Merge tag 'drm-for-v4.15' of git://people.freedesktop.org/~airlied/linux

Pull drm updates from Dave Airlie:
 "This is the main drm pull request for v4.15.

  Core:
   - Atomic object lifetime fixes
   - Atomic iterator improvements
   - Sparse/smatch fixes
   - Legacy kms ioctls to be interruptible
   - EDID override improvements
   - fb/gem helper cleanups
   - Simple outreachy patches
   - Documentation improvements
   - Fix dma-buf rcu races
   - DRM mode object leasing for improving VR use cases.
   - vgaarb improvements for non-x86 platforms.

  New driver:
   - tve200: Faraday Technology TVE200 block.

     This "TV Encoder" encodes a ITU-T BT.656 stream and can be found in
     the StorLink SL3516 (later Cortina Systems CS3516) as well as the
     Grain Media GM8180.

  New bridges:
   - SiI9234 support

  New panels:
   - S6E63J0X03, OTM8009A, Seiko 43WVF1G, 7" rpi touch panel, Toshiba
     LT089AC19000, Innolux AT043TN24

  i915:
   - Remove Coffeelake from alpha support
   - Cannonlake workarounds
   - Infoframe refactoring for DisplayPort
   - VBT updates
   - DisplayPort vswing/emph/buffer translation refactoring
   - CCS fixes
   - Restore GPU clock boost on missed vblanks
   - Scatter list updates for userptr allocations
   - Gen9+ transition watermarks
   - Display IPC (Isochronous Priority Control)
   - Private PAT management
   - GVT: improved error handling and pci config sanitizing
   - Execlist refactoring
   - Transparent Huge Page support
   - User defined priorities support
   - HuC/GuC firmware refactoring
   - DP MST fixes
   - eDP power sequencing fixes
   - Use RCU instead of stop_machine
   - PSR state tracking support
   - Eviction fixes
   - BDW DP aux channel timeout fixes
   - LSPCON fixes
   - Cannonlake PLL fixes

  amdgpu:
   - Per VM BO support
   - Powerplay cleanups
   - CI powerplay support
   - PASID mgr for kfd
   - SR-IOV fixes
   - initial GPU reset for vega10
   - Prime mmap support
   - TTM updates
   - Clock query interface for Raven
   - Fence to handle ioctl
   - UVD encode ring support on Polaris
   - Transparent huge page DMA support
   - Compute LRU pipe tweaks
   - BO flag to allow buffers to opt out of implicit sync
   - CTX priority setting API
   - VRAM lost infrastructure plumbing

  qxl:
   - fix flicker since atomic rework

  amdkfd:
   - Further improvements from internal AMD tree
   - Usermode events
   - Drop radeon support

  nouveau:
   - Pascal temperature sensor support
   - Improved BAR2 handling
   - MMU rework to support Pascal MMU

  exynos:
   - Improved HDMI/mixer support
   - HDMI audio interface support

  tegra:
   - Prep work for tegra186
   - Cleanup/fixes

  msm:
   - Preemption support for a5xx
   - Display fixes for 8x96 (snapdragon 820)
   - Async cursor plane fixes
   - FW loading rework
   - GPU debugging improvements

  vc4:
   - Prep for DSI panels
   - fix T-format tiling scanout
   - New madvise ioctl

  Rockchip:
   - LVDS support

  omapdrm:
   - omap4 HDMI CEC support

  etnaviv:
   - GPU performance counters groundwork

  sun4i:
   - refactor driver load + TCON backend
   - HDMI improvements
   - A31 support
   - Misc fixes

  udl:
   - Probe/EDID read fixes.

  tilcdc:
   - Misc fixes.

  pl111:
   - Support more variants

  adv7511:
   - Improve EDID handling.
   - HDMI CEC support

  sii8620:
   - Add remote control support"

* tag 'drm-for-v4.15' of git://people.freedesktop.org/~airlied/linux: (1480 commits)
  drm/rockchip: analogix_dp: Use mutex rather than spinlock
  drm/mode_object: fix documentation for object lookups.
  drm/i915: Reorder context-close to avoid calling i915_vma_close() under RCU
  drm/i915: Move init_clock_gating() back to where it was
  drm/i915: Prune the reservation shared fence array
  drm/i915: Idle the GPU before shinking everything
  drm/i915: Lock llist_del_first() vs llist_del_all()
  drm/i915: Calculate ironlake intermediate watermarks correctly, v2.
  drm/i915: Disable lazy PPGTT page table optimization for vGPU
  drm/i915/execlists: Remove the priority "optimisation"
  drm/i915: Filter out spurious execlists context-switch interrupts
  drm/amdgpu: use irq-safe lock for kiq->ring_lock
  drm/amdgpu: bypass lru touch for KIQ ring submission
  drm/amdgpu: Potential uninitialized variable in amdgpu_vm_update_directories()
  drm/amdgpu: potential uninitialized variable in amdgpu_vce_ring_parse_cs()
  drm/amd/powerplay: initialize a variable before using it
  drm/amd/powerplay: suppress KASAN out of bounds warning in vega10_populate_all_memory_levels
  drm/amd/amdgpu: fix evicted VRAM bo adjudgement condition
  drm/vblank: Tune drm_crtc_accurate_vblank_count() WARN down to a debug
  drm/rockchip: add CONFIG_OF dependency for lvds
  ...

18 files changed, 1111 insertions, 328 deletions

diff --git a/drivers/gpu/drm/etnaviv/Kconfig b/drivers/gpu/drm/etnaviv/Kconfig
index 38b477b5fbf9..a29b8f59eb15 100644
--- a/drivers/gpu/drm/etnaviv/Kconfig
+++ b/drivers/gpu/drm/etnaviv/Kconfig
@@ -7,8 +7,6 @@ config DRM_ETNAVIV
 	select SHMEM
 	select SYNC_FILE
 	select TMPFS
-	select IOMMU_API
-	select IOMMU_SUPPORT
 	select WANT_DEV_COREDUMP
 	select CMA if HAVE_DMA_CONTIGUOUS
 	select DMA_CMA if HAVE_DMA_CONTIGUOUS
diff --git a/drivers/gpu/drm/etnaviv/Makefile b/drivers/gpu/drm/etnaviv/Makefile
index ab3f551831d7..1281c8d4fae5 100644
--- a/drivers/gpu/drm/etnaviv/Makefile
+++ b/drivers/gpu/drm/etnaviv/Makefile
@@ -11,6 +11,7 @@ etnaviv-y := \
 	etnaviv_gpu.o \
 	etnaviv_iommu_v2.o \
 	etnaviv_iommu.o \
-	etnaviv_mmu.o
+	etnaviv_mmu.o \
+	etnaviv_perfmon.o
 
 obj-$(CONFIG_DRM_ETNAVIV)	+= etnaviv.o
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
index ed9588f36bc9..9e7098e3207f 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c
@@ -250,6 +250,42 @@ void etnaviv_buffer_end(struct etnaviv_gpu *gpu)
 	}
 }
 
+/* Append a 'sync point' to the ring buffer. */
+void etnaviv_sync_point_queue(struct etnaviv_gpu *gpu, unsigned int event)
+{
+	struct etnaviv_cmdbuf *buffer = gpu->buffer;
+	unsigned int waitlink_offset = buffer->user_size - 16;
+	u32 dwords, target;
+
+	/*
+	 * We need at most 3 dwords in the return target:
+	 * 1 event + 1 end + 1 wait + 1 link.
+	 */
+	dwords = 4;
+	target = etnaviv_buffer_reserve(gpu, buffer, dwords);
+
+	/* Signal sync point event */
+	CMD_LOAD_STATE(buffer, VIVS_GL_EVENT, VIVS_GL_EVENT_EVENT_ID(event) |
+		       VIVS_GL_EVENT_FROM_PE);
+
+	/* Stop the FE to 'pause' the GPU */
+	CMD_END(buffer);
+
+	/* Append waitlink */
+	CMD_WAIT(buffer);
+	CMD_LINK(buffer, 2, etnaviv_cmdbuf_get_va(buffer) +
+			    buffer->user_size - 4);
+
+	/*
+	 * Kick off the 'sync point' command by replacing the previous
+	 * WAIT with a link to the address in the ring buffer.
+	 */
+	etnaviv_buffer_replace_wait(buffer, waitlink_offset,
+				    VIV_FE_LINK_HEADER_OP_LINK |
+				    VIV_FE_LINK_HEADER_PREFETCH(dwords),
+				    target);
+}
+
 /* Append a command buffer to the ring buffer. */
 void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, unsigned int event,
 	struct etnaviv_cmdbuf *cmdbuf)
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.c b/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.c
index 633e0f07cbac..66ac79558bbd 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.c
@@ -19,6 +19,7 @@
 #include "etnaviv_cmdbuf.h"
 #include "etnaviv_gpu.h"
 #include "etnaviv_mmu.h"
+#include "etnaviv_perfmon.h"
 
 #define SUBALLOC_SIZE		SZ_256K
 #define SUBALLOC_GRANULE	SZ_4K
@@ -87,9 +88,10 @@ void etnaviv_cmdbuf_suballoc_destroy(struct etnaviv_cmdbuf_suballoc *suballoc)
 
 struct etnaviv_cmdbuf *
 etnaviv_cmdbuf_new(struct etnaviv_cmdbuf_suballoc *suballoc, u32 size,
-		   size_t nr_bos)
+		   size_t nr_bos, size_t nr_pmrs)
 {
 	struct etnaviv_cmdbuf *cmdbuf;
+	struct etnaviv_perfmon_request *pmrs;
 	size_t sz = size_vstruct(nr_bos, sizeof(cmdbuf->bo_map[0]),
 				 sizeof(*cmdbuf));
 	int granule_offs, order, ret;
@@ -98,6 +100,12 @@ etnaviv_cmdbuf_new(struct etnaviv_cmdbuf_suballoc *suballoc, u32 size,
 	if (!cmdbuf)
 		return NULL;
 
+	sz = sizeof(*pmrs) * nr_pmrs;
+	pmrs = kzalloc(sz, GFP_KERNEL);
+	if (!pmrs)
+		goto out_free_cmdbuf;
+
+	cmdbuf->pmrs = pmrs;
 	cmdbuf->suballoc = suballoc;
 	cmdbuf->size = size;
 
@@ -124,6 +132,10 @@ retry:
 	cmdbuf->vaddr = suballoc->vaddr + cmdbuf->suballoc_offset;
 
 	return cmdbuf;
+
+out_free_cmdbuf:
+	kfree(cmdbuf);
+	return NULL;
 }
 
 void etnaviv_cmdbuf_free(struct etnaviv_cmdbuf *cmdbuf)
@@ -139,6 +151,7 @@ void etnaviv_cmdbuf_free(struct etnaviv_cmdbuf *cmdbuf)
 	suballoc->free_space = 1;
 	mutex_unlock(&suballoc->lock);
 	wake_up_all(&suballoc->free_event);
+	kfree(cmdbuf->pmrs);
 	kfree(cmdbuf);
 }
 
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.h b/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.h
index 80d78076c679..b6348b9f2a9d 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_cmdbuf.h
@@ -21,6 +21,7 @@
 
 struct etnaviv_gpu;
 struct etnaviv_cmdbuf_suballoc;
+struct etnaviv_perfmon_request;
 
 struct etnaviv_cmdbuf {
 	/* suballocator this cmdbuf is allocated from */
@@ -38,6 +39,9 @@ struct etnaviv_cmdbuf {
 	u32 exec_state;
 	/* per GPU in-flight list */
 	struct list_head node;
+	/* perfmon requests */
+	unsigned int nr_pmrs;
+	struct etnaviv_perfmon_request *pmrs;
 	/* BOs attached to this command buffer */
 	unsigned int nr_bos;
 	struct etnaviv_vram_mapping *bo_map[0];
@@ -49,7 +53,7 @@ void etnaviv_cmdbuf_suballoc_destroy(struct etnaviv_cmdbuf_suballoc *suballoc);
 
 struct etnaviv_cmdbuf *
 etnaviv_cmdbuf_new(struct etnaviv_cmdbuf_suballoc *suballoc, u32 size,
-		   size_t nr_bos);
+		   size_t nr_bos, size_t nr_pmrs);
 void etnaviv_cmdbuf_free(struct etnaviv_cmdbuf *cmdbuf);
 
 u32 etnaviv_cmdbuf_get_va(struct etnaviv_cmdbuf *buf);
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.c b/drivers/gpu/drm/etnaviv/etnaviv_drv.c
index 2cb4773823c2..491eddf9b150 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_drv.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.c
@@ -23,6 +23,7 @@
 #include "etnaviv_gpu.h"
 #include "etnaviv_gem.h"
 #include "etnaviv_mmu.h"
+#include "etnaviv_perfmon.h"
 
 #ifdef CONFIG_DRM_ETNAVIV_REGISTER_LOGGING
 static bool reglog;
@@ -451,6 +452,46 @@ static int etnaviv_ioctl_gem_wait(struct drm_device *dev, void *data,
 	return ret;
 }
 
+static int etnaviv_ioctl_pm_query_dom(struct drm_device *dev, void *data,
+	struct drm_file *file)
+{
+	struct etnaviv_drm_private *priv = dev->dev_private;
+	struct drm_etnaviv_pm_domain *args = data;
+	struct etnaviv_gpu *gpu;
+
+	/* reject as long as the feature isn't stable */
+	return -EINVAL;
+
+	if (args->pipe >= ETNA_MAX_PIPES)
+		return -EINVAL;
+
+	gpu = priv->gpu[args->pipe];
+	if (!gpu)
+		return -ENXIO;
+
+	return etnaviv_pm_query_dom(gpu, args);
+}
+
+static int etnaviv_ioctl_pm_query_sig(struct drm_device *dev, void *data,
+	struct drm_file *file)
+{
+	struct etnaviv_drm_private *priv = dev->dev_private;
+	struct drm_etnaviv_pm_signal *args = data;
+	struct etnaviv_gpu *gpu;
+
+	/* reject as long as the feature isn't stable */
+	return -EINVAL;
+
+	if (args->pipe >= ETNA_MAX_PIPES)
+		return -EINVAL;
+
+	gpu = priv->gpu[args->pipe];
+	if (!gpu)
+		return -ENXIO;
+
+	return etnaviv_pm_query_sig(gpu, args);
+}
+
 static const struct drm_ioctl_desc etnaviv_ioctls[] = {
 #define ETNA_IOCTL(n, func, flags) \
 	DRM_IOCTL_DEF_DRV(ETNAVIV_##n, etnaviv_ioctl_##func, flags)
@@ -463,6 +504,8 @@ static const struct drm_ioctl_desc etnaviv_ioctls[] = {
 	ETNA_IOCTL(WAIT_FENCE,   wait_fence,   DRM_AUTH|DRM_RENDER_ALLOW),
 	ETNA_IOCTL(GEM_USERPTR,  gem_userptr,  DRM_AUTH|DRM_RENDER_ALLOW),
 	ETNA_IOCTL(GEM_WAIT,     gem_wait,     DRM_AUTH|DRM_RENDER_ALLOW),
+	ETNA_IOCTL(PM_QUERY_DOM, pm_query_dom, DRM_AUTH|DRM_RENDER_ALLOW),
+	ETNA_IOCTL(PM_QUERY_SIG, pm_query_sig, DRM_AUTH|DRM_RENDER_ALLOW),
 };
 
 static const struct vm_operations_struct vm_ops = {
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.h b/drivers/gpu/drm/etnaviv/etnaviv_drv.h
index 058389f93b69..d249acb6da08 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_drv.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.h
@@ -26,7 +26,6 @@
 #include <linux/pm_runtime.h>
 #include <linux/slab.h>
 #include <linux/list.h>
-#include <linux/iommu.h>
 #include <linux/types.h>
 #include <linux/sizes.h>
 
@@ -92,15 +91,12 @@ int etnaviv_gem_cpu_fini(struct drm_gem_object *obj);
 void etnaviv_gem_free_object(struct drm_gem_object *obj);
 int etnaviv_gem_new_handle(struct drm_device *dev, struct drm_file *file,
 		u32 size, u32 flags, u32 *handle);
-struct drm_gem_object *etnaviv_gem_new_locked(struct drm_device *dev,
-		u32 size, u32 flags);
-struct drm_gem_object *etnaviv_gem_new(struct drm_device *dev,
-		u32 size, u32 flags);
 int etnaviv_gem_new_userptr(struct drm_device *dev, struct drm_file *file,
 	uintptr_t ptr, u32 size, u32 flags, u32 *handle);
 u16 etnaviv_buffer_init(struct etnaviv_gpu *gpu);
 u16 etnaviv_buffer_config_mmuv2(struct etnaviv_gpu *gpu, u32 mtlb_addr, u32 safe_addr);
 void etnaviv_buffer_end(struct etnaviv_gpu *gpu);
+void etnaviv_sync_point_queue(struct etnaviv_gpu *gpu, unsigned int event);
 void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, unsigned int event,
 	struct etnaviv_cmdbuf *cmdbuf);
 void etnaviv_validate_init(void);
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem.c b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
index bcc8c2d7c7c9..daee3f1196df 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gem.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
@@ -704,25 +704,6 @@ int etnaviv_gem_new_handle(struct drm_device *dev, struct drm_file *file,
 	return ret;
 }
 
-struct drm_gem_object *etnaviv_gem_new(struct drm_device *dev,
-		u32 size, u32 flags)
-{
-	struct drm_gem_object *obj;
-	int ret;
-
-	obj = __etnaviv_gem_new(dev, size, flags);
-	if (IS_ERR(obj))
-		return obj;
-
-	ret = etnaviv_gem_obj_add(dev, obj);
-	if (ret < 0) {
-		drm_gem_object_put_unlocked(obj);
-		return ERR_PTR(ret);
-	}
-
-	return obj;
-}
-
 int etnaviv_gem_new_private(struct drm_device *dev, size_t size, u32 flags,
 	struct reservation_object *robj, const struct etnaviv_gem_ops *ops,
 	struct etnaviv_gem_object **res)
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c b/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c
index 46dfe0737f43..ff911541a190 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gem_submit.c
@@ -21,6 +21,7 @@
 #include "etnaviv_drv.h"
 #include "etnaviv_gpu.h"
 #include "etnaviv_gem.h"
+#include "etnaviv_perfmon.h"
 
 /*
  * Cmdstream submission:
@@ -283,6 +284,54 @@ static int submit_reloc(struct etnaviv_gem_submit *submit, void *stream,
 	return 0;
 }
 
+static int submit_perfmon_validate(struct etnaviv_gem_submit *submit,
+		struct etnaviv_cmdbuf *cmdbuf,
+		const struct drm_etnaviv_gem_submit_pmr *pmrs,
+		u32 nr_pms)
+{
+	u32 i;
+
+	for (i = 0; i < nr_pms; i++) {
+		const struct drm_etnaviv_gem_submit_pmr *r = pmrs + i;
+		struct etnaviv_gem_submit_bo *bo;
+		int ret;
+
+		ret = submit_bo(submit, r->read_idx, &bo);
+		if (ret)
+			return ret;
+
+		/* at offset 0 a sequence number gets stored used for userspace sync */
+		if (r->read_offset == 0) {
+			DRM_ERROR("perfmon request: offset is 0");
+			return -EINVAL;
+		}
+
+		if (r->read_offset >= bo->obj->base.size - sizeof(u32)) {
+			DRM_ERROR("perfmon request: offset %u outside object", i);
+			return -EINVAL;
+		}
+
+		if (r->flags & ~(ETNA_PM_PROCESS_PRE | ETNA_PM_PROCESS_POST)) {
+			DRM_ERROR("perfmon request: flags are not valid");
+			return -EINVAL;
+		}
+
+		if (etnaviv_pm_req_validate(r, cmdbuf->exec_state)) {
+			DRM_ERROR("perfmon request: domain or signal not valid");
+			return -EINVAL;
+		}
+
+		cmdbuf->pmrs[i].flags = r->flags;
+		cmdbuf->pmrs[i].domain = r->domain;
+		cmdbuf->pmrs[i].signal = r->signal;
+		cmdbuf->pmrs[i].sequence = r->sequence;
+		cmdbuf->pmrs[i].offset = r->read_offset;
+		cmdbuf->pmrs[i].bo_vma = etnaviv_gem_vmap(&bo->obj->base);
+	}
+
+	return 0;
+}
+
 static void submit_cleanup(struct etnaviv_gem_submit *submit)
 {
 	unsigned i;
@@ -306,6 +355,7 @@ int etnaviv_ioctl_gem_submit(struct drm_device *dev, void *data,
 	struct etnaviv_drm_private *priv = dev->dev_private;
 	struct drm_etnaviv_gem_submit *args = data;
 	struct drm_etnaviv_gem_submit_reloc *relocs;
+	struct drm_etnaviv_gem_submit_pmr *pmrs;
 	struct drm_etnaviv_gem_submit_bo *bos;
 	struct etnaviv_gem_submit *submit;
 	struct etnaviv_cmdbuf *cmdbuf;
@@ -347,11 +397,12 @@ int etnaviv_ioctl_gem_submit(struct drm_device *dev, void *data,
 	 */
 	bos = kvmalloc_array(args->nr_bos, sizeof(*bos), GFP_KERNEL);
 	relocs = kvmalloc_array(args->nr_relocs, sizeof(*relocs), GFP_KERNEL);
+	pmrs = kvmalloc_array(args->nr_pmrs, sizeof(*pmrs), GFP_KERNEL);
 	stream = kvmalloc_array(1, args->stream_size, GFP_KERNEL);
 	cmdbuf = etnaviv_cmdbuf_new(gpu->cmdbuf_suballoc,
 				    ALIGN(args->stream_size, 8) + 8,
-				    args->nr_bos);
-	if (!bos || !relocs || !stream || !cmdbuf) {
+				    args->nr_bos, args->nr_pmrs);
+	if (!bos || !relocs || !pmrs || !stream || !cmdbuf) {
 		ret = -ENOMEM;
 		goto err_submit_cmds;
 	}
@@ -373,6 +424,14 @@ int etnaviv_ioctl_gem_submit(struct drm_device *dev, void *data,
 		goto err_submit_cmds;
 	}
 
+	ret = copy_from_user(pmrs, u64_to_user_ptr(args->pmrs),
+			     args->nr_pmrs * sizeof(*pmrs));
+	if (ret) {
+		ret = -EFAULT;
+		goto err_submit_cmds;
+	}
+	cmdbuf->nr_pmrs = args->nr_pmrs;
+
 	ret = copy_from_user(stream, u64_to_user_ptr(args->stream),
 			     args->stream_size);
 	if (ret) {
@@ -441,6 +500,10 @@ int etnaviv_ioctl_gem_submit(struct drm_device *dev, void *data,
 	if (ret)
 		goto out;
 
+	ret = submit_perfmon_validate(submit, cmdbuf, pmrs, args->nr_pmrs);
+	if (ret)
+		goto out;
+
 	memcpy(cmdbuf->vaddr, stream, args->stream_size);
 	cmdbuf->user_size = ALIGN(args->stream_size, 8);
 
@@ -496,6 +559,8 @@ err_submit_cmds:
 		kvfree(bos);
 	if (relocs)
 		kvfree(relocs);
+	if (pmrs)
+		kvfree(pmrs);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
index 4b152e0d31a6..e19cbe05da2a 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
@@ -25,6 +25,7 @@
 #include "etnaviv_gpu.h"
 #include "etnaviv_gem.h"
 #include "etnaviv_mmu.h"
+#include "etnaviv_perfmon.h"
 #include "common.xml.h"
 #include "state.xml.h"
 #include "state_hi.xml.h"
@@ -420,9 +421,10 @@ static void etnaviv_gpu_update_clock(struct etnaviv_gpu *gpu)
 			     gpu->base_rate_shader >> gpu->freq_scale);
 	} else {
 		unsigned int fscale = 1 << (6 - gpu->freq_scale);
-		u32 clock = VIVS_HI_CLOCK_CONTROL_DISABLE_DEBUG_REGISTERS |
-			    VIVS_HI_CLOCK_CONTROL_FSCALE_VAL(fscale);
+		u32 clock = gpu_read(gpu, VIVS_HI_CLOCK_CONTROL);
 
+		clock &= ~VIVS_HI_CLOCK_CONTROL_FSCALE_VAL__MASK;
+		clock |= VIVS_HI_CLOCK_CONTROL_FSCALE_VAL(fscale);
 		etnaviv_gpu_load_clock(gpu, clock);
 	}
 }
@@ -433,24 +435,14 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 	unsigned long timeout;
 	bool failed = true;
 
-	/* TODO
-	 *
-	 * - clock gating
-	 * - puls eater
-	 * - what about VG?
-	 */
-
 	/* We hope that the GPU resets in under one second */
 	timeout = jiffies + msecs_to_jiffies(1000);
 
 	while (time_is_after_jiffies(timeout)) {
 		/* enable clock */
-		etnaviv_gpu_update_clock(gpu);
-
-		control = gpu_read(gpu, VIVS_HI_CLOCK_CONTROL);
-
-		/* Wait for stable clock.  Vivante's code waited for 1ms */
-		usleep_range(1000, 10000);
+		unsigned int fscale = 1 << (6 - gpu->freq_scale);
+		control = VIVS_HI_CLOCK_CONTROL_FSCALE_VAL(fscale);
+		etnaviv_gpu_load_clock(gpu, control);
 
 		/* isolate the GPU. */
 		control |= VIVS_HI_CLOCK_CONTROL_ISOLATE_GPU;
@@ -461,7 +453,7 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 		gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, control);
 
 		/* wait for reset. */
-		msleep(1);
+		usleep_range(10, 20);
 
 		/* reset soft reset bit. */
 		control &= ~VIVS_HI_CLOCK_CONTROL_SOFT_RESET;
@@ -490,6 +482,10 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu)
 			continue;
 		}
 
+		/* disable debug registers, as they are not normally needed */
+		control |= VIVS_HI_CLOCK_CONTROL_DISABLE_DEBUG_REGISTERS;
+		gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, control);
+
 		failed = false;
 		break;
 	}
@@ -721,7 +717,7 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 	}
 
 	/* Create buffer: */
-	gpu->buffer = etnaviv_cmdbuf_new(gpu->cmdbuf_suballoc, PAGE_SIZE, 0);
+	gpu->buffer = etnaviv_cmdbuf_new(gpu->cmdbuf_suballoc, PAGE_SIZE, 0, 0);
 	if (!gpu->buffer) {
 		ret = -ENOMEM;
 		dev_err(gpu->dev, "could not create command buffer\n");
@@ -739,10 +735,9 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu)
 	/* Setup event management */
 	spin_lock_init(&gpu->event_spinlock);
 	init_completion(&gpu->event_free);
-	for (i = 0; i < ARRAY_SIZE(gpu->event); i++) {
-		gpu->event[i].used = false;
+	bitmap_zero(gpu->event_bitmap, ETNA_NR_EVENTS);
+	for (i = 0; i < ARRAY_SIZE(gpu->event); i++)
 		complete(&gpu->event_free);
-	}
 
 	/* Now program the hardware */
 	mutex_lock(&gpu->lock);
@@ -926,7 +921,7 @@ static void recover_worker(struct work_struct *work)
 	struct etnaviv_gpu *gpu = container_of(work, struct etnaviv_gpu,
 					       recover_work);
 	unsigned long flags;
-	unsigned int i;
+	unsigned int i = 0;
 
 	dev_err(gpu->dev, "hangcheck recover!\n");
 
@@ -945,14 +940,12 @@ static void recover_worker(struct work_struct *work)
 
 	/* complete all events, the GPU won't do it after the reset */
 	spin_lock_irqsave(&gpu->event_spinlock, flags);
-	for (i = 0; i < ARRAY_SIZE(gpu->event); i++) {
-		if (!gpu->event[i].used)
-			continue;
+	for_each_set_bit_from(i, gpu->event_bitmap, ETNA_NR_EVENTS) {
 		dma_fence_signal(gpu->event[i].fence);
 		gpu->event[i].fence = NULL;
-		gpu->event[i].used = false;
 		complete(&gpu->event_free);
 	}
+	bitmap_zero(gpu->event_bitmap, ETNA_NR_EVENTS);
 	spin_unlock_irqrestore(&gpu->event_spinlock, flags);
 	gpu->completed_fence = gpu->active_fence;
 
@@ -1140,30 +1133,45 @@ int etnaviv_gpu_fence_sync_obj(struct etnaviv_gem_object *etnaviv_obj,
  * event management:
  */
 
-static unsigned int event_alloc(struct etnaviv_gpu *gpu)
+static int event_alloc(struct etnaviv_gpu *gpu, unsigned nr_events,
+	unsigned int *events)
 {
-	unsigned long ret, flags;
-	unsigned int i, event = ~0U;
+	unsigned long flags, timeout = msecs_to_jiffies(10 * 10000);
+	unsigned i, acquired = 0;
 
-	ret = wait_for_completion_timeout(&gpu->event_free,
-					  msecs_to_jiffies(10 * 10000));
-	if (!ret)
-		dev_err(gpu->dev, "wait_for_completion_timeout failed");
+	for (i = 0; i < nr_events; i++) {
+		unsigned long ret;
 
-	spin_lock_irqsave(&gpu->event_spinlock, flags);
+		ret = wait_for_completion_timeout(&gpu->event_free, timeout);
 
-	/* find first free event */
-	for (i = 0; i < ARRAY_SIZE(gpu->event); i++) {
-		if (gpu->event[i].used == false) {
-			gpu->event[i].used = true;
-			event = i;
-			break;
+		if (!ret) {
+			dev_err(gpu->dev, "wait_for_completion_timeout failed");
+			goto out;
 		}
+
+		acquired++;
+		timeout = ret;
+	}
+
+	spin_lock_irqsave(&gpu->event_spinlock, flags);
+
+	for (i = 0; i < nr_events; i++) {
+		int event = find_first_zero_bit(gpu->event_bitmap, ETNA_NR_EVENTS);
+
+		events[i] = event;
+		memset(&gpu->event[event], 0, sizeof(struct etnaviv_event));
+		set_bit(event, gpu->event_bitmap);
 	}
 
 	spin_unlock_irqrestore(&gpu->event_spinlock, flags);
 
-	return event;
+	return 0;
+
+out:
+	for (i = 0; i < acquired; i++)
+		complete(&gpu->event_free);
+
+	return -EBUSY;
 }
 
 static void event_free(struct etnaviv_gpu *gpu, unsigned int event)
@@ -1172,12 +1180,12 @@ static void event_free(struct etnaviv_gpu *gpu, unsigned int event)
 
 	spin_lock_irqsave(&gpu->event_spinlock, flags);
 
-	if (gpu->event[event].used == false) {
+	if (!test_bit(event, gpu->event_bitmap)) {
 		dev_warn(gpu->dev, "event %u is already marked as free",
 			 event);
 		spin_unlock_irqrestore(&gpu->event_spinlock, flags);
 	} else {
-		gpu->event[event].used = false;
+		clear_bit(event, gpu->event_bitmap);
 		spin_unlock_irqrestore(&gpu->event_spinlock, flags);
 
 		complete(&gpu->event_free);
@@ -1311,12 +1319,71 @@ void etnaviv_gpu_pm_put(struct etnaviv_gpu *gpu)
 	pm_runtime_put_autosuspend(gpu->dev);
 }
 
+static void sync_point_perfmon_sample(struct etnaviv_gpu *gpu,
+	struct etnaviv_event *event, unsigned int flags)
+{
+	const struct etnaviv_cmdbuf *cmdbuf = event->cmdbuf;
+	unsigned int i;
+
+	for (i = 0; i < cmdbuf->nr_pmrs; i++) {
+		const struct etnaviv_perfmon_request *pmr = cmdbuf->pmrs + i;
+
+		if (pmr->flags == flags)
+			etnaviv_perfmon_process(gpu, pmr);
+	}
+}
+
+static void sync_point_perfmon_sample_pre(struct etnaviv_gpu *gpu,
+	struct etnaviv_event *event)
+{
+	u32 val;
+
+	/* disable clock gating */
+	val = gpu_read(gpu, VIVS_PM_POWER_CONTROLS);
+	val &= ~VIVS_PM_POWER_CONTROLS_ENABLE_MODULE_CLOCK_GATING;
+	gpu_write(gpu, VIVS_PM_POWER_CONTROLS, val);
+
+	/* enable debug register */
+	val = gpu_read(gpu, VIVS_HI_CLOCK_CONTROL);
+	val &= ~VIVS_HI_CLOCK_CONTROL_DISABLE_DEBUG_REGISTERS;
+	gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, val);
+
+	sync_point_perfmon_sample(gpu, event, ETNA_PM_PROCESS_PRE);
+}
+
+static void sync_point_perfmon_sample_post(struct etnaviv_gpu *gpu,
+	struct etnaviv_event *event)
+{
+	const struct etnaviv_cmdbuf *cmdbuf = event->cmdbuf;
+	unsigned int i;
+	u32 val;
+
+	sync_point_perfmon_sample(gpu, event, ETNA_PM_PROCESS_POST);
+
+	for (i = 0; i < cmdbuf->nr_pmrs; i++) {
+		const struct etnaviv_perfmon_request *pmr = cmdbuf->pmrs + i;
+
+		*pmr->bo_vma = pmr->sequence;
+	}
+
+	/* disable debug register */
+	val = gpu_read(gpu, VIVS_HI_CLOCK_CONTROL);
+	val |= VIVS_HI_CLOCK_CONTROL_DISABLE_DEBUG_REGISTERS;
+	gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, val);
+
+	/* enable clock gating */
+	val = gpu_read(gpu, VIVS_PM_POWER_CONTROLS);
+	val |= VIVS_PM_POWER_CONTROLS_ENABLE_MODULE_CLOCK_GATING;
+	gpu_write(gpu, VIVS_PM_POWER_CONTROLS, val);
+}
+
+
 /* add bo's to gpu's ring, and kick gpu: */
 int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
 	struct etnaviv_gem_submit *submit, struct etnaviv_cmdbuf *cmdbuf)
 {
 	struct dma_fence *fence;
-	unsigned int event, i;
+	unsigned int i, nr_events = 1, event[3];
 	int ret;
 
 	ret = etnaviv_gpu_pm_get_sync(gpu);
@@ -1332,10 +1399,19 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
 	 *
 	 */
 
-	event = event_alloc(gpu);
-	if (unlikely(event == ~0U)) {
-		DRM_ERROR("no free event\n");
-		ret = -EBUSY;
+	/*
+	 * if there are performance monitor requests we need to have
+	 * - a sync point to re-configure gpu and process ETNA_PM_PROCESS_PRE
+	 *   requests.
+	 * - a sync point to re-configure gpu, process ETNA_PM_PROCESS_POST requests
+	 *   and update the sequence number for userspace.
+	 */
+	if (cmdbuf->nr_pmrs)
+		nr_events = 3;
+
+	ret = event_alloc(gpu, nr_events, event);
+	if (ret) {
+		DRM_ERROR("no free events\n");
 		goto out_pm_put;
 	}
 
@@ -1343,12 +1419,14 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
 
 	fence = etnaviv_gpu_fence_alloc(gpu);
 	if (!fence) {
-		event_free(gpu, event);
+		for (i = 0; i < nr_events; i++)
+			event_free(gpu, event[i]);
+
 		ret = -ENOMEM;
 		goto out_unlock;
 	}
 
-	gpu->event[event].fence = fence;
+	gpu->event[event[0]].fence = fence;
 	submit->fence = dma_fence_get(fence);
 	gpu->active_fence = submit->fence->seqno;
 
@@ -1358,7 +1436,19 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
 		gpu->lastctx = cmdbuf->ctx;
 	}
 
-	etnaviv_buffer_queue(gpu, event, cmdbuf);
+	if (cmdbuf->nr_pmrs) {
+		gpu->event[event[1]].sync_point = &sync_point_perfmon_sample_pre;
+		gpu->event[event[1]].cmdbuf = cmdbuf;
+		etnaviv_sync_point_queue(gpu, event[1]);
+	}
+
+	etnaviv_buffer_queue(gpu, event[0], cmdbuf);
+
+	if (cmdbuf->nr_pmrs) {
+		gpu->event[event[2]].sync_point = &sync_point_perfmon_sample_post;
+		gpu->event[event[2]].cmdbuf = cmdbuf;
+		etnaviv_sync_point_queue(gpu, event[2]);
+	}
 
 	cmdbuf->fence = fence;
 	list_add_tail(&cmdbuf->node, &gpu->active_cmd_list);
@@ -1394,6 +1484,24 @@ out_pm_put:
 	return ret;
 }
 
+static void etnaviv_process_sync_point(struct etnaviv_gpu *gpu,
+	struct etnaviv_event *event)
+{
+	u32 addr = gpu_read(gpu, VIVS_FE_DMA_ADDRESS);
+
+	event->sync_point(gpu, event);
+	etnaviv_gpu_start_fe(gpu, addr + 2, 2);
+}
+
+static void sync_point_worker(struct work_struct *work)
+{
+	struct etnaviv_gpu *gpu = container_of(work, struct etnaviv_gpu,
+					       sync_point_work);
+
+	etnaviv_process_sync_point(gpu, &gpu->event[gpu->sync_point_event]);
+	event_free(gpu, gpu->sync_point_event);
+}
+
 /*
  * Init/Cleanup:
  */
@@ -1440,7 +1548,15 @@ static irqreturn_t irq_handler(int irq, void *data)
 
 			dev_dbg(gpu->dev, "event %u\n", event);
 
+			if (gpu->event[event].sync_point) {
+				gpu->sync_point_event = event;
+				etnaviv_queue_work(gpu->drm, &gpu->sync_point_work);
+			}
+
 			fence = gpu->event[event].fence;
+			if (!fence)
+				continue;
+
 			gpu->event[event].fence = NULL;
 			dma_fence_signal(fence);
 
@@ -1645,6 +1761,7 @@ static int etnaviv_gpu_bind(struct device *dev, struct device *master,
 
 	INIT_LIST_HEAD(&gpu->active_cmd_list);
 	INIT_WORK(&gpu->retire_work, retire_worker);
+	INIT_WORK(&gpu->sync_point_work, sync_point_worker);
 	INIT_WORK(&gpu->recover_work, recover_worker);
 	init_waitqueue_head(&gpu->fence_event);
 
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
index 689cb8f3680c..4f10f147297a 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
@@ -88,13 +88,17 @@ struct etnaviv_chip_identity {
 };
 
 struct etnaviv_event {
-	bool used;
 	struct dma_fence *fence;
+	struct etnaviv_cmdbuf *cmdbuf;
+
+	void (*sync_point)(struct etnaviv_gpu *gpu, struct etnaviv_event *event);
 };
 
 struct etnaviv_cmdbuf_suballoc;
 struct etnaviv_cmdbuf;
 
+#define ETNA_NR_EVENTS 30
+
 struct etnaviv_gpu {
 	struct drm_device *drm;
 	struct thermal_cooling_device *cooling;
@@ -112,7 +116,8 @@ struct etnaviv_gpu {
 	u32 memory_base;
 
 	/* event management: */
-	struct etnaviv_event event[30];
+	DECLARE_BITMAP(event_bitmap, ETNA_NR_EVENTS);
+	struct etnaviv_event event[ETNA_NR_EVENTS];
 	struct completion event_free;
 	spinlock_t event_spinlock;
 
@@ -133,6 +138,10 @@ struct etnaviv_gpu {
 	/* worker for handling active-list retiring: */
 	struct work_struct retire_work;
 
+	/* worker for handling 'sync' points: */
+	struct work_struct sync_point_work;
+	int sync_point_event;
+
 	void __iomem *mmio;
 	int irq;
 
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
index 7a7c97f599d7..14e24ac6573f 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c
@@ -14,7 +14,6 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <linux/iommu.h>
 #include <linux/platform_device.h>
 #include <linux/sizes.h>
 #include <linux/slab.h>
@@ -31,174 +30,115 @@
 
 #define GPU_MEM_START	0x80000000
 
-struct etnaviv_iommu_domain_pgtable {
-	u32 *pgtable;
-	dma_addr_t paddr;
+struct etnaviv_iommuv1_domain {
+	struct etnaviv_iommu_domain base;
+	u32 *pgtable_cpu;
+	dma_addr_t pgtable_dma;
 };
 
-struct etnaviv_iommu_domain {
-	struct iommu_domain domain;
-	struct device *dev;
-	void *bad_page_cpu;
-	dma_addr_t bad_page_dma;
-	struct etnaviv_iommu_domain_pgtable pgtable;
-	spinlock_t map_lock;
-};
-
-static struct etnaviv_iommu_domain *to_etnaviv_domain(struct iommu_domain *domain)
-{
-	return container_of(domain, struct etnaviv_iommu_domain, domain);
-}
-
-static int pgtable_alloc(struct etnaviv_iommu_domain_pgtable *pgtable,
-			 size_t size)
-{
-	pgtable->pgtable = dma_alloc_coherent(NULL, size, &pgtable->paddr, GFP_KERNEL);
-	if (!pgtable->pgtable)
-		return -ENOMEM;
-
-	return 0;
-}
-
-static void pgtable_free(struct etnaviv_iommu_domain_pgtable *pgtable,
-			 size_t size)
+static struct etnaviv_iommuv1_domain *
+to_etnaviv_domain(struct etnaviv_iommu_domain *domain)
 {
-	dma_free_coherent(NULL, size, pgtable->pgtable, pgtable->paddr);
-}
-
-static u32 pgtable_read(struct etnaviv_iommu_domain_pgtable *pgtable,
-			   unsigned long iova)
-{
-	/* calcuate index into page table */
-	unsigned int index = (iova - GPU_MEM_START) / SZ_4K;
-	phys_addr_t paddr;
-
-	paddr = pgtable->pgtable[index];
-
-	return paddr;
+	return container_of(domain, struct etnaviv_iommuv1_domain, base);
 }
 
-static void pgtable_write(struct etnaviv_iommu_domain_pgtable *pgtable,
-			  unsigned long iova, phys_addr_t paddr)
-{
-	/* calcuate index into page table */
-	unsigned int index = (iova - GPU_MEM_START) / SZ_4K;
-
-	pgtable->pgtable[index] = paddr;
-}
-
-static int __etnaviv_iommu_init(struct etnaviv_iommu_domain *etnaviv_domain)
+static int __etnaviv_iommu_init(struct etnaviv_iommuv1_domain *etnaviv_domain)
 {
 	u32 *p;
-	int ret, i;
-
-	etnaviv_domain->bad_page_cpu = dma_alloc_coherent(etnaviv_domain->dev,
-						  SZ_4K,
-						  &etnaviv_domain->bad_page_dma,
-						  GFP_KERNEL);
-	if (!etnaviv_domain->bad_page_cpu)
+	int i;
+
+	etnaviv_domain->base.bad_page_cpu = dma_alloc_coherent(
+						etnaviv_domain->base.dev,
+						SZ_4K,
+						&etnaviv_domain->base.bad_page_dma,
+						GFP_KERNEL);
+	if (!etnaviv_domain->base.bad_page_cpu)
 		return -ENOMEM;
 
-	p = etnaviv_domain->bad_page_cpu;
+	p = etnaviv_domain->base.bad_page_cpu;
 	for (i = 0; i < SZ_4K / 4; i++)
 		*p++ = 0xdead55aa;
 
-	ret = pgtable_alloc(&etnaviv_domain->pgtable, PT_SIZE);
-	if (ret < 0) {
-		dma_free_coherent(etnaviv_domain->dev, SZ_4K,
-				  etnaviv_domain->bad_page_cpu,
-				  etnaviv_domain->bad_page_dma);
-		return ret;
+	etnaviv_domain->pgtable_cpu =
+			dma_alloc_coherent(etnaviv_domain->base.dev, PT_SIZE,
+					   &etnaviv_domain->pgtable_dma,
+					   GFP_KERNEL);
+	if (!etnaviv_domain->pgtable_cpu) {
+		dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
+				  etnaviv_domain->base.bad_page_cpu,
+				  etnaviv_domain->base.bad_page_dma);
+		return -ENOMEM;
 	}
 
 	for (i = 0; i < PT_ENTRIES; i++)
-		etnaviv_domain->pgtable.pgtable[i] =
-			etnaviv_domain->bad_page_dma;
-
-	spin_lock_init(&etnaviv_domain->map_lock);
+		etnaviv_domain->pgtable_cpu[i] =
+				etnaviv_domain->base.bad_page_dma;
 
 	return 0;
 }
 
-static void etnaviv_domain_free(struct iommu_domain *domain)
+static void etnaviv_iommuv1_domain_free(struct etnaviv_iommu_domain *domain)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain = to_etnaviv_domain(domain);
+	struct etnaviv_iommuv1_domain *etnaviv_domain =
+			to_etnaviv_domain(domain);
 
-	pgtable_free(&etnaviv_domain->pgtable, PT_SIZE);
+	dma_free_coherent(etnaviv_domain->base.dev, PT_SIZE,
+			  etnaviv_domain->pgtable_cpu,
+			  etnaviv_domain->pgtable_dma);
 
-	dma_free_coherent(etnaviv_domain->dev, SZ_4K,
-			  etnaviv_domain->bad_page_cpu,
-			  etnaviv_domain->bad_page_dma);
+	dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
+			  etnaviv_domain->base.bad_page_cpu,
+			  etnaviv_domain->base.bad_page_dma);
 
 	kfree(etnaviv_domain);
 }
 
-static int etnaviv_iommuv1_map(struct iommu_domain *domain, unsigned long iova,
-	   phys_addr_t paddr, size_t size, int prot)
+static int etnaviv_iommuv1_map(struct etnaviv_iommu_domain *domain,
+			       unsigned long iova, phys_addr_t paddr,
+			       size_t size, int prot)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain = to_etnaviv_domain(domain);
+	struct etnaviv_iommuv1_domain *etnaviv_domain = to_etnaviv_domain(domain);
+	unsigned int index = (iova - GPU_MEM_START) / SZ_4K;
 
 	if (size != SZ_4K)
 		return -EINVAL;
 
-	spin_lock(&etnaviv_domain->map_lock);
-	pgtable_write(&etnaviv_domain->pgtable, iova, paddr);
-	spin_unlock(&etnaviv_domain->map_lock);
+	etnaviv_domain->pgtable_cpu[index] = paddr;
 
 	return 0;
 }
 
-static size_t etnaviv_iommuv1_unmap(struct iommu_domain *domain,
+static size_t etnaviv_iommuv1_unmap(struct etnaviv_iommu_domain *domain,
 	unsigned long iova, size_t size)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain = to_etnaviv_domain(domain);
+	struct etnaviv_iommuv1_domain *etnaviv_domain =
+			to_etnaviv_domain(domain);
+	unsigned int index = (iova - GPU_MEM_START) / SZ_4K;
 
 	if (size != SZ_4K)
 		return -EINVAL;
 
-	spin_lock(&etnaviv_domain->map_lock);
-	pgtable_write(&etnaviv_domain->pgtable, iova,
-		      etnaviv_domain->bad_page_dma);
-	spin_unlock(&etnaviv_domain->map_lock);
+	etnaviv_domain->pgtable_cpu[index] = etnaviv_domain->base.bad_page_dma;
 
 	return SZ_4K;
 }
 
-static phys_addr_t etnaviv_iommu_iova_to_phys(struct iommu_domain *domain,
-	dma_addr_t iova)
-{
-	struct etnaviv_iommu_domain *etnaviv_domain = to_etnaviv_domain(domain);
-
-	return pgtable_read(&etnaviv_domain->pgtable, iova);
-}
-
-static size_t etnaviv_iommuv1_dump_size(struct iommu_domain *domain)
+static size_t etnaviv_iommuv1_dump_size(struct etnaviv_iommu_domain *domain)
 {
 	return PT_SIZE;
 }
 
-static void etnaviv_iommuv1_dump(struct iommu_domain *domain, void *buf)
+static void etnaviv_iommuv1_dump(struct etnaviv_iommu_domain *domain, void *buf)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain = to_etnaviv_domain(domain);
+	struct etnaviv_iommuv1_domain *etnaviv_domain =
+			to_etnaviv_domain(domain);
 
-	memcpy(buf, etnaviv_domain->pgtable.pgtable, PT_SIZE);
+	memcpy(buf, etnaviv_domain->pgtable_cpu, PT_SIZE);
 }
 
-static const struct etnaviv_iommu_ops etnaviv_iommu_ops = {
-	.ops = {
-		.domain_free = etnaviv_domain_free,
-		.map = etnaviv_iommuv1_map,
-		.unmap = etnaviv_iommuv1_unmap,
-		.iova_to_phys = etnaviv_iommu_iova_to_phys,
-		.pgsize_bitmap = SZ_4K,
-	},
-	.dump_size = etnaviv_iommuv1_dump_size,
-	.dump = etnaviv_iommuv1_dump,
-};
-
 void etnaviv_iommuv1_restore(struct etnaviv_gpu *gpu)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain =
+	struct etnaviv_iommuv1_domain *etnaviv_domain =
 			to_etnaviv_domain(gpu->mmu->domain);
 	u32 pgtable;
 
@@ -210,7 +150,7 @@ void etnaviv_iommuv1_restore(struct etnaviv_gpu *gpu)
 	gpu_write(gpu, VIVS_MC_MEMORY_BASE_ADDR_PE, gpu->memory_base);
 
 	/* set page table address in MC */
-	pgtable = (u32)etnaviv_domain->pgtable.paddr;
+	pgtable = (u32)etnaviv_domain->pgtable_dma;
 
 	gpu_write(gpu, VIVS_MC_MMU_FE_PAGE_TABLE, pgtable);
 	gpu_write(gpu, VIVS_MC_MMU_TX_PAGE_TABLE, pgtable);
@@ -219,28 +159,37 @@ void etnaviv_iommuv1_restore(struct etnaviv_gpu *gpu)
 	gpu_write(gpu, VIVS_MC_MMU_RA_PAGE_TABLE, pgtable);
 }
 
-struct iommu_domain *etnaviv_iommuv1_domain_alloc(struct etnaviv_gpu *gpu)
+const struct etnaviv_iommu_domain_ops etnaviv_iommuv1_ops = {
+	.free = etnaviv_iommuv1_domain_free,
+	.map = etnaviv_iommuv1_map,
+	.unmap = etnaviv_iommuv1_unmap,
+	.dump_size = etnaviv_iommuv1_dump_size,
+	.dump = etnaviv_iommuv1_dump,
+};
+
+struct etnaviv_iommu_domain *
+etnaviv_iommuv1_domain_alloc(struct etnaviv_gpu *gpu)
 {
-	struct etnaviv_iommu_domain *etnaviv_domain;
+	struct etnaviv_iommuv1_domain *etnaviv_domain;
+	struct etnaviv_iommu_domain *domain;
 	int ret;
 
 	etnaviv_domain = kzalloc(sizeof(*etnaviv_domain), GFP_KERNEL);
 	if (!etnaviv_domain)
 		return NULL;
 
-	etnaviv_domain->dev = gpu->dev;
+	domain = &etnaviv_domain->base;
 
-	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;
+	domain->dev = gpu->dev;
+	domain->base = GPU_MEM_START;
+	domain->size = PT_ENTRIES * SZ_4K;
+	domain->ops = &etnaviv_iommuv1_ops;
 
 	ret = __etnaviv_iommu_init(etnaviv_domain);
 	if (ret)
 		goto out_free;
 
-	return &etnaviv_domain->domain;
+	return &etnaviv_domain->base;
 
 out_free:
 	kfree(etnaviv_domain);
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu.h b/drivers/gpu/drm/etnaviv/etnaviv_iommu.h
index 8b51e7c16feb..01d59bf70d78 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_iommu.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu.h
@@ -18,11 +18,14 @@
 #define __ETNAVIV_IOMMU_H__
 
 struct etnaviv_gpu;
+struct etnaviv_iommu_domain;
 
-struct iommu_domain *etnaviv_iommuv1_domain_alloc(struct etnaviv_gpu *gpu);
+struct etnaviv_iommu_domain *
+etnaviv_iommuv1_domain_alloc(struct etnaviv_gpu *gpu);
 void etnaviv_iommuv1_restore(struct etnaviv_gpu *gpu);
 
-struct iommu_domain *etnaviv_iommuv2_domain_alloc(struct etnaviv_gpu *gpu);
+struct etnaviv_iommu_domain *
+etnaviv_iommuv2_domain_alloc(struct etnaviv_gpu *gpu);
 void etnaviv_iommuv2_restore(struct etnaviv_gpu *gpu);
 
 #endif /* __ETNAVIV_IOMMU_H__ */
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
index cbe447ac5974..fc60fc8ddbf0 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
@@ -14,7 +14,6 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#include <linux/iommu.h>
 #include <linux/platform_device.h>
 #include <linux/sizes.h>
 #include <linux/slab.h>
@@ -40,10 +39,7 @@
 #define MMUv2_MAX_STLB_ENTRIES		1024
 
 struct etnaviv_iommuv2_domain {
-	struct iommu_domain domain;
-	struct device *dev;
-	void *bad_page_cpu;
-	dma_addr_t bad_page_dma;
+	struct etnaviv_iommu_domain base;
 	/* M(aster) TLB aka first level pagetable */
 	u32 *mtlb_cpu;
 	dma_addr_t mtlb_dma;
@@ -52,13 +48,15 @@ struct etnaviv_iommuv2_domain {
 	dma_addr_t stlb_dma[1024];
 };
 
-static struct etnaviv_iommuv2_domain *to_etnaviv_domain(struct iommu_domain *domain)
+static struct etnaviv_iommuv2_domain *
+to_etnaviv_domain(struct etnaviv_iommu_domain *domain)
 {
-	return container_of(domain, struct etnaviv_iommuv2_domain, domain);
+	return container_of(domain, struct etnaviv_iommuv2_domain, base);
 }
 
-static int etnaviv_iommuv2_map(struct iommu_domain *domain, unsigned long iova,
-	   phys_addr_t paddr, size_t size, int prot)
+static int etnaviv_iommuv2_map(struct etnaviv_iommu_domain *domain,
+			       unsigned long iova, phys_addr_t paddr,
+			       size_t size, int prot)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
 			to_etnaviv_domain(domain);
@@ -68,7 +66,7 @@ static int etnaviv_iommuv2_map(struct iommu_domain *domain, unsigned long iova,
 	if (size != SZ_4K)
 		return -EINVAL;
 
-	if (prot & IOMMU_WRITE)
+	if (prot & ETNAVIV_PROT_WRITE)
 		entry |= MMUv2_PTE_WRITEABLE;
 
 	mtlb_entry = (iova & MMUv2_MTLB_MASK) >> MMUv2_MTLB_SHIFT;
@@ -79,8 +77,8 @@ static int etnaviv_iommuv2_map(struct iommu_domain *domain, unsigned long iova,
 	return 0;
 }
 
-static size_t etnaviv_iommuv2_unmap(struct iommu_domain *domain,
-	unsigned long iova, size_t size)
+static size_t etnaviv_iommuv2_unmap(struct etnaviv_iommu_domain *domain,
+				    unsigned long iova, size_t size)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
 			to_etnaviv_domain(domain);
@@ -97,38 +95,26 @@ static size_t etnaviv_iommuv2_unmap(struct iommu_domain *domain,
 	return SZ_4K;
 }
 
-static phys_addr_t etnaviv_iommuv2_iova_to_phys(struct iommu_domain *domain,
-	dma_addr_t iova)
-{
-	struct etnaviv_iommuv2_domain *etnaviv_domain =
-			to_etnaviv_domain(domain);
-	int mtlb_entry, stlb_entry;
-
-	mtlb_entry = (iova & MMUv2_MTLB_MASK) >> MMUv2_MTLB_SHIFT;
-	stlb_entry = (iova & MMUv2_STLB_MASK) >> MMUv2_STLB_SHIFT;
-
-	return etnaviv_domain->stlb_cpu[mtlb_entry][stlb_entry] & ~(SZ_4K - 1);
-}
-
 static int etnaviv_iommuv2_init(struct etnaviv_iommuv2_domain *etnaviv_domain)
 {
 	u32 *p;
 	int ret, i, j;
 
 	/* allocate scratch page */
-	etnaviv_domain->bad_page_cpu = dma_alloc_coherent(etnaviv_domain->dev,
-						  SZ_4K,
-						  &etnaviv_domain->bad_page_dma,
-						  GFP_KERNEL);
-	if (!etnaviv_domain->bad_page_cpu) {
+	etnaviv_domain->base.bad_page_cpu = dma_alloc_coherent(
+						etnaviv_domain->base.dev,
+						SZ_4K,
+						&etnaviv_domain->base.bad_page_dma,
+						GFP_KERNEL);
+	if (!etnaviv_domain->base.bad_page_cpu) {
 		ret = -ENOMEM;
 		goto fail_mem;
 	}
-	p = etnaviv_domain->bad_page_cpu;
+	p = etnaviv_domain->base.bad_page_cpu;
 	for (i = 0; i < SZ_4K / 4; i++)
 		*p++ = 0xdead55aa;
 
-	etnaviv_domain->mtlb_cpu = dma_alloc_coherent(etnaviv_domain->dev,
+	etnaviv_domain->mtlb_cpu = dma_alloc_coherent(etnaviv_domain->base.dev,
 						  SZ_4K,
 						  &etnaviv_domain->mtlb_dma,
 						  GFP_KERNEL);
@@ -140,7 +126,7 @@ static int etnaviv_iommuv2_init(struct etnaviv_iommuv2_domain *etnaviv_domain)
 	/* pre-populate STLB pages (may want to switch to on-demand later) */
 	for (i = 0; i < MMUv2_MAX_STLB_ENTRIES; i++) {
 		etnaviv_domain->stlb_cpu[i] =
-				dma_alloc_coherent(etnaviv_domain->dev,
+				dma_alloc_coherent(etnaviv_domain->base.dev,
 						   SZ_4K,
 						   &etnaviv_domain->stlb_dma[i],
 						   GFP_KERNEL);
@@ -159,19 +145,19 @@ static int etnaviv_iommuv2_init(struct etnaviv_iommuv2_domain *etnaviv_domain)
 	return 0;
 
 fail_mem:
-	if (etnaviv_domain->bad_page_cpu)
-		dma_free_coherent(etnaviv_domain->dev, SZ_4K,
-				  etnaviv_domain->bad_page_cpu,
-				  etnaviv_domain->bad_page_dma);
+	if (etnaviv_domain->base.bad_page_cpu)
+		dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
+				  etnaviv_domain->base.bad_page_cpu,
+				  etnaviv_domain->base.bad_page_dma);
 
 	if (etnaviv_domain->mtlb_cpu)
-		dma_free_coherent(etnaviv_domain->dev, SZ_4K,
+		dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
 				  etnaviv_domain->mtlb_cpu,
 				  etnaviv_domain->mtlb_dma);
 
 	for (i = 0; i < MMUv2_MAX_STLB_ENTRIES; i++) {
 		if (etnaviv_domain->stlb_cpu[i])
-			dma_free_coherent(etnaviv_domain->dev, SZ_4K,
+			dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
 					  etnaviv_domain->stlb_cpu[i],
 					  etnaviv_domain->stlb_dma[i]);
 	}
@@ -179,23 +165,23 @@ fail_mem:
 	return ret;
 }
 
-static void etnaviv_iommuv2_domain_free(struct iommu_domain *domain)
+static void etnaviv_iommuv2_domain_free(struct etnaviv_iommu_domain *domain)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
 			to_etnaviv_domain(domain);
 	int i;
 
-	dma_free_coherent(etnaviv_domain->dev, SZ_4K,
-			  etnaviv_domain->bad_page_cpu,
-			  etnaviv_domain->bad_page_dma);
+	dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
+			  etnaviv_domain->base.bad_page_cpu,
+			  etnaviv_domain->base.bad_page_dma);
 
-	dma_free_coherent(etnaviv_domain->dev, SZ_4K,
+	dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
 			  etnaviv_domain->mtlb_cpu,
 			  etnaviv_domain->mtlb_dma);
 
 	for (i = 0; i < MMUv2_MAX_STLB_ENTRIES; i++) {
 		if (etnaviv_domain->stlb_cpu[i])
-			dma_free_coherent(etnaviv_domain->dev, SZ_4K,
+			dma_free_coherent(etnaviv_domain->base.dev, SZ_4K,
 					  etnaviv_domain->stlb_cpu[i],
 					  etnaviv_domain->stlb_dma[i]);
 	}
@@ -203,7 +189,7 @@ static void etnaviv_iommuv2_domain_free(struct iommu_domain *domain)
 	vfree(etnaviv_domain);
 }
 
-static size_t etnaviv_iommuv2_dump_size(struct iommu_domain *domain)
+static size_t etnaviv_iommuv2_dump_size(struct etnaviv_iommu_domain *domain)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
 			to_etnaviv_domain(domain);
@@ -217,7 +203,7 @@ static size_t etnaviv_iommuv2_dump_size(struct iommu_domain *domain)
 	return dump_size;
 }
 
-static void etnaviv_iommuv2_dump(struct iommu_domain *domain, void *buf)
+static void etnaviv_iommuv2_dump(struct etnaviv_iommu_domain *domain, void *buf)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
 			to_etnaviv_domain(domain);
@@ -230,18 +216,6 @@ static void etnaviv_iommuv2_dump(struct iommu_domain *domain, void *buf)
 			memcpy(buf, etnaviv_domain->stlb_cpu[i], SZ_4K);
 }
 
-static const struct etnaviv_iommu_ops etnaviv_iommu_ops = {
-	.ops = {
-		.domain_free = etnaviv_iommuv2_domain_free,
-		.map = etnaviv_iommuv2_map,
-		.unmap = etnaviv_iommuv2_unmap,
-		.iova_to_phys = etnaviv_iommuv2_iova_to_phys,
-		.pgsize_bitmap = SZ_4K,
-	},
-	.dump_size = etnaviv_iommuv2_dump_size,
-	.dump = etnaviv_iommuv2_dump,
-};
-
 void etnaviv_iommuv2_restore(struct etnaviv_gpu *gpu)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain =
@@ -254,35 +228,45 @@ void etnaviv_iommuv2_restore(struct etnaviv_gpu *gpu)
 
 	prefetch = etnaviv_buffer_config_mmuv2(gpu,
 				(u32)etnaviv_domain->mtlb_dma,
-				(u32)etnaviv_domain->bad_page_dma);
+				(u32)etnaviv_domain->base.bad_page_dma);
 	etnaviv_gpu_start_fe(gpu, (u32)etnaviv_cmdbuf_get_pa(gpu->buffer),
 			     prefetch);
 	etnaviv_gpu_wait_idle(gpu, 100);
 
 	gpu_write(gpu, VIVS_MMUv2_CONTROL, VIVS_MMUv2_CONTROL_ENABLE);
 }
-struct iommu_domain *etnaviv_iommuv2_domain_alloc(struct etnaviv_gpu *gpu)
+
+const struct etnaviv_iommu_domain_ops etnaviv_iommuv2_ops = {
+	.free = etnaviv_iommuv2_domain_free,
+	.map = etnaviv_iommuv2_map,
+	.unmap = etnaviv_iommuv2_unmap,
+	.dump_size = etnaviv_iommuv2_dump_size,
+	.dump = etnaviv_iommuv2_dump,
+};
+
+struct etnaviv_iommu_domain *
+etnaviv_iommuv2_domain_alloc(struct etnaviv_gpu *gpu)
 {
 	struct etnaviv_iommuv2_domain *etnaviv_domain;
+	struct etnaviv_iommu_domain *domain;
 	int ret;
 
 	etnaviv_domain = vzalloc(sizeof(*etnaviv_domain));
 	if (!etnaviv_domain)
 		return NULL;
 
-	etnaviv_domain->dev = gpu->dev;
+	domain = &etnaviv_domain->base;
 
-	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 = 0;
-	etnaviv_domain->domain.geometry.aperture_end = ~0UL & ~(SZ_4K - 1);
+	domain->dev = gpu->dev;
+	domain->base = 0;
+	domain->size = (u64)SZ_1G * 4;
+	domain->ops = &etnaviv_iommuv2_ops;
 
 	ret = etnaviv_iommuv2_init(etnaviv_domain);
 	if (ret)
 		goto out_free;
 
-	return &etnaviv_domain->domain;
+	return &etnaviv_domain->base;
 
 out_free:
 	vfree(etnaviv_domain);
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
index f103e787de94..35074b944778 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.c
@@ -22,17 +22,64 @@
 #include "etnaviv_iommu.h"
 #include "etnaviv_mmu.h"
 
-static int etnaviv_fault_handler(struct iommu_domain *iommu, struct device *dev,
-		unsigned long iova, int flags, void *arg)
+static void etnaviv_domain_unmap(struct etnaviv_iommu_domain *domain,
+				 unsigned long iova, size_t size)
 {
-	DBG("*** fault: iova=%08lx, flags=%d", iova, flags);
-	return 0;
+	size_t unmapped_page, unmapped = 0;
+	size_t pgsize = SZ_4K;
+
+	if (!IS_ALIGNED(iova | size, pgsize)) {
+		pr_err("unaligned: iova 0x%lx size 0x%zx min_pagesz 0x%x\n",
+		       iova, size, pgsize);
+		return;
+	}
+
+	while (unmapped < size) {
+		unmapped_page = domain->ops->unmap(domain, iova, pgsize);
+		if (!unmapped_page)
+			break;
+
+		iova += unmapped_page;
+		unmapped += unmapped_page;
+	}
 }
 
-int etnaviv_iommu_map(struct etnaviv_iommu *iommu, u32 iova,
-		struct sg_table *sgt, unsigned len, int prot)
+static int etnaviv_domain_map(struct etnaviv_iommu_domain *domain,
+			      unsigned long iova, phys_addr_t paddr,
+			      size_t size, int prot)
 {
-	struct iommu_domain *domain = iommu->domain;
+	unsigned long orig_iova = iova;
+	size_t pgsize = SZ_4K;
+	size_t orig_size = size;
+	int ret = 0;
+
+	if (!IS_ALIGNED(iova | paddr | size, pgsize)) {
+		pr_err("unaligned: iova 0x%lx pa %pa size 0x%zx min_pagesz 0x%x\n",
+		       iova, &paddr, size, pgsize);
+		return -EINVAL;
+	}
+
+	while (size) {
+		ret = domain->ops->map(domain, iova, paddr, pgsize, prot);
+		if (ret)
+			break;
+
+		iova += pgsize;
+		paddr += pgsize;
+		size -= pgsize;
+	}
+
+	/* unroll mapping in case something went wrong */
+	if (ret)
+		etnaviv_domain_unmap(domain, orig_iova, orig_size - size);
+
+	return ret;
+}
+
+static int etnaviv_iommu_map(struct etnaviv_iommu *iommu, u32 iova,
+			     struct sg_table *sgt, unsigned len, int prot)
+{
+	struct etnaviv_iommu_domain *domain = iommu->domain;
 	struct scatterlist *sg;
 	unsigned int da = iova;
 	unsigned int i, j;
@@ -47,7 +94,7 @@ int etnaviv_iommu_map(struct etnaviv_iommu *iommu, u32 iova,
 
 		VERB("map[%d]: %08x %08x(%zx)", i, iova, pa, bytes);
 
-		ret = iommu_map(domain, da, pa, bytes, prot);
+		ret = etnaviv_domain_map(domain, da, pa, bytes, prot);
 		if (ret)
 			goto fail;
 
@@ -62,27 +109,24 @@ fail:
 	for_each_sg(sgt->sgl, sg, i, j) {
 		size_t bytes = sg_dma_len(sg) + sg->offset;
 
-		iommu_unmap(domain, da, bytes);
+		etnaviv_domain_unmap(domain, da, bytes);
 		da += bytes;
 	}
 	return ret;
 }
 
-int etnaviv_iommu_unmap(struct etnaviv_iommu *iommu, u32 iova,
-		struct sg_table *sgt, unsigned len)
+static void etnaviv_iommu_unmap(struct etnaviv_iommu *iommu, u32 iova,
+				struct sg_table *sgt, unsigned len)
 {
-	struct iommu_domain *domain = iommu->domain;
+	struct etnaviv_iommu_domain *domain = iommu->domain;
 	struct scatterlist *sg;
 	unsigned int da = iova;
 	int i;
 
 	for_each_sg(sgt->sgl, sg, sgt->nents, i) {
 		size_t bytes = sg_dma_len(sg) + sg->offset;
-		size_t unmapped;
 
-		unmapped = iommu_unmap(domain, da, bytes);
-		if (unmapped < bytes)
-			return unmapped;
+		etnaviv_domain_unmap(domain, da, bytes);
 
 		VERB("unmap[%d]: %08x(%zx)", i, iova, bytes);
 
@@ -90,8 +134,6 @@ int etnaviv_iommu_unmap(struct etnaviv_iommu *iommu, u32 iova,
 
 		da += bytes;
 	}
-
-	return 0;
 }
 
 static void etnaviv_iommu_remove_mapping(struct etnaviv_iommu *mmu,
@@ -237,7 +279,7 @@ int etnaviv_iommu_map_gem(struct etnaviv_iommu *mmu,
 	mmu->last_iova = node->start + etnaviv_obj->base.size;
 	mapping->iova = node->start;
 	ret = etnaviv_iommu_map(mmu, node->start, sgt, etnaviv_obj->base.size,
-				IOMMU_READ | IOMMU_WRITE);
+				ETNAVIV_PROT_READ | ETNAVIV_PROT_WRITE);
 
 	if (ret < 0) {
 		drm_mm_remove_node(node);
@@ -271,7 +313,7 @@ void etnaviv_iommu_unmap_gem(struct etnaviv_iommu *mmu,
 void etnaviv_iommu_destroy(struct etnaviv_iommu *mmu)
 {
 	drm_mm_takedown(&mmu->mm);
-	iommu_domain_free(mmu->domain);
+	mmu->domain->ops->free(mmu->domain);
 	kfree(mmu);
 }
 
@@ -303,11 +345,7 @@ struct etnaviv_iommu *etnaviv_iommu_new(struct etnaviv_gpu *gpu)
 	mutex_init(&mmu->lock);
 	INIT_LIST_HEAD(&mmu->mappings);
 
-	drm_mm_init(&mmu->mm, mmu->domain->geometry.aperture_start,
-		    mmu->domain->geometry.aperture_end -
-		    mmu->domain->geometry.aperture_start + 1);
-
-	iommu_set_fault_handler(mmu->domain, etnaviv_fault_handler, gpu->dev);
+	drm_mm_init(&mmu->mm, mmu->domain->base, mmu->domain->size);
 
 	return mmu;
 }
@@ -338,8 +376,8 @@ int etnaviv_iommu_get_suballoc_va(struct etnaviv_gpu *gpu, dma_addr_t paddr,
 			mutex_unlock(&mmu->lock);
 			return ret;
 		}
-		ret = iommu_map(mmu->domain, vram_node->start, paddr, size,
-				IOMMU_READ);
+		ret = etnaviv_domain_map(mmu->domain, vram_node->start, paddr,
+					 size, ETNAVIV_PROT_READ);
 		if (ret < 0) {
 			drm_mm_remove_node(vram_node);
 			mutex_unlock(&mmu->lock);
@@ -362,25 +400,17 @@ void etnaviv_iommu_put_suballoc_va(struct etnaviv_gpu *gpu,
 
 	if (mmu->version == ETNAVIV_IOMMU_V2) {
 		mutex_lock(&mmu->lock);
-		iommu_unmap(mmu->domain,iova, size);
+		etnaviv_domain_unmap(mmu->domain, iova, size);
 		drm_mm_remove_node(vram_node);
 		mutex_unlock(&mmu->lock);
 	}
 }
 size_t etnaviv_iommu_dump_size(struct etnaviv_iommu *iommu)
 {
-	struct etnaviv_iommu_ops *ops;
-
-	ops = container_of(iommu->domain->ops, struct etnaviv_iommu_ops, ops);
-
-	return ops->dump_size(iommu->domain);
+	return iommu->domain->ops->dump_size(iommu->domain);
 }
 
 void etnaviv_iommu_dump(struct etnaviv_iommu *iommu, void *buf)
 {
-	struct etnaviv_iommu_ops *ops;
-
-	ops = container_of(iommu->domain->ops, struct etnaviv_iommu_ops, ops);
-
-	ops->dump(iommu->domain, buf);
+	iommu->domain->ops->dump(iommu->domain, buf);
 }
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_mmu.h b/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
index 54be289e5981..ab603f5166b1 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_mmu.h
@@ -17,7 +17,8 @@
 #ifndef __ETNAVIV_MMU_H__
 #define __ETNAVIV_MMU_H__
 
-#include <linux/iommu.h>
+#define ETNAVIV_PROT_READ	(1 << 0)
+#define ETNAVIV_PROT_WRITE	(1 << 1)
 
 enum etnaviv_iommu_version {
 	ETNAVIV_IOMMU_V1 = 0,
@@ -26,16 +27,31 @@ enum etnaviv_iommu_version {
 
 struct etnaviv_gpu;
 struct etnaviv_vram_mapping;
+struct etnaviv_iommu_domain;
 
-struct etnaviv_iommu_ops {
-	struct iommu_ops ops;
-	size_t (*dump_size)(struct iommu_domain *);
-	void (*dump)(struct iommu_domain *, void *);
+struct etnaviv_iommu_domain_ops {
+	void (*free)(struct etnaviv_iommu_domain *);
+	int (*map)(struct etnaviv_iommu_domain *domain, unsigned long iova,
+		   phys_addr_t paddr, size_t size, int prot);
+	size_t (*unmap)(struct etnaviv_iommu_domain *domain, unsigned long iova,
+			size_t size);
+	size_t (*dump_size)(struct etnaviv_iommu_domain *);
+	void (*dump)(struct etnaviv_iommu_domain *, void *);
+};
+
+struct etnaviv_iommu_domain {
+	struct device *dev;
+	void *bad_page_cpu;
+	dma_addr_t bad_page_dma;
+	u64 base;
+	u64 size;
+
+	const struct etnaviv_iommu_domain_ops *ops;
 };
 
 struct etnaviv_iommu {
 	struct etnaviv_gpu *gpu;
-	struct iommu_domain *domain;
+	struct etnaviv_iommu_domain *domain;
 
 	enum etnaviv_iommu_version version;
 
@@ -49,18 +65,11 @@ struct etnaviv_iommu {
 
 struct etnaviv_gem_object;
 
-int etnaviv_iommu_attach(struct etnaviv_iommu *iommu, const char **names,
-	int cnt);
-int etnaviv_iommu_map(struct etnaviv_iommu *iommu, u32 iova,
-	struct sg_table *sgt, unsigned len, int prot);
-int etnaviv_iommu_unmap(struct etnaviv_iommu *iommu, u32 iova,
-	struct sg_table *sgt, unsigned len);
 int etnaviv_iommu_map_gem(struct etnaviv_iommu *mmu,
 	struct etnaviv_gem_object *etnaviv_obj, u32 memory_base,
 	struct etnaviv_vram_mapping *mapping);
 void etnaviv_iommu_unmap_gem(struct etnaviv_iommu *mmu,
 	struct etnaviv_vram_mapping *mapping);
-void etnaviv_iommu_destroy(struct etnaviv_iommu *iommu);
 
 int etnaviv_iommu_get_suballoc_va(struct etnaviv_gpu *gpu, dma_addr_t paddr,
 				  struct drm_mm_node *vram_node, size_t size,
@@ -73,6 +82,7 @@ size_t etnaviv_iommu_dump_size(struct etnaviv_iommu *iommu);
 void etnaviv_iommu_dump(struct etnaviv_iommu *iommu, void *buf);
 
 struct etnaviv_iommu *etnaviv_iommu_new(struct etnaviv_gpu *gpu);
+void etnaviv_iommu_destroy(struct etnaviv_iommu *iommu);
 void etnaviv_iommu_restore(struct etnaviv_gpu *gpu);
 
 #endif /* __ETNAVIV_MMU_H__ */
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_perfmon.c b/drivers/gpu/drm/etnaviv/etnaviv_perfmon.c
new file mode 100644
index 000000000000..768f5aafdd18
--- /dev/null
+++ b/drivers/gpu/drm/etnaviv/etnaviv_perfmon.c
@@ -0,0 +1,495 @@
+/*
+ * Copyright (C) 2017 Etnaviv Project
+ * Copyright (C) 2017 Zodiac Inflight Innovations
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "etnaviv_gpu.h"
+#include "etnaviv_perfmon.h"
+#include "state_hi.xml.h"
+
+struct etnaviv_pm_domain;
+
+struct etnaviv_pm_signal {
+	char name[64];
+	u32 data;
+
+	u32 (*sample)(struct etnaviv_gpu *gpu,
+	              const struct etnaviv_pm_domain *domain,
+	              const struct etnaviv_pm_signal *signal);
+};
+
+struct etnaviv_pm_domain {
+	char name[64];
+
+	/* profile register */
+	u32 profile_read;
+	u32 profile_config;
+
+	u8 nr_signals;
+	const struct etnaviv_pm_signal *signal;
+};
+
+struct etnaviv_pm_domain_meta {
+	const struct etnaviv_pm_domain *domains;
+	u32 nr_domains;
+};
+
+static u32 simple_reg_read(struct etnaviv_gpu *gpu,
+	const struct etnaviv_pm_domain *domain,
+	const struct etnaviv_pm_signal *signal)
+{
+	return gpu_read(gpu, signal->data);
+}
+
+static u32 perf_reg_read(struct etnaviv_gpu *gpu,
+	const struct etnaviv_pm_domain *domain,
+	const struct etnaviv_pm_signal *signal)
+{
+	gpu_write(gpu, domain->profile_config, signal->data);
+
+	return gpu_read(gpu, domain->profile_read);
+}
+
+static u32 pipe_reg_read(struct etnaviv_gpu *gpu,
+	const struct etnaviv_pm_domain *domain,
+	const struct etnaviv_pm_signal *signal)
+{
+	u32 clock = gpu_read(gpu, VIVS_HI_CLOCK_CONTROL);
+	u32 value = 0;
+	unsigned i;
+
+	for (i = 0; i < gpu->identity.pixel_pipes; i++) {
+		clock &= ~(VIVS_HI_CLOCK_CONTROL_DEBUG_PIXEL_PIPE__MASK);
+		clock |= VIVS_HI_CLOCK_CONTROL_DEBUG_PIXEL_PIPE(i);
+		gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, clock);
+		gpu_write(gpu, domain->profile_config, signal->data);
+		value += gpu_read(gpu, domain->profile_read);
+	}
+
+	/* switch back to pixel pipe 0 to prevent GPU hang */
+	clock &= ~(VIVS_HI_CLOCK_CONTROL_DEBUG_PIXEL_PIPE__MASK);
+	clock |= VIVS_HI_CLOCK_CONTROL_DEBUG_PIXEL_PIPE(0);
+	gpu_write(gpu, VIVS_HI_CLOCK_CONTROL, clock);
+
+	return value;
+}
+
+static const struct etnaviv_pm_domain doms_3d[] = {
+	{
+		.name = "HI",
+		.profile_read = VIVS_MC_PROFILE_HI_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG2,
+		.nr_signals = 5,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"TOTAL_CYCLES",
+				VIVS_HI_PROFILE_TOTAL_CYCLES,
+				&simple_reg_read
+			},
+			{
+				"IDLE_CYCLES",
+				VIVS_HI_PROFILE_IDLE_CYCLES,
+				&simple_reg_read
+			},
+			{
+				"AXI_CYCLES_READ_REQUEST_STALLED",
+				VIVS_MC_PROFILE_CONFIG2_HI_AXI_CYCLES_READ_REQUEST_STALLED,
+				&perf_reg_read
+			},
+			{
+				"AXI_CYCLES_WRITE_REQUEST_STALLED",
+				VIVS_MC_PROFILE_CONFIG2_HI_AXI_CYCLES_WRITE_REQUEST_STALLED,
+				&perf_reg_read
+			},
+			{
+				"AXI_CYCLES_WRITE_DATA_STALLED",
+				VIVS_MC_PROFILE_CONFIG2_HI_AXI_CYCLES_WRITE_DATA_STALLED,
+				&perf_reg_read
+			}
+		}
+	},
+	{
+		.name = "PE",
+		.profile_read = VIVS_MC_PROFILE_PE_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG0,
+		.nr_signals = 5,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"PIXEL_COUNT_KILLED_BY_COLOR_PIPE",
+				VIVS_MC_PROFILE_CONFIG0_PE_PIXEL_COUNT_KILLED_BY_COLOR_PIPE,
+				&pipe_reg_read
+			},
+			{
+				"PIXEL_COUNT_KILLED_BY_DEPTH_PIPE",
+				VIVS_MC_PROFILE_CONFIG0_PE_PIXEL_COUNT_KILLED_BY_DEPTH_PIPE,
+				&pipe_reg_read
+			},
+			{
+				"PIXEL_COUNT_DRAWN_BY_COLOR_PIPE",
+				VIVS_MC_PROFILE_CONFIG0_PE_PIXEL_COUNT_DRAWN_BY_COLOR_PIPE,
+				&pipe_reg_read
+			},
+			{
+				"PIXEL_COUNT_DRAWN_BY_DEPTH_PIPE",
+				VIVS_MC_PROFILE_CONFIG0_PE_PIXEL_COUNT_DRAWN_BY_DEPTH_PIPE,
+				&pipe_reg_read
+			}
+		}
+	},
+	{
+		.name = "SH",
+		.profile_read = VIVS_MC_PROFILE_SH_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG0,
+		.nr_signals = 9,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"SHADER_CYCLES",
+				VIVS_MC_PROFILE_CONFIG0_SH_SHADER_CYCLES,
+				&perf_reg_read
+			},
+			{
+				"PS_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_PS_INST_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"RENDERED_PIXEL_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_RENDERED_PIXEL_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"VS_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_VS_INST_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"RENDERED_VERTICE_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_RENDERED_VERTICE_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"VTX_BRANCH_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_VTX_BRANCH_INST_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"VTX_TEXLD_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_VTX_TEXLD_INST_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"PXL_BRANCH_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_PXL_BRANCH_INST_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"PXL_TEXLD_INST_COUNTER",
+				VIVS_MC_PROFILE_CONFIG0_SH_PXL_TEXLD_INST_COUNTER,
+				&pipe_reg_read
+			}
+		}
+	},
+	{
+		.name = "PA",
+		.profile_read = VIVS_MC_PROFILE_PA_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG1,
+		.nr_signals = 6,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"INPUT_VTX_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_INPUT_VTX_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"INPUT_PRIM_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_INPUT_PRIM_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"OUTPUT_PRIM_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_OUTPUT_PRIM_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"DEPTH_CLIPPED_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_DEPTH_CLIPPED_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"TRIVIAL_REJECTED_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_TRIVIAL_REJECTED_COUNTER,
+				&pipe_reg_read
+			},
+			{
+				"CULLED_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_PA_CULLED_COUNTER,
+				&pipe_reg_read
+			}
+		}
+	},
+	{
+		.name = "SE",
+		.profile_read = VIVS_MC_PROFILE_SE_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG1,
+		.nr_signals = 2,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"CULLED_TRIANGLE_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_SE_CULLED_TRIANGLE_COUNT,
+				&perf_reg_read
+			},
+			{
+				"CULLED_LINES_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_SE_CULLED_LINES_COUNT,
+				&perf_reg_read
+			}
+		}
+	},
+	{
+		.name = "RA",
+		.profile_read = VIVS_MC_PROFILE_RA_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG1,
+		.nr_signals = 7,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"VALID_PIXEL_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_RA_VALID_PIXEL_COUNT,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_QUAD_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_RA_TOTAL_QUAD_COUNT,
+				&perf_reg_read
+			},
+			{
+				"VALID_QUAD_COUNT_AFTER_EARLY_Z",
+				VIVS_MC_PROFILE_CONFIG1_RA_VALID_QUAD_COUNT_AFTER_EARLY_Z,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_PRIMITIVE_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_RA_TOTAL_PRIMITIVE_COUNT,
+				&perf_reg_read
+			},
+			{
+				"PIPE_CACHE_MISS_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_RA_PIPE_CACHE_MISS_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"PREFETCH_CACHE_MISS_COUNTER",
+				VIVS_MC_PROFILE_CONFIG1_RA_PREFETCH_CACHE_MISS_COUNTER,
+				&perf_reg_read
+			},
+			{
+				"CULLED_QUAD_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_RA_CULLED_QUAD_COUNT,
+				&perf_reg_read
+			}
+		}
+	},
+	{
+		.name = "TX",
+		.profile_read = VIVS_MC_PROFILE_TX_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG1,
+		.nr_signals = 9,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"TOTAL_BILINEAR_REQUESTS",
+				VIVS_MC_PROFILE_CONFIG1_TX_TOTAL_BILINEAR_REQUESTS,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_TRILINEAR_REQUESTS",
+				VIVS_MC_PROFILE_CONFIG1_TX_TOTAL_TRILINEAR_REQUESTS,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_DISCARDED_TEXTURE_REQUESTS",
+				VIVS_MC_PROFILE_CONFIG1_TX_TOTAL_DISCARDED_TEXTURE_REQUESTS,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_TEXTURE_REQUESTS",
+				VIVS_MC_PROFILE_CONFIG1_TX_TOTAL_TEXTURE_REQUESTS,
+				&perf_reg_read
+			},
+			{
+				"MEM_READ_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_TX_MEM_READ_COUNT,
+				&perf_reg_read
+			},
+			{
+				"MEM_READ_IN_8B_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_TX_MEM_READ_IN_8B_COUNT,
+				&perf_reg_read
+			},
+			{
+				"CACHE_MISS_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_TX_CACHE_MISS_COUNT,
+				&perf_reg_read
+			},
+			{
+				"CACHE_HIT_TEXEL_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_TX_CACHE_HIT_TEXEL_COUNT,
+				&perf_reg_read
+			},
+			{
+				"CACHE_MISS_TEXEL_COUNT",
+				VIVS_MC_PROFILE_CONFIG1_TX_CACHE_MISS_TEXEL_COUNT,
+				&perf_reg_read
+			}
+		}
+	},
+	{
+		.name = "MC",
+		.profile_read = VIVS_MC_PROFILE_MC_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG2,
+		.nr_signals = 3,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"TOTAL_READ_REQ_8B_FROM_PIPELINE",
+				VIVS_MC_PROFILE_CONFIG2_MC_TOTAL_READ_REQ_8B_FROM_PIPELINE,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_READ_REQ_8B_FROM_IP",
+				VIVS_MC_PROFILE_CONFIG2_MC_TOTAL_READ_REQ_8B_FROM_IP,
+				&perf_reg_read
+			},
+			{
+				"TOTAL_WRITE_REQ_8B_FROM_PIPELINE",
+				VIVS_MC_PROFILE_CONFIG2_MC_TOTAL_WRITE_REQ_8B_FROM_PIPELINE,
+				&perf_reg_read
+			}
+		}
+	}
+};
+
+static const struct etnaviv_pm_domain doms_2d[] = {
+	{
+		.name = "PE",
+		.profile_read = VIVS_MC_PROFILE_PE_READ,
+		.profile_config = VIVS_MC_PROFILE_CONFIG0,
+		.nr_signals = 1,
+		.signal = (const struct etnaviv_pm_signal[]) {
+			{
+				"PIXELS_RENDERED_2D",
+				VIVS_MC_PROFILE_CONFIG0_PE_PIXELS_RENDERED_2D,
+				&pipe_reg_read
+			}
+		}
+	}
+};
+
+static const struct etnaviv_pm_domain doms_vg[] = {
+};
+
+static const struct etnaviv_pm_domain_meta doms_meta[] = {
+	{
+		.nr_domains = ARRAY_SIZE(doms_3d),
+		.domains = &doms_3d[0]
+	},
+	{
+		.nr_domains = ARRAY_SIZE(doms_2d),
+		.domains = &doms_2d[0]
+	},
+	{
+		.nr_domains = ARRAY_SIZE(doms_vg),
+		.domains = &doms_vg[0]
+	}
+};
+
+int etnaviv_pm_query_dom(struct etnaviv_gpu *gpu,
+	struct drm_etnaviv_pm_domain *domain)
+{
+	const struct etnaviv_pm_domain_meta *meta = &doms_meta[domain->pipe];
+	const struct etnaviv_pm_domain *dom;
+
+	if (domain->iter >= meta->nr_domains)
+		return -EINVAL;
+
+	dom = meta->domains + domain->iter;
+
+	domain->id = domain->iter;
+	domain->nr_signals = dom->nr_signals;
+	strncpy(domain->name, dom->name, sizeof(domain->name));
+
+	domain->iter++;
+	if (domain->iter == meta->nr_domains)
+		domain->iter = 0xff;
+
+	return 0;
+}
+
+int etnaviv_pm_query_sig(struct etnaviv_gpu *gpu,
+	struct drm_etnaviv_pm_signal *signal)
+{
+	const struct etnaviv_pm_domain_meta *meta = &doms_meta[signal->pipe];
+	const struct etnaviv_pm_domain *dom;
+	const struct etnaviv_pm_signal *sig;
+
+	if (signal->domain >= meta->nr_domains)
+		return -EINVAL;
+
+	dom = meta->domains + signal->domain;
+
+	if (signal->iter > dom->nr_signals)
+		return -EINVAL;
+
+	sig = &dom->signal[signal->iter];
+
+	signal->id = signal->iter;
+	strncpy(signal->name, sig->name, sizeof(signal->name));
+
+	signal->iter++;
+	if (signal->iter == dom->nr_signals)
+		signal->iter = 0xffff;
+
+	return 0;
+}
+
+int etnaviv_pm_req_validate(const struct drm_etnaviv_gem_submit_pmr *r,
+	u32 exec_state)
+{
+	const struct etnaviv_pm_domain_meta *meta = &doms_meta[exec_state];
+	const struct etnaviv_pm_domain *dom;
+
+	if (r->domain >= meta->nr_domains)
+		return -EINVAL;
+
+	dom = meta->domains + r->domain;
+
+	if (r->signal > dom->nr_signals)
+		return -EINVAL;
+
+	return 0;
+}
+
+void etnaviv_perfmon_process(struct etnaviv_gpu *gpu,
+	const struct etnaviv_perfmon_request *pmr)
+{
+	const struct etnaviv_pm_domain_meta *meta = &doms_meta[gpu->exec_state];
+	const struct etnaviv_pm_domain *dom;
+	const struct etnaviv_pm_signal *sig;
+	u32 *bo = pmr->bo_vma;
+	u32 val;
+
+	dom = meta->domains + pmr->domain;
+	sig = &dom->signal[pmr->signal];
+	val = sig->sample(gpu, dom, sig);
+
+	*(bo + pmr->offset) = val;
+}
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_perfmon.h b/drivers/gpu/drm/etnaviv/etnaviv_perfmon.h
new file mode 100644
index 000000000000..35dce194cb00
--- /dev/null
+++ b/drivers/gpu/drm/etnaviv/etnaviv_perfmon.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright (C) 2017 Etnaviv Project
+ * Copyright (C) 2017 Zodiac Inflight Innovations
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __ETNAVIV_PERFMON_H__
+#define __ETNAVIV_PERFMON_H__
+
+struct etnaviv_gpu;
+struct drm_etnaviv_pm_domain;
+struct drm_etnaviv_pm_signal;
+
+struct etnaviv_perfmon_request
+{
+	u32 flags;
+	u8 domain;
+	u8 signal;
+	u32 sequence;
+
+	/* bo to store a value */
+	u32 *bo_vma;
+	u32 offset;
+};
+
+int etnaviv_pm_query_dom(struct etnaviv_gpu *gpu,
+	struct drm_etnaviv_pm_domain *domain);
+
+int etnaviv_pm_query_sig(struct etnaviv_gpu *gpu,
+	struct drm_etnaviv_pm_signal *signal);
+
+int etnaviv_pm_req_validate(const struct drm_etnaviv_gem_submit_pmr *r,
+	u32 exec_state);
+
+void etnaviv_perfmon_process(struct etnaviv_gpu *gpu,
+	const struct etnaviv_perfmon_request *pmr);
+
+#endif /* __ETNAVIV_PERFMON_H__ */