tegra_sor_soc {
 
 	const struct tegra_sor_hdmi_settings *settings;
 	unsigned int num_settings;
+
+	const u8 *xbar_cfg;
 };
 
 struct tegra_sor;
@@ -169,7 +173,9 @@ struct tegra_sor {
 
 	struct reset_control *rst;
 	struct clk *clk_parent;
+	struct clk *clk_brick;
 	struct clk *clk_safe;
+	struct clk *clk_src;
 	struct clk *clk_dp;
 	struct clk *clk;
 
@@ -190,6 +196,18 @@ struct tegra_sor {
 	struct regulator *hdmi_supply;
 };
 
+struct tegra_sor_state {
+	struct drm_connector_state base;
+
+	unsigned int bpc;
+};
+
+static inline struct tegra_sor_state *
+to_sor_state(struct drm_connector_state *state)
+{
+	return container_of(state, struct tegra_sor_state, base);
+}
+
 struct tegra_sor_config {
 	u32 bits_per_pixel;
 
@@ -225,6 +243,118 @@ static inline void tegra_sor_writel(struct tegra_sor *sor, u32 value,
 	writel(value, sor->regs + (offset << 2));
 }
 
+static int tegra_sor_set_parent_clock(struct tegra_sor *sor, struct clk *parent)
+{
+	int err;
+
+	clk_disable_unprepare(sor->clk);
+
+	err = clk_set_parent(sor->clk, parent);
+	if (err < 0)
+		return err;
+
+	err = clk_prepare_enable(sor->clk);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+struct tegra_clk_sor_brick {
+	struct clk_hw hw;
+	struct tegra_sor *sor;
+};
+
+static inline struct tegra_clk_sor_brick *to_brick(struct clk_hw *hw)
+{
+	return container_of(hw, struct tegra_clk_sor_brick, hw);
+}
+
+static const char * const tegra_clk_sor_brick_parents[] = {
+	"pll_d2_out0", "pll_dp"
+};
+
+static int tegra_clk_sor_brick_set_parent(struct clk_hw *hw, u8 index)
+{
+	struct tegra_clk_sor_brick *brick = to_brick(hw);
+	struct tegra_sor *sor = brick->sor;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+	value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK;
+
+	switch (index) {
+	case 0:
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK;
+		break;
+
+	case 1:
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK;
+		break;
+	}
+
+	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
+
+	return 0;
+}
+
+static u8 tegra_clk_sor_brick_get_parent(struct clk_hw *hw)
+{
+	struct tegra_clk_sor_brick *brick = to_brick(hw);
+	struct tegra_sor *sor = brick->sor;
+	u8 parent = U8_MAX;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+
+	switch (value & SOR_CLK_CNTRL_DP_CLK_SEL_MASK) {
+	case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK:
+	case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_PCLK:
+		parent = 0;
+		break;
+
+	case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK:
+	case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_DPCLK:
+		parent = 1;
+		break;
+	}
+
+	return parent;
+}
+
+static const struct clk_ops tegra_clk_sor_brick_ops = {
+	.set_parent = tegra_clk_sor_brick_set_parent,
+	.get_parent = tegra_clk_sor_brick_get_parent,
+};
+
+static struct clk *tegra_clk_sor_brick_register(struct tegra_sor *sor,
+						const char *name)
+{
+	struct tegra_clk_sor_brick *brick;
+	struct clk_init_data init;
+	struct clk *clk;
+
+	brick = devm_kzalloc(sor->dev, sizeof(*brick), GFP_KERNEL);
+	if (!brick)
+		return ERR_PTR(-ENOMEM);
+
+	brick->sor = sor;
+
+	init.name = name;
+	init.flags = 0;
+	init.parent_names = tegra_clk_sor_brick_parents;
+	init.num_parents = ARRAY_SIZE(tegra_clk_sor_brick_parents);
+	init.ops = &tegra_clk_sor_brick_ops;
+
+	brick->hw.init = &init;
+
+	clk = devm_clk_register(sor->dev, &brick->hw);
+	if (IS_ERR(clk))
+		kfree(brick);
+
+	return clk;
+}
+
 static int tegra_sor_dp_train_fast(struct tegra_sor *sor,
 				   struct drm_dp_link *link)
 {
@@ -569,10 +699,10 @@ static int tegra_sor_compute_params(struct tegra_sor *sor,
 	return false;
 }
 
-static int tegra_sor_calc_config(struct tegra_sor *sor,
-				 const struct drm_display_mode *mode,
-				 struct tegra_sor_config *config,
-				 struct drm_dp_link *link)
+static int tegra_sor_compute_config(struct tegra_sor *sor,
+				    const struct drm_display_mode *mode,
+				    struct tegra_sor_config *config,
+				    struct drm_dp_link *link)
 {
 	const u64 f = 100000, link_rate = link->rate * 1000;
 	const u64 pclk = mode->clock * 1000;
@@ -661,6 +791,135 @@ static int tegra_sor_calc_config(struct tegra_sor *sor,
 	return 0;
 }
 
+static void tegra_sor_apply_config(struct tegra_sor *sor,
+				   const struct tegra_sor_config *config)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
+	value &= ~SOR_DP_LINKCTL_TU_SIZE_MASK;
+	value |= SOR_DP_LINKCTL_TU_SIZE(config->tu_size);
+	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
+
+	value = tegra_sor_readl(sor, SOR_DP_CONFIG0);
+	value &= ~SOR_DP_CONFIG_WATERMARK_MASK;
+	value |= SOR_DP_CONFIG_WATERMARK(config->watermark);
+
+	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK;
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_COUNT(config->active_count);
+
+	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK;
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_FRAC(config->active_frac);
+
+	if (config->active_polarity)
+		value |= SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
+	else
+		value &= ~SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
+
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_ENABLE;
+	value |= SOR_DP_CONFIG_DISPARITY_NEGATIVE;
+	tegra_sor_writel(sor, value, SOR_DP_CONFIG0);
+
+	value = tegra_sor_readl(sor, SOR_DP_AUDIO_HBLANK_SYMBOLS);
+	value &= ~SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK;
+	value |= config->hblank_symbols & 0xffff;
+	tegra_sor_writel(sor, value, SOR_DP_AUDIO_HBLANK_SYMBOLS);
+
+	value = tegra_sor_readl(sor, SOR_DP_AUDIO_VBLANK_SYMBOLS);
+	value &= ~SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK;
+	value |= config->vblank_symbols & 0xffff;
+	tegra_sor_writel(sor, value, SOR_DP_AUDIO_VBLANK_SYMBOLS);
+}
+
+static void tegra_sor_mode_set(struct tegra_sor *sor,
+			       const struct drm_display_mode *mode,
+			       struct tegra_sor_state *state)
+{
+	struct tegra_dc *dc = to_tegra_dc(sor->output.encoder.crtc);
+	unsigned int vbe, vse, hbe, hse, vbs, hbs;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PIXELDEPTH_MASK;
+	value &= ~SOR_STATE_ASY_CRC_MODE_MASK;
+	value &= ~SOR_STATE_ASY_OWNER_MASK;
+
+	value |= SOR_STATE_ASY_CRC_MODE_COMPLETE |
+		 SOR_STATE_ASY_OWNER(dc->pipe + 1);
+
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		value &= ~SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		value &= ~SOR_STATE_ASY_VSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL;
+
+	switch (state->bpc) {
+	case 16:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_48_444;
+		break;
+
+	case 12:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_36_444;
+		break;
+
+	case 10:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_30_444;
+		break;
+
+	case 8:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
+		break;
+
+	case 6:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444;
+		break;
+
+	default:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
+		break;
+	}
+
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	/*
+	 * TODO: The video timing programming below doesn't seem to match the
+	 * register definitions.
+	 */
+
+	value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff);
+	tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe));
+
+	/* sync end = sync width - 1 */
+	vse = mode->vsync_end - mode->vsync_start - 1;
+	hse = mode->hsync_end - mode->hsync_start - 1;
+
+	value = ((vse & 0x7fff) << 16) | (hse & 0x7fff);
+	tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe));
+
+	/* blank end = sync end + back porch */
+	vbe = vse + (mode->vtotal - mode->vsync_end);
+	hbe = hse + (mode->htotal - mode->hsync_end);
+
+	value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff);
+	tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe));
+
+	/* blank start = blank end + active */
+	vbs = vbe + mode->vdisplay;
+	hbs = hbe + mode->hdisplay;
+
+	value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff);
+	tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe));
+
+	/* XXX interlacing support */
+	tegra_sor_writel(sor, 0x001, SOR_HEAD_STATE5(dc->pipe));
+}
+
 static int tegra_sor_detach(struct tegra_sor *sor)
 {
 	unsigned long value, timeout;
@@ -733,7 +992,8 @@ static int tegra_sor_power_down(struct tegra_sor *sor)
 	if ((value & SOR_PWR_TRIGGER) != 0)
 		return -ETIMEDOUT;
 
-	err = clk_set_parent(sor->clk, sor->clk_safe);
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
 	if (err < 0)
 		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
 
@@ -1038,6 +1298,22 @@ static void tegra_sor_debugfs_exit(struct tegra_sor *sor)
 	sor->debugfs = NULL;
 }
 
+static void tegra_sor_connector_reset(struct drm_connector *connector)
+{
+	struct tegra_sor_state *state;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return;
+
+	if (connector->state) {
+		__drm_atomic_helper_connector_destroy_state(connector->state);
+		kfree(connector->state);
+	}
+
+	__drm_atomic_helper_connector_reset(connector, &state->base);
+}
+
 static enum drm_connector_status
 tegra_sor_connector_detect(struct drm_connector *connector, bool force)
 {
@@ -1050,13 +1326,28 @@ tegra_sor_connector_detect(struct drm_connector *connector, bool force)
 	return tegra_output_connector_detect(connector, force);
 }
 
+static struct drm_connector_state *
+tegra_sor_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct tegra_sor_state *state = to_sor_state(connector->state);
+	struct tegra_sor_state *copy;
+
+	copy = kmemdup(state, sizeof(*state), GFP_KERNEL);
+	if (!copy)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector, &copy->base);
+
+	return &copy->base;
+}
+
 static const struct drm_connector_funcs tegra_sor_connector_funcs = {
 	.dpms = drm_atomic_helper_connector_dpms,
-	.reset = drm_atomic_helper_connector_reset,
+	.reset = tegra_sor_connector_reset,
 	.detect = tegra_sor_connector_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.destroy = tegra_output_connector_destroy,
-	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_duplicate_state = tegra_sor_connector_duplicate_state,
 	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
 };
 
@@ -1081,6 +1372,10 @@ static enum drm_mode_status
 tegra_sor_connector_mode_valid(struct drm_connector *connector,
 			       struct drm_display_mode *mode)
 {
+	/* HDMI 2.0 modes are not yet supported */
+	if (mode->clock > 340000)
+		return MODE_NOCLOCK;
+
 	return MODE_OK;
 }
 
@@ -1140,8 +1435,7 @@ static void tegra_sor_edp_disable(struct drm_encoder *encoder)
 	if (output->panel)
 		drm_panel_unprepare(output->panel);
 
-	reset_control_assert(sor->rst);
-	clk_disable_unprepare(sor->clk);
+	pm_runtime_put(sor->dev);
 }
 
 #if 0
@@ -1191,19 +1485,18 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
 	struct tegra_output *output = encoder_to_output(encoder);
 	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
-	unsigned int vbe, vse, hbe, hse, vbs, hbs, i;
 	struct tegra_sor *sor = to_sor(output);
 	struct tegra_sor_config config;
+	struct tegra_sor_state *state;
 	struct drm_dp_link link;
 	u8 rate, lanes;
+	unsigned int i;
 	int err = 0;
 	u32 value;
 
-	err = clk_prepare_enable(sor->clk);
-	if (err < 0)
-		dev_err(sor->dev, "failed to enable clock: %d\n", err);
+	state = to_sor_state(output->connector.state);
 
-	reset_control_deassert(sor->rst);
+	pm_runtime_get_sync(sor->dev);
 
 	if (output->panel)
 		drm_panel_prepare(output->panel);
@@ -1218,17 +1511,17 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 		return;
 	}
 
-	err = clk_set_parent(sor->clk, sor->clk_safe);
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
 	if (err < 0)
 		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
 
 	memset(&config, 0, sizeof(config));
-	config.bits_per_pixel = output->connector.display_info.bpc * 3;
+	config.bits_per_pixel = state->bpc * 3;
 
-	err = tegra_sor_calc_config(sor, mode, &config, &link);
+	err = tegra_sor_compute_config(sor, mode, &config, &link);
 	if (err < 0)
-		dev_err(sor->dev, "failed to compute link configuration: %d\n",
-			err);
+		dev_err(sor->dev, "failed to compute configuration: %d\n", err);
 
 	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
 	value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK;
@@ -1325,10 +1618,18 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 	value &= ~SOR_PLL2_PORT_POWERDOWN;
 	tegra_sor_writel(sor, value, SOR_PLL2);
 
-	/* switch to DP clock */
-	err = clk_set_parent(sor->clk, sor->clk_dp);
+	/* XXX not in TRM */
+	for (value = 0, i = 0; i < 5; i++)
+		value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) |
+			 SOR_XBAR_CTRL_LINK1_XSEL(i, i);
+
+	tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL);
+	tegra_sor_writel(sor, value, SOR_XBAR_CTRL);
+
+	/* switch to DP parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_dp);
 	if (err < 0)
-		dev_err(sor->dev, "failed to set DP parent clock: %d\n", err);
+		dev_err(sor->dev, "failed to set parent clock: %d\n", err);
 
 	/* power DP lanes */
 	value = tegra_sor_readl(sor, SOR_DP_PADCTL0);
@@ -1374,13 +1675,11 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 	value |= drm_dp_link_rate_to_bw_code(link.rate) << 2;
 	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
 
-	/* set linkctl */
+	tegra_sor_apply_config(sor, &config);
+
+	/* enable link */
 	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
 	value |= SOR_DP_LINKCTL_ENABLE;
-
-	value &= ~SOR_DP_LINKCTL_TU_SIZE_MASK;
-	value |= SOR_DP_LINKCTL_TU_SIZE(config.tu_size);
-
 	value |= SOR_DP_LINKCTL_ENHANCED_FRAME;
 	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
 
@@ -1393,35 +1692,6 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 
 	tegra_sor_writel(sor, value, SOR_DP_TPG);
 
-	value = tegra_sor_readl(sor, SOR_DP_CONFIG0);
-	value &= ~SOR_DP_CONFIG_WATERMARK_MASK;
-	value |= SOR_DP_CONFIG_WATERMARK(config.watermark);
-
-	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK;
-	value |= SOR_DP_CONFIG_ACTIVE_SYM_COUNT(config.active_count);
-
-	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK;
-	value |= SOR_DP_CONFIG_ACTIVE_SYM_FRAC(config.active_frac);
-
-	if (config.active_polarity)
-		value |= SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
-	else
-		value &= ~SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
-
-	value |= SOR_DP_CONFIG_ACTIVE_SYM_ENABLE;
-	value |= SOR_DP_CONFIG_DISPARITY_NEGATIVE;
-	tegra_sor_writel(sor, value, SOR_DP_CONFIG0);
-
-	value = tegra_sor_readl(sor, SOR_DP_AUDIO_HBLANK_SYMBOLS);
-	value &= ~SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK;
-	value |= config.hblank_symbols & 0xffff;
-	tegra_sor_writel(sor, value, SOR_DP_AUDIO_HBLANK_SYMBOLS);
-
-	value = tegra_sor_readl(sor, SOR_DP_AUDIO_VBLANK_SYMBOLS);
-	value &= ~SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK;
-	value |= config.vblank_symbols & 0xffff;
-	tegra_sor_writel(sor, value, SOR_DP_AUDIO_VBLANK_SYMBOLS);
-
 	/* enable pad calibration logic */
 	value = tegra_sor_readl(sor, SOR_DP_PADCTL0);
 	value |= SOR_DP_PADCTL_PAD_CAL_PD;
@@ -1477,75 +1747,19 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder)
 	if (err < 0)
 		dev_err(sor->dev, "failed to power up SOR: %d\n", err);
 
-	/*
-	 * configure panel (24bpp, vsync-, hsync-, DP-A protocol, complete
-	 * raster, associate with display controller)
-	 */
-	value = SOR_STATE_ASY_PROTOCOL_DP_A |
-		SOR_STATE_ASY_CRC_MODE_COMPLETE |
-		SOR_STATE_ASY_OWNER(dc->pipe + 1);
-
-	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
-		value &= ~SOR_STATE_ASY_HSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		value |= SOR_STATE_ASY_HSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
-		value &= ~SOR_STATE_ASY_VSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		value |= SOR_STATE_ASY_VSYNCPOL;
-
-	switch (config.bits_per_pixel) {
-	case 24:
-		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
-		break;
-
-	case 18:
-		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444;
-		break;
-
-	default:
-		BUG();
-		break;
-	}
-
-	tegra_sor_writel(sor, value, SOR_STATE1);
-
-	/*
-	 * TODO: The video timing programming below doesn't seem to match the
-	 * register definitions.
-	 */
-
-	value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe));
-
-	vse = mode->vsync_end - mode->vsync_start - 1;
-	hse = mode->hsync_end - mode->hsync_start - 1;
-
-	value = ((vse & 0x7fff) << 16) | (hse & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe));
-
-	vbe = vse + (mode->vsync_start - mode->vdisplay);
-	hbe = hse + (mode->hsync_start - mode->hdisplay);
-
-	value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe));
-
-	vbs = vbe + mode->vdisplay;
-	hbs = hbe + mode->hdisplay;
-
-	value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe));
-
-	tegra_sor_writel(sor, 0x1, SOR_HEAD_STATE5(dc->pipe));
-
 	/* CSTM (LVDS, link A/B, upper) */
 	value = SOR_CSTM_LVDS | SOR_CSTM_LINK_ACT_A | SOR_CSTM_LINK_ACT_B |
 		SOR_CSTM_UPPER;
 	tegra_sor_writel(sor, value, SOR_CSTM);
 
+	/* use DP-A protocol */
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PROTOCOL_MASK;
+	value |= SOR_STATE_ASY_PROTOCOL_DP_A;
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	tegra_sor_mode_set(sor, mode, state);
+
 	/* PWM setup */
 	err = tegra_sor_setup_pwm(sor, 250);
 	if (err < 0)
@@ -1577,11 +1791,15 @@ tegra_sor_encoder_atomic_check(struct drm_encoder *encoder,
 			       struct drm_connector_state *conn_state)
 {
 	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_sor_state *state = to_sor_state(conn_state);
 	struct tegra_dc *dc = to_tegra_dc(conn_state->crtc);
 	unsigned long pclk = crtc_state->mode.clock * 1000;
 	struct tegra_sor *sor = to_sor(output);
+	struct drm_display_info *info;
 	int err;
 
+	info = &output->connector.display_info;
+
 	err = tegra_dc_state_setup_clock(dc, crtc_state, sor->clk_parent,
 					 pclk, 0);
 	if (err < 0) {
@@ -1589,6 +1807,18 @@ tegra_sor_encoder_atomic_check(struct drm_encoder *encoder,
 		return err;
 	}
 
+	switch (info->bpc) {
+	case 8:
+	case 6:
+		state->bpc = info->bpc;
+		break;
+
+	default:
+		DRM_DEBUG_KMS("%u bits-per-color not supported\n", info->bpc);
+		state->bpc = 8;
+		break;
+	}
+
 	return 0;
 }
 
@@ -1751,9 +1981,7 @@ static void tegra_sor_hdmi_disable(struct drm_encoder *encoder)
 	if (err < 0)
 		dev_err(sor->dev, "failed to power off HDMI rail: %d\n", err);
 
-	reset_control_assert(sor->rst);
-	usleep_range(1000, 2000);
-	clk_disable_unprepare(sor->clk);
+	pm_runtime_put(sor->dev);
 }
 
 static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
@@ -1761,26 +1989,21 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
 	struct tegra_output *output = encoder_to_output(encoder);
 	unsigned int h_ref_to_sync = 1, pulse_start, max_ac;
 	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
-	unsigned int vbe, vse, hbe, hse, vbs, hbs, div;
 	struct tegra_sor_hdmi_settings *settings;
 	struct tegra_sor *sor = to_sor(output);
+	struct tegra_sor_state *state;
 	struct drm_display_mode *mode;
-	struct drm_display_info *info;
+	unsigned int div, i;
 	u32 value;
 	int err;
 
+	state = to_sor_state(output->connector.state);
 	mode = &encoder->crtc->state->adjusted_mode;
-	info = &output->connector.display_info;
 
-	err = clk_prepare_enable(sor->clk);
-	if (err < 0)
-		dev_err(sor->dev, "failed to enable clock: %d\n", err);
+	pm_runtime_get_sync(sor->dev);
 
-	usleep_range(1000, 2000);
-
-	reset_control_deassert(sor->rst);
-
-	err = clk_set_parent(sor->clk, sor->clk_safe);
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
 	if (err < 0)
 		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
 
@@ -1876,22 +2099,20 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
 	value = SOR_REFCLK_DIV_INT(div) | SOR_REFCLK_DIV_FRAC(div);
 	tegra_sor_writel(sor, value, SOR_REFCLK);
 
-	/* XXX don't hardcode */
-	value = SOR_XBAR_CTRL_LINK1_XSEL(4, 4) |
-		SOR_XBAR_CTRL_LINK1_XSEL(3, 3) |
-		SOR_XBAR_CTRL_LINK1_XSEL(2, 2) |
-		SOR_XBAR_CTRL_LINK1_XSEL(1, 1) |
-		SOR_XBAR_CTRL_LINK1_XSEL(0, 0) |
-		SOR_XBAR_CTRL_LINK0_XSEL(4, 4) |
-		SOR_XBAR_CTRL_LINK0_XSEL(3, 3) |
-		SOR_XBAR_CTRL_LINK0_XSEL(2, 0) |
-		SOR_XBAR_CTRL_LINK0_XSEL(1, 1) |
-		SOR_XBAR_CTRL_LINK0_XSEL(0, 2);
-	tegra_sor_writel(sor, value, SOR_XBAR_CTRL);
+	/* XXX not in TRM */
+	for (value = 0, i = 0; i < 5; i++)
+		value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) |
+			 SOR_XBAR_CTRL_LINK1_XSEL(i, i);
 
 	tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL);
+	tegra_sor_writel(sor, value, SOR_XBAR_CTRL);
 
-	err = clk_set_parent(sor->clk, sor->clk_parent);
+	/* switch to parent clock */
+	err = clk_set_parent(sor->clk_src, sor->clk_parent);
+	if (err < 0)
+		dev_err(sor->dev, "failed to set source clock: %d\n", err);
+
+	err = tegra_sor_set_parent_clock(sor, sor->clk_src);
 	if (err < 0)
 		dev_err(sor->dev, "failed to set parent clock: %d\n", err);
 
@@ -2001,7 +2222,7 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
 	value &= ~DITHER_CONTROL_MASK;
 	value &= ~BASE_COLOR_SIZE_MASK;
 
-	switch (info->bpc) {
+	switch (state->bpc) {
 	case 6:
 		value |= BASE_COLOR_SIZE_666;
 		break;
@@ -2011,7 +2232,8 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
 		break;
 
 	default:
-		WARN(1, "%u bits-per-color not supported\n", info->bpc);
+		WARN(1, "%u bits-per-color not supported\n", state->bpc);
+		value |= BASE_COLOR_SIZE_888;
 		break;
 	}
 
@@ -2021,83 +2243,19 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
 	if (err < 0)
 		dev_err(sor->dev, "failed to power up SOR: %d\n", err);
 
-	/* configure mode */
-	value = tegra_sor_readl(sor, SOR_STATE1);
-	value &= ~SOR_STATE_ASY_PIXELDEPTH_MASK;
-	value &= ~SOR_STATE_ASY_CRC_MODE_MASK;
-	value &= ~SOR_STATE_ASY_OWNER_MASK;
-
-	value |= SOR_STATE_ASY_CRC_MODE_COMPLETE |
-		 SOR_STATE_ASY_OWNER(dc->pipe + 1);
-
-	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
-		value &= ~SOR_STATE_ASY_HSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		value |= SOR_STATE_ASY_HSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
-		value &= ~SOR_STATE_ASY_VSYNCPOL;
-
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		value |= SOR_STATE_ASY_VSYNCPOL;
-
-	switch (info->bpc) {
-	case 8:
-		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
-		break;
-
-	case 6:
-		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444;
-		break;
-
-	default:
-		BUG();
-		break;
-	}
-
-	tegra_sor_writel(sor, value, SOR_STATE1);
-
+	/* configure dynamic range of output */
 	value = tegra_sor_readl(sor, SOR_HEAD_STATE0(dc->pipe));
 	value &= ~SOR_HEAD_STATE_RANGECOMPRESS_MASK;
 	value &= ~SOR_HEAD_STATE_DYNRANGE_MASK;
 	tegra_sor_writel(sor, value, SOR_HEAD_STATE0(dc->pipe));
 
+	/* configure colorspace */
 	value = tegra_sor_readl(sor, SOR_HEAD_STATE0(dc->pipe));
 	value &= ~SOR_HEAD_STATE_COLORSPACE_MASK;
 	value |= SOR_HEAD_STATE_COLORSPACE_RGB;
 	tegra_sor_writel(sor, value, SOR_HEAD_STATE0(dc->pipe));
 
-	/*
-	 * TODO: The video timing programming below doesn't seem to match the
-	 * register definitions.
-	 */
-
-	value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe));
-
-	/* sync end = sync width - 1 */
-	vse = mode->vsync_end - mode->vsync_start - 1;
-	hse = mode->hsync_end - mode->hsync_start - 1;
-
-	value = ((vse & 0x7fff) << 16) | (hse & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe));
-
-	/* blank end = sync end + back porch */
-	vbe = vse + (mode->vtotal - mode->vsync_end);
-	hbe = hse + (mode->htotal - mode->hsync_end);
-
-	value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe));
-
-	/* blank start = blank end + active */
-	vbs = vbe + mode->vdisplay;
-	hbs = hbe + mode->hdisplay;
-
-	value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff);
-	tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe));
-
-	tegra_sor_writel(sor, 0x1, SOR_HEAD_STATE5(dc->pipe));
+	tegra_sor_mode_set(sor, mode, state);
 
 	tegra_sor_update(sor);
 
@@ -2195,10 +2353,13 @@ static int tegra_sor_init(struct host1x_client *client)
 	 * XXX: Remove this reset once proper hand-over from firmware to
 	 * kernel is possible.
 	 */
-	err = reset_control_assert(sor->rst);
-	if (err < 0) {
-		dev_err(sor->dev, "failed to assert SOR reset: %d\n", err);
-		return err;
+	if (sor->rst) {
+		err = reset_control_assert(sor->rst);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to assert SOR reset: %d\n",
+				err);
+			return err;
+		}
 	}
 
 	err = clk_prepare_enable(sor->clk);
@@ -2209,10 +2370,13 @@ static int tegra_sor_init(struct host1x_client *client)
 
 	usleep_range(1000, 3000);
 
-	err = reset_control_deassert(sor->rst);
-	if (err < 0) {
-		dev_err(sor->dev, "failed to deassert SOR reset: %d\n", err);
-		return err;
+	if (sor->rst) {
+		err = reset_control_deassert(sor->rst);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to deassert SOR reset: %d\n",
+				err);
+			return err;
+		}
 	}
 
 	err = clk_prepare_enable(sor->clk_safe);
@@ -2323,11 +2487,16 @@ static const struct tegra_sor_ops tegra_sor_hdmi_ops = {
 	.remove = tegra_sor_hdmi_remove,
 };
 
+static const u8 tegra124_sor_xbar_cfg[5] = {
+	0, 1, 2, 3, 4
+};
+
 static const struct tegra_sor_soc tegra124_sor = {
 	.supports_edp = true,
 	.supports_lvds = true,
 	.supports_hdmi = false,
 	.supports_dp = false,
+	.xbar_cfg = tegra124_sor_xbar_cfg,
 };
 
 static const struct tegra_sor_soc tegra210_sor = {
@@ -2335,6 +2504,11 @@ static const struct tegra_sor_soc tegra210_sor = {
 	.supports_lvds = false,
 	.supports_hdmi = false,
 	.supports_dp = false,
+	.xbar_cfg = tegra124_sor_xbar_cfg,
+};
+
+static const u8 tegra210_sor_xbar_cfg[5] = {
+	2, 1, 0, 3, 4
 };
 
 static const struct tegra_sor_soc tegra210_sor1 = {
@@ -2345,6 +2519,8 @@ static const struct tegra_sor_soc tegra210_sor1 = {
 
 	.num_settings = ARRAY_SIZE(tegra210_sor_hdmi_defaults),
 	.settings = tegra210_sor_hdmi_defaults,
+
+	.xbar_cfg = tegra210_sor_xbar_cfg,
 };
 
 static const struct of_device_id tegra_sor_of_match[] = {
@@ -2434,11 +2610,14 @@ static int tegra_sor_probe(struct platform_device *pdev)
 		goto remove;
 	}
 
-	sor->rst = devm_reset_control_get(&pdev->dev, "sor");
-	if (IS_ERR(sor->rst)) {
-		err = PTR_ERR(sor->rst);
-		dev_err(&pdev->dev, "failed to get reset control: %d\n", err);
-		goto remove;
+	if (!pdev->dev.pm_domain) {
+		sor->rst = devm_reset_control_get(&pdev->dev, "sor");
+		if (IS_ERR(sor->rst)) {
+			err = PTR_ERR(sor->rst);
+			dev_err(&pdev->dev, "failed to get reset control: %d\n",
+				err);
+			goto remove;
+		}
 	}
 
 	sor->clk = devm_clk_get(&pdev->dev, NULL);
@@ -2448,6 +2627,16 @@ static int tegra_sor_probe(struct platform_device *pdev)
 		goto remove;
 	}
 
+	if (sor->soc->supports_hdmi || sor->soc->supports_dp) {
+		sor->clk_src = devm_clk_get(&pdev->dev, "source");
+		if (IS_ERR(sor->clk_src)) {
+			err = PTR_ERR(sor->clk_src);
+			dev_err(sor->dev, "failed to get source clock: %d\n",
+				err);
+			goto remove;
+		}
+	}
+
 	sor->clk_parent = devm_clk_get(&pdev->dev, "parent");
 	if (IS_ERR(sor->clk_parent)) {
 		err = PTR_ERR(sor->clk_parent);
@@ -2469,6 +2658,19 @@ static int tegra_sor_probe(struct platform_device *pdev)
 		goto remove;
 	}
 
+	platform_set_drvdata(pdev, sor);
+	pm_runtime_enable(&pdev->dev);
+
+	pm_runtime_get_sync(&pdev->dev);
+	sor->clk_brick = tegra_clk_sor_brick_register(sor, "sor1_brick");
+	pm_runtime_put(&pdev->dev);
+
+	if (IS_ERR(sor->clk_brick)) {
+		err = PTR_ERR(sor->clk_brick);
+		dev_err(&pdev->dev, "failed to register SOR clock: %d\n", err);
+		goto remove;
+	}
+
 	INIT_LIST_HEAD(&sor->client.list);
 	sor->client.ops = &sor_client_ops;
 	sor->client.dev = &pdev->dev;
@@ -2480,8 +2682,6 @@ static int tegra_sor_probe(struct platform_device *pdev)
 		goto remove;
 	}
 
-	platform_set_drvdata(pdev, sor);
-
 	return 0;
 
 remove:
@@ -2497,6 +2697,8 @@ static int tegra_sor_remove(struct platform_device *pdev)
 	struct tegra_sor *sor = platform_get_drvdata(pdev);
 	int err;
 
+	pm_runtime_disable(&pdev->dev);
+
 	err = host1x_client_unregister(&sor->client);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to unregister host1x client: %d\n",
@@ -2515,10 +2717,62 @@ static int tegra_sor_remove(struct platform_device *pdev)
 	return 0;
 }
 
+#ifdef CONFIG_PM
+static int tegra_sor_suspend(struct device *dev)
+{
+	struct tegra_sor *sor = dev_get_drvdata(dev);
+	int err;
+
+	if (sor->rst) {
+		err = reset_control_assert(sor->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to assert reset: %d\n", err);
+			return err;
+		}
+	}
+
+	usleep_range(1000, 2000);
+
+	clk_disable_unprepare(sor->clk);
+
+	return 0;
+}
+
+static int tegra_sor_resume(struct device *dev)
+{
+	struct tegra_sor *sor = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_prepare_enable(sor->clk);
+	if (err < 0) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		return err;
+	}
+
+	usleep_range(1000, 2000);
+
+	if (sor->rst) {
+		err = reset_control_deassert(sor->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to deassert reset: %d\n", err);
+			clk_disable_unprepare(sor->clk);
+			return err;
+		}
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops tegra_sor_pm_ops = {
+	SET_RUNTIME_PM_OPS(tegra_sor_suspend, tegra_sor_resume, NULL)
+};
+
 struct platform_driver tegra_sor_driver = {
 	.driver = {
 		.name = "tegra-sor",
 		.of_match_table = tegra_sor_of_match,
+		.pm = &tegra_sor_pm_ops,
 	},
 	.probe = tegra_sor_probe,
 	.remove = tegra_sor_remove,
diff --git a/drivers/gpu/drm/tegra/sor.h b/drivers/gpu/drm/tegra/sor.h
index 2d31d027e3f6..865c73b48968 100644
--- a/drivers/gpu/drm/tegra/sor.h
+++ b/drivers/gpu/drm/tegra/sor.h
@@ -27,6 +27,9 @@
 #define  SOR_STATE_ASY_PIXELDEPTH_MASK		(0xf << 17)
 #define  SOR_STATE_ASY_PIXELDEPTH_BPP_18_444	(0x2 << 17)
 #define  SOR_STATE_ASY_PIXELDEPTH_BPP_24_444	(0x5 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_30_444	(0x6 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_36_444	(0x8 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_48_444	(0x9 << 17)
 #define  SOR_STATE_ASY_VSYNCPOL			(1 << 13)
 #define  SOR_STATE_ASY_HSYNCPOL			(1 << 12)
 #define  SOR_STATE_ASY_PROTOCOL_MASK		(0xf << 8)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 4e55863e836b..b1c1d8f42c8c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -148,6 +148,7 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	BUG_ON(!list_empty(&bo->ddestroy));
 	ttm_tt_destroy(bo->ttm);
 	atomic_dec(&bo->glob->bo_count);
+	fence_put(bo->moving);
 	if (bo->resv == &bo->ttm_resv)
 		reservation_object_fini(&bo->ttm_resv);
 	mutex_destroy(&bo->wu_mutex);
@@ -358,7 +359,8 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 		ret = bdev->driver->move(bo, evict, interruptible,
 					 no_wait_gpu, mem);
 	else
-		ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, mem);
+		ret = ttm_bo_move_memcpy(bo, evict, interruptible,
+					 no_wait_gpu, mem);
 
 	if (ret) {
 		if (bdev->driver->move_notify) {
@@ -394,8 +396,7 @@ moved:
 
 out_err:
 	new_man = &bdev->man[bo->mem.mem_type];
-	if ((new_man->flags & TTM_MEMTYPE_FLAG_FIXED) && bo->ttm) {
-		ttm_tt_unbind(bo->ttm);
+	if (new_man->flags & TTM_MEMTYPE_FLAG_FIXED) {
 		ttm_tt_destroy(bo->ttm);
 		bo->ttm = NULL;
 	}
@@ -416,11 +417,8 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	if (bo->bdev->driver->move_notify)
 		bo->bdev->driver->move_notify(bo, NULL);
 
-	if (bo->ttm) {
-		ttm_tt_unbind(bo->ttm);
-		ttm_tt_destroy(bo->ttm);
-		bo->ttm = NULL;
-	}
+	ttm_tt_destroy(bo->ttm);
+	bo->ttm = NULL;
 	ttm_bo_mem_put(bo, &bo->mem);
 
 	ww_mutex_unlock (&bo->resv->lock);
@@ -686,15 +684,6 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	struct ttm_placement placement;
 	int ret = 0;
 
-	ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
-
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS) {
-			pr_err("Failed to expire sync object before buffer eviction\n");
-		}
-		goto out;
-	}
-
 	lockdep_assert_held(&bo->resv->lock.base);
 
 	evict_mem = bo->mem;
@@ -718,7 +707,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 
 	ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, interruptible,
 				     no_wait_gpu);
-	if (ret) {
+	if (unlikely(ret)) {
 		if (ret != -ERESTARTSYS)
 			pr_err("Buffer eviction failed\n");
 		ttm_bo_mem_put(bo, &evict_mem);
@@ -798,6 +787,34 @@ void ttm_bo_mem_put(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem)
 EXPORT_SYMBOL(ttm_bo_mem_put);
 
 /**
+ * Add the last move fence to the BO and reserve a new shared slot.
+ */
+static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo,
+				 struct ttm_mem_type_manager *man,
+				 struct ttm_mem_reg *mem)
+{
+	struct fence *fence;
+	int ret;
+
+	spin_lock(&man->move_lock);
+	fence = fence_get(man->move);
+	spin_unlock(&man->move_lock);
+
+	if (fence) {
+		reservation_object_add_shared_fence(bo->resv, fence);
+
+		ret = reservation_object_reserve_shared(bo->resv);
+		if (unlikely(ret))
+			return ret;
+
+		fence_put(bo->moving);
+		bo->moving = fence;
+	}
+
+	return 0;
+}
+
+/**
  * Repeatedly evict memory from the LRU for @mem_type until we create enough
  * space, or we've evicted everything and there isn't enough space.
  */
@@ -823,10 +840,8 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 		if (unlikely(ret != 0))
 			return ret;
 	} while (1);
-	if (mem->mm_node == NULL)
-		return -ENOMEM;
 	mem->mem_type = mem_type;
-	return 0;
+	return ttm_bo_add_move_fence(bo, man, mem);
 }
 
 static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man,
@@ -896,6 +911,10 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 	bool has_erestartsys = false;
 	int i, ret;
 
+	ret = reservation_object_reserve_shared(bo->resv);
+	if (unlikely(ret))
+		return ret;
+
 	mem->mm_node = NULL;
 	for (i = 0; i < placement->num_placement; ++i) {
 		const struct ttm_place *place = &placement->placement[i];
@@ -929,9 +948,15 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		ret = (*man->func->get_node)(man, bo, place, mem);
 		if (unlikely(ret))
 			return ret;
-		
-		if (mem->mm_node)
+
+		if (mem->mm_node) {
+			ret = ttm_bo_add_move_fence(bo, man, mem);
+			if (unlikely(ret)) {
+				(*man->func->put_node)(man, mem);
+				return ret;
+			}
 			break;
+		}
 	}
 
 	if ((type_ok && (mem_type == TTM_PL_SYSTEM)) || mem->mm_node) {
@@ -998,20 +1023,6 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 
 	lockdep_assert_held(&bo->resv->lock.base);
 
-	/*
-	 * Don't wait for the BO on initial allocation. This is important when
-	 * the BO has an imported reservation object.
-	 */
-	if (bo->mem.mem_type != TTM_PL_SYSTEM || bo->ttm != NULL) {
-		/*
-		 * FIXME: It's possible to pipeline buffer moves.
-		 * Have the driver move function wait for idle when necessary,
-		 * instead of doing it here.
-		 */
-		ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
-		if (ret)
-			return ret;
-	}
 	mem.num_pages = bo->num_pages;
 	mem.size = mem.num_pages << PAGE_SHIFT;
 	mem.page_alignment = bo->mem.page_alignment;
@@ -1163,7 +1174,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->mem.page_alignment = page_alignment;
 	bo->mem.bus.io_reserved_vm = false;
 	bo->mem.bus.io_reserved_count = 0;
-	bo->priv_flags = 0;
+	bo->moving = NULL;
 	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
 	bo->persistent_swap_storage = persistent_swap_storage;
 	bo->acc_size = acc_size;
@@ -1275,6 +1286,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 {
 	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
 	struct ttm_bo_global *glob = bdev->glob;
+	struct fence *fence;
 	int ret;
 
 	/*
@@ -1295,6 +1307,23 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 		spin_lock(&glob->lru_lock);
 	}
 	spin_unlock(&glob->lru_lock);
+
+	spin_lock(&man->move_lock);
+	fence = fence_get(man->move);
+	spin_unlock(&man->move_lock);
+
+	if (fence) {
+		ret = fence_wait(fence, false);
+		fence_put(fence);
+		if (ret) {
+			if (allow_errors) {
+				return ret;
+			} else {
+				pr_err("Cleanup eviction failed\n");
+			}
+		}
+	}
+
 	return 0;
 }
 
@@ -1314,6 +1343,7 @@ int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type)
 		       mem_type);
 		return ret;
 	}
+	fence_put(man->move);
 
 	man->use_type = false;
 	man->has_type = false;
@@ -1359,6 +1389,7 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
 	man->io_reserve_fastpath = true;
 	man->use_io_reserve_lru = false;
 	mutex_init(&man->io_reserve_mutex);
+	spin_lock_init(&man->move_lock);
 	INIT_LIST_HEAD(&man->io_reserve_lru);
 
 	ret = bdev->driver->init_mem_type(bdev, type, man);
@@ -1377,6 +1408,7 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
 	man->size = p_size;
 
 	INIT_LIST_HEAD(&man->lru);
+	man->move = NULL;
 
 	return 0;
 }
@@ -1570,47 +1602,17 @@ EXPORT_SYMBOL(ttm_bo_unmap_virtual);
 int ttm_bo_wait(struct ttm_buffer_object *bo,
 		bool interruptible, bool no_wait)
 {
-	struct reservation_object_list *fobj;
-	struct reservation_object *resv;
-	struct fence *excl;
-	long timeout = 15 * HZ;
-	int i;
-
-	resv = bo->resv;
-	fobj = reservation_object_get_list(resv);
-	excl = reservation_object_get_excl(resv);
-	if (excl) {
-		if (!fence_is_signaled(excl)) {
-			if (no_wait)
-				return -EBUSY;
-
-			timeout = fence_wait_timeout(excl,
-						     interruptible, timeout);
-		}
-	}
-
-	for (i = 0; fobj && timeout > 0 && i < fobj->shared_count; ++i) {
-		struct fence *fence;
-		fence = rcu_dereference_protected(fobj->shared[i],
-						reservation_object_held(resv));
-
-		if (!fence_is_signaled(fence)) {
-			if (no_wait)
-				return -EBUSY;
-
-			timeout = fence_wait_timeout(fence,
-						     interruptible, timeout);
-		}
-	}
+	long timeout = no_wait ? 0 : 15 * HZ;
 
+	timeout = reservation_object_wait_timeout_rcu(bo->resv, true,
+						      interruptible, timeout);
 	if (timeout < 0)
 		return timeout;
 
 	if (timeout == 0)
 		return -EBUSY;
 
-	reservation_object_add_excl_fence(resv, NULL);
-	clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
+	reservation_object_add_excl_fence(bo->resv, NULL);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_bo_wait);
@@ -1680,14 +1682,9 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 	ttm_bo_list_ref_sub(bo, put_count, true);
 
 	/**
-	 * Wait for GPU, then move to system cached.
+	 * Move to system cached
 	 */
 
-	ret = ttm_bo_wait(bo, false, false);
-
-	if (unlikely(ret != 0))
-		goto out;
-
 	if ((bo->mem.placement & swap_placement) != swap_placement) {
 		struct ttm_mem_reg evict_mem;
 
@@ -1702,6 +1699,14 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 			goto out;
 	}
 
+	/**
+	 * Make sure BO is idle.
+	 */
+
+	ret = ttm_bo_wait(bo, false, false);
+	if (unlikely(ret != 0))
+		goto out;
+
 	ttm_bo_unmap_virtual(bo);
 
 	/**
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index d9831559706e..4da0e784f9e7 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -53,7 +53,6 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 	int ret;
 
 	if (old_mem->mem_type != TTM_PL_SYSTEM) {
-		ttm_tt_unbind(ttm);
 		ttm_bo_free_old_node(bo);
 		ttm_flag_masked(&old_mem->placement, TTM_PL_FLAG_SYSTEM,
 				TTM_PL_MASK_MEM);
@@ -321,7 +320,8 @@ static int ttm_copy_ttm_io_page(struct ttm_tt *ttm, void *dst,
 }
 
 int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
-		       bool evict, bool no_wait_gpu,
+		       bool evict, bool interruptible,
+		       bool no_wait_gpu,
 		       struct ttm_mem_reg *new_mem)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
@@ -337,6 +337,10 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	unsigned long add = 0;
 	int dir;
 
+	ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+	if (ret)
+		return ret;
+
 	ret = ttm_mem_reg_ioremap(bdev, old_mem, &old_iomap);
 	if (ret)
 		return ret;
@@ -401,8 +405,7 @@ out2:
 	*old_mem = *new_mem;
 	new_mem->mm_node = NULL;
 
-	if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) && (ttm != NULL)) {
-		ttm_tt_unbind(ttm);
+	if (man->flags & TTM_MEMTYPE_FLAG_FIXED) {
 		ttm_tt_destroy(ttm);
 		bo->ttm = NULL;
 	}
@@ -462,6 +465,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	INIT_LIST_HEAD(&fbo->lru);
 	INIT_LIST_HEAD(&fbo->swap);
 	INIT_LIST_HEAD(&fbo->io_reserve_lru);
+	fbo->moving = NULL;
 	drm_vma_node_reset(&fbo->vma_node);
 	atomic_set(&fbo->cpu_writers, 0);
 
@@ -634,7 +638,6 @@ EXPORT_SYMBOL(ttm_bo_kunmap);
 int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 			      struct fence *fence,
 			      bool evict,
-			      bool no_wait_gpu,
 			      struct ttm_mem_reg *new_mem)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
@@ -649,9 +652,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 		if (ret)
 			return ret;
 
-		if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) &&
-		    (bo->ttm != NULL)) {
-			ttm_tt_unbind(bo->ttm);
+		if (man->flags & TTM_MEMTYPE_FLAG_FIXED) {
 			ttm_tt_destroy(bo->ttm);
 			bo->ttm = NULL;
 		}
@@ -665,7 +666,8 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 		 * operation has completed.
 		 */
 
-		set_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
+		fence_put(bo->moving);
+		bo->moving = fence_get(fence);
 
 		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
 		if (ret)
@@ -694,3 +696,95 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 	return 0;
 }
 EXPORT_SYMBOL(ttm_bo_move_accel_cleanup);
+
+int ttm_bo_pipeline_move(struct ttm_buffer_object *bo,
+			 struct fence *fence, bool evict,
+			 struct ttm_mem_reg *new_mem)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+	struct ttm_mem_reg *old_mem = &bo->mem;
+
+	struct ttm_mem_type_manager *from = &bdev->man[old_mem->mem_type];
+	struct ttm_mem_type_manager *to = &bdev->man[new_mem->mem_type];
+
+	int ret;
+
+	reservation_object_add_excl_fence(bo->resv, fence);
+
+	if (!evict) {
+		struct ttm_buffer_object *ghost_obj;
+
+		/**
+		 * This should help pipeline ordinary buffer moves.
+		 *
+		 * Hang old buffer memory on a new buffer object,
+		 * and leave it to be released when the GPU
+		 * operation has completed.
+		 */
+
+		fence_put(bo->moving);
+		bo->moving = fence_get(fence);
+
+		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
+		if (ret)
+			return ret;
+
+		reservation_object_add_excl_fence(ghost_obj->resv, fence);
+
+		/**
+		 * If we're not moving to fixed memory, the TTM object
+		 * needs to stay alive. Otherwhise hang it on the ghost
+		 * bo to be unbound and destroyed.
+		 */
+
+		if (!(to->flags & TTM_MEMTYPE_FLAG_FIXED))
+			ghost_obj->ttm = NULL;
+		else
+			bo->ttm = NULL;
+
+		ttm_bo_unreserve(ghost_obj);
+		ttm_bo_unref(&ghost_obj);
+
+	} else if (from->flags & TTM_MEMTYPE_FLAG_FIXED) {
+
+		/**
+		 * BO doesn't have a TTM we need to bind/unbind. Just remember
+		 * this eviction and free up the allocation
+		 */
+
+		spin_lock(&from->move_lock);
+		if (!from->move || fence_is_later(fence, from->move)) {
+			fence_put(from->move);
+			from->move = fence_get(fence);
+		}
+		spin_unlock(&from->move_lock);
+
+		ttm_bo_free_old_node(bo);
+
+		fence_put(bo->moving);
+		bo->moving = fence_get(fence);
+
+	} else {
+		/**
+		 * Last resort, wait for the move to be completed.
+		 *
+		 * Should never happen in pratice.
+		 */
+
+		ret = ttm_bo_wait(bo, false, false);
+		if (ret)
+			return ret;
+
+		if (to->flags & TTM_MEMTYPE_FLAG_FIXED) {
+			ttm_tt_destroy(bo->ttm);
+			bo->ttm = NULL;
+		}
+		ttm_bo_free_old_node(bo);
+	}
+
+	*old_mem = *new_mem;
+	new_mem->mm_node = NULL;
+
+	return 0;
+}
+EXPORT_SYMBOL(ttm_bo_pipeline_move);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 3216878bced3..a6ed9d5e5167 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -48,15 +48,14 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 {
 	int ret = 0;
 
-	if (likely(!test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)))
+	if (likely(!bo->moving))
 		goto out_unlock;
 
 	/*
 	 * Quick non-stalling check for idle.
 	 */
-	ret = ttm_bo_wait(bo, false, true);
-	if (likely(ret == 0))
-		goto out_unlock;
+	if (fence_is_signaled(bo->moving))
+		goto out_clear;
 
 	/*
 	 * If possible, avoid waiting for GPU with mmap_sem
@@ -68,17 +67,23 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 			goto out_unlock;
 
 		up_read(&vma->vm_mm->mmap_sem);
-		(void) ttm_bo_wait(bo, true, false);
+		(void) fence_wait(bo->moving, true);
 		goto out_unlock;
 	}
 
 	/*
 	 * Ordinary wait.
 	 */
-	ret = ttm_bo_wait(bo, true, false);
-	if (unlikely(ret != 0))
+	ret = fence_wait(bo->moving, true);
+	if (unlikely(ret != 0)) {
 		ret = (ret != -ERESTARTSYS) ? VM_FAULT_SIGBUS :
 			VM_FAULT_NOPAGE;
+		goto out_unlock;
+	}
+
+out_clear:
+	fence_put(bo->moving);
+	bo->moving = NULL;
 
 out_unlock:
 	return ret;
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 077ae9b2865d..d28d4333dcce 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -166,11 +166,15 @@ EXPORT_SYMBOL(ttm_tt_set_placement_caching);
 
 void ttm_tt_destroy(struct ttm_tt *ttm)
 {
-	if (unlikely(ttm == NULL))
+	int ret;
+
+	if (ttm == NULL)
 		return;
 
 	if (ttm->state == tt_bound) {
-		ttm_tt_unbind(ttm);
+		ret = ttm->func->unbind(ttm);
+		BUG_ON(ret);
+		ttm->state = tt_unbound;
 	}
 
 	if (ttm->state == tt_unbound)
@@ -251,17 +255,6 @@ void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma)
 }
 EXPORT_SYMBOL(ttm_dma_tt_fini);
 
-void ttm_tt_unbind(struct ttm_tt *ttm)
-{
-	int ret;
-
-	if (ttm->state == tt_bound) {
-		ret = ttm->func->unbind(ttm);
-		BUG_ON(ret);
-		ttm->state = tt_unbound;
-	}
-}
-
 int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem)
 {
 	int ret = 0;
diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c
index 4c0f26a644a3..8fc2b731b59a 100644
--- a/drivers/gpu/drm/vc4/vc4_crtc.c
+++ b/drivers/gpu/drm/vc4/vc4_crtc.c
@@ -46,12 +46,17 @@ struct vc4_crtc {
 	const struct vc4_crtc_data *data;
 	void __iomem *regs;
 
+	/* Timestamp at start of vblank irq - unaffected by lock delays. */
+	ktime_t t_vblank;
+
 	/* Which HVS channel we're using for our CRTC. */
 	int channel;
 
 	u8 lut_r[256];
 	u8 lut_g[256];
 	u8 lut_b[256];
+	/* Size in pixels of the COB memory allocated to this CRTC. */
+	u32 cob_size;
 
 	struct drm_pending_vblank_event *event;
 };
@@ -146,6 +151,144 @@ int vc4_crtc_debugfs_regs(struct seq_file *m, void *unused)
 }
 #endif
 
+int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id,
+			    unsigned int flags, int *vpos, int *hpos,
+			    ktime_t *stime, ktime_t *etime,
+			    const struct drm_display_mode *mode)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
+	u32 val;
+	int fifo_lines;
+	int vblank_lines;
+	int ret = 0;
+
+	/*
+	 * XXX Doesn't work well in interlaced mode yet, partially due
+	 * to problems in vc4 kms or drm core interlaced mode handling,
+	 * so disable for now in interlaced mode.
+	 */
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return ret;
+
+	/* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */
+
+	/* Get optional system timestamp before query. */
+	if (stime)
+		*stime = ktime_get();
+
+	/*
+	 * Read vertical scanline which is currently composed for our
+	 * pixelvalve by the HVS, and also the scaler status.
+	 */
+	val = HVS_READ(SCALER_DISPSTATX(vc4_crtc->channel));
+
+	/* Get optional system timestamp after query. */
+	if (etime)
+		*etime = ktime_get();
+
+	/* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */
+
+	/* Vertical position of hvs composed scanline. */
+	*vpos = VC4_GET_FIELD(val, SCALER_DISPSTATX_LINE);
+
+	/* No hpos info available. */
+	if (hpos)
+		*hpos = 0;
+
+	/* This is the offset we need for translating hvs -> pv scanout pos. */
+	fifo_lines = vc4_crtc->cob_size / mode->crtc_hdisplay;
+
+	if (fifo_lines > 0)
+		ret |= DRM_SCANOUTPOS_VALID;
+
+	/* HVS more than fifo_lines into frame for compositing? */
+	if (*vpos > fifo_lines) {
+		/*
+		 * We are in active scanout and can get some meaningful results
+		 * from HVS. The actual PV scanout can not trail behind more
+		 * than fifo_lines as that is the fifo's capacity. Assume that
+		 * in active scanout the HVS and PV work in lockstep wrt. HVS
+		 * refilling the fifo and PV consuming from the fifo, ie.
+		 * whenever the PV consumes and frees up a scanline in the
+		 * fifo, the HVS will immediately refill it, therefore
+		 * incrementing vpos. Therefore we choose HVS read position -
+		 * fifo size in scanlines as a estimate of the real scanout
+		 * position of the PV.
+		 */
+		*vpos -= fifo_lines + 1;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			*vpos /= 2;
+
+		ret |= DRM_SCANOUTPOS_ACCURATE;
+		return ret;
+	}
+
+	/*
+	 * Less: This happens when we are in vblank and the HVS, after getting
+	 * the VSTART restart signal from the PV, just started refilling its
+	 * fifo with new lines from the top-most lines of the new framebuffers.
+	 * The PV does not scan out in vblank, so does not remove lines from
+	 * the fifo, so the fifo will be full quickly and the HVS has to pause.
+	 * We can't get meaningful readings wrt. scanline position of the PV
+	 * and need to make things up in a approximative but consistent way.
+	 */
+	ret |= DRM_SCANOUTPOS_IN_VBLANK;
+	vblank_lines = mode->crtc_vtotal - mode->crtc_vdisplay;
+
+	if (flags & DRM_CALLED_FROM_VBLIRQ) {
+		/*
+		 * Assume the irq handler got called close to first
+		 * line of vblank, so PV has about a full vblank
+		 * scanlines to go, and as a base timestamp use the
+		 * one taken at entry into vblank irq handler, so it
+		 * is not affected by random delays due to lock
+		 * contention on event_lock or vblank_time lock in
+		 * the core.
+		 */
+		*vpos = -vblank_lines;
+
+		if (stime)
+			*stime = vc4_crtc->t_vblank;
+		if (etime)
+			*etime = vc4_crtc->t_vblank;
+
+		/*
+		 * If the HVS fifo is not yet full then we know for certain
+		 * we are at the very beginning of vblank, as the hvs just
+		 * started refilling, and the stime and etime timestamps
+		 * truly correspond to start of vblank.
+		 */
+		if ((val & SCALER_DISPSTATX_FULL) != SCALER_DISPSTATX_FULL)
+			ret |= DRM_SCANOUTPOS_ACCURATE;
+	} else {
+		/*
+		 * No clue where we are inside vblank. Return a vpos of zero,
+		 * which will cause calling code to just return the etime
+		 * timestamp uncorrected. At least this is no worse than the
+		 * standard fallback.
+		 */
+		*vpos = 0;
+	}
+
+	return ret;
+}
+
+int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id,
+				  int *max_error, struct timeval *vblank_time,
+				  unsigned flags)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
+	struct drm_crtc *crtc = &vc4_crtc->base;
+	struct drm_crtc_state *state = crtc->state;
+
+	/* Helper routine in DRM core does all the work: */
+	return drm_calc_vbltimestamp_from_scanoutpos(dev, crtc_id, max_error,
+						     vblank_time, flags,
+						     &state->adjusted_mode);
+}
+
 static void vc4_crtc_destroy(struct drm_crtc *crtc)
 {
 	drm_crtc_cleanup(crtc);
@@ -449,14 +592,6 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc,
 
 	WARN_ON_ONCE(dlist_next - dlist_start != vc4_state->mm.size);
 
-	HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
-		  vc4_state->mm.start);
-
-	if (debug_dump_regs) {
-		DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc));
-		vc4_hvs_dump_state(dev);
-	}
-
 	if (crtc->state->event) {
 		unsigned long flags;
 
@@ -466,8 +601,20 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc,
 
 		spin_lock_irqsave(&dev->event_lock, flags);
 		vc4_crtc->event = crtc->state->event;
-		spin_unlock_irqrestore(&dev->event_lock, flags);
 		crtc->state->event = NULL;
+
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	} else {
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+	}
+
+	if (debug_dump_regs) {
+		DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc));
+		vc4_hvs_dump_state(dev);
 	}
 }
 
@@ -493,12 +640,17 @@ static void vc4_crtc_handle_page_flip(struct vc4_crtc *vc4_crtc)
 {
 	struct drm_crtc *crtc = &vc4_crtc->base;
 	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+	u32 chan = vc4_crtc->channel;
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
-	if (vc4_crtc->event) {
+	if (vc4_crtc->event &&
+	    (vc4_state->mm.start == HVS_READ(SCALER_DISPLACTX(chan)))) {
 		drm_crtc_send_vblank_event(crtc, vc4_crtc->event);
 		vc4_crtc->event = NULL;
+		drm_crtc_vblank_put(crtc);
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 }
@@ -510,6 +662,7 @@ static irqreturn_t vc4_crtc_irq_handler(int irq, void *data)
 	irqreturn_t ret = IRQ_NONE;
 
 	if (stat & PV_INT_VFP_START) {
+		vc4_crtc->t_vblank = ktime_get();
 		CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START);
 		drm_crtc_handle_vblank(&vc4_crtc->base);
 		vc4_crtc_handle_page_flip(vc4_crtc);
@@ -549,6 +702,7 @@ vc4_async_page_flip_complete(struct vc4_seqno_cb *cb)
 		spin_unlock_irqrestore(&dev->event_lock, flags);
 	}
 
+	drm_crtc_vblank_put(crtc);
 	drm_framebuffer_unreference(flip_state->fb);
 	kfree(flip_state);
 
@@ -591,6 +745,8 @@ static int vc4_async_page_flip(struct drm_crtc *crtc,
 		return ret;
 	}
 
+	WARN_ON(drm_crtc_vblank_get(crtc) != 0);
+
 	/* Immediately update the plane's legacy fb pointer, so that later
 	 * modeset prep sees the state that will be present when the semaphore
 	 * is released.
@@ -711,6 +867,22 @@ static void vc4_set_crtc_possible_masks(struct drm_device *drm,
 	}
 }
 
+static void
+vc4_crtc_get_cob_allocation(struct vc4_crtc *vc4_crtc)
+{
+	struct drm_device *drm = vc4_crtc->base.dev;
+	struct vc4_dev *vc4 = to_vc4_dev(drm);
+	u32 dispbase = HVS_READ(SCALER_DISPBASEX(vc4_crtc->channel));
+	/* Top/base are supposed to be 4-pixel aligned, but the
+	 * Raspberry Pi firmware fills the low bits (which are
+	 * presumably ignored).
+	 */
+	u32 top = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_TOP) & ~3;
+	u32 base = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_BASE) & ~3;
+
+	vc4_crtc->cob_size = top - base + 4;
+}
+
 static int vc4_crtc_bind(struct device *dev, struct device *master, void *data)
 {
 	struct platform_device *pdev = to_platform_device(dev);
@@ -787,6 +959,8 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data)
 		crtc->cursor = cursor_plane;
 	}
 
+	vc4_crtc_get_cob_allocation(vc4_crtc);
+
 	CRTC_WRITE(PV_INTEN, 0);
 	CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START);
 	ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
diff --git a/drivers/gpu/drm/vc4/vc4_dpi.c b/drivers/gpu/drm/vc4/vc4_dpi.c
index dba1114297e4..275fedbdbd9e 100644
--- a/drivers/gpu/drm/vc4/vc4_dpi.c
+++ b/drivers/gpu/drm/vc4/vc4_dpi.c
@@ -227,14 +227,12 @@ static struct drm_connector *vc4_dpi_connector_init(struct drm_device *dev,
 {
 	struct drm_connector *connector = NULL;
 	struct vc4_dpi_connector *dpi_connector;
-	int ret = 0;
 
 	dpi_connector = devm_kzalloc(dev->dev, sizeof(*dpi_connector),
 				     GFP_KERNEL);
-	if (!dpi_connector) {
-		ret = -ENOMEM;
-		goto fail;
-	}
+	if (!dpi_connector)
+		return ERR_PTR(-ENOMEM);
+
 	connector = &dpi_connector->base;
 
 	dpi_connector->encoder = dpi->encoder;
@@ -251,12 +249,6 @@ static struct drm_connector *vc4_dpi_connector_init(struct drm_device *dev,
 	drm_mode_connector_attach_encoder(connector, dpi->encoder);
 
 	return connector;
-
- fail:
-	if (connector)
-		vc4_dpi_connector_destroy(connector);
-
-	return ERR_PTR(ret);
 }
 
 static const struct drm_encoder_funcs vc4_dpi_encoder_funcs = {
diff --git a/drivers/gpu/drm/vc4/vc4_drv.c b/drivers/gpu/drm/vc4/vc4_drv.c
index 493bb580a413..8b42d31a7f0e 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.c
+++ b/drivers/gpu/drm/vc4/vc4_drv.c
@@ -14,6 +14,7 @@
 #include <linux/module.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
 #include "drm_fb_cma_helper.h"
 
 #include "uapi/drm/vc4_drm.h"
@@ -43,6 +44,49 @@ void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index)
 	return map;
 }
 
+static int vc4_get_param_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file_priv)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct drm_vc4_get_param *args = data;
+	int ret;
+
+	if (args->pad != 0)
+		return -EINVAL;
+
+	switch (args->param) {
+	case DRM_VC4_PARAM_V3D_IDENT0:
+		ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
+		if (ret)
+			return ret;
+		args->value = V3D_READ(V3D_IDENT0);
+		pm_runtime_put(&vc4->v3d->pdev->dev);
+		break;
+	case DRM_VC4_PARAM_V3D_IDENT1:
+		ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
+		if (ret)
+			return ret;
+		args->value = V3D_READ(V3D_IDENT1);
+		pm_runtime_put(&vc4->v3d->pdev->dev);
+		break;
+	case DRM_VC4_PARAM_V3D_IDENT2:
+		ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
+		if (ret)
+			return ret;
+		args->value = V3D_READ(V3D_IDENT2);
+		pm_runtime_put(&vc4->v3d->pdev->dev);
+		break;
+	case DRM_VC4_PARAM_SUPPORTS_BRANCHES:
+		args->value = true;
+		break;
+	default:
+		DRM_DEBUG("Unknown parameter %d\n", args->param);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static void vc4_lastclose(struct drm_device *dev)
 {
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
@@ -65,14 +109,15 @@ static const struct file_operations vc4_drm_fops = {
 };
 
 static const struct drm_ioctl_desc vc4_drm_ioctls[] = {
-	DRM_IOCTL_DEF_DRV(VC4_SUBMIT_CL, vc4_submit_cl_ioctl, 0),
-	DRM_IOCTL_DEF_DRV(VC4_WAIT_SEQNO, vc4_wait_seqno_ioctl, 0),
-	DRM_IOCTL_DEF_DRV(VC4_WAIT_BO, vc4_wait_bo_ioctl, 0),
-	DRM_IOCTL_DEF_DRV(VC4_CREATE_BO, vc4_create_bo_ioctl, 0),
-	DRM_IOCTL_DEF_DRV(VC4_MMAP_BO, vc4_mmap_bo_ioctl, 0),
-	DRM_IOCTL_DEF_DRV(VC4_CREATE_SHADER_BO, vc4_create_shader_bo_ioctl, 0),
+	DRM_IOCTL_DEF_DRV(VC4_SUBMIT_CL, vc4_submit_cl_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(VC4_WAIT_SEQNO, vc4_wait_seqno_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(VC4_WAIT_BO, vc4_wait_bo_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(VC4_CREATE_BO, vc4_create_bo_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(VC4_MMAP_BO, vc4_mmap_bo_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(VC4_CREATE_SHADER_BO, vc4_create_shader_bo_ioctl, DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(VC4_GET_HANG_STATE, vc4_get_hang_state_ioctl,
 			  DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(VC4_GET_PARAM, vc4_get_param_ioctl, DRM_RENDER_ALLOW),
 };
 
 static struct drm_driver vc4_drm_driver = {
@@ -90,7 +135,9 @@ static struct drm_driver vc4_drm_driver = {
 
 	.enable_vblank = vc4_enable_vblank,
 	.disable_vblank = vc4_disable_vblank,
-	.get_vblank_counter = drm_vblank_count,
+	.get_vblank_counter = drm_vblank_no_hw_counter,
+	.get_scanout_position = vc4_crtc_get_scanoutpos,
+	.get_vblank_timestamp = vc4_crtc_get_vblank_timestamp,
 
 #if defined(CONFIG_DEBUG_FS)
 	.debugfs_init = vc4_debugfs_init,
@@ -194,8 +241,6 @@ static int vc4_drm_bind(struct device *dev)
 	vc4_bo_cache_init(drm);
 
 	drm_mode_config_init(drm);
-	if (ret)
-		goto unref;
 
 	vc4_gem_init(drm);
 
@@ -217,7 +262,6 @@ unbind_all:
 	component_unbind_all(dev, drm);
 gem_destroy:
 	vc4_gem_destroy(drm);
-unref:
 	drm_dev_unref(drm);
 	vc4_bo_cache_destroy(drm);
 	return ret;
@@ -245,8 +289,8 @@ static const struct component_master_ops vc4_drm_ops = {
 static struct platform_driver *const component_drivers[] = {
 	&vc4_hdmi_driver,
 	&vc4_dpi_driver,
-	&vc4_crtc_driver,
 	&vc4_hvs_driver,
+	&vc4_crtc_driver,
 	&vc4_v3d_driver,
 };
 
diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h
index c799baabc008..489e3de0c050 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.h
+++ b/drivers/gpu/drm/vc4/vc4_drv.h
@@ -355,6 +355,9 @@ struct vc4_validated_shader_info {
 	uint32_t uniforms_src_size;
 	uint32_t num_texture_samples;
 	struct vc4_texture_sample_info *texture_samples;
+
+	uint32_t num_uniform_addr_offsets;
+	uint32_t *uniform_addr_offsets;
 };
 
 /**
@@ -415,6 +418,13 @@ extern struct platform_driver vc4_crtc_driver;
 int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id);
 void vc4_disable_vblank(struct drm_device *dev, unsigned int crtc_id);
 int vc4_crtc_debugfs_regs(struct seq_file *m, void *arg);
+int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id,
+			    unsigned int flags, int *vpos, int *hpos,
+			    ktime_t *stime, ktime_t *etime,
+			    const struct drm_display_mode *mode);
+int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id,
+				  int *max_error, struct timeval *vblank_time,
+				  unsigned flags);
 
 /* vc4_debugfs.c */
 int vc4_debugfs_init(struct drm_minor *minor);
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
index 68df91c3f860..4452f3631cac 100644
--- a/drivers/gpu/drm/vc4/vc4_hdmi.c
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
@@ -456,12 +456,6 @@ static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data)
 	if (IS_ERR(hdmi->hd_regs))
 		return PTR_ERR(hdmi->hd_regs);
 
-	ddc_node = of_parse_phandle(dev->of_node, "ddc", 0);
-	if (!ddc_node) {
-		DRM_ERROR("Failed to find ddc node in device tree\n");
-		return -ENODEV;
-	}
-
 	hdmi->pixel_clock = devm_clk_get(dev, "pixel");
 	if (IS_ERR(hdmi->pixel_clock)) {
 		DRM_ERROR("Failed to get pixel clock\n");
@@ -473,7 +467,14 @@ static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data)
 		return PTR_ERR(hdmi->hsm_clock);
 	}
 
+	ddc_node = of_parse_phandle(dev->of_node, "ddc", 0);
+	if (!ddc_node) {
+		DRM_ERROR("Failed to find ddc node in device tree\n");
+		return -ENODEV;
+	}
+
 	hdmi->ddc = of_find_i2c_adapter_by_node(ddc_node);
+	of_node_put(ddc_node);
 	if (!hdmi->ddc) {
 		DRM_DEBUG("Failed to get ddc i2c adapter by node\n");
 		return -EPROBE_DEFER;
diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c
index 0fa3c8128875..4ac894d993cd 100644
--- a/drivers/gpu/drm/vc4/vc4_kms.c
+++ b/drivers/gpu/drm/vc4/vc4_kms.c
@@ -118,10 +118,18 @@ static int vc4_atomic_commit(struct drm_device *dev,
 		return -ENOMEM;
 
 	/* Make sure that any outstanding modesets have finished. */
-	ret = down_interruptible(&vc4->async_modeset);
-	if (ret) {
-		kfree(c);
-		return ret;
+	if (nonblock) {
+		ret = down_trylock(&vc4->async_modeset);
+		if (ret) {
+			kfree(c);
+			return -EBUSY;
+		}
+	} else {
+		ret = down_interruptible(&vc4->async_modeset);
+		if (ret) {
+			kfree(c);
+			return ret;
+		}
 	}
 
 	ret = drm_atomic_helper_prepare_planes(dev, state);
diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index 5d2c3d9fd17a..29e4b400e25e 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -94,6 +94,14 @@ static const struct hvs_format {
 		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true,
 	},
 	{
+		.drm = DRM_FORMAT_ABGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB, .has_alpha = true,
+	},
+	{
+		.drm = DRM_FORMAT_XBGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB, .has_alpha = false,
+	},
+	{
 		.drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565,
 		.pixel_order = HVS_PIXEL_ORDER_XRGB, .has_alpha = false,
 	},
diff --git a/drivers/gpu/drm/vc4/vc4_qpu_defines.h b/drivers/gpu/drm/vc4/vc4_qpu_defines.h
index d5c2f3c85ebb..f4e795a0d3f6 100644
--- a/drivers/gpu/drm/vc4/vc4_qpu_defines.h
+++ b/drivers/gpu/drm/vc4/vc4_qpu_defines.h
@@ -70,7 +70,7 @@ enum qpu_raddr {
 	QPU_R_ELEM_QPU = 38,
 	QPU_R_NOP,
 	QPU_R_XY_PIXEL_COORD = 41,
-	QPU_R_MS_REV_FLAGS = 41,
+	QPU_R_MS_REV_FLAGS = 42,
 	QPU_R_VPM = 48,
 	QPU_R_VPM_LD_BUSY,
 	QPU_R_VPM_LD_WAIT,
@@ -230,6 +230,15 @@ enum qpu_unpack_r4 {
 #define QPU_COND_MUL_SHIFT              46
 #define QPU_COND_MUL_MASK               QPU_MASK(48, 46)
 
+#define QPU_BRANCH_COND_SHIFT           52
+#define QPU_BRANCH_COND_MASK            QPU_MASK(55, 52)
+
+#define QPU_BRANCH_REL                  ((uint64_t)1 << 51)
+#define QPU_BRANCH_REG                  ((uint64_t)1 << 50)
+
+#define QPU_BRANCH_RADDR_A_SHIFT        45
+#define QPU_BRANCH_RADDR_A_MASK         QPU_MASK(49, 45)
+
 #define QPU_SF                          ((uint64_t)1 << 45)
 
 #define QPU_WADDR_ADD_SHIFT             38
@@ -261,4 +270,10 @@ enum qpu_unpack_r4 {
 #define QPU_OP_ADD_SHIFT                24
 #define QPU_OP_ADD_MASK                 QPU_MASK(28, 24)
 
+#define QPU_LOAD_IMM_SHIFT              0
+#define QPU_LOAD_IMM_MASK               QPU_MASK(31, 0)
+
+#define QPU_BRANCH_TARGET_SHIFT         0
+#define QPU_BRANCH_TARGET_MASK          QPU_MASK(31, 0)
+
 #endif /* VC4_QPU_DEFINES_H */
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index 6163b95c5411..160942a9180e 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -341,6 +341,10 @@
 #define SCALER_DISPLACT0                        0x00000030
 #define SCALER_DISPLACT1                        0x00000034
 #define SCALER_DISPLACT2                        0x00000038
+#define SCALER_DISPLACTX(x)			(SCALER_DISPLACT0 +	\
+						 (x) * (SCALER_DISPLACT1 - \
+							SCALER_DISPLACT0))
+
 #define SCALER_DISPCTRL0                        0x00000040
 # define SCALER_DISPCTRLX_ENABLE		BIT(31)
 # define SCALER_DISPCTRLX_RESET			BIT(30)
@@ -362,7 +366,6 @@
 # define SCALER_DISPBKGND_FILL			BIT(24)
 
 #define SCALER_DISPSTAT0                        0x00000048
-#define SCALER_DISPBASE0                        0x0000004c
 # define SCALER_DISPSTATX_MODE_MASK		VC4_MASK(31, 30)
 # define SCALER_DISPSTATX_MODE_SHIFT		30
 # define SCALER_DISPSTATX_MODE_DISABLED		0
@@ -371,6 +374,24 @@
 # define SCALER_DISPSTATX_MODE_EOF		3
 # define SCALER_DISPSTATX_FULL			BIT(29)
 # define SCALER_DISPSTATX_EMPTY			BIT(28)
+# define SCALER_DISPSTATX_FRAME_COUNT_MASK	VC4_MASK(17, 12)
+# define SCALER_DISPSTATX_FRAME_COUNT_SHIFT	12
+# define SCALER_DISPSTATX_LINE_MASK		VC4_MASK(11, 0)
+# define SCALER_DISPSTATX_LINE_SHIFT		0
+
+#define SCALER_DISPBASE0                        0x0000004c
+/* Last pixel in the COB (display FIFO memory) allocated to this HVS
+ * channel.  Must be 4-pixel aligned (and thus 4 pixels less than the
+ * next COB base).
+ */
+# define SCALER_DISPBASEX_TOP_MASK		VC4_MASK(31, 16)
+# define SCALER_DISPBASEX_TOP_SHIFT		16
+/* First pixel in the COB (display FIFO memory) allocated to this HVS
+ * channel.  Must be 4-pixel aligned.
+ */
+# define SCALER_DISPBASEX_BASE_MASK		VC4_MASK(15, 0)
+# define SCALER_DISPBASEX_BASE_SHIFT		0
+
 #define SCALER_DISPCTRL1                        0x00000050
 #define SCALER_DISPBKGND1                       0x00000054
 #define SCALER_DISPBKGNDX(x)			(SCALER_DISPBKGND0 +        \
@@ -381,6 +402,9 @@
 						 (x) * (SCALER_DISPSTAT1 - \
 							SCALER_DISPSTAT0))
 #define SCALER_DISPBASE1                        0x0000005c
+#define SCALER_DISPBASEX(x)			(SCALER_DISPBASE0 +        \
+						 (x) * (SCALER_DISPBASE1 - \
+							SCALER_DISPBASE0))
 #define SCALER_DISPCTRL2                        0x00000060
 #define SCALER_DISPCTRLX(x)			(SCALER_DISPCTRL0 +        \
 						 (x) * (SCALER_DISPCTRL1 - \
diff --git a/drivers/gpu/drm/vc4/vc4_validate.c b/drivers/gpu/drm/vc4/vc4_validate.c
index 24c2c746e8f3..9ce1d0adf882 100644
--- a/drivers/gpu/drm/vc4/vc4_validate.c
+++ b/drivers/gpu/drm/vc4/vc4_validate.c
@@ -802,7 +802,7 @@ validate_gl_shader_rec(struct drm_device *dev,
 		uint32_t src_offset = *(uint32_t *)(pkt_u + o);
 		uint32_t *texture_handles_u;
 		void *uniform_data_u;
-		uint32_t tex;
+		uint32_t tex, uni;
 
 		*(uint32_t *)(pkt_v + o) = bo[i]->paddr + src_offset;
 
@@ -840,6 +840,17 @@ validate_gl_shader_rec(struct drm_device *dev,
 			}
 		}
 
+		/* Fill in the uniform slots that need this shader's
+		 * start-of-uniforms address (used for resetting the uniform
+		 * stream in the presence of control flow).
+		 */
+		for (uni = 0;
+		     uni < validated_shader->num_uniform_addr_offsets;
+		     uni++) {
+			uint32_t o = validated_shader->uniform_addr_offsets[uni];
+			((uint32_t *)exec->uniforms_v)[o] = exec->uniforms_p;
+		}
+
 		*(uint32_t *)(pkt_v + o + 4) = exec->uniforms_p;
 
 		exec->uniforms_u += validated_shader->uniforms_src_size;
diff --git a/drivers/gpu/drm/vc4/vc4_validate_shaders.c b/drivers/gpu/drm/vc4/vc4_validate_shaders.c
index f67124b4c534..46527e989ce3 100644
--- a/drivers/gpu/drm/vc4/vc4_validate_shaders.c
+++ b/drivers/gpu/drm/vc4/vc4_validate_shaders.c
@@ -39,7 +39,17 @@
 #include "vc4_drv.h"
 #include "vc4_qpu_defines.h"
 
+#define LIVE_REG_COUNT (32 + 32 + 4)
+
 struct vc4_shader_validation_state {
+	/* Current IP being validated. */
+	uint32_t ip;
+
+	/* IP at the end of the BO, do not read shader[max_ip] */
+	uint32_t max_ip;
+
+	uint64_t *shader;
+
 	struct vc4_texture_sample_info tmu_setup[2];
 	int tmu_write_count[2];
 
@@ -49,8 +59,30 @@ struct vc4_shader_validation_state {
 	 *
 	 * This is used for the validation of direct address memory reads.
 	 */
-	uint32_t live_min_clamp_offsets[32 + 32 + 4];
-	bool live_max_clamp_regs[32 + 32 + 4];
+	uint32_t live_min_clamp_offsets[LIVE_REG_COUNT];
+	bool live_max_clamp_regs[LIVE_REG_COUNT];
+	uint32_t live_immediates[LIVE_REG_COUNT];
+
+	/* Bitfield of which IPs are used as branch targets.
+	 *
+	 * Used for validation that the uniform stream is updated at the right
+	 * points and clearing the texturing/clamping state.
+	 */
+	unsigned long *branch_targets;
+
+	/* Set when entering a basic block, and cleared when the uniform
+	 * address update is found.  This is used to make sure that we don't
+	 * read uniforms when the address is undefined.
+	 */
+	bool needs_uniform_address_update;
+
+	/* Set when we find a backwards branch.  If the branch is backwards,
+	 * the taraget is probably doing an address reset to read uniforms,
+	 * and so we need to be sure that a uniforms address is present in the
+	 * stream, even if the shader didn't need to read uniforms in later
+	 * basic blocks.
+	 */
+	bool needs_uniform_address_for_loop;
 };
 
 static uint32_t
@@ -129,11 +161,11 @@ record_texture_sample(struct vc4_validated_shader_info *validated_shader,
 }
 
 static bool
-check_tmu_write(uint64_t inst,
-		struct vc4_validated_shader_info *validated_shader,
+check_tmu_write(struct vc4_validated_shader_info *validated_shader,
 		struct vc4_shader_validation_state *validation_state,
 		bool is_mul)
 {
+	uint64_t inst = validation_state->shader[validation_state->ip];
 	uint32_t waddr = (is_mul ?
 			  QPU_GET_FIELD(inst, QPU_WADDR_MUL) :
 			  QPU_GET_FIELD(inst, QPU_WADDR_ADD));
@@ -162,7 +194,7 @@ check_tmu_write(uint64_t inst,
 			return false;
 		}
 
-		/* We assert that the the clamped address is the first
+		/* We assert that the clamped address is the first
 		 * argument, and the UBO base address is the second argument.
 		 * This is arbitrary, but simpler than supporting flipping the
 		 * two either way.
@@ -212,8 +244,14 @@ check_tmu_write(uint64_t inst,
 	/* Since direct uses a RADDR uniform reference, it will get counted in
 	 * check_instruction_reads()
 	 */
-	if (!is_direct)
+	if (!is_direct) {
+		if (validation_state->needs_uniform_address_update) {
+			DRM_ERROR("Texturing with undefined uniform address\n");
+			return false;
+		}
+
 		validated_shader->uniforms_size += 4;
+	}
 
 	if (submit) {
 		if (!record_texture_sample(validated_shader,
@@ -227,23 +265,138 @@ check_tmu_write(uint64_t inst,
 	return true;
 }
 
+static bool require_uniform_address_uniform(struct vc4_validated_shader_info *validated_shader)
+{
+	uint32_t o = validated_shader->num_uniform_addr_offsets;
+	uint32_t num_uniforms = validated_shader->uniforms_size / 4;
+
+	validated_shader->uniform_addr_offsets =
+		krealloc(validated_shader->uniform_addr_offsets,
+			 (o + 1) *
+			 sizeof(*validated_shader->uniform_addr_offsets),
+			 GFP_KERNEL);
+	if (!validated_shader->uniform_addr_offsets)
+		return false;
+
+	validated_shader->uniform_addr_offsets[o] = num_uniforms;
+	validated_shader->num_uniform_addr_offsets++;
+
+	return true;
+}
+
 static bool
-check_reg_write(uint64_t inst,
-		struct vc4_validated_shader_info *validated_shader,
+validate_uniform_address_write(struct vc4_validated_shader_info *validated_shader,
+			       struct vc4_shader_validation_state *validation_state,
+			       bool is_mul)
+{
+	uint64_t inst = validation_state->shader[validation_state->ip];
+	u32 add_b = QPU_GET_FIELD(inst, QPU_ADD_B);
+	u32 raddr_a = QPU_GET_FIELD(inst, QPU_RADDR_A);
+	u32 raddr_b = QPU_GET_FIELD(inst, QPU_RADDR_B);
+	u32 add_lri = raddr_add_a_to_live_reg_index(inst);
+	/* We want our reset to be pointing at whatever uniform follows the
+	 * uniforms base address.
+	 */
+	u32 expected_offset = validated_shader->uniforms_size + 4;
+
+	/* We only support absolute uniform address changes, and we
+	 * require that they be in the current basic block before any
+	 * of its uniform reads.
+	 *
+	 * One could potentially emit more efficient QPU code, by
+	 * noticing that (say) an if statement does uniform control
+	 * flow for all threads and that the if reads the same number
+	 * of uniforms on each side.  However, this scheme is easy to
+	 * validate so it's all we allow for now.
+	 */
+
+	if (QPU_GET_FIELD(inst, QPU_SIG) != QPU_SIG_NONE) {
+		DRM_ERROR("uniforms address change must be "
+			  "normal math\n");
+		return false;
+	}
+
+	if (is_mul || QPU_GET_FIELD(inst, QPU_OP_ADD) != QPU_A_ADD) {
+		DRM_ERROR("Uniform address reset must be an ADD.\n");
+		return false;
+	}
+
+	if (QPU_GET_FIELD(inst, QPU_COND_ADD) != QPU_COND_ALWAYS) {
+		DRM_ERROR("Uniform address reset must be unconditional.\n");
+		return false;
+	}
+
+	if (QPU_GET_FIELD(inst, QPU_PACK) != QPU_PACK_A_NOP &&
+	    !(inst & QPU_PM)) {
+		DRM_ERROR("No packing allowed on uniforms reset\n");
+		return false;
+	}
+
+	if (add_lri == -1) {
+		DRM_ERROR("First argument of uniform address write must be "
+			  "an immediate value.\n");
+		return false;
+	}
+
+	if (validation_state->live_immediates[add_lri] != expected_offset) {
+		DRM_ERROR("Resetting uniforms with offset %db instead of %db\n",
+			  validation_state->live_immediates[add_lri],
+			  expected_offset);
+		return false;
+	}
+
+	if (!(add_b == QPU_MUX_A && raddr_a == QPU_R_UNIF) &&
+	    !(add_b == QPU_MUX_B && raddr_b == QPU_R_UNIF)) {
+		DRM_ERROR("Second argument of uniform address write must be "
+			  "a uniform.\n");
+		return false;
+	}
+
+	validation_state->needs_uniform_address_update = false;
+	validation_state->needs_uniform_address_for_loop = false;
+	return require_uniform_address_uniform(validated_shader);
+}
+
+static bool
+check_reg_write(struct vc4_validated_shader_info *validated_shader,
 		struct vc4_shader_validation_state *validation_state,
 		bool is_mul)
 {
+	uint64_t inst = validation_state->shader[validation_state->ip];
 	uint32_t waddr = (is_mul ?
 			  QPU_GET_FIELD(inst, QPU_WADDR_MUL) :
 			  QPU_GET_FIELD(inst, QPU_WADDR_ADD));
+	uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG);
+	bool ws = inst & QPU_WS;
+	bool is_b = is_mul ^ ws;
+	u32 lri = waddr_to_live_reg_index(waddr, is_b);
+
+	if (lri != -1) {
+		uint32_t cond_add = QPU_GET_FIELD(inst, QPU_COND_ADD);
+		uint32_t cond_mul = QPU_GET_FIELD(inst, QPU_COND_MUL);
+
+		if (sig == QPU_SIG_LOAD_IMM &&
+		    QPU_GET_FIELD(inst, QPU_PACK) == QPU_PACK_A_NOP &&
+		    ((is_mul && cond_mul == QPU_COND_ALWAYS) ||
+		     (!is_mul && cond_add == QPU_COND_ALWAYS))) {
+			validation_state->live_immediates[lri] =
+				QPU_GET_FIELD(inst, QPU_LOAD_IMM);
+		} else {
+			validation_state->live_immediates[lri] = ~0;
+		}
+	}
 
 	switch (waddr) {
 	case QPU_W_UNIFORMS_ADDRESS:
-		/* XXX: We'll probably need to support this for reladdr, but
-		 * it's definitely a security-related one.
-		 */
-		DRM_ERROR("uniforms address load unsupported\n");
-		return false;
+		if (is_b) {
+			DRM_ERROR("relative uniforms address change "
+				  "unsupported\n");
+			return false;
+		}
+
+		return validate_uniform_address_write(validated_shader,
+						      validation_state,
+						      is_mul);
 
 	case QPU_W_TLB_COLOR_MS:
 	case QPU_W_TLB_COLOR_ALL:
@@ -261,7 +414,7 @@ check_reg_write(uint64_t inst,
 	case QPU_W_TMU1_T:
 	case QPU_W_TMU1_R:
 	case QPU_W_TMU1_B:
-		return check_tmu_write(inst, validated_shader, validation_state,
+		return check_tmu_write(validated_shader, validation_state,
 				       is_mul);
 
 	case QPU_W_HOST_INT:
@@ -294,10 +447,10 @@ check_reg_write(uint64_t inst,
 }
 
 static void
-track_live_clamps(uint64_t inst,
-		  struct vc4_validated_shader_info *validated_shader,
+track_live_clamps(struct vc4_validated_shader_info *validated_shader,
 		  struct vc4_shader_validation_state *validation_state)
 {
+	uint64_t inst = validation_state->shader[validation_state->ip];
 	uint32_t op_add = QPU_GET_FIELD(inst, QPU_OP_ADD);
 	uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD);
 	uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL);
@@ -369,10 +522,10 @@ track_live_clamps(uint64_t inst,
 }
 
 static bool
-check_instruction_writes(uint64_t inst,
-			 struct vc4_validated_shader_info *validated_shader,
+check_instruction_writes(struct vc4_validated_shader_info *validated_shader,
 			 struct vc4_shader_validation_state *validation_state)
 {
+	uint64_t inst = validation_state->shader[validation_state->ip];
 	uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD);
 	uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL);
 	bool ok;
@@ -382,20 +535,44 @@ check_instruction_writes(uint64_t inst,
 		return false;
 	}
 
-	ok = (check_reg_write(inst, validated_shader, validation_state,
-			      false) &&
-	      check_reg_write(inst, validated_shader, validation_state,
-			      true));
+	ok = (check_reg_write(validated_shader, validation_state, false) &&
+	      check_reg_write(validated_shader, validation_state, true));
 
-	track_live_clamps(inst, validated_shader, validation_state);
+	track_live_clamps(validated_shader, validation_state);
 
 	return ok;
 }
 
 static bool
-check_instruction_reads(uint64_t inst,
-			struct vc4_validated_shader_info *validated_shader)
+check_branch(uint64_t inst,
+	     struct vc4_validated_shader_info *validated_shader,
+	     struct vc4_shader_validation_state *validation_state,
+	     int ip)
+{
+	int32_t branch_imm = QPU_GET_FIELD(inst, QPU_BRANCH_TARGET);
+	uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD);
+	uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL);
+
+	if ((int)branch_imm < 0)
+		validation_state->needs_uniform_address_for_loop = true;
+
+	/* We don't want to have to worry about validation of this, and
+	 * there's no need for it.
+	 */
+	if (waddr_add != QPU_W_NOP || waddr_mul != QPU_W_NOP) {
+		DRM_ERROR("branch instruction at %d wrote a register.\n",
+			  validation_state->ip);
+		return false;
+	}
+
+	return true;
+}
+
+static bool
+check_instruction_reads(struct vc4_validated_shader_info *validated_shader,
+			struct vc4_shader_validation_state *validation_state)
 {
+	uint64_t inst = validation_state->shader[validation_state->ip];
 	uint32_t raddr_a = QPU_GET_FIELD(inst, QPU_RADDR_A);
 	uint32_t raddr_b = QPU_GET_FIELD(inst, QPU_RADDR_B);
 	uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG);
@@ -407,40 +584,204 @@ check_instruction_reads(uint64_t inst,
 		 * already be OOM.
 		 */
 		validated_shader->uniforms_size += 4;
+
+		if (validation_state->needs_uniform_address_update) {
+			DRM_ERROR("Uniform read with undefined uniform "
+				  "address\n");
+			return false;
+		}
+	}
+
+	return true;
+}
+
+/* Make sure that all branches are absolute and point within the shader, and
+ * note their targets for later.
+ */
+static bool
+vc4_validate_branches(struct vc4_shader_validation_state *validation_state)
+{
+	uint32_t max_branch_target = 0;
+	bool found_shader_end = false;
+	int ip;
+	int shader_end_ip = 0;
+	int last_branch = -2;
+
+	for (ip = 0; ip < validation_state->max_ip; ip++) {
+		uint64_t inst = validation_state->shader[ip];
+		int32_t branch_imm = QPU_GET_FIELD(inst, QPU_BRANCH_TARGET);
+		uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG);
+		uint32_t after_delay_ip = ip + 4;
+		uint32_t branch_target_ip;
+
+		if (sig == QPU_SIG_PROG_END) {
+			shader_end_ip = ip;
+			found_shader_end = true;
+			continue;
+		}
+
+		if (sig != QPU_SIG_BRANCH)
+			continue;
+
+		if (ip - last_branch < 4) {
+			DRM_ERROR("Branch at %d during delay slots\n", ip);
+			return false;
+		}
+		last_branch = ip;
+
+		if (inst & QPU_BRANCH_REG) {
+			DRM_ERROR("branching from register relative "
+				  "not supported\n");
+			return false;
+		}
+
+		if (!(inst & QPU_BRANCH_REL)) {
+			DRM_ERROR("relative branching required\n");
+			return false;
+		}
+
+		/* The actual branch target is the instruction after the delay
+		 * slots, plus whatever byte offset is in the low 32 bits of
+		 * the instruction.  Make sure we're not branching beyond the
+		 * end of the shader object.
+		 */
+		if (branch_imm % sizeof(inst) != 0) {
+			DRM_ERROR("branch target not aligned\n");
+			return false;
+		}
+
+		branch_target_ip = after_delay_ip + (branch_imm >> 3);
+		if (branch_target_ip >= validation_state->max_ip) {
+			DRM_ERROR("Branch at %d outside of shader (ip %d/%d)\n",
+				  ip, branch_target_ip,
+				  validation_state->max_ip);
+			return false;
+		}
+		set_bit(branch_target_ip, validation_state->branch_targets);
+
+		/* Make sure that the non-branching path is also not outside
+		 * the shader.
+		 */
+		if (after_delay_ip >= validation_state->max_ip) {
+			DRM_ERROR("Branch at %d continues past shader end "
+				  "(%d/%d)\n",
+				  ip, after_delay_ip, validation_state->max_ip);
+			return false;
+		}
+		set_bit(after_delay_ip, validation_state->branch_targets);
+		max_branch_target = max(max_branch_target, after_delay_ip);
+
+		/* There are two delay slots after program end is signaled
+		 * that are still executed, then we're finished.
+		 */
+		if (found_shader_end && ip == shader_end_ip + 2)
+			break;
+	}
+
+	if (max_branch_target > shader_end_ip) {
+		DRM_ERROR("Branch landed after QPU_SIG_PROG_END");
+		return false;
 	}
 
 	return true;
 }
 
+/* Resets any known state for the shader, used when we may be branched to from
+ * multiple locations in the program (or at shader start).
+ */
+static void
+reset_validation_state(struct vc4_shader_validation_state *validation_state)
+{
+	int i;
+
+	for (i = 0; i < 8; i++)
+		validation_state->tmu_setup[i / 4].p_offset[i % 4] = ~0;
+
+	for (i = 0; i < LIVE_REG_COUNT; i++) {
+		validation_state->live_min_clamp_offsets[i] = ~0;
+		validation_state->live_max_clamp_regs[i] = false;
+		validation_state->live_immediates[i] = ~0;
+	}
+}
+
+static bool
+texturing_in_progress(struct vc4_shader_validation_state *validation_state)
+{
+	return (validation_state->tmu_write_count[0] != 0 ||
+		validation_state->tmu_write_count[1] != 0);
+}
+
+static bool
+vc4_handle_branch_target(struct vc4_shader_validation_state *validation_state)
+{
+	uint32_t ip = validation_state->ip;
+
+	if (!test_bit(ip, validation_state->branch_targets))
+		return true;
+
+	if (texturing_in_progress(validation_state)) {
+		DRM_ERROR("Branch target landed during TMU setup\n");
+		return false;
+	}
+
+	/* Reset our live values tracking, since this instruction may have
+	 * multiple predecessors.
+	 *
+	 * One could potentially do analysis to determine that, for
+	 * example, all predecessors have a live max clamp in the same
+	 * register, but we don't bother with that.
+	 */
+	reset_validation_state(validation_state);
+
+	/* Since we've entered a basic block from potentially multiple
+	 * predecessors, we need the uniforms address to be updated before any
+	 * unforms are read.  We require that after any branch point, the next
+	 * uniform to be loaded is a uniform address offset.  That uniform's
+	 * offset will be marked by the uniform address register write
+	 * validation, or a one-off the end-of-program check.
+	 */
+	validation_state->needs_uniform_address_update = true;
+
+	return true;
+}
+
 struct vc4_validated_shader_info *
 vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
 {
 	bool found_shader_end = false;
 	int shader_end_ip = 0;
-	uint32_t ip, max_ip;
-	uint64_t *shader;
-	struct vc4_validated_shader_info *validated_shader;
+	uint32_t ip;
+	struct vc4_validated_shader_info *validated_shader = NULL;
 	struct vc4_shader_validation_state validation_state;
-	int i;
 
 	memset(&validation_state, 0, sizeof(validation_state));
+	validation_state.shader = shader_obj->vaddr;
+	validation_state.max_ip = shader_obj->base.size / sizeof(uint64_t);
 
-	for (i = 0; i < 8; i++)
-		validation_state.tmu_setup[i / 4].p_offset[i % 4] = ~0;
-	for (i = 0; i < ARRAY_SIZE(validation_state.live_min_clamp_offsets); i++)
-		validation_state.live_min_clamp_offsets[i] = ~0;
+	reset_validation_state(&validation_state);
 
-	shader = shader_obj->vaddr;
-	max_ip = shader_obj->base.size / sizeof(uint64_t);
+	validation_state.branch_targets =
+		kcalloc(BITS_TO_LONGS(validation_state.max_ip),
+			sizeof(unsigned long), GFP_KERNEL);
+	if (!validation_state.branch_targets)
+		goto fail;
 
 	validated_shader = kcalloc(1, sizeof(*validated_shader), GFP_KERNEL);
 	if (!validated_shader)
-		return NULL;
+		goto fail;
+
+	if (!vc4_validate_branches(&validation_state))
+		goto fail;
 
-	for (ip = 0; ip < max_ip; ip++) {
-		uint64_t inst = shader[ip];
+	for (ip = 0; ip < validation_state.max_ip; ip++) {
+		uint64_t inst = validation_state.shader[ip];
 		uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG);
 
+		validation_state.ip = ip;
+
+		if (!vc4_handle_branch_target(&validation_state))
+			goto fail;
+
 		switch (sig) {
 		case QPU_SIG_NONE:
 		case QPU_SIG_WAIT_FOR_SCOREBOARD:
@@ -450,13 +791,14 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
 		case QPU_SIG_LOAD_TMU1:
 		case QPU_SIG_PROG_END:
 		case QPU_SIG_SMALL_IMM:
-			if (!check_instruction_writes(inst, validated_shader,
+			if (!check_instruction_writes(validated_shader,
 						      &validation_state)) {
 				DRM_ERROR("Bad write at ip %d\n", ip);
 				goto fail;
 			}
 
-			if (!check_instruction_reads(inst, validated_shader))
+			if (!check_instruction_reads(validated_shader,
+						     &validation_state))
 				goto fail;
 
 			if (sig == QPU_SIG_PROG_END) {
@@ -467,13 +809,18 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
 			break;
 
 		case QPU_SIG_LOAD_IMM:
-			if (!check_instruction_writes(inst, validated_shader,
+			if (!check_instruction_writes(validated_shader,
 						      &validation_state)) {
 				DRM_ERROR("Bad LOAD_IMM write at ip %d\n", ip);
 				goto fail;
 			}
 			break;
 
+		case QPU_SIG_BRANCH:
+			if (!check_branch(inst, validated_shader,
+					  &validation_state, ip))
+				goto fail;
+			break;
 		default:
 			DRM_ERROR("Unsupported QPU signal %d at "
 				  "instruction %d\n", sig, ip);
@@ -487,13 +834,28 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
 			break;
 	}
 
-	if (ip == max_ip) {
+	if (ip == validation_state.max_ip) {
 		DRM_ERROR("shader failed to terminate before "
 			  "shader BO end at %zd\n",
 			  shader_obj->base.size);
 		goto fail;
 	}
 
+	/* If we did a backwards branch and we haven't emitted a uniforms
+	 * reset since then, we still need the uniforms stream to have the
+	 * uniforms address available so that the backwards branch can do its
+	 * uniforms reset.
+	 *
+	 * We could potentially prove that the backwards branch doesn't
+	 * contain any uses of uniforms until program exit, but that doesn't
+	 * seem to be worth the trouble.
+	 */
+	if (validation_state.needs_uniform_address_for_loop) {
+		if (!require_uniform_address_uniform(validated_shader))
+			goto fail;
+		validated_shader->uniforms_size += 4;
+	}
+
 	/* Again, no chance of integer overflow here because the worst case
 	 * scenario is 8 bytes of uniforms plus handles per 8-byte
 	 * instruction.
@@ -502,9 +864,12 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
 		(validated_shader->uniforms_size +
 		 4 * validated_shader->num_texture_samples);
 
+	kfree(validation_state.branch_targets);
+
 	return validated_shader;
 
 fail:
+	kfree(validation_state.branch_targets);
 	if (validated_shader) {
 		kfree(validated_shader->texture_samples);
 		kfree(validated_shader);
diff --git a/drivers/gpu/drm/virtio/virtgpu_ttm.c b/drivers/gpu/drm/virtio/virtgpu_ttm.c
index a0580815629f..80482ac5f95d 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ttm.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ttm.c
@@ -375,6 +375,12 @@ static int virtio_gpu_bo_move(struct ttm_buffer_object *bo,
 			      bool no_wait_gpu,
 			      struct ttm_mem_reg *new_mem)
 {
+	int ret;
+
+	ret = ttm_bo_wait(bo, interruptible, no_wait_gpu);
+	if (ret)
+		return ret;
+
 	virtio_gpu_move_null(bo, new_mem);
 	return 0;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c b/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c
index 6de283c8fa3e..f0374f9b56ca 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c
@@ -28,6 +28,7 @@
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
+#include <linux/frame.h>
 #include <asm/hypervisor.h>
 #include "drmP.h"
 #include "vmwgfx_msg.h"
@@ -194,7 +195,7 @@ static int vmw_send_msg(struct rpc_channel *channel, const char *msg)
 
 	return -EINVAL;
 }
-
+STACK_FRAME_NON_STANDARD(vmw_send_msg);
 
 
 /**
@@ -304,6 +305,7 @@ static int vmw_recv_msg(struct rpc_channel *channel, void **msg,
 
 	return 0;
 }
+STACK_FRAME_NON_STANDARD(vmw_recv_msg);
 
 
 /**
diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c
index a18db4d5347c..c5d82a8a2ec9 100644
--- a/drivers/gpu/host1x/cdma.c
+++ b/drivers/gpu/host1x/cdma.c
@@ -96,12 +96,12 @@ fail:
  */
 static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2)
 {
-	u32 pos = pb->pos;
-	u32 *p = (u32 *)((void *)pb->mapped + pos);
-	WARN_ON(pos == pb->fence);
+	u32 *p = (u32 *)((void *)pb->mapped + pb->pos);
+
+	WARN_ON(pb->pos == pb->fence);
 	*(p++) = op1;
 	*(p++) = op2;
-	pb->pos = (pos + 8) & (pb->size_bytes - 1);
+	pb->pos = (pb->pos + 8) & (pb->size_bytes - 1);
 }
 
 /*
@@ -134,14 +134,19 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma,
 				     enum cdma_event event)
 {
 	for (;;) {
+		struct push_buffer *pb = &cdma->push_buffer;
 		unsigned int space;
 
-		if (event == CDMA_EVENT_SYNC_QUEUE_EMPTY)
+		switch (event) {
+		case CDMA_EVENT_SYNC_QUEUE_EMPTY:
 			space = list_empty(&cdma->sync_queue) ? 1 : 0;
-		else if (event == CDMA_EVENT_PUSH_BUFFER_SPACE) {
-			struct push_buffer *pb = &cdma->push_buffer;
+			break;
+
+		case CDMA_EVENT_PUSH_BUFFER_SPACE:
 			space = host1x_pushbuffer_space(pb);
-		} else {
+			break;
+
+		default:
 			WARN_ON(1);
 			return -EINVAL;
 		}
@@ -159,12 +164,14 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma,
 			mutex_lock(&cdma->lock);
 			continue;
 		}
+
 		cdma->event = event;
 
 		mutex_unlock(&cdma->lock);
 		down(&cdma->sem);
 		mutex_lock(&cdma->lock);
 	}
+
 	return 0;
 }
 
@@ -234,6 +241,7 @@ static void update_cdma_locked(struct host1x_cdma *cdma)
 			/* Start timer on next pending syncpt */
 			if (job->timeout)
 				cdma_start_timer_locked(cdma, job);
+
 			break;
 		}
 
@@ -247,7 +255,9 @@ static void update_cdma_locked(struct host1x_cdma *cdma)
 		/* Pop push buffer slots */
 		if (job->num_slots) {
 			struct push_buffer *pb = &cdma->push_buffer;
+
 			host1x_pushbuffer_pop(pb, job->num_slots);
+
 			if (cdma->event == CDMA_EVENT_PUSH_BUFFER_SPACE)
 				signal = true;
 		}
@@ -269,11 +279,9 @@ static void update_cdma_locked(struct host1x_cdma *cdma)
 void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma,
 				   struct device *dev)
 {
-	u32 restart_addr;
-	u32 syncpt_incrs;
-	struct host1x_job *job = NULL;
-	u32 syncpt_val;
 	struct host1x *host1x = cdma_to_host1x(cdma);
+	u32 restart_addr, syncpt_incrs, syncpt_val;
+	struct host1x_job *job = NULL;
 
 	syncpt_val = host1x_syncpt_load(cdma->timeout.syncpt);
 
@@ -342,9 +350,11 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma,
 		syncpt_val += syncpt_incrs;
 	}
 
-	/* The following sumbits from the same client may be dependent on the
+	/*
+	 * The following sumbits from the same client may be dependent on the
 	 * failed submit and therefore they may fail. Force a small timeout
-	 * to make the queue cleanup faster */
+	 * to make the queue cleanup faster.
+	 */
 
 	list_for_each_entry_from(job, &cdma->sync_queue, list)
 		if (job->client == cdma->timeout.client)
@@ -375,6 +385,7 @@ int host1x_cdma_init(struct host1x_cdma *cdma)
 	err = host1x_pushbuffer_init(&cdma->push_buffer);
 	if (err)
 		return err;
+
 	return 0;
 }
 
@@ -410,6 +421,7 @@ int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job)
 		/* init state on first submit with timeout value */
 		if (!cdma->timeout.initialized) {
 			int err;
+
 			err = host1x_hw_cdma_timeout_init(host1x, cdma,
 							  job->syncpt_id);
 			if (err) {
@@ -418,6 +430,7 @@ int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job)
 			}
 		}
 	}
+
 	if (!cdma->running)
 		host1x_hw_cdma_start(host1x, cdma);
 
@@ -448,6 +461,7 @@ void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2)
 		slots_free = host1x_cdma_wait_locked(cdma,
 						CDMA_EVENT_PUSH_BUFFER_SPACE);
 	}
+
 	cdma->slots_free = slots_free - 1;
 	cdma->slots_used++;
 	host1x_pushbuffer_push(pb, op1, op2);
diff --git a/drivers/gpu/host1x/channel.c b/drivers/gpu/host1x/channel.c
index b4ae3affb987..8f437d924c10 100644
--- a/drivers/gpu/host1x/channel.c
+++ b/drivers/gpu/host1x/channel.c
@@ -83,9 +83,10 @@ EXPORT_SYMBOL(host1x_channel_put);
 struct host1x_channel *host1x_channel_request(struct device *dev)
 {
 	struct host1x *host = dev_get_drvdata(dev->parent);
-	int max_channels = host->info->nb_channels;
+	unsigned int max_channels = host->info->nb_channels;
 	struct host1x_channel *channel = NULL;
-	int index, err;
+	unsigned long index;
+	int err;
 
 	mutex_lock(&host->chlist_mutex);
 
diff --git a/drivers/gpu/host1x/debug.c b/drivers/gpu/host1x/debug.c
index ee3d12b51c50..d9330fcc62ad 100644
--- a/drivers/gpu/host1x/debug.c
+++ b/drivers/gpu/host1x/debug.c
@@ -39,6 +39,7 @@ void host1x_debug_output(struct output *o, const char *fmt, ...)
 	va_start(args, fmt);
 	len = vsnprintf(o->buf, sizeof(o->buf), fmt, args);
 	va_end(args);
+
 	o->fn(o->ctx, o->buf, len);
 }
 
@@ -48,13 +49,17 @@ static int show_channels(struct host1x_channel *ch, void *data, bool show_fifo)
 	struct output *o = data;
 
 	mutex_lock(&ch->reflock);
+
 	if (ch->refcount) {
 		mutex_lock(&ch->cdma.lock);
+
 		if (show_fifo)
 			host1x_hw_show_channel_fifo(m, ch, o);
+
 		host1x_hw_show_channel_cdma(m, ch, o);
 		mutex_unlock(&ch->cdma.lock);
 	}
+
 	mutex_unlock(&ch->reflock);
 
 	return 0;
@@ -62,22 +67,27 @@ static int show_channels(struct host1x_channel *ch, void *data, bool show_fifo)
 
 static void show_syncpts(struct host1x *m, struct output *o)
 {
-	int i;
+	unsigned int i;
+
 	host1x_debug_output(o, "---- syncpts ----\n");
+
 	for (i = 0; i < host1x_syncpt_nb_pts(m); i++) {
 		u32 max = host1x_syncpt_read_max(m->syncpt + i);
 		u32 min = host1x_syncpt_load(m->syncpt + i);
+
 		if (!min && !max)
 			continue;
-		host1x_debug_output(o, "id %d (%s) min %d max %d\n",
+
+		host1x_debug_output(o, "id %u (%s) min %d max %d\n",
 				    i, m->syncpt[i].name, min, max);
 	}
 
 	for (i = 0; i < host1x_syncpt_nb_bases(m); i++) {
 		u32 base_val;
+
 		base_val = host1x_syncpt_load_wait_base(m->syncpt + i);
 		if (base_val)
-			host1x_debug_output(o, "waitbase id %d val %d\n", i,
+			host1x_debug_output(o, "waitbase id %u val %d\n", i,
 					    base_val);
 	}
 
@@ -114,7 +124,9 @@ static int host1x_debug_show_all(struct seq_file *s, void *unused)
 		.fn = write_to_seqfile,
 		.ctx = s
 	};
+
 	show_all(s->private, &o);
+
 	return 0;
 }
 
@@ -124,7 +136,9 @@ static int host1x_debug_show(struct seq_file *s, void *unused)
 		.fn = write_to_seqfile,
 		.ctx = s
 	};
+
 	show_all_no_fifo(s->private, &o);
+
 	return 0;
 }
 
@@ -134,10 +148,10 @@ static int host1x_debug_open_all(struct inode *inode, struct file *file)
 }
 
 static const struct file_operations host1x_debug_all_fops = {
-	.open		= host1x_debug_open_all,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
+	.open = host1x_debug_open_all,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
 };
 
 static int host1x_debug_open(struct inode *inode, struct file *file)
@@ -146,10 +160,10 @@ static int host1x_debug_open(struct inode *inode, struct file *file)
 }
 
 static const struct file_operations host1x_debug_fops = {
-	.open		= host1x_debug_open,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
+	.open = host1x_debug_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
 };
 
 static void host1x_debugfs_init(struct host1x *host1x)
@@ -201,6 +215,7 @@ void host1x_debug_dump(struct host1x *host1x)
 	struct output o = {
 		.fn = write_to_printk
 	};
+
 	show_all(host1x, &o);
 }
 
@@ -209,5 +224,6 @@ void host1x_debug_dump_syncpts(struct host1x *host1x)
 	struct output o = {
 		.fn = write_to_printk
 	};
+
 	show_syncpts(host1x, &o);
 }
diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
index ff348690df94..a62317af76ad 100644
--- a/drivers/gpu/host1x/dev.c
+++ b/drivers/gpu/host1x/dev.c
@@ -63,13 +63,13 @@ u32 host1x_ch_readl(struct host1x_channel *ch, u32 r)
 }
 
 static const struct host1x_info host1x01_info = {
-	.nb_channels	= 8,
-	.nb_pts		= 32,
-	.nb_mlocks	= 16,
-	.nb_bases	= 8,
-	.init		= host1x01_init,
-	.sync_offset	= 0x3000,
-	.dma_mask	= DMA_BIT_MASK(32),
+	.nb_channels = 8,
+	.nb_pts = 32,
+	.nb_mlocks = 16,
+	.nb_bases = 8,
+	.init = host1x01_init,
+	.sync_offset = 0x3000,
+	.dma_mask = DMA_BIT_MASK(32),
 };
 
 static const struct host1x_info host1x02_info = {
@@ -102,7 +102,7 @@ static const struct host1x_info host1x05_info = {
 	.dma_mask = DMA_BIT_MASK(34),
 };
 
-static struct of_device_id host1x_of_match[] = {
+static const struct of_device_id host1x_of_match[] = {
 	{ .compatible = "nvidia,tegra210-host1x", .data = &host1x05_info, },
 	{ .compatible = "nvidia,tegra124-host1x", .data = &host1x04_info, },
 	{ .compatible = "nvidia,tegra114-host1x", .data = &host1x02_info, },
diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h
index dace124994bb..5220510f39da 100644
--- a/drivers/gpu/host1x/dev.h
+++ b/drivers/gpu/host1x/dev.h
@@ -45,7 +45,7 @@ struct host1x_cdma_ops {
 	void (*start)(struct host1x_cdma *cdma);
 	void (*stop)(struct host1x_cdma *cdma);
 	void (*flush)(struct  host1x_cdma *cdma);
-	int (*timeout_init)(struct host1x_cdma *cdma, u32 syncpt_id);
+	int (*timeout_init)(struct host1x_cdma *cdma, unsigned int syncpt);
 	void (*timeout_destroy)(struct host1x_cdma *cdma);
 	void (*freeze)(struct host1x_cdma *cdma);
 	void (*resume)(struct host1x_cdma *cdma, u32 getptr);
@@ -82,21 +82,21 @@ struct host1x_intr_ops {
 	int (*init_host_sync)(struct host1x *host, u32 cpm,
 		void (*syncpt_thresh_work)(struct work_struct *work));
 	void (*set_syncpt_threshold)(
-		struct host1x *host, u32 id, u32 thresh);
-	void (*enable_syncpt_intr)(struct host1x *host, u32 id);
-	void (*disable_syncpt_intr)(struct host1x *host, u32 id);
+		struct host1x *host, unsigned int id, u32 thresh);
+	void (*enable_syncpt_intr)(struct host1x *host, unsigned int id);
+	void (*disable_syncpt_intr)(struct host1x *host, unsigned int id);
 	void (*disable_all_syncpt_intrs)(struct host1x *host);
 	int (*free_syncpt_irq)(struct host1x *host);
 };
 
 struct host1x_info {
-	int	nb_channels;		/* host1x: num channels supported */
-	int	nb_pts;			/* host1x: num syncpoints supported */
-	int	nb_bases;		/* host1x: num syncpoints supported */
-	int	nb_mlocks;		/* host1x: number of mlocks */
-	int	(*init)(struct host1x *); /* initialize per SoC ops */
-	int	sync_offset;
-	u64	dma_mask;		/* mask of addressable memory */
+	unsigned int nb_channels; /* host1x: number of channels supported */
+	unsigned int nb_pts; /* host1x: number of syncpoints supported */
+	unsigned int nb_bases; /* host1x: number of syncpoint bases supported */
+	unsigned int nb_mlocks; /* host1x: number of mlocks supported */
+	int (*init)(struct host1x *host1x); /* initialize per SoC ops */
+	unsigned int sync_offset; /* offset of syncpoint registers */
+	u64 dma_mask; /* mask of addressable memory */
 };
 
 struct host1x {
@@ -109,7 +109,6 @@ struct host1x {
 	struct clk *clk;
 
 	struct mutex intr_mutex;
-	struct workqueue_struct *intr_wq;
 	int intr_syncpt_irq;
 
 	const struct host1x_syncpt_ops *syncpt_op;
@@ -183,19 +182,20 @@ static inline int host1x_hw_intr_init_host_sync(struct host1x *host, u32 cpm,
 }
 
 static inline void host1x_hw_intr_set_syncpt_threshold(struct host1x *host,
-						       u32 id, u32 thresh)
+						       unsigned int id,
+						       u32 thresh)
 {
 	host->intr_op->set_syncpt_threshold(host, id, thresh);
 }
 
 static inline void host1x_hw_intr_enable_syncpt_intr(struct host1x *host,
-						     u32 id)
+						     unsigned int id)
 {
 	host->intr_op->enable_syncpt_intr(host, id);
 }
 
 static inline void host1x_hw_intr_disable_syncpt_intr(struct host1x *host,
-						      u32 id)
+						      unsigned int id)
 {
 	host->intr_op->disable_syncpt_intr(host, id);
 }
@@ -212,9 +212,9 @@ static inline int host1x_hw_intr_free_syncpt_irq(struct host1x *host)
 
 static inline int host1x_hw_channel_init(struct host1x *host,
 					 struct host1x_channel *channel,
-					 int chid)
+					 unsigned int id)
 {
-	return host->channel_op->init(channel, host, chid);
+	return host->channel_op->init(channel, host, id);
 }
 
 static inline int host1x_hw_channel_submit(struct host1x *host,
@@ -243,9 +243,9 @@ static inline void host1x_hw_cdma_flush(struct host1x *host,
 
 static inline int host1x_hw_cdma_timeout_init(struct host1x *host,
 					      struct host1x_cdma *cdma,
-					      u32 syncpt_id)
+					      unsigned int syncpt)
 {
-	return host->cdma_op->timeout_init(cdma, syncpt_id);
+	return host->cdma_op->timeout_init(cdma, syncpt);
 }
 
 static inline void host1x_hw_cdma_timeout_destroy(struct host1x *host,
diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c
index 305ea8f3382d..659c1bbfeeba 100644
--- a/drivers/gpu/host1x/hw/cdma_hw.c
+++ b/drivers/gpu/host1x/hw/cdma_hw.c
@@ -41,7 +41,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
 {
 	struct host1x *host1x = cdma_to_host1x(cdma);
 	struct push_buffer *pb = &cdma->push_buffer;
-	u32 i;
+	unsigned int i;
 
 	for (i = 0; i < syncpt_incrs; i++)
 		host1x_syncpt_incr(cdma->timeout.syncpt);
@@ -58,6 +58,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
 			&pb->phys, getptr);
 		getptr = (getptr + 8) & (pb->size_bytes - 1);
 	}
+
 	wmb();
 }
 
@@ -162,12 +163,14 @@ static void cdma_stop(struct host1x_cdma *cdma)
 	struct host1x_channel *ch = cdma_to_channel(cdma);
 
 	mutex_lock(&cdma->lock);
+
 	if (cdma->running) {
 		host1x_cdma_wait_locked(cdma, CDMA_EVENT_SYNC_QUEUE_EMPTY);
 		host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
 				 HOST1X_CHANNEL_DMACTRL);
 		cdma->running = false;
 	}
+
 	mutex_unlock(&cdma->lock);
 }
 
@@ -213,11 +216,11 @@ static void cdma_resume(struct host1x_cdma *cdma, u32 getptr)
 	u32 cmdproc_stop;
 
 	dev_dbg(host1x->dev,
-		"resuming channel (id %d, DMAGET restart = 0x%x)\n",
+		"resuming channel (id %u, DMAGET restart = 0x%x)\n",
 		ch->id, getptr);
 
 	cmdproc_stop = host1x_sync_readl(host1x, HOST1X_SYNC_CMDPROC_STOP);
-	cmdproc_stop &= ~(BIT(ch->id));
+	cmdproc_stop &= ~BIT(ch->id);
 	host1x_sync_writel(host1x, cmdproc_stop, HOST1X_SYNC_CMDPROC_STOP);
 
 	cdma->torndown = false;
@@ -231,14 +234,11 @@ static void cdma_resume(struct host1x_cdma *cdma, u32 getptr)
  */
 static void cdma_timeout_handler(struct work_struct *work)
 {
+	u32 prev_cmdproc, cmdproc_stop, syncpt_val;
 	struct host1x_cdma *cdma;
 	struct host1x *host1x;
 	struct host1x_channel *ch;
 
-	u32 syncpt_val;
-
-	u32 prev_cmdproc, cmdproc_stop;
-
 	cdma = container_of(to_delayed_work(work), struct host1x_cdma,
 			    timeout.wq);
 	host1x = cdma_to_host1x(cdma);
@@ -277,9 +277,9 @@ static void cdma_timeout_handler(struct work_struct *work)
 		return;
 	}
 
-	dev_warn(host1x->dev, "%s: timeout: %d (%s), HW thresh %d, done %d\n",
-		__func__, cdma->timeout.syncpt->id, cdma->timeout.syncpt->name,
-		syncpt_val, cdma->timeout.syncpt_val);
+	dev_warn(host1x->dev, "%s: timeout: %u (%s), HW thresh %d, done %d\n",
+		 __func__, cdma->timeout.syncpt->id, cdma->timeout.syncpt->name,
+		 syncpt_val, cdma->timeout.syncpt_val);
 
 	/* stop HW, resetting channel/module */
 	host1x_hw_cdma_freeze(host1x, cdma);
@@ -291,7 +291,7 @@ static void cdma_timeout_handler(struct work_struct *work)
 /*
  * Init timeout resources
  */
-static int cdma_timeout_init(struct host1x_cdma *cdma, u32 syncpt_id)
+static int cdma_timeout_init(struct host1x_cdma *cdma, unsigned int syncpt)
 {
 	INIT_DELAYED_WORK(&cdma->timeout.wq, cdma_timeout_handler);
 	cdma->timeout.initialized = true;
@@ -306,6 +306,7 @@ static void cdma_timeout_destroy(struct host1x_cdma *cdma)
 {
 	if (cdma->timeout.initialized)
 		cancel_delayed_work(&cdma->timeout.wq);
+
 	cdma->timeout.initialized = false;
 }
 
diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c
index 946c332c3906..5e8df78b7acd 100644
--- a/drivers/gpu/host1x/hw/channel_hw.c
+++ b/drivers/gpu/host1x/hw/channel_hw.c
@@ -46,6 +46,7 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo,
 		 */
 		for (i = 0; i < words; i += TRACE_MAX_LENGTH) {
 			u32 num_words = min(words - i, TRACE_MAX_LENGTH);
+
 			offset += i * sizeof(u32);
 
 			trace_host1x_cdma_push_gather(dev_name(dev), bo,
@@ -66,6 +67,7 @@ static void submit_gathers(struct host1x_job *job)
 		struct host1x_job_gather *g = &job->gathers[i];
 		u32 op1 = host1x_opcode_gather(g->words);
 		u32 op2 = g->base + g->offset;
+
 		trace_write_gather(cdma, g->bo, g->offset, op1 & 0xffff);
 		host1x_cdma_push(cdma, op1, op2);
 	}
@@ -75,7 +77,8 @@ static inline void synchronize_syncpt_base(struct host1x_job *job)
 {
 	struct host1x *host = dev_get_drvdata(job->channel->dev->parent);
 	struct host1x_syncpt *sp = host->syncpt + job->syncpt_id;
-	u32 id, value;
+	unsigned int id;
+	u32 value;
 
 	value = host1x_syncpt_read_max(sp);
 	id = sp->base->id;
diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c
index cc3f1825c735..7a4a3286e4a7 100644
--- a/drivers/gpu/host1x/hw/debug_hw.c
+++ b/drivers/gpu/host1x/hw/debug_hw.c
@@ -40,8 +40,7 @@ enum {
 
 static unsigned int show_channel_command(struct output *o, u32 val)
 {
-	unsigned mask;
-	unsigned subop;
+	unsigned int mask, subop;
 
 	switch (val >> 28) {
 	case HOST1X_OPCODE_SETCLASS:
@@ -51,12 +50,11 @@ static unsigned int show_channel_command(struct output *o, u32 val)
 					    val >> 6 & 0x3ff,
 					    val >> 16 & 0xfff, mask);
 			return hweight8(mask);
-		} else {
-			host1x_debug_output(o, "SETCL(class=%03x)\n",
-					    val >> 6 & 0x3ff);
-			return 0;
 		}
 
+		host1x_debug_output(o, "SETCL(class=%03x)\n", val >> 6 & 0x3ff);
+		return 0;
+
 	case HOST1X_OPCODE_INCR:
 		host1x_debug_output(o, "INCR(offset=%03x, [",
 				    val >> 16 & 0xfff);
@@ -143,7 +141,8 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma)
 	struct host1x_job *job;
 
 	list_for_each_entry(job, &cdma->sync_queue, list) {
-		int i;
+		unsigned int i;
+
 		host1x_debug_output(o, "\n%p: JOB, syncpt_id=%d, syncpt_val=%d, first_get=%08x, timeout=%d num_slots=%d, num_handles=%d\n",
 				    job, job->syncpt_id, job->syncpt_end,
 				    job->first_get, job->timeout,
@@ -190,7 +189,7 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 	cbread = host1x_sync_readl(host, HOST1X_SYNC_CBREAD(ch->id));
 	cbstat = host1x_sync_readl(host, HOST1X_SYNC_CBSTAT(ch->id));
 
-	host1x_debug_output(o, "%d-%s: ", ch->id, dev_name(ch->dev));
+	host1x_debug_output(o, "%u-%s: ", ch->id, dev_name(ch->dev));
 
 	if (HOST1X_CHANNEL_DMACTRL_DMASTOP_V(dmactrl) ||
 	    !ch->cdma.push_buffer.mapped) {
@@ -200,14 +199,13 @@ static void host1x_debug_show_channel_cdma(struct host1x *host,
 
 	if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) == HOST1X_CLASS_HOST1X &&
 	    HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) ==
-	    HOST1X_UCLASS_WAIT_SYNCPT)
+			HOST1X_UCLASS_WAIT_SYNCPT)
 		host1x_debug_output(o, "waiting on syncpt %d val %d\n",
 				    cbread >> 24, cbread & 0xffffff);
 	else if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) ==
-	   HOST1X_CLASS_HOST1X &&
-	   HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) ==
-	   HOST1X_UCLASS_WAIT_SYNCPT_BASE) {
-
+				HOST1X_CLASS_HOST1X &&
+		 HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) ==
+				HOST1X_UCLASS_WAIT_SYNCPT_BASE) {
 		base = (cbread >> 16) & 0xff;
 		baseval =
 			host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(base));
@@ -236,7 +234,7 @@ static void host1x_debug_show_channel_fifo(struct host1x *host,
 	u32 val, rd_ptr, wr_ptr, start, end;
 	unsigned int data_count = 0;
 
-	host1x_debug_output(o, "%d: fifo:\n", ch->id);
+	host1x_debug_output(o, "%u: fifo:\n", ch->id);
 
 	val = host1x_ch_readl(ch, HOST1X_CHANNEL_FIFOSTAT);
 	host1x_debug_output(o, "FIFOSTAT %08x\n", val);
@@ -290,20 +288,22 @@ static void host1x_debug_show_channel_fifo(struct host1x *host,
 
 static void host1x_debug_show_mlocks(struct host1x *host, struct output *o)
 {
-	int i;
+	unsigned int i;
 
 	host1x_debug_output(o, "---- mlocks ----\n");
+
 	for (i = 0; i < host1x_syncpt_nb_mlocks(host); i++) {
 		u32 owner =
 			host1x_sync_readl(host, HOST1X_SYNC_MLOCK_OWNER(i));
 		if (HOST1X_SYNC_MLOCK_OWNER_CH_OWNS_V(owner))
-			host1x_debug_output(o, "%d: locked by channel %d\n",
+			host1x_debug_output(o, "%u: locked by channel %u\n",
 				i, HOST1X_SYNC_MLOCK_OWNER_CHID_V(owner));
 		else if (HOST1X_SYNC_MLOCK_OWNER_CPU_OWNS_V(owner))
-			host1x_debug_output(o, "%d: locked by cpu\n", i);
+			host1x_debug_output(o, "%u: locked by cpu\n", i);
 		else
-			host1x_debug_output(o, "%d: unlocked\n", i);
+			host1x_debug_output(o, "%u: unlocked\n", i);
 	}
+
 	host1x_debug_output(o, "\n");
 }
 
diff --git a/drivers/gpu/host1x/hw/intr_hw.c b/drivers/gpu/host1x/hw/intr_hw.c
index e1e31e9e67cd..dacb8009a605 100644
--- a/drivers/gpu/host1x/hw/intr_hw.c
+++ b/drivers/gpu/host1x/hw/intr_hw.c
@@ -38,14 +38,14 @@ static void host1x_intr_syncpt_handle(struct host1x_syncpt *syncpt)
 	host1x_sync_writel(host, BIT_MASK(id),
 		HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(BIT_WORD(id)));
 
-	queue_work(host->intr_wq, &syncpt->intr.work);
+	schedule_work(&syncpt->intr.work);
 }
 
 static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id)
 {
 	struct host1x *host = dev_id;
 	unsigned long reg;
-	int i, id;
+	unsigned int i, id;
 
 	for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); i++) {
 		reg = host1x_sync_readl(host,
@@ -62,7 +62,7 @@ static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id)
 
 static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host)
 {
-	u32 i;
+	unsigned int i;
 
 	for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); ++i) {
 		host1x_sync_writel(host, 0xffffffffu,
@@ -72,10 +72,12 @@ static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host)
 	}
 }
 
-static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm,
-	void (*syncpt_thresh_work)(struct work_struct *))
+static int
+_host1x_intr_init_host_sync(struct host1x *host, u32 cpm,
+			    void (*syncpt_thresh_work)(struct work_struct *))
 {
-	int i, err;
+	unsigned int i;
+	int err;
 
 	host1x_hw_intr_disable_all_syncpt_intrs(host);
 
@@ -106,18 +108,21 @@ static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm,
 }
 
 static void _host1x_intr_set_syncpt_threshold(struct host1x *host,
-	u32 id, u32 thresh)
+					      unsigned int id,
+					      u32 thresh)
 {
 	host1x_sync_writel(host, thresh, HOST1X_SYNC_SYNCPT_INT_THRESH(id));
 }
 
-static void _host1x_intr_enable_syncpt_intr(struct host1x *host, u32 id)
+static void _host1x_intr_enable_syncpt_intr(struct host1x *host,
+					    unsigned int id)
 {
 	host1x_sync_writel(host, BIT_MASK(id),
 		HOST1X_SYNC_SYNCPT_THRESH_INT_ENABLE_CPU0(BIT_WORD(id)));
 }
 
-static void _host1x_intr_disable_syncpt_intr(struct host1x *host, u32 id)
+static void _host1x_intr_disable_syncpt_intr(struct host1x *host,
+					     unsigned int id)
 {
 	host1x_sync_writel(host, BIT_MASK(id),
 		HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(BIT_WORD(id)));
@@ -127,8 +132,13 @@ static void _host1x_intr_disable_syncpt_intr(struct host1x *host, u32 id)
 
 static int _host1x_free_syncpt_irq(struct host1x *host)
 {
+	unsigned int i;
+
 	devm_free_irq(host->dev, host->intr_syncpt_irq, host);
-	flush_workqueue(host->intr_wq);
+
+	for (i = 0; i < host->info->nb_pts; i++)
+		cancel_work_sync(&host->syncpt[i].intr.work);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/host1x/hw/syncpt_hw.c b/drivers/gpu/host1x/hw/syncpt_hw.c
index 56e85395ac24..c93f74fcce72 100644
--- a/drivers/gpu/host1x/hw/syncpt_hw.c
+++ b/drivers/gpu/host1x/hw/syncpt_hw.c
@@ -26,8 +26,9 @@
  */
 static void syncpt_restore(struct host1x_syncpt *sp)
 {
+	u32 min = host1x_syncpt_read_min(sp);
 	struct host1x *host = sp->host;
-	int min = host1x_syncpt_read_min(sp);
+
 	host1x_sync_writel(host, min, HOST1X_SYNC_SYNCPT(sp->id));
 }
 
@@ -37,6 +38,7 @@ static void syncpt_restore(struct host1x_syncpt *sp)
 static void syncpt_restore_wait_base(struct host1x_syncpt *sp)
 {
 	struct host1x *host = sp->host;
+
 	host1x_sync_writel(host, sp->base_val,
 			   HOST1X_SYNC_SYNCPT_BASE(sp->id));
 }
@@ -47,6 +49,7 @@ static void syncpt_restore_wait_base(struct host1x_syncpt *sp)
 static void syncpt_read_wait_base(struct host1x_syncpt *sp)
 {
 	struct host1x *host = sp->host;
+
 	sp->base_val =
 		host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(sp->id));
 }
@@ -85,6 +88,7 @@ static int syncpt_cpu_incr(struct host1x_syncpt *sp)
 	if (!host1x_syncpt_client_managed(sp) &&
 	    host1x_syncpt_idle(sp))
 		return -EINVAL;
+
 	host1x_sync_writel(host, BIT_MASK(sp->id),
 			   HOST1X_SYNC_SYNCPT_CPU_INCR(reg_offset));
 	wmb();
@@ -95,10 +99,10 @@ static int syncpt_cpu_incr(struct host1x_syncpt *sp)
 /* remove a wait pointed to by patch_addr */
 static int syncpt_patch_wait(struct host1x_syncpt *sp, void *patch_addr)
 {
-	u32 override = host1x_class_host_wait_syncpt(
-		HOST1X_SYNCPT_RESERVED, 0);
+	u32 override = host1x_class_host_wait_syncpt(HOST1X_SYNCPT_RESERVED, 0);
 
 	*((u32 *)patch_addr) = override;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c
index 2491bf82e30c..8b4fad0ab35d 100644
--- a/drivers/gpu/host1x/intr.c
+++ b/drivers/gpu/host1x/intr.c
@@ -122,18 +122,20 @@ static void action_submit_complete(struct host1x_waitlist *waiter)
 static void action_wakeup(struct host1x_waitlist *waiter)
 {
 	wait_queue_head_t *wq = waiter->data;
+
 	wake_up(wq);
 }
 
 static void action_wakeup_interruptible(struct host1x_waitlist *waiter)
 {
 	wait_queue_head_t *wq = waiter->data;
+
 	wake_up_interruptible(wq);
 }
 
 typedef void (*action_handler)(struct host1x_waitlist *waiter);
 
-static action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = {
+static const action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = {
 	action_submit_complete,
 	action_wakeup,
 	action_wakeup_interruptible,
@@ -209,7 +211,7 @@ static void syncpt_thresh_work(struct work_struct *work)
 				host1x_syncpt_load(host->syncpt + id));
 }
 
-int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
+int host1x_intr_add_action(struct host1x *host, unsigned int id, u32 thresh,
 			   enum host1x_intr_action action, void *data,
 			   struct host1x_waitlist *waiter, void **ref)
 {
@@ -254,7 +256,7 @@ int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
 	return 0;
 }
 
-void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref)
+void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref)
 {
 	struct host1x_waitlist *waiter = ref;
 	struct host1x_syncpt *syncpt;
@@ -277,9 +279,6 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync)
 
 	mutex_init(&host->intr_mutex);
 	host->intr_syncpt_irq = irq_sync;
-	host->intr_wq = create_workqueue("host_syncpt");
-	if (!host->intr_wq)
-		return -ENOMEM;
 
 	for (id = 0; id < nb_pts; ++id) {
 		struct host1x_syncpt *syncpt = host->syncpt + id;
@@ -288,7 +287,7 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync)
 		INIT_LIST_HEAD(&syncpt->intr.wait_head);
 		snprintf(syncpt->intr.thresh_irq_name,
 			 sizeof(syncpt->intr.thresh_irq_name),
-			 "host1x_sp_%02d", id);
+			 "host1x_sp_%02u", id);
 	}
 
 	host1x_intr_start(host);
@@ -299,7 +298,6 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync)
 void host1x_intr_deinit(struct host1x *host)
 {
 	host1x_intr_stop(host);
-	destroy_workqueue(host->intr_wq);
 }
 
 void host1x_intr_start(struct host1x *host)
@@ -342,7 +340,7 @@ void host1x_intr_stop(struct host1x *host)
 		if (!list_empty(&syncpt[id].intr.wait_head)) {
 			/* output diagnostics */
 			mutex_unlock(&host->intr_mutex);
-			pr_warn("%s cannot stop syncpt intr id=%d\n",
+			pr_warn("%s cannot stop syncpt intr id=%u\n",
 				__func__, id);
 			return;
 		}
diff --git a/drivers/gpu/host1x/intr.h b/drivers/gpu/host1x/intr.h
index 2b8adf016a05..1370c2bb75b8 100644
--- a/drivers/gpu/host1x/intr.h
+++ b/drivers/gpu/host1x/intr.h
@@ -75,7 +75,7 @@ struct host1x_waitlist {
  *
  * This is a non-blocking api.
  */
-int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
+int host1x_intr_add_action(struct host1x *host, unsigned int id, u32 thresh,
 	enum host1x_intr_action action, void *data,
 	struct host1x_waitlist *waiter, void **ref);
 
@@ -84,7 +84,7 @@ int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
  * You must call this if you passed non-NULL as ref.
  * @ref the ref returned from host1x_intr_add_action()
  */
-void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref);
+void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref);
 
 /* Initialize host1x sync point interrupt */
 int host1x_intr_init(struct host1x *host, unsigned int irq_sync);
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index b4515d544039..a91b7c4a6110 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -161,7 +161,7 @@ static int do_waitchks(struct host1x_job *job, struct host1x *host,
 
 		if (host1x_syncpt_is_expired(sp, wait->thresh)) {
 			dev_dbg(host->dev,
-				"drop WAIT id %d (%s) thresh 0x%x, min 0x%x\n",
+				"drop WAIT id %u (%s) thresh 0x%x, min 0x%x\n",
 				wait->syncpt_id, sp->name, wait->thresh,
 				host1x_syncpt_read_min(sp));
 
@@ -464,6 +464,7 @@ static inline int copy_gathers(struct host1x_job *job, struct device *dev)
 
 	for (i = 0; i < job->num_gathers; i++) {
 		struct host1x_job_gather *g = &job->gathers[i];
+
 		size += g->words * sizeof(u32);
 	}
 
@@ -514,6 +515,7 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev)
 	bitmap_zero(waitchk_mask, host1x_syncpt_nb_pts(host));
 	for (i = 0; i < job->num_waitchk; i++) {
 		u32 syncpt_id = job->waitchk[i].syncpt_id;
+
 		if (syncpt_id < host1x_syncpt_nb_pts(host))
 			set_bit(syncpt_id, waitchk_mask);
 	}
@@ -571,14 +573,16 @@ void host1x_job_unpin(struct host1x_job *job)
 
 	for (i = 0; i < job->num_unpins; i++) {
 		struct host1x_job_unpin_data *unpin = &job->unpins[i];
+
 		host1x_bo_unpin(unpin->bo, unpin->sgt);
 		host1x_bo_put(unpin->bo);
 	}
+
 	job->num_unpins = 0;
 
 	if (job->gather_copy_size)
 		dma_free_wc(job->channel->dev, job->gather_copy_size,
-		            job->gather_copy_mapped, job->gather_copy);
+			    job->gather_copy_mapped, job->gather_copy);
 }
 EXPORT_SYMBOL(host1x_job_unpin);
 
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 6b7fdc1e2ed0..95589328ad52 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
@@ -73,7 +73,7 @@ static struct host1x_syncpt *host1x_syncpt_alloc(struct host1x *host,
 			return NULL;
 	}
 
-	name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id,
+	name = kasprintf(GFP_KERNEL, "%02u-%s", sp->id,
 			dev ? dev_name(dev) : NULL);
 	if (!name)
 		return NULL;
@@ -110,12 +110,14 @@ EXPORT_SYMBOL(host1x_syncpt_incr_max);
 void host1x_syncpt_restore(struct host1x *host)
 {
 	struct host1x_syncpt *sp_base = host->syncpt;
-	u32 i;
+	unsigned int i;
 
 	for (i = 0; i < host1x_syncpt_nb_pts(host); i++)
 		host1x_hw_syncpt_restore(host, sp_base + i);
+
 	for (i = 0; i < host1x_syncpt_nb_bases(host); i++)
 		host1x_hw_syncpt_restore_wait_base(host, sp_base + i);
+
 	wmb();
 }
 
@@ -126,7 +128,7 @@ void host1x_syncpt_restore(struct host1x *host)
 void host1x_syncpt_save(struct host1x *host)
 {
 	struct host1x_syncpt *sp_base = host->syncpt;
-	u32 i;
+	unsigned int i;
 
 	for (i = 0; i < host1x_syncpt_nb_pts(host); i++) {
 		if (host1x_syncpt_client_managed(sp_base + i))
@@ -146,6 +148,7 @@ void host1x_syncpt_save(struct host1x *host)
 u32 host1x_syncpt_load(struct host1x_syncpt *sp)
 {
 	u32 val;
+
 	val = host1x_hw_syncpt_load(sp->host, sp);
 	trace_host1x_syncpt_load_min(sp->id, val);
 
@@ -157,10 +160,9 @@ u32 host1x_syncpt_load(struct host1x_syncpt *sp)
  */
 u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp)
 {
-	u32 val;
 	host1x_hw_syncpt_load_wait_base(sp->host, sp);
-	val = sp->base_val;
-	return val;
+
+	return sp->base_val;
 }
 
 /*
@@ -179,6 +181,7 @@ EXPORT_SYMBOL(host1x_syncpt_incr);
 static bool syncpt_load_min_is_expired(struct host1x_syncpt *sp, u32 thresh)
 {
 	host1x_hw_syncpt_load(sp->host, sp);
+
 	return host1x_syncpt_is_expired(sp, thresh);
 }
 
@@ -186,7 +189,7 @@ static bool syncpt_load_min_is_expired(struct host1x_syncpt *sp, u32 thresh)
  * Main entrypoint for syncpoint value waits.
  */
 int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
-			u32 *value)
+		       u32 *value)
 {
 	DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq);
 	void *ref;
@@ -201,6 +204,7 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
 	if (host1x_syncpt_is_expired(sp, thresh)) {
 		if (value)
 			*value = host1x_syncpt_load(sp);
+
 		return 0;
 	}
 
@@ -209,6 +213,7 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
 	if (host1x_syncpt_is_expired(sp, thresh)) {
 		if (value)
 			*value = val;
+
 		goto done;
 	}
 
@@ -239,32 +244,42 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
 	/* wait for the syncpoint, or timeout, or signal */
 	while (timeout) {
 		long check = min_t(long, SYNCPT_CHECK_PERIOD, timeout);
-		int remain = wait_event_interruptible_timeout(wq,
+		int remain;
+
+		remain = wait_event_interruptible_timeout(wq,
 				syncpt_load_min_is_expired(sp, thresh),
 				check);
 		if (remain > 0 || host1x_syncpt_is_expired(sp, thresh)) {
 			if (value)
 				*value = host1x_syncpt_load(sp);
+
 			err = 0;
+
 			break;
 		}
+
 		if (remain < 0) {
 			err = remain;
 			break;
 		}
+
 		timeout -= check;
+
 		if (timeout && check_count <= MAX_STUCK_CHECK_COUNT) {
 			dev_warn(sp->host->dev,
-				"%s: syncpoint id %d (%s) stuck waiting %d, timeout=%ld\n",
+				"%s: syncpoint id %u (%s) stuck waiting %d, timeout=%ld\n",
 				 current->comm, sp->id, sp->name,
 				 thresh, timeout);
 
 			host1x_debug_dump_syncpts(sp->host);
+
 			if (check_count == MAX_STUCK_CHECK_COUNT)
 				host1x_debug_dump(sp->host);
+
 			check_count++;
 		}
 	}
+
 	host1x_intr_put_ref(sp->host, sp->id, ref);
 
 done:
@@ -279,7 +294,9 @@ bool host1x_syncpt_is_expired(struct host1x_syncpt *sp, u32 thresh)
 {
 	u32 current_val;
 	u32 future_val;
+
 	smp_rmb();
+
 	current_val = (u32)atomic_read(&sp->min_val);
 	future_val = (u32)atomic_read(&sp->max_val);
 
@@ -341,14 +358,14 @@ int host1x_syncpt_init(struct host1x *host)
 {
 	struct host1x_syncpt_base *bases;
 	struct host1x_syncpt *syncpt;
-	int i;
+	unsigned int i;
 
-	syncpt = devm_kzalloc(host->dev, sizeof(*syncpt) * host->info->nb_pts,
+	syncpt = devm_kcalloc(host->dev, host->info->nb_pts, sizeof(*syncpt),
 			      GFP_KERNEL);
 	if (!syncpt)
 		return -ENOMEM;
 
-	bases = devm_kzalloc(host->dev, sizeof(*bases) * host->info->nb_bases,
+	bases = devm_kcalloc(host->dev, host->info->nb_bases, sizeof(*bases),
 			     GFP_KERNEL);
 	if (!bases)
 		return -ENOMEM;
@@ -378,6 +395,7 @@ struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
 					    unsigned long flags)
 {
 	struct host1x *host = dev_get_drvdata(dev->parent);
+
 	return host1x_syncpt_alloc(host, dev, flags);
 }
 EXPORT_SYMBOL(host1x_syncpt_request);
@@ -398,8 +416,9 @@ EXPORT_SYMBOL(host1x_syncpt_free);
 
 void host1x_syncpt_deinit(struct host1x *host)
 {
-	int i;
 	struct host1x_syncpt *sp = host->syncpt;
+	unsigned int i;
+
 	for (i = 0; i < host->info->nb_pts; i++, sp++)
 		kfree(sp->name);
 }
@@ -407,10 +426,11 @@ void host1x_syncpt_deinit(struct host1x *host)
 /*
  * Read max. It indicates how many operations there are in queue, either in
  * channel or in a software thread.
- * */
+ */
 u32 host1x_syncpt_read_max(struct host1x_syncpt *sp)
 {
 	smp_rmb();
+
 	return (u32)atomic_read(&sp->max_val);
 }
 EXPORT_SYMBOL(host1x_syncpt_read_max);
@@ -421,6 +441,7 @@ EXPORT_SYMBOL(host1x_syncpt_read_max);
 u32 host1x_syncpt_read_min(struct host1x_syncpt *sp)
 {
 	smp_rmb();
+
 	return (u32)atomic_read(&sp->min_val);
 }
 EXPORT_SYMBOL(host1x_syncpt_read_min);
@@ -431,25 +452,26 @@ u32 host1x_syncpt_read(struct host1x_syncpt *sp)
 }
 EXPORT_SYMBOL(host1x_syncpt_read);
 
-int host1x_syncpt_nb_pts(struct host1x *host)
+unsigned int host1x_syncpt_nb_pts(struct host1x *host)
 {
 	return host->info->nb_pts;
 }
 
-int host1x_syncpt_nb_bases(struct host1x *host)
+unsigned int host1x_syncpt_nb_bases(struct host1x *host)
 {
 	return host->info->nb_bases;
 }
 
-int host1x_syncpt_nb_mlocks(struct host1x *host)
+unsigned int host1x_syncpt_nb_mlocks(struct host1x *host)
 {
 	return host->info->nb_mlocks;
 }
 
-struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id)
+struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, unsigned int id)
 {
 	if (host->info->nb_pts < id)
 		return NULL;
+
 	return host->syncpt + id;
 }
 EXPORT_SYMBOL(host1x_syncpt_get);
diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h
index 9056465ecd3f..f719205105ac 100644
--- a/drivers/gpu/host1x/syncpt.h
+++ b/drivers/gpu/host1x/syncpt.h
@@ -37,7 +37,7 @@ struct host1x_syncpt_base {
 };
 
 struct host1x_syncpt {
-	int id;
+	unsigned int id;
 	atomic_t min_val;
 	atomic_t max_val;
 	u32 base_val;
@@ -58,13 +58,13 @@ int host1x_syncpt_init(struct host1x *host);
 void host1x_syncpt_deinit(struct host1x *host);
 
 /* Return number of sync point supported. */
-int host1x_syncpt_nb_pts(struct host1x *host);
+unsigned int host1x_syncpt_nb_pts(struct host1x *host);
 
 /* Return number of wait bases supported. */
-int host1x_syncpt_nb_bases(struct host1x *host);
+unsigned int host1x_syncpt_nb_bases(struct host1x *host);
 
 /* Return number of mlocks supported. */
-int host1x_syncpt_nb_mlocks(struct host1x *host);
+unsigned int host1x_syncpt_nb_mlocks(struct host1x *host);
 
 /*
  * Check sync point sanity. If max is larger than min, there have too many
diff --git a/drivers/gpu/ipu-v3/ipu-dc.c b/drivers/gpu/ipu-v3/ipu-dc.c
index 2f29780e7c68..659475c1e44a 100644
--- a/drivers/gpu/ipu-v3/ipu-dc.c
+++ b/drivers/gpu/ipu-v3/ipu-dc.c
@@ -150,6 +150,9 @@ static void dc_write_tmpl(struct ipu_dc *dc, int word, u32 opcode, u32 operand,
 static int ipu_bus_format_to_map(u32 fmt)
 {
 	switch (fmt) {
+	default:
+		WARN_ON(1);
+		/* fall-through */
 	case MEDIA_BUS_FMT_RGB888_1X24:
 		return IPU_DC_MAP_RGB24;
 	case MEDIA_BUS_FMT_RGB565_1X16:
@@ -162,8 +165,6 @@ static int ipu_bus_format_to_map(u32 fmt)
 		return IPU_DC_MAP_LVDS666;
 	case MEDIA_BUS_FMT_BGR888_1X24:
 		return IPU_DC_MAP_BGR24;
-	default:
-		return -EINVAL;
 	}
 }
 
@@ -178,10 +179,6 @@ int ipu_dc_init_sync(struct ipu_dc *dc, struct ipu_di *di, bool interlaced,
 	dc->di = ipu_di_get_num(di);
 
 	map = ipu_bus_format_to_map(bus_format);
-	if (map < 0) {
-		dev_dbg(priv->dev, "IPU_DISP: No MAP\n");
-		return map;
-	}
 
 	/*
 	 * In interlaced mode we need more counters to create the asymmetric
diff --git a/drivers/gpu/ipu-v3/ipu-di.c b/drivers/gpu/ipu-v3/ipu-di.c
index 359268e3a166..a8d87ddd8a17 100644
--- a/drivers/gpu/ipu-v3/ipu-di.c
+++ b/drivers/gpu/ipu-v3/ipu-di.c
@@ -572,9 +572,6 @@ int ipu_di_init_sync_panel(struct ipu_di *di, struct ipu_di_signal_cfg *sig)
 	dev_dbg(di->ipu->dev, "disp %d: panel size = %d x %d\n",
 		di->id, sig->mode.hactive, sig->mode.vactive);
 
-	if ((sig->mode.vsync_len == 0) || (sig->mode.hsync_len == 0))
-		return -EINVAL;
-
 	dev_dbg(di->ipu->dev, "Clocks: IPU %luHz DI %luHz Needed %luHz\n",
 		clk_get_rate(di->clk_ipu),
 		clk_get_rate(di->clk_di),
diff --git a/drivers/gpu/ipu-v3/ipu-dmfc.c b/drivers/gpu/ipu-v3/ipu-dmfc.c
index 837b1ec22800..42705bb5aaa3 100644
--- a/drivers/gpu/ipu-v3/ipu-dmfc.c
+++ b/drivers/gpu/ipu-v3/ipu-dmfc.c
@@ -45,17 +45,6 @@
 #define DMFC_DP_CHAN_6B_24		16
 #define DMFC_DP_CHAN_6F_29		24
 
-#define DMFC_FIFO_SIZE_64		(3 << 3)
-#define DMFC_FIFO_SIZE_128		(2 << 3)
-#define DMFC_FIFO_SIZE_256		(1 << 3)
-#define DMFC_FIFO_SIZE_512		(0 << 3)
-
-#define DMFC_SEGMENT(x)			((x & 0x7) << 0)
-#define DMFC_BURSTSIZE_128		(0 << 6)
-#define DMFC_BURSTSIZE_64		(1 << 6)
-#define DMFC_BURSTSIZE_32		(2 << 6)
-#define DMFC_BURSTSIZE_16		(3 << 6)
-
 struct dmfc_channel_data {
 	int		ipu_channel;
 	unsigned long	channel_reg;
@@ -104,9 +93,6 @@ struct ipu_dmfc_priv;
 
 struct dmfc_channel {
 	unsigned			slots;
-	unsigned			slotmask;
-	unsigned			segment;
-	int				burstsize;
 	struct ipu_soc			*ipu;
 	struct ipu_dmfc_priv		*priv;
 	const struct dmfc_channel_data	*data;
@@ -117,7 +103,6 @@ struct ipu_dmfc_priv {
 	struct device *dev;
 	struct dmfc_channel channels[DMFC_NUM_CHANNELS];
 	struct mutex mutex;
-	unsigned long bandwidth_per_slot;
 	void __iomem *base;
 	int use_count;
 };
@@ -172,184 +157,6 @@ void ipu_dmfc_disable_channel(struct dmfc_channel *dmfc)
 }
 EXPORT_SYMBOL_GPL(ipu_dmfc_disable_channel);
 
-static int ipu_dmfc_setup_channel(struct dmfc_channel *dmfc, int slots,
-		int segment, int burstsize)
-{
-	struct ipu_dmfc_priv *priv = dmfc->priv;
-	u32 val, field;
-
-	dev_dbg(priv->dev,
-			"dmfc: using %d slots starting from segment %d for IPU channel %d\n",
-			slots, segment, dmfc->data->ipu_channel);
-
-	switch (slots) {
-	case 1:
-		field = DMFC_FIFO_SIZE_64;
-		break;
-	case 2:
-		field = DMFC_FIFO_SIZE_128;
-		break;
-	case 4:
-		field = DMFC_FIFO_SIZE_256;
-		break;
-	case 8:
-		field = DMFC_FIFO_SIZE_512;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	switch (burstsize) {
-	case 16:
-		field |= DMFC_BURSTSIZE_16;
-		break;
-	case 32:
-		field |= DMFC_BURSTSIZE_32;
-		break;
-	case 64:
-		field |= DMFC_BURSTSIZE_64;
-		break;
-	case 128:
-		field |= DMFC_BURSTSIZE_128;
-		break;
-	}
-
-	field |= DMFC_SEGMENT(segment);
-
-	val = readl(priv->base + dmfc->data->channel_reg);
-
-	val &= ~(0xff << dmfc->data->shift);
-	val |= field << dmfc->data->shift;
-
-	writel(val, priv->base + dmfc->data->channel_reg);
-
-	dmfc->slots = slots;
-	dmfc->segment = segment;
-	dmfc->burstsize = burstsize;
-	dmfc->slotmask = ((1 << slots) - 1) << segment;
-
-	return 0;
-}
-
-static int dmfc_bandwidth_to_slots(struct ipu_dmfc_priv *priv,
-		unsigned long bandwidth)
-{
-	int slots = 1;
-
-	while (slots * priv->bandwidth_per_slot < bandwidth)
-		slots *= 2;
-
-	return slots;
-}
-
-static int dmfc_find_slots(struct ipu_dmfc_priv *priv, int slots)
-{
-	unsigned slotmask_need, slotmask_used = 0;
-	int i, segment = 0;
-
-	slotmask_need = (1 << slots) - 1;
-
-	for (i = 0; i < DMFC_NUM_CHANNELS; i++)
-		slotmask_used |= priv->channels[i].slotmask;
-
-	while (slotmask_need <= 0xff) {
-		if (!(slotmask_used & slotmask_need))
-			return segment;
-
-		slotmask_need <<= 1;
-		segment++;
-	}
-
-	return -EBUSY;
-}
-
-void ipu_dmfc_free_bandwidth(struct dmfc_channel *dmfc)
-{
-	struct ipu_dmfc_priv *priv = dmfc->priv;
-	int i;
-
-	dev_dbg(priv->dev, "dmfc: freeing %d slots starting from segment %d\n",
-			dmfc->slots, dmfc->segment);
-
-	mutex_lock(&priv->mutex);
-
-	if (!dmfc->slots)
-		goto out;
-
-	dmfc->slotmask = 0;
-	dmfc->slots = 0;
-	dmfc->segment = 0;
-
-	for (i = 0; i < DMFC_NUM_CHANNELS; i++)
-		priv->channels[i].slotmask = 0;
-
-	for (i = 0; i < DMFC_NUM_CHANNELS; i++) {
-		if (priv->channels[i].slots > 0) {
-			priv->channels[i].segment =
-				dmfc_find_slots(priv, priv->channels[i].slots);
-			priv->channels[i].slotmask =
-				((1 << priv->channels[i].slots) - 1) <<
-				priv->channels[i].segment;
-		}
-	}
-
-	for (i = 0; i < DMFC_NUM_CHANNELS; i++) {
-		if (priv->channels[i].slots > 0)
-			ipu_dmfc_setup_channel(&priv->channels[i],
-					priv->channels[i].slots,
-					priv->channels[i].segment,
-					priv->channels[i].burstsize);
-	}
-out:
-	mutex_unlock(&priv->mutex);
-}
-EXPORT_SYMBOL_GPL(ipu_dmfc_free_bandwidth);
-
-int ipu_dmfc_alloc_bandwidth(struct dmfc_channel *dmfc,
-		unsigned long bandwidth_pixel_per_second, int burstsize)
-{
-	struct ipu_dmfc_priv *priv = dmfc->priv;
-	int slots = dmfc_bandwidth_to_slots(priv, bandwidth_pixel_per_second);
-	int segment = -1, ret = 0;
-
-	dev_dbg(priv->dev, "dmfc: trying to allocate %ldMpixel/s for IPU channel %d\n",
-			bandwidth_pixel_per_second / 1000000,
-			dmfc->data->ipu_channel);
-
-	ipu_dmfc_free_bandwidth(dmfc);
-
-	mutex_lock(&priv->mutex);
-
-	if (slots > 8) {
-		ret = -EBUSY;
-		goto out;
-	}
-
-	/* For the MEM_BG channel, first try to allocate twice the slots */
-	if (dmfc->data->ipu_channel == IPUV3_CHANNEL_MEM_BG_SYNC)
-		segment = dmfc_find_slots(priv, slots * 2);
-	else if (slots < 2)
-		/* Always allocate at least 128*4 bytes (2 slots) */
-		slots = 2;
-
-	if (segment >= 0)
-		slots *= 2;
-	else
-		segment = dmfc_find_slots(priv, slots);
-	if (segment < 0) {
-		ret = -EBUSY;
-		goto out;
-	}
-
-	ipu_dmfc_setup_channel(dmfc, slots, segment, burstsize);
-
-out:
-	mutex_unlock(&priv->mutex);
-
-	return ret;
-}
-EXPORT_SYMBOL_GPL(ipu_dmfc_alloc_bandwidth);
-
 void ipu_dmfc_config_wait4eot(struct dmfc_channel *dmfc, int width)
 {
 	struct ipu_dmfc_priv *priv = dmfc->priv;
@@ -384,7 +191,6 @@ EXPORT_SYMBOL_GPL(ipu_dmfc_get);
 
 void ipu_dmfc_put(struct dmfc_channel *dmfc)
 {
-	ipu_dmfc_free_bandwidth(dmfc);
 }
 EXPORT_SYMBOL_GPL(ipu_dmfc_put);
 
@@ -412,20 +218,15 @@ int ipu_dmfc_init(struct ipu_soc *ipu, struct device *dev, unsigned long base,
 		priv->channels[i].priv = priv;
 		priv->channels[i].ipu = ipu;
 		priv->channels[i].data = &dmfcdata[i];
-	}
-
-	writel(0x0, priv->base + DMFC_WR_CHAN);
-	writel(0x0, priv->base + DMFC_DP_CHAN);
 
-	/*
-	 * We have a total bandwidth of clkrate * 4pixel divided
-	 * into 8 slots.
-	 */
-	priv->bandwidth_per_slot = clk_get_rate(ipu_clk) * 4 / 8;
-
-	dev_dbg(dev, "dmfc: 8 slots with %ldMpixel/s bandwidth each\n",
-			priv->bandwidth_per_slot / 1000000);
+		if (dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_BG_SYNC ||
+		    dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_FG_SYNC ||
+		    dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_DC_SYNC)
+			priv->channels[i].slots = 2;
+	}
 
+	writel(0x00000050, priv->base + DMFC_WR_CHAN);
+	writel(0x00005654, priv->base + DMFC_DP_CHAN);
 	writel(0x202020f6, priv->base + DMFC_WR_CHAN_DEF);
 	writel(0x2020f6f6, priv->base + DMFC_DP_CHAN_DEF);
 	writel(0x00000003, priv->base + DMFC_GENERAL1);