diff options
Diffstat (limited to 'drivers/gpu/drm')
378 files changed, 6506 insertions, 2748 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index c8ad6bf6618a..bcef6ea4bcf9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -82,6 +82,7 @@ #include "amdgpu_bo_list.h" #include "amdgpu_gem.h" #include "amdgpu_doorbell.h" +#include "amdgpu_amdkfd.h" #define MAX_GPU_INSTANCE 16 @@ -163,6 +164,7 @@ extern int amdgpu_si_support; extern int amdgpu_cik_support; #endif +#define AMDGPU_VM_MAX_NUM_CTX 4096 #define AMDGPU_SG_THRESHOLD (256*1024*1024) #define AMDGPU_DEFAULT_GTT_SIZE_MB 3072ULL /* 3GB by default */ #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000 @@ -235,7 +237,7 @@ enum amdgpu_kiq_irq { #define MAX_KIQ_REG_WAIT 5000 /* in usecs, 5ms */ #define MAX_KIQ_REG_BAILOUT_INTERVAL 5 /* in msecs, 5ms */ -#define MAX_KIQ_REG_TRY 20 +#define MAX_KIQ_REG_TRY 80 /* 20 -> 80 */ int amdgpu_device_ip_set_clockgating_state(void *dev, enum amd_ip_block_type block_type, @@ -862,6 +864,9 @@ struct amdgpu_device { /* GDS */ struct amdgpu_gds gds; + /* KFD */ + struct amdgpu_kfd_dev kfd; + /* display related functionality */ struct amdgpu_display_manager dm; @@ -875,9 +880,6 @@ struct amdgpu_device { atomic64_t visible_pin_size; atomic64_t gart_pin_size; - /* amdkfd interface */ - struct kfd_dev *kfd; - /* soc15 register offset based on ip, instance and segment */ uint32_t *reg_offset[MAX_HWIP][HWIP_MAX_INSTANCE]; @@ -910,7 +912,9 @@ struct amdgpu_device { bool in_gpu_reset; struct mutex lock_reset; struct amdgpu_doorbell_index doorbell_index; + int asic_reset_res; + struct work_struct xgmi_reset_work; }; static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c index 47db65926d71..4376b17ca594 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c @@ -886,6 +886,5 @@ void amdgpu_acpi_get_backlight_caps(struct amdgpu_device *adev, void amdgpu_acpi_fini(struct amdgpu_device *adev) { unregister_acpi_notifier(&adev->acpi_nb); - if (adev->atif) - kfree(adev->atif); + kfree(adev->atif); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index d693b8047653..2dfaf158ef07 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -26,15 +26,26 @@ #include "amdgpu.h" #include "amdgpu_gfx.h" #include <linux/module.h> +#include <linux/dma-buf.h> const struct kgd2kfd_calls *kgd2kfd; static const unsigned int compute_vmid_bitmap = 0xFF00; +/* Total memory size in system memory and all GPU VRAM. Used to + * estimate worst case amount of memory to reserve for page tables + */ +uint64_t amdgpu_amdkfd_total_mem_size; + int amdgpu_amdkfd_init(void) { + struct sysinfo si; int ret; + si_meminfo(&si); + amdgpu_amdkfd_total_mem_size = si.totalram - si.totalhigh; + amdgpu_amdkfd_total_mem_size *= si.mem_unit; + #ifdef CONFIG_HSA_AMD ret = kgd2kfd_init(KFD_INTERFACE_VERSION, &kgd2kfd); if (ret) @@ -87,8 +98,11 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) return; } - adev->kfd = kgd2kfd->probe((struct kgd_dev *)adev, - adev->pdev, kfd2kgd); + adev->kfd.dev = kgd2kfd->probe((struct kgd_dev *)adev, + adev->pdev, kfd2kgd); + + if (adev->kfd.dev) + amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size; } /** @@ -128,7 +142,8 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) { int i, n; int last_valid_bit; - if (adev->kfd) { + + if (adev->kfd.dev) { struct kgd2kfd_shared_resources gpu_resources = { .compute_vmid_bitmap = compute_vmid_bitmap, .num_pipe_per_mec = adev->gfx.mec.num_pipe_per_mec, @@ -167,7 +182,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) &gpu_resources.doorbell_start_offset); if (adev->asic_type < CHIP_VEGA10) { - kgd2kfd->device_init(adev->kfd, &gpu_resources); + kgd2kfd->device_init(adev->kfd.dev, &gpu_resources); return; } @@ -196,37 +211,37 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) gpu_resources.reserved_doorbell_mask = 0x1e0; gpu_resources.reserved_doorbell_val = 0x0e0; - kgd2kfd->device_init(adev->kfd, &gpu_resources); + kgd2kfd->device_init(adev->kfd.dev, &gpu_resources); } } void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev) { - if (adev->kfd) { - kgd2kfd->device_exit(adev->kfd); - adev->kfd = NULL; + if (adev->kfd.dev) { + kgd2kfd->device_exit(adev->kfd.dev); + adev->kfd.dev = NULL; } } void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev, const void *ih_ring_entry) { - if (adev->kfd) - kgd2kfd->interrupt(adev->kfd, ih_ring_entry); + if (adev->kfd.dev) + kgd2kfd->interrupt(adev->kfd.dev, ih_ring_entry); } void amdgpu_amdkfd_suspend(struct amdgpu_device *adev) { - if (adev->kfd) - kgd2kfd->suspend(adev->kfd); + if (adev->kfd.dev) + kgd2kfd->suspend(adev->kfd.dev); } int amdgpu_amdkfd_resume(struct amdgpu_device *adev) { int r = 0; - if (adev->kfd) - r = kgd2kfd->resume(adev->kfd); + if (adev->kfd.dev) + r = kgd2kfd->resume(adev->kfd.dev); return r; } @@ -235,8 +250,8 @@ int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev) { int r = 0; - if (adev->kfd) - r = kgd2kfd->pre_reset(adev->kfd); + if (adev->kfd.dev) + r = kgd2kfd->pre_reset(adev->kfd.dev); return r; } @@ -245,8 +260,8 @@ int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev) { int r = 0; - if (adev->kfd) - r = kgd2kfd->post_reset(adev->kfd); + if (adev->kfd.dev) + r = kgd2kfd->post_reset(adev->kfd.dev); return r; } @@ -419,6 +434,62 @@ void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info) cu_info->lds_size = acu_info.lds_size; } +int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd, + struct kgd_dev **dma_buf_kgd, + uint64_t *bo_size, void *metadata_buffer, + size_t buffer_size, uint32_t *metadata_size, + uint32_t *flags) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)kgd; + struct dma_buf *dma_buf; + struct drm_gem_object *obj; + struct amdgpu_bo *bo; + uint64_t metadata_flags; + int r = -EINVAL; + + dma_buf = dma_buf_get(dma_buf_fd); + if (IS_ERR(dma_buf)) + return PTR_ERR(dma_buf); + + if (dma_buf->ops != &amdgpu_dmabuf_ops) + /* Can't handle non-graphics buffers */ + goto out_put; + + obj = dma_buf->priv; + if (obj->dev->driver != adev->ddev->driver) + /* Can't handle buffers from different drivers */ + goto out_put; + + adev = obj->dev->dev_private; + bo = gem_to_amdgpu_bo(obj); + if (!(bo->preferred_domains & (AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT))) + /* Only VRAM and GTT BOs are supported */ + goto out_put; + + r = 0; + if (dma_buf_kgd) + *dma_buf_kgd = (struct kgd_dev *)adev; + if (bo_size) + *bo_size = amdgpu_bo_size(bo); + if (metadata_size) + *metadata_size = bo->metadata_size; + if (metadata_buffer) + r = amdgpu_bo_get_metadata(bo, metadata_buffer, buffer_size, + metadata_size, &metadata_flags); + if (flags) { + *flags = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ? + ALLOC_MEM_FLAGS_VRAM : ALLOC_MEM_FLAGS_GTT; + + if (bo->flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) + *flags |= ALLOC_MEM_FLAGS_PUBLIC; + } + +out_put: + dma_buf_put(dma_buf); + return r; +} + uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd) { struct amdgpu_device *adev = (struct amdgpu_device *)kgd; @@ -501,7 +572,7 @@ void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle) bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid) { - if (adev->kfd) { + if (adev->kfd.dev) { if ((1 << vmid) & compute_vmid_bitmap) return true; } @@ -515,7 +586,7 @@ bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm) return false; } -void amdgpu_amdkfd_unreserve_system_memory_limit(struct amdgpu_bo *bo) +void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo) { } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index bcf587b4ba98..70429f7aa9a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -27,7 +27,6 @@ #include <linux/types.h> #include <linux/mm.h> -#include <linux/mmu_context.h> #include <linux/workqueue.h> #include <kgd_kfd_interface.h> #include <drm/ttm/ttm_execbuf_util.h> @@ -35,6 +34,7 @@ #include "amdgpu_vm.h" extern const struct kgd2kfd_calls *kgd2kfd; +extern uint64_t amdgpu_amdkfd_total_mem_size; struct amdgpu_device; @@ -77,6 +77,11 @@ struct amdgpu_amdkfd_fence { char timeline_name[TASK_COMM_LEN]; }; +struct amdgpu_kfd_dev { + struct kfd_dev *dev; + uint64_t vram_used; +}; + struct amdgpu_amdkfd_fence *amdgpu_amdkfd_fence_create(u64 context, struct mm_struct *mm); bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm); @@ -144,6 +149,11 @@ uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd); uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd); void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info); +int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd, + struct kgd_dev **dmabuf_kgd, + uint64_t *bo_size, void *metadata_buffer, + size_t buffer_size, uint32_t *metadata_size, + uint32_t *flags); uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd); uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd); @@ -195,7 +205,13 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *process_info, int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, struct kfd_vm_fault_info *info); +int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, + struct dma_buf *dmabuf, + uint64_t va, void *vm, + struct kgd_mem **mem, uint64_t *size, + uint64_t *mmap_offset); + void amdgpu_amdkfd_gpuvm_init_mem_limits(void); -void amdgpu_amdkfd_unreserve_system_memory_limit(struct amdgpu_bo *bo); +void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo); #endif /* AMDGPU_AMDKFD_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c index 72a357dae070..ff7fac7df34b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c @@ -23,6 +23,7 @@ #include <linux/fdtable.h> #include <linux/uaccess.h> #include <linux/firmware.h> +#include <linux/mmu_context.h> #include <drm/drmP.h> #include "amdgpu.h" #include "amdgpu_amdkfd.h" diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c index 0e2a56b6a9b6..56ea929f524b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c @@ -24,6 +24,7 @@ #include <linux/fdtable.h> #include <linux/uaccess.h> #include <linux/firmware.h> +#include <linux/mmu_context.h> #include <drm/drmP.h> #include "amdgpu.h" #include "amdgpu_amdkfd.h" diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c index 03b604c96d94..5c51d4910650 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c @@ -26,6 +26,7 @@ #include <linux/fdtable.h> #include <linux/uaccess.h> #include <linux/firmware.h> +#include <linux/mmu_context.h> #include <drm/drmP.h> #include "amdgpu.h" #include "amdgpu_amdkfd.h" diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index f3129b912714..be1ab43473c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -25,6 +25,7 @@ #include <linux/list.h> #include <linux/pagemap.h> #include <linux/sched/mm.h> +#include <linux/dma-buf.h> #include <drm/drmP.h> #include "amdgpu_object.h" #include "amdgpu_vm.h" @@ -110,17 +111,17 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) (kfd_mem_limit.max_ttm_mem_limit >> 20)); } -static int amdgpu_amdkfd_reserve_system_mem_limit(struct amdgpu_device *adev, +static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev, uint64_t size, u32 domain, bool sg) { - size_t acc_size, system_mem_needed, ttm_mem_needed; + size_t acc_size, system_mem_needed, ttm_mem_needed, vram_needed; + uint64_t reserved_for_pt = amdgpu_amdkfd_total_mem_size >> 9; int ret = 0; acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size, sizeof(struct amdgpu_bo)); - spin_lock(&kfd_mem_limit.mem_limit_lock); - + vram_needed = 0; if (domain == AMDGPU_GEM_DOMAIN_GTT) { /* TTM GTT memory */ system_mem_needed = acc_size + size; @@ -133,23 +134,30 @@ static int amdgpu_amdkfd_reserve_system_mem_limit(struct amdgpu_device *adev, /* VRAM and SG */ system_mem_needed = acc_size; ttm_mem_needed = acc_size; + if (domain == AMDGPU_GEM_DOMAIN_VRAM) + vram_needed = size; } + spin_lock(&kfd_mem_limit.mem_limit_lock); + if ((kfd_mem_limit.system_mem_used + system_mem_needed > - kfd_mem_limit.max_system_mem_limit) || - (kfd_mem_limit.ttm_mem_used + ttm_mem_needed > - kfd_mem_limit.max_ttm_mem_limit)) + kfd_mem_limit.max_system_mem_limit) || + (kfd_mem_limit.ttm_mem_used + ttm_mem_needed > + kfd_mem_limit.max_ttm_mem_limit) || + (adev->kfd.vram_used + vram_needed > + adev->gmc.real_vram_size - reserved_for_pt)) { ret = -ENOMEM; - else { + } else { kfd_mem_limit.system_mem_used += system_mem_needed; kfd_mem_limit.ttm_mem_used += ttm_mem_needed; + adev->kfd.vram_used += vram_needed; } spin_unlock(&kfd_mem_limit.mem_limit_lock); return ret; } -static void unreserve_system_mem_limit(struct amdgpu_device *adev, +static void unreserve_mem_limit(struct amdgpu_device *adev, uint64_t size, u32 domain, bool sg) { size_t acc_size; @@ -167,6 +175,11 @@ static void unreserve_system_mem_limit(struct amdgpu_device *adev, } else { kfd_mem_limit.system_mem_used -= acc_size; kfd_mem_limit.ttm_mem_used -= acc_size; + if (domain == AMDGPU_GEM_DOMAIN_VRAM) { + adev->kfd.vram_used -= size; + WARN_ONCE(adev->kfd.vram_used < 0, + "kfd VRAM memory accounting unbalanced"); + } } WARN_ONCE(kfd_mem_limit.system_mem_used < 0, "kfd system memory accounting unbalanced"); @@ -176,29 +189,18 @@ static void unreserve_system_mem_limit(struct amdgpu_device *adev, spin_unlock(&kfd_mem_limit.mem_limit_lock); } -void amdgpu_amdkfd_unreserve_system_memory_limit(struct amdgpu_bo *bo) +void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo) { - spin_lock(&kfd_mem_limit.mem_limit_lock); + struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + u32 domain = bo->preferred_domains; + bool sg = (bo->preferred_domains == AMDGPU_GEM_DOMAIN_CPU); if (bo->flags & AMDGPU_AMDKFD_USERPTR_BO) { - kfd_mem_limit.system_mem_used -= - (bo->tbo.acc_size + amdgpu_bo_size(bo)); - kfd_mem_limit.ttm_mem_used -= bo->tbo.acc_size; - } else if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT) { - kfd_mem_limit.system_mem_used -= - (bo->tbo.acc_size + amdgpu_bo_size(bo)); - kfd_mem_limit.ttm_mem_used -= - (bo->tbo.acc_size + amdgpu_bo_size(bo)); - } else { - kfd_mem_limit.system_mem_used -= bo->tbo.acc_size; - kfd_mem_limit.ttm_mem_used -= bo->tbo.acc_size; + domain = AMDGPU_GEM_DOMAIN_CPU; + sg = false; } - WARN_ONCE(kfd_mem_limit.system_mem_used < 0, - "kfd system memory accounting unbalanced"); - WARN_ONCE(kfd_mem_limit.ttm_mem_used < 0, - "kfd TTM memory accounting unbalanced"); - spin_unlock(&kfd_mem_limit.mem_limit_lock); + unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg); } @@ -535,7 +537,7 @@ static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem, struct amdgpu_bo *bo = mem->bo; INIT_LIST_HEAD(&entry->head); - entry->shared = true; + entry->num_shared = 1; entry->bo = &bo->tbo; mutex_lock(&process_info->lock); if (userptr) @@ -676,7 +678,7 @@ static int reserve_bo_and_vm(struct kgd_mem *mem, ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; - ctx->kfd_bo.tv.shared = true; + ctx->kfd_bo.tv.num_shared = 1; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); @@ -740,7 +742,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem, ctx->kfd_bo.priority = 0; ctx->kfd_bo.tv.bo = &bo->tbo; - ctx->kfd_bo.tv.shared = true; + ctx->kfd_bo.tv.num_shared = 1; ctx->kfd_bo.user_pages = NULL; list_add(&ctx->kfd_bo.tv.head, &ctx->list); @@ -885,6 +887,24 @@ update_gpuvm_pte_failed: return ret; } +static struct sg_table *create_doorbell_sg(uint64_t addr, uint32_t size) +{ + struct sg_table *sg = kmalloc(sizeof(*sg), GFP_KERNEL); + + if (!sg) + return NULL; + if (sg_alloc_table(sg, 1, GFP_KERNEL)) { + kfree(sg); + return NULL; + } + sg->sgl->dma_address = addr; + sg->sgl->length = size; +#ifdef CONFIG_NEED_SG_DMA_LENGTH + sg->sgl->dma_length = size; +#endif + return sg; +} + static int process_validate_vms(struct amdkfd_process_info *process_info) { struct amdgpu_vm *peer_vm; @@ -1168,6 +1188,8 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; + enum ttm_bo_type bo_type = ttm_bo_type_device; + struct sg_table *sg = NULL; uint64_t user_addr = 0; struct amdgpu_bo *bo; struct amdgpu_bo_param bp; @@ -1196,13 +1218,25 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( if (!offset || !*offset) return -EINVAL; user_addr = *offset; + } else if (flags & ALLOC_MEM_FLAGS_DOORBELL) { + domain = AMDGPU_GEM_DOMAIN_GTT; + alloc_domain = AMDGPU_GEM_DOMAIN_CPU; + bo_type = ttm_bo_type_sg; + alloc_flags = 0; + if (size > UINT_MAX) + return -EINVAL; + sg = create_doorbell_sg(*offset, size); + if (!sg) + return -ENOMEM; } else { return -EINVAL; } *mem = kzalloc(sizeof(struct kgd_mem), GFP_KERNEL); - if (!*mem) - return -ENOMEM; + if (!*mem) { + ret = -ENOMEM; + goto err; + } INIT_LIST_HEAD(&(*mem)->bo_va_list); mutex_init(&(*mem)->lock); (*mem)->aql_queue = !!(flags & ALLOC_MEM_FLAGS_AQL_QUEUE_MEM); @@ -1235,8 +1269,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( amdgpu_sync_create(&(*mem)->sync); - ret = amdgpu_amdkfd_reserve_system_mem_limit(adev, size, - alloc_domain, false); + ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, alloc_domain, !!sg); if (ret) { pr_debug("Insufficient system memory\n"); goto err_reserve_limit; @@ -1250,7 +1283,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( bp.byte_align = byte_align; bp.domain = alloc_domain; bp.flags = alloc_flags; - bp.type = ttm_bo_type_device; + bp.type = bo_type; bp.resv = NULL; ret = amdgpu_bo_create(adev, &bp, &bo); if (ret) { @@ -1258,6 +1291,10 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( domain_string(alloc_domain), ret); goto err_bo_create; } + if (bo_type == ttm_bo_type_sg) { + bo->tbo.sg = sg; + bo->tbo.ttm->sg = sg; + } bo->kfd_bo = *mem; (*mem)->bo = bo; if (user_addr) @@ -1289,10 +1326,15 @@ allocate_init_user_pages_failed: /* Don't unreserve system mem limit twice */ goto err_reserve_limit; err_bo_create: - unreserve_system_mem_limit(adev, size, alloc_domain, false); + unreserve_mem_limit(adev, size, alloc_domain, !!sg); err_reserve_limit: mutex_destroy(&(*mem)->lock); kfree(*mem); +err: + if (sg) { + sg_free_table(sg); + kfree(sg); + } return ret; } @@ -1362,6 +1404,14 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( /* Free the sync object */ amdgpu_sync_free(&mem->sync); + /* If the SG is not NULL, it's one we created for a doorbell + * BO. We need to free it. + */ + if (mem->bo->tbo.sg) { + sg_free_table(mem->bo->tbo.sg); + kfree(mem->bo->tbo.sg); + } + /* Free the BO*/ amdgpu_bo_unref(&mem->bo); mutex_destroy(&mem->lock); @@ -1664,6 +1714,60 @@ int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd, return 0; } +int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, + struct dma_buf *dma_buf, + uint64_t va, void *vm, + struct kgd_mem **mem, uint64_t *size, + uint64_t *mmap_offset) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)kgd; + struct drm_gem_object *obj; + struct amdgpu_bo *bo; + struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; + + if (dma_buf->ops != &amdgpu_dmabuf_ops) + /* Can't handle non-graphics buffers */ + return -EINVAL; + + obj = dma_buf->priv; + if (obj->dev->dev_private != adev) + /* Can't handle buffers from other devices */ + return -EINVAL; + + bo = gem_to_amdgpu_bo(obj); + if (!(bo->preferred_domains & (AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT))) + /* Only VRAM and GTT BOs are supported */ + return -EINVAL; + + *mem = kzalloc(sizeof(struct kgd_mem), GFP_KERNEL); + if (!*mem) + return -ENOMEM; + + if (size) + *size = amdgpu_bo_size(bo); + + if (mmap_offset) + *mmap_offset = amdgpu_bo_mmap_offset(bo); + + INIT_LIST_HEAD(&(*mem)->bo_va_list); + mutex_init(&(*mem)->lock); + (*mem)->mapping_flags = + AMDGPU_VM_PAGE_READABLE | AMDGPU_VM_PAGE_WRITEABLE | + AMDGPU_VM_PAGE_EXECUTABLE | AMDGPU_VM_MTYPE_NC; + + (*mem)->bo = amdgpu_bo_ref(bo); + (*mem)->va = va; + (*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ? + AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT; + (*mem)->mapped_to_gpu_memory = 0; + (*mem)->process_info = avm->process_info; + add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, false); + amdgpu_sync_create(&(*mem)->sync); + + return 0; +} + /* Evict a userptr BO by stopping the queues if necessary * * Runs in MMU notifier, may be in RECLAIM_FS context. This means it @@ -1830,7 +1934,7 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) validate_list.head) { list_add_tail(&mem->resv_list.head, &resv_list); mem->resv_list.bo = mem->validate_list.bo; - mem->resv_list.shared = mem->validate_list.shared; + mem->resv_list.num_shared = mem->validate_list.num_shared; } /* Reserve all BOs and page tables for validation */ @@ -2049,7 +2153,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef) list_add_tail(&mem->resv_list.head, &ctx.list); mem->resv_list.bo = mem->validate_list.bo; - mem->resv_list.shared = mem->validate_list.shared; + mem->resv_list.num_shared = mem->validate_list.num_shared; } ret = ttm_eu_reserve_buffers(&ctx.ticket, &ctx.list, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c index a028661d9e20..92b11de19581 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c @@ -576,6 +576,7 @@ static const struct amdgpu_px_quirk amdgpu_px_quirk_list[] = { { 0x1002, 0x6900, 0x1028, 0x0812, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0x1002, 0x6900, 0x1028, 0x0813, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0x1002, 0x6900, 0x1025, 0x125A, AMDGPU_PX_QUIRK_FORCE_ATPX }, + { 0x1002, 0x6900, 0x17AA, 0x3806, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0, 0, 0, 0, 0 }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index 14d2982a47cc..5c79da8e1150 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -118,7 +118,6 @@ int amdgpu_bo_list_create(struct amdgpu_device *adev, struct drm_file *filp, entry->priority = min(info[i].bo_priority, AMDGPU_BO_LIST_MAX_PRIORITY); entry->tv.bo = &bo->tbo; - entry->tv.shared = !bo->prime_shared_count; if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GDS) list->gds_obj = bo; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c index ceadeeadfa56..387f1cf1dc20 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c @@ -381,7 +381,8 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, (adev->pdev->revision == 0xe7) || (adev->pdev->revision == 0xef))) || ((adev->pdev->device == 0x6fdf) && - (adev->pdev->revision == 0xef))) { + ((adev->pdev->revision == 0xef) || + (adev->pdev->revision == 0xff)))) { info->is_kicker = true; strcpy(fw_name, "amdgpu/polaris10_k_smc.bin"); } else if ((adev->pdev->device == 0x67df) && diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index dc54e9efd910..1c49b8266d69 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -50,7 +50,8 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); p->uf_entry.priority = 0; p->uf_entry.tv.bo = &bo->tbo; - p->uf_entry.tv.shared = true; + /* One for TTM and one for the CS job */ + p->uf_entry.tv.num_shared = 2; p->uf_entry.user_pages = NULL; drm_gem_object_put_unlocked(gobj); @@ -124,14 +125,14 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs goto free_chunk; } + mutex_lock(&p->ctx->lock); + /* skip guilty context job */ if (atomic_read(&p->ctx->guilty) == 1) { ret = -ECANCELED; goto free_chunk; } - mutex_lock(&p->ctx->lock); - /* get chunks */ chunk_array_user = u64_to_user_ptr(cs->in.chunks); if (copy_from_user(chunk_array, chunk_array_user, @@ -598,6 +599,10 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, return r; } + /* One for TTM and one for the CS job */ + amdgpu_bo_list_for_each_entry(e, p->bo_list) + e->tv.num_shared = 2; + amdgpu_bo_list_get_list(p->bo_list, &p->validated); if (p->bo_list->first_userptr != p->bo_list->num_entries) p->mn = amdgpu_mn_get(p->adev, AMDGPU_MN_TYPE_GFX); @@ -717,8 +722,14 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, gws = p->bo_list->gws_obj; oa = p->bo_list->oa_obj; - amdgpu_bo_list_for_each_entry(e, p->bo_list) - e->bo_va = amdgpu_vm_bo_find(vm, ttm_to_amdgpu_bo(e->tv.bo)); + amdgpu_bo_list_for_each_entry(e, p->bo_list) { + struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo); + + /* Make sure we use the exclusive slot for shared BOs */ + if (bo->prime_shared_count) + e->tv.num_shared = 0; + e->bo_va = amdgpu_vm_bo_find(vm, bo); + } if (gds) { p->job->gds_base = amdgpu_bo_gpu_offset(gds) >> PAGE_SHIFT; @@ -955,10 +966,6 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p) if (r) return r; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); - if (r) - return r; - p->job->vm_pd_addr = amdgpu_gmc_pd_addr(vm->root.base.bo); if (amdgpu_vm_debug) { @@ -1421,6 +1428,9 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data, if (IS_ERR(fence)) return PTR_ERR(fence); + if (!fence) + fence = dma_fence_get_stub(); + switch (info->in.what) { case AMDGPU_FENCE_TO_HANDLE_GET_SYNCOBJ: r = drm_syncobj_create(&syncobj, 0, fence); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c index 5b550706ee76..7e22be7ca68a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c @@ -74,7 +74,7 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, INIT_LIST_HEAD(&list); INIT_LIST_HEAD(&csa_tv.head); csa_tv.bo = &bo->tbo; - csa_tv.shared = true; + csa_tv.num_shared = 1; list_add(&csa_tv.head, &list); amdgpu_vm_get_pd_bo(vm, &list, &pd); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index 95f4c4139fc6..d85184b5b35c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -248,7 +248,7 @@ static int amdgpu_ctx_alloc(struct amdgpu_device *adev, return -ENOMEM; mutex_lock(&mgr->lock); - r = idr_alloc(&mgr->ctx_handles, ctx, 1, 0, GFP_KERNEL); + r = idr_alloc(&mgr->ctx_handles, ctx, 1, AMDGPU_VM_MAX_NUM_CTX, GFP_KERNEL); if (r < 0) { mutex_unlock(&mgr->lock); kfree(ctx); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index c75badfa5c4c..7ff3a28fc903 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -515,7 +515,6 @@ void amdgpu_device_pci_config_reset(struct amdgpu_device *adev) */ static int amdgpu_device_doorbell_init(struct amdgpu_device *adev) { - amdgpu_asic_init_doorbell_index(adev); /* No doorbell on SI hardware generation */ if (adev->asic_type < CHIP_BONAIRE) { @@ -529,6 +528,8 @@ static int amdgpu_device_doorbell_init(struct amdgpu_device *adev) if (pci_resource_flags(adev->pdev, 2) & IORESOURCE_UNSET) return -EINVAL; + amdgpu_asic_init_doorbell_index(adev); + /* doorbell bar mapping */ adev->doorbell.base = pci_resource_start(adev->pdev, 2); adev->doorbell.size = pci_resource_len(adev->pdev, 2); @@ -1700,8 +1701,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) amdgpu_xgmi_add_device(adev); amdgpu_amdkfd_device_init(adev); - if (amdgpu_sriov_vf(adev)) + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_init_data_exchange(adev); amdgpu_virt_release_full_gpu(adev, true); + } return 0; } @@ -1864,6 +1867,9 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev) { int i, r; + if (adev->gmc.xgmi.num_physical_nodes > 1) + amdgpu_xgmi_remove_device(adev); + amdgpu_amdkfd_device_fini(adev); amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE); @@ -2353,6 +2359,19 @@ bool amdgpu_device_has_dc_support(struct amdgpu_device *adev) return amdgpu_device_asic_has_dc_support(adev->asic_type); } + +static void amdgpu_device_xgmi_reset_func(struct work_struct *__work) +{ + struct amdgpu_device *adev = + container_of(__work, struct amdgpu_device, xgmi_reset_work); + + adev->asic_reset_res = amdgpu_asic_reset(adev); + if (adev->asic_reset_res) + DRM_WARN("ASIC reset failed with err r, %d for drm dev, %s", + adev->asic_reset_res, adev->ddev->unique); +} + + /** * amdgpu_device_init - initialize the driver * @@ -2451,6 +2470,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work, amdgpu_device_delay_enable_gfx_off); + INIT_WORK(&adev->xgmi_reset_work, amdgpu_device_xgmi_reset_func); + adev->gfx.gfx_off_req_count = 1; adev->pm.ac_power = power_supply_is_system_supplied() > 0 ? true : false; @@ -2613,9 +2634,6 @@ fence_driver_init: goto failed; } - if (amdgpu_sriov_vf(adev)) - amdgpu_virt_init_data_exchange(adev); - amdgpu_fbdev_init(adev); r = amdgpu_pm_sysfs_init(adev); @@ -2779,7 +2797,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon) struct drm_framebuffer *fb = crtc->primary->fb; struct amdgpu_bo *robj; - if (amdgpu_crtc->cursor_bo) { + if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) { struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo); r = amdgpu_bo_reserve(aobj, true); if (r == 0) { @@ -2887,7 +2905,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon) list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - if (amdgpu_crtc->cursor_bo) { + if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) { struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo); r = amdgpu_bo_reserve(aobj, true); if (r == 0) { @@ -3207,6 +3225,7 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, r = amdgpu_ib_ring_tests(adev); error: + amdgpu_virt_init_data_exchange(adev); amdgpu_virt_release_full_gpu(adev, true); if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) { atomic_inc(&adev->vram_lost_counter); @@ -3239,6 +3258,8 @@ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev) if (amdgpu_gpu_recovery == -1) { switch (adev->asic_type) { + case CHIP_BONAIRE: + case CHIP_HAWAII: case CHIP_TOPAZ: case CHIP_TONGA: case CHIP_FIJI: @@ -3328,10 +3349,31 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, */ if (need_full_reset) { list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { - r = amdgpu_asic_reset(tmp_adev); - if (r) - DRM_WARN("ASIC reset failed with err r, %d for drm dev, %s", + /* For XGMI run all resets in parallel to speed up the process */ + if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) { + if (!queue_work(system_highpri_wq, &tmp_adev->xgmi_reset_work)) + r = -EALREADY; + } else + r = amdgpu_asic_reset(tmp_adev); + + if (r) { + DRM_ERROR("ASIC reset failed with err r, %d for drm dev, %s", r, tmp_adev->ddev->unique); + break; + } + } + + /* For XGMI wait for all PSP resets to complete before proceed */ + if (!r) { + list_for_each_entry(tmp_adev, device_list_handle, + gmc.xgmi.head) { + if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) { + flush_work(&tmp_adev->xgmi_reset_work); + r = tmp_adev->asic_reset_res; + if (r) + break; + } + } } } @@ -3434,14 +3476,16 @@ static void amdgpu_device_lock_adev(struct amdgpu_device *adev) mutex_lock(&adev->lock_reset); atomic_inc(&adev->gpu_reset_counter); adev->in_gpu_reset = 1; - /* Block kfd */ - amdgpu_amdkfd_pre_reset(adev); + /* Block kfd: SRIOV would do it separately */ + if (!amdgpu_sriov_vf(adev)) + amdgpu_amdkfd_pre_reset(adev); } static void amdgpu_device_unlock_adev(struct amdgpu_device *adev) { - /*unlock kfd */ - amdgpu_amdkfd_post_reset(adev); + /*unlock kfd: SRIOV would do it separately */ + if (!amdgpu_sriov_vf(adev)) + amdgpu_amdkfd_post_reset(adev); amdgpu_vf_error_trans_all(adev); adev->in_gpu_reset = 0; mutex_unlock(&adev->lock_reset); @@ -3518,8 +3562,6 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ if (tmp_adev == adev) continue; - dev_info(tmp_adev->dev, "GPU reset begin for drm dev %s!\n", adev->ddev->unique); - amdgpu_device_lock_adev(tmp_adev); r = amdgpu_device_pre_asic_reset(tmp_adev, NULL, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 15ce7e681d67..b083b219b1a9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -188,10 +188,12 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc, goto cleanup; } - r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev)); - if (unlikely(r != 0)) { - DRM_ERROR("failed to pin new abo buffer before flip\n"); - goto unreserve; + if (!adev->enable_virtual_display) { + r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev)); + if (unlikely(r != 0)) { + DRM_ERROR("failed to pin new abo buffer before flip\n"); + goto unreserve; + } } r = amdgpu_ttm_alloc_gart(&new_abo->tbo); @@ -211,7 +213,8 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc, amdgpu_bo_get_tiling_flags(new_abo, &tiling_flags); amdgpu_bo_unreserve(new_abo); - work->base = amdgpu_bo_gpu_offset(new_abo); + if (!adev->enable_virtual_display) + work->base = amdgpu_bo_gpu_offset(new_abo); work->target_vblank = target - (uint32_t)drm_crtc_vblank_count(crtc) + amdgpu_get_vblank_counter_kms(dev, work->crtc_id); @@ -242,9 +245,10 @@ pflip_cleanup: goto cleanup; } unpin: - if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) { - DRM_ERROR("failed to unpin new abo in error path\n"); - } + if (!adev->enable_virtual_display) + if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) + DRM_ERROR("failed to unpin new abo in error path\n"); + unreserve: amdgpu_bo_unreserve(new_abo); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 90f474f98b6e..c806f984bcc5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -865,6 +865,7 @@ static const struct pci_device_id pciidlist[] = { /* VEGAM */ {0x1002, 0x694C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGAM}, {0x1002, 0x694E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGAM}, + {0x1002, 0x694F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGAM}, /* Vega 10 */ {0x1002, 0x6860, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, {0x1002, 0x6861, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, @@ -873,7 +874,13 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x6864, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, {0x1002, 0x6867, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, {0x1002, 0x6868, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x6869, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x686a, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x686b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, {0x1002, 0x686c, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x686d, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x686e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, + {0x1002, 0x686f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, {0x1002, 0x687f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10}, /* Vega 12 */ {0x1002, 0x69A0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA12}, @@ -886,6 +893,7 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x66A1, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, {0x1002, 0x66A2, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, {0x1002, 0x66A3, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, + {0x1002, 0x66A4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, {0x1002, 0x66A7, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, {0x1002, 0x66AF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA20}, /* Raven */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 7b3d1ebda9df..f4f00217546e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -169,7 +169,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj, INIT_LIST_HEAD(&duplicates); tv.bo = &bo->tbo; - tv.shared = true; + tv.num_shared = 1; list_add(&tv.head, &list); amdgpu_vm_get_pd_bo(vm, &list, &vm_pd); @@ -604,7 +604,10 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, return -ENOENT; abo = gem_to_amdgpu_bo(gobj); tv.bo = &abo->tbo; - tv.shared = !!(abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID); + if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) + tv.num_shared = 1; + else + tv.num_shared = 0; list_add(&tv.head, &list); } else { gobj = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h index d63daba9b17c..f1ddfc50bcc7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.h @@ -54,6 +54,8 @@ void *amdgpu_gem_prime_vmap(struct drm_gem_object *obj); void amdgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr); int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma); +extern const struct dma_buf_ops amdgpu_dmabuf_ops; + /* * GEM objects. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h index 8c57924c075f..81e6070d255b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h @@ -99,6 +99,7 @@ struct amdgpu_xgmi { unsigned num_physical_nodes; /* gpu list in the same hive */ struct list_head head; + bool supported; }; struct amdgpu_gmc { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index 9ce8c93ec19b..f877bb78d10a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h @@ -51,14 +51,12 @@ struct amdgpu_ih_ring { struct amdgpu_ih_funcs { /* ring read/write ptr handling, called from interrupt context */ u32 (*get_wptr)(struct amdgpu_device *adev); - bool (*prescreen_iv)(struct amdgpu_device *adev); void (*decode_iv)(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry); void (*set_rptr)(struct amdgpu_device *adev); }; #define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev)) -#define amdgpu_ih_prescreen_iv(adev) (adev)->irq.ih_funcs->prescreen_iv((adev)) #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv)) #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index 6b6524f04ce0..b7968f426862 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -145,13 +145,6 @@ static void amdgpu_irq_callback(struct amdgpu_device *adev, u32 ring_index = ih->rptr >> 2; struct amdgpu_iv_entry entry; - /* Prescreening of high-frequency interrupts */ - if (!amdgpu_ih_prescreen_iv(adev)) - return; - - /* Before dispatching irq to IP blocks, send it to amdkfd */ - amdgpu_amdkfd_interrupt(adev, (const void *) &ih->ring[ring_index]); - entry.iv_entry = (const uint32_t *)&ih->ring[ring_index]; amdgpu_ih_decode_iv(adev, &entry); @@ -371,39 +364,38 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, unsigned client_id = entry->client_id; unsigned src_id = entry->src_id; struct amdgpu_irq_src *src; + bool handled = false; int r; trace_amdgpu_iv(entry); if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) { DRM_DEBUG("Invalid client_id in IV: %d\n", client_id); - return; - } - if (src_id >= AMDGPU_MAX_IRQ_SRC_ID) { + } else if (src_id >= AMDGPU_MAX_IRQ_SRC_ID) { DRM_DEBUG("Invalid src_id in IV: %d\n", src_id); - return; - } - if (adev->irq.virq[src_id]) { + } else if (adev->irq.virq[src_id]) { generic_handle_irq(irq_find_mapping(adev->irq.domain, src_id)); - } else { - if (!adev->irq.client[client_id].sources) { - DRM_DEBUG("Unregistered interrupt client_id: %d src_id: %d\n", - client_id, src_id); - return; - } - src = adev->irq.client[client_id].sources[src_id]; - if (!src) { - DRM_DEBUG("Unhandled interrupt src_id: %d\n", src_id); - return; - } + } else if (!adev->irq.client[client_id].sources) { + DRM_DEBUG("Unregistered interrupt client_id: %d src_id: %d\n", + client_id, src_id); + } else if ((src = adev->irq.client[client_id].sources[src_id])) { r = src->funcs->process(adev, src, entry); - if (r) + if (r < 0) DRM_ERROR("error processing interrupt (%d)\n", r); + else if (r) + handled = true; + + } else { + DRM_DEBUG("Unhandled interrupt src_id: %d\n", src_id); } + + /* Send it to amdkfd as well if it isn't already handled */ + if (!handled) + amdgpu_amdkfd_interrupt(adev, entry->iv_entry); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index e0af44fd6a0c..0a17fb1af204 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -32,6 +32,9 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job) { struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched); struct amdgpu_job *job = to_amdgpu_job(s_job); + struct amdgpu_task_info ti; + + memset(&ti, 0, sizeof(struct amdgpu_task_info)); if (amdgpu_ring_soft_recovery(ring, job->vmid, s_job->s_fence->parent)) { DRM_ERROR("ring %s timeout, but soft recovered\n", @@ -39,9 +42,12 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job) return; } + amdgpu_vm_get_task_info(ring->adev, job->pasid, &ti); DRM_ERROR("ring %s timeout, signaled seq=%u, emitted seq=%u\n", job->base.sched->name, atomic_read(&ring->fence_drv.last_seq), ring->fence_drv.sync_seq); + DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n", + ti.process_name, ti.tgid, ti.task_name, ti.pid); if (amdgpu_device_should_recover_gpu(ring->adev)) amdgpu_device_gpu_recover(ring->adev, job); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index e55508b39496..3e6823fdd939 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -238,44 +238,40 @@ static void amdgpu_mn_invalidate_node(struct amdgpu_mn_node *node, * amdgpu_mn_invalidate_range_start_gfx - callback to notify about mm change * * @mn: our notifier - * @mm: the mm this callback is about - * @start: start of updated range - * @end: end of updated range + * @range: mmu notifier context * * Block for operations on BOs to finish and mark pages as accessed and * potentially dirty. */ static int amdgpu_mn_invalidate_range_start_gfx(struct mmu_notifier *mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end, - bool blockable) + const struct mmu_notifier_range *range) { struct amdgpu_mn *amn = container_of(mn, struct amdgpu_mn, mn); struct interval_tree_node *it; + unsigned long end; /* notification is exclusive, but interval is inclusive */ - end -= 1; + end = range->end - 1; /* TODO we should be able to split locking for interval tree and * amdgpu_mn_invalidate_node */ - if (amdgpu_mn_read_lock(amn, blockable)) + if (amdgpu_mn_read_lock(amn, range->blockable)) return -EAGAIN; - it = interval_tree_iter_first(&amn->objects, start, end); + it = interval_tree_iter_first(&amn->objects, range->start, end); while (it) { struct amdgpu_mn_node *node; - if (!blockable) { + if (!range->blockable) { amdgpu_mn_read_unlock(amn); return -EAGAIN; } node = container_of(it, struct amdgpu_mn_node, it); - it = interval_tree_iter_next(it, start, end); + it = interval_tree_iter_next(it, range->start, end); - amdgpu_mn_invalidate_node(node, start, end); + amdgpu_mn_invalidate_node(node, range->start, end); } return 0; @@ -294,39 +290,38 @@ static int amdgpu_mn_invalidate_range_start_gfx(struct mmu_notifier *mn, * are restorted in amdgpu_mn_invalidate_range_end_hsa. */ static int amdgpu_mn_invalidate_range_start_hsa(struct mmu_notifier *mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end, - bool blockable) + const struct mmu_notifier_range *range) { struct amdgpu_mn *amn = container_of(mn, struct amdgpu_mn, mn); struct interval_tree_node *it; + unsigned long end; /* notification is exclusive, but interval is inclusive */ - end -= 1; + end = range->end - 1; - if (amdgpu_mn_read_lock(amn, blockable)) + if (amdgpu_mn_read_lock(amn, range->blockable)) return -EAGAIN; - it = interval_tree_iter_first(&amn->objects, start, end); + it = interval_tree_iter_first(&amn->objects, range->start, end); while (it) { struct amdgpu_mn_node *node; struct amdgpu_bo *bo; - if (!blockable) { + if (!range->blockable) { amdgpu_mn_read_unlock(amn); return -EAGAIN; } node = container_of(it, struct amdgpu_mn_node, it); - it = interval_tree_iter_next(it, start, end); + it = interval_tree_iter_next(it, range->start, end); list_for_each_entry(bo, &node->bos, mn_list) { struct kgd_mem *mem = bo->kfd_bo; if (amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, - start, end)) - amdgpu_amdkfd_evict_userptr(mem, mm); + range->start, + end)) + amdgpu_amdkfd_evict_userptr(mem, range->mm); } } @@ -344,9 +339,7 @@ static int amdgpu_mn_invalidate_range_start_hsa(struct mmu_notifier *mn, * Release the lock again to allow new command submissions. */ static void amdgpu_mn_invalidate_range_end(struct mmu_notifier *mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end) + const struct mmu_notifier_range *range) { struct amdgpu_mn *amn = container_of(mn, struct amdgpu_mn, mn); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index cf768acb51dc..728e15e5d68a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -81,7 +81,7 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo) amdgpu_bo_subtract_pin_size(bo); if (bo->kfd_bo) - amdgpu_amdkfd_unreserve_system_memory_limit(bo); + amdgpu_amdkfd_unreserve_memory_limit(bo); amdgpu_bo_kunmap(bo); @@ -608,53 +608,6 @@ int amdgpu_bo_create(struct amdgpu_device *adev, } /** - * amdgpu_bo_backup_to_shadow - Backs up an &amdgpu_bo buffer object - * @adev: amdgpu device object - * @ring: amdgpu_ring for the engine handling the buffer operations - * @bo: &amdgpu_bo buffer to be backed up - * @resv: reservation object with embedded fence - * @fence: dma_fence associated with the operation - * @direct: whether to submit the job directly - * - * Copies an &amdgpu_bo buffer object to its shadow object. - * Not used for now. - * - * Returns: - * 0 for success or a negative error code on failure. - */ -int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct amdgpu_bo *bo, - struct reservation_object *resv, - struct dma_fence **fence, - bool direct) - -{ - struct amdgpu_bo *shadow = bo->shadow; - uint64_t bo_addr, shadow_addr; - int r; - - if (!shadow) - return -EINVAL; - - bo_addr = amdgpu_bo_gpu_offset(bo); - shadow_addr = amdgpu_bo_gpu_offset(bo->shadow); - - r = reservation_object_reserve_shared(bo->tbo.resv, 1); - if (r) - goto err; - - r = amdgpu_copy_buffer(ring, bo_addr, shadow_addr, - amdgpu_bo_size(bo), resv, fence, - direct, false); - if (!r) - amdgpu_bo_fence(bo, *fence, true); - -err: - return r; -} - -/** * amdgpu_bo_validate - validate an &amdgpu_bo buffer object * @bo: pointer to the buffer object * @@ -959,7 +912,7 @@ int amdgpu_bo_unpin(struct amdgpu_bo *bo) struct ttm_operation_ctx ctx = { false, false }; int r, i; - if (!bo->pin_count) { + if (WARN_ON_ONCE(!bo->pin_count)) { dev_warn(adev->dev, "%p unpin not necessary\n", bo); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 7d3312d0da11..9291c2f837e9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -267,11 +267,6 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo); void amdgpu_bo_fence(struct amdgpu_bo *bo, struct dma_fence *fence, bool shared); u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo); -int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct amdgpu_bo *bo, - struct reservation_object *resv, - struct dma_fence **fence, bool direct); int amdgpu_bo_validate(struct amdgpu_bo *bo); int amdgpu_bo_restore_shadow(struct amdgpu_bo *shadow, struct dma_fence **fence); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 1f61ed95727c..6896dec97fc7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -2008,6 +2008,7 @@ void amdgpu_pm_print_power_states(struct amdgpu_device *adev) int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) { + struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; int ret; if (adev->pm.sysfs_initialized) @@ -2091,12 +2092,14 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) "pp_power_profile_mode\n"); return ret; } - ret = device_create_file(adev->dev, - &dev_attr_pp_od_clk_voltage); - if (ret) { - DRM_ERROR("failed to create device file " - "pp_od_clk_voltage\n"); - return ret; + if (hwmgr->od_enabled) { + ret = device_create_file(adev->dev, + &dev_attr_pp_od_clk_voltage); + if (ret) { + DRM_ERROR("failed to create device file " + "pp_od_clk_voltage\n"); + return ret; + } } ret = device_create_file(adev->dev, &dev_attr_gpu_busy_percent); @@ -2118,6 +2121,8 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) { + struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; + if (adev->pm.dpm_enabled == 0) return; @@ -2138,8 +2143,9 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) device_remove_file(adev->dev, &dev_attr_pp_mclk_od); device_remove_file(adev->dev, &dev_attr_pp_power_profile_mode); - device_remove_file(adev->dev, - &dev_attr_pp_od_clk_voltage); + if (hwmgr->od_enabled) + device_remove_file(adev->dev, + &dev_attr_pp_od_clk_voltage); device_remove_file(adev->dev, &dev_attr_gpu_busy_percent); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c index 3e44d889f7af..71913a18d142 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c @@ -39,8 +39,6 @@ #include <drm/amdgpu_drm.h> #include <linux/dma-buf.h> -static const struct dma_buf_ops amdgpu_dmabuf_ops; - /** * amdgpu_gem_prime_get_sg_table - &drm_driver.gem_prime_get_sg_table * implementation @@ -332,7 +330,7 @@ static int amdgpu_gem_begin_cpu_access(struct dma_buf *dma_buf, return ret; } -static const struct dma_buf_ops amdgpu_dmabuf_ops = { +const struct dma_buf_ops amdgpu_dmabuf_ops = { .attach = amdgpu_gem_map_attach, .detach = amdgpu_gem_map_detach, .map_dma_buf = drm_gem_map_dma_buf, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index e05dc66b1090..8fab0d637ee5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -155,10 +155,14 @@ psp_cmd_submit_buf(struct psp_context *psp, return ret; } -static void psp_prep_tmr_cmd_buf(struct psp_gfx_cmd_resp *cmd, +static void psp_prep_tmr_cmd_buf(struct psp_context *psp, + struct psp_gfx_cmd_resp *cmd, uint64_t tmr_mc, uint32_t size) { - cmd->cmd_id = GFX_CMD_ID_SETUP_TMR; + if (psp_support_vmr_ring(psp)) + cmd->cmd_id = GFX_CMD_ID_SETUP_VMR; + else + cmd->cmd_id = GFX_CMD_ID_SETUP_TMR; cmd->cmd.cmd_setup_tmr.buf_phy_addr_lo = lower_32_bits(tmr_mc); cmd->cmd.cmd_setup_tmr.buf_phy_addr_hi = upper_32_bits(tmr_mc); cmd->cmd.cmd_setup_tmr.buf_size = size; @@ -192,7 +196,7 @@ static int psp_tmr_load(struct psp_context *psp) if (!cmd) return -ENOMEM; - psp_prep_tmr_cmd_buf(cmd, psp->tmr_mc_addr, PSP_TMR_SIZE); + psp_prep_tmr_cmd_buf(psp, cmd, psp->tmr_mc_addr, PSP_TMR_SIZE); DRM_INFO("reserve 0x%x from 0x%llx for PSP TMR SIZE\n", PSP_TMR_SIZE, psp->tmr_mc_addr); @@ -536,8 +540,10 @@ static int psp_load_fw(struct amdgpu_device *adev) int ret; struct psp_context *psp = &adev->psp; - if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset != 0) + if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset) { + psp_ring_destroy(psp, PSP_RING_TYPE__KM); goto skip_memalloc; + } psp->cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL); if (!psp->cmd) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index 9ec5d1a666a6..3ee573b4016e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -83,12 +83,13 @@ struct psp_funcs enum AMDGPU_UCODE_ID ucode_type); bool (*smu_reload_quirk)(struct psp_context *psp); int (*mode1_reset)(struct psp_context *psp); - uint64_t (*xgmi_get_node_id)(struct psp_context *psp); - uint64_t (*xgmi_get_hive_id)(struct psp_context *psp); + int (*xgmi_get_node_id)(struct psp_context *psp, uint64_t *node_id); + int (*xgmi_get_hive_id)(struct psp_context *psp, uint64_t *hive_id); int (*xgmi_get_topology_info)(struct psp_context *psp, int number_devices, struct psp_xgmi_topology_info *topology); int (*xgmi_set_topology_info)(struct psp_context *psp, int number_devices, struct psp_xgmi_topology_info *topology); + bool (*support_vmr_ring)(struct psp_context *psp); }; struct psp_xgmi_context { @@ -192,12 +193,14 @@ struct psp_xgmi_topology_info { ((psp)->funcs->bootloader_load_sos ? (psp)->funcs->bootloader_load_sos((psp)) : 0) #define psp_smu_reload_quirk(psp) \ ((psp)->funcs->smu_reload_quirk ? (psp)->funcs->smu_reload_quirk((psp)) : false) +#define psp_support_vmr_ring(psp) \ + ((psp)->funcs->support_vmr_ring ? (psp)->funcs->support_vmr_ring((psp)) : false) #define psp_mode1_reset(psp) \ ((psp)->funcs->mode1_reset ? (psp)->funcs->mode1_reset((psp)) : false) -#define psp_xgmi_get_node_id(psp) \ - ((psp)->funcs->xgmi_get_node_id ? (psp)->funcs->xgmi_get_node_id((psp)) : 0) -#define psp_xgmi_get_hive_id(psp) \ - ((psp)->funcs->xgmi_get_hive_id ? (psp)->funcs->xgmi_get_hive_id((psp)) : 0) +#define psp_xgmi_get_node_id(psp, node_id) \ + ((psp)->funcs->xgmi_get_node_id ? (psp)->funcs->xgmi_get_node_id((psp), (node_id)) : -EINVAL) +#define psp_xgmi_get_hive_id(psp, hive_id) \ + ((psp)->funcs->xgmi_get_hive_id ? (psp)->funcs->xgmi_get_hive_id((psp), (hive_id)) : -EINVAL) #define psp_xgmi_get_topology_info(psp, num_device, topology) \ ((psp)->funcs->xgmi_get_topology_info ? \ (psp)->funcs->xgmi_get_topology_info((psp), (num_device), (topology)) : -EINVAL) @@ -217,7 +220,6 @@ extern const struct amdgpu_ip_block_version psp_v10_0_ip_block; int psp_gpu_reset(struct amdgpu_device *adev); int psp_xgmi_invoke(struct psp_context *psp, uint32_t ta_cmd_id); - extern const struct amdgpu_ip_block_version psp_v11_0_ip_block; #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 5b75bdc8dc28..335a0edf114b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -397,7 +397,7 @@ bool amdgpu_ring_soft_recovery(struct amdgpu_ring *ring, unsigned int vmid, { ktime_t deadline = ktime_add_us(ktime_get(), 10000); - if (!ring->funcs->soft_recovery) + if (!ring->funcs->soft_recovery || !fence) return false; atomic_inc(&ring->adev->gpu_reset_counter); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index 0beb01fef83f..d87e828a084b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -29,7 +29,7 @@ #include <drm/drm_print.h> /* max number of rings */ -#define AMDGPU_MAX_RINGS 21 +#define AMDGPU_MAX_RINGS 23 #define AMDGPU_MAX_GFX_RINGS 1 #define AMDGPU_MAX_COMPUTE_RINGS 8 #define AMDGPU_MAX_VCE_RINGS 3 diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index e2e42e3fbcf3..ecf6f96df2ad 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -262,7 +262,7 @@ static int amdgpu_vcn_pause_dpg_mode(struct amdgpu_device *adev, ring = &adev->vcn.ring_dec; WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, - RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2)); + RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2) & 0x7FFFFFFF); SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); @@ -322,7 +322,7 @@ static int amdgpu_vcn_pause_dpg_mode(struct amdgpu_device *adev, ring = &adev->vcn.ring_dec; WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, - RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2)); + RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2) & 0x7FFFFFFF); SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); @@ -396,16 +396,26 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { struct dpg_pause_state new_state; + unsigned int fences = 0; + unsigned int i; - if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) + for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + fences += amdgpu_fence_count_emitted(&adev->vcn.ring_enc[i]); + } + if (fences) new_state.fw_based = VCN_DPG_STATE__PAUSE; else - new_state.fw_based = adev->vcn.pause_state.fw_based; + new_state.fw_based = VCN_DPG_STATE__UNPAUSE; - if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) + if (amdgpu_fence_count_emitted(&adev->vcn.ring_jpeg)) new_state.jpeg = VCN_DPG_STATE__PAUSE; else - new_state.jpeg = adev->vcn.pause_state.jpeg; + new_state.jpeg = VCN_DPG_STATE__UNPAUSE; + + if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) + new_state.fw_based = VCN_DPG_STATE__PAUSE; + else if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) + new_state.jpeg = VCN_DPG_STATE__PAUSE; amdgpu_vcn_pause_dpg_mode(adev, &new_state); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 58a2363040dd..d2ea5ce2cefb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -181,7 +181,7 @@ static unsigned amdgpu_vm_num_entries(struct amdgpu_device *adev, if (level == adev->vm_manager.root_level) /* For the root directory */ - return round_up(adev->vm_manager.max_pfn, 1 << shift) >> shift; + return round_up(adev->vm_manager.max_pfn, 1ULL << shift) >> shift; else if (level != AMDGPU_VM_PTB) /* Everything in between */ return 512; @@ -617,7 +617,8 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, { entry->priority = 0; entry->tv.bo = &vm->root.base.bo->tbo; - entry->tv.shared = true; + /* One for the VM updates, one for TTM and one for the CS job */ + entry->tv.num_shared = 3; entry->user_pages = NULL; list_add(&entry->tv.head, validated); } @@ -773,10 +774,6 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, ring = container_of(vm->entity.rq->sched, struct amdgpu_ring, sched); - r = reservation_object_reserve_shared(bo->tbo.resv, 1); - if (r) - return r; - r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (r) goto error; @@ -850,9 +847,6 @@ static void amdgpu_vm_bo_param(struct amdgpu_device *adev, struct amdgpu_vm *vm, bp->size = amdgpu_vm_bo_size(adev, level); bp->byte_align = AMDGPU_GPU_PAGE_SIZE; bp->domain = AMDGPU_GEM_DOMAIN_VRAM; - if (bp->size <= PAGE_SIZE && adev->asic_type >= CHIP_VEGA10 && - adev->flags & AMD_IS_APU) - bp->domain |= AMDGPU_GEM_DOMAIN_GTT; bp->domain = amdgpu_bo_get_preferred_pin_domain(adev, bp->domain); bp->flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | AMDGPU_GEM_CREATE_CPU_GTT_USWC; @@ -1656,9 +1650,11 @@ static int amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params, if (!amdgpu_vm_pt_descendant(adev, &cursor)) return -ENOENT; continue; - } else if (frag >= parent_shift) { + } else if (frag >= parent_shift && + cursor.level - 1 != adev->vm_manager.root_level) { /* If the fragment size is even larger than the parent - * shift we should go up one level and check it again. + * shift we should go up one level and check it again + * unless one level up is the root level. */ if (!amdgpu_vm_pt_ancestor(&cursor)) return -ENOENT; @@ -1666,10 +1662,10 @@ static int amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params, } /* Looks good so far, calculate parameters for the update */ - incr = AMDGPU_GPU_PAGE_SIZE << shift; + incr = (uint64_t)AMDGPU_GPU_PAGE_SIZE << shift; mask = amdgpu_vm_entries_mask(adev, cursor.level); pe_start = ((cursor.pfn >> shift) & mask) * 8; - entry_end = (mask + 1) << shift; + entry_end = (uint64_t)(mask + 1) << shift; entry_end += cursor.pfn & ~(entry_end - 1); entry_end = min(entry_end, end); @@ -1682,7 +1678,7 @@ static int amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params, flags | AMDGPU_PTE_FRAG(frag)); pe_start += nptes * 8; - dst += nptes * AMDGPU_GPU_PAGE_SIZE << shift; + dst += (uint64_t)nptes * AMDGPU_GPU_PAGE_SIZE << shift; frag_start = upd_end; if (frag_start >= frag_end) { @@ -1842,10 +1838,6 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, if (r) goto error_free; - r = reservation_object_reserve_shared(vm->root.base.bo->tbo.resv, 1); - if (r) - goto error_free; - r = amdgpu_vm_update_ptes(¶ms, start, last + 1, addr, flags); if (r) goto error_free; @@ -3026,6 +3018,10 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, if (r) goto error_free_root; + r = reservation_object_reserve_shared(root->tbo.resv, 1); + if (r) + goto error_unreserve; + r = amdgpu_vm_clear_bo(adev, vm, root, adev->vm_manager.root_level, vm->pte_support_ats); @@ -3055,7 +3051,6 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, } INIT_KFIFO(vm->faults); - vm->fault_credit = 16; return 0; @@ -3268,42 +3263,6 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) } /** - * amdgpu_vm_pasid_fault_credit - Check fault credit for given PASID - * - * @adev: amdgpu_device pointer - * @pasid: PASID do identify the VM - * - * This function is expected to be called in interrupt context. - * - * Returns: - * True if there was fault credit, false otherwise - */ -bool amdgpu_vm_pasid_fault_credit(struct amdgpu_device *adev, - unsigned int pasid) -{ - struct amdgpu_vm *vm; - - spin_lock(&adev->vm_manager.pasid_lock); - vm = idr_find(&adev->vm_manager.pasid_idr, pasid); - if (!vm) { - /* VM not found, can't track fault credit */ - spin_unlock(&adev->vm_manager.pasid_lock); - return true; - } - - /* No lock needed. only accessed by IRQ handler */ - if (!vm->fault_credit) { - /* Too many faults in this VM */ - spin_unlock(&adev->vm_manager.pasid_lock); - return false; - } - - vm->fault_credit--; - spin_unlock(&adev->vm_manager.pasid_lock); - return true; -} - -/** * amdgpu_vm_manager_init - init the VM manager * * @adev: amdgpu_device pointer diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index 2a8898d19c8b..e8dcfd59fc93 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -229,9 +229,6 @@ struct amdgpu_vm { /* Up to 128 pending retry page faults */ DECLARE_KFIFO(faults, u64, 128); - /* Limit non-retry fault storms */ - unsigned int fault_credit; - /* Points to the KFD process VM info */ struct amdkfd_process_info *process_info; @@ -299,8 +296,6 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, unsigned int pasid); void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm); void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm); -bool amdgpu_vm_pasid_fault_credit(struct amdgpu_device *adev, - unsigned int pasid); void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct list_head *validated, struct amdgpu_bo_list_entry *entry); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index fb37e69f1bba..8a8bc60cb6b4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -78,7 +78,7 @@ int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_dev adev->gmc.xgmi.node_id, adev->gmc.xgmi.hive_id, ret); else - dev_info(adev->dev, "XGMI: Add node %d to hive 0x%llx.\n", + dev_info(adev->dev, "XGMI: Set topology for node %d, hive 0x%llx.\n", adev->gmc.xgmi.physical_node_id, adev->gmc.xgmi.hive_id); @@ -94,11 +94,22 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev) int count = 0, ret = -EINVAL; - if ((adev->asic_type < CHIP_VEGA20) || - (adev->flags & AMD_IS_APU) ) + if (!adev->gmc.xgmi.supported) return 0; - adev->gmc.xgmi.node_id = psp_xgmi_get_node_id(&adev->psp); - adev->gmc.xgmi.hive_id = psp_xgmi_get_hive_id(&adev->psp); + + ret = psp_xgmi_get_node_id(&adev->psp, &adev->gmc.xgmi.node_id); + if (ret) { + dev_err(adev->dev, + "XGMI: Failed to get node id\n"); + return ret; + } + + ret = psp_xgmi_get_hive_id(&adev->psp, &adev->gmc.xgmi.hive_id); + if (ret) { + dev_err(adev->dev, + "XGMI: Failed to get hive id\n"); + return ret; + } mutex_lock(&xgmi_mutex); hive = amdgpu_get_xgmi_hive(adev); @@ -135,3 +146,23 @@ exit: mutex_unlock(&xgmi_mutex); return ret; } + +void amdgpu_xgmi_remove_device(struct amdgpu_device *adev) +{ + struct amdgpu_hive_info *hive; + + if (!adev->gmc.xgmi.supported) + return; + + mutex_lock(&xgmi_mutex); + + hive = amdgpu_get_xgmi_hive(adev); + if (!hive) + goto exit; + + if (!(hive->number_devices--)) + mutex_destroy(&hive->hive_lock); + +exit: + mutex_unlock(&xgmi_mutex); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h index 6335bfdcc51d..6151eb9c8ad3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h @@ -35,5 +35,6 @@ struct amdgpu_hive_info { struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev); int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_device *adev); int amdgpu_xgmi_add_device(struct amdgpu_device *adev); +void amdgpu_xgmi_remove_device(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c index b5775c6a857b..8a8b4967a101 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c @@ -228,34 +228,6 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev) * [127:96] - reserved */ -/** - * cik_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool cik_ih_prescreen_iv(struct amdgpu_device *adev) -{ - u32 ring_index = adev->irq.ih.rptr >> 2; - u16 pasid; - - switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { - case 146: - case 147: - pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; - if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) - return true; - break; - default: - /* Not a VM fault */ - return true; - } - - adev->irq.ih.rptr += 16; - return false; -} - /** * cik_ih_decode_iv - decode an interrupt vector * @@ -461,7 +433,6 @@ static const struct amd_ip_funcs cik_ih_ip_funcs = { static const struct amdgpu_ih_funcs cik_ih_funcs = { .get_wptr = cik_ih_get_wptr, - .prescreen_iv = cik_ih_prescreen_iv, .decode_iv = cik_ih_decode_iv, .set_rptr = cik_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c index df5ac4d85a00..9d3ea298e116 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c @@ -208,34 +208,6 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev) } /** - * cz_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool cz_ih_prescreen_iv(struct amdgpu_device *adev) -{ - u32 ring_index = adev->irq.ih.rptr >> 2; - u16 pasid; - - switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { - case 146: - case 147: - pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; - if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) - return true; - break; - default: - /* Not a VM fault */ - return true; - } - - adev->irq.ih.rptr += 16; - return false; -} - -/** * cz_ih_decode_iv - decode an interrupt vector * * @adev: amdgpu_device pointer @@ -442,7 +414,6 @@ static const struct amd_ip_funcs cz_ih_ip_funcs = { static const struct amdgpu_ih_funcs cz_ih_funcs = { .get_wptr = cz_ih_get_wptr, - .prescreen_iv = cz_ih_prescreen_iv, .decode_iv = cz_ih_decode_iv, .set_rptr = cz_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c index fdace004544d..e4cc1d48eaab 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c @@ -167,19 +167,6 @@ static void dce_virtual_crtc_disable(struct drm_crtc *crtc) struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); - if (crtc->primary->fb) { - int r; - struct amdgpu_bo *abo; - - abo = gem_to_amdgpu_bo(crtc->primary->fb->obj[0]); - r = amdgpu_bo_reserve(abo, true); - if (unlikely(r)) - DRM_ERROR("failed to reserve abo before unpin\n"); - else { - amdgpu_bo_unpin(abo); - amdgpu_bo_unreserve(abo); - } - } amdgpu_crtc->pll_id = ATOM_PPLL_INVALID; amdgpu_crtc->encoder = NULL; @@ -692,7 +679,9 @@ static int dce_virtual_pageflip(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->ddev->event_lock, flags); drm_crtc_vblank_put(&amdgpu_crtc->base); - schedule_work(&works->unpin_work); + amdgpu_bo_unref(&works->old_abo); + kfree(works->shared); + kfree(works); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 1454fc306783..57cb3a51bda7 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -4068,6 +4068,11 @@ static void gfx_v8_0_rlc_start(struct amdgpu_device *adev) static int gfx_v8_0_rlc_resume(struct amdgpu_device *adev) { + if (amdgpu_sriov_vf(adev)) { + gfx_v8_0_init_csb(adev); + return 0; + } + adev->gfx.rlc.funcs->stop(adev); adev->gfx.rlc.funcs->reset(adev); gfx_v8_0_init_pg(adev); @@ -4228,7 +4233,6 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev) u32 tmp; u32 rb_bufsz; u64 rb_addr, rptr_addr, wptr_gpu_addr; - int r; /* Set the write pointer delay */ WREG32(mmCP_RB_WPTR_DELAY, 0); @@ -4273,9 +4277,8 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev) amdgpu_ring_clear_ring(ring); gfx_v8_0_cp_gfx_start(adev); ring->sched.ready = true; - r = amdgpu_ring_test_helper(ring); - return r; + return 0; } static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) @@ -4364,10 +4367,9 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev) amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr)); } - r = amdgpu_ring_test_helper(kiq_ring); - if (r) - DRM_ERROR("KCQ enable failed\n"); - return r; + amdgpu_ring_commit(kiq_ring); + + return 0; } static int gfx_v8_0_deactivate_hqd(struct amdgpu_device *adev, u32 req) @@ -4704,16 +4706,32 @@ static int gfx_v8_0_kcq_resume(struct amdgpu_device *adev) if (r) goto done; - /* Test KCQs - reversing the order of rings seems to fix ring test failure - * after GPU reset - */ - for (i = adev->gfx.num_compute_rings - 1; i >= 0; i--) { +done: + return r; +} + +static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev) +{ + int r, i; + struct amdgpu_ring *ring; + + /* collect all the ring_tests here, gfx, kiq, compute */ + ring = &adev->gfx.gfx_ring[0]; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + + ring = &adev->gfx.kiq.ring; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { ring = &adev->gfx.compute_ring[i]; - r = amdgpu_ring_test_helper(ring); + amdgpu_ring_test_helper(ring); } -done: - return r; + return 0; } static int gfx_v8_0_cp_resume(struct amdgpu_device *adev) @@ -4734,6 +4752,11 @@ static int gfx_v8_0_cp_resume(struct amdgpu_device *adev) r = gfx_v8_0_kcq_resume(adev); if (r) return r; + + r = gfx_v8_0_cp_test_all_rings(adev); + if (r) + return r; + gfx_v8_0_enable_gui_idle_interrupt(adev, true); return 0; @@ -4947,14 +4970,13 @@ static bool gfx_v8_0_check_soft_reset(void *handle) static int gfx_v8_0_pre_soft_reset(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - u32 grbm_soft_reset = 0, srbm_soft_reset = 0; + u32 grbm_soft_reset = 0; if ((!adev->gfx.grbm_soft_reset) && (!adev->gfx.srbm_soft_reset)) return 0; grbm_soft_reset = adev->gfx.grbm_soft_reset; - srbm_soft_reset = adev->gfx.srbm_soft_reset; /* stop the rlc */ adev->gfx.rlc.funcs->stop(adev); @@ -5051,14 +5073,13 @@ static int gfx_v8_0_soft_reset(void *handle) static int gfx_v8_0_post_soft_reset(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - u32 grbm_soft_reset = 0, srbm_soft_reset = 0; + u32 grbm_soft_reset = 0; if ((!adev->gfx.grbm_soft_reset) && (!adev->gfx.srbm_soft_reset)) return 0; grbm_soft_reset = adev->gfx.grbm_soft_reset; - srbm_soft_reset = adev->gfx.srbm_soft_reset; if (REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP) || REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CPF) || @@ -5083,6 +5104,8 @@ static int gfx_v8_0_post_soft_reset(void *handle) REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX)) gfx_v8_0_cp_gfx_resume(adev); + gfx_v8_0_cp_test_all_rings(adev); + adev->gfx.rlc.funcs->start(adev); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index af8ccb014be3..fbca0494f871 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -86,6 +86,7 @@ MODULE_FIRMWARE("amdgpu/picasso_me.bin"); MODULE_FIRMWARE("amdgpu/picasso_mec.bin"); MODULE_FIRMWARE("amdgpu/picasso_mec2.bin"); MODULE_FIRMWARE("amdgpu/picasso_rlc.bin"); +MODULE_FIRMWARE("amdgpu/picasso_rlc_am4.bin"); MODULE_FIRMWARE("amdgpu/raven2_ce.bin"); MODULE_FIRMWARE("amdgpu/raven2_pfp.bin"); @@ -112,7 +113,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_0[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x4a2c0e68), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0xb5d3f197), SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_CACHE_INVALIDATION, 0x3fff3af3, 0x19200000), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) }; static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] = @@ -134,10 +138,7 @@ static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_UTCL1_CNTL2, 0x00030000, 0x00020000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x0000000f, 0x01000107), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x00001800, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080) }; static const struct soc15_reg_golden golden_settings_gc_9_0_vg20[] = @@ -645,7 +646,20 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev) adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version); adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version); - snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name); + /* + * For Picasso && AM4 SOCKET board, we use picasso_rlc_am4.bin + * instead of picasso_rlc.bin. + * Judgment method: + * PCO AM4: revision >= 0xC8 && revision <= 0xCF + * or revision >= 0xD8 && revision <= 0xDF + * otherwise is PCO FP5 + */ + if (!strcmp(chip_name, "picasso") && + (((adev->pdev->revision >= 0xC8) && (adev->pdev->revision <= 0xCF)) || + ((adev->pdev->revision >= 0xD8) && (adev->pdev->revision <= 0xDF)))) + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc_am4.bin", chip_name); + else + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name); err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev); if (err) goto out; @@ -2326,12 +2340,13 @@ static void gfx_v9_0_rlc_start(struct amdgpu_device *adev) #endif WREG32_FIELD15(GC, 0, RLC_CNTL, RLC_ENABLE_F32, 1); + udelay(50); /* carrizo do enable cp interrupt after cp inited */ - if (!(adev->flags & AMD_IS_APU)) + if (!(adev->flags & AMD_IS_APU)) { gfx_v9_0_enable_gui_idle_interrupt(adev, true); - - udelay(50); + udelay(50); + } #ifdef AMDGPU_RLC_DEBUG_RETRY /* RLC_GPM_GENERAL_6 : RLC Ucode version */ @@ -3572,6 +3587,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev { uint32_t data, def; + amdgpu_gfx_rlc_enter_safe_mode(adev); + /* It is disabled by HW by default */ if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { /* 1 - RLC_CGTT_MGCG_OVERRIDE */ @@ -3636,6 +3653,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev WREG32_SOC15(GC, 0, mmCP_MEM_SLP_CNTL, data); } } + + amdgpu_gfx_rlc_exit_safe_mode(adev); } static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 531aaf377592..1ad7e6b8ed1d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -56,6 +56,9 @@ MODULE_FIRMWARE("amdgpu/tonga_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris11_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris10_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris12_mc.bin"); +MODULE_FIRMWARE("amdgpu/polaris11_k_mc.bin"); +MODULE_FIRMWARE("amdgpu/polaris10_k_mc.bin"); +MODULE_FIRMWARE("amdgpu/polaris12_k_mc.bin"); static const u32 golden_settings_tonga_a11[] = { @@ -224,13 +227,39 @@ static int gmc_v8_0_init_microcode(struct amdgpu_device *adev) chip_name = "tonga"; break; case CHIP_POLARIS11: - chip_name = "polaris11"; + if (((adev->pdev->device == 0x67ef) && + ((adev->pdev->revision == 0xe0) || + (adev->pdev->revision == 0xe5))) || + ((adev->pdev->device == 0x67ff) && + ((adev->pdev->revision == 0xcf) || + (adev->pdev->revision == 0xef) || + (adev->pdev->revision == 0xff)))) + chip_name = "polaris11_k"; + else if ((adev->pdev->device == 0x67ef) && + (adev->pdev->revision == 0xe2)) + chip_name = "polaris11_k"; + else + chip_name = "polaris11"; break; case CHIP_POLARIS10: - chip_name = "polaris10"; + if ((adev->pdev->device == 0x67df) && + ((adev->pdev->revision == 0xe1) || + (adev->pdev->revision == 0xf7))) + chip_name = "polaris10_k"; + else + chip_name = "polaris10"; break; case CHIP_POLARIS12: - chip_name = "polaris12"; + if (((adev->pdev->device == 0x6987) && + ((adev->pdev->revision == 0xc0) || + (adev->pdev->revision == 0xc3))) || + ((adev->pdev->device == 0x6981) && + ((adev->pdev->revision == 0x00) || + (adev->pdev->revision == 0x01) || + (adev->pdev->revision == 0x10)))) + chip_name = "polaris12_k"; + else + chip_name = "polaris12"; break; case CHIP_FIJI: case CHIP_CARRIZO: @@ -337,7 +366,7 @@ static int gmc_v8_0_polaris_mc_load_microcode(struct amdgpu_device *adev) const struct mc_firmware_header_v1_0 *hdr; const __le32 *fw_data = NULL; const __le32 *io_mc_regs = NULL; - u32 data, vbios_version; + u32 data; int i, ucode_size, regs_size; /* Skip MC ucode loading on SR-IOV capable boards. @@ -348,13 +377,6 @@ static int gmc_v8_0_polaris_mc_load_microcode(struct amdgpu_device *adev) if (amdgpu_sriov_bios(adev)) return 0; - WREG32(mmMC_SEQ_IO_DEBUG_INDEX, 0x9F); - data = RREG32(mmMC_SEQ_IO_DEBUG_DATA); - vbios_version = data & 0xf; - - if (vbios_version == 0) - return 0; - if (!adev->gmc.fw) return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 3a4e5d8d5162..bacdaef77b6c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -244,6 +244,62 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, return 0; } +/** + * vega10_ih_prescreen_iv - prescreen an interrupt vector + * + * @adev: amdgpu_device pointer + * + * Returns true if the interrupt vector should be further processed. + */ +static bool gmc_v9_0_prescreen_iv(struct amdgpu_device *adev, + struct amdgpu_iv_entry *entry, + uint64_t addr) +{ + struct amdgpu_vm *vm; + u64 key; + int r; + + /* No PASID, can't identify faulting process */ + if (!entry->pasid) + return true; + + /* Not a retry fault */ + if (!(entry->src_data[1] & 0x80)) + return true; + + /* Track retry faults in per-VM fault FIFO. */ + spin_lock(&adev->vm_manager.pasid_lock); + vm = idr_find(&adev->vm_manager.pasid_idr, entry->pasid); + if (!vm) { + /* VM not found, process it normally */ + spin_unlock(&adev->vm_manager.pasid_lock); + return true; + } + + key = AMDGPU_VM_FAULT(entry->pasid, addr); + r = amdgpu_vm_add_fault(vm->fault_hash, key); + + /* Hash table is full or the fault is already being processed, + * ignore further page faults + */ + if (r != 0) { + spin_unlock(&adev->vm_manager.pasid_lock); + return false; + } + /* No locking required with single writer and single reader */ + r = kfifo_put(&vm->faults, key); + if (!r) { + /* FIFO is full. Ignore it until there is space */ + amdgpu_vm_clear_fault(vm->fault_hash, key); + spin_unlock(&adev->vm_manager.pasid_lock); + return false; + } + + spin_unlock(&adev->vm_manager.pasid_lock); + /* It's the first fault for this address, process it normally */ + return true; +} + static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) @@ -255,6 +311,9 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, addr = (u64)entry->src_data[0] << 12; addr |= ((u64)entry->src_data[1] & 0xf) << 44; + if (!gmc_v9_0_prescreen_iv(adev, entry, addr)) + return 1; /* This also prevents sending it to KFD */ + if (!amdgpu_sriov_vf(adev)) { status = RREG32(hub->vm_l2_pro_fault_status); WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1); @@ -659,37 +718,46 @@ static bool gmc_v9_0_keep_stolen_memory(struct amdgpu_device *adev) } } -static int gmc_v9_0_late_init(void *handle) +static int gmc_v9_0_allocate_vm_inv_eng(struct amdgpu_device *adev) { - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - /* - * The latest engine allocation on gfx9 is: - * Engine 0, 1: idle - * Engine 2, 3: firmware - * Engine 4~13: amdgpu ring, subject to change when ring number changes - * Engine 14~15: idle - * Engine 16: kfd tlb invalidation - * Engine 17: Gart flushes - */ - unsigned vm_inv_eng[AMDGPU_MAX_VMHUBS] = { 4, 4 }; + struct amdgpu_ring *ring; + unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] = + {GFXHUB_FREE_VM_INV_ENGS_BITMAP, MMHUB_FREE_VM_INV_ENGS_BITMAP}; unsigned i; - int r; + unsigned vmhub, inv_eng; - if (!gmc_v9_0_keep_stolen_memory(adev)) - amdgpu_bo_late_init(adev); + for (i = 0; i < adev->num_rings; ++i) { + ring = adev->rings[i]; + vmhub = ring->funcs->vmhub; - for(i = 0; i < adev->num_rings; ++i) { - struct amdgpu_ring *ring = adev->rings[i]; - unsigned vmhub = ring->funcs->vmhub; + inv_eng = ffs(vm_inv_engs[vmhub]); + if (!inv_eng) { + dev_err(adev->dev, "no VM inv eng for ring %s\n", + ring->name); + return -EINVAL; + } + + ring->vm_inv_eng = inv_eng - 1; + change_bit(inv_eng - 1, (unsigned long *)(&vm_inv_engs[vmhub])); - ring->vm_inv_eng = vm_inv_eng[vmhub]++; dev_info(adev->dev, "ring %s uses VM inv eng %u on hub %u\n", ring->name, ring->vm_inv_eng, ring->funcs->vmhub); } - /* Engine 16 is used for KFD and 17 for GART flushes */ - for(i = 0; i < AMDGPU_MAX_VMHUBS; ++i) - BUG_ON(vm_inv_eng[i] > 16); + return 0; +} + +static int gmc_v9_0_late_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + if (!gmc_v9_0_keep_stolen_memory(adev)) + amdgpu_bo_late_init(adev); + + r = gmc_v9_0_allocate_vm_inv_eng(adev); + if (r) + return r; if (adev->asic_type == CHIP_VEGA10 && !amdgpu_sriov_vf(adev)) { r = gmc_v9_0_ecc_available(adev); @@ -902,6 +970,9 @@ static int gmc_v9_0_sw_init(void *handle) /* This interrupt is VMC page fault.*/ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VMC, VMC_1_0__SRCID__VM_FAULT, &adev->gmc.vm_fault); + if (r) + return r; + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_UTCL2, UTCL2_1_0__SRCID__FAULT, &adev->gmc.vm_fault); @@ -934,7 +1005,7 @@ static int gmc_v9_0_sw_init(void *handle) } adev->need_swiotlb = drm_get_max_iomem() > ((u64)1 << dma_bits); - if (adev->asic_type == CHIP_VEGA20) { + if (adev->gmc.xgmi.supported) { r = gfxhub_v1_1_get_xgmi_info(adev); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h index b030ca5ea107..5c8deac65580 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h @@ -24,6 +24,16 @@ #ifndef __GMC_V9_0_H__ #define __GMC_V9_0_H__ + /* + * The latest engine allocation on gfx9 is: + * Engine 2, 3: firmware + * Engine 0, 1, 4~16: amdgpu ring, + * subject to change when ring number changes + * Engine 17: Gart flushes + */ +#define GFXHUB_FREE_VM_INV_ENGS_BITMAP 0x1FFF3 +#define MMHUB_FREE_VM_INV_ENGS_BITMAP 0x1FFF3 + extern const struct amd_ip_funcs gmc_v9_0_ip_funcs; extern const struct amdgpu_ip_block_version gmc_v9_0_ip_block; diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c index cf0fc61aebe6..a3984d10b604 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c @@ -208,34 +208,6 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev) } /** - * iceland_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool iceland_ih_prescreen_iv(struct amdgpu_device *adev) -{ - u32 ring_index = adev->irq.ih.rptr >> 2; - u16 pasid; - - switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { - case 146: - case 147: - pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; - if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) - return true; - break; - default: - /* Not a VM fault */ - return true; - } - - adev->irq.ih.rptr += 16; - return false; -} - -/** * iceland_ih_decode_iv - decode an interrupt vector * * @adev: amdgpu_device pointer @@ -440,7 +412,6 @@ static const struct amd_ip_funcs iceland_ih_ip_funcs = { static const struct amdgpu_ih_funcs iceland_ih_funcs = { .get_wptr = iceland_ih_get_wptr, - .prescreen_iv = iceland_ih_prescreen_iv, .decode_iv = iceland_ih_decode_iv, .set_rptr = iceland_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index 8cbb4655896a..b11a1c17a7f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -174,7 +174,7 @@ static int xgpu_ai_send_access_requests(struct amdgpu_device *adev, return r; } /* Retrieve checksum from mailbox2 */ - if (req == IDH_REQ_GPU_INIT_ACCESS) { + if (req == IDH_REQ_GPU_INIT_ACCESS || req == IDH_REQ_GPU_RESET_ACCESS) { adev->virt.fw_reserve.checksum_key = RREG32_NO_KIQ(SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_MSGBUF_RCV_DW2)); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c index 6f9c54978cc1..accdedd63c98 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c @@ -32,6 +32,7 @@ #define smnCPM_CONTROL 0x11180460 #define smnPCIE_CNTL2 0x11180070 #define smnPCIE_CONFIG_CNTL 0x11180044 +#define smnPCIE_CI_CNTL 0x11180080 static u32 nbio_v6_1_get_rev_id(struct amdgpu_device *adev) { @@ -270,6 +271,12 @@ static void nbio_v6_1_init_registers(struct amdgpu_device *adev) if (def != data) WREG32_PCIE(smnPCIE_CONFIG_CNTL, data); + + def = data = RREG32_PCIE(smnPCIE_CI_CNTL); + data = REG_SET_FIELD(data, PCIE_CI_CNTL, CI_SLV_ORDERING_DIS, 1); + + if (def != data) + WREG32_PCIE(smnPCIE_CI_CNTL, data); } const struct amdgpu_nbio_funcs nbio_v6_1_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c index f8cee95d61cc..4cd31a276dcd 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c @@ -31,6 +31,7 @@ #define smnCPM_CONTROL 0x11180460 #define smnPCIE_CNTL2 0x11180070 +#define smnPCIE_CI_CNTL 0x11180080 static u32 nbio_v7_4_get_rev_id(struct amdgpu_device *adev) { @@ -222,7 +223,13 @@ static void nbio_v7_4_detect_hw_virt(struct amdgpu_device *adev) static void nbio_v7_4_init_registers(struct amdgpu_device *adev) { + uint32_t def, data; + + def = data = RREG32_PCIE(smnPCIE_CI_CNTL); + data = REG_SET_FIELD(data, PCIE_CI_CNTL, CI_SLV_ORDERING_DIS, 1); + if (def != data) + WREG32_PCIE(smnPCIE_CI_CNTL, data); } const struct amdgpu_nbio_funcs nbio_v7_4_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h index 882bd83a28c4..0de00fbe9233 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h +++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h @@ -43,6 +43,8 @@ enum psp_gfx_crtl_cmd_id GFX_CTRL_CMD_ID_ENABLE_INT = 0x00050000, /* enable PSP-to-Gfx interrupt */ GFX_CTRL_CMD_ID_DISABLE_INT = 0x00060000, /* disable PSP-to-Gfx interrupt */ GFX_CTRL_CMD_ID_MODE1_RST = 0x00070000, /* trigger the Mode 1 reset */ + GFX_CTRL_CMD_ID_CONSUME_CMD = 0x000A0000, /* send interrupt to psp for updating write pointer of vf */ + GFX_CTRL_CMD_ID_DESTROY_GPCOM_RING = 0x000C0000, /* destroy GPCOM ring */ GFX_CTRL_CMD_ID_MAX = 0x000F0000, /* max command ID */ }; @@ -89,7 +91,8 @@ enum psp_gfx_cmd_id GFX_CMD_ID_LOAD_IP_FW = 0x00000006, /* load HW IP FW */ GFX_CMD_ID_DESTROY_TMR = 0x00000007, /* destroy TMR region */ GFX_CMD_ID_SAVE_RESTORE = 0x00000008, /* save/restore HW IP FW */ - + GFX_CMD_ID_SETUP_VMR = 0x00000009, /* setup VMR region */ + GFX_CMD_ID_DESTROY_VMR = 0x0000000A, /* destroy VMR region */ }; diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index e5dd052d9e06..0c6e7f9b143f 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -34,6 +34,7 @@ #include "nbio/nbio_7_4_offset.h" MODULE_FIRMWARE("amdgpu/vega20_sos.bin"); +MODULE_FIRMWARE("amdgpu/vega20_asd.bin"); MODULE_FIRMWARE("amdgpu/vega20_ta.bin"); /* address block */ @@ -100,6 +101,7 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) char fw_name[30]; int err = 0; const struct psp_firmware_header_v1_0 *sos_hdr; + const struct psp_firmware_header_v1_0 *asd_hdr; const struct ta_firmware_header_v1_0 *ta_hdr; DRM_DEBUG("\n"); @@ -132,14 +134,30 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) adev->psp.sos_start_addr = (uint8_t *)adev->psp.sys_start_addr + le32_to_cpu(sos_hdr->sos_offset_bytes); + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_asd.bin", chip_name); + err = request_firmware(&adev->psp.asd_fw, fw_name, adev->dev); + if (err) + goto out1; + + err = amdgpu_ucode_validate(adev->psp.asd_fw); + if (err) + goto out1; + + asd_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.asd_fw->data; + adev->psp.asd_fw_version = le32_to_cpu(asd_hdr->header.ucode_version); + adev->psp.asd_feature_version = le32_to_cpu(asd_hdr->ucode_feature_version); + adev->psp.asd_ucode_size = le32_to_cpu(asd_hdr->header.ucode_size_bytes); + adev->psp.asd_start_addr = (uint8_t *)asd_hdr + + le32_to_cpu(asd_hdr->header.ucode_array_offset_bytes); + snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ta.bin", chip_name); err = request_firmware(&adev->psp.ta_fw, fw_name, adev->dev); if (err) - goto out; + goto out2; err = amdgpu_ucode_validate(adev->psp.ta_fw); if (err) - goto out; + goto out2; ta_hdr = (const struct ta_firmware_header_v1_0 *)adev->psp.ta_fw->data; adev->psp.ta_xgmi_ucode_version = le32_to_cpu(ta_hdr->ta_xgmi_ucode_version); @@ -148,14 +166,18 @@ static int psp_v11_0_init_microcode(struct psp_context *psp) le32_to_cpu(ta_hdr->header.ucode_array_offset_bytes); return 0; + +out2: + release_firmware(adev->psp.ta_fw); + adev->psp.ta_fw = NULL; +out1: + release_firmware(adev->psp.asd_fw); + adev->psp.asd_fw = NULL; out: - if (err) { - dev_err(adev->dev, - "psp v11.0: Failed to load firmware \"%s\"\n", - fw_name); - release_firmware(adev->psp.sos_fw); - adev->psp.sos_fw = NULL; - } + dev_err(adev->dev, + "psp v11.0: Failed to load firmware \"%s\"\n", fw_name); + release_firmware(adev->psp.sos_fw); + adev->psp.sos_fw = NULL; return err; } @@ -171,8 +193,11 @@ static int psp_v11_0_bootloader_load_sysdrv(struct psp_context *psp) * are already been loaded. */ sol_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81); - if (sol_reg) + if (sol_reg) { + psp->sos_fw_version = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_58); + printk("sos fw version = 0x%x.\n", psp->sos_fw_version); return 0; + } /* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */ ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35), @@ -288,6 +313,13 @@ static int psp_v11_0_ring_init(struct psp_context *psp, return 0; } +static bool psp_v11_0_support_vmr_ring(struct psp_context *psp) +{ + if (amdgpu_sriov_vf(psp->adev) && psp->sos_fw_version > 0x80045) + return true; + return false; +} + static int psp_v11_0_ring_create(struct psp_context *psp, enum psp_ring_type ring_type) { @@ -296,26 +328,47 @@ static int psp_v11_0_ring_create(struct psp_context *psp, struct psp_ring *ring = &psp->km_ring; struct amdgpu_device *adev = psp->adev; - /* Write low address of the ring to C2PMSG_69 */ - psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg); - /* Write high address of the ring to C2PMSG_70 */ - psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg); - /* Write size of ring to C2PMSG_71 */ - psp_ring_reg = ring->ring_size; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg); - /* Write the ring initialization command to C2PMSG_64 */ - psp_ring_reg = ring_type; - psp_ring_reg = psp_ring_reg << 16; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg); - - /* there might be handshake issue with hardware which needs delay */ - mdelay(20); - - /* Wait for response flag (bit 31) in C2PMSG_64 */ - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), - 0x80000000, 0x8000FFFF, false); + if (psp_v11_0_support_vmr_ring(psp)) { + /* Write low address of the ring to C2PMSG_102 */ + psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_102, psp_ring_reg); + /* Write high address of the ring to C2PMSG_103 */ + psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_103, psp_ring_reg); + + /* Write the ring initialization command to C2PMSG_101 */ + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_101, + GFX_CTRL_CMD_ID_INIT_GPCOM_RING); + + /* there might be handshake issue with hardware which needs delay */ + mdelay(20); + + /* Wait for response flag (bit 31) in C2PMSG_101 */ + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101), + 0x80000000, 0x8000FFFF, false); + + } else { + /* Write low address of the ring to C2PMSG_69 */ + psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_69, psp_ring_reg); + /* Write high address of the ring to C2PMSG_70 */ + psp_ring_reg = upper_32_bits(ring->ring_mem_mc_addr); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_70, psp_ring_reg); + /* Write size of ring to C2PMSG_71 */ + psp_ring_reg = ring->ring_size; + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_71, psp_ring_reg); + /* Write the ring initialization command to C2PMSG_64 */ + psp_ring_reg = ring_type; + psp_ring_reg = psp_ring_reg << 16; + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, psp_ring_reg); + + /* there might be handshake issue with hardware which needs delay */ + mdelay(20); + + /* Wait for response flag (bit 31) in C2PMSG_64 */ + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), + 0x80000000, 0x8000FFFF, false); + } return ret; } @@ -326,15 +379,24 @@ static int psp_v11_0_ring_stop(struct psp_context *psp, int ret = 0; struct amdgpu_device *adev = psp->adev; - /* Write the ring destroy command to C2PMSG_64 */ - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_DESTROY_RINGS); + /* Write the ring destroy command*/ + if (psp_v11_0_support_vmr_ring(psp)) + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_101, + GFX_CTRL_CMD_ID_DESTROY_GPCOM_RING); + else + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, + GFX_CTRL_CMD_ID_DESTROY_RINGS); /* there might be handshake issue with hardware which needs delay */ mdelay(20); - /* Wait for response flag (bit 31) in C2PMSG_64 */ - ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), - 0x80000000, 0x80000000, false); + /* Wait for response flag (bit 31) */ + if (psp_v11_0_support_vmr_ring(psp)) + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_101), + 0x80000000, 0x80000000, false); + else + ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_64), + 0x80000000, 0x80000000, false); return ret; } @@ -373,7 +435,10 @@ static int psp_v11_0_cmd_submit(struct psp_context *psp, uint32_t rb_frame_size_dw = sizeof(struct psp_gfx_rb_frame) / 4; /* KM (GPCOM) prepare write pointer */ - psp_write_ptr_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_67); + if (psp_v11_0_support_vmr_ring(psp)) + psp_write_ptr_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_102); + else + psp_write_ptr_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_67); /* Update KM RB frame pointer to new frame */ /* write_frame ptr increments by size of rb_frame in bytes */ @@ -402,7 +467,11 @@ static int psp_v11_0_cmd_submit(struct psp_context *psp, /* Update the write Pointer in DWORDs */ psp_write_ptr_reg = (psp_write_ptr_reg + rb_frame_size_dw) % ring_size_dw; - WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_67, psp_write_ptr_reg); + if (psp_v11_0_support_vmr_ring(psp)) { + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_102, psp_write_ptr_reg); + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_101, GFX_CTRL_CMD_ID_CONSUME_CMD); + } else + WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_67, psp_write_ptr_reg); return 0; } @@ -547,7 +616,7 @@ static int psp_v11_0_mode1_reset(struct psp_context *psp) /*send the mode 1 reset command*/ WREG32(offset, GFX_CTRL_CMD_ID_MODE1_RST); - mdelay(1000); + msleep(500); offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33); @@ -640,7 +709,7 @@ static int psp_v11_0_xgmi_set_topology_info(struct psp_context *psp, return psp_xgmi_invoke(psp, TA_COMMAND_XGMI__SET_TOPOLOGY_INFO); } -static u64 psp_v11_0_xgmi_get_hive_id(struct psp_context *psp) +static int psp_v11_0_xgmi_get_hive_id(struct psp_context *psp, uint64_t *hive_id) { struct ta_xgmi_shared_memory *xgmi_cmd; int ret; @@ -653,12 +722,14 @@ static u64 psp_v11_0_xgmi_get_hive_id(struct psp_context *psp) /* Invoke xgmi ta to get hive id */ ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id); if (ret) - return 0; - else - return xgmi_cmd->xgmi_out_message.get_hive_id.hive_id; + return ret; + + *hive_id = xgmi_cmd->xgmi_out_message.get_hive_id.hive_id; + + return 0; } -static u64 psp_v11_0_xgmi_get_node_id(struct psp_context *psp) +static int psp_v11_0_xgmi_get_node_id(struct psp_context *psp, uint64_t *node_id) { struct ta_xgmi_shared_memory *xgmi_cmd; int ret; @@ -671,9 +742,11 @@ static u64 psp_v11_0_xgmi_get_node_id(struct psp_context *psp) /* Invoke xgmi ta to get the node id */ ret = psp_xgmi_invoke(psp, xgmi_cmd->cmd_id); if (ret) - return 0; - else - return xgmi_cmd->xgmi_out_message.get_node_id.node_id; + return ret; + + *node_id = xgmi_cmd->xgmi_out_message.get_node_id.node_id; + + return 0; } static const struct psp_funcs psp_v11_0_funcs = { @@ -692,6 +765,7 @@ static const struct psp_funcs psp_v11_0_funcs = { .xgmi_set_topology_info = psp_v11_0_xgmi_set_topology_info, .xgmi_get_hive_id = psp_v11_0_xgmi_get_hive_id, .xgmi_get_node_id = psp_v11_0_xgmi_get_node_id, + .support_vmr_ring = psp_v11_0_support_vmr_ring, }; void psp_v11_0_set_psp_funcs(struct psp_context *psp) diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c index 7efb823dd3b1..79694ff16969 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c @@ -240,8 +240,11 @@ static int psp_v3_1_bootloader_load_sos(struct psp_context *psp) * are already been loaded. */ sol_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81); - if (sol_reg) + if (sol_reg) { + psp->sos_fw_version = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_58); + printk("sos fw version = 0x%x.\n", psp->sos_fw_version); return 0; + } /* Wait for bootloader to signify that is ready having bit 31 of C2PMSG_35 set to 1 */ ret = psp_wait_for(psp, SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_35), @@ -592,7 +595,7 @@ static int psp_v3_1_mode1_reset(struct psp_context *psp) /*send the mode 1 reset command*/ WREG32(offset, GFX_CTRL_CMD_ID_MODE1_RST); - mdelay(1000); + msleep(500); offset = SOC15_REG_OFFSET(MP0, 0, mmMP0_SMN_C2PMSG_33); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 4b6d3e5c821f..6811a5d05b27 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -78,7 +78,6 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC1_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_PAGE, 0x000003ff, 0x000003c0), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_WATERMK, 0xfc000000, 0x00000000), - SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831f07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CLK_CTRL, 0xffffffff, 0x3f000100), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_IB_CNTL, 0x800f0100, 0x00000100), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000), @@ -96,6 +95,7 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = { static const struct soc15_reg_golden golden_settings_sdma_vg10[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104002), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002), + SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104002), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002) }; @@ -103,6 +103,7 @@ static const struct soc15_reg_golden golden_settings_sdma_vg10[] = { static const struct soc15_reg_golden golden_settings_sdma_vg12[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104001), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001), + SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104001), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001) }; @@ -1458,8 +1459,7 @@ static bool sdma_v4_0_fw_support_paging_queue(struct amdgpu_device *adev) /*return fw_version >= 31;*/ return false; case CHIP_VEGA20: - /*return fw_version >= 115;*/ - return false; + return fw_version >= 123; default: return false; } @@ -1706,13 +1706,15 @@ static int sdma_v4_0_process_trap_irq(struct amdgpu_device *adev, amdgpu_fence_process(&adev->sdma.instance[instance].ring); break; case 1: - /* XXX compute */ + if (adev->asic_type == CHIP_VEGA20) + amdgpu_fence_process(&adev->sdma.instance[instance].page); break; case 2: /* XXX compute */ break; case 3: - amdgpu_fence_process(&adev->sdma.instance[instance].page); + if (adev->asic_type != CHIP_VEGA20) + amdgpu_fence_process(&adev->sdma.instance[instance].page); break; } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c index b3d7d9f83202..2938fb9f17cc 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c @@ -118,19 +118,6 @@ static u32 si_ih_get_wptr(struct amdgpu_device *adev) return (wptr & adev->irq.ih.ptr_mask); } -/** - * si_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool si_ih_prescreen_iv(struct amdgpu_device *adev) -{ - /* Process all interrupts */ - return true; -} - static void si_ih_decode_iv(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { @@ -301,7 +288,6 @@ static const struct amd_ip_funcs si_ih_ip_funcs = { static const struct amdgpu_ih_funcs si_ih_funcs = { .get_wptr = si_ih_get_wptr, - .prescreen_iv = si_ih_prescreen_iv, .decode_iv = si_ih_decode_iv, .set_rptr = si_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 83624e150ca7..8849b74078d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -507,6 +507,9 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) return -EINVAL; } + if (adev->asic_type == CHIP_VEGA20) + adev->gmc.xgmi.supported = true; + if (adev->flags & AMD_IS_APU) adev->nbio_funcs = &nbio_v7_0_funcs; else if (adev->asic_type == CHIP_VEGA20) diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h index 958b10a57073..49c262540940 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15_common.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15_common.h @@ -49,14 +49,19 @@ #define SOC15_WAIT_ON_RREG(ip, inst, reg, expected_value, mask, ret) \ do { \ + uint32_t old_ = 0; \ uint32_t tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \ uint32_t loop = adev->usec_timeout; \ while ((tmp_ & (mask)) != (expected_value)) { \ - udelay(2); \ + if (old_ != tmp_) { \ + loop = adev->usec_timeout; \ + old_ = tmp_; \ + } else \ + udelay(1); \ tmp_ = RREG32(adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg); \ loop--; \ if (!loop) { \ - DRM_ERROR("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \ + DRM_WARN("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \ inst, #reg, (unsigned)expected_value, (unsigned)(tmp_ & (mask))); \ ret = -ETIMEDOUT; \ break; \ diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c index dcdbb4d72472..15da06ddeb75 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c @@ -219,34 +219,6 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev) } /** - * tonga_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool tonga_ih_prescreen_iv(struct amdgpu_device *adev) -{ - u32 ring_index = adev->irq.ih.rptr >> 2; - u16 pasid; - - switch (le32_to_cpu(adev->irq.ih.ring[ring_index]) & 0xff) { - case 146: - case 147: - pasid = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]) >> 16; - if (!pasid || amdgpu_vm_pasid_fault_credit(adev, pasid)) - return true; - break; - default: - /* Not a VM fault */ - return true; - } - - adev->irq.ih.rptr += 16; - return false; -} - -/** * tonga_ih_decode_iv - decode an interrupt vector * * @adev: amdgpu_device pointer @@ -506,7 +478,6 @@ static const struct amd_ip_funcs tonga_ih_ip_funcs = { static const struct amdgpu_ih_funcs tonga_ih_funcs = { .get_wptr = tonga_ih_get_wptr, - .prescreen_iv = tonga_ih_prescreen_iv, .decode_iv = tonga_ih_decode_iv, .set_rptr = tonga_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index 089645e78f98..aef924026a28 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -435,7 +435,7 @@ static int uvd_v7_0_sw_init(void *handle) continue; if (!amdgpu_sriov_vf(adev)) { ring = &adev->uvd.inst[j].ring; - sprintf(ring->name, "uvd<%d>", j); + sprintf(ring->name, "uvd_%d", ring->me); r = amdgpu_ring_init(adev, ring, 512, &adev->uvd.inst[j].irq, 0); if (r) return r; @@ -443,7 +443,7 @@ static int uvd_v7_0_sw_init(void *handle) for (i = 0; i < adev->uvd.num_enc_rings; ++i) { ring = &adev->uvd.inst[j].ring_enc[i]; - sprintf(ring->name, "uvd_enc%d<%d>", i, j); + sprintf(ring->name, "uvd_enc_%d.%d", ring->me, i); if (amdgpu_sriov_vf(adev)) { ring->use_doorbell = true; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index c1a03505f956..89bb2fef90eb 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -48,6 +48,7 @@ static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev); static void vcn_v1_0_set_jpeg_ring_funcs(struct amdgpu_device *adev); static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev); static void vcn_v1_0_jpeg_ring_set_patch_ring(struct amdgpu_ring *ring, uint32_t ptr); +static int vcn_v1_0_set_powergating_state(void *handle, enum amd_powergating_state state); /** * vcn_v1_0_early_init - set function pointers @@ -213,8 +214,9 @@ static int vcn_v1_0_hw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct amdgpu_ring *ring = &adev->vcn.ring_dec; - if (RREG32_SOC15(VCN, 0, mmUVD_STATUS)) - vcn_v1_0_stop(adev); + if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || + RREG32_SOC15(VCN, 0, mmUVD_STATUS)) + vcn_v1_0_set_powergating_state(adev, AMD_PG_STATE_GATE); ring->sched.ready = false; @@ -1086,7 +1088,8 @@ static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev) WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0, ~UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK); - /* initialize wptr */ + /* initialize JPEG wptr */ + ring = &adev->vcn.ring_jpeg; ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR); /* copy patch commands to the jpeg ring */ @@ -1158,21 +1161,29 @@ static int vcn_v1_0_stop_spg_mode(struct amdgpu_device *adev) static int vcn_v1_0_stop_dpg_mode(struct amdgpu_device *adev) { int ret_code = 0; + uint32_t tmp; /* Wait for power status to be UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF */ SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); - if (!ret_code) { - int tmp = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR) & 0x7FFFFFFF; - /* wait for read ptr to be equal to write ptr */ - SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code); + /* wait for read ptr to be equal to write ptr */ + tmp = RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RB_RPTR, tmp, 0xFFFFFFFF, ret_code); - SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, - UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, - UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); - } + tmp = RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RB_RPTR2, tmp, 0xFFFFFFFF, ret_code); + + tmp = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR); + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_JRBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code); + + tmp = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR) & 0x7FFFFFFF; + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code); + + SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS, + UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF, + UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code); /* disable dynamic power gating mode */ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0, diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index d84b687240d1..2c250b01a903 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -220,90 +220,6 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev) } /** - * vega10_ih_prescreen_iv - prescreen an interrupt vector - * - * @adev: amdgpu_device pointer - * - * Returns true if the interrupt vector should be further processed. - */ -static bool vega10_ih_prescreen_iv(struct amdgpu_device *adev) -{ - u32 ring_index = adev->irq.ih.rptr >> 2; - u32 dw0, dw3, dw4, dw5; - u16 pasid; - u64 addr, key; - struct amdgpu_vm *vm; - int r; - - dw0 = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw3 = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - dw4 = le32_to_cpu(adev->irq.ih.ring[ring_index + 4]); - dw5 = le32_to_cpu(adev->irq.ih.ring[ring_index + 5]); - - /* Filter retry page faults, let only the first one pass. If - * there are too many outstanding faults, ignore them until - * some faults get cleared. - */ - switch (dw0 & 0xff) { - case SOC15_IH_CLIENTID_VMC: - case SOC15_IH_CLIENTID_UTCL2: - break; - default: - /* Not a VM fault */ - return true; - } - - pasid = dw3 & 0xffff; - /* No PASID, can't identify faulting process */ - if (!pasid) - return true; - - /* Not a retry fault, check fault credit */ - if (!(dw5 & 0x80)) { - if (!amdgpu_vm_pasid_fault_credit(adev, pasid)) - goto ignore_iv; - return true; - } - - /* Track retry faults in per-VM fault FIFO. */ - spin_lock(&adev->vm_manager.pasid_lock); - vm = idr_find(&adev->vm_manager.pasid_idr, pasid); - addr = ((u64)(dw5 & 0xf) << 44) | ((u64)dw4 << 12); - key = AMDGPU_VM_FAULT(pasid, addr); - if (!vm) { - /* VM not found, process it normally */ - spin_unlock(&adev->vm_manager.pasid_lock); - return true; - } else { - r = amdgpu_vm_add_fault(vm->fault_hash, key); - - /* Hash table is full or the fault is already being processed, - * ignore further page faults - */ - if (r != 0) { - spin_unlock(&adev->vm_manager.pasid_lock); - goto ignore_iv; - } - } - /* No locking required with single writer and single reader */ - r = kfifo_put(&vm->faults, key); - if (!r) { - /* FIFO is full. Ignore it until there is space */ - amdgpu_vm_clear_fault(vm->fault_hash, key); - spin_unlock(&adev->vm_manager.pasid_lock); - goto ignore_iv; - } - - spin_unlock(&adev->vm_manager.pasid_lock); - /* It's the first fault for this address, process it normally */ - return true; - -ignore_iv: - adev->irq.ih.rptr += 32; - return false; -} - -/** * vega10_ih_decode_iv - decode an interrupt vector * * @adev: amdgpu_device pointer @@ -487,7 +403,6 @@ const struct amd_ip_funcs vega10_ih_ip_funcs = { static const struct amdgpu_ih_funcs vega10_ih_funcs = { .get_wptr = vega10_ih_get_wptr, - .prescreen_iv = vega10_ih_prescreen_iv, .decode_iv = vega10_ih_decode_iv, .set_rptr = vega10_ih_set_rptr }; diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index ff2906c215fa..77e367459101 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -87,9 +87,9 @@ static u32 vi_pcie_rreg(struct amdgpu_device *adev, u32 reg) u32 r; spin_lock_irqsave(&adev->pcie_idx_lock, flags); - WREG32(mmPCIE_INDEX, reg); - (void)RREG32(mmPCIE_INDEX); - r = RREG32(mmPCIE_DATA); + WREG32_NO_KIQ(mmPCIE_INDEX, reg); + (void)RREG32_NO_KIQ(mmPCIE_INDEX); + r = RREG32_NO_KIQ(mmPCIE_DATA); spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); return r; } @@ -99,10 +99,10 @@ static void vi_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v) unsigned long flags; spin_lock_irqsave(&adev->pcie_idx_lock, flags); - WREG32(mmPCIE_INDEX, reg); - (void)RREG32(mmPCIE_INDEX); - WREG32(mmPCIE_DATA, v); - (void)RREG32(mmPCIE_DATA); + WREG32_NO_KIQ(mmPCIE_INDEX, reg); + (void)RREG32_NO_KIQ(mmPCIE_INDEX); + WREG32_NO_KIQ(mmPCIE_DATA, v); + (void)RREG32_NO_KIQ(mmPCIE_DATA); spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); } @@ -123,8 +123,8 @@ static void vi_smc_wreg(struct amdgpu_device *adev, u32 reg, u32 v) unsigned long flags; spin_lock_irqsave(&adev->smc_idx_lock, flags); - WREG32(mmSMC_IND_INDEX_11, (reg)); - WREG32(mmSMC_IND_DATA_11, (v)); + WREG32_NO_KIQ(mmSMC_IND_INDEX_11, (reg)); + WREG32_NO_KIQ(mmSMC_IND_DATA_11, (v)); spin_unlock_irqrestore(&adev->smc_idx_lock, flags); } diff --git a/drivers/gpu/drm/amd/amdkfd/Kconfig b/drivers/gpu/drm/amd/amdkfd/Kconfig index fbf0ee5201c3..c3613604a4f8 100644 --- a/drivers/gpu/drm/amd/amdkfd/Kconfig +++ b/drivers/gpu/drm/amd/amdkfd/Kconfig @@ -4,8 +4,8 @@ config HSA_AMD bool "HSA kernel driver for AMD GPU devices" - depends on DRM_AMDGPU && X86_64 - imply AMD_IOMMU_V2 + depends on DRM_AMDGPU && (X86_64 || ARM64) + imply AMD_IOMMU_V2 if X86_64 select MMU_NOTIFIER help Enable this if you want to use HSA features on AMD GPU devices. diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index 5f4062b41add..083bd8114db1 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -33,6 +33,7 @@ #include <linux/time.h> #include <linux/mm.h> #include <linux/mman.h> +#include <linux/dma-buf.h> #include <asm/processor.h> #include "kfd_priv.h" #include "kfd_device_queue_manager.h" @@ -157,8 +158,7 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties, } if ((args->ring_base_address) && - (!access_ok(VERIFY_WRITE, - (const void __user *) args->ring_base_address, + (!access_ok((const void __user *) args->ring_base_address, sizeof(uint64_t)))) { pr_err("Can't access ring base address\n"); return -EFAULT; @@ -169,31 +169,27 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties, return -EINVAL; } - if (!access_ok(VERIFY_WRITE, - (const void __user *) args->read_pointer_address, + if (!access_ok((const void __user *) args->read_pointer_address, sizeof(uint32_t))) { pr_err("Can't access read pointer\n"); return -EFAULT; } - if (!access_ok(VERIFY_WRITE, - (const void __user *) args->write_pointer_address, + if (!access_ok((const void __user *) args->write_pointer_address, sizeof(uint32_t))) { pr_err("Can't access write pointer\n"); return -EFAULT; } if (args->eop_buffer_address && - !access_ok(VERIFY_WRITE, - (const void __user *) args->eop_buffer_address, + !access_ok((const void __user *) args->eop_buffer_address, sizeof(uint32_t))) { pr_debug("Can't access eop buffer"); return -EFAULT; } if (args->ctx_save_restore_address && - !access_ok(VERIFY_WRITE, - (const void __user *) args->ctx_save_restore_address, + !access_ok((const void __user *) args->ctx_save_restore_address, sizeof(uint32_t))) { pr_debug("Can't access ctx save restore buffer"); return -EFAULT; @@ -364,8 +360,7 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p, } if ((args->ring_base_address) && - (!access_ok(VERIFY_WRITE, - (const void __user *) args->ring_base_address, + (!access_ok((const void __user *) args->ring_base_address, sizeof(uint64_t)))) { pr_err("Can't access ring base address\n"); return -EFAULT; @@ -1273,6 +1268,12 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep, return -EINVAL; } + if (flags & KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL) { + if (args->size != kfd_doorbell_process_slice(dev)) + return -EINVAL; + offset = kfd_get_process_doorbells(dev, p); + } + mutex_lock(&p->mutex); pdd = kfd_bind_process_to_device(dev, p); @@ -1550,6 +1551,115 @@ copy_from_user_failed: return err; } +static int kfd_ioctl_get_dmabuf_info(struct file *filep, + struct kfd_process *p, void *data) +{ + struct kfd_ioctl_get_dmabuf_info_args *args = data; + struct kfd_dev *dev = NULL; + struct kgd_dev *dma_buf_kgd; + void *metadata_buffer = NULL; + uint32_t flags; + unsigned int i; + int r; + + /* Find a KFD GPU device that supports the get_dmabuf_info query */ + for (i = 0; kfd_topology_enum_kfd_devices(i, &dev) == 0; i++) + if (dev) + break; + if (!dev) + return -EINVAL; + + if (args->metadata_ptr) { + metadata_buffer = kzalloc(args->metadata_size, GFP_KERNEL); + if (!metadata_buffer) + return -ENOMEM; + } + + /* Get dmabuf info from KGD */ + r = amdgpu_amdkfd_get_dmabuf_info(dev->kgd, args->dmabuf_fd, + &dma_buf_kgd, &args->size, + metadata_buffer, args->metadata_size, + &args->metadata_size, &flags); + if (r) + goto exit; + + /* Reverse-lookup gpu_id from kgd pointer */ + dev = kfd_device_by_kgd(dma_buf_kgd); + if (!dev) { + r = -EINVAL; + goto exit; + } + args->gpu_id = dev->id; + args->flags = flags; + + /* Copy metadata buffer to user mode */ + if (metadata_buffer) { + r = copy_to_user((void __user *)args->metadata_ptr, + metadata_buffer, args->metadata_size); + if (r != 0) + r = -EFAULT; + } + +exit: + kfree(metadata_buffer); + + return r; +} + +static int kfd_ioctl_import_dmabuf(struct file *filep, + struct kfd_process *p, void *data) +{ + struct kfd_ioctl_import_dmabuf_args *args = data; + struct kfd_process_device *pdd; + struct dma_buf *dmabuf; + struct kfd_dev *dev; + int idr_handle; + uint64_t size; + void *mem; + int r; + + dev = kfd_device_by_id(args->gpu_id); + if (!dev) + return -EINVAL; + + dmabuf = dma_buf_get(args->dmabuf_fd); + if (IS_ERR(dmabuf)) + return PTR_ERR(dmabuf); + + mutex_lock(&p->mutex); + + pdd = kfd_bind_process_to_device(dev, p); + if (IS_ERR(pdd)) { + r = PTR_ERR(pdd); + goto err_unlock; + } + + r = amdgpu_amdkfd_gpuvm_import_dmabuf(dev->kgd, dmabuf, + args->va_addr, pdd->vm, + (struct kgd_mem **)&mem, &size, + NULL); + if (r) + goto err_unlock; + + idr_handle = kfd_process_device_create_obj_handle(pdd, mem); + if (idr_handle < 0) { + r = -EFAULT; + goto err_free; + } + + mutex_unlock(&p->mutex); + + args->handle = MAKE_HANDLE(args->gpu_id, idr_handle); + + return 0; + +err_free: + amdgpu_amdkfd_gpuvm_free_memory_of_gpu(dev->kgd, (struct kgd_mem *)mem); +err_unlock: + mutex_unlock(&p->mutex); + return r; +} + #define AMDKFD_IOCTL_DEF(ioctl, _func, _flags) \ [_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, \ .cmd_drv = 0, .name = #ioctl} @@ -1635,7 +1745,13 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { kfd_ioctl_set_cu_mask, 0), AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_QUEUE_WAVE_STATE, - kfd_ioctl_get_queue_wave_state, 0) + kfd_ioctl_get_queue_wave_state, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_DMABUF_INFO, + kfd_ioctl_get_dmabuf_info, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_IMPORT_DMABUF, + kfd_ioctl_import_dmabuf, 0), }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c index c02adbbeef2a..5d85ff341385 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c @@ -853,7 +853,7 @@ static int kfd_fill_mem_info_for_cpu(int numa_node_id, int *avail_size, */ pgdat = NODE_DATA(numa_node_id); for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) - mem_in_bytes += pgdat->node_zones[zone_type].managed_pages; + mem_in_bytes += zone_managed_pages(&pgdat->node_zones[zone_type]); mem_in_bytes <<= PAGE_SHIFT; sub_type_hdr->length_low = lower_32_bits(mem_in_bytes); @@ -863,6 +863,7 @@ static int kfd_fill_mem_info_for_cpu(int numa_node_id, int *avail_size, return 0; } +#if CONFIG_X86_64 static int kfd_fill_iolink_info_for_cpu(int numa_node_id, int *avail_size, uint32_t *num_entries, struct crat_subtype_iolink *sub_type_hdr) @@ -905,6 +906,7 @@ static int kfd_fill_iolink_info_for_cpu(int numa_node_id, int *avail_size, return 0; } +#endif /* kfd_create_vcrat_image_cpu - Create Virtual CRAT for CPU * @@ -920,7 +922,9 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) struct crat_subtype_generic *sub_type_hdr; int avail_size = *size; int numa_node_id; +#ifdef CONFIG_X86_64 uint32_t entries = 0; +#endif int ret = 0; if (!pcrat_image || avail_size < VCRAT_SIZE_FOR_CPU) @@ -982,6 +986,7 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) sub_type_hdr->length); /* Fill in Subtype: IO Link */ +#ifdef CONFIG_X86_64 ret = kfd_fill_iolink_info_for_cpu(numa_node_id, &avail_size, &entries, (struct crat_subtype_iolink *)sub_type_hdr); @@ -992,6 +997,9 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr + sub_type_hdr->length * entries); +#else + pr_info("IO link not available for non x86 platforms\n"); +#endif crat_table->num_domains++; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 9ed14a11afa2..8be9677c0c07 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -378,7 +378,13 @@ static const struct kfd_deviceid supported_devices[] = { { 0x6864, &vega10_device_info }, /* Vega10 */ { 0x6867, &vega10_device_info }, /* Vega10 */ { 0x6868, &vega10_device_info }, /* Vega10 */ + { 0x6869, &vega10_device_info }, /* Vega10 */ + { 0x686A, &vega10_device_info }, /* Vega10 */ + { 0x686B, &vega10_device_info }, /* Vega10 */ { 0x686C, &vega10_vf_device_info }, /* Vega10 vf*/ + { 0x686D, &vega10_device_info }, /* Vega10 */ + { 0x686E, &vega10_device_info }, /* Vega10 */ + { 0x686F, &vega10_device_info }, /* Vega10 */ { 0x687F, &vega10_device_info }, /* Vega10 */ { 0x69A0, &vega12_device_info }, /* Vega12 */ { 0x69A1, &vega12_device_info }, /* Vega12 */ @@ -389,6 +395,7 @@ static const struct kfd_deviceid supported_devices[] = { { 0x66a1, &vega20_device_info }, /* Vega20 */ { 0x66a2, &vega20_device_info }, /* Vega20 */ { 0x66a3, &vega20_device_info }, /* Vega20 */ + { 0x66a4, &vega20_device_info }, /* Vega20 */ { 0x66a7, &vega20_device_info }, /* Vega20 */ { 0x66af, &vega20_device_info } /* Vega20 */ }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index dec8e64f36bd..0689d4ccbbc0 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -793,6 +793,7 @@ struct kfd_topology_device *kfd_topology_device_by_proximity_domain( struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id); struct kfd_dev *kfd_device_by_id(uint32_t gpu_id); struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev); +struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd); int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev); int kfd_numa_node_to_apic_id(int numa_node_id); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index aa793fcbbdcc..09da91644f9f 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -101,7 +101,25 @@ struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev) down_read(&topology_lock); list_for_each_entry(top_dev, &topology_device_list, list) - if (top_dev->gpu->pdev == pdev) { + if (top_dev->gpu && top_dev->gpu->pdev == pdev) { + device = top_dev->gpu; + break; + } + + up_read(&topology_lock); + + return device; +} + +struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd) +{ + struct kfd_topology_device *top_dev; + struct kfd_dev *device = NULL; + + down_read(&topology_lock); + + list_for_each_entry(top_dev, &topology_device_list, list) + if (top_dev->gpu && top_dev->gpu->kgd == kgd) { device = top_dev->gpu; break; } @@ -1075,8 +1093,6 @@ static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu) * the GPU device is not already present in the topology device * list then return NULL. This means a new topology device has to * be created for this GPU. - * TODO: Rather than assiging @gpu to first topology device withtout - * gpu attached, it will better to have more stringent check. */ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) { @@ -1084,12 +1100,20 @@ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) struct kfd_topology_device *out_dev = NULL; down_write(&topology_lock); - list_for_each_entry(dev, &topology_device_list, list) + list_for_each_entry(dev, &topology_device_list, list) { + /* Discrete GPUs need their own topology device list + * entries. Don't assign them to CPU/APU nodes. + */ + if (!gpu->device_info->needs_iommu_device && + dev->node_props.cpu_cores_count) + continue; + if (!dev->gpu && (dev->node_props.simd_count > 0)) { dev->gpu = gpu; out_dev = dev; break; } + } up_write(&topology_lock); return out_dev; } @@ -1374,7 +1398,6 @@ int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev) static int kfd_cpumask_to_apic_id(const struct cpumask *cpumask) { - const struct cpuinfo_x86 *cpuinfo; int first_cpu_of_numa_node; if (!cpumask || cpumask == cpu_none_mask) @@ -1382,9 +1405,11 @@ static int kfd_cpumask_to_apic_id(const struct cpumask *cpumask) first_cpu_of_numa_node = cpumask_first(cpumask); if (first_cpu_of_numa_node >= nr_cpu_ids) return -1; - cpuinfo = &cpu_data(first_cpu_of_numa_node); - - return cpuinfo->apicid; +#ifdef CONFIG_X86_64 + return cpu_data(first_cpu_of_numa_node).apicid; +#else + return first_cpu_of_numa_node; +#endif } /* kfd_numa_node_to_apic_id - Returns the APIC ID of the first logical processor diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 32e791d9b9a8..f4fa40c387d3 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -23,6 +23,9 @@ * */ +/* The caprices of the preprocessor require that this be declared right here */ +#define CREATE_TRACE_POINTS + #include "dm_services_types.h" #include "dc.h" #include "dc/inc/core_types.h" @@ -54,6 +57,7 @@ #include <drm/drmP.h> #include <drm/drm_atomic.h> +#include <drm/drm_atomic_uapi.h> #include <drm/drm_atomic_helper.h> #include <drm/drm_dp_mst_helper.h> #include <drm/drm_fb_helper.h> @@ -72,6 +76,7 @@ #include "modules/inc/mod_freesync.h" #include "modules/power/power_helpers.h" +#include "modules/inc/mod_info_packet.h" #define FIRMWARE_RAVEN_DMCU "amdgpu/raven_dmcu.bin" MODULE_FIRMWARE(FIRMWARE_RAVEN_DMCU); @@ -129,6 +134,8 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state); static int amdgpu_dm_atomic_check(struct drm_device *dev, struct drm_atomic_state *state); +static void handle_cursor_update(struct drm_plane *plane, + struct drm_plane_state *old_plane_state); @@ -324,12 +331,29 @@ static void dm_crtc_high_irq(void *interrupt_params) struct common_irq_params *irq_params = interrupt_params; struct amdgpu_device *adev = irq_params->adev; struct amdgpu_crtc *acrtc; + struct dm_crtc_state *acrtc_state; acrtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_VBLANK); if (acrtc) { drm_crtc_handle_vblank(&acrtc->base); amdgpu_dm_crtc_handle_crc_irq(&acrtc->base); + + acrtc_state = to_dm_crtc_state(acrtc->base.state); + + if (acrtc_state->stream && + acrtc_state->vrr_params.supported && + acrtc_state->freesync_config.state == VRR_STATE_ACTIVE_VARIABLE) { + mod_freesync_handle_v_update( + adev->dm.freesync_module, + acrtc_state->stream, + &acrtc_state->vrr_params); + + dc_stream_adjust_vmin_vmax( + adev->dm.dc, + acrtc_state->stream, + &acrtc_state->vrr_params.adjust); + } } } @@ -398,6 +422,8 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) /* Zero all the fields */ memset(&init_data, 0, sizeof(init_data)); + mutex_init(&adev->dm.dc_lock); + if(amdgpu_dm_irq_init(adev)) { DRM_ERROR("amdgpu: failed to initialize DM IRQ support.\n"); goto error; @@ -512,6 +538,9 @@ static void amdgpu_dm_fini(struct amdgpu_device *adev) /* DC Destroy TODO: Replace destroy DAL */ if (adev->dm.dc) dc_destroy(&adev->dm.dc); + + mutex_destroy(&adev->dm.dc_lock); + return; } @@ -670,22 +699,36 @@ static void s3_handle_mst(struct drm_device *dev, bool suspend) { struct amdgpu_dm_connector *aconnector; struct drm_connector *connector; + struct drm_dp_mst_topology_mgr *mgr; + int ret; + bool need_hotplug = false; drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - aconnector = to_amdgpu_dm_connector(connector); - if (aconnector->dc_link->type == dc_connection_mst_branch && - !aconnector->mst_port) { + list_for_each_entry(connector, &dev->mode_config.connector_list, + head) { + aconnector = to_amdgpu_dm_connector(connector); + if (aconnector->dc_link->type != dc_connection_mst_branch || + aconnector->mst_port) + continue; - if (suspend) - drm_dp_mst_topology_mgr_suspend(&aconnector->mst_mgr); - else - drm_dp_mst_topology_mgr_resume(&aconnector->mst_mgr); - } + mgr = &aconnector->mst_mgr; + + if (suspend) { + drm_dp_mst_topology_mgr_suspend(mgr); + } else { + ret = drm_dp_mst_topology_mgr_resume(mgr); + if (ret < 0) { + drm_dp_mst_topology_mgr_set_mst(mgr, false); + need_hotplug = true; + } + } } drm_modeset_unlock(&dev->mode_config.connection_mutex); + + if (need_hotplug) + drm_kms_helper_hotplug_event(dev); } /** @@ -869,7 +912,6 @@ static int dm_resume(void *handle) struct drm_plane_state *new_plane_state; struct dm_plane_state *dm_new_plane_state; enum dc_connection_type new_connection_type = dc_connection_none; - int ret; int i; /* power on hardware */ @@ -942,13 +984,13 @@ static int dm_resume(void *handle) } } - ret = drm_atomic_helper_resume(ddev, dm->cached_state); + drm_atomic_helper_resume(ddev, dm->cached_state); dm->cached_state = NULL; amdgpu_dm_irq_resume_late(adev); - return ret; + return 0; } /** @@ -1730,7 +1772,7 @@ static int amdgpu_dm_backlight_update_status(struct backlight_device *bd) + caps.min_input_signal * 0x101; if (dc_link_set_backlight_level(dm->backlight_link, - brightness, 0, 0)) + brightness, 0)) return 0; else return 1; @@ -2698,9 +2740,9 @@ static void fill_audio_info(struct audio_info *audio_info, cea_revision = drm_connector->display_info.cea_rev; - strncpy(audio_info->display_name, + strscpy(audio_info->display_name, edid_caps->display_name, - AUDIO_INFO_DISPLAY_NAME_SIZE_IN_CHARS - 1); + AUDIO_INFO_DISPLAY_NAME_SIZE_IN_CHARS); if (cea_revision >= 3) { audio_info->mode_count = edid_caps->audio_mode_count; @@ -2930,6 +2972,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, if (dm_state && dm_state->freesync_capable) stream->ignore_msa_timing_param = true; + finish: if (sink && sink->sink_signal == SIGNAL_TYPE_VIRTUAL && aconnector->base.force != DRM_FORCE_ON) dc_sink_release(sink); @@ -2996,11 +3039,12 @@ dm_crtc_duplicate_state(struct drm_crtc *crtc) dc_stream_retain(state->stream); } - state->adjust = cur->adjust; + state->vrr_params = cur->vrr_params; state->vrr_infopacket = cur->vrr_infopacket; state->abm_level = cur->abm_level; state->vrr_supported = cur->vrr_supported; state->freesync_config = cur->freesync_config; + state->crc_enabled = cur->crc_enabled; /* TODO Duplicate dc_stream after objects are stream object is flattened */ @@ -3094,10 +3138,8 @@ int amdgpu_dm_connector_atomic_set_property(struct drm_connector *connector, rmx_type = RMX_FULL; break; case DRM_MODE_SCALE_NONE: - rmx_type = RMX_OFF; - break; default: - rmx_type = RMX_ASPECT; + rmx_type = RMX_OFF; break; } @@ -3210,10 +3252,11 @@ void amdgpu_dm_connector_funcs_reset(struct drm_connector *connector) state = kzalloc(sizeof(*state), GFP_KERNEL); if (state) { - state->scaling = RMX_ASPECT; + state->scaling = RMX_OFF; state->underscan_enable = false; state->underscan_hborder = 0; state->underscan_vborder = 0; + state->max_bpc = 8; __drm_atomic_helper_connector_reset(connector, &state->base); } @@ -3235,6 +3278,11 @@ amdgpu_dm_connector_atomic_duplicate_state(struct drm_connector *connector) new_state->freesync_capable = state->freesync_capable; new_state->abm_level = state->abm_level; + new_state->scaling = state->scaling; + new_state->underscan_enable = state->underscan_enable; + new_state->underscan_hborder = state->underscan_hborder; + new_state->underscan_vborder = state->underscan_vborder; + new_state->max_bpc = state->max_bpc; return &new_state->base; } @@ -3607,10 +3655,53 @@ static int dm_plane_atomic_check(struct drm_plane *plane, return -EINVAL; } +static int dm_plane_atomic_async_check(struct drm_plane *plane, + struct drm_plane_state *new_plane_state) +{ + struct drm_plane_state *old_plane_state = + drm_atomic_get_old_plane_state(new_plane_state->state, plane); + + /* Only support async updates on cursor planes. */ + if (plane->type != DRM_PLANE_TYPE_CURSOR) + return -EINVAL; + + /* + * DRM calls prepare_fb and cleanup_fb on new_plane_state for + * async commits so don't allow fb changes. + */ + if (old_plane_state->fb != new_plane_state->fb) + return -EINVAL; + + return 0; +} + +static void dm_plane_atomic_async_update(struct drm_plane *plane, + struct drm_plane_state *new_state) +{ + struct drm_plane_state *old_state = + drm_atomic_get_old_plane_state(new_state->state, plane); + + if (plane->state->fb != new_state->fb) + drm_atomic_set_fb_for_plane(plane->state, new_state->fb); + + plane->state->src_x = new_state->src_x; + plane->state->src_y = new_state->src_y; + plane->state->src_w = new_state->src_w; + plane->state->src_h = new_state->src_h; + plane->state->crtc_x = new_state->crtc_x; + plane->state->crtc_y = new_state->crtc_y; + plane->state->crtc_w = new_state->crtc_w; + plane->state->crtc_h = new_state->crtc_h; + + handle_cursor_update(plane, old_state); +} + static const struct drm_plane_helper_funcs dm_plane_helper_funcs = { .prepare_fb = dm_plane_helper_prepare_fb, .cleanup_fb = dm_plane_helper_cleanup_fb, .atomic_check = dm_plane_atomic_check, + .atomic_async_check = dm_plane_atomic_async_check, + .atomic_async_update = dm_plane_atomic_async_update }; /* @@ -3824,7 +3915,7 @@ amdgpu_dm_create_common_mode(struct drm_encoder *encoder, mode->hdisplay = hdisplay; mode->vdisplay = vdisplay; mode->type &= ~DRM_MODE_TYPE_PREFERRED; - strncpy(mode->name, name, DRM_DISPLAY_MODE_LEN); + strscpy(mode->name, name, DRM_DISPLAY_MODE_LEN); return mode; @@ -4299,6 +4390,7 @@ static int get_cursor_position(struct drm_plane *plane, struct drm_crtc *crtc, static void handle_cursor_update(struct drm_plane *plane, struct drm_plane_state *old_plane_state) { + struct amdgpu_device *adev = plane->dev->dev_private; struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(plane->state->fb); struct drm_crtc *crtc = afb ? plane->state->crtc : old_plane_state->crtc; struct dm_crtc_state *crtc_state = crtc ? to_dm_crtc_state(crtc->state) : NULL; @@ -4323,9 +4415,12 @@ static void handle_cursor_update(struct drm_plane *plane, if (!position.enable) { /* turn off cursor */ - if (crtc_state && crtc_state->stream) + if (crtc_state && crtc_state->stream) { + mutex_lock(&adev->dm.dc_lock); dc_stream_set_cursor_position(crtc_state->stream, &position); + mutex_unlock(&adev->dm.dc_lock); + } return; } @@ -4343,6 +4438,7 @@ static void handle_cursor_update(struct drm_plane *plane, attributes.pitch = attributes.width; if (crtc_state->stream) { + mutex_lock(&adev->dm.dc_lock); if (!dc_stream_set_cursor_attributes(crtc_state->stream, &attributes)) DRM_ERROR("DC failed to set cursor attributes\n"); @@ -4350,6 +4446,7 @@ static void handle_cursor_update(struct drm_plane *plane, if (!dc_stream_set_cursor_position(crtc_state->stream, &position)) DRM_ERROR("DC failed to set cursor position\n"); + mutex_unlock(&adev->dm.dc_lock); } } @@ -4388,9 +4485,11 @@ struct dc_stream_status *dc_state_get_stream_status( static void update_freesync_state_on_stream( struct amdgpu_display_manager *dm, struct dm_crtc_state *new_crtc_state, - struct dc_stream_state *new_stream) + struct dc_stream_state *new_stream, + struct dc_plane_state *surface, + u32 flip_timestamp_in_us) { - struct mod_vrr_params vrr = {0}; + struct mod_vrr_params vrr_params = new_crtc_state->vrr_params; struct dc_info_packet vrr_infopacket = {0}; struct mod_freesync_config config = new_crtc_state->freesync_config; @@ -4417,43 +4516,52 @@ static void update_freesync_state_on_stream( mod_freesync_build_vrr_params(dm->freesync_module, new_stream, - &config, &vrr); + &config, &vrr_params); + + if (surface) { + mod_freesync_handle_preflip( + dm->freesync_module, + surface, + new_stream, + flip_timestamp_in_us, + &vrr_params); + } mod_freesync_build_vrr_infopacket( dm->freesync_module, new_stream, - &vrr, - packet_type_vrr, - transfer_func_unknown, + &vrr_params, + PACKET_TYPE_VRR, + TRANSFER_FUNC_UNKNOWN, &vrr_infopacket); new_crtc_state->freesync_timing_changed = - (memcmp(&new_crtc_state->adjust, - &vrr.adjust, - sizeof(vrr.adjust)) != 0); + (memcmp(&new_crtc_state->vrr_params.adjust, + &vrr_params.adjust, + sizeof(vrr_params.adjust)) != 0); new_crtc_state->freesync_vrr_info_changed = (memcmp(&new_crtc_state->vrr_infopacket, &vrr_infopacket, sizeof(vrr_infopacket)) != 0); - new_crtc_state->adjust = vrr.adjust; + new_crtc_state->vrr_params = vrr_params; new_crtc_state->vrr_infopacket = vrr_infopacket; - new_stream->adjust = new_crtc_state->adjust; + new_stream->adjust = new_crtc_state->vrr_params.adjust; new_stream->vrr_infopacket = vrr_infopacket; if (new_crtc_state->freesync_vrr_info_changed) DRM_DEBUG_KMS("VRR packet update: crtc=%u enabled=%d state=%d", new_crtc_state->base.crtc->base.id, (int)new_crtc_state->base.vrr_enabled, - (int)vrr.state); + (int)vrr_params.state); if (new_crtc_state->freesync_timing_changed) DRM_DEBUG_KMS("VRR timing update: crtc=%u min=%u max=%u\n", new_crtc_state->base.crtc->base.id, - vrr.adjust.v_total_min, - vrr.adjust.v_total_max); + vrr_params.adjust.v_total_min, + vrr_params.adjust.v_total_max); } /* @@ -4467,6 +4575,7 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, struct dc_state *state) { unsigned long flags; + uint64_t timestamp_ns; uint32_t target_vblank; int r, vpos, hpos; struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc); @@ -4480,6 +4589,7 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, struct dc_stream_update stream_update = {0}; struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state); struct dc_stream_status *stream_status; + struct dc_plane_state *surface; /* Prepare wait for target vblank early - before the fence-waits */ @@ -4529,6 +4639,9 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, addr.address.grph.addr.high_part = upper_32_bits(afb->address); addr.flip_immediate = async_flip; + timestamp_ns = ktime_get_ns(); + addr.flip_timestamp_in_us = div_u64(timestamp_ns, 1000); + if (acrtc->base.state->event) prepare_flip_isr(acrtc); @@ -4542,8 +4655,10 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, return; } - surface_updates->surface = stream_status->plane_states[0]; - if (!surface_updates->surface) { + surface = stream_status->plane_states[0]; + surface_updates->surface = surface; + + if (!surface) { DRM_ERROR("No surface for CRTC: id=%d\n", acrtc->crtc_id); return; @@ -4554,7 +4669,9 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, update_freesync_state_on_stream( &adev->dm, acrtc_state, - acrtc_state->stream); + acrtc_state->stream, + surface, + addr.flip_timestamp_in_us); if (acrtc_state->freesync_timing_changed) stream_update.adjust = @@ -4565,6 +4682,16 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, &acrtc_state->stream->vrr_infopacket; } + /* Update surface timing information. */ + surface->time.time_elapsed_in_us[surface->time.index] = + addr.flip_timestamp_in_us - surface->time.prev_update_time_in_us; + surface->time.prev_update_time_in_us = addr.flip_timestamp_in_us; + surface->time.index++; + if (surface->time.index >= DC_PLANE_UPDATE_TIMES_MAX) + surface->time.index = 0; + + mutex_lock(&adev->dm.dc_lock); + dc_commit_updates_for_stream(adev->dm.dc, surface_updates, 1, @@ -4572,6 +4699,7 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, &stream_update, &surface_updates->surface, state); + mutex_unlock(&adev->dm.dc_lock); DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x \n", __func__, @@ -4586,6 +4714,7 @@ static void amdgpu_dm_do_flip(struct drm_crtc *crtc, * with a dc_plane_state and follow the atomic model a bit more closely here. */ static bool commit_planes_to_stream( + struct amdgpu_display_manager *dm, struct dc *dc, struct dc_plane_state **plane_states, uint8_t new_plane_count, @@ -4662,11 +4791,13 @@ static bool commit_planes_to_stream( updates[i].scaling_info = &scaling_info[i]; } + mutex_lock(&dm->dc_lock); dc_commit_updates_for_stream( dc, updates, new_plane_count, dc_stream, stream_update, plane_states, state); + mutex_unlock(&dm->dc_lock); kfree(flip_addr); kfree(plane_info); @@ -4772,7 +4903,8 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, dc_stream_attach->abm_level = acrtc_state->abm_level; - if (false == commit_planes_to_stream(dm->dc, + if (false == commit_planes_to_stream(dm, + dm->dc, plane_states_constructed, planes_count, acrtc_state, @@ -4942,7 +5074,9 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) if (dc_state) { dm_enable_per_frame_crtc_master_sync(dc_state); + mutex_lock(&dm->dc_lock); WARN_ON(!dc_commit_state(dm->dc, dc_state)); + mutex_unlock(&dm->dc_lock); } for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) { @@ -5004,6 +5138,7 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) /*TODO How it works with MPO ?*/ if (!commit_planes_to_stream( + dm, dm->dc, status->plane_states, status->plane_count, @@ -5248,6 +5383,7 @@ static void get_freesync_config_for_crtc( config.max_refresh_in_uhz = aconnector->max_vfreq * 1000000; config.vsif_supported = true; + config.btr = true; } new_crtc_state->freesync_config = config; @@ -5258,8 +5394,8 @@ static void reset_freesync_config_for_crtc( { new_crtc_state->vrr_supported = false; - memset(&new_crtc_state->adjust, 0, - sizeof(new_crtc_state->adjust)); + memset(&new_crtc_state->vrr_params, 0, + sizeof(new_crtc_state->vrr_params)); memset(&new_crtc_state->vrr_infopacket, 0, sizeof(new_crtc_state->vrr_infopacket)); } @@ -5797,7 +5933,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { if (!drm_atomic_crtc_needs_modeset(new_crtc_state) && !new_crtc_state->color_mgmt_changed && - !new_crtc_state->vrr_enabled) + old_crtc_state->vrr_enabled == new_crtc_state->vrr_enabled) continue; if (!new_crtc_state->enable) @@ -5896,6 +6032,13 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, ret = -EINVAL; goto fail; } + } else if (state->legacy_cursor_update) { + /* + * This is a fast cursor update coming from the plane update + * helper, check if it can be done asynchronously for better + * performance. + */ + state->async_update = !drm_atomic_helper_async_check(dev, state); } /* Must be success */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h index 4326dc256491..fbd161ddc3f4 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -135,6 +135,14 @@ struct amdgpu_display_manager { struct drm_modeset_lock atomic_obj_lock; /** + * @dc_lock: + * + * Guards access to DC functions that can issue register write + * sequences. + */ + struct mutex dc_lock; + + /** * @irq_handler_list_low_tab: * * Low priority IRQ handler table. @@ -260,7 +268,7 @@ struct dm_crtc_state { bool vrr_supported; struct mod_freesync_config freesync_config; - struct dc_crtc_timing_adjust adjust; + struct mod_vrr_params vrr_params; struct dc_info_packet vrr_infopacket; int abm_level; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index d02c32a1039c..1b0d209d8367 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -342,10 +342,9 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, master->connector_id); aconnector->mst_encoder = dm_dp_create_fake_mst_encoder(master); + drm_connector_attach_encoder(&aconnector->base, + &aconnector->mst_encoder->base); - /* - * TODO: understand why this one is needed - */ drm_object_attach_property( &connector->base, dev->mode_config.path_property, diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h new file mode 100644 index 000000000000..d898981684d5 --- /dev/null +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_trace.h @@ -0,0 +1,104 @@ +/* + * Copyright 2018 Advanced Micro Devices, 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: AMD + * + */ + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM amdgpu_dm + +#if !defined(_AMDGPU_DM_TRACE_H) || defined(TRACE_HEADER_MULTI_READ) +#define _AMDGPU_DM_TRACE_H_ + +#include <linux/tracepoint.h> + +TRACE_EVENT(amdgpu_dc_rreg, + TP_PROTO(unsigned long *read_count, uint32_t reg, uint32_t value), + TP_ARGS(read_count, reg, value), + TP_STRUCT__entry( + __field(uint32_t, reg) + __field(uint32_t, value) + ), + TP_fast_assign( + __entry->reg = reg; + __entry->value = value; + *read_count = *read_count + 1; + ), + TP_printk("reg=0x%08lx, value=0x%08lx", + (unsigned long)__entry->reg, + (unsigned long)__entry->value) +); + +TRACE_EVENT(amdgpu_dc_wreg, + TP_PROTO(unsigned long *write_count, uint32_t reg, uint32_t value), + TP_ARGS(write_count, reg, value), + TP_STRUCT__entry( + __field(uint32_t, reg) + __field(uint32_t, value) + ), + TP_fast_assign( + __entry->reg = reg; + __entry->value = value; + *write_count = *write_count + 1; + ), + TP_printk("reg=0x%08lx, value=0x%08lx", + (unsigned long)__entry->reg, + (unsigned long)__entry->value) +); + + +TRACE_EVENT(amdgpu_dc_performance, + TP_PROTO(unsigned long read_count, unsigned long write_count, + unsigned long *last_read, unsigned long *last_write, + const char *func, unsigned int line), + TP_ARGS(read_count, write_count, last_read, last_write, func, line), + TP_STRUCT__entry( + __field(uint32_t, reads) + __field(uint32_t, writes) + __field(uint32_t, read_delta) + __field(uint32_t, write_delta) + __string(func, func) + __field(uint32_t, line) + ), + TP_fast_assign( + __entry->reads = read_count; + __entry->writes = write_count; + __entry->read_delta = read_count - *last_read; + __entry->write_delta = write_count - *last_write; + __assign_str(func, func); + __entry->line = line; + *last_read = read_count; + *last_write = write_count; + ), + TP_printk("%s:%d reads=%08ld (%08ld total), writes=%08ld (%08ld total)", + __get_str(func), __entry->line, + (unsigned long)__entry->read_delta, + (unsigned long)__entry->reads, + (unsigned long)__entry->write_delta, + (unsigned long)__entry->writes) +); +#endif /* _AMDGPU_DM_TRACE_H_ */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#define TRACE_INCLUDE_FILE amdgpu_dm_trace +#include <trace/define_trace.h> diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c index 751bb614fc0e..c513ab6f3843 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c @@ -638,6 +638,7 @@ static enum bp_result get_ss_info_v4_1( { enum bp_result result = BP_RESULT_OK; struct atom_display_controller_info_v4_1 *disp_cntl_tbl = NULL; + struct atom_smu_info_v3_3 *smu_info = NULL; if (!ss_info) return BP_RESULT_BADINPUT; @@ -650,6 +651,7 @@ static enum bp_result get_ss_info_v4_1( if (!disp_cntl_tbl) return BP_RESULT_BADBIOSTABLE; + ss_info->type.STEP_AND_DELAY_INFO = false; ss_info->spread_percentage_divider = 1000; /* BIOS no longer uses target clock. Always enable for now */ @@ -688,6 +690,19 @@ static enum bp_result get_ss_info_v4_1( */ result = BP_RESULT_UNSUPPORTED; break; + case AS_SIGNAL_TYPE_XGMI: + smu_info = GET_IMAGE(struct atom_smu_info_v3_3, + DATA_TABLES(smu_info)); + if (!smu_info) + return BP_RESULT_BADBIOSTABLE; + + ss_info->spread_spectrum_percentage = + smu_info->waflclk_ss_percentage; + ss_info->spread_spectrum_range = + smu_info->gpuclk_ss_rate_10hz * 10; + if (smu_info->waflclk_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE) + ss_info->type.CENTER_MODE = true; + break; default: result = BP_RESULT_UNSUPPORTED; } diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c index 65b006ad372e..8196f3bb10c7 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c @@ -67,6 +67,7 @@ bool dal_bios_parser_init_cmd_tbl_helper2( return true; #endif case DCE_VERSION_12_0: + case DCE_VERSION_12_1: *h = dal_cmd_tbl_helper_dce112_get_table2(); return true; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index dba6b57830c7..5fd52094d459 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -62,6 +62,55 @@ const static char DC_BUILD_ID[] = "production-build"; +/** + * DOC: Overview + * + * DC is the OS-agnostic component of the amdgpu DC driver. + * + * DC maintains and validates a set of structs representing the state of the + * driver and writes that state to AMD hardware + * + * Main DC HW structs: + * + * struct dc - The central struct. One per driver. Created on driver load, + * destroyed on driver unload. + * + * struct dc_context - One per driver. + * Used as a backpointer by most other structs in dc. + * + * struct dc_link - One per connector (the physical DP, HDMI, miniDP, or eDP + * plugpoints). Created on driver load, destroyed on driver unload. + * + * struct dc_sink - One per display. Created on boot or hotplug. + * Destroyed on shutdown or hotunplug. A dc_link can have a local sink + * (the display directly attached). It may also have one or more remote + * sinks (in the Multi-Stream Transport case) + * + * struct resource_pool - One per driver. Represents the hw blocks not in the + * main pipeline. Not directly accessible by dm. + * + * Main dc state structs: + * + * These structs can be created and destroyed as needed. There is a full set of + * these structs in dc->current_state representing the currently programmed state. + * + * struct dc_state - The global DC state to track global state information, + * such as bandwidth values. + * + * struct dc_stream_state - Represents the hw configuration for the pipeline from + * a framebuffer to a display. Maps one-to-one with dc_sink. + * + * struct dc_plane_state - Represents a framebuffer. Each stream has at least one, + * and may have more in the Multi-Plane Overlay case. + * + * struct resource_context - Represents the programmable state of everything in + * the resource_pool. Not directly accessible by dm. + * + * struct pipe_ctx - A member of struct resource_context. Represents the + * internal hardware pipeline components. Each dc_plane_state has either + * one or two (in the pipe-split case). + */ + /******************************************************************************* * Private functions ******************************************************************************/ @@ -102,10 +151,6 @@ static bool create_links( return false; } - if (connectors_num == 0 && num_virtual_links == 0) { - dm_error("DC: Number of connectors is zero!\n"); - } - dm_output_to_console( "DC: %s: connectors_num: physical:%d, virtual:%d\n", __func__, @@ -175,6 +220,17 @@ failed_alloc: return false; } +static struct dc_perf_trace *dc_perf_trace_create(void) +{ + return kzalloc(sizeof(struct dc_perf_trace), GFP_KERNEL); +} + +static void dc_perf_trace_destroy(struct dc_perf_trace **perf_trace) +{ + kfree(*perf_trace); + *perf_trace = NULL; +} + /** ***************************************************************************** * Function: dc_stream_adjust_vmin_vmax @@ -240,7 +296,7 @@ bool dc_stream_get_crtc_position(struct dc *dc, } /** - * dc_stream_configure_crc: Configure CRC capture for the given stream. + * dc_stream_configure_crc() - Configure CRC capture for the given stream. * @dc: DC Object * @stream: The stream to configure CRC on. * @enable: Enable CRC if true, disable otherwise. @@ -292,7 +348,7 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream, } /** - * dc_stream_get_crc: Get CRC values for the given stream. + * dc_stream_get_crc() - Get CRC values for the given stream. * @dc: DC object * @stream: The DC stream state of the stream to get CRCs from. * @r_cr, g_y, b_cb: CRC values for the three channels are stored here. @@ -536,6 +592,8 @@ static void destruct(struct dc *dc) if (dc->ctx->created_bios) dal_bios_parser_destroy(&dc->ctx->dc_bios); + dc_perf_trace_destroy(&dc->ctx->perf_trace); + kfree(dc->ctx); dc->ctx = NULL; @@ -659,6 +717,12 @@ static bool construct(struct dc *dc, goto fail; } + dc_ctx->perf_trace = dc_perf_trace_create(); + if (!dc_ctx->perf_trace) { + ASSERT_CRITICAL(false); + goto fail; + } + /* Create GPIO service */ dc_ctx->gpio_service = dal_gpio_service_create( dc_version, @@ -1329,6 +1393,11 @@ static enum surface_update_type check_update_surfaces_for_stream( return overall_type; } +/** + * dc_check_update_surfaces_for_stream() - Determine update type (fast, med, or full) + * + * See :c:type:`enum surface_update_type <surface_update_type>` for explanation of update types + */ enum surface_update_type dc_check_update_surfaces_for_stream( struct dc *dc, struct dc_surface_update *updates, @@ -1398,7 +1467,8 @@ static void commit_planes_do_stream_update(struct dc *dc, if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) || stream_update->vrr_infopacket || - stream_update->vsc_infopacket) { + stream_update->vsc_infopacket || + stream_update->vsp_infopacket) { resource_build_info_frame(pipe_ctx); dc->hwss.update_info_frame(pipe_ctx); } @@ -1409,6 +1479,14 @@ static void commit_planes_do_stream_update(struct dc *dc, if (stream_update->output_csc_transform) dc_stream_program_csc_matrix(dc, stream); + if (stream_update->dither_option) { + resource_build_bit_depth_reduction_params(pipe_ctx->stream, + &pipe_ctx->stream->bit_depth_params); + pipe_ctx->stream_res.opp->funcs->opp_program_fmt(pipe_ctx->stream_res.opp, + &stream->bit_depth_params, + &stream->clamping); + } + /* Full fe update*/ if (update_type == UPDATE_TYPE_FAST) continue; @@ -1492,9 +1570,6 @@ static void commit_planes_for_stream(struct dc *dc, } } - if (update_type == UPDATE_TYPE_FULL) - context_timing_trace(dc, &context->res_ctx); - // Update Type FAST, Surface updates if (update_type == UPDATE_TYPE_FAST) { /* Lock the top pipe while updating plane addrs, since freesync requires @@ -1631,6 +1706,9 @@ enum dc_irq_source dc_interrupt_to_irq_source( return dal_irq_service_to_irq_source(dc->res_pool->irqs, src_id, ext_id); } +/** + * dc_interrupt_set() - Enable/disable an AMD hw interrupt source + */ bool dc_interrupt_set(struct dc *dc, enum dc_irq_source src, bool enable) { @@ -1724,6 +1802,11 @@ static bool link_add_remote_sink_helper(struct dc_link *dc_link, struct dc_sink return true; } +/** + * dc_link_add_remote_sink() - Create a sink and attach it to an existing link + * + * EDID length is in bytes + */ struct dc_sink *dc_link_add_remote_sink( struct dc_link *link, const uint8_t *edid, @@ -1782,6 +1865,12 @@ fail_add_sink: return NULL; } +/** + * dc_link_remove_remote_sink() - Remove a remote sink from a dc_link + * + * Note that this just removes the struct dc_sink - it doesn't + * program hardware or alter other members of dc_link + */ void dc_link_remove_remote_sink(struct dc_link *link, struct dc_sink *sink) { int i; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c index 948596a02392..b0265dbebd4c 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c @@ -198,6 +198,13 @@ static bool program_hpd_filter( return result; } +/** + * dc_link_detect_sink() - Determine if there is a sink connected + * + * @type: Returned connection type + * Does not detect downstream devices, such as MST sinks + * or display connected through active dongles + */ bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type) { uint32_t is_hpd_high = 0; @@ -208,6 +215,9 @@ bool dc_link_detect_sink(struct dc_link *link, enum dc_connection_type *type) return true; } + if (link->connector_signal == SIGNAL_TYPE_EDP) + link->dc->hwss.edp_wait_for_hpd_ready(link, true); + /* todo: may need to lock gpio access */ hpd_pin = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service); if (hpd_pin == NULL) @@ -324,15 +334,15 @@ static enum signal_type get_basic_signal_type( return SIGNAL_TYPE_NONE; } -/* - * @brief - * Check whether there is a dongle on DP connector +/** + * dc_link_is_dp_sink_present() - Check if there is a native DP + * or passive DP-HDMI dongle connected */ bool dc_link_is_dp_sink_present(struct dc_link *link) { enum gpio_result gpio_result; uint32_t clock_pin = 0; - + uint8_t retry = 0; struct ddc *ddc; enum connector_id connector_id = @@ -361,11 +371,22 @@ bool dc_link_is_dp_sink_present(struct dc_link *link) return present; } - /* Read GPIO: DP sink is present if both clock and data pins are zero */ - /* [anaumov] in DAL2, there was no check for GPIO failure */ - - gpio_result = dal_gpio_get_value(ddc->pin_clock, &clock_pin); - ASSERT(gpio_result == GPIO_RESULT_OK); + /* + * Read GPIO: DP sink is present if both clock and data pins are zero + * + * [W/A] plug-unplug DP cable, sometimes customer board has + * one short pulse on clk_pin(1V, < 1ms). DP will be config to HDMI/DVI + * then monitor can't br light up. Add retry 3 times + * But in real passive dongle, it need additional 3ms to detect + */ + do { + gpio_result = dal_gpio_get_value(ddc->pin_clock, &clock_pin); + ASSERT(gpio_result == GPIO_RESULT_OK); + if (clock_pin) + udelay(1000); + else + break; + } while (retry++ < 3); present = (gpio_result == GPIO_RESULT_OK) && !clock_pin; @@ -593,6 +614,14 @@ static bool is_same_edid(struct dc_edid *old_edid, struct dc_edid *new_edid) return (memcmp(old_edid->raw_edid, new_edid->raw_edid, new_edid->length) == 0); } +/** + * dc_link_detect() - Detect if a sink is attached to a given link + * + * link->local_sink is created or destroyed as needed. + * + * This does not create remote sinks but will trigger DM + * to start MST detection if a branch is detected. + */ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) { struct dc_sink_init_data sink_init_data = { 0 }; @@ -688,12 +717,26 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) if (memcmp(&link->dpcd_caps, &prev_dpcd_caps, sizeof(struct dpcd_caps))) same_dpcd = false; } - /* Active dongle downstream unplug */ + /* Active dongle plug in without display or downstream unplug*/ if (link->type == dc_connection_active_dongle && link->dpcd_caps.sink_count. bits.SINK_COUNT == 0) { - if (prev_sink != NULL) + if (prev_sink != NULL) { + /* Downstream unplug */ dc_sink_release(prev_sink); + } else { + /* Empty dongle plug in */ + for (i = 0; i < LINK_TRAINING_MAX_VERIFY_RETRY; i++) { + int fail_count = 0; + + dp_verify_link_cap(link, + &link->reported_link_cap, + &fail_count); + + if (fail_count == 0) + break; + } + } return true; } @@ -2147,8 +2190,7 @@ int dc_link_get_backlight_level(const struct dc_link *link) bool dc_link_set_backlight_level(const struct dc_link *link, uint32_t backlight_pwm_u16_16, - uint32_t frame_ramp, - const struct dc_stream_state *stream) + uint32_t frame_ramp) { struct dc *core_dc = link->ctx->dc; struct abm *abm = core_dc->res_pool->abm; @@ -2163,10 +2205,6 @@ bool dc_link_set_backlight_level(const struct dc_link *link, (abm->funcs->set_backlight_level_pwm == NULL)) return false; - if (stream) - ((struct dc_stream_state *)stream)->bl_pwm_level = - backlight_pwm_u16_16; - use_smooth_brightness = dmcu->funcs->is_dmcu_initialized(dmcu); DC_LOG_BACKLIGHT("New Backlight level: %d (0x%X)\n", @@ -2594,11 +2632,6 @@ void core_link_enable_stream( if (dc_is_dp_signal(pipe_ctx->stream->signal)) enable_stream_features(pipe_ctx); - - dc_link_set_backlight_level(pipe_ctx->stream->sink->link, - pipe_ctx->stream->bl_pwm_level, - 0, - pipe_ctx->stream); } } @@ -2607,11 +2640,11 @@ void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option) { struct dc *core_dc = pipe_ctx->stream->ctx->dc; + core_dc->hwss.blank_stream(pipe_ctx); + if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) deallocate_mst_payload(pipe_ctx); - core_dc->hwss.blank_stream(pipe_ctx); - core_dc->hwss.disable_stream(pipe_ctx, option); disable_link(pipe_ctx->stream->sink->link, pipe_ctx->stream->signal); diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c index 4d1f8ac069c1..0caacb60b02f 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c @@ -1089,6 +1089,121 @@ static struct dc_link_settings get_max_link_cap(struct dc_link *link) return max_link_cap; } +static enum dc_status read_hpd_rx_irq_data( + struct dc_link *link, + union hpd_irq_data *irq_data) +{ + static enum dc_status retval; + + /* The HW reads 16 bytes from 200h on HPD, + * but if we get an AUX_DEFER, the HW cannot retry + * and this causes the CTS tests 4.3.2.1 - 3.2.4 to + * fail, so we now explicitly read 6 bytes which is + * the req from the above mentioned test cases. + * + * For DP 1.4 we need to read those from 2002h range. + */ + if (link->dpcd_caps.dpcd_rev.raw < DPCD_REV_14) + retval = core_link_read_dpcd( + link, + DP_SINK_COUNT, + irq_data->raw, + sizeof(union hpd_irq_data)); + else { + /* Read 14 bytes in a single read and then copy only the required fields. + * This is more efficient than doing it in two separate AUX reads. */ + + uint8_t tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI + 1]; + + retval = core_link_read_dpcd( + link, + DP_SINK_COUNT_ESI, + tmp, + sizeof(tmp)); + + if (retval != DC_OK) + return retval; + + irq_data->bytes.sink_cnt.raw = tmp[DP_SINK_COUNT_ESI - DP_SINK_COUNT_ESI]; + irq_data->bytes.device_service_irq.raw = tmp[DP_DEVICE_SERVICE_IRQ_VECTOR_ESI0 - DP_SINK_COUNT_ESI]; + irq_data->bytes.lane01_status.raw = tmp[DP_LANE0_1_STATUS_ESI - DP_SINK_COUNT_ESI]; + irq_data->bytes.lane23_status.raw = tmp[DP_LANE2_3_STATUS_ESI - DP_SINK_COUNT_ESI]; + irq_data->bytes.lane_status_updated.raw = tmp[DP_LANE_ALIGN_STATUS_UPDATED_ESI - DP_SINK_COUNT_ESI]; + irq_data->bytes.sink_status.raw = tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI]; + } + + return retval; +} + +static bool hpd_rx_irq_check_link_loss_status( + struct dc_link *link, + union hpd_irq_data *hpd_irq_dpcd_data) +{ + uint8_t irq_reg_rx_power_state = 0; + enum dc_status dpcd_result = DC_ERROR_UNEXPECTED; + union lane_status lane_status; + uint32_t lane; + bool sink_status_changed; + bool return_code; + + sink_status_changed = false; + return_code = false; + + if (link->cur_link_settings.lane_count == 0) + return return_code; + + /*1. Check that Link Status changed, before re-training.*/ + + /*parse lane status*/ + for (lane = 0; lane < link->cur_link_settings.lane_count; lane++) { + /* check status of lanes 0,1 + * changed DpcdAddress_Lane01Status (0x202) + */ + lane_status.raw = get_nibble_at_index( + &hpd_irq_dpcd_data->bytes.lane01_status.raw, + lane); + + if (!lane_status.bits.CHANNEL_EQ_DONE_0 || + !lane_status.bits.CR_DONE_0 || + !lane_status.bits.SYMBOL_LOCKED_0) { + /* if one of the channel equalization, clock + * recovery or symbol lock is dropped + * consider it as (link has been + * dropped) dp sink status has changed + */ + sink_status_changed = true; + break; + } + } + + /* Check interlane align.*/ + if (sink_status_changed || + !hpd_irq_dpcd_data->bytes.lane_status_updated.bits.INTERLANE_ALIGN_DONE) { + + DC_LOG_HW_HPD_IRQ("%s: Link Status changed.\n", __func__); + + return_code = true; + + /*2. Check that we can handle interrupt: Not in FS DOS, + * Not in "Display Timeout" state, Link is trained. + */ + dpcd_result = core_link_read_dpcd(link, + DP_SET_POWER, + &irq_reg_rx_power_state, + sizeof(irq_reg_rx_power_state)); + + if (dpcd_result != DC_OK) { + DC_LOG_HW_HPD_IRQ("%s: DPCD read failed to obtain power state.\n", + __func__); + } else { + if (irq_reg_rx_power_state != DP_SET_POWER_D0) + return_code = false; + } + } + + return return_code; +} + bool dp_verify_link_cap( struct dc_link *link, struct dc_link_settings *known_limit_link_setting, @@ -1104,12 +1219,14 @@ bool dp_verify_link_cap( struct clock_source *dp_cs; enum clock_source_id dp_cs_id = CLOCK_SOURCE_ID_EXTERNAL; enum link_training_result status; + union hpd_irq_data irq_data; if (link->dc->debug.skip_detection_link_training) { link->verified_link_cap = *known_limit_link_setting; return true; } + memset(&irq_data, 0, sizeof(irq_data)); success = false; skip_link_training = false; @@ -1168,9 +1285,15 @@ bool dp_verify_link_cap( (*fail_count)++; } - if (success) + if (success) { link->verified_link_cap = *cur; - + udelay(1000); + if (read_hpd_rx_irq_data(link, &irq_data) == DC_OK) + if (hpd_rx_irq_check_link_loss_status( + link, + &irq_data)) + (*fail_count)++; + } /* always disable the link before trying another * setting or before returning we'll enable it later * based on the actual mode we're driving @@ -1572,122 +1695,6 @@ void decide_link_settings(struct dc_stream_state *stream, } /*************************Short Pulse IRQ***************************/ - -static bool hpd_rx_irq_check_link_loss_status( - struct dc_link *link, - union hpd_irq_data *hpd_irq_dpcd_data) -{ - uint8_t irq_reg_rx_power_state = 0; - enum dc_status dpcd_result = DC_ERROR_UNEXPECTED; - union lane_status lane_status; - uint32_t lane; - bool sink_status_changed; - bool return_code; - - sink_status_changed = false; - return_code = false; - - if (link->cur_link_settings.lane_count == 0) - return return_code; - - /*1. Check that Link Status changed, before re-training.*/ - - /*parse lane status*/ - for (lane = 0; lane < link->cur_link_settings.lane_count; lane++) { - /* check status of lanes 0,1 - * changed DpcdAddress_Lane01Status (0x202) - */ - lane_status.raw = get_nibble_at_index( - &hpd_irq_dpcd_data->bytes.lane01_status.raw, - lane); - - if (!lane_status.bits.CHANNEL_EQ_DONE_0 || - !lane_status.bits.CR_DONE_0 || - !lane_status.bits.SYMBOL_LOCKED_0) { - /* if one of the channel equalization, clock - * recovery or symbol lock is dropped - * consider it as (link has been - * dropped) dp sink status has changed - */ - sink_status_changed = true; - break; - } - } - - /* Check interlane align.*/ - if (sink_status_changed || - !hpd_irq_dpcd_data->bytes.lane_status_updated.bits.INTERLANE_ALIGN_DONE) { - - DC_LOG_HW_HPD_IRQ("%s: Link Status changed.\n", __func__); - - return_code = true; - - /*2. Check that we can handle interrupt: Not in FS DOS, - * Not in "Display Timeout" state, Link is trained. - */ - dpcd_result = core_link_read_dpcd(link, - DP_SET_POWER, - &irq_reg_rx_power_state, - sizeof(irq_reg_rx_power_state)); - - if (dpcd_result != DC_OK) { - DC_LOG_HW_HPD_IRQ("%s: DPCD read failed to obtain power state.\n", - __func__); - } else { - if (irq_reg_rx_power_state != DP_SET_POWER_D0) - return_code = false; - } - } - - return return_code; -} - -static enum dc_status read_hpd_rx_irq_data( - struct dc_link *link, - union hpd_irq_data *irq_data) -{ - static enum dc_status retval; - - /* The HW reads 16 bytes from 200h on HPD, - * but if we get an AUX_DEFER, the HW cannot retry - * and this causes the CTS tests 4.3.2.1 - 3.2.4 to - * fail, so we now explicitly read 6 bytes which is - * the req from the above mentioned test cases. - * - * For DP 1.4 we need to read those from 2002h range. - */ - if (link->dpcd_caps.dpcd_rev.raw < DPCD_REV_14) - retval = core_link_read_dpcd( - link, - DP_SINK_COUNT, - irq_data->raw, - sizeof(union hpd_irq_data)); - else { - /* Read 14 bytes in a single read and then copy only the required fields. - * This is more efficient than doing it in two separate AUX reads. */ - - uint8_t tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI + 1]; - - retval = core_link_read_dpcd( - link, - DP_SINK_COUNT_ESI, - tmp, - sizeof(tmp)); - - if (retval != DC_OK) - return retval; - - irq_data->bytes.sink_cnt.raw = tmp[DP_SINK_COUNT_ESI - DP_SINK_COUNT_ESI]; - irq_data->bytes.device_service_irq.raw = tmp[DP_DEVICE_SERVICE_IRQ_VECTOR_ESI0 - DP_SINK_COUNT_ESI]; - irq_data->bytes.lane01_status.raw = tmp[DP_LANE0_1_STATUS_ESI - DP_SINK_COUNT_ESI]; - irq_data->bytes.lane23_status.raw = tmp[DP_LANE2_3_STATUS_ESI - DP_SINK_COUNT_ESI]; - irq_data->bytes.lane_status_updated.raw = tmp[DP_LANE_ALIGN_STATUS_UPDATED_ESI - DP_SINK_COUNT_ESI]; - irq_data->bytes.sink_status.raw = tmp[DP_SINK_STATUS_ESI - DP_SINK_COUNT_ESI]; - } - - return retval; -} - static bool allow_hpd_rx_irq(const struct dc_link *link) { /* @@ -2196,7 +2203,7 @@ static void get_active_converter_info( } if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_11) { - uint8_t det_caps[4]; + uint8_t det_caps[16]; /* CTS 4.2.2.7 expects source to read Detailed Capabilities Info : 00080h-0008F.*/ union dwnstream_port_caps_byte0 *port_caps = (union dwnstream_port_caps_byte0 *)det_caps; core_link_read_dpcd(link, DP_DOWNSTREAM_PORT_0, @@ -2240,7 +2247,8 @@ static void get_active_converter_info( translate_dpcd_max_bpc( hdmi_color_caps.bits.MAX_BITS_PER_COLOR_COMPONENT); - link->dpcd_caps.dongle_caps.extendedCapValid = true; + if (link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk != 0) + link->dpcd_caps.dongle_caps.extendedCapValid = true; } break; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c index 82cd1d6e6e59..0065ec7d5330 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c @@ -96,6 +96,7 @@ void dp_enable_link_phy( link_settings, clock_source); } + link->cur_link_settings = *link_settings; dp_receiver_power_ctrl(link, true); } @@ -307,6 +308,7 @@ void dp_retrain_link_dp_test(struct dc_link *link, link->link_enc, link_setting, pipes[i].clock_source->id); + link->cur_link_settings = *link_setting; dp_receiver_power_ctrl(link, true); @@ -316,7 +318,6 @@ void dp_retrain_link_dp_test(struct dc_link *link, skip_video_pattern, LINK_TRAINING_ATTEMPTS); - link->cur_link_settings = *link_setting; link->dc->hwss.enable_stream(&pipes[i]); diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 0bb844a7b990..76137df74a53 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -83,7 +83,10 @@ enum dce_version resource_parse_asic_id(struct hw_asic_id asic_id) dc_version = DCE_VERSION_11_22; break; case FAMILY_AI: - dc_version = DCE_VERSION_12_0; + if (ASICREV_IS_VEGA20_P(asic_id.hw_internal_rev)) + dc_version = DCE_VERSION_12_1; + else + dc_version = DCE_VERSION_12_0; break; #if defined(CONFIG_DRM_AMD_DC_DCN1_0) case FAMILY_RV: @@ -136,6 +139,7 @@ struct resource_pool *dc_create_resource_pool( num_virtual_links, dc); break; case DCE_VERSION_12_0: + case DCE_VERSION_12_1: res_pool = dce120_create_resource_pool( num_virtual_links, dc); break; @@ -1447,6 +1451,14 @@ static bool are_stream_backends_same( return true; } +/** + * dc_is_stream_unchanged() - Compare two stream states for equivalence. + * + * Checks if there a difference between the two states + * that would require a mode change. + * + * Does not compare cursor position or attributes. + */ bool dc_is_stream_unchanged( struct dc_stream_state *old_stream, struct dc_stream_state *stream) { @@ -1457,6 +1469,9 @@ bool dc_is_stream_unchanged( return true; } +/** + * dc_is_stream_scaling_unchanged() - Compare scaling rectangles of two streams. + */ bool dc_is_stream_scaling_unchanged( struct dc_stream_state *old_stream, struct dc_stream_state *stream) { @@ -1616,6 +1631,9 @@ bool resource_is_stream_unchanged( return false; } +/** + * dc_add_stream_to_ctx() - Add a new dc_stream_state to a dc_state. + */ enum dc_status dc_add_stream_to_ctx( struct dc *dc, struct dc_state *new_ctx, @@ -1640,6 +1658,9 @@ enum dc_status dc_add_stream_to_ctx( return res; } +/** + * dc_remove_stream_from_ctx() - Remove a stream from a dc_state. + */ enum dc_status dc_remove_stream_from_ctx( struct dc *dc, struct dc_state *new_ctx, @@ -1860,6 +1881,12 @@ enum dc_status resource_map_pool_resources( return DC_ERROR_UNEXPECTED; } +/** + * dc_resource_state_copy_construct_current() - Creates a new dc_state from existing state + * Is a shallow copy. Increments refcounts on existing streams and planes. + * @dc: copy out of dc->current_state + * @dst_ctx: copy into this + */ void dc_resource_state_copy_construct_current( const struct dc *dc, struct dc_state *dst_ctx) @@ -1875,6 +1902,14 @@ void dc_resource_state_construct( dst_ctx->dccg = dc->res_pool->clk_mgr; } +/** + * dc_validate_global_state() - Determine if HW can support a given state + * Checks HW resource availability and bandwidth requirement. + * @dc: dc struct for this driver + * @new_ctx: state to be validated + * + * Return: DC_OK if the result can be programmed. Otherwise, an error code. + */ enum dc_status dc_validate_global_state( struct dc *dc, struct dc_state *new_ctx) @@ -2202,113 +2237,15 @@ static void set_vendor_info_packet( struct dc_info_packet *info_packet, struct dc_stream_state *stream) { - uint32_t length = 0; - bool hdmi_vic_mode = false; - uint8_t checksum = 0; - uint32_t i = 0; - enum dc_timing_3d_format format; - // Can be different depending on packet content /*todo*/ - // unsigned int length = pPathMode->dolbyVision ? 24 : 5; - - info_packet->valid = false; - - format = stream->timing.timing_3d_format; - if (stream->view_format == VIEW_3D_FORMAT_NONE) - format = TIMING_3D_FORMAT_NONE; - - /* Can be different depending on packet content */ - length = 5; - - if (stream->timing.hdmi_vic != 0 - && stream->timing.h_total >= 3840 - && stream->timing.v_total >= 2160) - hdmi_vic_mode = true; - - /* According to HDMI 1.4a CTS, VSIF should be sent - * for both 3D stereo and HDMI VIC modes. - * For all other modes, there is no VSIF sent. */ + /* SPD info packet for FreeSync */ - if (format == TIMING_3D_FORMAT_NONE && !hdmi_vic_mode) + /* Check if Freesync is supported. Return if false. If true, + * set the corresponding bit in the info packet + */ + if (!stream->vsp_infopacket.valid) return; - /* 24bit IEEE Registration identifier (0x000c03). LSB first. */ - info_packet->sb[1] = 0x03; - info_packet->sb[2] = 0x0C; - info_packet->sb[3] = 0x00; - - /*PB4: 5 lower bytes = 0 (reserved). 3 higher bits = HDMI_Video_Format. - * The value for HDMI_Video_Format are: - * 0x0 (0b000) - No additional HDMI video format is presented in this - * packet - * 0x1 (0b001) - Extended resolution format present. 1 byte of HDMI_VIC - * parameter follows - * 0x2 (0b010) - 3D format indication present. 3D_Structure and - * potentially 3D_Ext_Data follows - * 0x3..0x7 (0b011..0b111) - reserved for future use */ - if (format != TIMING_3D_FORMAT_NONE) - info_packet->sb[4] = (2 << 5); - else if (hdmi_vic_mode) - info_packet->sb[4] = (1 << 5); - - /* PB5: If PB4 claims 3D timing (HDMI_Video_Format = 0x2): - * 4 lower bites = 0 (reserved). 4 higher bits = 3D_Structure. - * The value for 3D_Structure are: - * 0x0 - Frame Packing - * 0x1 - Field Alternative - * 0x2 - Line Alternative - * 0x3 - Side-by-Side (full) - * 0x4 - L + depth - * 0x5 - L + depth + graphics + graphics-depth - * 0x6 - Top-and-Bottom - * 0x7 - Reserved for future use - * 0x8 - Side-by-Side (Half) - * 0x9..0xE - Reserved for future use - * 0xF - Not used */ - switch (format) { - case TIMING_3D_FORMAT_HW_FRAME_PACKING: - case TIMING_3D_FORMAT_SW_FRAME_PACKING: - info_packet->sb[5] = (0x0 << 4); - break; - - case TIMING_3D_FORMAT_SIDE_BY_SIDE: - case TIMING_3D_FORMAT_SBS_SW_PACKED: - info_packet->sb[5] = (0x8 << 4); - length = 6; - break; - - case TIMING_3D_FORMAT_TOP_AND_BOTTOM: - case TIMING_3D_FORMAT_TB_SW_PACKED: - info_packet->sb[5] = (0x6 << 4); - break; - - default: - break; - } - - /*PB5: If PB4 is set to 0x1 (extended resolution format) - * fill PB5 with the correct HDMI VIC code */ - if (hdmi_vic_mode) - info_packet->sb[5] = stream->timing.hdmi_vic; - - /* Header */ - info_packet->hb0 = HDMI_INFOFRAME_TYPE_VENDOR; /* VSIF packet type. */ - info_packet->hb1 = 0x01; /* Version */ - - /* 4 lower bits = Length, 4 higher bits = 0 (reserved) */ - info_packet->hb2 = (uint8_t) (length); - - /* Calculate checksum */ - checksum = 0; - checksum += info_packet->hb0; - checksum += info_packet->hb1; - checksum += info_packet->hb2; - - for (i = 1; i <= length; i++) - checksum += info_packet->sb[i]; - - info_packet->sb[0] = (uint8_t) (0x100 - checksum); - - info_packet->valid = true; + *info_packet = stream->vsp_infopacket; } static void set_spd_info_packet( @@ -2364,10 +2301,6 @@ void dc_resource_state_destruct(struct dc_state *context) } } -/* - * Copy src_ctx into dst_ctx and retain all surfaces and streams referenced - * by the src_ctx - */ void dc_resource_state_copy_construct( const struct dc_state *src_ctx, struct dc_state *dst_ctx) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c index 780838a05f44..66e5c4623a49 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c @@ -170,7 +170,7 @@ struct dc_stream_status *dc_stream_get_status( } /** - * Update the cursor attributes and set cursor surface address + * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address */ bool dc_stream_set_cursor_attributes( struct dc_stream_state *stream, diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index dea8bc39c688..4b5bbb13ce7f 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -39,7 +39,7 @@ #include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.06" +#define DC_VER "3.2.08" #define MAX_SURFACES 3 #define MAX_STREAMS 6 diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c index fcfd50b5dba0..4842d2378bbf 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c @@ -234,14 +234,14 @@ uint32_t generic_reg_wait(const struct dc_context *ctx, if (field_value == condition_value) { if (i * delay_between_poll_us > 1000 && !IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) - dm_output_to_console("REG_WAIT taking a while: %dms in %s line:%d\n", + DC_LOG_DC("REG_WAIT taking a while: %dms in %s line:%d\n", delay_between_poll_us * i / 1000, func_name, line); return reg_val; } } - dm_error("REG_WAIT timeout %dus * %d tries - %s line:%d\n", + DC_LOG_WARNING("REG_WAIT timeout %dus * %d tries - %s line:%d\n", delay_between_poll_us, time_out_num_tries, func_name, line); diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h index 7825e4b5e97c..e72fce4eca65 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h @@ -192,7 +192,6 @@ enum surface_pixel_format { /*swaped & float*/ SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F, /*grow graphics here if necessary */ - SURFACE_PIXEL_FORMAT_VIDEO_AYCrCb8888, SURFACE_PIXEL_FORMAT_VIDEO_BEGIN, SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr = SURFACE_PIXEL_FORMAT_VIDEO_BEGIN, @@ -200,6 +199,7 @@ enum surface_pixel_format { SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr, SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb, SURFACE_PIXEL_FORMAT_SUBSAMPLE_END, + SURFACE_PIXEL_FORMAT_VIDEO_AYCrCb8888, SURFACE_PIXEL_FORMAT_INVALID /*grow 444 video here if necessary */ @@ -358,15 +358,16 @@ union dc_tiling_info { } gfx8; struct { + enum swizzle_mode_values swizzle; unsigned int num_pipes; - unsigned int num_banks; + unsigned int max_compressed_frags; unsigned int pipe_interleave; + + unsigned int num_banks; unsigned int num_shader_engines; unsigned int num_rb_per_se; - unsigned int max_compressed_frags; bool shaderEnable; - enum swizzle_mode_values swizzle; bool meta_linear; bool rb_aligned; bool pipe_aligned; diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h index 8738f27a8708..b2243e0dad1f 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_link.h +++ b/drivers/gpu/drm/amd/display/dc/dc_link.h @@ -128,8 +128,10 @@ struct dc_link { const struct dc_link_status *dc_link_get_status(const struct dc_link *dc_link); -/* - * Return an enumerated dc_link. dc_link order is constant and determined at +/** + * dc_get_link_at_index() - Return an enumerated dc_link. + * + * dc_link order is constant and determined at * boot time. They cannot be created or destroyed. * Use dc_get_caps() to get number of links. */ @@ -144,8 +146,7 @@ static inline struct dc_link *dc_get_link_at_index(struct dc *dc, uint32_t link_ */ bool dc_link_set_backlight_level(const struct dc_link *dc_link, uint32_t backlight_pwm_u16_16, - uint32_t frame_ramp, - const struct dc_stream_state *stream); + uint32_t frame_ramp); int dc_link_get_backlight_level(const struct dc_link *dc_link); diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index 771d9f17e26e..d70c9e1cda3d 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -56,6 +56,7 @@ struct dc_stream_state { struct dc_crtc_timing_adjust adjust; struct dc_info_packet vrr_infopacket; struct dc_info_packet vsc_infopacket; + struct dc_info_packet vsp_infopacket; struct rect src; /* composition area */ struct rect dst; /* stream addressable area */ @@ -90,7 +91,6 @@ struct dc_stream_state { /* DMCU info */ unsigned int abm_level; - unsigned int bl_pwm_level; /* from core_stream struct */ struct dc_context *ctx; @@ -129,11 +129,13 @@ struct dc_stream_update { struct dc_crtc_timing_adjust *adjust; struct dc_info_packet *vrr_infopacket; struct dc_info_packet *vsc_infopacket; + struct dc_info_packet *vsp_infopacket; bool *dpms_off; struct colorspace_transform *gamut_remap; enum dc_color_space *output_color_space; + enum dc_dither_option *dither_option; struct dc_csc_transform *output_csc_transform; diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 6e12d640d020..0b20ae23f169 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -73,10 +73,18 @@ struct hw_asic_id { void *atombios_base_address; }; +struct dc_perf_trace { + unsigned long read_count; + unsigned long write_count; + unsigned long last_entry_read; + unsigned long last_entry_write; +}; + struct dc_context { struct dc *dc; void *driver_context; /* e.g. amdgpu_device */ + struct dc_perf_trace *perf_trace; void *cgs_device; enum dce_environment dce_environment; @@ -191,7 +199,6 @@ union display_content_support { }; struct dc_panel_patch { - unsigned int disconnect_delay; unsigned int dppowerup_delay; unsigned int extra_t12_ms; }; diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c index bd22f51813bf..afd287f08bc9 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c @@ -676,6 +676,11 @@ static void dce112_update_clocks(struct clk_mgr *clk_mgr, { struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr); struct dm_pp_power_level_change_request level_change_req; + int unpatched_disp_clk = context->bw.dce.dispclk_khz; + + /*TODO: W/A for dal3 linux, investigate why this works */ + if (!clk_mgr_dce->dfs_bypass_active) + context->bw.dce.dispclk_khz = context->bw.dce.dispclk_khz * 115 / 100; level_change_req.power_level = dce_get_required_clocks_state(clk_mgr, context); /* get max clock state from PPLIB */ @@ -690,6 +695,8 @@ static void dce112_update_clocks(struct clk_mgr *clk_mgr, clk_mgr->clks.dispclk_khz = context->bw.dce.dispclk_khz; } dce11_pplib_apply_display_requirements(clk_mgr->ctx->dc, context); + + context->bw.dce.dispclk_khz = unpatched_disp_clk; } static void dce12_update_clocks(struct clk_mgr *clk_mgr, diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c index bc50a8e25f4f..87771676acac 100644 --- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce100/dce100_hw_sequencer.c @@ -117,6 +117,18 @@ void dce100_prepare_bandwidth( false); } +void dce100_optimize_bandwidth( + struct dc *dc, + struct dc_state *context) +{ + dce110_set_safe_displaymarks(&context->res_ctx, dc->res_pool); + + dc->res_pool->clk_mgr->funcs->update_clocks( + dc->res_pool->clk_mgr, + context, + true); +} + /**************************************************************************/ void dce100_hw_sequencer_construct(struct dc *dc) @@ -125,6 +137,6 @@ void dce100_hw_sequencer_construct(struct dc *dc) dc->hwss.enable_display_power_gating = dce100_enable_display_power_gating; dc->hwss.prepare_bandwidth = dce100_prepare_bandwidth; - dc->hwss.optimize_bandwidth = dce100_prepare_bandwidth; + dc->hwss.optimize_bandwidth = dce100_optimize_bandwidth; } diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c index 1f7f25013217..52d50e24a995 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c @@ -64,65 +64,37 @@ static const struct dce110_compressor_reg_offsets reg_offsets[] = { static const uint32_t dce11_one_lpt_channel_max_resolution = 2560 * 1600; -enum fbc_idle_force { - /* Bit 0 - Display registers updated */ - FBC_IDLE_FORCE_DISPLAY_REGISTER_UPDATE = 0x00000001, - - /* Bit 2 - FBC_GRPH_COMP_EN register updated */ - FBC_IDLE_FORCE_GRPH_COMP_EN = 0x00000002, - /* Bit 3 - FBC_SRC_SEL register updated */ - FBC_IDLE_FORCE_SRC_SEL_CHANGE = 0x00000004, - /* Bit 4 - FBC_MIN_COMPRESSION register updated */ - FBC_IDLE_FORCE_MIN_COMPRESSION_CHANGE = 0x00000008, - /* Bit 5 - FBC_ALPHA_COMP_EN register updated */ - FBC_IDLE_FORCE_ALPHA_COMP_EN = 0x00000010, - /* Bit 6 - FBC_ZERO_ALPHA_CHUNK_SKIP_EN register updated */ - FBC_IDLE_FORCE_ZERO_ALPHA_CHUNK_SKIP_EN = 0x00000020, - /* Bit 7 - FBC_FORCE_COPY_TO_COMP_BUF register updated */ - FBC_IDLE_FORCE_FORCE_COPY_TO_COMP_BUF = 0x00000040, - - /* Bit 24 - Memory write to region 0 defined by MC registers. */ - FBC_IDLE_FORCE_MEMORY_WRITE_TO_REGION0 = 0x01000000, - /* Bit 25 - Memory write to region 1 defined by MC registers */ - FBC_IDLE_FORCE_MEMORY_WRITE_TO_REGION1 = 0x02000000, - /* Bit 26 - Memory write to region 2 defined by MC registers */ - FBC_IDLE_FORCE_MEMORY_WRITE_TO_REGION2 = 0x04000000, - /* Bit 27 - Memory write to region 3 defined by MC registers. */ - FBC_IDLE_FORCE_MEMORY_WRITE_TO_REGION3 = 0x08000000, - - /* Bit 28 - Memory write from any client other than MCIF */ - FBC_IDLE_FORCE_MEMORY_WRITE_OTHER_THAN_MCIF = 0x10000000, - /* Bit 29 - CG statics screen signal is inactive */ - FBC_IDLE_FORCE_CG_STATIC_SCREEN_IS_INACTIVE = 0x20000000, -}; - - static uint32_t align_to_chunks_number_per_line(uint32_t pixels) { return 256 * ((pixels + 255) / 256); } -static void reset_lb_on_vblank(struct dc_context *ctx) +static void reset_lb_on_vblank(struct compressor *compressor, uint32_t crtc_inst) { - uint32_t value, frame_count; + uint32_t value; + uint32_t frame_count; + uint32_t status_pos; uint32_t retry = 0; - uint32_t status_pos = - dm_read_reg(ctx, mmCRTC_STATUS_POSITION); + struct dce110_compressor *cp110 = TO_DCE110_COMPRESSOR(compressor); + + cp110->offsets = reg_offsets[crtc_inst]; + + status_pos = dm_read_reg(compressor->ctx, DCP_REG(mmCRTC_STATUS_POSITION)); /* Only if CRTC is enabled and counter is moving we wait for one frame. */ - if (status_pos != dm_read_reg(ctx, mmCRTC_STATUS_POSITION)) { + if (status_pos != dm_read_reg(compressor->ctx, DCP_REG(mmCRTC_STATUS_POSITION))) { /* Resetting LB on VBlank */ - value = dm_read_reg(ctx, mmLB_SYNC_RESET_SEL); + value = dm_read_reg(compressor->ctx, DCP_REG(mmLB_SYNC_RESET_SEL)); set_reg_field_value(value, 3, LB_SYNC_RESET_SEL, LB_SYNC_RESET_SEL); set_reg_field_value(value, 1, LB_SYNC_RESET_SEL, LB_SYNC_RESET_SEL2); - dm_write_reg(ctx, mmLB_SYNC_RESET_SEL, value); + dm_write_reg(compressor->ctx, DCP_REG(mmLB_SYNC_RESET_SEL), value); - frame_count = dm_read_reg(ctx, mmCRTC_STATUS_FRAME_COUNT); + frame_count = dm_read_reg(compressor->ctx, DCP_REG(mmCRTC_STATUS_FRAME_COUNT)); for (retry = 10000; retry > 0; retry--) { - if (frame_count != dm_read_reg(ctx, mmCRTC_STATUS_FRAME_COUNT)) + if (frame_count != dm_read_reg(compressor->ctx, DCP_REG(mmCRTC_STATUS_FRAME_COUNT))) break; udelay(10); } @@ -130,13 +102,11 @@ static void reset_lb_on_vblank(struct dc_context *ctx) dm_error("Frame count did not increase for 100ms.\n"); /* Resetting LB on VBlank */ - value = dm_read_reg(ctx, mmLB_SYNC_RESET_SEL); + value = dm_read_reg(compressor->ctx, DCP_REG(mmLB_SYNC_RESET_SEL)); set_reg_field_value(value, 2, LB_SYNC_RESET_SEL, LB_SYNC_RESET_SEL); set_reg_field_value(value, 0, LB_SYNC_RESET_SEL, LB_SYNC_RESET_SEL2); - dm_write_reg(ctx, mmLB_SYNC_RESET_SEL, value); - + dm_write_reg(compressor->ctx, DCP_REG(mmLB_SYNC_RESET_SEL), value); } - } static void wait_for_fbc_state_changed( @@ -226,10 +196,10 @@ void dce110_compressor_enable_fbc( uint32_t addr; uint32_t value, misc_value; - addr = mmFBC_CNTL; value = dm_read_reg(compressor->ctx, addr); set_reg_field_value(value, 1, FBC_CNTL, FBC_GRPH_COMP_EN); + /* params->inst is valid HW CRTC instance start from 0 */ set_reg_field_value( value, params->inst, @@ -238,8 +208,10 @@ void dce110_compressor_enable_fbc( /* Keep track of enum controller_id FBC is attached to */ compressor->is_enabled = true; - compressor->attached_inst = params->inst; - cp110->offsets = reg_offsets[params->inst]; + /* attached_inst is SW CRTC instance start from 1 + * 0 = CONTROLLER_ID_UNDEFINED means not attached crtc + */ + compressor->attached_inst = params->inst + CONTROLLER_ID_D0; /* Toggle it as there is bug in HW */ set_reg_field_value(value, 0, FBC_CNTL, FBC_GRPH_COMP_EN); @@ -268,9 +240,10 @@ void dce110_compressor_enable_fbc( void dce110_compressor_disable_fbc(struct compressor *compressor) { struct dce110_compressor *cp110 = TO_DCE110_COMPRESSOR(compressor); + uint32_t crtc_inst = 0; if (compressor->options.bits.FBC_SUPPORT) { - if (dce110_compressor_is_fbc_enabled_in_hw(compressor, NULL)) { + if (dce110_compressor_is_fbc_enabled_in_hw(compressor, &crtc_inst)) { uint32_t reg_data; /* Turn off compression */ reg_data = dm_read_reg(compressor->ctx, mmFBC_CNTL); @@ -284,8 +257,10 @@ void dce110_compressor_disable_fbc(struct compressor *compressor) wait_for_fbc_state_changed(cp110, false); } - /* Sync line buffer - dce100/110 only*/ - reset_lb_on_vblank(compressor->ctx); + /* Sync line buffer which fbc was attached to dce100/110 only */ + if (crtc_inst > CONTROLLER_ID_UNDEFINED && crtc_inst < CONTROLLER_ID_D3) + reset_lb_on_vblank(compressor, + crtc_inst - CONTROLLER_ID_D0); } } @@ -328,6 +303,8 @@ void dce110_compressor_program_compressed_surface_address_and_pitch( uint32_t compressed_surf_address_low_part = compressor->compr_surface_address.addr.low_part; + cp110->offsets = reg_offsets[params->inst]; + /* Clear content first. */ dm_write_reg( compressor->ctx, @@ -410,13 +387,7 @@ void dce110_compressor_set_fbc_invalidation_triggers( value = dm_read_reg(compressor->ctx, addr); set_reg_field_value( value, - fbc_trigger | - FBC_IDLE_FORCE_GRPH_COMP_EN | - FBC_IDLE_FORCE_SRC_SEL_CHANGE | - FBC_IDLE_FORCE_MIN_COMPRESSION_CHANGE | - FBC_IDLE_FORCE_ALPHA_COMP_EN | - FBC_IDLE_FORCE_ZERO_ALPHA_CHUNK_SKIP_EN | - FBC_IDLE_FORCE_FORCE_COPY_TO_COMP_BUF, + fbc_trigger, FBC_IDLE_FORCE_CLEAR_MASK, FBC_IDLE_FORCE_CLEAR_MASK); dm_write_reg(compressor->ctx, addr, value); @@ -549,7 +520,7 @@ void dce110_compressor_construct(struct dce110_compressor *compressor, compressor->base.channel_interleave_size = 0; compressor->base.dram_channels_num = 0; compressor->base.lpt_channels_num = 0; - compressor->base.attached_inst = 0; + compressor->base.attached_inst = CONTROLLER_ID_UNDEFINED; compressor->base.is_enabled = false; compressor->base.funcs = &dce110_compressor_funcs; diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c index 2f062bacd78a..8f09b8625c5d 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c @@ -1000,7 +1000,7 @@ void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx) pipe_ctx->stream_res.audio->funcs->az_enable(pipe_ctx->stream_res.audio); - if (num_audio == 1 && pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) + if (num_audio >= 1 && pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) /*this is the first audio. apply the PME w/a in order to wake AZ from D3*/ pp_smu->set_pme_wa_enable(&pp_smu->pp_smu); /* un-mute audio */ @@ -1017,6 +1017,8 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control( pipe_ctx->stream_res.stream_enc, true); if (pipe_ctx->stream_res.audio) { + struct pp_smu_funcs_rv *pp_smu = dc->res_pool->pp_smu; + if (option != KEEP_ACQUIRED_RESOURCE || !dc->debug.az_endpoint_mute_only) { /*only disalbe az_endpoint if power down or free*/ @@ -1036,6 +1038,9 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, pipe_ctx->stream_res.audio, false); pipe_ctx->stream_res.audio = NULL; } + if (pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) + /*this is the first audio. apply the PME w/a in order to wake AZ from D3*/ + pp_smu->set_pme_wa_enable(&pp_smu->pp_smu); /* TODO: notify audio driver for if audio modes list changed * add audio mode list change flag */ @@ -1267,10 +1272,19 @@ static void program_scaler(const struct dc *dc, pipe_ctx->plane_res.scl_data.lb_params.depth, &pipe_ctx->stream->bit_depth_params); - if (pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color) + if (pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color) { + /* + * The way 420 is packed, 2 channels carry Y component, 1 channel + * alternate between Cb and Cr, so both channels need the pixel + * value for Y + */ + if (pipe_ctx->stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) + color.color_r_cr = color.color_g_y; + pipe_ctx->stream_res.tg->funcs->set_overscan_blank_color( pipe_ctx->stream_res.tg, &color); + } pipe_ctx->plane_res.xfm->funcs->transform_set_scaler(pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data); @@ -1766,12 +1780,13 @@ static void set_static_screen_control(struct pipe_ctx **pipe_ctx, * Check if FBC can be enabled */ static bool should_enable_fbc(struct dc *dc, - struct dc_state *context, - uint32_t *pipe_idx) + struct dc_state *context, + uint32_t *pipe_idx) { uint32_t i; struct pipe_ctx *pipe_ctx = NULL; struct resource_context *res_ctx = &context->res_ctx; + unsigned int underlay_idx = dc->res_pool->underlay_pipe_index; ASSERT(dc->fbc_compressor); @@ -1786,14 +1801,28 @@ static bool should_enable_fbc(struct dc *dc, for (i = 0; i < dc->res_pool->pipe_count; i++) { if (res_ctx->pipe_ctx[i].stream) { + pipe_ctx = &res_ctx->pipe_ctx[i]; - *pipe_idx = i; - break; + + if (!pipe_ctx) + continue; + + /* fbc not applicable on underlay pipe */ + if (pipe_ctx->pipe_idx != underlay_idx) { + *pipe_idx = i; + break; + } } } - /* Pipe context should be found */ - ASSERT(pipe_ctx); + if (i == dc->res_pool->pipe_count) + return false; + + if (!pipe_ctx->stream->sink) + return false; + + if (!pipe_ctx->stream->sink->link) + return false; /* Only supports eDP */ if (pipe_ctx->stream->sink->link->connector_signal != SIGNAL_TYPE_EDP) @@ -1817,8 +1846,9 @@ static bool should_enable_fbc(struct dc *dc, /* * Enable FBC */ -static void enable_fbc(struct dc *dc, - struct dc_state *context) +static void enable_fbc( + struct dc *dc, + struct dc_state *context) { uint32_t pipe_idx = 0; @@ -1828,10 +1858,9 @@ static void enable_fbc(struct dc *dc, struct compressor *compr = dc->fbc_compressor; struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[pipe_idx]; - params.source_view_width = pipe_ctx->stream->timing.h_addressable; params.source_view_height = pipe_ctx->stream->timing.v_addressable; - + params.inst = pipe_ctx->stream_res.tg->inst; compr->compr_surface_address.quad_part = dc->ctx->fbc_gpu_addr; compr->funcs->surface_address_and_pitch(compr, ¶ms); @@ -2046,10 +2075,10 @@ enum dc_status dce110_apply_ctx_to_hw( return status; } - dcb->funcs->set_scratch_critical_state(dcb, false); - if (dc->fbc_compressor) - enable_fbc(dc, context); + enable_fbc(dc, dc->current_state); + + dcb->funcs->set_scratch_critical_state(dcb, false); return DC_OK; } @@ -2408,7 +2437,6 @@ static void dce110_program_front_end_for_pipe( struct dc_plane_state *plane_state = pipe_ctx->plane_state; struct xfm_grph_csc_adjustment adjust; struct out_csc_color_matrix tbl_entry; - unsigned int underlay_idx = dc->res_pool->underlay_pipe_index; unsigned int i; DC_LOGGER_INIT(); memset(&tbl_entry, 0, sizeof(tbl_entry)); @@ -2449,15 +2477,6 @@ static void dce110_program_front_end_for_pipe( program_scaler(dc, pipe_ctx); - /* fbc not applicable on Underlay pipe */ - if (dc->fbc_compressor && old_pipe->stream && - pipe_ctx->pipe_idx != underlay_idx) { - if (plane_state->tiling_info.gfx8.array_mode == DC_ARRAY_LINEAR_GENERAL) - dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor); - else - enable_fbc(dc, dc->current_state); - } - mi->funcs->mem_input_program_surface_config( mi, plane_state->format, @@ -2534,6 +2553,9 @@ static void dce110_apply_ctx_for_surface( if (num_planes == 0) return; + if (dc->fbc_compressor) + dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor); + for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; @@ -2576,6 +2598,9 @@ static void dce110_apply_ctx_for_surface( (pipe_ctx->plane_state || old_pipe_ctx->plane_state)) dc->hwss.pipe_control_lock(dc, pipe_ctx, false); } + + if (dc->fbc_compressor) + enable_fbc(dc, dc->current_state); } static void dce110_power_down_fe(struct dc *dc, struct pipe_ctx *pipe_ctx) diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c index f9d7d2c26cc2..54abedbf1b43 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c @@ -328,12 +328,10 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, *smu_req_cur = smu_req; } - static const struct clk_mgr_funcs dcn1_funcs = { .get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz, .update_clocks = dcn1_update_clocks }; - struct clk_mgr *dcn1_clk_mgr_create(struct dc_context *ctx) { struct dc_debug_options *debug = &ctx->dc->debug; @@ -373,3 +371,5 @@ struct clk_mgr *dcn1_clk_mgr_create(struct dc_context *ctx) return &clk_mgr_dce->base; } + + diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.h index 9dbaf6578006..a995eda443a3 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.h @@ -28,6 +28,12 @@ #include "../dce/dce_clk_mgr.h" +struct clk_bypass { + uint32_t dcfclk_bypass; + uint32_t dispclk_pypass; + uint32_t dprefclk_bypass; +}; + void dcn1_pplib_apply_display_requirements( struct dc *dc, struct dc_state *context); diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c index 3eea44092a04..7469333a2c8a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c @@ -324,7 +324,7 @@ bool cm_helper_translate_curve_to_hw_format( if (output_tf == NULL || lut_params == NULL || output_tf->type == TF_TYPE_BYPASS) return false; - PERF_TRACE(); + PERF_TRACE_CTX(output_tf->ctx); corner_points = lut_params->corner_points; rgb_resulted = lut_params->rgb_resulted; @@ -513,7 +513,7 @@ bool cm_helper_translate_curve_to_degamma_hw_format( if (output_tf == NULL || lut_params == NULL || output_tf->type == TF_TYPE_BYPASS) return false; - PERF_TRACE(); + PERF_TRACE_CTX(output_tf->ctx); corner_points = lut_params->corner_points; rgb_resulted = lut_params->rgb_resulted; diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c index dcb3c5530236..cd1ebe57ed59 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c @@ -463,7 +463,7 @@ void dpp1_set_cursor_position( if (src_y_offset >= (int)param->viewport.height) cur_en = 0; /* not visible beyond bottom edge*/ - if (src_y_offset < 0) + if (src_y_offset + (int)height <= 0) cur_en = 0; /* not visible beyond top edge*/ REG_UPDATE(CURSOR0_CONTROL, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c index 345af015d061..d1acd7165bc8 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c @@ -1140,7 +1140,7 @@ void hubp1_cursor_set_position( if (src_y_offset >= (int)param->viewport.height) cur_en = 0; /* not visible beyond bottom edge*/ - if (src_y_offset < 0) //+ (int)hubp->curs_attr.height + if (src_y_offset + (int)hubp->curs_attr.height <= 0) cur_en = 0; /* not visible beyond top edge*/ if (cur_en && REG_READ(CURSOR_SURFACE_ADDRESS) == 0) diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c index 0bd33a713836..58a12ddf12f3 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c @@ -2159,6 +2159,15 @@ static void dcn10_blank_pixel_data( color_space = stream->output_color_space; color_space_to_black_color(dc, color_space, &black_color); + /* + * The way 420 is packed, 2 channels carry Y component, 1 channel + * alternate between Cb and Cr, so both channels need the pixel + * value for Y + */ + if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) + black_color.color_r_cr = black_color.color_g_y; + + if (stream_res->tg->funcs->set_blank_color) stream_res->tg->funcs->set_blank_color( stream_res->tg, @@ -2346,27 +2355,22 @@ static void dcn10_apply_ctx_for_surface( top_pipe_to_program->plane_state->update_flags.bits.full_update) for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - + tg = pipe_ctx->stream_res.tg; /* Skip inactive pipes and ones already updated */ - if (!pipe_ctx->stream || pipe_ctx->stream == stream) + if (!pipe_ctx->stream || pipe_ctx->stream == stream + || !pipe_ctx->plane_state + || !tg->funcs->is_tg_enabled(tg)) continue; - pipe_ctx->stream_res.tg->funcs->lock(pipe_ctx->stream_res.tg); + tg->funcs->lock(tg); pipe_ctx->plane_res.hubp->funcs->hubp_setup_interdependent( pipe_ctx->plane_res.hubp, &pipe_ctx->dlg_regs, &pipe_ctx->ttu_regs); - } - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - - if (!pipe_ctx->stream || pipe_ctx->stream == stream) - continue; - - dcn10_pipe_control_lock(dc, pipe_ctx, false); - } + tg->funcs->unlock(tg); + } if (num_planes == 0) false_optc_underflow_wa(dc, stream, tg); diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c index 211bb240a720..cd469014baa3 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c @@ -44,6 +44,7 @@ #include "dcn10_hubp.h" #include "dcn10_hubbub.h" #include "dcn10_cm_common.h" +#include "dcn10_clk_mgr.h" static unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, ...) { @@ -463,19 +464,22 @@ static unsigned int dcn10_get_otg_states(struct dc *dc, char *pBuf, unsigned int static unsigned int dcn10_get_clock_states(struct dc *dc, char *pBuf, unsigned int bufSize) { unsigned int chars_printed = 0; + unsigned int remaining_buffer = bufSize; - chars_printed = snprintf_count(pBuf, bufSize, "dcfclk_khz,dcfclk_deep_sleep_khz,dispclk_khz," - "dppclk_khz,max_supported_dppclk_khz,fclk_khz,socclk_khz\n" - "%d,%d,%d,%d,%d,%d,%d\n", + chars_printed = snprintf_count(pBuf, bufSize, "dcfclk,dcfclk_deep_sleep,dispclk," + "dppclk,fclk,socclk\n" + "%d,%d,%d,%d,%d,%d\n", dc->current_state->bw.dcn.clk.dcfclk_khz, dc->current_state->bw.dcn.clk.dcfclk_deep_sleep_khz, dc->current_state->bw.dcn.clk.dispclk_khz, dc->current_state->bw.dcn.clk.dppclk_khz, - dc->current_state->bw.dcn.clk.max_supported_dppclk_khz, dc->current_state->bw.dcn.clk.fclk_khz, dc->current_state->bw.dcn.clk.socclk_khz); - return chars_printed; + remaining_buffer -= chars_printed; + pBuf += chars_printed; + + return bufSize - remaining_buffer; } static void dcn10_clear_otpc_underflow(struct dc *dc) @@ -538,16 +542,16 @@ void dcn10_get_hw_state(struct dc *dc, char *pBuf, unsigned int bufSize, unsigne * Bit 0 - 15: Hardware block mask * Bit 15: 1 = Invariant Only, 0 = All */ - const unsigned int DC_HW_STATE_MASK_HUBBUB = 0x1; - const unsigned int DC_HW_STATE_MASK_HUBP = 0x2; - const unsigned int DC_HW_STATE_MASK_RQ = 0x4; - const unsigned int DC_HW_STATE_MASK_DLG = 0x8; - const unsigned int DC_HW_STATE_MASK_TTU = 0x10; - const unsigned int DC_HW_STATE_MASK_CM = 0x20; - const unsigned int DC_HW_STATE_MASK_MPCC = 0x40; - const unsigned int DC_HW_STATE_MASK_OTG = 0x80; - const unsigned int DC_HW_STATE_MASK_CLOCKS = 0x100; - const unsigned int DC_HW_STATE_INVAR_ONLY = 0x8000; + const unsigned int DC_HW_STATE_MASK_HUBBUB = 0x1; + const unsigned int DC_HW_STATE_MASK_HUBP = 0x2; + const unsigned int DC_HW_STATE_MASK_RQ = 0x4; + const unsigned int DC_HW_STATE_MASK_DLG = 0x8; + const unsigned int DC_HW_STATE_MASK_TTU = 0x10; + const unsigned int DC_HW_STATE_MASK_CM = 0x20; + const unsigned int DC_HW_STATE_MASK_MPCC = 0x40; + const unsigned int DC_HW_STATE_MASK_OTG = 0x80; + const unsigned int DC_HW_STATE_MASK_CLOCKS = 0x100; + const unsigned int DC_HW_STATE_INVAR_ONLY = 0x8000; unsigned int chars_printed = 0; unsigned int remaining_buf_size = bufSize; @@ -603,6 +607,9 @@ void dcn10_get_hw_state(struct dc *dc, char *pBuf, unsigned int bufSize, unsigne remaining_buf_size -= chars_printed; } - if ((mask & DC_HW_STATE_MASK_CLOCKS) && remaining_buf_size > 0) + if ((mask & DC_HW_STATE_MASK_CLOCKS) && remaining_buf_size > 0) { chars_printed = dcn10_get_clock_states(dc, pBuf, remaining_buf_size); + pBuf += chars_printed; + remaining_buf_size -= chars_printed; + } } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c index 47dbe4bb294a..5d4772dec0ba 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c @@ -202,7 +202,6 @@ enum dcn10_clk_src_array_id { #define MMHUB_SR(reg_name)\ .reg_name = MMHUB_BASE(mm ## reg_name ## _BASE_IDX) + \ mm ## reg_name - /* macros to expend register list macro defined in HW object header file * end *********************/ @@ -436,7 +435,6 @@ static const struct dcn_optc_mask tg_mask = { TG_COMMON_MASK_SH_LIST_DCN1_0(_MASK) }; - static const struct bios_registers bios_regs = { NBIO_SR(BIOS_SCRATCH_0), NBIO_SR(BIOS_SCRATCH_3), @@ -497,7 +495,6 @@ static const struct dce110_clk_src_mask cs_mask = { CS_COMMON_MASK_SH_LIST_DCN1_0(_MASK) }; - static const struct resource_caps res_cap = { .num_timing_generator = 4, .num_opp = 4, @@ -1277,7 +1274,6 @@ static bool construct( goto fail; } } - pool->base.clk_mgr = dcn1_clk_mgr_create(ctx); if (pool->base.clk_mgr == NULL) { dm_error("DC: failed to create display clock!\n"); diff --git a/drivers/gpu/drm/amd/display/dc/dm_event_log.h b/drivers/gpu/drm/amd/display/dc/dm_event_log.h index 34a701ca879e..65663f4d93e1 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_event_log.h +++ b/drivers/gpu/drm/amd/display/dc/dm_event_log.h @@ -33,6 +33,7 @@ #define EVENT_LOG_AUX_REQ(ddc, type, action, address, len, data) #define EVENT_LOG_AUX_REP(ddc, type, replyStatus, len, data) +#define EVENT_LOG_CUST_MSG(tag, a, ...) #endif diff --git a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h index beb08fd12b1d..0029a39efb1c 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h +++ b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h @@ -102,7 +102,7 @@ struct pp_smu_funcs_rv { */ void (*set_display_count)(struct pp_smu *pp, int count); - /* which SMU message? are reader and writer WM separate SMU msg? */ + /* reader and writer WM's are sent together as part of one table*/ /* * PPSMC_MSG_SetDriverDramAddrHigh * PPSMC_MSG_SetDriverDramAddrLow diff --git a/drivers/gpu/drm/amd/display/dc/dm_services.h b/drivers/gpu/drm/amd/display/dc/dm_services.h index 28128c02de00..1961cc6d9143 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_services.h +++ b/drivers/gpu/drm/amd/display/dc/dm_services.h @@ -31,6 +31,8 @@ #define __DM_SERVICES_H__ +#include "amdgpu_dm_trace.h" + /* TODO: remove when DC is complete. */ #include "dm_services_types.h" #include "logger_interface.h" @@ -70,6 +72,7 @@ static inline uint32_t dm_read_reg_func( } #endif value = cgs_read_register(ctx->cgs_device, address); + trace_amdgpu_dc_rreg(&ctx->perf_trace->read_count, address, value); return value; } @@ -90,6 +93,7 @@ static inline void dm_write_reg_func( } #endif cgs_write_register(ctx->cgs_device, address, value); + trace_amdgpu_dc_wreg(&ctx->perf_trace->write_count, address, value); } static inline uint32_t dm_read_index_reg( @@ -351,8 +355,12 @@ unsigned long long dm_get_elapse_time_in_ns(struct dc_context *ctx, /* * performance tracing */ -void dm_perf_trace_timestamp(const char *func_name, unsigned int line); -#define PERF_TRACE() dm_perf_trace_timestamp(__func__, __LINE__) +#define PERF_TRACE() trace_amdgpu_dc_performance(CTX->perf_trace->read_count,\ + CTX->perf_trace->write_count, &CTX->perf_trace->last_entry_read,\ + &CTX->perf_trace->last_entry_write, __func__, __LINE__) +#define PERF_TRACE_CTX(__CTX) trace_amdgpu_dc_performance(__CTX->perf_trace->read_count,\ + __CTX->perf_trace->write_count, &__CTX->perf_trace->last_entry_read,\ + &__CTX->perf_trace->last_entry_write, __func__, __LINE__) /* diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c index f20161c5706d..dada04296025 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c @@ -56,7 +56,6 @@ struct gpio_service *dal_gpio_service_create( struct dc_context *ctx) { struct gpio_service *service; - uint32_t index_of_id; service = kzalloc(sizeof(struct gpio_service), GFP_KERNEL); @@ -78,44 +77,33 @@ struct gpio_service *dal_gpio_service_create( goto failure_1; } - /* allocate and initialize business storage */ + /* allocate and initialize busyness storage */ { - const uint32_t bits_per_uint = sizeof(uint32_t) << 3; - index_of_id = 0; service->ctx = ctx; do { uint32_t number_of_bits = service->factory.number_of_pins[index_of_id]; + uint32_t i = 0; - uint32_t number_of_uints = - (number_of_bits + bits_per_uint - 1) / - bits_per_uint; - - uint32_t *slot; - - if (number_of_bits) { - uint32_t index_of_uint = 0; + if (number_of_bits) { + service->busyness[index_of_id] = + kcalloc(number_of_bits, sizeof(char), + GFP_KERNEL); - slot = kcalloc(number_of_uints, - sizeof(uint32_t), - GFP_KERNEL); - - if (!slot) { + if (!service->busyness[index_of_id]) { BREAK_TO_DEBUGGER(); goto failure_2; } do { - slot[index_of_uint] = 0; - - ++index_of_uint; - } while (index_of_uint < number_of_uints); - } else - slot = NULL; - - service->busyness[index_of_id] = slot; + service->busyness[index_of_id][i] = 0; + ++i; + } while (i < number_of_bits); + } else { + service->busyness[index_of_id] = NULL; + } ++index_of_id; } while (index_of_id < GPIO_ID_COUNT); @@ -125,13 +113,8 @@ struct gpio_service *dal_gpio_service_create( failure_2: while (index_of_id) { - uint32_t *slot; - --index_of_id; - - slot = service->busyness[index_of_id]; - - kfree(slot); + kfree(service->busyness[index_of_id]); } failure_1: @@ -169,9 +152,7 @@ void dal_gpio_service_destroy( uint32_t index_of_id = 0; do { - uint32_t *slot = (*ptr)->busyness[index_of_id]; - - kfree(slot); + kfree((*ptr)->busyness[index_of_id]); ++index_of_id; } while (index_of_id < GPIO_ID_COUNT); @@ -192,11 +173,7 @@ static bool is_pin_busy( enum gpio_id id, uint32_t en) { - const uint32_t bits_per_uint = sizeof(uint32_t) << 3; - - const uint32_t *slot = service->busyness[id] + (en / bits_per_uint); - - return 0 != (*slot & (1 << (en % bits_per_uint))); + return service->busyness[id][en]; } static void set_pin_busy( @@ -204,10 +181,7 @@ static void set_pin_busy( enum gpio_id id, uint32_t en) { - const uint32_t bits_per_uint = sizeof(uint32_t) << 3; - - service->busyness[id][en / bits_per_uint] |= - (1 << (en % bits_per_uint)); + service->busyness[id][en] = true; } static void set_pin_free( @@ -215,10 +189,7 @@ static void set_pin_free( enum gpio_id id, uint32_t en) { - const uint32_t bits_per_uint = sizeof(uint32_t) << 3; - - service->busyness[id][en / bits_per_uint] &= - ~(1 << (en % bits_per_uint)); + service->busyness[id][en] = false; } enum gpio_result dal_gpio_service_open( diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h index c7f3081f59cc..1d501a43d13b 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h @@ -36,10 +36,9 @@ struct gpio_service { /* * @brief * Business storage. - * For each member of 'enum gpio_id', - * store array of bits (packed into uint32_t slots), - * index individual bit by 'en' value */ - uint32_t *busyness[GPIO_ID_COUNT]; + * one byte For each member of 'enum gpio_id' + */ + char *busyness[GPIO_ID_COUNT]; }; enum gpio_result dal_gpio_service_open( diff --git a/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c b/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c index a683f4102e65..c2028c4744a6 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/hw_factory.c @@ -79,6 +79,7 @@ bool dal_hw_factory_init( dal_hw_factory_dce110_init(factory); return true; case DCE_VERSION_12_0: + case DCE_VERSION_12_1: dal_hw_factory_dce120_init(factory); return true; #if defined(CONFIG_DRM_AMD_DC_DCN1_0) diff --git a/drivers/gpu/drm/amd/display/dc/gpio/hw_translate.c b/drivers/gpu/drm/amd/display/dc/gpio/hw_translate.c index 096f45628630..236ca28784a9 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/hw_translate.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/hw_translate.c @@ -76,6 +76,7 @@ bool dal_hw_translate_init( dal_hw_translate_dce110_init(translate); return true; case DCE_VERSION_12_0: + case DCE_VERSION_12_1: dal_hw_translate_dce120_init(translate); return true; #if defined(CONFIG_DRM_AMD_DC_DCN1_0) diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c index e56093f26eed..1ad6e49102ff 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c +++ b/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c @@ -90,6 +90,7 @@ struct i2caux *dal_i2caux_create( case DCE_VERSION_10_0: return dal_i2caux_dce100_create(ctx); case DCE_VERSION_12_0: + case DCE_VERSION_12_1: return dal_i2caux_dce120_create(ctx); #if defined(CONFIG_DRM_AMD_DC_DCN1_0) case DCN_VERSION_1_0: diff --git a/drivers/gpu/drm/amd/display/dc/inc/compressor.h b/drivers/gpu/drm/amd/display/dc/inc/compressor.h index bcb18f5e1e60..7a147a9762a0 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/compressor.h +++ b/drivers/gpu/drm/amd/display/dc/inc/compressor.h @@ -77,6 +77,7 @@ struct compressor_funcs { }; struct compressor { struct dc_context *ctx; + /* CONTROLLER_ID_D0 + instance, CONTROLLER_ID_UNDEFINED = 0 */ uint32_t attached_inst; bool is_enabled; const struct compressor_funcs *funcs; diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h index e3ee96afa60e..b168a5e9dd9d 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -272,6 +272,17 @@ union bw_context { struct dce_bw_output dce; }; +/** + * struct dc_state - The full description of a state requested by a user + * + * @streams: Stream properties + * @stream_status: The planes on a given stream + * @res_ctx: Persistent state of resources + * @bw: The output from bandwidth and watermark calculations + * @pp_display_cfg: PowerPlay clocks and settings + * @dcn_bw_vars: non-stack memory to support bandwidth calculations + * + */ struct dc_state { struct dc_stream_state *streams[MAX_PIPES]; struct dc_stream_status stream_status[MAX_PIPES]; @@ -279,7 +290,6 @@ struct dc_state { struct resource_context res_ctx; - /* The output from BW and WM calculations. */ union bw_context bw; /* Note: these are big structures, do *not* put on stack! */ diff --git a/drivers/gpu/drm/amd/display/include/bios_parser_types.h b/drivers/gpu/drm/amd/display/include/bios_parser_types.h index f8dbfa5b89f2..7fd78a696800 100644 --- a/drivers/gpu/drm/amd/display/include/bios_parser_types.h +++ b/drivers/gpu/drm/amd/display/include/bios_parser_types.h @@ -41,6 +41,7 @@ enum as_signal_type { AS_SIGNAL_TYPE_LVDS, AS_SIGNAL_TYPE_DISPLAY_PORT, AS_SIGNAL_TYPE_GPU_PLL, + AS_SIGNAL_TYPE_XGMI, AS_SIGNAL_TYPE_UNKNOWN }; diff --git a/drivers/gpu/drm/amd/display/include/dal_types.h b/drivers/gpu/drm/amd/display/include/dal_types.h index 89627133e188..f5bd869d4320 100644 --- a/drivers/gpu/drm/amd/display/include/dal_types.h +++ b/drivers/gpu/drm/amd/display/include/dal_types.h @@ -42,6 +42,7 @@ enum dce_version { DCE_VERSION_11_2, DCE_VERSION_11_22, DCE_VERSION_12_0, + DCE_VERSION_12_1, DCE_VERSION_MAX, DCN_VERSION_1_0, #if defined(CONFIG_DRM_AMD_DC_DCN1_01) diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c index bbecbaefb741..479b77c2e89e 100644 --- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c +++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c @@ -1761,7 +1761,7 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, struct pwl_float_data *rgb_user = NULL; struct pwl_float_data_ex *curve = NULL; - struct gamma_pixel *axix_x = NULL; + struct gamma_pixel *axis_x = NULL; struct pixel_gamma_point *coeff = NULL; enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_SRGB; bool ret = false; @@ -1787,10 +1787,10 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, GFP_KERNEL); if (!curve) goto curve_alloc_fail; - axix_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axix_x), + axis_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axis_x), GFP_KERNEL); - if (!axix_x) - goto axix_x_alloc_fail; + if (!axis_x) + goto axis_x_alloc_fail; coeff = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*coeff), GFP_KERNEL); if (!coeff) @@ -1803,7 +1803,7 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, tf = input_tf->tf; build_evenly_distributed_points( - axix_x, + axis_x, ramp->num_entries, dividers); @@ -1828,7 +1828,7 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, tf_pts->x_point_at_y1_blue = 1; map_regamma_hw_to_x_user(ramp, coeff, rgb_user, - coordinates_x, axix_x, curve, + coordinates_x, axis_x, curve, MAX_HW_POINTS, tf_pts, mapUserRamp && ramp->type != GAMMA_CUSTOM); if (ramp->type == GAMMA_CUSTOM) @@ -1838,8 +1838,8 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, kvfree(coeff); coeff_alloc_fail: - kvfree(axix_x); -axix_x_alloc_fail: + kvfree(axis_x); +axis_x_alloc_fail: kvfree(curve); curve_alloc_fail: kvfree(rgb_user); diff --git a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c index 620a171620ee..1544ed3f1747 100644 --- a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c +++ b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c @@ -608,12 +608,12 @@ static void build_vrr_infopacket_data(const struct mod_vrr_params *vrr, static void build_vrr_infopacket_fs2_data(enum color_transfer_func app_tf, struct dc_info_packet *infopacket) { - if (app_tf != transfer_func_unknown) { + if (app_tf != TRANSFER_FUNC_UNKNOWN) { infopacket->valid = true; infopacket->sb[6] |= 0x08; // PB6 = [Bit 3 = Native Color Active] - if (app_tf == transfer_func_gamma_22) { + if (app_tf == TRANSFER_FUNC_GAMMA_22) { infopacket->sb[9] |= 0x04; // PB6 = [Bit 2 = Gamma 2.2 EOTF Active] } } @@ -688,11 +688,11 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, return; switch (packet_type) { - case packet_type_fs2: + case PACKET_TYPE_FS2: build_vrr_infopacket_v2(stream->signal, vrr, app_tf, infopacket); break; - case packet_type_vrr: - case packet_type_fs1: + case PACKET_TYPE_VRR: + case PACKET_TYPE_FS1: default: build_vrr_infopacket_v1(stream->signal, vrr, infopacket); } diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_info_packet.h b/drivers/gpu/drm/amd/display/modules/inc/mod_info_packet.h index 786b34380f85..5b1c9a4c7643 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_info_packet.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_info_packet.h @@ -26,15 +26,13 @@ #ifndef MOD_INFO_PACKET_H_ #define MOD_INFO_PACKET_H_ -struct info_packet_inputs { - const struct dc_stream_state *pStream; -}; +#include "mod_shared.h" -struct info_packets { - struct dc_info_packet *pVscInfoPacket; -}; +//Forward Declarations +struct dc_stream_state; +struct dc_info_packet; -void mod_build_infopackets(struct info_packet_inputs *inputs, - struct info_packets *info_packets); +void mod_build_vsc_infopacket(const struct dc_stream_state *stream, + struct dc_info_packet *info_packet); #endif diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h index 238c431ae483..1bd02c0ac30c 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h @@ -23,27 +23,26 @@ * */ - #ifndef MOD_SHARED_H_ #define MOD_SHARED_H_ enum color_transfer_func { - transfer_func_unknown, - transfer_func_srgb, - transfer_func_bt709, - transfer_func_pq2084, - transfer_func_pq2084_interim, - transfer_func_linear_0_1, - transfer_func_linear_0_125, - transfer_func_dolbyvision, - transfer_func_gamma_22, - transfer_func_gamma_26 + TRANSFER_FUNC_UNKNOWN, + TRANSFER_FUNC_SRGB, + TRANSFER_FUNC_BT709, + TRANSFER_FUNC_PQ2084, + TRANSFER_FUNC_PQ2084_INTERIM, + TRANSFER_FUNC_LINEAR_0_1, + TRANSFER_FUNC_LINEAR_0_125, + TRANSFER_FUNC_GAMMA_22, + TRANSFER_FUNC_GAMMA_26 }; enum vrr_packet_type { - packet_type_vrr, - packet_type_fs1, - packet_type_fs2 + PACKET_TYPE_VRR, + PACKET_TYPE_FS1, + PACKET_TYPE_FS2 }; + #endif /* MOD_SHARED_H_ */ diff --git a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c index ff8bfb9b43b0..db06fab2ad5c 100644 --- a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c +++ b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c @@ -25,6 +25,10 @@ #include "mod_info_packet.h" #include "core_types.h" +#include "dc_types.h" +#include "mod_shared.h" + +#define HDMI_INFOFRAME_TYPE_VENDOR 0x81 enum ColorimetryRGBDP { ColorimetryRGB_DP_sRGB = 0, @@ -41,7 +45,7 @@ enum ColorimetryYCCDP { ColorimetryYCC_DP_ITU2020YCbCr = 7, }; -static void mod_build_vsc_infopacket(const struct dc_stream_state *stream, +void mod_build_vsc_infopacket(const struct dc_stream_state *stream, struct dc_info_packet *info_packet) { unsigned int vscPacketRevision = 0; @@ -159,7 +163,7 @@ static void mod_build_vsc_infopacket(const struct dc_stream_state *stream, * DPCD register is exposed in the new Extended Receiver Capability field for DPCD Rev. 1.4 * (and higher). When MISC1. bit 6. is Set to 1, a Source device uses a VSC SDP to indicate * the Pixel Encoding/Colorimetry Format and that a Sink device must ignore MISC1, bit 7, and - * MISC0, bits 7:1 (MISC1, bit 7. and MISC0, bits 7:1 become “don’t care”).) + * MISC0, bits 7:1 (MISC1, bit 7. and MISC0, bits 7:1 become "don't care").) */ if (vscPacketRevision == 0x5) { /* Secondary-data Packet ID = 0 */ @@ -320,10 +324,3 @@ static void mod_build_vsc_infopacket(const struct dc_stream_state *stream, } -void mod_build_infopackets(struct info_packet_inputs *inputs, - struct info_packets *info_packets) -{ - if (info_packets->pVscInfoPacket != NULL) - mod_build_vsc_infopacket(inputs->pStream, info_packets->pVscInfoPacket); -} - diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index 00f63b7dd32f..c11a443dcbc8 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -57,6 +57,7 @@ static const unsigned char abm_config[abm_defines_max_config][abm_defines_max_le #define NUM_POWER_FN_SEGS 8 #define NUM_BL_CURVE_SEGS 16 +#pragma pack(push, 1) /* NOTE: iRAM is 256B in size */ struct iram_table_v_2 { /* flags */ @@ -100,6 +101,7 @@ struct iram_table_v_2 { uint8_t dummy8; /* 0xfe */ uint8_t dummy9; /* 0xff */ }; +#pragma pack(pop) static uint16_t backlight_8_to_16(unsigned int backlight_8bit) { diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_offset.h new file mode 100644 index 000000000000..8f515875a34d --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_offset.h @@ -0,0 +1,32 @@ +/* + * Copyright (C) 2018 Advanced Micro Devices, 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) 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. + */ +#ifndef _mmhub_9_4_0_OFFSET_HEADER +#define _mmhub_9_4_0_OFFSET_HEADER + + +// addressBlock: mmhub_utcl2_vmsharedpfdec +// base address: 0x6a040 +#define mmMC_VM_XGMI_LFB_CNTL 0x0823 +#define mmMC_VM_XGMI_LFB_CNTL_BASE_IDX 0 +#define mmMC_VM_XGMI_LFB_SIZE 0x0824 +#define mmMC_VM_XGMI_LFB_SIZE_BASE_IDX 0 + +#endif diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_sh_mask.h new file mode 100644 index 000000000000..0a6b072d191e --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_0_sh_mask.h @@ -0,0 +1,35 @@ +/* + * Copyright (C) 2018 Advanced Micro Devices, 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) 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. + */ +#ifndef _mmhub_9_4_0_SH_MASK_HEADER +#define _mmhub_9_4_0_SH_MASK_HEADER + + +// addressBlock: mmhub_utcl2_vmsharedpfdec +//MC_VM_XGMI_LFB_CNTL +#define MC_VM_XGMI_LFB_CNTL__PF_LFB_REGION__SHIFT 0x0 +#define MC_VM_XGMI_LFB_CNTL__PF_MAX_REGION__SHIFT 0x4 +#define MC_VM_XGMI_LFB_CNTL__PF_LFB_REGION_MASK 0x00000007L +#define MC_VM_XGMI_LFB_CNTL__PF_MAX_REGION_MASK 0x00000070L +//MC_VM_XGMI_LFB_SIZE +#define MC_VM_XGMI_LFB_SIZE__PF_LFB_SIZE__SHIFT 0x0 +#define MC_VM_XGMI_LFB_SIZE__PF_LFB_SIZE_MASK 0x0000FFFFL + +#endif diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h index 58ac0b90c310..8154d67388cc 100644 --- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h @@ -188,8 +188,8 @@ struct tile_config { */ #define ALLOC_MEM_FLAGS_VRAM (1 << 0) #define ALLOC_MEM_FLAGS_GTT (1 << 1) -#define ALLOC_MEM_FLAGS_USERPTR (1 << 2) /* TODO */ -#define ALLOC_MEM_FLAGS_DOORBELL (1 << 3) /* TODO */ +#define ALLOC_MEM_FLAGS_USERPTR (1 << 2) +#define ALLOC_MEM_FLAGS_DOORBELL (1 << 3) /* * Allocation flags attributes/access options. diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h index 980e696989b1..789c4f288485 100644 --- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h @@ -127,12 +127,13 @@ enum amd_pp_task { }; enum PP_SMC_POWER_PROFILE { - PP_SMC_POWER_PROFILE_FULLSCREEN3D = 0x0, - PP_SMC_POWER_PROFILE_POWERSAVING = 0x1, - PP_SMC_POWER_PROFILE_VIDEO = 0x2, - PP_SMC_POWER_PROFILE_VR = 0x3, - PP_SMC_POWER_PROFILE_COMPUTE = 0x4, - PP_SMC_POWER_PROFILE_CUSTOM = 0x5, + PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT = 0x0, + PP_SMC_POWER_PROFILE_FULLSCREEN3D = 0x1, + PP_SMC_POWER_PROFILE_POWERSAVING = 0x2, + PP_SMC_POWER_PROFILE_VIDEO = 0x3, + PP_SMC_POWER_PROFILE_VR = 0x4, + PP_SMC_POWER_PROFILE_COMPUTE = 0x5, + PP_SMC_POWER_PROFILE_CUSTOM = 0x6, }; enum { @@ -276,6 +277,10 @@ struct amd_pm_funcs { struct amd_pp_simple_clock_info *clocks); int (*notify_smu_enable_pwe)(void *handle); int (*enable_mgpu_fan_boost)(void *handle); + int (*set_active_display_count)(void *handle, uint32_t count); + int (*set_hard_min_dcefclk_by_freq)(void *handle, uint32_t clock); + int (*set_hard_min_fclk_by_freq)(void *handle, uint32_t clock); + int (*set_min_deep_sleep_dcefclk)(void *handle, uint32_t clock); }; #endif diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index b68c2e0fef01..9bc27f468d5b 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -725,7 +725,7 @@ static int pp_dpm_force_clock_level(void *handle, } if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) { - pr_info("force clock level is for dpm manual mode only.\n"); + pr_debug("force clock level is for dpm manual mode only.\n"); return -EINVAL; } @@ -899,7 +899,7 @@ static int pp_set_power_profile_mode(void *handle, long *input, uint32_t size) } if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) { - pr_info("power profile setting is for manual dpm mode only.\n"); + pr_debug("power profile setting is for manual dpm mode only.\n"); return ret; } @@ -1072,7 +1072,7 @@ static int pp_get_current_clocks(void *handle, &hw_clocks, PHM_PerformanceLevelDesignation_Activity); if (ret) { - pr_info("Error in phm_get_clock_info \n"); + pr_debug("Error in phm_get_clock_info \n"); mutex_unlock(&hwmgr->smu_lock); return -EINVAL; } @@ -1332,6 +1332,78 @@ static int pp_enable_mgpu_fan_boost(void *handle) return 0; } +static int pp_set_min_deep_sleep_dcefclk(void *handle, uint32_t clock) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk == NULL) { + pr_debug("%s was not implemented.\n", __func__); + return -EINVAL;; + } + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk(hwmgr, clock); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_set_hard_min_dcefclk_by_freq(void *handle, uint32_t clock) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->set_hard_min_dcefclk_by_freq == NULL) { + pr_debug("%s was not implemented.\n", __func__); + return -EINVAL;; + } + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->set_hard_min_dcefclk_by_freq(hwmgr, clock); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_set_hard_min_fclk_by_freq(void *handle, uint32_t clock) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->set_hard_min_fclk_by_freq == NULL) { + pr_debug("%s was not implemented.\n", __func__); + return -EINVAL;; + } + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->set_hard_min_fclk_by_freq(hwmgr, clock); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_set_active_display_count(void *handle, uint32_t count) +{ + struct pp_hwmgr *hwmgr = handle; + int ret = 0; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + mutex_lock(&hwmgr->smu_lock); + ret = phm_set_active_display_count(hwmgr, count); + mutex_unlock(&hwmgr->smu_lock); + + return ret; +} + static const struct amd_pm_funcs pp_dpm_funcs = { .load_firmware = pp_dpm_load_fw, .wait_for_fw_loading_complete = pp_dpm_fw_loading_complete, @@ -1378,4 +1450,8 @@ static const struct amd_pm_funcs pp_dpm_funcs = { .get_display_mode_validation_clocks = pp_get_display_mode_validation_clocks, .notify_smu_enable_pwe = pp_notify_smu_enable_pwe, .enable_mgpu_fan_boost = pp_enable_mgpu_fan_boost, + .set_active_display_count = pp_set_active_display_count, + .set_min_deep_sleep_dcefclk = pp_set_min_deep_sleep_dcefclk, + .set_hard_min_dcefclk_by_freq = pp_set_hard_min_dcefclk_by_freq, + .set_hard_min_fclk_by_freq = pp_set_hard_min_fclk_by_freq, }; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c index 85119c2bdcc8..1f92a9f4c9e3 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c @@ -80,7 +80,9 @@ int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr) PHM_FUNC_CHECK(hwmgr); adev = hwmgr->adev; - if (smum_is_dpm_running(hwmgr) && !amdgpu_passthrough(adev)) { + /* Skip for suspend/resume case */ + if (smum_is_dpm_running(hwmgr) && !amdgpu_passthrough(adev) + && adev->in_suspend) { pr_info("dpm has been enabled\n"); return 0; } @@ -286,8 +288,8 @@ int phm_store_dal_configuration_data(struct pp_hwmgr *hwmgr, if (display_config == NULL) return -EINVAL; - if (NULL != hwmgr->hwmgr_func->set_deep_sleep_dcefclk) - hwmgr->hwmgr_func->set_deep_sleep_dcefclk(hwmgr, display_config->min_dcef_deep_sleep_set_clk); + if (NULL != hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk) + hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk(hwmgr, display_config->min_dcef_deep_sleep_set_clk); for (index = 0; index < display_config->num_path_including_non_display; index++) { if (display_config->displays[index].controller_id != 0) @@ -478,3 +480,44 @@ int phm_disable_smc_firmware_ctf(struct pp_hwmgr *hwmgr) return hwmgr->hwmgr_func->disable_smc_firmware_ctf(hwmgr); } + +int phm_set_active_display_count(struct pp_hwmgr *hwmgr, uint32_t count) +{ + PHM_FUNC_CHECK(hwmgr); + + if (!hwmgr->hwmgr_func->set_active_display_count) + return -EINVAL; + + return hwmgr->hwmgr_func->set_active_display_count(hwmgr, count); +} + +int phm_set_min_deep_sleep_dcefclk(struct pp_hwmgr *hwmgr, uint32_t clock) +{ + PHM_FUNC_CHECK(hwmgr); + + if (!hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk) + return -EINVAL; + + return hwmgr->hwmgr_func->set_min_deep_sleep_dcefclk(hwmgr, clock); +} + +int phm_set_hard_min_dcefclk_by_freq(struct pp_hwmgr *hwmgr, uint32_t clock) +{ + PHM_FUNC_CHECK(hwmgr); + + if (!hwmgr->hwmgr_func->set_hard_min_dcefclk_by_freq) + return -EINVAL; + + return hwmgr->hwmgr_func->set_hard_min_dcefclk_by_freq(hwmgr, clock); +} + +int phm_set_hard_min_fclk_by_freq(struct pp_hwmgr *hwmgr, uint32_t clock) +{ + PHM_FUNC_CHECK(hwmgr); + + if (!hwmgr->hwmgr_func->set_hard_min_fclk_by_freq) + return -EINVAL; + + return hwmgr->hwmgr_func->set_hard_min_fclk_by_freq(hwmgr, clock); +} + diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c index 47ac92369739..310b102a9292 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c @@ -64,17 +64,19 @@ static int ci_set_asic_special_caps(struct pp_hwmgr *hwmgr); static void hwmgr_init_workload_prority(struct pp_hwmgr *hwmgr) { - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 2; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 0; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO] = 1; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VR] = 3; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 4; - - hwmgr->workload_setting[0] = PP_SMC_POWER_PROFILE_POWERSAVING; - hwmgr->workload_setting[1] = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->workload_setting[2] = PP_SMC_POWER_PROFILE_FULLSCREEN3D; - hwmgr->workload_setting[3] = PP_SMC_POWER_PROFILE_VR; - hwmgr->workload_setting[4] = PP_SMC_POWER_PROFILE_COMPUTE; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT] = 0; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 1; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 2; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO] = 3; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VR] = 4; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 5; + + hwmgr->workload_setting[0] = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->workload_setting[1] = PP_SMC_POWER_PROFILE_FULLSCREEN3D; + hwmgr->workload_setting[2] = PP_SMC_POWER_PROFILE_POWERSAVING; + hwmgr->workload_setting[3] = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_setting[4] = PP_SMC_POWER_PROFILE_VR; + hwmgr->workload_setting[5] = PP_SMC_POWER_PROFILE_COMPUTE; } int hwmgr_early_init(struct pp_hwmgr *hwmgr) @@ -352,6 +354,9 @@ int hwmgr_handle_task(struct pp_hwmgr *hwmgr, enum amd_pp_task task_id, switch (task_id) { case AMD_PP_TASK_DISPLAY_CONFIG_CHANGE: + ret = phm_pre_display_configuration_changed(hwmgr); + if (ret) + return ret; ret = phm_set_cpu_power_state(hwmgr); if (ret) return ret; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c index 91ffb7bc4ee7..56437866d120 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c @@ -265,8 +265,6 @@ int psm_adjust_power_state_dynamic(struct pp_hwmgr *hwmgr, bool skip, if (skip) return 0; - phm_pre_display_configuration_changed(hwmgr); - phm_display_configuration_changed(hwmgr); if (hwmgr->ps) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c index dd18cb710391..f95c5f50eb0f 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c @@ -216,12 +216,12 @@ static inline uint32_t convert_10k_to_mhz(uint32_t clock) return (clock + 99) / 100; } -static int smu10_set_deep_sleep_dcefclk(struct pp_hwmgr *hwmgr, uint32_t clock) +static int smu10_set_min_deep_sleep_dcefclk(struct pp_hwmgr *hwmgr, uint32_t clock) { struct smu10_hwmgr *smu10_data = (struct smu10_hwmgr *)(hwmgr->backend); if (smu10_data->need_min_deep_sleep_dcefclk && - smu10_data->deep_sleep_dcefclk != convert_10k_to_mhz(clock)) { + smu10_data->deep_sleep_dcefclk != convert_10k_to_mhz(clock)) { smu10_data->deep_sleep_dcefclk = convert_10k_to_mhz(clock); smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetMinDeepSleepDcefclk, @@ -230,6 +230,34 @@ static int smu10_set_deep_sleep_dcefclk(struct pp_hwmgr *hwmgr, uint32_t clock) return 0; } +static int smu10_set_hard_min_dcefclk_by_freq(struct pp_hwmgr *hwmgr, uint32_t clock) +{ + struct smu10_hwmgr *smu10_data = (struct smu10_hwmgr *)(hwmgr->backend); + + if (smu10_data->dcf_actual_hard_min_freq && + smu10_data->dcf_actual_hard_min_freq != convert_10k_to_mhz(clock)) { + smu10_data->dcf_actual_hard_min_freq = convert_10k_to_mhz(clock); + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_SetHardMinDcefclkByFreq, + smu10_data->dcf_actual_hard_min_freq); + } + return 0; +} + +static int smu10_set_hard_min_fclk_by_freq(struct pp_hwmgr *hwmgr, uint32_t clock) +{ + struct smu10_hwmgr *smu10_data = (struct smu10_hwmgr *)(hwmgr->backend); + + if (smu10_data->f_actual_hard_min_freq && + smu10_data->f_actual_hard_min_freq != convert_10k_to_mhz(clock)) { + smu10_data->f_actual_hard_min_freq = convert_10k_to_mhz(clock); + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_SetHardMinFclkByFreq, + smu10_data->f_actual_hard_min_freq); + } + return 0; +} + static int smu10_set_active_display_count(struct pp_hwmgr *hwmgr, uint32_t count) { struct smu10_hwmgr *smu10_data = (struct smu10_hwmgr *)(hwmgr->backend); @@ -1206,7 +1234,7 @@ static const struct pp_hwmgr_func smu10_hwmgr_funcs = { .get_max_high_clocks = smu10_get_max_high_clocks, .read_sensor = smu10_read_sensor, .set_active_display_count = smu10_set_active_display_count, - .set_deep_sleep_dcefclk = smu10_set_deep_sleep_dcefclk, + .set_min_deep_sleep_dcefclk = smu10_set_min_deep_sleep_dcefclk, .dynamic_state_management_enable = smu10_enable_dpm_tasks, .power_off_asic = smu10_power_off_asic, .asic_setup = smu10_setup_asic_task, @@ -1217,6 +1245,8 @@ static const struct pp_hwmgr_func smu10_hwmgr_funcs = { .display_clock_voltage_request = smu10_display_clock_voltage_request, .powergate_gfx = smu10_gfx_off_control, .powergate_sdma = smu10_powergate_sdma, + .set_hard_min_dcefclk_by_freq = smu10_set_hard_min_dcefclk_by_freq, + .set_hard_min_fclk_by_freq = smu10_set_hard_min_fclk_by_freq, }; int smu10_init_function_pointers(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c index 5dcd21d29dbf..c8f5c00dd1e7 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c @@ -77,8 +77,9 @@ #define PCIE_BUS_CLK 10000 #define TCLK (PCIE_BUS_CLK / 10) -static const struct profile_mode_setting smu7_profiling[6] = - {{1, 0, 100, 30, 1, 0, 100, 10}, +static const struct profile_mode_setting smu7_profiling[7] = + {{0, 0, 0, 0, 0, 0, 0, 0}, + {1, 0, 100, 30, 1, 0, 100, 10}, {1, 10, 0, 30, 0, 0, 0, 0}, {0, 0, 0, 0, 1, 10, 16, 31}, {1, 0, 11, 50, 1, 0, 100, 10}, @@ -2859,7 +2860,10 @@ static int smu7_vblank_too_short(struct pp_hwmgr *hwmgr, case CHIP_POLARIS10: case CHIP_POLARIS11: case CHIP_POLARIS12: - switch_limit_us = data->is_memory_gddr5 ? 190 : 150; + if (hwmgr->is_kicker) + switch_limit_us = data->is_memory_gddr5 ? 450 : 150; + else + switch_limit_us = data->is_memory_gddr5 ? 190 : 150; break; case CHIP_VEGAM: switch_limit_us = 30; @@ -3589,8 +3593,10 @@ static int smu7_find_dpm_states_clocks_in_dpm_table(struct pp_hwmgr *hwmgr, cons } if (i >= sclk_table->count) { - data->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; - sclk_table->dpm_levels[i-1].value = sclk; + if (sclk > sclk_table->dpm_levels[i-1].value) { + data->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; + sclk_table->dpm_levels[i-1].value = sclk; + } } else { /* TODO: Check SCLK in DAL's minimum clocks * in case DeepSleep divider update is required. @@ -3607,8 +3613,10 @@ static int smu7_find_dpm_states_clocks_in_dpm_table(struct pp_hwmgr *hwmgr, cons } if (i >= mclk_table->count) { - data->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; - mclk_table->dpm_levels[i-1].value = mclk; + if (mclk > mclk_table->dpm_levels[i-1].value) { + data->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; + mclk_table->dpm_levels[i-1].value = mclk; + } } if (data->display_timing.num_existing_displays != hwmgr->display_config->num_display) @@ -4219,9 +4227,17 @@ static int smu7_check_mc_firmware(struct pp_hwmgr *hwmgr) if (tmp & (1 << 23)) { data->mem_latency_high = MEM_LATENCY_HIGH; data->mem_latency_low = MEM_LATENCY_LOW; + if ((hwmgr->chip_id == CHIP_POLARIS10) || + (hwmgr->chip_id == CHIP_POLARIS11) || + (hwmgr->chip_id == CHIP_POLARIS12)) + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_EnableFFC); } else { data->mem_latency_high = 330; data->mem_latency_low = 330; + if ((hwmgr->chip_id == CHIP_POLARIS10) || + (hwmgr->chip_id == CHIP_POLARIS11) || + (hwmgr->chip_id == CHIP_POLARIS12)) + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_DisableFFC); } return 0; @@ -4874,7 +4890,8 @@ static int smu7_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) uint32_t i, size = 0; uint32_t len; - static const char *profile_name[6] = {"3D_FULL_SCREEN", + static const char *profile_name[7] = {"BOOTUP_DEFAULT", + "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", "VR", diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c index e2bc6e0c229f..91e3bbe6d61d 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c @@ -804,9 +804,9 @@ static int vega10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) hwmgr->backend = data; - hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO]; - hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; + hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; vega10_set_default_registry_data(hwmgr); data->disable_dpm_mask = 0xff; @@ -3266,8 +3266,10 @@ static int vega10_find_dpm_states_clocks_in_dpm_table(struct pp_hwmgr *hwmgr, co } if (i >= sclk_table->count) { - data->need_update_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; - sclk_table->dpm_levels[i-1].value = sclk; + if (sclk > sclk_table->dpm_levels[i-1].value) { + data->need_update_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; + sclk_table->dpm_levels[i-1].value = sclk; + } } for (i = 0; i < mclk_table->count; i++) { @@ -3276,8 +3278,10 @@ static int vega10_find_dpm_states_clocks_in_dpm_table(struct pp_hwmgr *hwmgr, co } if (i >= mclk_table->count) { - data->need_update_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; - mclk_table->dpm_levels[i-1].value = mclk; + if (mclk > mclk_table->dpm_levels[i-1].value) { + data->need_update_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; + mclk_table->dpm_levels[i-1].value = mclk; + } } if (data->display_timing.num_existing_displays != hwmgr->display_config->num_display) @@ -4664,13 +4668,15 @@ static int vega10_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) { struct vega10_hwmgr *data = hwmgr->backend; uint32_t i, size = 0; - static const uint8_t profile_mode_setting[5][4] = {{70, 60, 1, 3,}, + static const uint8_t profile_mode_setting[6][4] = {{70, 60, 0, 0,}, + {70, 60, 1, 3,}, {90, 60, 0, 0,}, {70, 60, 0, 0,}, {70, 90, 0, 0,}, {30, 60, 0, 6,}, }; - static const char *profile_name[6] = {"3D_FULL_SCREEN", + static const char *profile_name[7] = {"BOOTUP_DEFAULT", + "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", "VR", diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c index b8747a5c9204..99d596dc0e89 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c @@ -32,6 +32,7 @@ #include "vega10_pptable.h" #define NUM_DSPCLK_LEVELS 8 +#define VEGA10_ENGINECLOCK_HARDMAX 198000 static void set_hw_cap(struct pp_hwmgr *hwmgr, bool enable, enum phm_platform_caps cap) @@ -258,7 +259,26 @@ static int init_over_drive_limits( struct pp_hwmgr *hwmgr, const ATOM_Vega10_POWERPLAYTABLE *powerplay_table) { - hwmgr->platform_descriptor.overdriveLimit.engineClock = + const ATOM_Vega10_GFXCLK_Dependency_Table *gfxclk_dep_table = + (const ATOM_Vega10_GFXCLK_Dependency_Table *) + (((unsigned long) powerplay_table) + + le16_to_cpu(powerplay_table->usGfxclkDependencyTableOffset)); + bool is_acg_enabled = false; + ATOM_Vega10_GFXCLK_Dependency_Record_V2 *patom_record_v2; + + if (gfxclk_dep_table->ucRevId == 1) { + patom_record_v2 = + (ATOM_Vega10_GFXCLK_Dependency_Record_V2 *)gfxclk_dep_table->entries; + is_acg_enabled = + (bool)patom_record_v2[gfxclk_dep_table->ucNumEntries-1].ucACGEnable; + } + + if (powerplay_table->ulMaxODEngineClock > VEGA10_ENGINECLOCK_HARDMAX && + !is_acg_enabled) + hwmgr->platform_descriptor.overdriveLimit.engineClock = + VEGA10_ENGINECLOCK_HARDMAX; + else + hwmgr->platform_descriptor.overdriveLimit.engineClock = le32_to_cpu(powerplay_table->ulMaxODEngineClock); hwmgr->platform_descriptor.overdriveLimit.memoryClock = le32_to_cpu(powerplay_table->ulMaxODMemoryClock); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c index 54364444ecd1..0c8212902275 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c @@ -753,6 +753,22 @@ static int vega12_init_smc_table(struct pp_hwmgr *hwmgr) return 0; } +static int vega12_run_acg_btc(struct pp_hwmgr *hwmgr) +{ + uint32_t result; + + PP_ASSERT_WITH_CODE( + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_RunAcgBtc) == 0, + "[Run_ACG_BTC] Attempt to run ACG BTC failed!", + return -EINVAL); + + result = smum_get_argument(hwmgr); + PP_ASSERT_WITH_CODE(result == 1, + "Failed to run ACG BTC!", return -EINVAL); + + return 0; +} + static int vega12_set_allowed_featuresmask(struct pp_hwmgr *hwmgr) { struct vega12_hwmgr *data = @@ -931,6 +947,11 @@ static int vega12_enable_dpm_tasks(struct pp_hwmgr *hwmgr) "Failed to initialize SMC table!", result = tmp_result); + tmp_result = vega12_run_acg_btc(hwmgr); + PP_ASSERT_WITH_CODE(!tmp_result, + "Failed to run ACG BTC!", + result = tmp_result); + result = vega12_enable_all_smu_features(hwmgr); PP_ASSERT_WITH_CODE(!result, "Failed to enable all smu features!", diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c index 2679d1240fa1..82935a3bd950 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c @@ -49,6 +49,10 @@ #include "soc15_common.h" #include "smuio/smuio_9_0_offset.h" #include "smuio/smuio_9_0_sh_mask.h" +#include "nbio/nbio_7_4_sh_mask.h" + +#define smnPCIE_LC_SPEED_CNTL 0x11140290 +#define smnPCIE_LC_LINK_WIDTH_CNTL 0x11140288 static void vega20_set_default_registry_data(struct pp_hwmgr *hwmgr) { @@ -130,7 +134,7 @@ static void vega20_set_default_registry_data(struct pp_hwmgr *hwmgr) data->registry_data.disable_auto_wattman = 1; data->registry_data.auto_wattman_debug = 0; data->registry_data.auto_wattman_sample_period = 100; - data->registry_data.fclk_gfxclk_ratio = 0x3F6CCCCD; + data->registry_data.fclk_gfxclk_ratio = 0; data->registry_data.auto_wattman_threshold = 50; data->registry_data.gfxoff_controlled_by_driver = 1; data->gfxoff_allowed = false; @@ -386,9 +390,9 @@ static int vega20_hwmgr_backend_init(struct pp_hwmgr *hwmgr) hwmgr->backend = data; - hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO]; - hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; + hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; vega20_set_default_registry_data(hwmgr); @@ -976,6 +980,9 @@ static int vega20_od8_set_feature_capabilities( pp_table->FanZeroRpmEnable) od_settings->overdrive8_capabilities |= OD8_FAN_ZERO_RPM_CONTROL; + if (!od_settings->overdrive8_capabilities) + hwmgr->od_enabled = false; + return 0; } @@ -1660,14 +1667,15 @@ static uint32_t vega20_find_highest_dpm_level( return i; } -static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr) +static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr, uint32_t feature_mask) { struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); uint32_t min_freq; int ret = 0; - if (data->smu_features[GNLD_DPM_GFXCLK].enabled) { + if (data->smu_features[GNLD_DPM_GFXCLK].enabled && + (feature_mask & FEATURE_DPM_GFXCLK_MASK)) { min_freq = data->dpm_table.gfx_table.dpm_state.soft_min_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( hwmgr, PPSMC_MSG_SetSoftMinByFreq, @@ -1676,23 +1684,18 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_UCLK].enabled) { + if (data->smu_features[GNLD_DPM_UCLK].enabled && + (feature_mask & FEATURE_DPM_UCLK_MASK)) { min_freq = data->dpm_table.mem_table.dpm_state.soft_min_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( hwmgr, PPSMC_MSG_SetSoftMinByFreq, (PPCLK_UCLK << 16) | (min_freq & 0xffff))), "Failed to set soft min memclk !", return ret); - - min_freq = data->dpm_table.mem_table.dpm_state.hard_min_level; - PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( - hwmgr, PPSMC_MSG_SetHardMinByFreq, - (PPCLK_UCLK << 16) | (min_freq & 0xffff))), - "Failed to set hard min memclk !", - return ret); } - if (data->smu_features[GNLD_DPM_UVD].enabled) { + if (data->smu_features[GNLD_DPM_UVD].enabled && + (feature_mask & FEATURE_DPM_UVD_MASK)) { min_freq = data->dpm_table.vclk_table.dpm_state.soft_min_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1710,7 +1713,8 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_VCE].enabled) { + if (data->smu_features[GNLD_DPM_VCE].enabled && + (feature_mask & FEATURE_DPM_VCE_MASK)) { min_freq = data->dpm_table.eclk_table.dpm_state.soft_min_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1720,7 +1724,8 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_SOCCLK].enabled) { + if (data->smu_features[GNLD_DPM_SOCCLK].enabled && + (feature_mask & FEATURE_DPM_SOCCLK_MASK)) { min_freq = data->dpm_table.soc_table.dpm_state.soft_min_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1733,14 +1738,15 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr) return ret; } -static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr) +static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr, uint32_t feature_mask) { struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); uint32_t max_freq; int ret = 0; - if (data->smu_features[GNLD_DPM_GFXCLK].enabled) { + if (data->smu_features[GNLD_DPM_GFXCLK].enabled && + (feature_mask & FEATURE_DPM_GFXCLK_MASK)) { max_freq = data->dpm_table.gfx_table.dpm_state.soft_max_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1750,7 +1756,8 @@ static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_UCLK].enabled) { + if (data->smu_features[GNLD_DPM_UCLK].enabled && + (feature_mask & FEATURE_DPM_UCLK_MASK)) { max_freq = data->dpm_table.mem_table.dpm_state.soft_max_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1760,7 +1767,8 @@ static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_UVD].enabled) { + if (data->smu_features[GNLD_DPM_UVD].enabled && + (feature_mask & FEATURE_DPM_UVD_MASK)) { max_freq = data->dpm_table.vclk_table.dpm_state.soft_max_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1777,7 +1785,8 @@ static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_VCE].enabled) { + if (data->smu_features[GNLD_DPM_VCE].enabled && + (feature_mask & FEATURE_DPM_VCE_MASK)) { max_freq = data->dpm_table.eclk_table.dpm_state.soft_max_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -1787,7 +1796,8 @@ static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr) return ret); } - if (data->smu_features[GNLD_DPM_SOCCLK].enabled) { + if (data->smu_features[GNLD_DPM_SOCCLK].enabled && + (feature_mask & FEATURE_DPM_SOCCLK_MASK)) { max_freq = data->dpm_table.soc_table.dpm_state.soft_max_level; PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( @@ -2126,12 +2136,12 @@ static int vega20_force_dpm_highest(struct pp_hwmgr *hwmgr) data->dpm_table.mem_table.dpm_state.soft_max_level = data->dpm_table.mem_table.dpm_levels[soft_level].value; - ret = vega20_upload_dpm_min_level(hwmgr); + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to highest!", return ret); - ret = vega20_upload_dpm_max_level(hwmgr); + ret = vega20_upload_dpm_max_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload dpm max level to highest!", return ret); @@ -2158,12 +2168,12 @@ static int vega20_force_dpm_lowest(struct pp_hwmgr *hwmgr) data->dpm_table.mem_table.dpm_state.soft_max_level = data->dpm_table.mem_table.dpm_levels[soft_level].value; - ret = vega20_upload_dpm_min_level(hwmgr); + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to highest!", return ret); - ret = vega20_upload_dpm_max_level(hwmgr); + ret = vega20_upload_dpm_max_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload dpm max level to highest!", return ret); @@ -2176,12 +2186,12 @@ static int vega20_unforce_dpm_levels(struct pp_hwmgr *hwmgr) { int ret = 0; - ret = vega20_upload_dpm_min_level(hwmgr); + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload DPM Bootup Levels!", return ret); - ret = vega20_upload_dpm_max_level(hwmgr); + ret = vega20_upload_dpm_max_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload DPM Max Levels!", return ret); @@ -2234,17 +2244,24 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; + if (soft_max_level >= data->dpm_table.gfx_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.gfx_table.count - 1); + return -EINVAL; + } + data->dpm_table.gfx_table.dpm_state.soft_min_level = data->dpm_table.gfx_table.dpm_levels[soft_min_level].value; data->dpm_table.gfx_table.dpm_state.soft_max_level = data->dpm_table.gfx_table.dpm_levels[soft_max_level].value; - ret = vega20_upload_dpm_min_level(hwmgr); + ret = vega20_upload_dpm_min_level(hwmgr, FEATURE_DPM_GFXCLK_MASK); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to lowest!", return ret); - ret = vega20_upload_dpm_max_level(hwmgr); + ret = vega20_upload_dpm_max_level(hwmgr, FEATURE_DPM_GFXCLK_MASK); PP_ASSERT_WITH_CODE(!ret, "Failed to upload dpm max level to highest!", return ret); @@ -2254,17 +2271,24 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; + if (soft_max_level >= data->dpm_table.mem_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.mem_table.count - 1); + return -EINVAL; + } + data->dpm_table.mem_table.dpm_state.soft_min_level = data->dpm_table.mem_table.dpm_levels[soft_min_level].value; data->dpm_table.mem_table.dpm_state.soft_max_level = data->dpm_table.mem_table.dpm_levels[soft_max_level].value; - ret = vega20_upload_dpm_min_level(hwmgr); + ret = vega20_upload_dpm_min_level(hwmgr, FEATURE_DPM_UCLK_MASK); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to lowest!", return ret); - ret = vega20_upload_dpm_max_level(hwmgr); + ret = vega20_upload_dpm_max_level(hwmgr, FEATURE_DPM_UCLK_MASK); PP_ASSERT_WITH_CODE(!ret, "Failed to upload dpm max level to highest!", return ret); @@ -2272,6 +2296,18 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, break; case PP_PCIE: + soft_min_level = mask ? (ffs(mask) - 1) : 0; + soft_max_level = mask ? (fls(mask) - 1) : 0; + if (soft_min_level >= NUM_LINK_LEVELS || + soft_max_level >= NUM_LINK_LEVELS) + return -EINVAL; + + ret = smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_SetMinLinkDpmByIndex, soft_min_level); + PP_ASSERT_WITH_CODE(!ret, + "Failed to set min link dpm level!", + return ret); + break; default: @@ -2748,9 +2784,14 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr, data->od8_settings.od8_settings_array; OverDriveTable_t *od_table = &(data->smc_state_table.overdrive_table); + struct phm_ppt_v3_information *pptable_information = + (struct phm_ppt_v3_information *)hwmgr->pptable; + PPTable_t *pptable = (PPTable_t *)pptable_information->smc_pptable; + struct amdgpu_device *adev = hwmgr->adev; struct pp_clock_levels_with_latency clocks; int i, now, size = 0; int ret = 0; + uint32_t gen_speed, lane_width; switch (type) { case PP_SCLK: @@ -2788,6 +2829,28 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr, break; case PP_PCIE: + gen_speed = (RREG32_PCIE(smnPCIE_LC_SPEED_CNTL) & + PSWUSP0_PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK) + >> PSWUSP0_PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT; + lane_width = (RREG32_PCIE(smnPCIE_LC_LINK_WIDTH_CNTL) & + PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK) + >> PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT; + for (i = 0; i < NUM_LINK_LEVELS; i++) + size += sprintf(buf + size, "%d: %s %s %dMhz %s\n", i, + (pptable->PcieGenSpeed[i] == 0) ? "2.5GT/s," : + (pptable->PcieGenSpeed[i] == 1) ? "5.0GT/s," : + (pptable->PcieGenSpeed[i] == 2) ? "8.0GT/s," : + (pptable->PcieGenSpeed[i] == 3) ? "16.0GT/s," : "", + (pptable->PcieLaneCount[i] == 1) ? "x1" : + (pptable->PcieLaneCount[i] == 2) ? "x2" : + (pptable->PcieLaneCount[i] == 3) ? "x4" : + (pptable->PcieLaneCount[i] == 4) ? "x8" : + (pptable->PcieLaneCount[i] == 5) ? "x12" : + (pptable->PcieLaneCount[i] == 6) ? "x16" : "", + pptable->LclkFreq[i], + (gen_speed == pptable->PcieGenSpeed[i]) && + (lane_width == pptable->PcieLaneCount[i]) ? + "*" : ""); break; case OD_SCLK: @@ -3208,6 +3271,9 @@ static int conv_power_profile_to_pplib_workload(int power_profile) int pplib_workload = 0; switch (power_profile) { + case PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT: + pplib_workload = WORKLOAD_DEFAULT_BIT; + break; case PP_SMC_POWER_PROFILE_FULLSCREEN3D: pplib_workload = WORKLOAD_PPLIB_FULL_SCREEN_3D_BIT; break; @@ -3237,6 +3303,7 @@ static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) uint32_t i, size = 0; uint16_t workload_type = 0; static const char *profile_name[] = { + "BOOTUP_DEFAULT", "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h index 54fd0125d9cf..f4dab979a3a1 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h @@ -463,5 +463,8 @@ extern int phm_display_clock_voltage_request(struct pp_hwmgr *hwmgr, extern int phm_get_max_high_clocks(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks); extern int phm_disable_smc_firmware_ctf(struct pp_hwmgr *hwmgr); + +extern int phm_set_active_display_count(struct pp_hwmgr *hwmgr, uint32_t count); + #endif /* _HARDWARE_MANAGER_H_ */ diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h index fb0f96f7cdbc..8cb831b6a016 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h @@ -309,7 +309,7 @@ struct pp_hwmgr_func { int (*avfs_control)(struct pp_hwmgr *hwmgr, bool enable); int (*disable_smc_firmware_ctf)(struct pp_hwmgr *hwmgr); int (*set_active_display_count)(struct pp_hwmgr *hwmgr, uint32_t count); - int (*set_deep_sleep_dcefclk)(struct pp_hwmgr *hwmgr, uint32_t clock); + int (*set_min_deep_sleep_dcefclk)(struct pp_hwmgr *hwmgr, uint32_t clock); int (*start_thermal_controller)(struct pp_hwmgr *hwmgr, struct PP_TemperatureRange *range); int (*notify_cac_buffer_info)(struct pp_hwmgr *hwmgr, uint32_t virtual_addr_low, @@ -332,6 +332,8 @@ struct pp_hwmgr_func { int (*smus_notify_pwe)(struct pp_hwmgr *hwmgr); int (*powergate_sdma)(struct pp_hwmgr *hwmgr, bool bgate); int (*enable_mgpu_fan_boost)(struct pp_hwmgr *hwmgr); + int (*set_hard_min_dcefclk_by_freq)(struct pp_hwmgr *hwmgr, uint32_t clock); + int (*set_hard_min_fclk_by_freq)(struct pp_hwmgr *hwmgr, uint32_t clock); }; struct pp_table_func { @@ -703,7 +705,7 @@ enum PP_TABLE_VERSION { /** * The main hardware manager structure. */ -#define Workload_Policy_Max 5 +#define Workload_Policy_Max 6 struct pp_hwmgr { void *adev; diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu7_ppsmc.h b/drivers/gpu/drm/amd/powerplay/inc/smu7_ppsmc.h index 62f36ba2435b..6e19f4c7cf8f 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smu7_ppsmc.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smu7_ppsmc.h @@ -386,6 +386,8 @@ typedef uint16_t PPSMC_Result; #define PPSMC_MSG_AgmResetPsm ((uint16_t) 0x403) #define PPSMC_MSG_ReadVftCell ((uint16_t) 0x404) +#define PPSMC_MSG_ApplyAvfsCksOffVoltage ((uint16_t) 0x415) + #define PPSMC_MSG_GFX_CU_PG_ENABLE ((uint16_t) 0x280) #define PPSMC_MSG_GFX_CU_PG_DISABLE ((uint16_t) 0x281) #define PPSMC_MSG_GetCurrPkgPwr ((uint16_t) 0x282) @@ -395,6 +397,9 @@ typedef uint16_t PPSMC_Result; #define PPSMC_MSG_SetVBITimeout ((uint16_t) 0x306) +#define PPSMC_MSG_EnableFFC ((uint16_t) 0x307) +#define PPSMC_MSG_DisableFFC ((uint16_t) 0x308) + #define PPSMC_MSG_EnableDpmDidt ((uint16_t) 0x309) #define PPSMC_MSG_DisableDpmDidt ((uint16_t) 0x30A) diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c index 2b2c26616902..52abca065764 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c @@ -1528,8 +1528,21 @@ static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr) efuse = efuse >> 24; if (hwmgr->chip_id == CHIP_POLARIS10) { - min = 1000; - max = 2300; + if (hwmgr->is_kicker) { + min = 1200; + max = 2500; + } else { + min = 1000; + max = 2300; + } + } else if (hwmgr->chip_id == CHIP_POLARIS11) { + if (hwmgr->is_kicker) { + min = 900; + max = 2100; + } else { + min = 1100; + max = 2100; + } } else { min = 1100; max = 2100; @@ -1626,6 +1639,7 @@ static int polaris10_populate_avfs_parameters(struct pp_hwmgr *hwmgr) { struct smu7_hwmgr *data = (struct smu7_hwmgr *)(hwmgr->backend); struct polaris10_smumgr *smu_data = (struct polaris10_smumgr *)(hwmgr->smu_backend); + struct amdgpu_device *adev = hwmgr->adev; SMU74_Discrete_DpmTable *table = &(smu_data->smc_state_table); int result = 0; @@ -1646,6 +1660,59 @@ static int polaris10_populate_avfs_parameters(struct pp_hwmgr *hwmgr) result = atomctrl_get_avfs_information(hwmgr, &avfs_params); if (0 == result) { + if (((adev->pdev->device == 0x67ef) && + ((adev->pdev->revision == 0xe0) || + (adev->pdev->revision == 0xe5))) || + ((adev->pdev->device == 0x67ff) && + ((adev->pdev->revision == 0xcf) || + (adev->pdev->revision == 0xef) || + (adev->pdev->revision == 0xff)))) { + avfs_params.ucEnableApplyAVFS_CKS_OFF_Voltage = 1; + if ((adev->pdev->device == 0x67ef && adev->pdev->revision == 0xe5) || + (adev->pdev->device == 0x67ff && adev->pdev->revision == 0xef)) { + if ((avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0 == 0xEA522DD3) && + (avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1 == 0x5645A) && + (avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2 == 0x33F9E) && + (avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1 == 0xFFFFC5CC) && + (avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2 == 0x1B1A) && + (avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b == 0xFFFFFCED)) { + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0 = 0xF718F1D4; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1 = 0x323FD; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2 = 0x1E455; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1 = 0; + avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2 = 0; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b = 0x23; + } + } + } else if (hwmgr->chip_id == CHIP_POLARIS12 && !hwmgr->is_kicker) { + avfs_params.ucEnableApplyAVFS_CKS_OFF_Voltage = 1; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0 = 0xF6B024DD; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1 = 0x3005E; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2 = 0x18A5F; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1 = 0x315; + avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2 = 0xFED1; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b = 0x3B; + } else if (((adev->pdev->device == 0x67df) && + ((adev->pdev->revision == 0xe0) || + (adev->pdev->revision == 0xe3) || + (adev->pdev->revision == 0xe4) || + (adev->pdev->revision == 0xe5) || + (adev->pdev->revision == 0xe7) || + (adev->pdev->revision == 0xef))) || + ((adev->pdev->device == 0x6fdf) && + ((adev->pdev->revision == 0xef) || + (adev->pdev->revision == 0xff)))) { + avfs_params.ucEnableApplyAVFS_CKS_OFF_Voltage = 1; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0 = 0xF843B66B; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1 = 0x59CB5; + avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2 = 0xFFFF287F; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1 = 0; + avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2 = 0xFF23; + avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b = 0x58; + } + } + + 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); @@ -1984,6 +2051,12 @@ int polaris10_thermal_avfs_enable(struct pp_hwmgr *hwmgr) smum_send_msg_to_smc(hwmgr, PPSMC_MSG_EnableAvfs); + /* Apply avfs cks-off voltages to avoid the overshoot + * when switching to the highest sclk frequency + */ + if (data->apply_avfs_cks_off_voltage) + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_ApplyAvfsCksOffVoltage); + return 0; } diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c index 892c1d9304bb..642d0e70d0f8 100644 --- a/drivers/gpu/drm/armada/armada_gem.c +++ b/drivers/gpu/drm/armada/armada_gem.c @@ -334,7 +334,7 @@ int armada_gem_pwrite_ioctl(struct drm_device *dev, void *data, ptr = (char __user *)(uintptr_t)args->ptr; - if (!access_ok(VERIFY_READ, ptr, args->size)) + if (!access_ok(ptr, args->size)) return -EFAULT; ret = fault_in_pages_readable(ptr, args->size); diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c index 0cd827e11fa2..de26df0c6044 100644 --- a/drivers/gpu/drm/ast/ast_fb.c +++ b/drivers/gpu/drm/ast/ast_fb.c @@ -263,6 +263,7 @@ static void ast_fbdev_destroy(struct drm_device *dev, { struct ast_framebuffer *afb = &afbdev->afb; + drm_crtc_force_disable_all(dev); drm_fb_helper_unregister_fbi(&afbdev->helper); if (afb->obj) { diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c index dac355812adc..373700c05a00 100644 --- a/drivers/gpu/drm/ast/ast_main.c +++ b/drivers/gpu/drm/ast/ast_main.c @@ -583,7 +583,8 @@ void ast_driver_unload(struct drm_device *dev) drm_mode_config_cleanup(dev); ast_mm_fini(ast); - pci_iounmap(dev->pdev, ast->ioregs); + if (ast->ioregs != ast->regs + AST_IO_MM_OFFSET) + pci_iounmap(dev->pdev, ast->ioregs); pci_iounmap(dev->pdev, ast->regs); kfree(ast); } diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c index 7c6ac3cadb6b..8bb355d5d43d 100644 --- a/drivers/gpu/drm/ast/ast_mode.c +++ b/drivers/gpu/drm/ast/ast_mode.c @@ -973,9 +973,21 @@ static int get_clock(void *i2c_priv) { struct ast_i2c_chan *i2c = i2c_priv; struct ast_private *ast = i2c->dev->dev_private; - uint32_t val; + uint32_t val, val2, count, pass; + + count = 0; + pass = 0; + val = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x10) >> 4) & 0x01; + do { + val2 = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x10) >> 4) & 0x01; + if (val == val2) { + pass++; + } else { + pass = 0; + val = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x10) >> 4) & 0x01; + } + } while ((pass < 5) && (count++ < 0x10000)); - val = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x10) >> 4; return val & 1 ? 1 : 0; } @@ -983,9 +995,21 @@ static int get_data(void *i2c_priv) { struct ast_i2c_chan *i2c = i2c_priv; struct ast_private *ast = i2c->dev->dev_private; - uint32_t val; + uint32_t val, val2, count, pass; + + count = 0; + pass = 0; + val = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x20) >> 5) & 0x01; + do { + val2 = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x20) >> 5) & 0x01; + if (val == val2) { + pass++; + } else { + pass = 0; + val = (ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x20) >> 5) & 0x01; + } + } while ((pass < 5) && (count++ < 0x10000)); - val = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x20) >> 5; return val & 1 ? 1 : 0; } @@ -998,7 +1022,7 @@ static void set_clock(void *i2c_priv, int clock) for (i = 0; i < 0x10000; i++) { ujcrb7 = ((clock & 0x01) ? 0 : 1); - ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0xfe, ujcrb7); + ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0xf4, ujcrb7); jtemp = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x01); if (ujcrb7 == jtemp) break; @@ -1014,7 +1038,7 @@ static void set_data(void *i2c_priv, int data) for (i = 0; i < 0x10000; i++) { ujcrb7 = ((data & 0x01) ? 0 : 1) << 2; - ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0xfb, ujcrb7); + ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0xf1, ujcrb7); jtemp = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb7, 0x04); if (ujcrb7 == jtemp) break; diff --git a/drivers/gpu/drm/bridge/tc358767.c b/drivers/gpu/drm/bridge/tc358767.c index 8e28e738cb52..e6403b9549f1 100644 --- a/drivers/gpu/drm/bridge/tc358767.c +++ b/drivers/gpu/drm/bridge/tc358767.c @@ -98,6 +98,8 @@ #define DP0_STARTVAL 0x064c #define DP0_ACTIVEVAL 0x0650 #define DP0_SYNCVAL 0x0654 +#define SYNCVAL_HS_POL_ACTIVE_LOW (1 << 15) +#define SYNCVAL_VS_POL_ACTIVE_LOW (1 << 31) #define DP0_MISC 0x0658 #define TU_SIZE_RECOMMENDED (63) /* LSCLK cycles per TU */ #define BPC_6 (0 << 5) @@ -142,6 +144,8 @@ #define DP0_LTLOOPCTRL 0x06d8 #define DP0_SNKLTCTRL 0x06e4 +#define DP1_SRCCTRL 0x07a0 + /* PHY */ #define DP_PHY_CTRL 0x0800 #define DP_PHY_RST BIT(28) /* DP PHY Global Soft Reset */ @@ -150,6 +154,7 @@ #define PHY_M1_RST BIT(12) /* Reset PHY1 Main Channel */ #define PHY_RDY BIT(16) /* PHY Main Channels Ready */ #define PHY_M0_RST BIT(8) /* Reset PHY0 Main Channel */ +#define PHY_2LANE BIT(2) /* PHY Enable 2 lanes */ #define PHY_A0_EN BIT(1) /* PHY Aux Channel0 Enable */ #define PHY_M0_EN BIT(0) /* PHY Main Channel0 Enable */ @@ -540,6 +545,7 @@ static int tc_aux_link_setup(struct tc_data *tc) unsigned long rate; u32 value; int ret; + u32 dp_phy_ctrl; rate = clk_get_rate(tc->refclk); switch (rate) { @@ -564,7 +570,10 @@ static int tc_aux_link_setup(struct tc_data *tc) value |= SYSCLK_SEL_LSCLK | LSCLK_DIV_2; tc_write(SYS_PLLPARAM, value); - tc_write(DP_PHY_CTRL, BGREN | PWR_SW_EN | BIT(2) | PHY_A0_EN); + dp_phy_ctrl = BGREN | PWR_SW_EN | PHY_A0_EN; + if (tc->link.base.num_lanes == 2) + dp_phy_ctrl |= PHY_2LANE; + tc_write(DP_PHY_CTRL, dp_phy_ctrl); /* * Initially PLLs are in bypass. Force PLL parameter update, @@ -719,7 +728,9 @@ static int tc_set_video_mode(struct tc_data *tc, struct drm_display_mode *mode) tc_write(DP0_ACTIVEVAL, (mode->vdisplay << 16) | (mode->hdisplay)); - tc_write(DP0_SYNCVAL, (vsync_len << 16) | (hsync_len << 0)); + tc_write(DP0_SYNCVAL, (vsync_len << 16) | (hsync_len << 0) | + ((mode->flags & DRM_MODE_FLAG_NHSYNC) ? SYNCVAL_HS_POL_ACTIVE_LOW : 0) | + ((mode->flags & DRM_MODE_FLAG_NVSYNC) ? SYNCVAL_VS_POL_ACTIVE_LOW : 0)); tc_write(DPIPXLFMT, VS_POL_ACTIVE_LOW | HS_POL_ACTIVE_LOW | DE_POL_ACTIVE_HIGH | SUB_CFG_TYPE_CONFIG1 | DPI_BPP_RGB888); @@ -829,12 +840,11 @@ static int tc_main_link_setup(struct tc_data *tc) if (!tc->mode) return -EINVAL; - /* from excel file - DP0_SrcCtrl */ - tc_write(DP0_SRCCTRL, DP0_SRCCTRL_SCRMBLDIS | DP0_SRCCTRL_EN810B | - DP0_SRCCTRL_LANESKEW | DP0_SRCCTRL_LANES_2 | - DP0_SRCCTRL_BW27 | DP0_SRCCTRL_AUTOCORRECT); - /* from excel file - DP1_SrcCtrl */ - tc_write(0x07a0, 0x00003083); + tc_write(DP0_SRCCTRL, tc_srcctrl(tc)); + /* SSCG and BW27 on DP1 must be set to the same as on DP0 */ + tc_write(DP1_SRCCTRL, + (tc->link.spread ? DP0_SRCCTRL_SSCG : 0) | + ((tc->link.base.rate != 162000) ? DP0_SRCCTRL_BW27 : 0)); rate = clk_get_rate(tc->refclk); switch (rate) { @@ -855,8 +865,11 @@ static int tc_main_link_setup(struct tc_data *tc) } value |= SYSCLK_SEL_LSCLK | LSCLK_DIV_2; tc_write(SYS_PLLPARAM, value); + /* Setup Main Link */ - dp_phy_ctrl = BGREN | PWR_SW_EN | BIT(2) | PHY_A0_EN | PHY_M0_EN; + dp_phy_ctrl = BGREN | PWR_SW_EN | PHY_A0_EN | PHY_M0_EN; + if (tc->link.base.num_lanes == 2) + dp_phy_ctrl |= PHY_2LANE; tc_write(DP_PHY_CTRL, dp_phy_ctrl); msleep(100); @@ -1105,10 +1118,20 @@ static bool tc_bridge_mode_fixup(struct drm_bridge *bridge, static enum drm_mode_status tc_connector_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + struct tc_data *tc = connector_to_tc(connector); + u32 req, avail; + u32 bits_per_pixel = 24; + /* DPI interface clock limitation: upto 154 MHz */ if (mode->clock > 154000) return MODE_CLOCK_HIGH; + req = mode->clock * bits_per_pixel / 8; + avail = tc->link.base.num_lanes * tc->link.base.rate; + + if (req > avail) + return MODE_BAD; + return MODE_OK; } @@ -1186,7 +1209,8 @@ static int tc_bridge_attach(struct drm_bridge *bridge) /* Create eDP connector */ drm_connector_helper_add(&tc->connector, &tc_connector_helper_funcs); ret = drm_connector_init(drm, &tc->connector, &tc_connector_funcs, - DRM_MODE_CONNECTOR_eDP); + tc->panel ? DRM_MODE_CONNECTOR_eDP : + DRM_MODE_CONNECTOR_DisplayPort); if (ret) return ret; @@ -1195,6 +1219,10 @@ static int tc_bridge_attach(struct drm_bridge *bridge) drm_display_info_set_bus_formats(&tc->connector.display_info, &bus_format, 1); + tc->connector.display_info.bus_flags = + DRM_BUS_FLAG_DE_HIGH | + DRM_BUS_FLAG_PIXDATA_NEGEDGE | + DRM_BUS_FLAG_SYNC_NEGEDGE; drm_connector_attach_encoder(&tc->connector, tc->bridge.encoder); return 0; diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi86.c b/drivers/gpu/drm/bridge/ti-sn65dsi86.c index 680566d97adc..10243965ee7c 100644 --- a/drivers/gpu/drm/bridge/ti-sn65dsi86.c +++ b/drivers/gpu/drm/bridge/ti-sn65dsi86.c @@ -54,7 +54,7 @@ #define SN_AUX_ADDR_7_0_REG 0x76 #define SN_AUX_LENGTH_REG 0x77 #define SN_AUX_CMD_REG 0x78 -#define AUX_CMD_SEND BIT(1) +#define AUX_CMD_SEND BIT(0) #define AUX_CMD_REQ(x) ((x) << 4) #define SN_AUX_RDATA_REG(x) (0x79 + (x)) #define SN_SSC_CONFIG_REG 0x93 diff --git a/drivers/gpu/drm/drm_atomic_state_helper.c b/drivers/gpu/drm/drm_atomic_state_helper.c index 60bd7d708e35..4985384e51f6 100644 --- a/drivers/gpu/drm/drm_atomic_state_helper.c +++ b/drivers/gpu/drm/drm_atomic_state_helper.c @@ -241,6 +241,7 @@ void __drm_atomic_helper_plane_duplicate_state(struct drm_plane *plane, state->fence = NULL; state->commit = NULL; + state->fb_damage_clips = NULL; } EXPORT_SYMBOL(__drm_atomic_helper_plane_duplicate_state); @@ -285,6 +286,8 @@ void __drm_atomic_helper_plane_destroy_state(struct drm_plane_state *state) if (state->commit) drm_crtc_commit_put(state->commit); + + drm_property_blob_put(state->fb_damage_clips); } EXPORT_SYMBOL(__drm_atomic_helper_plane_destroy_state); diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c index c40889888a16..9a1f41adfc67 100644 --- a/drivers/gpu/drm/drm_atomic_uapi.c +++ b/drivers/gpu/drm/drm_atomic_uapi.c @@ -1296,12 +1296,11 @@ int drm_mode_atomic_ioctl(struct drm_device *dev, (arg->flags & DRM_MODE_PAGE_FLIP_EVENT)) return -EINVAL; - drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE); - state = drm_atomic_state_alloc(dev); if (!state) return -ENOMEM; + drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE); state->acquire_ctx = &ctx; state->allow_modeset = !!(arg->flags & DRM_MODE_ATOMIC_ALLOW_MODESET); diff --git a/drivers/gpu/drm/drm_auth.c b/drivers/gpu/drm/drm_auth.c index d9c0f7573905..1669c42c40ed 100644 --- a/drivers/gpu/drm/drm_auth.c +++ b/drivers/gpu/drm/drm_auth.c @@ -142,6 +142,7 @@ static int drm_new_set_master(struct drm_device *dev, struct drm_file *fpriv) lockdep_assert_held_once(&dev->master_mutex); + WARN_ON(fpriv->is_master); old_master = fpriv->master; fpriv->master = drm_master_create(dev); if (!fpriv->master) { @@ -170,6 +171,7 @@ out_err: /* drop references and restore old master on failure */ drm_master_put(&fpriv->master); fpriv->master = old_master; + fpriv->is_master = 0; return ret; } diff --git a/drivers/gpu/drm/drm_damage_helper.c b/drivers/gpu/drm/drm_damage_helper.c index 05c8e7267165..31032407254d 100644 --- a/drivers/gpu/drm/drm_damage_helper.c +++ b/drivers/gpu/drm/drm_damage_helper.c @@ -178,7 +178,7 @@ int drm_atomic_helper_dirtyfb(struct drm_framebuffer *fb, state = drm_atomic_state_alloc(fb->dev); if (!state) { ret = -ENOMEM; - goto out; + goto out_drop_locks; } state->acquire_ctx = &ctx; @@ -238,6 +238,7 @@ out: kfree(rects); drm_atomic_state_put(state); +out_drop_locks: drm_modeset_drop_locks(&ctx); drm_modeset_acquire_fini(&ctx); @@ -250,7 +251,7 @@ EXPORT_SYMBOL(drm_atomic_helper_dirtyfb); * drm_atomic_helper_damage_iter_init - Initialize the damage iterator. * @iter: The iterator to initialize. * @old_state: Old plane state for validation. - * @new_state: Plane state from which to iterate the damage clips. + * @state: Plane state from which to iterate the damage clips. * * Initialize an iterator, which clips plane damage * &drm_plane_state.fb_damage_clips to plane &drm_plane_state.src. This iterator diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 2d6c491a0542..516e82d0ed50 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -1273,6 +1273,8 @@ static const struct dpcd_quirk dpcd_quirk_list[] = { { OUI(0x00, 0x22, 0xb9), DEVICE_ID_ANY, true, BIT(DP_DPCD_QUIRK_CONSTANT_N) }, /* LG LP140WF6-SPM1 eDP panel */ { OUI(0x00, 0x22, 0xb9), DEVICE_ID('s', 'i', 'v', 'a', 'r', 'T'), false, BIT(DP_DPCD_QUIRK_CONSTANT_N) }, + /* Apple panels need some additional handling to support PSR */ + { OUI(0x00, 0x10, 0xfa), DEVICE_ID_ANY, false, BIT(DP_DPCD_QUIRK_NO_PSR) } }; #undef OUI diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index 5e9ca6f96379..d73703a695e8 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -71,7 +71,7 @@ MODULE_PARM_DESC(drm_fbdev_overalloc, #if IS_ENABLED(CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM) static bool drm_leak_fbdev_smem = false; module_param_unsafe(drm_leak_fbdev_smem, bool, 0600); -MODULE_PARM_DESC(fbdev_emulation, +MODULE_PARM_DESC(drm_leak_fbdev_smem, "Allow unsafe leaking fbdev physical smem address [default=false]"); #endif @@ -1621,6 +1621,64 @@ static bool drm_fb_pixel_format_equal(const struct fb_var_screeninfo *var_1, var_1->transp.msb_right == var_2->transp.msb_right; } +static void drm_fb_helper_fill_pixel_fmt(struct fb_var_screeninfo *var, + u8 depth) +{ + switch (depth) { + case 8: + var->red.offset = 0; + var->green.offset = 0; + var->blue.offset = 0; + var->red.length = 8; /* 8bit DAC */ + var->green.length = 8; + var->blue.length = 8; + var->transp.offset = 0; + var->transp.length = 0; + break; + case 15: + var->red.offset = 10; + var->green.offset = 5; + var->blue.offset = 0; + var->red.length = 5; + var->green.length = 5; + var->blue.length = 5; + var->transp.offset = 15; + var->transp.length = 1; + break; + case 16: + var->red.offset = 11; + var->green.offset = 5; + var->blue.offset = 0; + var->red.length = 5; + var->green.length = 6; + var->blue.length = 5; + var->transp.offset = 0; + break; + case 24: + var->red.offset = 16; + var->green.offset = 8; + var->blue.offset = 0; + var->red.length = 8; + var->green.length = 8; + var->blue.length = 8; + var->transp.offset = 0; + var->transp.length = 0; + break; + case 32: + var->red.offset = 16; + var->green.offset = 8; + var->blue.offset = 0; + var->red.length = 8; + var->green.length = 8; + var->blue.length = 8; + var->transp.offset = 24; + var->transp.length = 8; + break; + default: + break; + } +} + /** * drm_fb_helper_check_var - implementation for &fb_ops.fb_check_var * @var: screeninfo to check @@ -1632,9 +1690,14 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var, struct drm_fb_helper *fb_helper = info->par; struct drm_framebuffer *fb = fb_helper->fb; - if (var->pixclock != 0 || in_dbg_master()) + if (in_dbg_master()) return -EINVAL; + if (var->pixclock != 0) { + DRM_DEBUG("fbdev emulation doesn't support changing the pixel clock, value of pixclock is ignored\n"); + var->pixclock = 0; + } + if ((drm_format_info_block_width(fb->format, 0) > 1) || (drm_format_info_block_height(fb->format, 0) > 1)) return -EINVAL; @@ -1655,6 +1718,20 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var, } /* + * Workaround for SDL 1.2, which is known to be setting all pixel format + * fields values to zero in some cases. We treat this situation as a + * kind of "use some reasonable autodetected values". + */ + if (!var->red.offset && !var->green.offset && + !var->blue.offset && !var->transp.offset && + !var->red.length && !var->green.length && + !var->blue.length && !var->transp.length && + !var->red.msb_right && !var->green.msb_right && + !var->blue.msb_right && !var->transp.msb_right) { + drm_fb_helper_fill_pixel_fmt(var, fb->format->depth); + } + + /* * drm fbdev emulation doesn't support changing the pixel format at all, * so reject all pixel format changing requests. */ @@ -1967,59 +2044,7 @@ void drm_fb_helper_fill_var(struct fb_info *info, struct drm_fb_helper *fb_helpe info->var.yoffset = 0; info->var.activate = FB_ACTIVATE_NOW; - switch (fb->format->depth) { - case 8: - info->var.red.offset = 0; - info->var.green.offset = 0; - info->var.blue.offset = 0; - info->var.red.length = 8; /* 8bit DAC */ - info->var.green.length = 8; - info->var.blue.length = 8; - info->var.transp.offset = 0; - info->var.transp.length = 0; - break; - case 15: - info->var.red.offset = 10; - info->var.green.offset = 5; - info->var.blue.offset = 0; - info->var.red.length = 5; - info->var.green.length = 5; - info->var.blue.length = 5; - info->var.transp.offset = 15; - info->var.transp.length = 1; - break; - case 16: - info->var.red.offset = 11; - info->var.green.offset = 5; - info->var.blue.offset = 0; - info->var.red.length = 5; - info->var.green.length = 6; - info->var.blue.length = 5; - info->var.transp.offset = 0; - break; - case 24: - info->var.red.offset = 16; - info->var.green.offset = 8; - info->var.blue.offset = 0; - info->var.red.length = 8; - info->var.green.length = 8; - info->var.blue.length = 8; - info->var.transp.offset = 0; - info->var.transp.length = 0; - break; - case 32: - info->var.red.offset = 16; - info->var.green.offset = 8; - info->var.blue.offset = 0; - info->var.red.length = 8; - info->var.green.length = 8; - info->var.blue.length = 8; - info->var.transp.offset = 24; - info->var.transp.length = 8; - break; - default: - break; - } + drm_fb_helper_fill_pixel_fmt(&info->var, fb->format->depth); info->var.xres = fb_width; info->var.yres = fb_height; diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c index ffa8dc35515f..46f48f245eb5 100644 --- a/drivers/gpu/drm/drm_file.c +++ b/drivers/gpu/drm/drm_file.c @@ -525,7 +525,7 @@ ssize_t drm_read(struct file *filp, char __user *buffer, struct drm_device *dev = file_priv->minor->dev; ssize_t ret; - if (!access_ok(VERIFY_WRITE, buffer, count)) + if (!access_ok(buffer, count)) return -EFAULT; ret = mutex_lock_interruptible(&file_priv->event_read_lock); diff --git a/drivers/gpu/drm/drm_internal.h b/drivers/gpu/drm/drm_internal.h index c7a7d7ce5d1c..d9caf205e0b3 100644 --- a/drivers/gpu/drm/drm_internal.h +++ b/drivers/gpu/drm/drm_internal.h @@ -99,6 +99,8 @@ struct device *drm_sysfs_minor_alloc(struct drm_minor *minor); int drm_sysfs_connector_add(struct drm_connector *connector); void drm_sysfs_connector_remove(struct drm_connector *connector); +void drm_sysfs_lease_event(struct drm_device *dev); + /* drm_gem.c */ int drm_gem_init(struct drm_device *dev); void drm_gem_destroy(struct drm_device *dev); diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c index 94bd872d56c4..7e6746b2d704 100644 --- a/drivers/gpu/drm/drm_ioctl.c +++ b/drivers/gpu/drm/drm_ioctl.c @@ -37,6 +37,7 @@ #include <linux/pci.h> #include <linux/export.h> +#include <linux/nospec.h> /** * DOC: getunique and setversion story @@ -800,13 +801,17 @@ long drm_ioctl(struct file *filp, if (is_driver_ioctl) { /* driver ioctl */ - if (nr - DRM_COMMAND_BASE >= dev->driver->num_ioctls) + unsigned int index = nr - DRM_COMMAND_BASE; + + if (index >= dev->driver->num_ioctls) goto err_i1; - ioctl = &dev->driver->ioctls[nr - DRM_COMMAND_BASE]; + index = array_index_nospec(index, dev->driver->num_ioctls); + ioctl = &dev->driver->ioctls[index]; } else { /* core ioctl */ if (nr >= DRM_CORE_IOCTL_COUNT) goto err_i1; + nr = array_index_nospec(nr, DRM_CORE_IOCTL_COUNT); ioctl = &drm_ioctls[nr]; } @@ -888,6 +893,7 @@ bool drm_ioctl_flags(unsigned int nr, unsigned int *flags) if (nr >= DRM_CORE_IOCTL_COUNT) return false; + nr = array_index_nospec(nr, DRM_CORE_IOCTL_COUNT); *flags = drm_ioctls[nr].flags; return true; diff --git a/drivers/gpu/drm/drm_lease.c b/drivers/gpu/drm/drm_lease.c index 3650d3c46718..99cba8ea5d82 100644 --- a/drivers/gpu/drm/drm_lease.c +++ b/drivers/gpu/drm/drm_lease.c @@ -292,7 +292,7 @@ void drm_lease_destroy(struct drm_master *master) if (master->lessor) { /* Tell the master to check the lessee list */ - drm_sysfs_hotplug_event(dev); + drm_sysfs_lease_event(dev); drm_master_put(&master->lessor); } diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c index cd9bc0ce9be0..004191d01772 100644 --- a/drivers/gpu/drm/drm_mode_object.c +++ b/drivers/gpu/drm/drm_mode_object.c @@ -459,11 +459,11 @@ static int set_property_atomic(struct drm_mode_object *obj, struct drm_modeset_acquire_ctx ctx; int ret; - drm_modeset_acquire_init(&ctx, 0); - state = drm_atomic_state_alloc(dev); if (!state) return -ENOMEM; + + drm_modeset_acquire_init(&ctx, 0); state->acquire_ctx = &ctx; retry: if (prop == state->dev->mode_config.dpms_property) { diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c index a9d9df6c85ad..693748ad8b88 100644 --- a/drivers/gpu/drm/drm_pci.c +++ b/drivers/gpu/drm/drm_pci.c @@ -61,8 +61,9 @@ drm_dma_handle_t *drm_pci_alloc(struct drm_device * dev, size_t size, size_t ali return NULL; dmah->size = size; - dmah->vaddr = dma_zalloc_coherent(&dev->pdev->dev, size, &dmah->busaddr, - GFP_KERNEL | __GFP_COMP); + dmah->vaddr = dma_alloc_coherent(&dev->pdev->dev, size, + &dmah->busaddr, + GFP_KERNEL | __GFP_COMP); if (dmah->vaddr == NULL) { kfree(dmah); diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c index b3c1daad1169..ecb7b33002bb 100644 --- a/drivers/gpu/drm/drm_sysfs.c +++ b/drivers/gpu/drm/drm_sysfs.c @@ -301,6 +301,16 @@ void drm_sysfs_connector_remove(struct drm_connector *connector) connector->kdev = NULL; } +void drm_sysfs_lease_event(struct drm_device *dev) +{ + char *event_string = "LEASE=1"; + char *envp[] = { event_string, NULL }; + + DRM_DEBUG("generating lease event\n"); + + kobject_uevent_env(&dev->primary->kdev->kobj, KOBJ_CHANGE, envp); +} + /** * drm_sysfs_hotplug_event - generate a DRM uevent * @dev: DRM device diff --git a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c index 7fea74861a87..160ce3c060a5 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c @@ -439,6 +439,4 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state, if (drm_debug & DRM_UT_DRIVER) etnaviv_buffer_dump(gpu, buffer, 0, 0x50); - - gpu->lastctx = cmdbuf->ctx; } diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.c b/drivers/gpu/drm/etnaviv/etnaviv_drv.c index 52802e6049e0..18c27f795cf6 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_drv.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.c @@ -72,14 +72,8 @@ static void etnaviv_postclose(struct drm_device *dev, struct drm_file *file) for (i = 0; i < ETNA_MAX_PIPES; i++) { struct etnaviv_gpu *gpu = priv->gpu[i]; - if (gpu) { - mutex_lock(&gpu->lock); - if (gpu->lastctx == ctx) - gpu->lastctx = NULL; - mutex_unlock(&gpu->lock); - + if (gpu) drm_sched_entity_destroy(&ctx->sched_entity[i]); - } } kfree(ctx); @@ -345,7 +339,6 @@ static int etnaviv_ioctl_gem_userptr(struct drm_device *dev, void *data, struct drm_file *file) { struct drm_etnaviv_gem_userptr *args = data; - int access; if (args->flags & ~(ETNA_USERPTR_READ|ETNA_USERPTR_WRITE) || args->flags == 0) @@ -357,12 +350,7 @@ static int etnaviv_ioctl_gem_userptr(struct drm_device *dev, void *data, args->user_ptr & ~PAGE_MASK) return -EINVAL; - if (args->flags & ETNA_USERPTR_WRITE) - access = VERIFY_WRITE; - else - access = VERIFY_READ; - - if (!access_ok(access, (void __user *)(unsigned long)args->user_ptr, + if (!access_ok((void __user *)(unsigned long)args->user_ptr, args->user_size)) return -EFAULT; @@ -523,7 +511,7 @@ static int etnaviv_bind(struct device *dev) if (!priv) { dev_err(dev, "failed to allocate private data\n"); ret = -ENOMEM; - goto out_unref; + goto out_put; } drm->dev_private = priv; @@ -549,7 +537,7 @@ out_register: component_unbind_all(dev, drm); out_bind: kfree(priv); -out_unref: +out_put: drm_dev_put(drm); return ret; diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.h b/drivers/gpu/drm/etnaviv/etnaviv_drv.h index 8d02d1b7dcf5..4bf698de5996 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_drv.h +++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.h @@ -107,17 +107,6 @@ static inline size_t size_vstruct(size_t nelem, size_t elem_size, size_t base) return base + nelem * elem_size; } -/* returns true if fence a comes after fence b */ -static inline bool fence_after(u32 a, u32 b) -{ - return (s32)(a - b) > 0; -} - -static inline bool fence_after_eq(u32 a, u32 b) -{ - return (s32)(a - b) >= 0; -} - /* * Etnaviv timeouts are specified wrt CLOCK_MONOTONIC, not jiffies. * We need to calculate the timeout in terms of number of jiffies diff --git a/drivers/gpu/drm/etnaviv/etnaviv_dump.c b/drivers/gpu/drm/etnaviv/etnaviv_dump.c index 9146e30e24a6..3fbb4855396c 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_dump.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_dump.c @@ -118,6 +118,7 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) unsigned int n_obj, n_bomap_pages; size_t file_size, mmu_size; __le64 *bomap, *bomap_start; + unsigned long flags; /* Only catch the first event, or when manually re-armed */ if (!etnaviv_dump_core) @@ -134,13 +135,13 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) mmu_size + gpu->buffer.size; /* Add in the active command buffers */ - spin_lock(&gpu->sched.job_list_lock); + spin_lock_irqsave(&gpu->sched.job_list_lock, flags); list_for_each_entry(s_job, &gpu->sched.ring_mirror_list, node) { submit = to_etnaviv_submit(s_job); file_size += submit->cmdbuf.size; n_obj++; } - spin_unlock(&gpu->sched.job_list_lock); + spin_unlock_irqrestore(&gpu->sched.job_list_lock, flags); /* Add in the active buffer objects */ list_for_each_entry(vram, &gpu->mmu->mappings, mmu_node) { @@ -182,14 +183,14 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) gpu->buffer.size, etnaviv_cmdbuf_get_va(&gpu->buffer)); - spin_lock(&gpu->sched.job_list_lock); + spin_lock_irqsave(&gpu->sched.job_list_lock, flags); list_for_each_entry(s_job, &gpu->sched.ring_mirror_list, node) { submit = to_etnaviv_submit(s_job); etnaviv_core_dump_mem(&iter, ETDUMP_BUF_CMD, submit->cmdbuf.vaddr, submit->cmdbuf.size, etnaviv_cmdbuf_get_va(&submit->cmdbuf)); } - spin_unlock(&gpu->sched.job_list_lock); + spin_unlock_irqrestore(&gpu->sched.job_list_lock, flags); /* Reserve space for the bomap */ if (n_bomap_pages) { diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c index f225fbc6edd2..6904535475de 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c @@ -3,10 +3,12 @@ * Copyright (C) 2015-2018 Etnaviv Project */ +#include <linux/clk.h> #include <linux/component.h> #include <linux/dma-fence.h> #include <linux/moduleparam.h> #include <linux/of_device.h> +#include <linux/regulator/consumer.h> #include <linux/thermal.h> #include "etnaviv_cmdbuf.h" @@ -976,7 +978,6 @@ int etnaviv_gpu_debugfs(struct etnaviv_gpu *gpu, struct seq_file *m) void etnaviv_gpu_recover_hang(struct etnaviv_gpu *gpu) { - unsigned long flags; unsigned int i = 0; dev_err(gpu->dev, "recover hung GPU!\n"); @@ -989,15 +990,13 @@ void etnaviv_gpu_recover_hang(struct etnaviv_gpu *gpu) etnaviv_hw_reset(gpu); /* complete all events, the GPU won't do it after the reset */ - spin_lock_irqsave(&gpu->event_spinlock, flags); + spin_lock(&gpu->event_spinlock); for_each_set_bit_from(i, gpu->event_bitmap, ETNA_NR_EVENTS) 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; + spin_unlock(&gpu->event_spinlock); etnaviv_gpu_hw_init(gpu); - gpu->lastctx = NULL; gpu->exec_state = -1; mutex_unlock(&gpu->lock); @@ -1032,7 +1031,7 @@ static bool etnaviv_fence_signaled(struct dma_fence *fence) { struct etnaviv_fence *f = to_etnaviv_fence(fence); - return fence_completed(f->gpu, f->base.seqno); + return (s32)(f->gpu->completed_fence - f->base.seqno) >= 0; } static void etnaviv_fence_release(struct dma_fence *fence) @@ -1071,6 +1070,12 @@ static struct dma_fence *etnaviv_gpu_fence_alloc(struct etnaviv_gpu *gpu) return &f->base; } +/* returns true if fence a comes after fence b */ +static inline bool fence_after(u32 a, u32 b) +{ + return (s32)(a - b) > 0; +} + /* * event management: */ @@ -1078,7 +1083,7 @@ static struct dma_fence *etnaviv_gpu_fence_alloc(struct etnaviv_gpu *gpu) static int event_alloc(struct etnaviv_gpu *gpu, unsigned nr_events, unsigned int *events) { - unsigned long flags, timeout = msecs_to_jiffies(10 * 10000); + unsigned long timeout = msecs_to_jiffies(10 * 10000); unsigned i, acquired = 0; for (i = 0; i < nr_events; i++) { @@ -1095,7 +1100,7 @@ static int event_alloc(struct etnaviv_gpu *gpu, unsigned nr_events, timeout = ret; } - spin_lock_irqsave(&gpu->event_spinlock, flags); + spin_lock(&gpu->event_spinlock); for (i = 0; i < nr_events; i++) { int event = find_first_zero_bit(gpu->event_bitmap, ETNA_NR_EVENTS); @@ -1105,7 +1110,7 @@ static int event_alloc(struct etnaviv_gpu *gpu, unsigned nr_events, set_bit(event, gpu->event_bitmap); } - spin_unlock_irqrestore(&gpu->event_spinlock, flags); + spin_unlock(&gpu->event_spinlock); return 0; @@ -1118,18 +1123,11 @@ out: static void event_free(struct etnaviv_gpu *gpu, unsigned int event) { - unsigned long flags; - - spin_lock_irqsave(&gpu->event_spinlock, flags); - 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 { clear_bit(event, gpu->event_bitmap); - spin_unlock_irqrestore(&gpu->event_spinlock, flags); - complete(&gpu->event_free); } } @@ -1306,8 +1304,6 @@ struct dma_fence *etnaviv_gpu_submit(struct etnaviv_gem_submit *submit) goto out_unlock; } - gpu->active_fence = gpu_fence->seqno; - if (submit->nr_pmrs) { gpu->event[event[1]].sync_point = &sync_point_perfmon_sample_pre; kref_get(&submit->refcount); @@ -1549,7 +1545,6 @@ static int etnaviv_gpu_hw_resume(struct etnaviv_gpu *gpu) etnaviv_gpu_update_clock(gpu); etnaviv_gpu_hw_init(gpu); - gpu->lastctx = NULL; gpu->exec_state = -1; mutex_unlock(&gpu->lock); @@ -1806,8 +1801,8 @@ static int etnaviv_gpu_rpm_suspend(struct device *dev) struct etnaviv_gpu *gpu = dev_get_drvdata(dev); u32 idle, mask; - /* If we have outstanding fences, we're not idle */ - if (gpu->completed_fence != gpu->active_fence) + /* If there are any jobs in the HW queue, we're not idle */ + if (atomic_read(&gpu->sched.hw_rq_count)) return -EBUSY; /* Check whether the hardware (except FE) is idle */ diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h index 9a75a6937268..9bcf151f706b 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h @@ -6,9 +6,6 @@ #ifndef __ETNAVIV_GPU_H__ #define __ETNAVIV_GPU_H__ -#include <linux/clk.h> -#include <linux/regulator/consumer.h> - #include "etnaviv_cmdbuf.h" #include "etnaviv_drv.h" @@ -88,6 +85,8 @@ struct etnaviv_event { struct etnaviv_cmdbuf_suballoc; struct etnaviv_cmdbuf; +struct regulator; +struct clk; #define ETNA_NR_EVENTS 30 @@ -98,7 +97,6 @@ struct etnaviv_gpu { struct mutex lock; struct etnaviv_chip_identity identity; enum etnaviv_sec_mode sec_mode; - struct etnaviv_file_private *lastctx; struct workqueue_struct *wq; struct drm_gpu_scheduler sched; @@ -121,7 +119,6 @@ struct etnaviv_gpu { struct mutex fence_lock; struct idr fence_idr; u32 next_fence; - u32 active_fence; u32 completed_fence; wait_queue_head_t fence_event; u64 fence_context; @@ -161,11 +158,6 @@ static inline u32 gpu_read(struct etnaviv_gpu *gpu, u32 reg) return readl(gpu->mmio + reg); } -static inline bool fence_completed(struct etnaviv_gpu *gpu, u32 fence) -{ - return fence_after_eq(gpu->completed_fence, fence); -} - int etnaviv_gpu_get_param(struct etnaviv_gpu *gpu, u32 param, u64 *value); int etnaviv_gpu_init(struct etnaviv_gpu *gpu); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c index e3d6a8584715..786a8ee6f10f 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c @@ -228,6 +228,21 @@ static const uint32_t fimd_formats[] = { DRM_FORMAT_ARGB8888, }; +static const unsigned int capabilities[WINDOWS_NR] = { + 0, + EXYNOS_DRM_PLANE_CAP_WIN_BLEND | EXYNOS_DRM_PLANE_CAP_PIX_BLEND, + EXYNOS_DRM_PLANE_CAP_WIN_BLEND | EXYNOS_DRM_PLANE_CAP_PIX_BLEND, + EXYNOS_DRM_PLANE_CAP_WIN_BLEND | EXYNOS_DRM_PLANE_CAP_PIX_BLEND, + EXYNOS_DRM_PLANE_CAP_WIN_BLEND | EXYNOS_DRM_PLANE_CAP_PIX_BLEND, +}; + +static inline void fimd_set_bits(struct fimd_context *ctx, u32 reg, u32 mask, + u32 val) +{ + val = (val & mask) | (readl(ctx->regs + reg) & ~mask); + writel(val, ctx->regs + reg); +} + static int fimd_enable_vblank(struct exynos_drm_crtc *crtc) { struct fimd_context *ctx = crtc->ctx; @@ -551,13 +566,88 @@ static void fimd_commit(struct exynos_drm_crtc *crtc) writel(val, ctx->regs + VIDCON0); } +static void fimd_win_set_bldeq(struct fimd_context *ctx, unsigned int win, + unsigned int alpha, unsigned int pixel_alpha) +{ + u32 mask = BLENDEQ_A_FUNC_F(0xf) | BLENDEQ_B_FUNC_F(0xf); + u32 val = 0; + + switch (pixel_alpha) { + case DRM_MODE_BLEND_PIXEL_NONE: + case DRM_MODE_BLEND_COVERAGE: + val |= BLENDEQ_A_FUNC_F(BLENDEQ_ALPHA_A); + val |= BLENDEQ_B_FUNC_F(BLENDEQ_ONE_MINUS_ALPHA_A); + break; + case DRM_MODE_BLEND_PREMULTI: + default: + if (alpha != DRM_BLEND_ALPHA_OPAQUE) { + val |= BLENDEQ_A_FUNC_F(BLENDEQ_ALPHA0); + val |= BLENDEQ_B_FUNC_F(BLENDEQ_ONE_MINUS_ALPHA_A); + } else { + val |= BLENDEQ_A_FUNC_F(BLENDEQ_ONE); + val |= BLENDEQ_B_FUNC_F(BLENDEQ_ONE_MINUS_ALPHA_A); + } + break; + } + fimd_set_bits(ctx, BLENDEQx(win), mask, val); +} -static void fimd_win_set_pixfmt(struct fimd_context *ctx, unsigned int win, - uint32_t pixel_format, int width) +static void fimd_win_set_bldmod(struct fimd_context *ctx, unsigned int win, + unsigned int alpha, unsigned int pixel_alpha) { - unsigned long val; + u32 win_alpha_l = (alpha >> 8) & 0xf; + u32 win_alpha_h = alpha >> 12; + u32 val = 0; - val = WINCONx_ENWIN; + switch (pixel_alpha) { + case DRM_MODE_BLEND_PIXEL_NONE: + break; + case DRM_MODE_BLEND_COVERAGE: + case DRM_MODE_BLEND_PREMULTI: + default: + val |= WINCON1_ALPHA_SEL; + val |= WINCON1_BLD_PIX; + val |= WINCON1_ALPHA_MUL; + break; + } + fimd_set_bits(ctx, WINCON(win), WINCONx_BLEND_MODE_MASK, val); + + /* OSD alpha */ + val = VIDISD14C_ALPHA0_R(win_alpha_h) | + VIDISD14C_ALPHA0_G(win_alpha_h) | + VIDISD14C_ALPHA0_B(win_alpha_h) | + VIDISD14C_ALPHA1_R(0x0) | + VIDISD14C_ALPHA1_G(0x0) | + VIDISD14C_ALPHA1_B(0x0); + writel(val, ctx->regs + VIDOSD_C(win)); + + val = VIDW_ALPHA_R(win_alpha_l) | VIDW_ALPHA_G(win_alpha_l) | + VIDW_ALPHA_B(win_alpha_l); + writel(val, ctx->regs + VIDWnALPHA0(win)); + + val = VIDW_ALPHA_R(0x0) | VIDW_ALPHA_G(0x0) | + VIDW_ALPHA_B(0x0); + writel(val, ctx->regs + VIDWnALPHA1(win)); + + fimd_set_bits(ctx, BLENDCON, BLENDCON_NEW_MASK, + BLENDCON_NEW_8BIT_ALPHA_VALUE); +} + +static void fimd_win_set_pixfmt(struct fimd_context *ctx, unsigned int win, + struct drm_framebuffer *fb, int width) +{ + struct exynos_drm_plane plane = ctx->planes[win]; + struct exynos_drm_plane_state *state = + to_exynos_plane_state(plane.base.state); + uint32_t pixel_format = fb->format->format; + unsigned int alpha = state->base.alpha; + u32 val = WINCONx_ENWIN; + unsigned int pixel_alpha; + + if (fb->format->has_alpha) + pixel_alpha = state->base.pixel_blend_mode; + else + pixel_alpha = DRM_MODE_BLEND_PIXEL_NONE; /* * In case of s3c64xx, window 0 doesn't support alpha channel. @@ -591,8 +681,7 @@ static void fimd_win_set_pixfmt(struct fimd_context *ctx, unsigned int win, break; case DRM_FORMAT_ARGB8888: default: - val |= WINCON1_BPPMODE_25BPP_A1888 - | WINCON1_BLD_PIX | WINCON1_ALPHA_SEL; + val |= WINCON1_BPPMODE_25BPP_A1888; val |= WINCONx_WSWP; val |= WINCONx_BURSTLEN_16WORD; break; @@ -610,25 +699,12 @@ static void fimd_win_set_pixfmt(struct fimd_context *ctx, unsigned int win, val &= ~WINCONx_BURSTLEN_MASK; val |= WINCONx_BURSTLEN_4WORD; } - - writel(val, ctx->regs + WINCON(win)); + fimd_set_bits(ctx, WINCON(win), ~WINCONx_BLEND_MODE_MASK, val); /* hardware window 0 doesn't support alpha channel. */ if (win != 0) { - /* OSD alpha */ - val = VIDISD14C_ALPHA0_R(0xf) | - VIDISD14C_ALPHA0_G(0xf) | - VIDISD14C_ALPHA0_B(0xf) | - VIDISD14C_ALPHA1_R(0xf) | - VIDISD14C_ALPHA1_G(0xf) | - VIDISD14C_ALPHA1_B(0xf); - - writel(val, ctx->regs + VIDOSD_C(win)); - - val = VIDW_ALPHA_R(0xf) | VIDW_ALPHA_G(0xf) | - VIDW_ALPHA_G(0xf); - writel(val, ctx->regs + VIDWnALPHA0(win)); - writel(val, ctx->regs + VIDWnALPHA1(win)); + fimd_win_set_bldmod(ctx, win, alpha, pixel_alpha); + fimd_win_set_bldeq(ctx, win, alpha, pixel_alpha); } } @@ -785,7 +861,7 @@ static void fimd_update_plane(struct exynos_drm_crtc *crtc, DRM_DEBUG_KMS("osd size = 0x%x\n", (unsigned int)val); } - fimd_win_set_pixfmt(ctx, win, fb->format->format, state->src.w); + fimd_win_set_pixfmt(ctx, win, fb, state->src.w); /* hardware window 0 doesn't support color key. */ if (win != 0) @@ -987,6 +1063,7 @@ static int fimd_bind(struct device *dev, struct device *master, void *data) ctx->configs[i].num_pixel_formats = ARRAY_SIZE(fimd_formats); ctx->configs[i].zpos = i; ctx->configs[i].type = fimd_win_types[i]; + ctx->configs[i].capabilities = capabilities[i]; ret = exynos_plane_init(drm_dev, &ctx->planes[i], i, &ctx->configs[i]); if (ret) diff --git a/drivers/gpu/drm/i915/Kconfig b/drivers/gpu/drm/i915/Kconfig index 33a458b7f1fc..148be8e1a090 100644 --- a/drivers/gpu/drm/i915/Kconfig +++ b/drivers/gpu/drm/i915/Kconfig @@ -131,5 +131,5 @@ config DRM_I915_GVT_KVMGT menu "drm/i915 Debugging" depends on DRM_I915 depends on EXPERT -source drivers/gpu/drm/i915/Kconfig.debug +source "drivers/gpu/drm/i915/Kconfig.debug" endmenu diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index fe754022e356..359d37d5c958 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -61,10 +61,12 @@ static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm) } mutex_lock(&dev_priv->drm.struct_mutex); + mmio_hw_access_pre(dev_priv); ret = i915_gem_gtt_insert(&dev_priv->ggtt.vm, node, size, I915_GTT_PAGE_SIZE, I915_COLOR_UNEVICTABLE, start, end, flags); + mmio_hw_access_post(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); if (ret) gvt_err("fail to alloc %s gm space from host\n", diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c index 77edbfcb0f75..77ae634eb11c 100644 --- a/drivers/gpu/drm/i915/gvt/cmd_parser.c +++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c @@ -1900,11 +1900,11 @@ static struct cmd_info cmd_info[] = { {"MI_URB_CLEAR", OP_MI_URB_CLEAR, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL}, - {"ME_SEMAPHORE_SIGNAL", OP_MI_SEMAPHORE_SIGNAL, F_LEN_VAR, R_ALL, + {"MI_SEMAPHORE_SIGNAL", OP_MI_SEMAPHORE_SIGNAL, F_LEN_VAR, R_ALL, D_BDW_PLUS, 0, 8, NULL}, - {"ME_SEMAPHORE_WAIT", OP_MI_SEMAPHORE_WAIT, F_LEN_VAR, R_ALL, D_BDW_PLUS, - ADDR_FIX_1(2), 8, cmd_handler_mi_semaphore_wait}, + {"MI_SEMAPHORE_WAIT", OP_MI_SEMAPHORE_WAIT, F_LEN_VAR, R_ALL, + D_BDW_PLUS, ADDR_FIX_1(2), 8, cmd_handler_mi_semaphore_wait}, {"MI_STORE_DATA_IMM", OP_MI_STORE_DATA_IMM, F_LEN_VAR, R_ALL, D_BDW_PLUS, ADDR_FIX_1(1), 10, cmd_handler_mi_store_data_imm}, diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.c b/drivers/gpu/drm/i915/gvt/fb_decoder.c index 481896fb712a..85e6736f0a32 100644 --- a/drivers/gpu/drm/i915/gvt/fb_decoder.c +++ b/drivers/gpu/drm/i915/gvt/fb_decoder.c @@ -235,7 +235,7 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu, plane->bpp = skl_pixel_formats[fmt].bpp; plane->drm_format = skl_pixel_formats[fmt].drm_format; } else { - plane->tiled = !!(val & DISPPLANE_TILED); + plane->tiled = val & DISPPLANE_TILED; fmt = bdw_format_to_drm(val & DISPPLANE_PIXFORMAT_MASK); plane->bpp = bdw_pixel_formats[fmt].bpp; plane->drm_format = bdw_pixel_formats[fmt].drm_format; diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index 58e166effa45..c7103dd2d8d5 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -2447,10 +2447,11 @@ static void intel_vgpu_destroy_all_ppgtt_mm(struct intel_vgpu *vgpu) static void intel_vgpu_destroy_ggtt_mm(struct intel_vgpu *vgpu) { - struct intel_gvt_partial_pte *pos; + struct intel_gvt_partial_pte *pos, *next; - list_for_each_entry(pos, - &vgpu->gtt.ggtt_mm->ggtt_mm.partial_pte_list, list) { + list_for_each_entry_safe(pos, next, + &vgpu->gtt.ggtt_mm->ggtt_mm.partial_pte_list, + list) { gvt_dbg_mm("partial PTE update on hold 0x%lx : 0x%llx\n", pos->offset, pos->data); kfree(pos); diff --git a/drivers/gpu/drm/i915/gvt/gvt.c b/drivers/gpu/drm/i915/gvt/gvt.c index 6ef5a7fc70df..733a2a0d0c30 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.c +++ b/drivers/gpu/drm/i915/gvt/gvt.c @@ -437,7 +437,7 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) ret = intel_gvt_debugfs_init(gvt); if (ret) - gvt_err("debugfs registeration failed, go on.\n"); + gvt_err("debugfs registration failed, go on.\n"); gvt_dbg_core("gvt device initialization is done\n"); dev_priv->gvt = gvt; diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index 31f6cdbe5c42..b4ab1dad0143 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -159,6 +159,10 @@ struct intel_vgpu_submission { struct kmem_cache *workloads; atomic_t running_workload_num; struct i915_gem_context *shadow_ctx; + union { + u64 i915_context_pml4; + u64 i915_context_pdps[GEN8_3LVL_PDPES]; + }; DECLARE_BITMAP(shadow_ctx_desc_updated, I915_NUM_ENGINES); DECLARE_BITMAP(tlb_handle_pending, I915_NUM_ENGINES); void *ring_scan_buffer[I915_NUM_ENGINES]; diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index aa280bb07125..e9f343b124b0 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -475,6 +475,7 @@ static i915_reg_t force_nonpriv_white_list[] = { _MMIO(0x7704), _MMIO(0x7708), _MMIO(0x770c), + _MMIO(0x83a8), _MMIO(0xb110), GEN8_L3SQCREG4,//_MMIO(0xb118) _MMIO(0xe100), @@ -2798,6 +2799,7 @@ static int init_broadwell_mmio_info(struct intel_gvt *gvt) MMIO_DFH(_MMIO(0xe2a0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL); MMIO_DFH(_MMIO(0xe2b0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL); MMIO_DFH(_MMIO(0xe2c0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL); + MMIO_DFH(_MMIO(0x21f0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL); return 0; } diff --git a/drivers/gpu/drm/i915/gvt/hypercall.h b/drivers/gpu/drm/i915/gvt/hypercall.h index 5af11cf1b482..e1675a00df12 100644 --- a/drivers/gpu/drm/i915/gvt/hypercall.h +++ b/drivers/gpu/drm/i915/gvt/hypercall.h @@ -41,7 +41,7 @@ struct intel_gvt_mpt { int (*host_init)(struct device *dev, void *gvt, const void *ops); void (*host_exit)(struct device *dev, void *gvt); int (*attach_vgpu)(void *vgpu, unsigned long *handle); - void (*detach_vgpu)(unsigned long handle); + void (*detach_vgpu)(void *vgpu); int (*inject_msi)(unsigned long handle, u32 addr, u16 data); unsigned long (*from_virt_to_mfn)(void *p); int (*enable_page_track)(unsigned long handle, u64 gfn); diff --git a/drivers/gpu/drm/i915/gvt/interrupt.c b/drivers/gpu/drm/i915/gvt/interrupt.c index 5daa23ae566b..6b9d1354ff29 100644 --- a/drivers/gpu/drm/i915/gvt/interrupt.c +++ b/drivers/gpu/drm/i915/gvt/interrupt.c @@ -126,7 +126,7 @@ static const char * const irq_name[INTEL_GVT_EVENT_MAX] = { [FDI_RX_INTERRUPTS_TRANSCODER_C] = "FDI RX Interrupts Combined C", [AUDIO_CP_CHANGE_TRANSCODER_C] = "Audio CP Change Transcoder C", [AUDIO_CP_REQUEST_TRANSCODER_C] = "Audio CP Request Transcoder C", - [ERR_AND_DBG] = "South Error and Debug Interupts Combined", + [ERR_AND_DBG] = "South Error and Debug Interrupts Combined", [GMBUS] = "Gmbus", [SDVO_B_HOTPLUG] = "SDVO B hotplug", [CRT_HOTPLUG] = "CRT Hotplug", diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index c1072143da1d..dd3dfd00f4e6 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -996,7 +996,7 @@ static int intel_vgpu_mmap(struct mdev_device *mdev, struct vm_area_struct *vma) { unsigned int index; u64 virtaddr; - unsigned long req_size, pgoff = 0; + unsigned long req_size, pgoff, req_start; pgprot_t pg_prot; struct intel_vgpu *vgpu = mdev_get_drvdata(mdev); @@ -1014,7 +1014,17 @@ static int intel_vgpu_mmap(struct mdev_device *mdev, struct vm_area_struct *vma) pg_prot = vma->vm_page_prot; virtaddr = vma->vm_start; req_size = vma->vm_end - vma->vm_start; - pgoff = vgpu_aperture_pa_base(vgpu) >> PAGE_SHIFT; + pgoff = vma->vm_pgoff & + ((1U << (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT)) - 1); + req_start = pgoff << PAGE_SHIFT; + + if (!intel_vgpu_in_aperture(vgpu, req_start)) + return -EINVAL; + if (req_start + req_size > + vgpu_aperture_offset(vgpu) + vgpu_aperture_sz(vgpu)) + return -EINVAL; + + pgoff = (gvt_aperture_pa_base(vgpu->gvt) >> PAGE_SHIFT) + pgoff; return remap_pfn_range(vma, virtaddr, pgoff, req_size, pg_prot); } @@ -1662,9 +1672,21 @@ static int kvmgt_attach_vgpu(void *vgpu, unsigned long *handle) return 0; } -static void kvmgt_detach_vgpu(unsigned long handle) +static void kvmgt_detach_vgpu(void *p_vgpu) { - /* nothing to do here */ + int i; + struct intel_vgpu *vgpu = (struct intel_vgpu *)p_vgpu; + + if (!vgpu->vdev.region) + return; + + for (i = 0; i < vgpu->vdev.num_regions; i++) + if (vgpu->vdev.region[i].ops->release) + vgpu->vdev.region[i].ops->release(vgpu, + &vgpu->vdev.region[i]); + vgpu->vdev.num_regions = 0; + kfree(vgpu->vdev.region); + vgpu->vdev.region = NULL; } static int kvmgt_inject_msi(unsigned long handle, u32 addr, u16 data) diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c index 36a5147cd01e..d6e02c15ef97 100644 --- a/drivers/gpu/drm/i915/gvt/mmio_context.c +++ b/drivers/gpu/drm/i915/gvt/mmio_context.c @@ -158,6 +158,8 @@ static void load_render_mocs(struct drm_i915_private *dev_priv) int ring_id, i; for (ring_id = 0; ring_id < ARRAY_SIZE(regs); ring_id++) { + if (!HAS_ENGINE(dev_priv, ring_id)) + continue; offset.reg = regs[ring_id]; for (i = 0; i < GEN9_MOCS_SIZE; i++) { gen9_render_mocs.control_table[ring_id][i] = diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h index 67f19992b226..3ed34123d8d1 100644 --- a/drivers/gpu/drm/i915/gvt/mpt.h +++ b/drivers/gpu/drm/i915/gvt/mpt.h @@ -101,7 +101,7 @@ static inline void intel_gvt_hypervisor_detach_vgpu(struct intel_vgpu *vgpu) if (!intel_gvt_host.mpt->detach_vgpu) return; - intel_gvt_host.mpt->detach_vgpu(vgpu->handle); + intel_gvt_host.mpt->detach_vgpu(vgpu); } #define MSI_CAP_CONTROL(offset) (offset + 2) diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index b8fbe3fabea3..55bb7885e228 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -332,6 +332,9 @@ static void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx) i915_gem_object_unpin_map(wa_ctx->indirect_ctx.obj); i915_gem_object_put(wa_ctx->indirect_ctx.obj); + + wa_ctx->indirect_ctx.obj = NULL; + wa_ctx->indirect_ctx.shadow_va = NULL; } static int set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload, @@ -356,6 +359,33 @@ static int set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload, return 0; } +static int +intel_gvt_workload_req_alloc(struct intel_vgpu_workload *workload) +{ + struct intel_vgpu *vgpu = workload->vgpu; + struct intel_vgpu_submission *s = &vgpu->submission; + struct i915_gem_context *shadow_ctx = s->shadow_ctx; + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; + struct intel_engine_cs *engine = dev_priv->engine[workload->ring_id]; + struct i915_request *rq; + int ret = 0; + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + if (workload->req) + goto out; + + rq = i915_request_alloc(engine, shadow_ctx); + if (IS_ERR(rq)) { + gvt_vgpu_err("fail to allocate gem request\n"); + ret = PTR_ERR(rq); + goto out; + } + workload->req = i915_request_get(rq); +out: + return ret; +} + /** * intel_gvt_scan_and_shadow_workload - audit the workload by scanning and * shadow it as well, include ringbuffer,wa_ctx and ctx. @@ -372,12 +402,11 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload) struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; struct intel_engine_cs *engine = dev_priv->engine[workload->ring_id]; struct intel_context *ce; - struct i915_request *rq; int ret; lockdep_assert_held(&dev_priv->drm.struct_mutex); - if (workload->req) + if (workload->shadow) return 0; ret = set_context_ppgtt_from_shadow(workload, shadow_ctx); @@ -417,22 +446,8 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload) goto err_shadow; } - rq = i915_request_alloc(engine, shadow_ctx); - if (IS_ERR(rq)) { - gvt_vgpu_err("fail to allocate gem request\n"); - ret = PTR_ERR(rq); - goto err_shadow; - } - workload->req = i915_request_get(rq); - - ret = populate_shadow_context(workload); - if (ret) - goto err_req; - + workload->shadow = true; return 0; -err_req: - rq = fetch_and_zero(&workload->req); - i915_request_put(rq); err_shadow: release_shadow_wa_ctx(&workload->wa_ctx); err_unpin: @@ -671,23 +686,31 @@ static int dispatch_workload(struct intel_vgpu_workload *workload) mutex_lock(&vgpu->vgpu_lock); mutex_lock(&dev_priv->drm.struct_mutex); + ret = intel_gvt_workload_req_alloc(workload); + if (ret) + goto err_req; + ret = intel_gvt_scan_and_shadow_workload(workload); if (ret) goto out; - ret = prepare_workload(workload); + ret = populate_shadow_context(workload); + if (ret) { + release_shadow_wa_ctx(&workload->wa_ctx); + goto out; + } + ret = prepare_workload(workload); out: - if (ret) - workload->status = ret; - if (!IS_ERR_OR_NULL(workload->req)) { gvt_dbg_sched("ring id %d submit workload to i915 %p\n", ring_id, workload->req); i915_request_add(workload->req); workload->dispatched = true; } - +err_req: + if (ret) + workload->status = ret; mutex_unlock(&dev_priv->drm.struct_mutex); mutex_unlock(&vgpu->vgpu_lock); return ret; @@ -891,11 +914,6 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) list_del_init(&workload->list); - if (!workload->status) { - release_shadow_batch_buffer(workload); - release_shadow_wa_ctx(&workload->wa_ctx); - } - if (workload->status || (vgpu->resetting_eng & ENGINE_MASK(ring_id))) { /* if workload->status is not successful means HW GPU * has occurred GPU hang or something wrong with i915/GVT, @@ -1079,6 +1097,21 @@ err: return ret; } +static void +i915_context_ppgtt_root_restore(struct intel_vgpu_submission *s) +{ + struct i915_hw_ppgtt *i915_ppgtt = s->shadow_ctx->ppgtt; + int i; + + if (i915_vm_is_48bit(&i915_ppgtt->vm)) + px_dma(&i915_ppgtt->pml4) = s->i915_context_pml4; + else { + for (i = 0; i < GEN8_3LVL_PDPES; i++) + px_dma(i915_ppgtt->pdp.page_directory[i]) = + s->i915_context_pdps[i]; + } +} + /** * intel_vgpu_clean_submission - free submission-related resource for vGPU * @vgpu: a vGPU @@ -1091,6 +1124,7 @@ void intel_vgpu_clean_submission(struct intel_vgpu *vgpu) struct intel_vgpu_submission *s = &vgpu->submission; intel_vgpu_select_submission_ops(vgpu, ALL_ENGINES, 0); + i915_context_ppgtt_root_restore(s); i915_gem_context_put(s->shadow_ctx); kmem_cache_destroy(s->workloads); } @@ -1116,6 +1150,21 @@ void intel_vgpu_reset_submission(struct intel_vgpu *vgpu, s->ops->reset(vgpu, engine_mask); } +static void +i915_context_ppgtt_root_save(struct intel_vgpu_submission *s) +{ + struct i915_hw_ppgtt *i915_ppgtt = s->shadow_ctx->ppgtt; + int i; + + if (i915_vm_is_48bit(&i915_ppgtt->vm)) + s->i915_context_pml4 = px_dma(&i915_ppgtt->pml4); + else { + for (i = 0; i < GEN8_3LVL_PDPES; i++) + s->i915_context_pdps[i] = + px_dma(i915_ppgtt->pdp.page_directory[i]); + } +} + /** * intel_vgpu_setup_submission - setup submission-related resource for vGPU * @vgpu: a vGPU @@ -1138,6 +1187,8 @@ int intel_vgpu_setup_submission(struct intel_vgpu *vgpu) if (IS_ERR(s->shadow_ctx)) return PTR_ERR(s->shadow_ctx); + i915_context_ppgtt_root_save(s); + bitmap_zero(s->shadow_ctx_desc_updated, I915_NUM_ENGINES); s->workloads = kmem_cache_create_usercopy("gvt-g_vgpu_workload", @@ -1230,6 +1281,9 @@ void intel_vgpu_destroy_workload(struct intel_vgpu_workload *workload) { struct intel_vgpu_submission *s = &workload->vgpu->submission; + release_shadow_batch_buffer(workload); + release_shadow_wa_ctx(&workload->wa_ctx); + if (workload->shadow_mm) intel_vgpu_mm_put(workload->shadow_mm); diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h index ca5529d0e48e..2065cba59aab 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.h +++ b/drivers/gpu/drm/i915/gvt/scheduler.h @@ -83,6 +83,7 @@ struct intel_vgpu_workload { struct i915_request *req; /* if this workload has been dispatched to i915? */ bool dispatched; + bool shadow; /* if workload has done shadow of guest request */ int status; struct intel_vgpu_mm *shadow_mm; diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 38dcee1ca062..40a61ef9aac1 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -984,8 +984,8 @@ static int i915_gpu_info_open(struct inode *inode, struct file *file) intel_runtime_pm_get(i915); gpu = i915_capture_gpu_state(i915); intel_runtime_pm_put(i915); - if (!gpu) - return -ENOMEM; + if (IS_ERR(gpu)) + return PTR_ERR(gpu); file->private_data = gpu; return 0; @@ -1018,7 +1018,13 @@ i915_error_state_write(struct file *filp, static int i915_error_state_open(struct inode *inode, struct file *file) { - file->private_data = i915_first_error_state(inode->i_private); + struct i915_gpu_state *error; + + error = i915_first_error_state(inode->i_private); + if (IS_ERR(error)) + return PTR_ERR(error); + + file->private_data = error; return 0; } diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index d36a9755ad91..216f52b744a6 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1282,8 +1282,7 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data, if (args->size == 0) return 0; - if (!access_ok(VERIFY_WRITE, - u64_to_user_ptr(args->data_ptr), + if (!access_ok(u64_to_user_ptr(args->data_ptr), args->size)) return -EFAULT; @@ -1609,9 +1608,7 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, if (args->size == 0) return 0; - if (!access_ok(VERIFY_READ, - u64_to_user_ptr(args->data_ptr), - args->size)) + if (!access_ok(u64_to_user_ptr(args->data_ptr), args->size)) return -EFAULT; obj = i915_gem_object_lookup(file, args->handle); @@ -2559,7 +2556,7 @@ static int i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj) * If there's no chance of allocating enough pages for the whole * object, bail early. */ - if (page_count > totalram_pages) + if (page_count > totalram_pages()) return -ENOMEM; st = kmalloc(sizeof(*st), GFP_KERNEL); diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 10a4afb4f235..485b259127c3 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -26,7 +26,7 @@ * */ -#include <linux/dma_remapping.h> +#include <linux/intel-iommu.h> #include <linux/reservation.h> #include <linux/sync_file.h> #include <linux/uaccess.h> @@ -1268,7 +1268,7 @@ relocate_entry(struct i915_vma *vma, else if (gen >= 4) len = 4; else - len = 6; + len = 3; batch = reloc_gpu(eb, vma, len); if (IS_ERR(batch)) @@ -1309,11 +1309,6 @@ relocate_entry(struct i915_vma *vma, *batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL; *batch++ = addr; *batch++ = target_offset; - - /* And again for good measure (blb/pnv) */ - *batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL; - *batch++ = addr; - *batch++ = target_offset; } goto out; @@ -1452,7 +1447,7 @@ static int eb_relocate_vma(struct i915_execbuffer *eb, struct i915_vma *vma) * to read. However, if the array is not writable the user loses * the updated relocation values. */ - if (unlikely(!access_ok(VERIFY_READ, urelocs, remain*sizeof(*urelocs)))) + if (unlikely(!access_ok(urelocs, remain*sizeof(*urelocs)))) return -EFAULT; do { @@ -1559,7 +1554,7 @@ static int check_relocations(const struct drm_i915_gem_exec_object2 *entry) addr = u64_to_user_ptr(entry->relocs_ptr); size *= sizeof(struct drm_i915_gem_relocation_entry); - if (!access_ok(VERIFY_READ, addr, size)) + if (!access_ok(addr, size)) return -EFAULT; end = addr + size; @@ -1610,6 +1605,7 @@ static int eb_copy_relocations(const struct i915_execbuffer *eb) (char __user *)urelocs + copied, len)) { end_user: + user_access_end(); kvfree(relocs); err = -EFAULT; goto err; @@ -1628,7 +1624,9 @@ end_user: * happened we would make the mistake of assuming that the * relocations were valid. */ - user_access_begin(); + if (!user_access_begin(urelocs, size)) + goto end_user; + for (copied = 0; copied < nreloc; copied++) unsafe_put_user(-1, &urelocs[copied].presumed_offset, @@ -2095,7 +2093,7 @@ get_fence_array(struct drm_i915_gem_execbuffer2 *args, return ERR_PTR(-EINVAL); user = u64_to_user_ptr(args->cliprects_ptr); - if (!access_ok(VERIFY_READ, user, nfences * sizeof(*user))) + if (!access_ok(user, nfences * sizeof(*user))) return ERR_PTR(-EFAULT); fences = kvmalloc_array(nfences, sizeof(*fences), @@ -2610,7 +2608,16 @@ i915_gem_execbuffer2_ioctl(struct drm_device *dev, void *data, unsigned int i; /* Copy the new buffer offsets back to the user's exec list. */ - user_access_begin(); + /* + * Note: count * sizeof(*user_exec_list) does not overflow, + * because we checked 'count' in check_buffer_count(). + * + * And this range already got effectively checked earlier + * when we did the "copy_from_user()" above. + */ + if (!user_access_begin(user_exec_list, count * sizeof(*user_exec_list))) + goto end_user; + for (i = 0; i < args->buffer_count; i++) { if (!(exec2_list[i].offset & UPDATE)) continue; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index add1fe7aeb93..bd17dd1f5da5 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -2075,6 +2075,7 @@ static struct i915_vma *pd_vma_create(struct gen6_hw_ppgtt *ppgtt, int size) int gen6_ppgtt_pin(struct i915_hw_ppgtt *base) { struct gen6_hw_ppgtt *ppgtt = to_gen6_ppgtt(base); + int err; /* * Workaround the limited maximum vma->pin_count and the aliasing_ppgtt @@ -2090,9 +2091,17 @@ int gen6_ppgtt_pin(struct i915_hw_ppgtt *base) * allocator works in address space sizes, so it's multiplied by page * size. We allocate at the top of the GTT to avoid fragmentation. */ - return i915_vma_pin(ppgtt->vma, - 0, GEN6_PD_ALIGN, - PIN_GLOBAL | PIN_HIGH); + err = i915_vma_pin(ppgtt->vma, + 0, GEN6_PD_ALIGN, + PIN_GLOBAL | PIN_HIGH); + if (err) + goto unpin; + + return 0; + +unpin: + ppgtt->pin_count = 0; + return err; } void gen6_ppgtt_unpin(struct i915_hw_ppgtt *base) diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c index 2c9b284036d1..9558582c105e 100644 --- a/drivers/gpu/drm/i915/i915_gem_userptr.c +++ b/drivers/gpu/drm/i915/i915_gem_userptr.c @@ -113,27 +113,25 @@ static void del_object(struct i915_mmu_object *mo) } static int i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end, - bool blockable) + const struct mmu_notifier_range *range) { struct i915_mmu_notifier *mn = container_of(_mn, struct i915_mmu_notifier, mn); struct i915_mmu_object *mo; struct interval_tree_node *it; LIST_HEAD(cancelled); + unsigned long end; if (RB_EMPTY_ROOT(&mn->objects.rb_root)) return 0; /* interval ranges are inclusive, but invalidate range is exclusive */ - end--; + end = range->end - 1; spin_lock(&mn->lock); - it = interval_tree_iter_first(&mn->objects, start, end); + it = interval_tree_iter_first(&mn->objects, range->start, end); while (it) { - if (!blockable) { + if (!range->blockable) { spin_unlock(&mn->lock); return -EAGAIN; } @@ -151,7 +149,7 @@ static int i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, queue_work(mn->wq, &mo->work); list_add(&mo->link, &cancelled); - it = interval_tree_iter_next(it, start, end); + it = interval_tree_iter_next(it, range->start, end); } list_for_each_entry(mo, &cancelled, link) del_object(mo); @@ -791,8 +789,7 @@ i915_gem_userptr_ioctl(struct drm_device *dev, if (offset_in_page(args->user_ptr | args->user_size)) return -EINVAL; - if (!access_ok(args->flags & I915_USERPTR_READ_ONLY ? VERIFY_READ : VERIFY_WRITE, - (char __user *)(unsigned long)args->user_ptr, args->user_size)) + if (!access_ok((char __user *)(unsigned long)args->user_ptr, args->user_size)) return -EFAULT; if (args->flags & I915_USERPTR_READ_ONLY) { diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 07465123c166..3f9ce403c755 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1907,9 +1907,16 @@ i915_capture_gpu_state(struct drm_i915_private *i915) { struct i915_gpu_state *error; + /* Check if GPU capture has been disabled */ + error = READ_ONCE(i915->gpu_error.first_error); + if (IS_ERR(error)) + return error; + error = kzalloc(sizeof(*error), GFP_ATOMIC); - if (!error) - return NULL; + if (!error) { + i915_disable_error_state(i915, -ENOMEM); + return ERR_PTR(-ENOMEM); + } kref_init(&error->ref); error->i915 = i915; @@ -1945,11 +1952,8 @@ void i915_capture_error_state(struct drm_i915_private *i915, return; error = i915_capture_gpu_state(i915); - if (!error) { - DRM_DEBUG_DRIVER("out of memory, not capturing error state\n"); - i915_disable_error_state(i915, -ENOMEM); + if (IS_ERR(error)) return; - } i915_error_capture_msg(i915, error, engine_mask, error_msg); DRM_INFO("%s\n", error->error_msg); @@ -1987,7 +1991,7 @@ i915_first_error_state(struct drm_i915_private *i915) spin_lock_irq(&i915->gpu_error.lock); error = i915->gpu_error.first_error; - if (error) + if (!IS_ERR_OR_NULL(error)) i915_gpu_state_get(error); spin_unlock_irq(&i915->gpu_error.lock); @@ -2000,10 +2004,11 @@ void i915_reset_error_state(struct drm_i915_private *i915) spin_lock_irq(&i915->gpu_error.lock); error = i915->gpu_error.first_error; - i915->gpu_error.first_error = NULL; + if (error != ERR_PTR(-ENODEV)) /* if disabled, always disabled */ + i915->gpu_error.first_error = NULL; spin_unlock_irq(&i915->gpu_error.lock); - if (!IS_ERR(error)) + if (!IS_ERR_OR_NULL(error)) i915_gpu_state_put(error); } diff --git a/drivers/gpu/drm/i915/i915_ioc32.c b/drivers/gpu/drm/i915/i915_ioc32.c index 0e5c580d117c..e869daf9c8a9 100644 --- a/drivers/gpu/drm/i915/i915_ioc32.c +++ b/drivers/gpu/drm/i915/i915_ioc32.c @@ -52,7 +52,7 @@ static int compat_i915_getparam(struct file *file, unsigned int cmd, return -EFAULT; request = compat_alloc_user_space(sizeof(*request)); - if (!access_ok(VERIFY_WRITE, request, sizeof(*request)) || + if (!access_ok(request, sizeof(*request)) || __put_user(req32.param, &request->param) || __put_user((void __user *)(unsigned long)req32.value, &request->value)) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 4529edfdcfc8..2b2eb57ca71f 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -3052,7 +3052,7 @@ static struct i915_oa_reg *alloc_oa_regs(struct drm_i915_private *dev_priv, if (!n_regs) return NULL; - if (!access_ok(VERIFY_READ, regs, n_regs * sizeof(u32) * 2)) + if (!access_ok(regs, n_regs * sizeof(u32) * 2)) return ERR_PTR(-EFAULT); /* No is_valid function means we're not allowing any register to be programmed. */ diff --git a/drivers/gpu/drm/i915/i915_query.c b/drivers/gpu/drm/i915/i915_query.c index 6fc4b8eeab42..fe56465cdfd6 100644 --- a/drivers/gpu/drm/i915/i915_query.c +++ b/drivers/gpu/drm/i915/i915_query.c @@ -46,7 +46,7 @@ static int query_topology_info(struct drm_i915_private *dev_priv, if (topo.flags != 0) return -EINVAL; - if (!access_ok(VERIFY_WRITE, u64_to_user_ptr(query_item->data_ptr), + if (!access_ok(u64_to_user_ptr(query_item->data_ptr), total_length)) return -EFAULT; diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index 535caebd9813..c0cfe7ae2ba5 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -521,7 +521,9 @@ static ssize_t error_state_read(struct file *filp, struct kobject *kobj, ssize_t ret; gpu = i915_first_error_state(i915); - if (gpu) { + if (IS_ERR(gpu)) { + ret = PTR_ERR(gpu); + } else if (gpu) { ret = i915_gpu_state_copy_to_buffer(gpu, buf, off, count); i915_gpu_state_put(gpu); } else { diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 07c861884c70..3da9c0f9e948 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -46,7 +46,7 @@ #include <drm/drm_plane_helper.h> #include <drm/drm_rect.h> #include <drm/drm_atomic_uapi.h> -#include <linux/dma_remapping.h> +#include <linux/intel-iommu.h> #include <linux/reservation.h> /* Primary plane formats for gen <= 3 */ diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index d7fa301b5ec7..eab9341a5152 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -303,6 +303,7 @@ static void __unwind_incomplete_requests(struct intel_engine_cs *engine) */ if (!(prio & I915_PRIORITY_NEWCLIENT)) { prio |= I915_PRIORITY_NEWCLIENT; + active->sched.attr.priority = prio; list_move_tail(&active->sched.link, i915_sched_lookup_priolist(engine, prio)); } @@ -398,8 +399,13 @@ static u64 execlists_update_context(struct i915_request *rq) * may not be visible to the HW prior to the completion of the UC * register write and that we may begin execution from the context * before its image is complete leading to invalid PD chasing. + * + * Furthermore, Braswell, at least, wants a full mb to be sure that + * the writes are coherent in memory (visible to the GPU) prior to + * execution, and not just visible to other CPUs (as is the result of + * wmb). */ - wmb(); + mb(); return ce->lrc_desc; } @@ -640,6 +646,9 @@ static void execlists_dequeue(struct intel_engine_cs *engine) int i; priolist_for_each_request_consume(rq, rn, p, i) { + GEM_BUG_ON(last && + need_preempt(engine, last, rq_prio(rq))); + /* * Can we combine this request with the current port? * It has to be the same context/ringbuffer and not @@ -2239,6 +2248,8 @@ static int logical_ring_init(struct intel_engine_cs *engine) if (ret) return ret; + intel_engine_init_workarounds(engine); + if (HAS_LOGICAL_RING_ELSQ(i915)) { execlists->submit_reg = i915->regs + i915_mmio_reg_offset(RING_EXECLIST_SQ_CONTENTS(engine)); @@ -2305,7 +2316,6 @@ int logical_render_ring_init(struct intel_engine_cs *engine) } intel_engine_init_whitelist(engine); - intel_engine_init_workarounds(engine); return 0; } diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 419e56342523..f71970df9936 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -274,10 +274,16 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) DRM_DEBUG_KMS("eDP panel supports PSR version %x\n", intel_dp->psr_dpcd[0]); + if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_NO_PSR)) { + DRM_DEBUG_KMS("PSR support not currently available for this panel\n"); + return; + } + if (!(intel_dp->edp_dpcd[1] & DP_EDP_SET_POWER_CAP)) { DRM_DEBUG_KMS("Panel lacks power state control, PSR cannot be enabled\n"); return; } + dev_priv->psr.sink_support = true; dev_priv->psr.sink_sync_latency = intel_dp_get_sink_sync_latency(intel_dp); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index c5eb26a7ee79..fbeaec3994e7 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -69,19 +69,28 @@ unsigned int intel_ring_update_space(struct intel_ring *ring) static int gen2_render_ring_flush(struct i915_request *rq, u32 mode) { + unsigned int num_store_dw; u32 cmd, *cs; cmd = MI_FLUSH; - + num_store_dw = 0; if (mode & EMIT_INVALIDATE) cmd |= MI_READ_FLUSH; + if (mode & EMIT_FLUSH) + num_store_dw = 4; - cs = intel_ring_begin(rq, 2); + cs = intel_ring_begin(rq, 2 + 3 * num_store_dw); if (IS_ERR(cs)) return PTR_ERR(cs); *cs++ = cmd; - *cs++ = MI_NOOP; + while (num_store_dw--) { + *cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL; + *cs++ = i915_scratch_offset(rq->i915); + *cs++ = 0; + } + *cs++ = MI_FLUSH | MI_NO_WRITE_FLUSH; + intel_ring_advance(rq, cs); return 0; diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c index 69fe86b30fbb..a9ed0ecc94e2 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c @@ -170,7 +170,7 @@ static int igt_ppgtt_alloc(void *arg) * This should ensure that we do not run into the oomkiller during * the test and take down the machine wilfully. */ - limit = totalram_pages << PAGE_SHIFT; + limit = totalram_pages() << PAGE_SHIFT; limit = min(ppgtt->vm.total, limit); /* Check we can allocate the entire range */ @@ -1244,7 +1244,7 @@ static int exercise_mock(struct drm_i915_private *i915, u64 hole_start, u64 hole_end, unsigned long end_time)) { - const u64 limit = totalram_pages << PAGE_SHIFT; + const u64 limit = totalram_pages() << PAGE_SHIFT; struct i915_gem_context *ctx; struct i915_hw_ppgtt *ppgtt; IGT_TIMEOUT(end_time); diff --git a/drivers/gpu/drm/mediatek/mtk_dsi.c b/drivers/gpu/drm/mediatek/mtk_dsi.c index 66df1b177959..27b507eb4a99 100644 --- a/drivers/gpu/drm/mediatek/mtk_dsi.c +++ b/drivers/gpu/drm/mediatek/mtk_dsi.c @@ -818,10 +818,13 @@ static int mtk_dsi_create_conn_enc(struct drm_device *drm, struct mtk_dsi *dsi) dsi->encoder.possible_crtcs = 1; /* If there's a bridge, attach to it and let it create the connector */ - ret = drm_bridge_attach(&dsi->encoder, dsi->bridge, NULL); - if (ret) { - DRM_ERROR("Failed to attach bridge to drm\n"); - + if (dsi->bridge) { + ret = drm_bridge_attach(&dsi->encoder, dsi->bridge, NULL); + if (ret) { + DRM_ERROR("Failed to attach bridge to drm\n"); + goto err_encoder_cleanup; + } + } else { /* Otherwise create our own connector and attach to a panel */ ret = mtk_dsi_create_connector(drm, dsi); if (ret) diff --git a/drivers/gpu/drm/meson/meson_crtc.c b/drivers/gpu/drm/meson/meson_crtc.c index d78168f979db..4f5c67f70c4d 100644 --- a/drivers/gpu/drm/meson/meson_crtc.c +++ b/drivers/gpu/drm/meson/meson_crtc.c @@ -106,6 +106,8 @@ static void meson_crtc_atomic_enable(struct drm_crtc *crtc, writel_bits_relaxed(VPP_POSTBLEND_ENABLE, VPP_POSTBLEND_ENABLE, priv->io_base + _REG(VPP_MISC)); + drm_crtc_vblank_on(crtc); + priv->viu.osd1_enabled = true; } @@ -117,6 +119,8 @@ static void meson_crtc_atomic_disable(struct drm_crtc *crtc, DRM_DEBUG_DRIVER("\n"); + drm_crtc_vblank_off(crtc); + priv->viu.osd1_enabled = false; priv->viu.osd1_commit = false; diff --git a/drivers/gpu/drm/meson/meson_drv.c b/drivers/gpu/drm/meson/meson_drv.c index 3ee4d4a4ecba..12ff47b13668 100644 --- a/drivers/gpu/drm/meson/meson_drv.c +++ b/drivers/gpu/drm/meson/meson_drv.c @@ -75,6 +75,10 @@ static const struct drm_mode_config_funcs meson_mode_config_funcs = { .fb_create = drm_gem_fb_create, }; +static const struct drm_mode_config_helper_funcs meson_mode_config_helpers = { + .atomic_commit_tail = drm_atomic_helper_commit_tail_rpm, +}; + static irqreturn_t meson_irq(int irq, void *arg) { struct drm_device *dev = arg; @@ -266,6 +270,7 @@ static int meson_drv_bind_master(struct device *dev, bool has_components) drm->mode_config.max_width = 3840; drm->mode_config.max_height = 2160; drm->mode_config.funcs = &meson_mode_config_funcs; + drm->mode_config.helper_private = &meson_mode_config_helpers; /* Hardware Initialization */ @@ -388,8 +393,10 @@ static int meson_probe_remote(struct platform_device *pdev, remote_node = of_graph_get_remote_port_parent(ep); if (!remote_node || remote_node == parent || /* Ignore parent endpoint */ - !of_device_is_available(remote_node)) + !of_device_is_available(remote_node)) { + of_node_put(remote_node); continue; + } count += meson_probe_remote(pdev, match, remote, remote_node); @@ -408,10 +415,13 @@ static int meson_drv_probe(struct platform_device *pdev) for_each_endpoint_of_node(np, ep) { remote = of_graph_get_remote_port_parent(ep); - if (!remote || !of_device_is_available(remote)) + if (!remote || !of_device_is_available(remote)) { + of_node_put(remote); continue; + } count += meson_probe_remote(pdev, &match, np, remote); + of_node_put(remote); } if (count && !match) diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c b/drivers/gpu/drm/meson/meson_dw_hdmi.c index d8c5cc34e22e..807111ebfdd9 100644 --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c @@ -696,6 +696,7 @@ static const struct regmap_config meson_dw_hdmi_regmap_config = { .reg_read = meson_dw_hdmi_reg_read, .reg_write = meson_dw_hdmi_reg_write, .max_register = 0x10000, + .fast_io = true, }; static bool meson_hdmi_connector_is_available(struct device *dev) diff --git a/drivers/gpu/drm/meson/meson_venc.c b/drivers/gpu/drm/meson/meson_venc.c index e95e0e7a7fa1..0ba04f6813e6 100644 --- a/drivers/gpu/drm/meson/meson_venc.c +++ b/drivers/gpu/drm/meson/meson_venc.c @@ -71,6 +71,7 @@ */ /* HHI Registers */ +#define HHI_GCLK_MPEG2 0x148 /* 0x52 offset in data sheet */ #define HHI_VDAC_CNTL0 0x2F4 /* 0xbd offset in data sheet */ #define HHI_VDAC_CNTL1 0x2F8 /* 0xbe offset in data sheet */ #define HHI_HDMI_PHY_CNTL0 0x3a0 /* 0xe8 offset in data sheet */ @@ -840,6 +841,7 @@ struct meson_hdmi_venc_vic_mode { { 5, &meson_hdmi_encp_mode_1080i60 }, { 20, &meson_hdmi_encp_mode_1080i50 }, { 32, &meson_hdmi_encp_mode_1080p24 }, + { 33, &meson_hdmi_encp_mode_1080p50 }, { 34, &meson_hdmi_encp_mode_1080p30 }, { 31, &meson_hdmi_encp_mode_1080p50 }, { 16, &meson_hdmi_encp_mode_1080p60 }, @@ -1659,10 +1661,12 @@ unsigned int meson_venci_get_field(struct meson_drm *priv) void meson_venc_enable_vsync(struct meson_drm *priv) { writel_relaxed(2, priv->io_base + _REG(VENC_INTCTRL)); + regmap_update_bits(priv->hhi, HHI_GCLK_MPEG2, BIT(25), BIT(25)); } void meson_venc_disable_vsync(struct meson_drm *priv) { + regmap_update_bits(priv->hhi, HHI_GCLK_MPEG2, BIT(25), 0); writel_relaxed(0, priv->io_base + _REG(VENC_INTCTRL)); } diff --git a/drivers/gpu/drm/meson/meson_viu.c b/drivers/gpu/drm/meson/meson_viu.c index 0ba87ff95530..e46e05f50bad 100644 --- a/drivers/gpu/drm/meson/meson_viu.c +++ b/drivers/gpu/drm/meson/meson_viu.c @@ -184,18 +184,18 @@ void meson_viu_set_osd_lut(struct meson_drm *priv, enum viu_lut_sel_e lut_sel, if (lut_sel == VIU_LUT_OSD_OETF) { writel(0, priv->io_base + _REG(addr_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_OETF_LUT_SIZE / 2); i++) writel(r_map[i * 2] | (r_map[i * 2 + 1] << 16), priv->io_base + _REG(data_port)); writel(r_map[OSD_OETF_LUT_SIZE - 1] | (g_map[0] << 16), priv->io_base + _REG(data_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_OETF_LUT_SIZE / 2); i++) writel(g_map[i * 2 + 1] | (g_map[i * 2 + 2] << 16), priv->io_base + _REG(data_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_OETF_LUT_SIZE / 2); i++) writel(b_map[i * 2] | (b_map[i * 2 + 1] << 16), priv->io_base + _REG(data_port)); @@ -211,18 +211,18 @@ void meson_viu_set_osd_lut(struct meson_drm *priv, enum viu_lut_sel_e lut_sel, } else if (lut_sel == VIU_LUT_OSD_EOTF) { writel(0, priv->io_base + _REG(addr_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_EOTF_LUT_SIZE / 2); i++) writel(r_map[i * 2] | (r_map[i * 2 + 1] << 16), priv->io_base + _REG(data_port)); writel(r_map[OSD_EOTF_LUT_SIZE - 1] | (g_map[0] << 16), priv->io_base + _REG(data_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_EOTF_LUT_SIZE / 2); i++) writel(g_map[i * 2 + 1] | (g_map[i * 2 + 2] << 16), priv->io_base + _REG(data_port)); - for (i = 0; i < 20; i++) + for (i = 0; i < (OSD_EOTF_LUT_SIZE / 2); i++) writel(b_map[i * 2] | (b_map[i * 2 + 1] << 16), priv->io_base + _REG(data_port)); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index e942316b3ba7..ce1b3cc4bf6d 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -927,26 +927,6 @@ static int a6xx_gmu_memory_probe(struct a6xx_gmu *gmu) return ret; } -/* Get the list of RPMh voltage levels from cmd-db */ -static int a6xx_gmu_rpmh_arc_cmds(const char *id, void *vals, int size) -{ - u32 len = cmd_db_read_aux_data_len(id); - - if (!len) - return 0; - - if (WARN_ON(len > size)) - return -EINVAL; - - cmd_db_read_aux_data(id, vals, len); - - /* - * The data comes back as an array of unsigned shorts so adjust the - * count accordingly - */ - return len >> 1; -} - /* Return the 'arc-level' for the given frequency */ static u32 a6xx_gmu_get_arc_level(struct device *dev, unsigned long freq) { @@ -974,11 +954,30 @@ static u32 a6xx_gmu_get_arc_level(struct device *dev, unsigned long freq) } static int a6xx_gmu_rpmh_arc_votes_init(struct device *dev, u32 *votes, - unsigned long *freqs, int freqs_count, - u16 *pri, int pri_count, - u16 *sec, int sec_count) + unsigned long *freqs, int freqs_count, const char *id) { int i, j; + const u16 *pri, *sec; + size_t pri_count, sec_count; + + pri = cmd_db_read_aux_data(id, &pri_count); + if (IS_ERR(pri)) + return PTR_ERR(pri); + /* + * The data comes back as an array of unsigned shorts so adjust the + * count accordingly + */ + pri_count >>= 1; + if (!pri_count) + return -EINVAL; + + sec = cmd_db_read_aux_data("mx.lvl", &sec_count); + if (IS_ERR(sec)) + return PTR_ERR(sec); + + sec_count >>= 1; + if (!sec_count) + return -EINVAL; /* Construct a vote for each frequency */ for (i = 0; i < freqs_count; i++) { @@ -1037,25 +1036,15 @@ static int a6xx_gmu_rpmh_votes_init(struct a6xx_gmu *gmu) struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; struct msm_gpu *gpu = &adreno_gpu->base; - - u16 gx[16], cx[16], mx[16]; - u32 gxcount, cxcount, mxcount; int ret; - /* Get the list of available voltage levels for each component */ - gxcount = a6xx_gmu_rpmh_arc_cmds("gfx.lvl", gx, sizeof(gx)); - cxcount = a6xx_gmu_rpmh_arc_cmds("cx.lvl", cx, sizeof(cx)); - mxcount = a6xx_gmu_rpmh_arc_cmds("mx.lvl", mx, sizeof(mx)); - /* Build the GX votes */ ret = a6xx_gmu_rpmh_arc_votes_init(&gpu->pdev->dev, gmu->gx_arc_votes, - gmu->gpu_freqs, gmu->nr_gpu_freqs, - gx, gxcount, mx, mxcount); + gmu->gpu_freqs, gmu->nr_gpu_freqs, "gfx.lvl"); /* Build the CX votes */ ret |= a6xx_gmu_rpmh_arc_votes_init(gmu->dev, gmu->cx_arc_votes, - gmu->gmu_freqs, gmu->nr_gmu_freqs, - cx, cxcount, mx, mxcount); + gmu->gmu_freqs, gmu->nr_gmu_freqs, "cx.lvl"); return ret; } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c index a6f0c38a0a95..9be7c355debd 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c @@ -1421,7 +1421,6 @@ struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane *plane, NULL); drm_crtc_helper_add(crtc, &dpu_crtc_helper_funcs); - plane->crtc = crtc; /* save user friendly CRTC name for later */ snprintf(dpu_crtc->name, DPU_CRTC_NAME_SIZE, "crtc%u", crtc->base.id); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c index 0dda4a603685..36158b7d99cd 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c @@ -459,8 +459,6 @@ static void dpu_encoder_destroy(struct drm_encoder *drm_enc) drm_encoder_cleanup(drm_enc); mutex_destroy(&dpu_enc->enc_lock); - - kfree(dpu_enc); } void dpu_encoder_helper_split_config( diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_formats.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_formats.c index 0aa9b8e1ae70..0874f0a53bf9 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_formats.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_formats.c @@ -216,7 +216,7 @@ static const struct dpu_format dpu_format_map[] = { INTERLEAVED_RGB_FMT(XBGR8888, COLOR_8BIT, COLOR_8BIT, COLOR_8BIT, COLOR_8BIT, C2_R_Cr, C0_G_Y, C1_B_Cb, C3_ALPHA, 4, - true, 4, 0, + false, 4, 0, DPU_FETCH_LINEAR, 1), INTERLEAVED_RGB_FMT(RGBA8888, diff --git a/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c b/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c index e41f278e4e62..aabab6311043 100644 --- a/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c +++ b/drivers/gpu/drm/msm/dsi/pll/dsi_pll_10nm.c @@ -39,6 +39,8 @@ #define DSI_PIXEL_PLL_CLK 1 #define NUM_PROVIDED_CLKS 2 +#define VCO_REF_CLK_RATE 19200000 + struct dsi_pll_regs { u32 pll_prop_gain_rate; u32 pll_lockdet_rate; @@ -316,7 +318,7 @@ static int dsi_pll_10nm_vco_set_rate(struct clk_hw *hw, unsigned long rate, parent_rate); pll_10nm->vco_current_rate = rate; - pll_10nm->vco_ref_clk_rate = parent_rate; + pll_10nm->vco_ref_clk_rate = VCO_REF_CLK_RATE; dsi_pll_setup_config(pll_10nm); diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c index f0725761b327..e247d6942a49 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi.c @@ -332,6 +332,12 @@ int msm_hdmi_modeset_init(struct hdmi *hdmi, goto fail; } + ret = msm_hdmi_hpd_enable(hdmi->connector); + if (ret < 0) { + DRM_DEV_ERROR(&hdmi->pdev->dev, "failed to enable HPD: %d\n", ret); + goto fail; + } + encoder->bridge = hdmi->bridge; priv->bridges[priv->num_bridges++] = hdmi->bridge; @@ -571,7 +577,7 @@ static int msm_hdmi_bind(struct device *dev, struct device *master, void *data) { struct drm_device *drm = dev_get_drvdata(master); struct msm_drm_private *priv = drm->dev_private; - static struct hdmi_platform_config *hdmi_cfg; + struct hdmi_platform_config *hdmi_cfg; struct hdmi *hdmi; struct device_node *of_node = dev->of_node; int i, err; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h index accc9a61611d..5c5df6ab2a57 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.h +++ b/drivers/gpu/drm/msm/hdmi/hdmi.h @@ -245,6 +245,7 @@ void msm_hdmi_bridge_destroy(struct drm_bridge *bridge); void msm_hdmi_connector_irq(struct drm_connector *connector); struct drm_connector *msm_hdmi_connector_init(struct hdmi *hdmi); +int msm_hdmi_hpd_enable(struct drm_connector *connector); /* * i2c adapter for ddc: diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c index 99f2f1038258..a6eeab2c4dc3 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c @@ -167,8 +167,9 @@ static void enable_hpd_clocks(struct hdmi *hdmi, bool enable) } } -static int hpd_enable(struct hdmi_connector *hdmi_connector) +int msm_hdmi_hpd_enable(struct drm_connector *connector) { + struct hdmi_connector *hdmi_connector = to_hdmi_connector(connector); struct hdmi *hdmi = hdmi_connector->hdmi; const struct hdmi_platform_config *config = hdmi->config; struct device *dev = &hdmi->pdev->dev; @@ -450,7 +451,6 @@ struct drm_connector *msm_hdmi_connector_init(struct hdmi *hdmi) { struct drm_connector *connector = NULL; struct hdmi_connector *hdmi_connector; - int ret; hdmi_connector = kzalloc(sizeof(*hdmi_connector), GFP_KERNEL); if (!hdmi_connector) @@ -471,12 +471,6 @@ struct drm_connector *msm_hdmi_connector_init(struct hdmi *hdmi) connector->interlace_allowed = 0; connector->doublescan_allowed = 0; - ret = hpd_enable(hdmi_connector); - if (ret) { - dev_err(&hdmi->pdev->dev, "failed to enable HPD: %d\n", ret); - return ERR_PTR(ret); - } - drm_connector_attach_encoder(connector, hdmi->encoder); return connector; diff --git a/drivers/gpu/drm/msm/msm_atomic.c b/drivers/gpu/drm/msm/msm_atomic.c index ff291dd0744b..f5b1256e32b6 100644 --- a/drivers/gpu/drm/msm/msm_atomic.c +++ b/drivers/gpu/drm/msm/msm_atomic.c @@ -34,7 +34,12 @@ static void msm_atomic_wait_for_commit_done(struct drm_device *dev, if (!new_crtc_state->active) continue; + if (drm_crtc_vblank_get(crtc)) + continue; + kms->funcs->wait_for_crtc_commit_done(kms, crtc); + + drm_crtc_vblank_put(crtc); } } diff --git a/drivers/gpu/drm/msm/msm_debugfs.c b/drivers/gpu/drm/msm/msm_debugfs.c index 42a2cba78998..fb423d309e91 100644 --- a/drivers/gpu/drm/msm/msm_debugfs.c +++ b/drivers/gpu/drm/msm/msm_debugfs.c @@ -84,7 +84,7 @@ static int msm_gpu_open(struct inode *inode, struct file *file) ret = mutex_lock_interruptible(&dev->struct_mutex); if (ret) - return ret; + goto free_priv; pm_runtime_get_sync(&gpu->pdev->dev); show_priv->state = gpu->funcs->gpu_state_get(gpu); @@ -94,13 +94,20 @@ static int msm_gpu_open(struct inode *inode, struct file *file) if (IS_ERR(show_priv->state)) { ret = PTR_ERR(show_priv->state); - kfree(show_priv); - return ret; + goto free_priv; } show_priv->dev = dev; - return single_open(file, msm_gpu_show, show_priv); + ret = single_open(file, msm_gpu_show, show_priv); + if (ret) + goto free_priv; + + return 0; + +free_priv: + kfree(show_priv); + return ret; } static const struct file_operations msm_gpu_fops = { diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c index 6265be8f4119..d2cdc7b553fe 100644 --- a/drivers/gpu/drm/msm/msm_drv.c +++ b/drivers/gpu/drm/msm/msm_drv.c @@ -563,17 +563,18 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) kthread_run(kthread_worker_fn, &priv->disp_thread[i].worker, "crtc_commit:%d", priv->disp_thread[i].crtc_id); - ret = sched_setscheduler(priv->disp_thread[i].thread, - SCHED_FIFO, ¶m); - if (ret) - pr_warn("display thread priority update failed: %d\n", - ret); - if (IS_ERR(priv->disp_thread[i].thread)) { DRM_DEV_ERROR(dev, "failed to create crtc_commit kthread\n"); priv->disp_thread[i].thread = NULL; + goto err_msm_uninit; } + ret = sched_setscheduler(priv->disp_thread[i].thread, + SCHED_FIFO, ¶m); + if (ret) + dev_warn(dev, "disp_thread set priority failed: %d\n", + ret); + /* initialize event thread */ priv->event_thread[i].crtc_id = priv->crtcs[i]->base.id; kthread_init_worker(&priv->event_thread[i].worker); @@ -582,6 +583,11 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) kthread_run(kthread_worker_fn, &priv->event_thread[i].worker, "crtc_event:%d", priv->event_thread[i].crtc_id); + if (IS_ERR(priv->event_thread[i].thread)) { + DRM_DEV_ERROR(dev, "failed to create crtc_event kthread\n"); + priv->event_thread[i].thread = NULL; + goto err_msm_uninit; + } /** * event thread should also run at same priority as disp_thread @@ -591,34 +597,10 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) * failure at crtc commit level. */ ret = sched_setscheduler(priv->event_thread[i].thread, - SCHED_FIFO, ¶m); + SCHED_FIFO, ¶m); if (ret) - pr_warn("display event thread priority update failed: %d\n", - ret); - - if (IS_ERR(priv->event_thread[i].thread)) { - dev_err(dev, "failed to create crtc_event kthread\n"); - priv->event_thread[i].thread = NULL; - } - - if ((!priv->disp_thread[i].thread) || - !priv->event_thread[i].thread) { - /* clean up previously created threads if any */ - for ( ; i >= 0; i--) { - if (priv->disp_thread[i].thread) { - kthread_stop( - priv->disp_thread[i].thread); - priv->disp_thread[i].thread = NULL; - } - - if (priv->event_thread[i].thread) { - kthread_stop( - priv->event_thread[i].thread); - priv->event_thread[i].thread = NULL; - } - } - goto err_msm_uninit; - } + dev_warn(dev, "event_thread set priority failed:%d\n", + ret); } ret = drm_vblank_init(ddev, priv->num_crtcs); diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index 3cbed4acb0f4..12b983fc0b56 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -77,7 +77,7 @@ void msm_gem_submit_free(struct msm_gem_submit *submit) static inline unsigned long __must_check copy_from_user_inatomic(void *to, const void __user *from, unsigned long n) { - if (access_ok(VERIFY_READ, from, n)) + if (access_ok(from, n)) return __copy_from_user_inatomic(to, from, n); return -EFAULT; } @@ -321,6 +321,9 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob uint32_t *ptr; int ret = 0; + if (!nr_relocs) + return 0; + if (offset % 4) { DRM_ERROR("non-aligned cmdstream buffer: %u\n", offset); return -EINVAL; @@ -415,7 +418,6 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data, struct msm_file_private *ctx = file->driver_priv; struct msm_gem_submit *submit; struct msm_gpu *gpu = priv->gpu; - struct dma_fence *in_fence = NULL; struct sync_file *sync_file = NULL; struct msm_gpu_submitqueue *queue; struct msm_ringbuffer *ring; @@ -453,6 +455,8 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data, args->nr_bos, args->nr_cmds); if (args->flags & MSM_SUBMIT_FENCE_FD_IN) { + struct dma_fence *in_fence; + in_fence = sync_file_get_fence(args->fence_fd); if (!in_fence) @@ -462,11 +466,13 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data, * Wait if the fence is from a foreign context, or if the fence * array contains any fence from a foreign context. */ - if (!dma_fence_match_context(in_fence, ring->fctx->context)) { + ret = 0; + if (!dma_fence_match_context(in_fence, ring->fctx->context)) ret = dma_fence_wait(in_fence, true); - if (ret) - return ret; - } + + dma_fence_put(in_fence); + if (ret) + return ret; } ret = mutex_lock_interruptible(&dev->struct_mutex); @@ -595,8 +601,6 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data, } out: - if (in_fence) - dma_fence_put(in_fence); submit_cleanup(submit); if (ret) msm_gem_submit_free(submit); diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c index abefba5b14e7..10babd18e286 100644 --- a/drivers/gpu/drm/msm/msm_gpu.c +++ b/drivers/gpu/drm/msm/msm_gpu.c @@ -347,6 +347,10 @@ static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, { struct msm_gpu_state *state; + /* Check if the target supports capturing crash state */ + if (!gpu->funcs->gpu_state_get) + return; + /* Only save one crash state at a time */ if (gpu->crashstate) return; @@ -439,10 +443,9 @@ static void recover_worker(struct work_struct *work) if (submit) { struct task_struct *task; - rcu_read_lock(); - task = pid_task(submit->pid, PIDTYPE_PID); + task = get_pid_task(submit->pid, PIDTYPE_PID); if (task) { - comm = kstrdup(task->comm, GFP_ATOMIC); + comm = kstrdup(task->comm, GFP_KERNEL); /* * So slightly annoying, in other paths like @@ -455,10 +458,10 @@ static void recover_worker(struct work_struct *work) * about the submit going away. */ mutex_unlock(&dev->struct_mutex); - cmd = kstrdup_quotable_cmdline(task, GFP_ATOMIC); + cmd = kstrdup_quotable_cmdline(task, GFP_KERNEL); + put_task_struct(task); mutex_lock(&dev->struct_mutex); } - rcu_read_unlock(); if (comm && cmd) { DRM_DEV_ERROR(dev->dev, "%s: offending task: %s (%s)\n", diff --git a/drivers/gpu/drm/msm/msm_iommu.c b/drivers/gpu/drm/msm/msm_iommu.c index 9c313cb129ee..4d62790cd425 100644 --- a/drivers/gpu/drm/msm/msm_iommu.c +++ b/drivers/gpu/drm/msm/msm_iommu.c @@ -66,7 +66,7 @@ static int msm_iommu_map(struct msm_mmu *mmu, uint64_t iova, // pm_runtime_get_sync(mmu->dev); ret = iommu_map_sg(iommu->domain, iova, sgt->sgl, sgt->nents, prot); // pm_runtime_put_sync(mmu->dev); - WARN_ON(ret < 0); + WARN_ON(!ret); return (ret == len) ? 0 : -EINVAL; } diff --git a/drivers/gpu/drm/msm/msm_rd.c b/drivers/gpu/drm/msm/msm_rd.c index d990b5f5154c..d21172933d92 100644 --- a/drivers/gpu/drm/msm/msm_rd.c +++ b/drivers/gpu/drm/msm/msm_rd.c @@ -321,10 +321,11 @@ static void snapshot_buf(struct msm_rd_state *rd, uint64_t iova, uint32_t size) { struct msm_gem_object *obj = submit->bos[idx].obj; + unsigned offset = 0; const char *buf; if (iova) { - buf += iova - submit->bos[idx].iova; + offset = iova - submit->bos[idx].iova; } else { iova = submit->bos[idx].iova; size = obj->base.size; @@ -345,6 +346,8 @@ static void snapshot_buf(struct msm_rd_state *rd, if (IS_ERR(buf)) return; + buf += offset; + rd_write_section(rd, RD_BUFFER_CONTENTS, buf, size); msm_gem_put_vaddr(&obj->base); diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig index 4b75ad40dd80..432c440223bb 100644 --- a/drivers/gpu/drm/nouveau/Kconfig +++ b/drivers/gpu/drm/nouveau/Kconfig @@ -4,7 +4,8 @@ config DRM_NOUVEAU select FW_LOADER select DRM_KMS_HELPER select DRM_TTM - select FB_BACKLIGHT if DRM_NOUVEAU_BACKLIGHT + select BACKLIGHT_CLASS_DEVICE if DRM_NOUVEAU_BACKLIGHT + select BACKLIGHT_LCD_SUPPORT if DRM_NOUVEAU_BACKLIGHT select ACPI_VIDEO if ACPI && X86 && BACKLIGHT_CLASS_DEVICE && INPUT select X86_PLATFORM_DEVICES if ACPI && X86 select ACPI_WMI if ACPI && X86 diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c index 70dce544984e..1727d399833c 100644 --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c @@ -67,7 +67,7 @@ nv04_display_create(struct drm_device *dev) for (i = 0; i < dcb->entries; i++) { struct dcb_output *dcbent = &dcb->entry[i]; - connector = nouveau_connector_create(dev, dcbent->connector); + connector = nouveau_connector_create(dev, dcbent); if (IS_ERR(connector)) continue; diff --git a/drivers/gpu/drm/nouveau/dispnv50/Kbuild b/drivers/gpu/drm/nouveau/dispnv50/Kbuild index 849b0f45afb8..3d074aa31173 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/Kbuild +++ b/drivers/gpu/drm/nouveau/dispnv50/Kbuild @@ -7,6 +7,7 @@ nouveau-y += dispnv50/core827d.o nouveau-y += dispnv50/core907d.o nouveau-y += dispnv50/core917d.o nouveau-y += dispnv50/corec37d.o +nouveau-y += dispnv50/corec57d.o nouveau-y += dispnv50/dac507d.o nouveau-y += dispnv50/dac907d.o @@ -23,12 +24,14 @@ nouveau-y += dispnv50/head827d.o nouveau-y += dispnv50/head907d.o nouveau-y += dispnv50/head917d.o nouveau-y += dispnv50/headc37d.o +nouveau-y += dispnv50/headc57d.o nouveau-y += dispnv50/wimm.o nouveau-y += dispnv50/wimmc37b.o nouveau-y += dispnv50/wndw.o nouveau-y += dispnv50/wndwc37e.o +nouveau-y += dispnv50/wndwc57e.o nouveau-y += dispnv50/base.o nouveau-y += dispnv50/base507c.o diff --git a/drivers/gpu/drm/nouveau/dispnv50/atom.h b/drivers/gpu/drm/nouveau/dispnv50/atom.h index 908feb1fc60f..a194990d2b0d 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/atom.h +++ b/drivers/gpu/drm/nouveau/dispnv50/atom.h @@ -54,9 +54,10 @@ struct nv50_head_atom { u64 offset:40; u8 buffer:1; u8 mode:4; - u8 size:2; + u16 size:11; u8 range:2; u8 output_mode:2; + void (*load)(struct drm_color_lut *, int size, void __iomem *); } olut; struct { @@ -169,9 +170,11 @@ struct nv50_wndw_atom { u8 buffer:1; u8 enable:2; u8 mode:4; - u8 size:2; + u16 size:11; u8 range:2; u8 output_mode:2; + void (*load)(struct drm_color_lut *, int size, + void __iomem *); } i; } xlut; diff --git a/drivers/gpu/drm/nouveau/dispnv50/base907c.c b/drivers/gpu/drm/nouveau/dispnv50/base907c.c index a562fc94ce59..049ce6da321c 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/base907c.c +++ b/drivers/gpu/drm/nouveau/dispnv50/base907c.c @@ -80,6 +80,7 @@ base907c_ilut(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) { asyw->xlut.i.mode = 7; asyw->xlut.i.enable = 2; + asyw->xlut.i.load = head907d_olut_load; } const struct nv50_wndw_func diff --git a/drivers/gpu/drm/nouveau/dispnv50/core.c b/drivers/gpu/drm/nouveau/dispnv50/core.c index f3c49adb1bdb..c25e0ebe3c92 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core.c +++ b/drivers/gpu/drm/nouveau/dispnv50/core.c @@ -42,6 +42,7 @@ nv50_core_new(struct nouveau_drm *drm, struct nv50_core **pcore) int version; int (*new)(struct nouveau_drm *, s32, struct nv50_core **); } cores[] = { + { TU104_DISP_CORE_CHANNEL_DMA, 0, corec57d_new }, { GV100_DISP_CORE_CHANNEL_DMA, 0, corec37d_new }, { GP102_DISP_CORE_CHANNEL_DMA, 0, core917d_new }, { GP100_DISP_CORE_CHANNEL_DMA, 0, core917d_new }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/core.h b/drivers/gpu/drm/nouveau/dispnv50/core.h index 8470df9dd13d..df8336b593f7 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core.h +++ b/drivers/gpu/drm/nouveau/dispnv50/core.h @@ -46,5 +46,9 @@ extern const struct nv50_outp_func sor907d; int core917d_new(struct nouveau_drm *, s32, struct nv50_core **); int corec37d_new(struct nouveau_drm *, s32, struct nv50_core **); +int corec37d_ntfy_wait_done(struct nouveau_bo *, u32, struct nvif_device *); +void corec37d_update(struct nv50_core *, u32 *, bool); extern const struct nv50_outp_func sorc37d; + +int corec57d_new(struct nouveau_drm *, s32, struct nv50_core **); #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c index b5c17c948918..7860774b65bc 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/corec37d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/corec37d.c @@ -24,7 +24,7 @@ #include <nouveau_bo.h> -static void +void corec37d_update(struct nv50_core *core, u32 *interlock, bool ntfy) { u32 *push; @@ -71,7 +71,7 @@ corec37d_ntfy_init(struct nouveau_bo *bo, u32 offset) nouveau_bo_wr32(bo, offset / 4 + 3, 0x00000000); } -void +static void corec37d_init(struct nv50_core *core) { const u32 windows = 8; /*XXX*/ diff --git a/drivers/gpu/drm/nouveau/dispnv50/corec57d.c b/drivers/gpu/drm/nouveau/dispnv50/corec57d.c new file mode 100644 index 000000000000..b606d68cda10 --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/corec57d.c @@ -0,0 +1,61 @@ +/* + * Copyright 2018 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. + */ +#include "core.h" +#include "head.h" + +static void +corec57d_init(struct nv50_core *core) +{ + const u32 windows = 8; /*XXX*/ + u32 *push, i; + if ((push = evo_wait(&core->chan, 2 + 6 * windows + 2))) { + evo_mthd(push, 0x0208, 1); + evo_data(push, core->chan.sync.handle); + for (i = 0; i < windows; i++) { + evo_mthd(push, 0x1000 + (i * 0x080), 3); + evo_data(push, i >> 1); + evo_data(push, 0x0000000f); + evo_data(push, 0x00000000); + evo_mthd(push, 0x1010 + (i * 0x080), 1); + evo_data(push, 0x00117fff); + } + evo_mthd(push, 0x0200, 1); + evo_data(push, 0x00000001); + evo_kick(push, &core->chan); + } +} + +static const struct nv50_core_func +corec57d = { + .init = corec57d_init, + .ntfy_init = corec37d_ntfy_init, + .ntfy_wait_done = corec37d_ntfy_wait_done, + .update = corec37d_update, + .head = &headc57d, + .sor = &sorc37d, +}; + +int +corec57d_new(struct nouveau_drm *drm, s32 oclass, struct nv50_core **pcore) +{ + return core507d_new_(&corec57d, drm, oclass, pcore); +} diff --git a/drivers/gpu/drm/nouveau/dispnv50/curs.c b/drivers/gpu/drm/nouveau/dispnv50/curs.c index f592087338c4..cb6e4d2b1b45 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/curs.c +++ b/drivers/gpu/drm/nouveau/dispnv50/curs.c @@ -31,6 +31,7 @@ nv50_curs_new(struct nouveau_drm *drm, int head, struct nv50_wndw **pwndw) int version; int (*new)(struct nouveau_drm *, int, s32, struct nv50_wndw **); } curses[] = { + { TU104_DISP_CURSOR, 0, cursc37a_new }, { GV100_DISP_CURSOR, 0, cursc37a_new }, { GK104_DISP_CURSOR, 0, curs907a_new }, { GF110_DISP_CURSOR, 0, curs907a_new }, diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c index 6cbbae3f438b..134701a837c8 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/disp.c +++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c @@ -198,6 +198,22 @@ nv50_dmac_create(struct nvif_device *device, struct nvif_object *disp, /****************************************************************************** * EVO channel helpers *****************************************************************************/ +static void +evo_flush(struct nv50_dmac *dmac) +{ + /* Push buffer fetches are not coherent with BAR1, we need to ensure + * writes have been flushed right through to VRAM before writing PUT. + */ + if (dmac->push.type & NVIF_MEM_VRAM) { + struct nvif_device *device = dmac->base.device; + nvif_wr32(&device->object, 0x070000, 0x00000001); + nvif_msec(device, 2000, + if (!(nvif_rd32(&device->object, 0x070000) & 0x00000002)) + break; + ); + } +} + u32 * evo_wait(struct nv50_dmac *evoc, int nr) { @@ -208,6 +224,7 @@ evo_wait(struct nv50_dmac *evoc, int nr) mutex_lock(&dmac->lock); if (put + nr >= (PAGE_SIZE / 4) - 8) { dmac->ptr[put] = 0x20000000; + evo_flush(dmac); nvif_wr32(&dmac->base.user, 0x0000, 0x00000000); if (nvif_msec(device, 2000, @@ -230,17 +247,7 @@ evo_kick(u32 *push, struct nv50_dmac *evoc) { struct nv50_dmac *dmac = evoc; - /* Push buffer fetches are not coherent with BAR1, we need to ensure - * writes have been flushed right through to VRAM before writing PUT. - */ - if (dmac->push.type & NVIF_MEM_VRAM) { - struct nvif_device *device = dmac->base.device; - nvif_wr32(&device->object, 0x070000, 0x00000001); - nvif_msec(device, 2000, - if (!(nvif_rd32(&device->object, 0x070000) & 0x00000002)) - break; - ); - } + evo_flush(dmac); nvif_wr32(&dmac->base.user, 0x0000, (push - dmac->ptr) << 2); mutex_unlock(&dmac->lock); @@ -1255,8 +1262,16 @@ nv50_mstm_fini(struct nv50_mstm *mstm) static void nv50_mstm_init(struct nv50_mstm *mstm) { - if (mstm && mstm->mgr.mst_state) - drm_dp_mst_topology_mgr_resume(&mstm->mgr); + int ret; + + if (!mstm || !mstm->mgr.mst_state) + return; + + ret = drm_dp_mst_topology_mgr_resume(&mstm->mgr); + if (ret == -1) { + drm_dp_mst_topology_mgr_set_mst(&mstm->mgr, false); + drm_kms_helper_hotplug_event(mstm->mgr.dev); + } } static void @@ -1264,6 +1279,7 @@ nv50_mstm_del(struct nv50_mstm **pmstm) { struct nv50_mstm *mstm = *pmstm; if (mstm) { + drm_dp_mst_topology_mgr_destroy(&mstm->mgr); kfree(*pmstm); *pmstm = NULL; } @@ -2293,7 +2309,7 @@ nv50_display_create(struct drm_device *dev) /* create encoder/connector objects based on VBIOS DCB table */ for (i = 0, dcbe = &dcb->entry[0]; i < dcb->entries; i++, dcbe++) { - connector = nouveau_connector_create(dev, dcbe->connector); + connector = nouveau_connector_create(dev, dcbe); if (IS_ERR(connector)) continue; diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.h b/drivers/gpu/drm/nouveau/dispnv50/disp.h index e48c5eb35b49..2216c58620c2 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/disp.h +++ b/drivers/gpu/drm/nouveau/dispnv50/disp.h @@ -45,6 +45,8 @@ struct nv50_disp_interlock { void corec37d_ntfy_init(struct nouveau_bo *, u32); +void head907d_olut_load(struct drm_color_lut *, int size, void __iomem *); + struct nv50_chan { struct nvif_object user; struct nvif_device *device; diff --git a/drivers/gpu/drm/nouveau/dispnv50/head.c b/drivers/gpu/drm/nouveau/dispnv50/head.c index 4f57e5379796..ac97ebce5b35 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head.c +++ b/drivers/gpu/drm/nouveau/dispnv50/head.c @@ -50,9 +50,9 @@ nv50_head_flush_set(struct nv50_head *head, struct nv50_head_atom *asyh) if (asyh->set.core ) head->func->core_set(head, asyh); if (asyh->set.olut ) { asyh->olut.offset = nv50_lut_load(&head->olut, - asyh->olut.mode <= 1, asyh->olut.buffer, - asyh->state.gamma_lut); + asyh->state.gamma_lut, + asyh->olut.load); head->func->olut_set(head, asyh); } if (asyh->set.curs ) head->func->curs_set(head, asyh); @@ -210,7 +210,7 @@ nv50_head_atomic_check_lut(struct nv50_head *head, } } - if (!olut) { + if (!olut && !head->func->olut_identity) { asyh->olut.handle = 0; return 0; } diff --git a/drivers/gpu/drm/nouveau/dispnv50/head.h b/drivers/gpu/drm/nouveau/dispnv50/head.h index 37b3248c6dae..d1c002f534d4 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head.h +++ b/drivers/gpu/drm/nouveau/dispnv50/head.h @@ -21,6 +21,7 @@ struct nv50_head_func { void (*view)(struct nv50_head *, struct nv50_head_atom *); void (*mode)(struct nv50_head *, struct nv50_head_atom *); void (*olut)(struct nv50_head *, struct nv50_head_atom *); + bool olut_identity; void (*olut_set)(struct nv50_head *, struct nv50_head_atom *); void (*olut_clr)(struct nv50_head *); void (*core_calc)(struct nv50_head *, struct nv50_head_atom *); @@ -75,4 +76,14 @@ int head917d_curs_layout(struct nv50_head *, struct nv50_wndw_atom *, struct nv50_head_atom *); extern const struct nv50_head_func headc37d; +void headc37d_view(struct nv50_head *, struct nv50_head_atom *); +void headc37d_core_set(struct nv50_head *, struct nv50_head_atom *); +void headc37d_core_clr(struct nv50_head *); +int headc37d_curs_format(struct nv50_head *, struct nv50_wndw_atom *, + struct nv50_head_atom *); +void headc37d_curs_set(struct nv50_head *, struct nv50_head_atom *); +void headc37d_curs_clr(struct nv50_head *); +void headc37d_dither(struct nv50_head *, struct nv50_head_atom *); + +extern const struct nv50_head_func headc57d; #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/head507d.c b/drivers/gpu/drm/nouveau/dispnv50/head507d.c index 51bc5996fd37..7561be5ca707 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head507d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/head507d.c @@ -254,6 +254,23 @@ head507d_olut_set(struct nv50_head *head, struct nv50_head_atom *asyh) } } +static void +head507d_olut_load(struct drm_color_lut *in, int size, void __iomem *mem) +{ + for (; size--; in++, mem += 8) { + writew(drm_color_lut_extract(in-> red, 11) << 3, mem + 0); + writew(drm_color_lut_extract(in->green, 11) << 3, mem + 2); + writew(drm_color_lut_extract(in-> blue, 11) << 3, mem + 4); + } + + /* INTERPOLATE modes require a "next" entry to interpolate with, + * so we replicate the last entry to deal with this for now. + */ + writew(readw(mem - 8), mem + 0); + writew(readw(mem - 6), mem + 2); + writew(readw(mem - 4), mem + 4); +} + void head507d_olut(struct nv50_head *head, struct nv50_head_atom *asyh) { @@ -261,6 +278,8 @@ head507d_olut(struct nv50_head *head, struct nv50_head_atom *asyh) asyh->olut.mode = 0; else asyh->olut.mode = 1; + + asyh->olut.load = head507d_olut_load; } void diff --git a/drivers/gpu/drm/nouveau/dispnv50/head907d.c b/drivers/gpu/drm/nouveau/dispnv50/head907d.c index 633907163eb1..c2d09dd97b1f 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/head907d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/head907d.c @@ -214,9 +214,27 @@ head907d_olut_set(struct nv50_head *head, struct nv50_head_atom *asyh) } void +head907d_olut_load(struct drm_color_lut *in, int size, void __iomem *mem) +{ + for (; size--; in++, mem += 8) { + writew(drm_color_lut_extract(in-> red, 14) + 0x6000, mem + 0); + writew(drm_color_lut_extract(in->green, 14) + 0x6000, mem + 2); + writew(drm_color_lut_extract(in-> blue, 14) + 0x6000, mem + 4); + } + + /* INTERPOLATE modes require a "next" entry to interpolate with, + * so we replicate the last entry to deal with this for now. + */ + writew(readw(mem - 8), mem + 0); + writew(readw(mem - 6), mem + 2); + writew(readw(mem - 4), mem + 4); +} + +void head907d_olut(struct nv50_head *head, struct nv50_head_atom *asyh) { asyh->olut.mode = 7; + asyh->olut.load = head907d_olut_load; } void diff --git a/drivers/gpu/drm/nouveau/dispnv50/headc37d.c b/drivers/gpu/drm/nouveau/dispnv50/headc37d.c index 989c14083066..ef6a99d95a9c 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/headc37d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/headc37d.c @@ -65,7 +65,7 @@ headc37d_procamp(struct nv50_head *head, struct nv50_head_atom *asyh) } } -static void +void headc37d_dither(struct nv50_head *head, struct nv50_head_atom *asyh) { struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; @@ -79,7 +79,7 @@ headc37d_dither(struct nv50_head *head, struct nv50_head_atom *asyh) } } -static void +void headc37d_curs_clr(struct nv50_head *head) { struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; @@ -93,7 +93,7 @@ headc37d_curs_clr(struct nv50_head *head) } } -static void +void headc37d_curs_set(struct nv50_head *head, struct nv50_head_atom *asyh) { struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; @@ -112,7 +112,7 @@ headc37d_curs_set(struct nv50_head *head, struct nv50_head_atom *asyh) } } -static int +int headc37d_curs_format(struct nv50_head *head, struct nv50_wndw_atom *asyw, struct nv50_head_atom *asyh) { @@ -155,6 +155,7 @@ headc37d_olut(struct nv50_head *head, struct nv50_head_atom *asyh) asyh->olut.size = 0; asyh->olut.range = 0; asyh->olut.output_mode = 1; + asyh->olut.load = head907d_olut_load; } static void @@ -181,7 +182,7 @@ headc37d_mode(struct nv50_head *head, struct nv50_head_atom *asyh) } } -static void +void headc37d_view(struct nv50_head *head, struct nv50_head_atom *asyh) { struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; diff --git a/drivers/gpu/drm/nouveau/dispnv50/headc57d.c b/drivers/gpu/drm/nouveau/dispnv50/headc57d.c new file mode 100644 index 000000000000..32a7f9e85fb0 --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/headc57d.c @@ -0,0 +1,206 @@ +/* + * Copyright 2018 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. + */ +#include "head.h" +#include "atom.h" +#include "core.h" + +static void +headc57d_or(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; + u32 *push; + if ((push = evo_wait(core, 2))) { + /*XXX: This is a dirty hack until OR depth handling is + * improved later for deep colour etc. + */ + switch (asyh->or.depth) { + case 6: asyh->or.depth = 5; break; + case 5: asyh->or.depth = 4; break; + case 2: asyh->or.depth = 1; break; + case 0: asyh->or.depth = 4; break; + default: + WARN_ON(1); + break; + } + + evo_mthd(push, 0x2004 + (head->base.index * 0x400), 1); + evo_data(push, 0xfc000001 | + asyh->or.depth << 4 | + asyh->or.nvsync << 3 | + asyh->or.nhsync << 2); + evo_kick(push, core); + } +} + +static void +headc57d_procamp(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; + u32 *push; + if ((push = evo_wait(core, 2))) { + evo_mthd(push, 0x2000 + (head->base.index * 0x400), 1); +#if 0 + evo_data(push, 0x80000000 | + asyh->procamp.sat.sin << 16 | + asyh->procamp.sat.cos << 4); +#else + evo_data(push, 0); +#endif + evo_kick(push, core); + } +} + +void +headc57d_olut_clr(struct nv50_head *head) +{ + struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; + u32 *push; + if ((push = evo_wait(core, 2))) { + evo_mthd(push, 0x2288 + (head->base.index * 0x400), 1); + evo_data(push, 0x00000000); + evo_kick(push, core); + } +} + +void +headc57d_olut_set(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; + u32 *push; + if ((push = evo_wait(core, 4))) { + evo_mthd(push, 0x2280 + (head->base.index * 0x400), 4); + evo_data(push, asyh->olut.size << 8 | + asyh->olut.mode << 2 | + asyh->olut.output_mode); + evo_data(push, 0xffffffff); /* FP_NORM_SCALE. */ + evo_data(push, asyh->olut.handle); + evo_data(push, asyh->olut.offset >> 8); + evo_kick(push, core); + } +} + +static void +headc57d_olut_load_8(struct drm_color_lut *in, int size, void __iomem *mem) +{ + memset_io(mem, 0x00, 0x20); /* VSS header. */ + mem += 0x20; + + while (size--) { + u16 r = drm_color_lut_extract(in-> red + 0, 16); + u16 g = drm_color_lut_extract(in->green + 0, 16); + u16 b = drm_color_lut_extract(in-> blue + 0, 16); + u16 ri = 0, gi = 0, bi = 0, i; + + if (in++, size) { + ri = (drm_color_lut_extract(in-> red, 16) - r) / 4; + gi = (drm_color_lut_extract(in->green, 16) - g) / 4; + bi = (drm_color_lut_extract(in-> blue, 16) - b) / 4; + } + + for (i = 0; i < 4; i++, mem += 8) { + writew(r + ri * i, mem + 0); + writew(g + gi * i, mem + 2); + writew(b + bi * i, mem + 4); + } + } + + /* INTERPOLATE modes require a "next" entry to interpolate with, + * so we replicate the last entry to deal with this for now. + */ + writew(readw(mem - 8), mem + 0); + writew(readw(mem - 6), mem + 2); + writew(readw(mem - 4), mem + 4); +} + +static void +headc57d_olut_load(struct drm_color_lut *in, int size, void __iomem *mem) +{ + memset_io(mem, 0x00, 0x20); /* VSS header. */ + mem += 0x20; + + for (; size--; in++, mem += 0x08) { + writew(drm_color_lut_extract(in-> red, 16), mem + 0); + writew(drm_color_lut_extract(in->green, 16), mem + 2); + writew(drm_color_lut_extract(in-> blue, 16), mem + 4); + } + + /* INTERPOLATE modes require a "next" entry to interpolate with, + * so we replicate the last entry to deal with this for now. + */ + writew(readw(mem - 8), mem + 0); + writew(readw(mem - 6), mem + 2); + writew(readw(mem - 4), mem + 4); +} + +void +headc57d_olut(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + asyh->olut.mode = 2; /* DIRECT10 */ + asyh->olut.size = 4 /* VSS header. */ + 1024 + 1 /* Entries. */; + asyh->olut.output_mode = 1; /* INTERPOLATE_ENABLE. */ + if (asyh->state.gamma_lut && + asyh->state.gamma_lut->length / sizeof(struct drm_color_lut) == 256) + asyh->olut.load = headc57d_olut_load_8; + else + asyh->olut.load = headc57d_olut_load; +} + +static void +headc57d_mode(struct nv50_head *head, struct nv50_head_atom *asyh) +{ + struct nv50_dmac *core = &nv50_disp(head->base.base.dev)->core->chan; + struct nv50_head_mode *m = &asyh->mode; + u32 *push; + if ((push = evo_wait(core, 12))) { + evo_mthd(push, 0x2064 + (head->base.index * 0x400), 5); + evo_data(push, (m->v.active << 16) | m->h.active ); + evo_data(push, (m->v.synce << 16) | m->h.synce ); + evo_data(push, (m->v.blanke << 16) | m->h.blanke ); + evo_data(push, (m->v.blanks << 16) | m->h.blanks ); + evo_data(push, (m->v.blank2e << 16) | m->v.blank2s); + evo_mthd(push, 0x200c + (head->base.index * 0x400), 1); + evo_data(push, m->clock * 1000); + evo_mthd(push, 0x2028 + (head->base.index * 0x400), 1); + evo_data(push, m->clock * 1000); + /*XXX: HEAD_USAGE_BOUNDS, doesn't belong here. */ + evo_mthd(push, 0x2030 + (head->base.index * 0x400), 1); + evo_data(push, 0x00001014); + evo_kick(push, core); + } +} + +const struct nv50_head_func +headc57d = { + .view = headc37d_view, + .mode = headc57d_mode, + .olut = headc57d_olut, + .olut_identity = true, + .olut_set = headc57d_olut_set, + .olut_clr = headc57d_olut_clr, + .curs_layout = head917d_curs_layout, + .curs_format = headc37d_curs_format, + .curs_set = headc37d_curs_set, + .curs_clr = headc37d_curs_clr, + .dither = headc37d_dither, + .procamp = headc57d_procamp, + .or = headc57d_or, +}; diff --git a/drivers/gpu/drm/nouveau/dispnv50/lut.c b/drivers/gpu/drm/nouveau/dispnv50/lut.c index a6b96ae2a22f..994def4fd51a 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/lut.c +++ b/drivers/gpu/drm/nouveau/dispnv50/lut.c @@ -29,45 +29,29 @@ #include <nvif/class.h> u32 -nv50_lut_load(struct nv50_lut *lut, bool legacy, int buffer, - struct drm_property_blob *blob) +nv50_lut_load(struct nv50_lut *lut, int buffer, struct drm_property_blob *blob, + void (*load)(struct drm_color_lut *, int, void __iomem *)) { - struct drm_color_lut *in = (struct drm_color_lut *)blob->data; + struct drm_color_lut *in = blob ? blob->data : NULL; void __iomem *mem = lut->mem[buffer].object.map.ptr; - const int size = blob->length / sizeof(*in); - int bits, shift, i; - u16 zero, r, g, b; - u32 addr = lut->mem[buffer].addr; - - /* This can't happen.. But it shuts the compiler up. */ - if (WARN_ON(size != 256)) - return 0; + const u32 addr = lut->mem[buffer].addr; + int i; - if (legacy) { - bits = 11; - shift = 3; - zero = 0x0000; + if (!in) { + in = kvmalloc_array(1024, sizeof(*in), GFP_KERNEL); + if (!WARN_ON(!in)) { + for (i = 0; i < 1024; i++) { + in[i].red = + in[i].green = + in[i].blue = (i << 16) >> 10; + } + load(in, 1024, mem); + kvfree(in); + } } else { - bits = 14; - shift = 0; - zero = 0x6000; - } - - for (i = 0; i < size; i++) { - r = (drm_color_lut_extract(in[i]. red, bits) + zero) << shift; - g = (drm_color_lut_extract(in[i].green, bits) + zero) << shift; - b = (drm_color_lut_extract(in[i]. blue, bits) + zero) << shift; - writew(r, mem + (i * 0x08) + 0); - writew(g, mem + (i * 0x08) + 2); - writew(b, mem + (i * 0x08) + 4); + load(in, blob->length / sizeof(*in), mem); } - /* INTERPOLATE modes require a "next" entry to interpolate with, - * so we replicate the last entry to deal with this for now. - */ - writew(r, mem + (i * 0x08) + 0); - writew(g, mem + (i * 0x08) + 2); - writew(b, mem + (i * 0x08) + 4); return addr; } diff --git a/drivers/gpu/drm/nouveau/dispnv50/lut.h b/drivers/gpu/drm/nouveau/dispnv50/lut.h index 6d7b8352e4cb..b3b9040cfe9a 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/lut.h +++ b/drivers/gpu/drm/nouveau/dispnv50/lut.h @@ -2,6 +2,7 @@ #define __NV50_KMS_LUT_H__ #include <nvif/mem.h> struct drm_property_blob; +struct drm_color_lut; struct nv50_disp; struct nv50_lut { @@ -10,6 +11,6 @@ struct nv50_lut { int nv50_lut_init(struct nv50_disp *, struct nvif_mmu *, struct nv50_lut *); void nv50_lut_fini(struct nv50_lut *); -u32 nv50_lut_load(struct nv50_lut *, bool legacy, int buffer, - struct drm_property_blob *); +u32 nv50_lut_load(struct nv50_lut *, int buffer, struct drm_property_blob *, + void (*)(struct drm_color_lut *, int size, void __iomem *)); #endif diff --git a/drivers/gpu/drm/nouveau/dispnv50/wimm.c b/drivers/gpu/drm/nouveau/dispnv50/wimm.c index fc36e0696407..bc9eeaf212ae 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wimm.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wimm.c @@ -31,6 +31,7 @@ nv50_wimm_init(struct nouveau_drm *drm, struct nv50_wndw *wndw) int version; int (*init)(struct nouveau_drm *, s32, struct nv50_wndw *); } wimms[] = { + { TU104_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, { GV100_DISP_WINDOW_IMM_CHANNEL_DMA, 0, wimmc37b_init }, {} }; diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.c b/drivers/gpu/drm/nouveau/dispnv50/wndw.c index 2187922e8dc2..ba9eea2ff16b 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndw.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.c @@ -139,10 +139,8 @@ nv50_wndw_flush_set(struct nv50_wndw *wndw, u32 *interlock, if (asyw->set.xlut ) { if (asyw->ilut) { asyw->xlut.i.offset = - nv50_lut_load(&wndw->ilut, - asyw->xlut.i.mode <= 1, - asyw->xlut.i.buffer, - asyw->ilut); + nv50_lut_load(&wndw->ilut, asyw->xlut.i.buffer, + asyw->ilut, asyw->xlut.i.load); } wndw->func->xlut_set(wndw, asyw); } @@ -322,6 +320,11 @@ nv50_wndw_atomic_check_lut(struct nv50_wndw *wndw, asyh->wndw.olut &= ~BIT(wndw->id); } + if (!ilut && wndw->func->ilut_identity) { + static struct drm_property_blob dummy = {}; + ilut = &dummy; + } + /* Recalculate LUT state. */ memset(&asyw->xlut, 0x00, sizeof(asyw->xlut)); if ((asyw->ilut = wndw->func->ilut ? ilut : NULL)) { @@ -623,6 +626,7 @@ nv50_wndw_new(struct nouveau_drm *drm, enum drm_plane_type type, int index, int (*new)(struct nouveau_drm *, enum drm_plane_type, int, s32, struct nv50_wndw **); } wndws[] = { + { TU104_DISP_WINDOW_CHANNEL_DMA, 0, wndwc57e_new }, { GV100_DISP_WINDOW_CHANNEL_DMA, 0, wndwc37e_new }, {} }; diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.h b/drivers/gpu/drm/nouveau/dispnv50/wndw.h index b0b6428034b0..03f3d8dc235a 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndw.h +++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.h @@ -65,6 +65,7 @@ struct nv50_wndw_func { int (*ntfy_wait_begun)(struct nouveau_bo *, u32 offset, struct nvif_device *); void (*ilut)(struct nv50_wndw *, struct nv50_wndw_atom *); + bool ilut_identity; bool olut_core; void (*xlut_set)(struct nv50_wndw *, struct nv50_wndw_atom *); void (*xlut_clr)(struct nv50_wndw *); @@ -90,6 +91,23 @@ extern const struct nv50_wimm_func curs507a; int wndwc37e_new(struct nouveau_drm *, enum drm_plane_type, int, s32, struct nv50_wndw **); +int wndwc37e_new_(const struct nv50_wndw_func *, struct nouveau_drm *, + enum drm_plane_type type, int index, s32 oclass, u32 heads, + struct nv50_wndw **); +int wndwc37e_acquire(struct nv50_wndw *, struct nv50_wndw_atom *, + struct nv50_head_atom *); +void wndwc37e_release(struct nv50_wndw *, struct nv50_wndw_atom *, + struct nv50_head_atom *); +void wndwc37e_sema_set(struct nv50_wndw *, struct nv50_wndw_atom *); +void wndwc37e_sema_clr(struct nv50_wndw *); +void wndwc37e_ntfy_set(struct nv50_wndw *, struct nv50_wndw_atom *); +void wndwc37e_ntfy_clr(struct nv50_wndw *); +void wndwc37e_image_set(struct nv50_wndw *, struct nv50_wndw_atom *); +void wndwc37e_image_clr(struct nv50_wndw *); +void wndwc37e_update(struct nv50_wndw *, u32 *); + +int wndwc57e_new(struct nouveau_drm *, enum drm_plane_type, int, s32, + struct nv50_wndw **); int nv50_wndw_new(struct nouveau_drm *, enum drm_plane_type, int index, struct nv50_wndw **); diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c index 44afb0f069a5..e52a85c83f7a 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c +++ b/drivers/gpu/drm/nouveau/dispnv50/wndwc37e.c @@ -61,9 +61,10 @@ wndwc37e_ilut(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) asyw->xlut.i.size = 0; asyw->xlut.i.range = 0; asyw->xlut.i.output_mode = 1; + asyw->xlut.i.load = head907d_olut_load; } -static void +void wndwc37e_image_clr(struct nv50_wndw *wndw) { u32 *push; @@ -76,7 +77,7 @@ wndwc37e_image_clr(struct nv50_wndw *wndw) } } -static void +void wndwc37e_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) { u32 *push; @@ -117,7 +118,7 @@ wndwc37e_image_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) evo_kick(push, &wndw->wndw); } -static void +void wndwc37e_ntfy_clr(struct nv50_wndw *wndw) { u32 *push; @@ -128,7 +129,7 @@ wndwc37e_ntfy_clr(struct nv50_wndw *wndw) } } -static void +void wndwc37e_ntfy_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) { u32 *push; @@ -140,7 +141,7 @@ wndwc37e_ntfy_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) } } -static void +void wndwc37e_sema_clr(struct nv50_wndw *wndw) { u32 *push; @@ -151,7 +152,7 @@ wndwc37e_sema_clr(struct nv50_wndw *wndw) } } -static void +void wndwc37e_sema_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) { u32 *push; @@ -165,7 +166,7 @@ wndwc37e_sema_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) } } -static void +void wndwc37e_update(struct nv50_wndw *wndw, u32 *interlock) { u32 *push; @@ -183,13 +184,13 @@ wndwc37e_update(struct nv50_wndw *wndw, u32 *interlock) } } -static void +void wndwc37e_release(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw, struct nv50_head_atom *asyh) { } -static int +int wndwc37e_acquire(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw, struct nv50_head_atom *asyh) { @@ -236,7 +237,7 @@ wndwc37e = { .update = wndwc37e_update, }; -static int +int wndwc37e_new_(const struct nv50_wndw_func *func, struct nouveau_drm *drm, enum drm_plane_type type, int index, s32 oclass, u32 heads, struct nv50_wndw **pwndw) diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c b/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c new file mode 100644 index 000000000000..ba89f1a5fcfa --- /dev/null +++ b/drivers/gpu/drm/nouveau/dispnv50/wndwc57e.c @@ -0,0 +1,133 @@ +/* + * Copyright 2018 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. + */ +#include "wndw.h" +#include "atom.h" + +#include <drm/drm_atomic_helper.h> +#include <drm/drm_plane_helper.h> +#include <nouveau_bo.h> + +#include <nvif/clc37e.h> + +static void +wndwc57e_ilut_clr(struct nv50_wndw *wndw) +{ + u32 *push; + if ((push = evo_wait(&wndw->wndw, 2))) { + evo_mthd(push, 0x0444, 1); + evo_data(push, 0x00000000); + evo_kick(push, &wndw->wndw); + } +} + +static void +wndwc57e_ilut_set(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) +{ + u32 *push; + if ((push = evo_wait(&wndw->wndw, 4))) { + evo_mthd(push, 0x0440, 3); + evo_data(push, asyw->xlut.i.size << 8 | + asyw->xlut.i.mode << 2 | + asyw->xlut.i.output_mode); + evo_data(push, asyw->xlut.handle); + evo_data(push, asyw->xlut.i.offset >> 8); + evo_kick(push, &wndw->wndw); + } +} + +static u16 +fixedU0_16_FP16(u16 fixed) +{ + int sign = 0, exp = 0, man = 0; + if (fixed) { + while (--exp && !(fixed & 0x8000)) + fixed <<= 1; + man = ((fixed << 1) & 0xffc0) >> 6; + exp += 15; + } + return (sign << 15) | (exp << 10) | man; +} + +static void +wndwc57e_ilut_load(struct drm_color_lut *in, int size, void __iomem *mem) +{ + memset_io(mem, 0x00, 0x20); /* VSS header. */ + mem += 0x20; + + for (; size--; in++, mem += 0x08) { + u16 r = fixedU0_16_FP16(drm_color_lut_extract(in-> red, 16)); + u16 g = fixedU0_16_FP16(drm_color_lut_extract(in->green, 16)); + u16 b = fixedU0_16_FP16(drm_color_lut_extract(in-> blue, 16)); + writew(r, mem + 0); + writew(g, mem + 2); + writew(b, mem + 4); + } + + /* INTERPOLATE modes require a "next" entry to interpolate with, + * so we replicate the last entry to deal with this for now. + */ + writew(readw(mem - 8), mem + 0); + writew(readw(mem - 6), mem + 2); + writew(readw(mem - 4), mem + 4); +} + +static void +wndwc57e_ilut(struct nv50_wndw *wndw, struct nv50_wndw_atom *asyw) +{ + u16 size = asyw->ilut->length / sizeof(struct drm_color_lut); + if (size == 256) { + asyw->xlut.i.mode = 1; /* DIRECT8. */ + } else { + asyw->xlut.i.mode = 2; /* DIRECT10. */ + size = 1024; + } + asyw->xlut.i.size = 4 /* VSS header. */ + size + 1 /* Entries. */; + asyw->xlut.i.output_mode = 0; /* INTERPOLATE_DISABLE. */ + asyw->xlut.i.load = wndwc57e_ilut_load; +} + +static const struct nv50_wndw_func +wndwc57e = { + .acquire = wndwc37e_acquire, + .release = wndwc37e_release, + .sema_set = wndwc37e_sema_set, + .sema_clr = wndwc37e_sema_clr, + .ntfy_set = wndwc37e_ntfy_set, + .ntfy_clr = wndwc37e_ntfy_clr, + .ntfy_reset = corec37d_ntfy_init, + .ntfy_wait_begun = base507c_ntfy_wait_begun, + .ilut = wndwc57e_ilut, + .ilut_identity = true, + .xlut_set = wndwc57e_ilut_set, + .xlut_clr = wndwc57e_ilut_clr, + .image_set = wndwc37e_image_set, + .image_clr = wndwc37e_image_clr, + .update = wndwc37e_update, +}; + +int +wndwc57e_new(struct nouveau_drm *drm, enum drm_plane_type type, int index, + s32 oclass, struct nv50_wndw **pwndw) +{ + return wndwc37e_new_(&wndwc57e, drm, type, index, oclass, + BIT(index >> 1), pwndw); +} diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h index 4f5233107f5f..4cbed0329367 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h +++ b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h @@ -32,6 +32,7 @@ struct nv_device_info_v0 { #define NV_DEVICE_INFO_V0_MAXWELL 0x09 #define NV_DEVICE_INFO_V0_PASCAL 0x0a #define NV_DEVICE_INFO_V0_VOLTA 0x0b +#define NV_DEVICE_INFO_V0_TURING 0x0c __u8 family; __u8 pad06[2]; __u64 ram_size; diff --git a/drivers/gpu/drm/nouveau/include/nvif/cla06f.h b/drivers/gpu/drm/nouveau/include/nvif/cla06f.h index fbfcffc5feb2..81401eb970ea 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/cla06f.h +++ b/drivers/gpu/drm/nouveau/include/nvif/cla06f.h @@ -4,12 +4,13 @@ struct kepler_channel_gpfifo_a_v0 { __u8 version; - __u8 pad01[1]; + __u8 priv; __u16 chid; __u32 ilength; __u64 ioffset; __u64 runlist; __u64 vmm; + __u64 inst; }; #define NVA06F_V0_NTFY_NON_STALL_INTERRUPT 0x00 diff --git a/drivers/gpu/drm/nouveau/include/nvif/class.h b/drivers/gpu/drm/nouveau/include/nvif/class.h index 6db56bd7d67e..1d82cbf70cf4 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/class.h +++ b/drivers/gpu/drm/nouveau/include/nvif/class.h @@ -68,7 +68,8 @@ #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 VOLTA_CHANNEL_GPFIFO_A /* cla06f.h */ 0x0000c36f +#define VOLTA_CHANNEL_GPFIFO_A /* clc36f.h */ 0x0000c36f +#define TURING_CHANNEL_GPFIFO_A /* clc36f.h */ 0x0000c46f #define NV50_DISP /* cl5070.h */ 0x00005070 #define G82_DISP /* cl5070.h */ 0x00008270 @@ -83,6 +84,7 @@ #define GP100_DISP /* cl5070.h */ 0x00009770 #define GP102_DISP /* cl5070.h */ 0x00009870 #define GV100_DISP /* cl5070.h */ 0x0000c370 +#define TU104_DISP /* cl5070.h */ 0x0000c570 #define NV31_MPEG 0x00003174 #define G82_MPEG 0x00008274 @@ -95,6 +97,7 @@ #define GF110_DISP_CURSOR /* cl507a.h */ 0x0000907a #define GK104_DISP_CURSOR /* cl507a.h */ 0x0000917a #define GV100_DISP_CURSOR /* cl507a.h */ 0x0000c37a +#define TU104_DISP_CURSOR /* cl507a.h */ 0x0000c57a #define NV50_DISP_OVERLAY /* cl507b.h */ 0x0000507b #define G82_DISP_OVERLAY /* cl507b.h */ 0x0000827b @@ -103,6 +106,7 @@ #define GK104_DISP_OVERLAY /* cl507b.h */ 0x0000917b #define GV100_DISP_WINDOW_IMM_CHANNEL_DMA /* clc37b.h */ 0x0000c37b +#define TU104_DISP_WINDOW_IMM_CHANNEL_DMA /* clc37b.h */ 0x0000c57b #define NV50_DISP_BASE_CHANNEL_DMA /* cl507c.h */ 0x0000507c #define G82_DISP_BASE_CHANNEL_DMA /* cl507c.h */ 0x0000827c @@ -125,6 +129,7 @@ #define GP100_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000977d #define GP102_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000987d #define GV100_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000c37d +#define TU104_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000c57d #define NV50_DISP_OVERLAY_CHANNEL_DMA /* cl507e.h */ 0x0000507e #define G82_DISP_OVERLAY_CHANNEL_DMA /* cl507e.h */ 0x0000827e @@ -134,6 +139,7 @@ #define GK104_DISP_OVERLAY_CONTROL_DMA /* cl507e.h */ 0x0000917e #define GV100_DISP_WINDOW_CHANNEL_DMA /* clc37e.h */ 0x0000c37e +#define TU104_DISP_WINDOW_CHANNEL_DMA /* clc37e.h */ 0x0000c57e #define NV50_TESLA 0x00005097 #define G82_TESLA 0x00008297 @@ -183,6 +189,7 @@ #define PASCAL_DMA_COPY_A 0x0000c0b5 #define PASCAL_DMA_COPY_B 0x0000c1b5 #define VOLTA_DMA_COPY_A 0x0000c3b5 +#define TURING_DMA_COPY_A 0x0000c5b5 #define FERMI_DECOMPRESS 0x000090b8 diff --git a/drivers/gpu/drm/nouveau/include/nvif/clc36f.h b/drivers/gpu/drm/nouveau/include/nvif/clc36f.h new file mode 100644 index 000000000000..6b14d7e3f6bb --- /dev/null +++ b/drivers/gpu/drm/nouveau/include/nvif/clc36f.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __NVIF_CLC36F_H__ +#define __NVIF_CLC36F_H__ + +struct volta_channel_gpfifo_a_v0 { + __u8 version; + __u8 priv; + __u16 chid; + __u32 ilength; + __u64 ioffset; + __u64 runlist; + __u64 vmm; + __u64 inst; + __u32 token; +}; + +#define NVC36F_V0_NTFY_NON_STALL_INTERRUPT 0x00 +#define NVC36F_V0_NTFY_KILLED 0x01 +#endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h index d83d834b7452..72e4dc1f0236 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h @@ -61,7 +61,11 @@ enum nvkm_devidx { NVKM_ENGINE_NVENC2, NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC2, - NVKM_ENGINE_NVDEC, + NVKM_ENGINE_NVDEC0, + NVKM_ENGINE_NVDEC1, + NVKM_ENGINE_NVDEC2, + NVKM_ENGINE_NVDEC_LAST = NVKM_ENGINE_NVDEC2, + NVKM_ENGINE_PM, NVKM_ENGINE_SEC, NVKM_ENGINE_SEC2, @@ -114,6 +118,7 @@ struct nvkm_device { GM100 = 0x110, GP100 = 0x130, GV100 = 0x140, + TU100 = 0x160, } card_type; u32 chipset; u8 chiprev; @@ -163,7 +168,7 @@ struct nvkm_device { struct nvkm_engine *msppp; struct nvkm_engine *msvld; struct nvkm_engine *nvenc[3]; - struct nvkm_nvdec *nvdec; + struct nvkm_nvdec *nvdec[3]; struct nvkm_pm *pm; struct nvkm_engine *sec; struct nvkm_sec2 *sec2; @@ -235,7 +240,7 @@ struct nvkm_device_chip { int (*msppp )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*msvld )(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_nvdec **); + int (*nvdec[3])(struct nvkm_device *, int idx, struct nvkm_nvdec **); int (*pm )(struct nvkm_device *, int idx, struct nvkm_pm **); int (*sec )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*sec2 )(struct nvkm_device *, int idx, struct nvkm_sec2 **); diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/memory.h b/drivers/gpu/drm/nouveau/include/nvkm/core/memory.h index 05f505de0075..f34c80310861 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/memory.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/memory.h @@ -29,6 +29,7 @@ struct nvkm_memory_func { void *(*dtor)(struct nvkm_memory *); enum nvkm_memory_target (*target)(struct nvkm_memory *); u8 (*page)(struct nvkm_memory *); + u64 (*bar2)(struct nvkm_memory *); u64 (*addr)(struct nvkm_memory *); u64 (*size)(struct nvkm_memory *); void (*boot)(struct nvkm_memory *, struct nvkm_vmm *); @@ -56,6 +57,7 @@ void nvkm_memory_tags_put(struct nvkm_memory *, struct nvkm_device *, #define nvkm_memory_target(p) (p)->func->target(p) #define nvkm_memory_page(p) (p)->func->page(p) +#define nvkm_memory_bar2(p) (p)->func->bar2(p) #define nvkm_memory_addr(p) (p)->func->addr(p) #define nvkm_memory_size(p) (p)->func->size(p) #define nvkm_memory_boot(p,v) (p)->func->boot((p),(v)) diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h index fc295e1faa19..86abe76023c2 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h @@ -11,4 +11,5 @@ int gm200_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gp100_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gp102_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gv100_ce_new(struct nvkm_device *, int, struct nvkm_engine **); +int tu104_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 ef7dc0844d26..5ca86e178bb9 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h @@ -36,4 +36,5 @@ int gm200_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gp100_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gp102_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gv100_disp_new(struct nvkm_device *, int, struct nvkm_disp **); +int tu104_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 7e39fbed2519..3b2b685778eb 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h @@ -74,4 +74,5 @@ int gm20b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gp100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gp10b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gv100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); +int tu104_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h index f6bd94c7e0f7..fd9d713b611c 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bar.h @@ -16,8 +16,10 @@ struct nvkm_bar { }; struct nvkm_vmm *nvkm_bar_bar1_vmm(struct nvkm_device *); +void nvkm_bar_bar1_reset(struct nvkm_device *); void nvkm_bar_bar2_init(struct nvkm_device *); void nvkm_bar_bar2_fini(struct nvkm_device *); +void nvkm_bar_bar2_reset(struct nvkm_device *); struct nvkm_vmm *nvkm_bar_bar2_vmm(struct nvkm_device *); void nvkm_bar_flush(struct nvkm_bar *); @@ -27,4 +29,5 @@ int gf100_bar_new(struct nvkm_device *, int, struct nvkm_bar **); int gk20a_bar_new(struct nvkm_device *, int, struct nvkm_bar **); int gm107_bar_new(struct nvkm_device *, int, struct nvkm_bar **); int gm20b_bar_new(struct nvkm_device *, int, struct nvkm_bar **); +int tu104_bar_new(struct nvkm_device *, int, struct nvkm_bar **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/M0203.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/M0203.h index 703a5b524b96..425ccc47e3b7 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/M0203.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/M0203.h @@ -12,11 +12,14 @@ u32 nvbios_M0203Tp(struct nvkm_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_M0203T *); struct nvbios_M0203E { -#define M0203E_TYPE_DDR2 0x0 -#define M0203E_TYPE_DDR3 0x1 -#define M0203E_TYPE_GDDR3 0x2 -#define M0203E_TYPE_GDDR5 0x3 -#define M0203E_TYPE_SKIP 0xf +#define M0203E_TYPE_DDR2 0x0 +#define M0203E_TYPE_DDR3 0x1 +#define M0203E_TYPE_GDDR3 0x2 +#define M0203E_TYPE_GDDR5 0x3 +#define M0203E_TYPE_HBM2 0x6 +#define M0203E_TYPE_GDDR5X 0x8 +#define M0203E_TYPE_GDDR6 0x9 +#define M0203E_TYPE_SKIP 0xf u8 type; u8 strap; u8 group; diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h index ed9e0a6a0011..8463b421d345 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/conn.h @@ -20,6 +20,7 @@ enum dcb_connector_type { DCB_CONNECTOR_DMS59_DP0 = 0x64, DCB_CONNECTOR_DMS59_DP1 = 0x65, DCB_CONNECTOR_WFD = 0x70, + DCB_CONNECTOR_USB_C = 0x71, DCB_CONNECTOR_NONE = 0xff }; diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h index 486e7635c29d..1b71812a790b 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/devinit.h @@ -31,4 +31,5 @@ int gf100_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **); int gm107_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **); int gm200_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **); int gv100_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **); +int tu104_devinit_new(struct nvkm_device *, int, struct nvkm_devinit **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h index 5a77498fe6a0..127f48066026 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fault.h @@ -30,4 +30,5 @@ struct nvkm_fault_data { int gp100_fault_new(struct nvkm_device *, int, struct nvkm_fault **); int gv100_fault_new(struct nvkm_device *, int, struct nvkm_fault **); +int tu104_fault_new(struct nvkm_device *, int, struct nvkm_fault **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h index 96ccc624ee81..27298f8b7ead 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h @@ -105,7 +105,10 @@ enum nvkm_ram_type { NVKM_RAM_TYPE_GDDR2, NVKM_RAM_TYPE_GDDR3, NVKM_RAM_TYPE_GDDR4, - NVKM_RAM_TYPE_GDDR5 + NVKM_RAM_TYPE_GDDR5, + NVKM_RAM_TYPE_GDDR5X, + NVKM_RAM_TYPE_GDDR6, + NVKM_RAM_TYPE_HBM2, }; struct nvkm_ram { diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h index 61c93c86e2e2..b66dedd8abb6 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h @@ -31,4 +31,5 @@ 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 **); int gp10b_mc_new(struct nvkm_device *, int, struct nvkm_mc **); +int tu104_mc_new(struct nvkm_device *, int, struct nvkm_mc **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h index 688595545e21..0a0e064f22e5 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mmu.h @@ -130,4 +130,5 @@ int gm20b_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **); int gp100_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **); int gp10b_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **); int gv100_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **); +int tu104_mmu_new(struct nvkm_device *, int, struct nvkm_mmu **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h index e9b0746826ca..3693ebf371b6 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h @@ -28,6 +28,18 @@ struct nvkm_timer { u64 nvkm_timer_read(struct nvkm_timer *); void nvkm_timer_alarm(struct nvkm_timer *, u32 nsec, struct nvkm_alarm *); +struct nvkm_timer_wait { + struct nvkm_timer *tmr; + u64 limit; + u64 time0; + u64 time1; + int reads; +}; + +void nvkm_timer_wait_init(struct nvkm_device *, u64 nsec, + struct nvkm_timer_wait *); +s64 nvkm_timer_wait_test(struct nvkm_timer_wait *); + /* Delay based on GPU time (ie. PTIMER). * * Will return -ETIMEDOUT unless the loop was terminated with 'break', @@ -38,21 +50,17 @@ void nvkm_timer_alarm(struct nvkm_timer *, u32 nsec, struct nvkm_alarm *); */ #define NVKM_DELAY _warn = false; #define nvkm_nsec(d,n,cond...) ({ \ - struct nvkm_device *_device = (d); \ - struct nvkm_timer *_tmr = _device->timer; \ - u64 _nsecs = (n), _time0 = nvkm_timer_read(_tmr); \ - s64 _taken = 0; \ + struct nvkm_timer_wait _wait; \ bool _warn = true; \ + s64 _taken = 0; \ \ + nvkm_timer_wait_init((d), (n), &_wait); \ do { \ cond \ - } while (_taken = nvkm_timer_read(_tmr) - _time0, _taken < _nsecs); \ + } while ((_taken = nvkm_timer_wait_test(&_wait)) >= 0); \ \ - if (_taken >= _nsecs) { \ - if (_warn) \ - dev_WARN(_device->dev, "timeout\n"); \ - _taken = -ETIMEDOUT; \ - } \ + if (_warn && _taken < 0) \ + dev_WARN(_wait.tmr->subdev.device->dev, "timeout\n"); \ _taken; \ }) #define nvkm_usec(d,u,cond...) nvkm_nsec((d), (u) * 1000, ##cond) diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c index e67a471331b5..b06cdac8f3a2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_abi16.c +++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c @@ -306,7 +306,7 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS) /* create channel object and initialise dma and fence management */ ret = nouveau_channel_new(drm, device, init->fb_ctxdma_handle, - init->tt_ctxdma_handle, &chan->chan); + init->tt_ctxdma_handle, false, &chan->chan); if (ret) goto done; diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c index 5f5be6368aed..c7a94c94dbf3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_backlight.c +++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c @@ -253,6 +253,9 @@ nouveau_backlight_init(struct drm_connector *connector) case NV_DEVICE_INFO_V0_FERMI: case NV_DEVICE_INFO_V0_KEPLER: case NV_DEVICE_INFO_V0_MAXWELL: + case NV_DEVICE_INFO_V0_PASCAL: + case NV_DEVICE_INFO_V0_VOLTA: + case NV_DEVICE_INFO_V0_TURING: ret = nv50_backlight_init(nv_encoder, &props, &ops); break; default: diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 7214022dfb91..73eff52036d2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -1141,6 +1141,8 @@ 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, 0xc5b5, nve0_bo_move_copy, nve0_bo_move_init }, + { "GRCE", 0, 0xc5b5, nve0_bo_move_copy, nvc0_bo_move_init }, { "COPY", 4, 0xc3b5, nve0_bo_move_copy, nve0_bo_move_init }, { "GRCE", 0, 0xc3b5, nve0_bo_move_copy, nvc0_bo_move_init }, { "COPY", 4, 0xc1b5, nve0_bo_move_copy, nve0_bo_move_init }, diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c index 92d3115f96b5..668afbc29c3e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_chan.c +++ b/drivers/gpu/drm/nouveau/nouveau_chan.c @@ -29,6 +29,7 @@ #include <nvif/cl506f.h> #include <nvif/cl906f.h> #include <nvif/cla06f.h> +#include <nvif/clc36f.h> #include <nvif/ioctl.h> /*XXX*/ @@ -217,10 +218,11 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nvif_device *device, static int nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device, - u64 runlist, struct nouveau_channel **pchan) + u64 runlist, bool priv, struct nouveau_channel **pchan) { struct nouveau_cli *cli = (void *)device->object.client; - static const u16 oclasses[] = { VOLTA_CHANNEL_GPFIFO_A, + static const u16 oclasses[] = { TURING_CHANNEL_GPFIFO_A, + VOLTA_CHANNEL_GPFIFO_A, PASCAL_CHANNEL_GPFIFO_A, MAXWELL_CHANNEL_GPFIFO_A, KEPLER_CHANNEL_GPFIFO_B, @@ -234,6 +236,7 @@ nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device, struct nv50_channel_gpfifo_v0 nv50; struct fermi_channel_gpfifo_v0 fermi; struct kepler_channel_gpfifo_a_v0 kepler; + struct volta_channel_gpfifo_a_v0 volta; } args; struct nouveau_channel *chan; u32 size; @@ -247,12 +250,22 @@ nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device, /* create channel object */ do { + if (oclass[0] >= VOLTA_CHANNEL_GPFIFO_A) { + args.volta.version = 0; + args.volta.ilength = 0x02000; + args.volta.ioffset = 0x10000 + chan->push.addr; + args.volta.runlist = runlist; + args.volta.vmm = nvif_handle(&cli->vmm.vmm.object); + args.volta.priv = priv; + size = sizeof(args.volta); + } else if (oclass[0] >= KEPLER_CHANNEL_GPFIFO_A) { args.kepler.version = 0; args.kepler.ilength = 0x02000; args.kepler.ioffset = 0x10000 + chan->push.addr; args.kepler.runlist = runlist; args.kepler.vmm = nvif_handle(&cli->vmm.vmm.object); + args.kepler.priv = priv; size = sizeof(args.kepler); } else if (oclass[0] >= FERMI_CHANNEL_GPFIFO) { @@ -273,13 +286,20 @@ nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device, ret = nvif_object_init(&device->object, 0, *oclass++, &args, size, &chan->user); if (ret == 0) { - if (chan->user.oclass >= KEPLER_CHANNEL_GPFIFO_A) + if (chan->user.oclass >= VOLTA_CHANNEL_GPFIFO_A) { + chan->chid = args.volta.chid; + chan->inst = args.volta.inst; + chan->token = args.volta.token; + } else + if (chan->user.oclass >= KEPLER_CHANNEL_GPFIFO_A) { chan->chid = args.kepler.chid; - else - if (chan->user.oclass >= FERMI_CHANNEL_GPFIFO) + chan->inst = args.kepler.inst; + } else + if (chan->user.oclass >= FERMI_CHANNEL_GPFIFO) { chan->chid = args.fermi.chid; - else + } else { chan->chid = args.nv50.chid; + } return ret; } } while (*oclass); @@ -448,7 +468,8 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart) int nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device, - u32 arg0, u32 arg1, struct nouveau_channel **pchan) + u32 arg0, u32 arg1, bool priv, + struct nouveau_channel **pchan) { struct nouveau_cli *cli = (void *)device->object.client; bool super; @@ -458,7 +479,7 @@ nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device, super = cli->base.super; cli->base.super = true; - ret = nouveau_channel_ind(drm, device, arg0, pchan); + ret = nouveau_channel_ind(drm, device, arg0, priv, pchan); if (ret) { NV_PRINTK(dbg, cli, "ib channel create, %d\n", ret); ret = nouveau_channel_dma(drm, device, pchan); diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.h b/drivers/gpu/drm/nouveau/nouveau_chan.h index 64454c2ebd90..28418f4e5748 100644 --- a/drivers/gpu/drm/nouveau/nouveau_chan.h +++ b/drivers/gpu/drm/nouveau/nouveau_chan.h @@ -10,6 +10,8 @@ struct nouveau_channel { struct nouveau_drm *drm; int chid; + u64 inst; + u32 token; struct nvif_object vram; struct nvif_object gart; @@ -48,7 +50,8 @@ struct nouveau_channel { int nouveau_channels_init(struct nouveau_drm *); int nouveau_channel_new(struct nouveau_drm *, struct nvif_device *, - u32 arg0, u32 arg1, struct nouveau_channel **); + u32 arg0, u32 arg1, bool priv, + struct nouveau_channel **); void nouveau_channel_del(struct nouveau_channel **); int nouveau_channel_idle(struct nouveau_channel *); diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index fd80661dff92..3f463c91314a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -403,6 +403,7 @@ nouveau_connector_destroy(struct drm_connector *connector) if (nv_connector->aux.transfer) { drm_dp_cec_unregister_connector(&nv_connector->aux); drm_dp_aux_unregister(&nv_connector->aux); + kfree(nv_connector->aux.name); } kfree(connector); } @@ -1218,7 +1219,8 @@ drm_conntype_from_dcb(enum dcb_connector_type dcb) case DCB_CONNECTOR_LVDS_SPWG: return DRM_MODE_CONNECTOR_LVDS; case DCB_CONNECTOR_DMS59_DP0: case DCB_CONNECTOR_DMS59_DP1: - case DCB_CONNECTOR_DP : return DRM_MODE_CONNECTOR_DisplayPort; + case DCB_CONNECTOR_DP : + case DCB_CONNECTOR_USB_C : return DRM_MODE_CONNECTOR_DisplayPort; case DCB_CONNECTOR_eDP : return DRM_MODE_CONNECTOR_eDP; case DCB_CONNECTOR_HDMI_0 : case DCB_CONNECTOR_HDMI_1 : @@ -1232,7 +1234,8 @@ drm_conntype_from_dcb(enum dcb_connector_type dcb) } struct drm_connector * -nouveau_connector_create(struct drm_device *dev, int index) +nouveau_connector_create(struct drm_device *dev, + const struct dcb_output *dcbe) { const struct drm_connector_funcs *funcs = &nouveau_connector_funcs; struct nouveau_drm *drm = nouveau_drm(dev); @@ -1240,6 +1243,8 @@ nouveau_connector_create(struct drm_device *dev, int index) struct nouveau_connector *nv_connector = NULL; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; + char aux_name[48] = {0}; + int index = dcbe->connector; int type, ret = 0; bool dummy; @@ -1342,6 +1347,9 @@ nouveau_connector_create(struct drm_device *dev, int index) case DRM_MODE_CONNECTOR_eDP: nv_connector->aux.dev = dev->dev; nv_connector->aux.transfer = nouveau_connector_aux_xfer; + snprintf(aux_name, sizeof(aux_name), "sor-%04x-%04x", + dcbe->hasht, dcbe->hashm); + nv_connector->aux.name = kstrdup(aux_name, GFP_KERNEL); ret = drm_dp_aux_register(&nv_connector->aux); if (ret) { NV_ERROR(drm, "failed to register aux channel\n"); diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h index f57ef35b1e5e..f43a8d63aef8 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.h +++ b/drivers/gpu/drm/nouveau/nouveau_connector.h @@ -38,6 +38,7 @@ #include "nouveau_encoder.h" struct nvkm_i2c_port; +struct dcb_output; #ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT struct nouveau_backlight; @@ -113,7 +114,7 @@ nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc) } struct drm_connector * -nouveau_connector_create(struct drm_device *, int index); +nouveau_connector_create(struct drm_device *, const struct dcb_output *); extern int nouveau_tv_disable; extern int nouveau_ignorelid; diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c index 9109b69cd052..88a52f6b39fe 100644 --- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c +++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c @@ -47,6 +47,26 @@ nouveau_debugfs_vbios_image(struct seq_file *m, void *data) } static int +nouveau_debugfs_strap_peek(struct seq_file *m, void *data) +{ + struct drm_info_node *node = m->private; + struct nouveau_drm *drm = nouveau_drm(node->minor->dev); + int ret; + + ret = pm_runtime_get_sync(drm->dev->dev); + if (ret < 0 && ret != -EACCES) + return ret; + + seq_printf(m, "0x%08x\n", + nvif_rd32(&drm->client.device.object, 0x101000)); + + pm_runtime_mark_last_busy(drm->dev->dev); + pm_runtime_put_autosuspend(drm->dev->dev); + + return 0; +} + +static int nouveau_debugfs_pstate_get(struct seq_file *m, void *data) { struct drm_device *drm = m->private; @@ -185,7 +205,8 @@ static const struct file_operations nouveau_pstate_fops = { }; static struct drm_info_list nouveau_debugfs_list[] = { - { "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL }, + { "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL }, + { "strap_peek", nouveau_debugfs_strap_peek, 0, NULL }, }; #define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list) @@ -199,8 +220,9 @@ static const struct nouveau_debugfs_files { int nouveau_drm_debugfs_init(struct drm_minor *minor) { + struct nouveau_drm *drm = nouveau_drm(minor->dev); struct dentry *dentry; - int i; + int i, ret; for (i = 0; i < ARRAY_SIZE(nouveau_debugfs_files); i++) { dentry = debugfs_create_file(nouveau_debugfs_files[i].name, @@ -211,9 +233,23 @@ nouveau_drm_debugfs_init(struct drm_minor *minor) return -ENOMEM; } - return drm_debugfs_create_files(nouveau_debugfs_list, - NOUVEAU_DEBUGFS_ENTRIES, - minor->debugfs_root, minor); + ret = drm_debugfs_create_files(nouveau_debugfs_list, + NOUVEAU_DEBUGFS_ENTRIES, + minor->debugfs_root, minor); + if (ret) + return ret; + + /* Set the size of the vbios since we know it, and it's confusing to + * userspace if it wants to seek() but the file has a length of 0 + */ + dentry = debugfs_lookup("vbios.rom", minor->debugfs_root); + if (!dentry) + return 0; + + d_inode(dentry)->i_size = drm->vbios.length; + dput(dentry); + + return 0; } int diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c index 945afd34138e..078f65d849ce 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.c +++ b/drivers/gpu/drm/nouveau/nouveau_dma.c @@ -101,7 +101,7 @@ nv50_dma_push(struct nouveau_channel *chan, u64 offset, int length) nvif_wr32(&chan->user, 0x8c, chan->dma.ib_put); if (user->func && user->func->doorbell) - user->func->doorbell(user, chan->chid); + user->func->doorbell(user, chan->token); chan->dma.ib_free--; } diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 2b2baf6e0e0d..f900e94592f8 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -353,6 +353,7 @@ nouveau_accel_init(struct nouveau_drm *drm) case MAXWELL_CHANNEL_GPFIFO_A: case PASCAL_CHANNEL_GPFIFO_A: case VOLTA_CHANNEL_GPFIFO_A: + case TURING_CHANNEL_GPFIFO_A: ret = nvc0_fence_create(drm); break; default: @@ -370,7 +371,7 @@ nouveau_accel_init(struct nouveau_drm *drm) if (device->info.family >= NV_DEVICE_INFO_V0_KEPLER) { ret = nouveau_channel_new(drm, &drm->client.device, nvif_fifo_runlist_ce(device), 0, - &drm->cechan); + true, &drm->cechan); if (ret) NV_ERROR(drm, "failed to create ce channel, %d\n", ret); @@ -381,7 +382,8 @@ nouveau_accel_init(struct nouveau_drm *drm) device->info.chipset != 0xaa && device->info.chipset != 0xac) { ret = nouveau_channel_new(drm, &drm->client.device, - NvDmaFB, NvDmaTT, &drm->cechan); + NvDmaFB, NvDmaTT, false, + &drm->cechan); if (ret) NV_ERROR(drm, "failed to create ce channel, %d\n", ret); @@ -393,7 +395,7 @@ nouveau_accel_init(struct nouveau_drm *drm) } ret = nouveau_channel_new(drm, &drm->client.device, - arg0, arg1, &drm->channel); + arg0, arg1, false, &drm->channel); if (ret) { NV_ERROR(drm, "failed to create kernel channel, %d\n", ret); nouveau_accel_fini(drm); @@ -1171,10 +1173,16 @@ nouveau_platform_device_create(const struct nvkm_device_tegra_func *func, goto err_free; } + err = nouveau_drm_device_init(drm); + if (err) + goto err_put; + platform_set_drvdata(pdev, drm); return drm; +err_put: + drm_dev_put(drm); err_free: nvkm_device_del(pdevice); diff --git a/drivers/gpu/drm/nouveau/nouveau_vmm.h b/drivers/gpu/drm/nouveau/nouveau_vmm.h index 7e3b118cf7c4..ede872f6f668 100644 --- a/drivers/gpu/drm/nouveau/nouveau_vmm.h +++ b/drivers/gpu/drm/nouveau/nouveau_vmm.h @@ -25,7 +25,6 @@ void nouveau_vma_unmap(struct nouveau_vma *); struct nouveau_vmm { struct nouveau_cli *cli; struct nvif_vmm vmm; - struct nvkm_vm *vm; }; int nouveau_vmm_init(struct nouveau_cli *, s32 oclass, struct nouveau_vmm *); diff --git a/drivers/gpu/drm/nouveau/nvif/disp.c b/drivers/gpu/drm/nouveau/nvif/disp.c index 18c7d064f75c..ef97dd223a32 100644 --- a/drivers/gpu/drm/nouveau/nvif/disp.c +++ b/drivers/gpu/drm/nouveau/nvif/disp.c @@ -34,6 +34,7 @@ int nvif_disp_ctor(struct nvif_device *device, s32 oclass, struct nvif_disp *disp) { static const struct nvif_mclass disps[] = { + { TU104_DISP, -1 }, { GV100_DISP, -1 }, { GP102_DISP, -1 }, { GP100_DISP, -1 }, diff --git a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c index 03f676c18aad..c61b467cf45e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c +++ b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c @@ -79,7 +79,9 @@ nvkm_subdev_name[NVKM_SUBDEV_NR] = { [NVKM_ENGINE_NVENC0 ] = "nvenc0", [NVKM_ENGINE_NVENC1 ] = "nvenc1", [NVKM_ENGINE_NVENC2 ] = "nvenc2", - [NVKM_ENGINE_NVDEC ] = "nvdec", + [NVKM_ENGINE_NVDEC0 ] = "nvdec0", + [NVKM_ENGINE_NVDEC1 ] = "nvdec1", + [NVKM_ENGINE_NVDEC2 ] = "nvdec2", [NVKM_ENGINE_PM ] = "pm", [NVKM_ENGINE_SEC ] = "sec", [NVKM_ENGINE_SEC2 ] = "sec2", diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild index 80d784441904..177a23301d6a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild @@ -6,3 +6,4 @@ nvkm-y += nvkm/engine/ce/gm200.o nvkm-y += nvkm/engine/ce/gp100.o nvkm-y += nvkm/engine/ce/gp102.o nvkm-y += nvkm/engine/ce/gv100.o +nvkm-y += nvkm/engine/ce/tu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu104.c new file mode 100644 index 000000000000..3c25043bbb33 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/tu104.c @@ -0,0 +1,40 @@ +/* + * Copyright 2018 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. + */ +#include "priv.h" + +#include <nvif/class.h> + +static const struct nvkm_engine_func +tu104_ce = { + .intr = gp100_ce_intr, + .sclass = { + { -1, -1, TURING_DMA_COPY_A }, + {} + } +}; + +int +tu104_ce_new(struct nvkm_device *device, int index, + struct nvkm_engine **pengine) +{ + return nvkm_engine_new_(&tu104_ce, device, index, true, pengine); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c index e294013426ce..d9edb5785813 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c @@ -2221,7 +2221,7 @@ nv132_chipset = { .dma = gf119_dma_new, .fifo = gp100_fifo_new, .gr = gp102_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, .sw = gf100_sw_new, }; @@ -2257,7 +2257,7 @@ nv134_chipset = { .dma = gf119_dma_new, .fifo = gp100_fifo_new, .gr = gp104_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, .sw = gf100_sw_new, }; @@ -2293,7 +2293,7 @@ nv136_chipset = { .dma = gf119_dma_new, .fifo = gp100_fifo_new, .gr = gp104_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, .sw = gf100_sw_new, }; @@ -2329,7 +2329,7 @@ nv137_chipset = { .dma = gf119_dma_new, .fifo = gp100_fifo_new, .gr = gp107_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, .sw = gf100_sw_new, }; @@ -2365,7 +2365,7 @@ nv138_chipset = { .dma = gf119_dma_new, .fifo = gp100_fifo_new, .gr = gp107_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, .sw = gf100_sw_new, }; @@ -2430,10 +2430,106 @@ nv140_chipset = { .dma = gv100_dma_new, .fifo = gv100_fifo_new, .gr = gv100_gr_new, - .nvdec = gp102_nvdec_new, + .nvdec[0] = gp102_nvdec_new, .sec2 = gp102_sec2_new, }; +static const struct nvkm_device_chip +nv162_chipset = { + .name = "TU102", + .bar = tu104_bar_new, + .bios = nvkm_bios_new, + .bus = gf100_bus_new, + .devinit = tu104_devinit_new, + .fault = tu104_fault_new, + .fb = gv100_fb_new, + .fuse = gm107_fuse_new, + .gpio = gk104_gpio_new, + .i2c = gm200_i2c_new, + .ibus = gm200_ibus_new, + .imem = nv50_instmem_new, + .ltc = gp102_ltc_new, + .mc = tu104_mc_new, + .mmu = tu104_mmu_new, + .pci = gp100_pci_new, + .pmu = gp102_pmu_new, + .therm = gp100_therm_new, + .timer = gk20a_timer_new, + .top = gk104_top_new, + .ce[0] = tu104_ce_new, + .ce[1] = tu104_ce_new, + .ce[2] = tu104_ce_new, + .ce[3] = tu104_ce_new, + .ce[4] = tu104_ce_new, + .disp = tu104_disp_new, + .dma = gv100_dma_new, + .fifo = tu104_fifo_new, +}; + +static const struct nvkm_device_chip +nv164_chipset = { + .name = "TU104", + .bar = tu104_bar_new, + .bios = nvkm_bios_new, + .bus = gf100_bus_new, + .devinit = tu104_devinit_new, + .fault = tu104_fault_new, + .fb = gv100_fb_new, + .fuse = gm107_fuse_new, + .gpio = gk104_gpio_new, + .i2c = gm200_i2c_new, + .ibus = gm200_ibus_new, + .imem = nv50_instmem_new, + .ltc = gp102_ltc_new, + .mc = tu104_mc_new, + .mmu = tu104_mmu_new, + .pci = gp100_pci_new, + .pmu = gp102_pmu_new, + .therm = gp100_therm_new, + .timer = gk20a_timer_new, + .top = gk104_top_new, + .ce[0] = tu104_ce_new, + .ce[1] = tu104_ce_new, + .ce[2] = tu104_ce_new, + .ce[3] = tu104_ce_new, + .ce[4] = tu104_ce_new, + .disp = tu104_disp_new, + .dma = gv100_dma_new, + .fifo = tu104_fifo_new, +}; + +static const struct nvkm_device_chip +nv166_chipset = { + .name = "TU106", + .bar = tu104_bar_new, + .bios = nvkm_bios_new, + .bus = gf100_bus_new, + .devinit = tu104_devinit_new, + .fault = tu104_fault_new, + .fb = gv100_fb_new, + .fuse = gm107_fuse_new, + .gpio = gk104_gpio_new, + .i2c = gm200_i2c_new, + .ibus = gm200_ibus_new, + .imem = nv50_instmem_new, + .ltc = gp102_ltc_new, + .mc = tu104_mc_new, + .mmu = tu104_mmu_new, + .pci = gp100_pci_new, + .pmu = gp102_pmu_new, + .therm = gp100_therm_new, + .timer = gk20a_timer_new, + .top = gk104_top_new, + .ce[0] = tu104_ce_new, + .ce[1] = tu104_ce_new, + .ce[2] = tu104_ce_new, + .ce[3] = tu104_ce_new, + .ce[4] = tu104_ce_new, + .disp = tu104_disp_new, + .dma = gv100_dma_new, + .fifo = tu104_fifo_new, +}; + static int nvkm_device_event_ctor(struct nvkm_object *object, void *data, u32 size, struct nvkm_notify *notify) @@ -2529,7 +2625,9 @@ nvkm_device_engine(struct nvkm_device *device, int index) _(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->engine); + _(NVDEC0 , device->nvdec[0], &device->nvdec[0]->engine); + _(NVDEC1 , device->nvdec[1], &device->nvdec[1]->engine); + _(NVDEC2 , device->nvdec[2], &device->nvdec[2]->engine); _(PM , device->pm , &device->pm->engine); _(SEC , device->sec , device->sec); _(SEC2 , device->sec2 , &device->sec2->engine); @@ -2791,6 +2889,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x120: device->card_type = GM100; break; case 0x130: device->card_type = GP100; break; case 0x140: device->card_type = GV100; break; + case 0x160: device->card_type = TU100; break; default: break; } @@ -2883,6 +2982,9 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x138: device->chip = &nv138_chipset; break; case 0x13b: device->chip = &nv13b_chipset; break; case 0x140: device->chip = &nv140_chipset; break; + case 0x162: device->chip = &nv162_chipset; break; + case 0x164: device->chip = &nv164_chipset; break; + case 0x166: device->chip = &nv166_chipset; break; default: nvdev_error(device, "unknown chipset (%08x)\n", boot0); goto done; @@ -2988,7 +3090,9 @@ nvkm_device_ctor(const struct nvkm_device_func *func, _(NVKM_ENGINE_NVENC0 , nvenc[0]); _(NVKM_ENGINE_NVENC1 , nvenc[1]); _(NVKM_ENGINE_NVENC2 , nvenc[2]); - _(NVKM_ENGINE_NVDEC , nvdec); + _(NVKM_ENGINE_NVDEC0 , nvdec[0]); + _(NVKM_ENGINE_NVDEC1 , nvdec[1]); + _(NVKM_ENGINE_NVDEC2 , nvdec[2]); _(NVKM_ENGINE_PM , pm); _(NVKM_ENGINE_SEC , sec); _(NVKM_ENGINE_SEC2 , sec2); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c index dde6bbafa709..092ddc4ffefa 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c @@ -91,7 +91,7 @@ nvkm_udevice_info_v1(struct nvkm_device *device, case ENGINE_A(MSENC ); break; case ENGINE_A(VIC ); break; case ENGINE_A(SEC2 ); break; - case ENGINE_A(NVDEC ); break; + case ENGINE_B(NVDEC ); break; case ENGINE_B(NVENC ); break; default: args->mthd = NV_DEVICE_INFO_INVALID; @@ -175,6 +175,7 @@ nvkm_udevice_info(struct nvkm_udevice *udev, void *data, u32 size) case GM100: args->v0.family = NV_DEVICE_INFO_V0_MAXWELL; break; case GP100: args->v0.family = NV_DEVICE_INFO_V0_PASCAL; break; case GV100: args->v0.family = NV_DEVICE_INFO_V0_VOLTA; break; + case TU100: args->v0.family = NV_DEVICE_INFO_V0_TURING; 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 8089ac9a12e2..c6a257ba4347 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild @@ -15,6 +15,7 @@ nvkm-y += nvkm/engine/disp/gm200.o nvkm-y += nvkm/engine/disp/gp100.o nvkm-y += nvkm/engine/disp/gp102.o nvkm-y += nvkm/engine/disp/gv100.o +nvkm-y += nvkm/engine/disp/tu104.o nvkm-y += nvkm/engine/disp/vga.o nvkm-y += nvkm/engine/disp/head.o @@ -38,6 +39,7 @@ nvkm-y += nvkm/engine/disp/sorgk104.o nvkm-y += nvkm/engine/disp/sorgm107.o nvkm-y += nvkm/engine/disp/sorgm200.o nvkm-y += nvkm/engine/disp/sorgv100.o +nvkm-y += nvkm/engine/disp/sortu104.o nvkm-y += nvkm/engine/disp/outp.o nvkm-y += nvkm/engine/disp/dp.o @@ -69,6 +71,7 @@ nvkm-y += nvkm/engine/disp/rootgm200.o nvkm-y += nvkm/engine/disp/rootgp100.o nvkm-y += nvkm/engine/disp/rootgp102.o nvkm-y += nvkm/engine/disp/rootgv100.o +nvkm-y += nvkm/engine/disp/roottu104.o nvkm-y += nvkm/engine/disp/channv50.o nvkm-y += nvkm/engine/disp/changf119.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c index d0a7e3456da1..47be0ba4aebe 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gv100.c @@ -28,7 +28,7 @@ #include <core/gpuobj.h> #include <subdev/timer.h> -static int +int gv100_disp_wndw_cnt(struct nvkm_disp *disp, unsigned long *pmask) { struct nvkm_device *device = disp->engine.subdev.device; @@ -36,7 +36,7 @@ gv100_disp_wndw_cnt(struct nvkm_disp *disp, unsigned long *pmask) return (nvkm_rd32(device, 0x610074) & 0x03f00000) >> 20; } -static void +void gv100_disp_super(struct work_struct *work) { struct nv50_disp *disp = @@ -257,7 +257,7 @@ gv100_disp_intr_head_timing(struct nv50_disp *disp, int head) } } -static void +void gv100_disp_intr(struct nv50_disp *disp) { struct nvkm_subdev *subdev = &disp->base.engine.subdev; @@ -297,7 +297,7 @@ gv100_disp_intr(struct nv50_disp *disp) nvkm_warn(subdev, "intr %08x\n", stat); } -static void +void gv100_disp_fini(struct nv50_disp *disp) { struct nvkm_device *device = disp->base.engine.subdev.device; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h index 0f0c86c32ec3..790e42f460fd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ior.h @@ -144,6 +144,11 @@ void gm200_sor_route_set(struct nvkm_outp *, struct nvkm_ior *); int gm200_sor_route_get(struct nvkm_outp *, int *); void gm200_sor_dp_drive(struct nvkm_ior *, int, int, int, int, int); +void gv100_sor_state(struct nvkm_ior *, struct nvkm_ior_state *); +void gv100_sor_dp_audio(struct nvkm_ior *, int, bool); +void gv100_sor_dp_audio_sym(struct nvkm_ior *, int, u16, u32); +void gv100_sor_dp_watermark(struct nvkm_ior *, int, u8); + void g84_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); void gt215_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); void gf119_hdmi_ctrl(struct nvkm_ior *, int, bool, u8, u8, u8 *, u8 , u8 *, u8); @@ -195,4 +200,6 @@ int gm200_sor_new(struct nvkm_disp *, int); int gv100_sor_cnt(struct nvkm_disp *, unsigned long *); int gv100_sor_new(struct nvkm_disp *, int); + +int tu104_sor_new(struct nvkm_disp *, int); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h index 8580382ab248..c36a8a7cafa1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h @@ -78,6 +78,11 @@ void gf119_disp_intr(struct nv50_disp *); void gf119_disp_super(struct work_struct *); void gf119_disp_intr_error(struct nv50_disp *, int); +void gv100_disp_fini(struct nv50_disp *); +void gv100_disp_intr(struct nv50_disp *); +void gv100_disp_super(struct work_struct *); +int gv100_disp_wndw_cnt(struct nvkm_disp *, unsigned long *); + void nv50_disp_dptmds_war_2(struct nv50_disp *, struct dcb_output *); void nv50_disp_dptmds_war_3(struct nv50_disp *, struct dcb_output *); void nv50_disp_update_sppll1(struct nv50_disp *); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h index 6ca4f9184b51..97de928cbde1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h @@ -37,4 +37,5 @@ 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 gp102_disp_root_oclass; extern const struct nvkm_disp_oclass gv100_disp_root_oclass; +extern const struct nvkm_disp_oclass tu104_disp_root_oclass; #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/roottu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/roottu104.c new file mode 100644 index 000000000000..ad438c62f66c --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/roottu104.c @@ -0,0 +1,52 @@ +/* + * Copyright 2018 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. + */ +#include "rootnv50.h" +#include "channv50.h" + +#include <nvif/class.h> + +static const struct nv50_disp_root_func +tu104_disp_root = { + .user = { + {{0,0,TU104_DISP_CURSOR }, gv100_disp_curs_new }, + {{0,0,TU104_DISP_WINDOW_IMM_CHANNEL_DMA}, gv100_disp_wimm_new }, + {{0,0,TU104_DISP_CORE_CHANNEL_DMA }, gv100_disp_core_new }, + {{0,0,TU104_DISP_WINDOW_CHANNEL_DMA }, gv100_disp_wndw_new }, + {} + }, +}; + +static int +tu104_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_(&tu104_disp_root, disp, oclass, + data, size, pobject); +} + +const struct nvkm_disp_oclass +tu104_disp_root_oclass = { + .base.oclass = TU104_DISP, + .base.minver = -1, + .base.maxver = -1, + .ctor = tu104_disp_root_new, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c index 8ba881a729ee..b0597ff9a714 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgv100.c @@ -23,7 +23,7 @@ #include <subdev/timer.h> -static void +void gv100_sor_dp_watermark(struct nvkm_ior *sor, int head, u8 watermark) { struct nvkm_device *device = sor->disp->engine.subdev.device; @@ -31,7 +31,7 @@ gv100_sor_dp_watermark(struct nvkm_ior *sor, int head, u8 watermark) nvkm_mask(device, 0x616550 + hoff, 0x0c00003f, 0x08000000 | watermark); } -static void +void gv100_sor_dp_audio_sym(struct nvkm_ior *sor, int head, u16 h, u32 v) { struct nvkm_device *device = sor->disp->engine.subdev.device; @@ -40,7 +40,7 @@ gv100_sor_dp_audio_sym(struct nvkm_ior *sor, int head, u16 h, u32 v) nvkm_mask(device, 0x61656c + hoff, 0x00ffffff, v); } -static void +void gv100_sor_dp_audio(struct nvkm_ior *sor, int head, bool enable) { struct nvkm_device *device = sor->disp->engine.subdev.device; @@ -54,7 +54,7 @@ gv100_sor_dp_audio(struct nvkm_ior *sor, int head, bool enable) ); } -static void +void gv100_sor_state(struct nvkm_ior *sor, struct nvkm_ior_state *state) { struct nvkm_device *device = sor->disp->engine.subdev.device; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu104.c new file mode 100644 index 000000000000..df026a525ef1 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sortu104.c @@ -0,0 +1,97 @@ +/* + * Copyright 2018 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. + */ +#include "ior.h" + +#include <subdev/timer.h> + +static void +tu104_sor_dp_vcpi(struct nvkm_ior *sor, int head, + u8 slot, u8 slot_nr, u16 pbn, u16 aligned) +{ + struct nvkm_device *device = sor->disp->engine.subdev.device; + const u32 hoff = head * 0x800; + + nvkm_mask(device, 0x61657c + hoff, 0xffffffff, (aligned << 16) | pbn); + nvkm_mask(device, 0x616578 + hoff, 0x00003f3f, (slot_nr << 8) | slot); +} + +static int +tu104_sor_dp_links(struct nvkm_ior *sor, struct nvkm_i2c_aux *aux) +{ + struct nvkm_device *device = sor->disp->engine.subdev.device; + const u32 soff = nv50_ior_base(sor); + const u32 loff = nv50_sor_link(sor); + u32 dpctrl = 0x00000000; + u32 clksor = 0x00000000; + + clksor |= sor->dp.bw << 18; + dpctrl |= ((1 << sor->dp.nr) - 1) << 16; + if (sor->dp.mst) + dpctrl |= 0x40000000; + if (sor->dp.ef) + dpctrl |= 0x00004000; + + nvkm_mask(device, 0x612300 + soff, 0x007c0000, clksor); + + /*XXX*/ + nvkm_msec(device, 40, NVKM_DELAY); + nvkm_mask(device, 0x612300 + soff, 0x00030000, 0x00010000); + nvkm_mask(device, 0x61c10c + loff, 0x00000003, 0x00000001); + + nvkm_mask(device, 0x61c10c + loff, 0x401f4000, dpctrl); + return 0; +} + +static const struct nvkm_ior_func +tu104_sor = { + .route = { + .get = gm200_sor_route_get, + .set = gm200_sor_route_set, + }, + .state = gv100_sor_state, + .power = nv50_sor_power, + .clock = gf119_sor_clock, + .hdmi = { + .ctrl = gv100_hdmi_ctrl, + }, + .dp = { + .lanes = { 0, 1, 2, 3 }, + .links = tu104_sor_dp_links, + .power = g94_sor_dp_power, + .pattern = gm107_sor_dp_pattern, + .drive = gm200_sor_dp_drive, + .vcpi = tu104_sor_dp_vcpi, + .audio = gv100_sor_dp_audio, + .audio_sym = gv100_sor_dp_audio_sym, + .watermark = gv100_sor_dp_watermark, + }, + .hda = { + .hpd = gf119_hda_hpd, + .eld = gf119_hda_eld, + }, +}; + +int +tu104_sor_new(struct nvkm_disp *disp, int id) +{ + return nvkm_ior_new_(&tu104_sor, disp, SOR, id); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu104.c new file mode 100644 index 000000000000..13fa21459d38 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/tu104.c @@ -0,0 +1,152 @@ +/* + * Copyright 2018 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. + */ +#include "nv50.h" +#include "head.h" +#include "ior.h" +#include "channv50.h" +#include "rootnv50.h" + +#include <core/gpuobj.h> +#include <subdev/timer.h> + +static int +tu104_disp_init(struct nv50_disp *disp) +{ + struct nvkm_device *device = disp->base.engine.subdev.device; + struct nvkm_head *head; + int i, j; + u32 tmp; + + /* Claim ownership of display. */ + if (nvkm_rd32(device, 0x6254e8) & 0x00000002) { + nvkm_mask(device, 0x6254e8, 0x00000001, 0x00000000); + if (nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x6254e8) & 0x00000002)) + break; + ) < 0) + return -EBUSY; + } + + /* Lock pin capabilities. */ + tmp = 0x00000021; /*XXX*/ + nvkm_wr32(device, 0x640008, tmp); + + /* SOR capabilities. */ + for (i = 0; i < disp->sor.nr; i++) { + tmp = nvkm_rd32(device, 0x61c000 + (i * 0x800)); + nvkm_mask(device, 0x640000, 0x00000100 << i, 0x00000100 << i); + nvkm_wr32(device, 0x640144 + (i * 0x08), tmp); + } + + /* Head capabilities. */ + list_for_each_entry(head, &disp->base.head, head) { + const int id = head->id; + + /* RG. */ + tmp = nvkm_rd32(device, 0x616300 + (id * 0x800)); + nvkm_wr32(device, 0x640048 + (id * 0x020), tmp); + + /* POSTCOMP. */ + for (j = 0; j < 5 * 4; j += 4) { + tmp = nvkm_rd32(device, 0x616140 + (id * 0x800) + j); + nvkm_wr32(device, 0x640680 + (id * 0x20) + j, tmp); + } + } + + /* Window capabilities. */ + for (i = 0; i < disp->wndw.nr; i++) { + nvkm_mask(device, 0x640004, 1 << i, 1 << i); + for (j = 0; j < 6 * 4; j += 4) { + tmp = nvkm_rd32(device, 0x630100 + (i * 0x800) + j); + nvkm_mask(device, 0x640780 + (i * 0x20) + j, 0xffffffff, tmp); + } + nvkm_mask(device, 0x64000c, 0x00000100, 0x00000100); + } + + /* IHUB capabilities. */ + for (i = 0; i < 3; i++) { + tmp = nvkm_rd32(device, 0x62e000 + (i * 0x04)); + nvkm_wr32(device, 0x640010 + (i * 0x04), tmp); + } + + nvkm_mask(device, 0x610078, 0x00000001, 0x00000001); + + /* Setup instance memory. */ + switch (nvkm_memory_target(disp->inst->memory)) { + case NVKM_MEM_TARGET_VRAM: tmp = 0x00000001; break; + case NVKM_MEM_TARGET_NCOH: tmp = 0x00000002; break; + case NVKM_MEM_TARGET_HOST: tmp = 0x00000003; break; + default: + break; + } + nvkm_wr32(device, 0x610010, 0x00000008 | tmp); + nvkm_wr32(device, 0x610014, disp->inst->addr >> 16); + + /* CTRL_DISP: AWAKEN, ERROR, SUPERVISOR[1-3]. */ + nvkm_wr32(device, 0x611cf0, 0x00000187); /* MSK. */ + nvkm_wr32(device, 0x611db0, 0x00000187); /* EN. */ + + /* EXC_OTHER: CURSn, CORE. */ + nvkm_wr32(device, 0x611cec, disp->head.mask << 16 | + 0x00000001); /* MSK. */ + nvkm_wr32(device, 0x611dac, 0x00000000); /* EN. */ + + /* EXC_WINIM. */ + nvkm_wr32(device, 0x611ce8, disp->wndw.mask); /* MSK. */ + nvkm_wr32(device, 0x611da8, 0x00000000); /* EN. */ + + /* EXC_WIN. */ + nvkm_wr32(device, 0x611ce4, disp->wndw.mask); /* MSK. */ + nvkm_wr32(device, 0x611da4, 0x00000000); /* EN. */ + + /* HEAD_TIMING(n): VBLANK. */ + list_for_each_entry(head, &disp->base.head, head) { + const u32 hoff = head->id * 4; + nvkm_wr32(device, 0x611cc0 + hoff, 0x00000004); /* MSK. */ + nvkm_wr32(device, 0x611d80 + hoff, 0x00000000); /* EN. */ + } + + /* OR. */ + nvkm_wr32(device, 0x611cf4, 0x00000000); /* MSK. */ + nvkm_wr32(device, 0x611db4, 0x00000000); /* EN. */ + return 0; +} + +static const struct nv50_disp_func +tu104_disp = { + .init = tu104_disp_init, + .fini = gv100_disp_fini, + .intr = gv100_disp_intr, + .uevent = &gv100_disp_chan_uevent, + .super = gv100_disp_super, + .root = &tu104_disp_root_oclass, + .wndw = { .cnt = gv100_disp_wndw_cnt }, + .head = { .cnt = gv100_head_cnt, .new = gv100_head_new }, + .sor = { .cnt = gv100_sor_cnt, .new = tu104_sor_new }, + .ramht_size = 0x2000, +}; + +int +tu104_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp) +{ + return nv50_disp_new_(&tu104_disp, device, index, pdisp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/wndwgv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/wndwgv100.c index 98911805aabf..5d3b641dbb14 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/wndwgv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/wndwgv100.c @@ -118,7 +118,7 @@ gv100_disp_wndw_mthd_base = { const struct nv50_disp_chan_mthd gv100_disp_wndw_mthd = { - .name = "Base", + .name = "Window", .addr = 0x001000, .prev = 0x000800, .data = { diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c b/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c index 816ccaedfc73..8675613e142b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/falcon.c @@ -22,6 +22,7 @@ #include <engine/falcon.h> #include <core/gpuobj.h> +#include <subdev/mc.h> #include <subdev/timer.h> #include <engine/fifo.h> @@ -107,8 +108,10 @@ nvkm_falcon_fini(struct nvkm_engine *engine, bool suspend) } } - nvkm_mask(device, base + 0x048, 0x00000003, 0x00000000); - nvkm_wr32(device, base + 0x014, 0xffffffff); + if (nvkm_mc_enabled(device, engine->subdev.index)) { + nvkm_mask(device, base + 0x048, 0x00000003, 0x00000000); + nvkm_wr32(device, base + 0x014, 0xffffffff); + } return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild index f00408577a6a..87d8e054e40a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild @@ -16,6 +16,7 @@ nvkm-y += nvkm/engine/fifo/gm20b.o nvkm-y += nvkm/engine/fifo/gp100.o nvkm-y += nvkm/engine/fifo/gp10b.o nvkm-y += nvkm/engine/fifo/gv100.o +nvkm-y += nvkm/engine/fifo/tu104.o nvkm-y += nvkm/engine/fifo/chan.o nvkm-y += nvkm/engine/fifo/channv50.o @@ -33,5 +34,7 @@ nvkm-y += nvkm/engine/fifo/gpfifog84.o nvkm-y += nvkm/engine/fifo/gpfifogf100.o nvkm-y += nvkm/engine/fifo/gpfifogk104.o nvkm-y += nvkm/engine/fifo/gpfifogv100.o +nvkm-y += nvkm/engine/fifo/gpfifotu104.o nvkm-y += nvkm/engine/fifo/usergv100.o +nvkm-y += nvkm/engine/fifo/usertu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h index 3ffef236189e..2c7c5afc1ea5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/chan.h @@ -17,6 +17,7 @@ struct nvkm_fifo_chan_func { bool suspend); int (*object_ctor)(struct nvkm_fifo_chan *, struct nvkm_object *); void (*object_dtor)(struct nvkm_fifo_chan *, int); + u32 (*submit_token)(struct nvkm_fifo_chan *); }; int nvkm_fifo_chan_ctor(const struct nvkm_fifo_chan_func *, struct nvkm_fifo *, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h index 8e28ba6b2307..a14545d871d8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h @@ -14,6 +14,8 @@ struct gk104_fifo_chan { struct list_head head; bool killed; + struct nvkm_memory *mthd; + struct { struct nvkm_gpuobj *inst; struct nvkm_vma *vma; @@ -36,4 +38,15 @@ int gk104_fifo_gpfifo_kick_locked(struct gk104_fifo_chan *); int gv100_fifo_gpfifo_new(struct gk104_fifo *, const struct nvkm_oclass *, void *data, u32 size, struct nvkm_object **); +int gv100_fifo_gpfifo_new_(const struct nvkm_fifo_chan_func *, + struct gk104_fifo *, u64 *, u16 *, u64, u64, u64, + u64 *, bool, u32 *, const struct nvkm_oclass *, + struct nvkm_object **); +int gv100_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *, + struct nvkm_engine *); +int gv100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *, + struct nvkm_engine *, bool); + +int tu104_fifo_gpfifo_new(struct gk104_fifo *, const struct nvkm_oclass *, + void *data, u32 size, struct nvkm_object **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c index f69576868164..10a2e7039a75 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gf100.c @@ -346,10 +346,10 @@ gf100_fifo_intr_fault(struct gf100_fifo *fifo, int unit) if (eu && eu->data2) { switch (eu->data2) { case NVKM_SUBDEV_BAR: - nvkm_mask(device, 0x001704, 0x00000000, 0x00000000); + nvkm_bar_bar1_reset(device); break; case NVKM_SUBDEV_INSTMEM: - nvkm_mask(device, 0x001714, 0x00000000, 0x00000000); + nvkm_bar_bar2_reset(device); break; case NVKM_ENGINE_IFB: nvkm_mask(device, 0x001718, 0x00000000, 0x00000000); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c index afccf9721cf0..1053fe796466 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c @@ -149,16 +149,41 @@ gk104_fifo_uevent_init(struct nvkm_fifo *fifo) } void -gk104_fifo_runlist_commit(struct gk104_fifo *fifo, int runl) +gk104_fifo_runlist_commit(struct gk104_fifo *fifo, int runl, + struct nvkm_memory *mem, int nr) +{ + struct nvkm_subdev *subdev = &fifo->base.engine.subdev; + struct nvkm_device *device = subdev->device; + int target; + + switch (nvkm_memory_target(mem)) { + case NVKM_MEM_TARGET_VRAM: target = 0; break; + case NVKM_MEM_TARGET_NCOH: target = 3; break; + default: + WARN_ON(1); + return; + } + + nvkm_wr32(device, 0x002270, (nvkm_memory_addr(mem) >> 12) | + (target << 28)); + nvkm_wr32(device, 0x002274, (runl << 20) | nr); + + if (nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x002284 + (runl * 0x08)) & 0x00100000)) + break; + ) < 0) + nvkm_error(subdev, "runlist %d update timeout\n", runl); +} + +void +gk104_fifo_runlist_update(struct gk104_fifo *fifo, int runl) { const struct gk104_fifo_runlist_func *func = fifo->func->runlist; struct gk104_fifo_chan *chan; struct nvkm_subdev *subdev = &fifo->base.engine.subdev; - struct nvkm_device *device = subdev->device; struct nvkm_memory *mem; struct nvkm_fifo_cgrp *cgrp; int nr = 0; - int target; mutex_lock(&subdev->mutex); mem = fifo->runlist[runl].mem[fifo->runlist[runl].next]; @@ -177,24 +202,7 @@ gk104_fifo_runlist_commit(struct gk104_fifo *fifo, int runl) } nvkm_done(mem); - switch (nvkm_memory_target(mem)) { - case NVKM_MEM_TARGET_VRAM: target = 0; break; - case NVKM_MEM_TARGET_NCOH: target = 3; break; - default: - WARN_ON(1); - goto unlock; - } - - nvkm_wr32(device, 0x002270, (nvkm_memory_addr(mem) >> 12) | - (target << 28)); - nvkm_wr32(device, 0x002274, (runl << 20) | nr); - - if (nvkm_msec(device, 2000, - if (!(nvkm_rd32(device, 0x002284 + (runl * 0x08)) & 0x00100000)) - break; - ) < 0) - nvkm_error(subdev, "runlist %d update timeout\n", runl); -unlock: + func->commit(fifo, runl, mem, nr); mutex_unlock(&subdev->mutex); } @@ -238,6 +246,29 @@ const struct gk104_fifo_runlist_func gk104_fifo_runlist = { .size = 8, .chan = gk104_fifo_runlist_chan, + .commit = gk104_fifo_runlist_commit, +}; + +void +gk104_fifo_pbdma_init(struct gk104_fifo *fifo) +{ + struct nvkm_device *device = fifo->base.engine.subdev.device; + nvkm_wr32(device, 0x000204, (1 << fifo->pbdma_nr) - 1); +} + +int +gk104_fifo_pbdma_nr(struct gk104_fifo *fifo) +{ + struct nvkm_device *device = fifo->base.engine.subdev.device; + /* Determine number of PBDMAs by checking valid enable bits. */ + nvkm_wr32(device, 0x000204, 0xffffffff); + return hweight32(nvkm_rd32(device, 0x000204)); +} + +const struct gk104_fifo_pbdma_func +gk104_fifo_pbdma = { + .nr = gk104_fifo_pbdma_nr, + .init = gk104_fifo_pbdma_init, }; static void @@ -267,7 +298,7 @@ gk104_fifo_recover_work(struct work_struct *w) } for (todo = runm; runl = __ffs(todo), todo; todo &= ~BIT(runl)) - gk104_fifo_runlist_commit(fifo, runl); + gk104_fifo_runlist_update(fifo, runl); nvkm_wr32(device, 0x00262c, runm); nvkm_mask(device, 0x002630, runm, 0x00000000); @@ -456,10 +487,10 @@ gk104_fifo_fault(struct nvkm_fifo *base, struct nvkm_fault_data *info) if (ee && ee->data2) { switch (ee->data2) { case NVKM_SUBDEV_BAR: - nvkm_mask(device, 0x001704, 0x00000000, 0x00000000); + nvkm_bar_bar1_reset(device); break; case NVKM_SUBDEV_INSTMEM: - nvkm_mask(device, 0x001714, 0x00000000, 0x00000000); + nvkm_bar_bar2_reset(device); break; case NVKM_ENGINE_IFB: nvkm_mask(device, 0x001718, 0x00000000, 0x00000000); @@ -904,9 +935,7 @@ gk104_fifo_oneinit(struct nvkm_fifo *base) enum nvkm_devidx engidx; u32 *map; - /* Determine number of PBDMAs by checking valid enable bits. */ - nvkm_wr32(device, 0x000204, 0xffffffff); - fifo->pbdma_nr = hweight32(nvkm_rd32(device, 0x000204)); + fifo->pbdma_nr = fifo->func->pbdma->nr(fifo); nvkm_debug(subdev, "%d PBDMA(s)\n", fifo->pbdma_nr); /* Read PBDMA->runlist(s) mapping from HW. */ @@ -978,7 +1007,7 @@ gk104_fifo_init(struct nvkm_fifo *base) int i; /* Enable PBDMAs. */ - nvkm_wr32(device, 0x000204, (1 << fifo->pbdma_nr) - 1); + fifo->func->pbdma->init(fifo); /* PBDMA[n] */ for (i = 0; i < fifo->pbdma_nr; i++) { @@ -995,8 +1024,8 @@ gk104_fifo_init(struct nvkm_fifo *base) nvkm_wr32(device, 0x002254, 0x10000000 | fifo->user.bar->addr >> 12); - if (fifo->func->init_pbdma_timeout) - fifo->func->init_pbdma_timeout(fifo); + if (fifo->func->pbdma->init_timeout) + fifo->func->pbdma->init_timeout(fifo); nvkm_wr32(device, 0x002100, 0xffffffff); nvkm_wr32(device, 0x002140, 0x7fffffff); @@ -1175,6 +1204,7 @@ gk104_fifo_fault_gpcclient[] = { static const struct gk104_fifo_func gk104_fifo = { + .pbdma = &gk104_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gk104_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h index d295b81e18d6..d4e565658f46 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.h @@ -45,7 +45,11 @@ struct gk104_fifo { }; struct gk104_fifo_func { - void (*init_pbdma_timeout)(struct gk104_fifo *); + const struct gk104_fifo_pbdma_func { + int (*nr)(struct gk104_fifo *); + void (*init)(struct gk104_fifo *); + void (*init_timeout)(struct gk104_fifo *); + } *pbdma; struct { const struct nvkm_enum *access; @@ -61,6 +65,8 @@ struct gk104_fifo_func { struct nvkm_memory *, u32 offset); void (*chan)(struct gk104_fifo_chan *, struct nvkm_memory *, u32 offset); + void (*commit)(struct gk104_fifo *, int runl, + struct nvkm_memory *, int entries); } *runlist; struct gk104_fifo_user_user { @@ -81,8 +87,11 @@ int gk104_fifo_new_(const struct gk104_fifo_func *, struct nvkm_device *, int index, int nr, struct nvkm_fifo **); void gk104_fifo_runlist_insert(struct gk104_fifo *, struct gk104_fifo_chan *); void gk104_fifo_runlist_remove(struct gk104_fifo *, struct gk104_fifo_chan *); -void gk104_fifo_runlist_commit(struct gk104_fifo *, int runl); +void gk104_fifo_runlist_update(struct gk104_fifo *, int runl); +extern const struct gk104_fifo_pbdma_func gk104_fifo_pbdma; +int gk104_fifo_pbdma_nr(struct gk104_fifo *); +void gk104_fifo_pbdma_init(struct gk104_fifo *); extern const struct nvkm_enum gk104_fifo_fault_access[]; extern const struct nvkm_enum gk104_fifo_fault_engine[]; extern const struct nvkm_enum gk104_fifo_fault_reason[]; @@ -91,15 +100,30 @@ extern const struct nvkm_enum gk104_fifo_fault_gpcclient[]; extern const struct gk104_fifo_runlist_func gk104_fifo_runlist; void gk104_fifo_runlist_chan(struct gk104_fifo_chan *, struct nvkm_memory *, u32); +void gk104_fifo_runlist_commit(struct gk104_fifo *, int runl, + struct nvkm_memory *, int); extern const struct gk104_fifo_runlist_func gk110_fifo_runlist; void gk110_fifo_runlist_cgrp(struct nvkm_fifo_cgrp *, struct nvkm_memory *, u32); -void gk208_fifo_init_pbdma_timeout(struct gk104_fifo *); +extern const struct gk104_fifo_pbdma_func gk208_fifo_pbdma; +void gk208_fifo_pbdma_init_timeout(struct gk104_fifo *); extern const struct nvkm_enum gm107_fifo_fault_engine[]; extern const struct gk104_fifo_runlist_func gm107_fifo_runlist; +extern const struct gk104_fifo_pbdma_func gm200_fifo_pbdma; +int gm200_fifo_pbdma_nr(struct gk104_fifo *); + extern const struct nvkm_enum gp100_fifo_fault_engine[]; + +extern const struct nvkm_enum gv100_fifo_fault_access[]; +extern const struct nvkm_enum gv100_fifo_fault_reason[]; +extern const struct nvkm_enum gv100_fifo_fault_hubclient[]; +extern const struct nvkm_enum gv100_fifo_fault_gpcclient[]; +void gv100_fifo_runlist_cgrp(struct nvkm_fifo_cgrp *, + struct nvkm_memory *, u32); +void gv100_fifo_runlist_chan(struct gk104_fifo_chan *, + struct nvkm_memory *, u32); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c index ac7655a130fb..8adfa6b182cb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk110.c @@ -43,10 +43,12 @@ gk110_fifo_runlist = { .size = 8, .cgrp = gk110_fifo_runlist_cgrp, .chan = gk104_fifo_runlist_chan, + .commit = gk104_fifo_runlist_commit, }; static const struct gk104_fifo_func gk110_fifo = { + .pbdma = &gk104_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gk104_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c index 5ea7e452cc66..9553fb4af601 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk208.c @@ -27,7 +27,7 @@ #include <nvif/class.h> void -gk208_fifo_init_pbdma_timeout(struct gk104_fifo *fifo) +gk208_fifo_pbdma_init_timeout(struct gk104_fifo *fifo) { struct nvkm_device *device = fifo->base.engine.subdev.device; int i; @@ -36,9 +36,16 @@ gk208_fifo_init_pbdma_timeout(struct gk104_fifo *fifo) nvkm_wr32(device, 0x04012c + (i * 0x2000), 0x0000ffff); } +const struct gk104_fifo_pbdma_func +gk208_fifo_pbdma = { + .nr = gk104_fifo_pbdma_nr, + .init = gk104_fifo_pbdma_init, + .init_timeout = gk208_fifo_pbdma_init_timeout, +}; + static const struct gk104_fifo_func gk208_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gk208_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gk104_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c index 535a0eb67a5f..a4c6ac3cd6c7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk20a.c @@ -26,7 +26,7 @@ static const struct gk104_fifo_func gk20a_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gk208_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gk104_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c index 79ae19b1db67..acf230764cb0 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm107.c @@ -41,6 +41,7 @@ gm107_fifo_runlist = { .size = 8, .cgrp = gk110_fifo_runlist_cgrp, .chan = gm107_fifo_runlist_chan, + .commit = gk104_fifo_runlist_commit, }; const struct nvkm_enum @@ -68,7 +69,7 @@ gm107_fifo_fault_engine[] = { static const struct gk104_fifo_func gm107_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gk208_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gm107_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c index 49565faa854d..b96c1c5d6577 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm200.c @@ -26,9 +26,23 @@ #include <nvif/class.h> +int +gm200_fifo_pbdma_nr(struct gk104_fifo *fifo) +{ + struct nvkm_device *device = fifo->base.engine.subdev.device; + return nvkm_rd32(device, 0x002004) & 0x000000ff; +} + +const struct gk104_fifo_pbdma_func +gm200_fifo_pbdma = { + .nr = gm200_fifo_pbdma_nr, + .init = gk104_fifo_pbdma_init, + .init_timeout = gk208_fifo_pbdma_init_timeout, +}; + static const struct gk104_fifo_func gm200_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gm200_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gm107_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c index 46736513bd11..a49539b9e4ec 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gm20b.c @@ -26,7 +26,7 @@ static const struct gk104_fifo_func gm20b_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gm200_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gm107_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c index e2f8f9087d7c..54377e0f6a88 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c @@ -52,7 +52,7 @@ gp100_fifo_fault_engine[] = { static const struct gk104_fifo_func gp100_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gm200_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gp100_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c index 7733bf7c6545..778ba7e46fb3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp10b.c @@ -26,7 +26,7 @@ static const struct gk104_fifo_func gp10b_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gm200_fifo_pbdma, .fault.access = gk104_fifo_fault_access, .fault.engine = gp100_fifo_fault_engine, .fault.reason = gk104_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c index 118b37aea318..728a1edbf98c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogk104.c @@ -85,7 +85,7 @@ gk104_fifo_gpfifo_engine_addr(struct nvkm_engine *engine) case NVKM_ENGINE_MSVLD : return 0x0270; case NVKM_ENGINE_VIC : return 0x0280; case NVKM_ENGINE_MSENC : return 0x0290; - case NVKM_ENGINE_NVDEC : return 0x02100270; + case NVKM_ENGINE_NVDEC0: return 0x02100270; case NVKM_ENGINE_NVENC0: return 0x02100290; case NVKM_ENGINE_NVENC1: return 0x0210; default: @@ -192,7 +192,7 @@ gk104_fifo_gpfifo_fini(struct nvkm_fifo_chan *base) gk104_fifo_runlist_remove(fifo, chan); nvkm_mask(device, 0x800004 + coff, 0x00000800, 0x00000800); gk104_fifo_gpfifo_kick(chan); - gk104_fifo_runlist_commit(fifo, chan->runl); + gk104_fifo_runlist_update(fifo, chan->runl); } nvkm_wr32(device, 0x800000 + coff, 0x00000000); @@ -213,7 +213,7 @@ gk104_fifo_gpfifo_init(struct nvkm_fifo_chan *base) if (list_empty(&chan->head) && !chan->killed) { gk104_fifo_runlist_insert(fifo, chan); nvkm_mask(device, 0x800004 + coff, 0x00000400, 0x00000400); - gk104_fifo_runlist_commit(fifo, chan->runl); + gk104_fifo_runlist_update(fifo, chan->runl); nvkm_mask(device, 0x800004 + coff, 0x00000400, 0x00000400); } } @@ -222,6 +222,7 @@ void * gk104_fifo_gpfifo_dtor(struct nvkm_fifo_chan *base) { struct gk104_fifo_chan *chan = gk104_fifo_chan(base); + nvkm_memory_unref(&chan->mthd); kfree(chan->cgrp); return chan; } @@ -240,7 +241,7 @@ gk104_fifo_gpfifo_func = { static int gk104_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, - u64 vmm, u64 ioffset, u64 ilength, + u64 vmm, u64 ioffset, u64 ilength, u64 *inst, bool priv, const struct nvkm_oclass *oclass, struct nvkm_object **pobject) { @@ -279,6 +280,7 @@ gk104_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, return ret; *chid = chan->base.chid; + *inst = chan->base.inst->addr; /* Hack to support GPUs where even individual channels should be * part of a channel group. @@ -315,6 +317,7 @@ gk104_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, nvkm_wo32(chan->base.inst, 0x94, 0x30000001); nvkm_wo32(chan->base.inst, 0x9c, 0x00000100); nvkm_wo32(chan->base.inst, 0xac, 0x0000001f); + nvkm_wo32(chan->base.inst, 0xe4, priv ? 0x00000020 : 0x00000000); nvkm_wo32(chan->base.inst, 0xe8, chan->base.chid); nvkm_wo32(chan->base.inst, 0xb8, 0xf8000000); nvkm_wo32(chan->base.inst, 0xf8, 0x10003080); /* 0x002310 */ @@ -337,15 +340,19 @@ gk104_fifo_gpfifo_new(struct gk104_fifo *fifo, const struct nvkm_oclass *oclass, if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, false))) { nvif_ioctl(parent, "create channel gpfifo vers %d vmm %llx " "ioffset %016llx ilength %08x " - "runlist %016llx\n", + "runlist %016llx priv %d\n", args->v0.version, args->v0.vmm, args->v0.ioffset, - args->v0.ilength, args->v0.runlist); + args->v0.ilength, args->v0.runlist, args->v0.priv); + if (args->v0.priv && !oclass->client->super) + return -EINVAL; return gk104_fifo_gpfifo_new_(fifo, &args->v0.runlist, &args->v0.chid, args->v0.vmm, args->v0.ioffset, args->v0.ilength, + &args->v0.inst, + args->v0.priv, oclass, pobject); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c index 9598853ced56..a7462cf59d65 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogv100.c @@ -25,9 +25,15 @@ #include <core/client.h> #include <core/gpuobj.h> -#include <nvif/cla06f.h> +#include <nvif/clc36f.h> #include <nvif/unpack.h> +static u32 +gv100_fifo_gpfifo_submit_token(struct nvkm_fifo_chan *chan) +{ + return chan->chid; +} + static int gv100_fifo_gpfifo_engine_valid(struct gk104_fifo_chan *chan, bool ce, bool valid) { @@ -56,7 +62,7 @@ gv100_fifo_gpfifo_engine_valid(struct gk104_fifo_chan *chan, bool ce, bool valid return ret; } -static int +int gv100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base, struct nvkm_engine *engine, bool suspend) { @@ -79,7 +85,7 @@ gv100_fifo_gpfifo_engine_fini(struct nvkm_fifo_chan *base, return ret; } -static int +int gv100_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *base, struct nvkm_engine *engine) { @@ -100,8 +106,8 @@ gv100_fifo_gpfifo_engine_init(struct nvkm_fifo_chan *base, return gv100_fifo_gpfifo_engine_valid(chan, false, true); } -const struct nvkm_fifo_chan_func -gv100_fifo_gpfifo_func = { +static const struct nvkm_fifo_chan_func +gv100_fifo_gpfifo = { .dtor = gk104_fifo_gpfifo_dtor, .init = gk104_fifo_gpfifo_init, .fini = gk104_fifo_gpfifo_fini, @@ -110,19 +116,23 @@ gv100_fifo_gpfifo_func = { .engine_dtor = gk104_fifo_gpfifo_engine_dtor, .engine_init = gv100_fifo_gpfifo_engine_init, .engine_fini = gv100_fifo_gpfifo_engine_fini, + .submit_token = gv100_fifo_gpfifo_submit_token, }; -static int -gv100_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, - u64 vmm, u64 ioffset, u64 ilength, - const struct nvkm_oclass *oclass, +int +gv100_fifo_gpfifo_new_(const struct nvkm_fifo_chan_func *func, + struct gk104_fifo *fifo, u64 *runlists, u16 *chid, + u64 vmm, u64 ioffset, u64 ilength, u64 *inst, bool priv, + u32 *token, const struct nvkm_oclass *oclass, struct nvkm_object **pobject) { + struct nvkm_device *device = fifo->base.engine.subdev.device; struct gk104_fifo_chan *chan; int runlist = ffs(*runlists) -1, ret, i; unsigned long engm; u64 subdevs = 0; - u64 usermem; + u64 usermem, mthd; + u32 size; if (!vmm || runlist < 0 || runlist >= fifo->runlist_nr) return -EINVAL; @@ -142,14 +152,15 @@ gv100_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, chan->runl = runlist; INIT_LIST_HEAD(&chan->head); - ret = nvkm_fifo_chan_ctor(&gv100_fifo_gpfifo_func, &fifo->base, - 0x1000, 0x1000, true, vmm, 0, subdevs, - 1, fifo->user.bar->addr, 0x200, + ret = nvkm_fifo_chan_ctor(func, &fifo->base, 0x1000, 0x1000, true, vmm, + 0, subdevs, 1, fifo->user.bar->addr, 0x200, oclass, &chan->base); if (ret) return ret; *chid = chan->base.chid; + *inst = chan->base.inst->addr; + *token = chan->base.func->submit_token(&chan->base); /* Hack to support GPUs where even individual channels should be * part of a channel group. @@ -173,6 +184,20 @@ gv100_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, nvkm_done(fifo->user.mem); usermem = nvkm_memory_addr(fifo->user.mem) + usermem; + /* Allocate fault method buffer (magics come from nvgpu). */ + size = nvkm_rd32(device, 0x104028); /* NV_PCE_PCE_MAP */ + size = 27 * 5 * (((9 + 1 + 3) * hweight32(size)) + 2); + size = roundup(size, PAGE_SIZE); + + ret = nvkm_memory_new(device, NVKM_MEM_TARGET_INST, size, 0x1000, true, + &chan->mthd); + if (ret) + return ret; + + mthd = nvkm_memory_bar2(chan->mthd); + if (mthd == ~0ULL) + return -EFAULT; + /* RAMFC */ nvkm_kmap(chan->base.inst); nvkm_wo32(chan->base.inst, 0x008, lower_32_bits(usermem)); @@ -184,13 +209,13 @@ gv100_fifo_gpfifo_new_(struct gk104_fifo *fifo, u64 *runlists, u16 *chid, (ilength << 16)); nvkm_wo32(chan->base.inst, 0x084, 0x20400000); nvkm_wo32(chan->base.inst, 0x094, 0x30000001); - nvkm_wo32(chan->base.inst, 0x0e4, 0x00000020); + nvkm_wo32(chan->base.inst, 0x0e4, priv ? 0x00000020 : 0x00000000); nvkm_wo32(chan->base.inst, 0x0e8, chan->base.chid); - nvkm_wo32(chan->base.inst, 0x0f4, 0x00001100); + nvkm_wo32(chan->base.inst, 0x0f4, 0x00001000); nvkm_wo32(chan->base.inst, 0x0f8, 0x10003080); nvkm_mo32(chan->base.inst, 0x218, 0x00000000, 0x00000000); - nvkm_wo32(chan->base.inst, 0x220, 0x020a1000); - nvkm_wo32(chan->base.inst, 0x224, 0x00000000); + nvkm_wo32(chan->base.inst, 0x220, lower_32_bits(mthd)); + nvkm_wo32(chan->base.inst, 0x224, upper_32_bits(mthd)); nvkm_done(chan->base.inst); return gv100_fifo_gpfifo_engine_valid(chan, true, true); } @@ -201,7 +226,7 @@ gv100_fifo_gpfifo_new(struct gk104_fifo *fifo, const struct nvkm_oclass *oclass, { struct nvkm_object *parent = oclass->parent; union { - struct kepler_channel_gpfifo_a_v0 v0; + struct volta_channel_gpfifo_a_v0 v0; } *args = data; int ret = -ENOSYS; @@ -209,15 +234,20 @@ gv100_fifo_gpfifo_new(struct gk104_fifo *fifo, const struct nvkm_oclass *oclass, if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, false))) { nvif_ioctl(parent, "create channel gpfifo vers %d vmm %llx " "ioffset %016llx ilength %08x " - "runlist %016llx\n", + "runlist %016llx priv %d\n", args->v0.version, args->v0.vmm, args->v0.ioffset, - args->v0.ilength, args->v0.runlist); - return gv100_fifo_gpfifo_new_(fifo, + args->v0.ilength, args->v0.runlist, args->v0.priv); + if (args->v0.priv && !oclass->client->super) + return -EINVAL; + return gv100_fifo_gpfifo_new_(&gv100_fifo_gpfifo, fifo, &args->v0.runlist, &args->v0.chid, args->v0.vmm, args->v0.ioffset, args->v0.ilength, + &args->v0.inst, + args->v0.priv, + &args->v0.token, oclass, pobject); } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifotu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifotu104.c new file mode 100644 index 000000000000..ff70484dd01a --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifotu104.c @@ -0,0 +1,83 @@ +/* + * Copyright 2018 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. + */ +#include "changk104.h" +#include "cgrp.h" + +#include <core/client.h> +#include <core/gpuobj.h> + +#include <nvif/clc36f.h> +#include <nvif/unpack.h> + +static u32 +tu104_fifo_gpfifo_submit_token(struct nvkm_fifo_chan *base) +{ + struct gk104_fifo_chan *chan = gk104_fifo_chan(base); + return (chan->runl << 16) | chan->base.chid; +} + +static const struct nvkm_fifo_chan_func +tu104_fifo_gpfifo = { + .dtor = gk104_fifo_gpfifo_dtor, + .init = gk104_fifo_gpfifo_init, + .fini = gk104_fifo_gpfifo_fini, + .ntfy = gf100_fifo_chan_ntfy, + .engine_ctor = gk104_fifo_gpfifo_engine_ctor, + .engine_dtor = gk104_fifo_gpfifo_engine_dtor, + .engine_init = gv100_fifo_gpfifo_engine_init, + .engine_fini = gv100_fifo_gpfifo_engine_fini, + .submit_token = tu104_fifo_gpfifo_submit_token, +}; + +int +tu104_fifo_gpfifo_new(struct gk104_fifo *fifo, const struct nvkm_oclass *oclass, + void *data, u32 size, struct nvkm_object **pobject) +{ + struct nvkm_object *parent = oclass->parent; + union { + struct volta_channel_gpfifo_a_v0 v0; + } *args = data; + int ret = -ENOSYS; + + nvif_ioctl(parent, "create channel gpfifo size %d\n", size); + if (!(ret = nvif_unpack(ret, &data, &size, args->v0, 0, 0, false))) { + nvif_ioctl(parent, "create channel gpfifo vers %d vmm %llx " + "ioffset %016llx ilength %08x " + "runlist %016llx priv %d\n", + args->v0.version, args->v0.vmm, args->v0.ioffset, + args->v0.ilength, args->v0.runlist, args->v0.priv); + if (args->v0.priv && !oclass->client->super) + return -EINVAL; + return gv100_fifo_gpfifo_new_(&tu104_fifo_gpfifo, fifo, + &args->v0.runlist, + &args->v0.chid, + args->v0.vmm, + args->v0.ioffset, + args->v0.ilength, + &args->v0.inst, + args->v0.priv, + &args->v0.token, + oclass, pobject); + } + + return ret; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c index 4e1d159c0ae7..6ee1bb32a071 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gv100.c @@ -28,7 +28,7 @@ #include <nvif/class.h> -static void +void gv100_fifo_runlist_chan(struct gk104_fifo_chan *chan, struct nvkm_memory *memory, u32 offset) { @@ -42,7 +42,7 @@ gv100_fifo_runlist_chan(struct gk104_fifo_chan *chan, nvkm_wo32(memory, offset + 0xc, upper_32_bits(inst)); } -static void +void gv100_fifo_runlist_cgrp(struct nvkm_fifo_cgrp *cgrp, struct nvkm_memory *memory, u32 offset) { @@ -57,9 +57,10 @@ gv100_fifo_runlist = { .size = 16, .cgrp = gv100_fifo_runlist_cgrp, .chan = gv100_fifo_runlist_chan, + .commit = gk104_fifo_runlist_commit, }; -static const struct nvkm_enum +const struct nvkm_enum gv100_fifo_fault_gpcclient[] = { { 0x00, "T1_0" }, { 0x01, "T1_1" }, @@ -161,7 +162,7 @@ gv100_fifo_fault_gpcclient[] = { {} }; -static const struct nvkm_enum +const struct nvkm_enum gv100_fifo_fault_hubclient[] = { { 0x00, "VIP" }, { 0x01, "CE0" }, @@ -223,7 +224,7 @@ gv100_fifo_fault_hubclient[] = { {} }; -static const struct nvkm_enum +const struct nvkm_enum gv100_fifo_fault_reason[] = { { 0x00, "PDE" }, { 0x01, "PDE_SIZE" }, @@ -271,7 +272,7 @@ gv100_fifo_fault_engine[] = { {} }; -static const struct nvkm_enum +const struct nvkm_enum gv100_fifo_fault_access[] = { { 0x0, "VIRT_READ" }, { 0x1, "VIRT_WRITE" }, @@ -287,7 +288,7 @@ gv100_fifo_fault_access[] = { static const struct gk104_fifo_func gv100_fifo = { - .init_pbdma_timeout = gk208_fifo_init_pbdma_timeout, + .pbdma = &gm200_fifo_pbdma, .fault.access = gv100_fifo_fault_access, .fault.engine = gv100_fifo_fault_engine, .fault.reason = gv100_fifo_fault_reason, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu104.c new file mode 100644 index 000000000000..98c80705bc61 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/tu104.c @@ -0,0 +1,116 @@ +/* + * Copyright 2018 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. + */ +#include "gk104.h" +#include "cgrp.h" +#include "changk104.h" +#include "user.h" + +#include <core/gpuobj.h> + +#include <nvif/class.h> + +static void +tu104_fifo_runlist_commit(struct gk104_fifo *fifo, int runl, + struct nvkm_memory *mem, int nr) +{ + struct nvkm_device *device = fifo->base.engine.subdev.device; + u64 addr = nvkm_memory_addr(mem); + /*XXX: target? */ + + nvkm_wr32(device, 0x002b00 + (runl * 0x10), lower_32_bits(addr)); + nvkm_wr32(device, 0x002b04 + (runl * 0x10), upper_32_bits(addr)); + nvkm_wr32(device, 0x002b08 + (runl * 0x10), nr); + + /*XXX: how to wait? can you even wait? */ +} + +const struct gk104_fifo_runlist_func +tu104_fifo_runlist = { + .size = 16, + .cgrp = gv100_fifo_runlist_cgrp, + .chan = gv100_fifo_runlist_chan, + .commit = tu104_fifo_runlist_commit, +}; + +static const struct nvkm_enum +tu104_fifo_fault_engine[] = { + { 0x01, "DISPLAY" }, + { 0x03, "PTP" }, + { 0x06, "PWR_PMU" }, + { 0x08, "IFB", NULL, NVKM_ENGINE_IFB }, + { 0x09, "PERF" }, + { 0x1f, "PHYSICAL" }, + { 0x20, "HOST0" }, + { 0x21, "HOST1" }, + { 0x22, "HOST2" }, + { 0x23, "HOST3" }, + { 0x24, "HOST4" }, + { 0x25, "HOST5" }, + { 0x26, "HOST6" }, + { 0x27, "HOST7" }, + { 0x28, "HOST8" }, + { 0x29, "HOST9" }, + { 0x2a, "HOST10" }, + { 0x2b, "HOST11" }, + { 0x2c, "HOST12" }, + { 0x2d, "HOST13" }, + { 0x2e, "HOST14" }, + { 0x80, "BAR1", NULL, NVKM_SUBDEV_BAR }, + { 0xc0, "BAR2", NULL, NVKM_SUBDEV_INSTMEM }, + {} +}; + +static void +tu104_fifo_pbdma_init(struct gk104_fifo *fifo) +{ + struct nvkm_device *device = fifo->base.engine.subdev.device; + const u32 mask = (1 << fifo->pbdma_nr) - 1; + /*XXX: this is a bit of a guess at this point in time. */ + nvkm_mask(device, 0xb65000, 0x80000fff, 0x80000000 | mask); +} + +static const struct gk104_fifo_pbdma_func +tu104_fifo_pbdma = { + .nr = gm200_fifo_pbdma_nr, + .init = tu104_fifo_pbdma_init, + .init_timeout = gk208_fifo_pbdma_init_timeout, +}; + +static const struct gk104_fifo_func +tu104_fifo = { + .pbdma = &tu104_fifo_pbdma, + .fault.access = gv100_fifo_fault_access, + .fault.engine = tu104_fifo_fault_engine, + .fault.reason = gv100_fifo_fault_reason, + .fault.hubclient = gv100_fifo_fault_hubclient, + .fault.gpcclient = gv100_fifo_fault_gpcclient, + .runlist = &tu104_fifo_runlist, + .user = {{-1,-1,VOLTA_USERMODE_A }, tu104_fifo_user_new }, + .chan = {{ 0, 0,TURING_CHANNEL_GPFIFO_A}, tu104_fifo_gpfifo_new }, + .cgrp_force = true, +}; + +int +tu104_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo) +{ + return gk104_fifo_new_(&tu104_fifo, device, index, 4096, pfifo); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/user.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/user.h index ed840921ebe8..14b0c6bde8eb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/user.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/user.h @@ -3,4 +3,6 @@ #include "priv.h" int gv100_fifo_user_new(const struct nvkm_oclass *, void *, u32, struct nvkm_object **); +int tu104_fifo_user_new(const struct nvkm_oclass *, void *, u32, + struct nvkm_object **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/usertu104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/usertu104.c new file mode 100644 index 000000000000..8f98548a21f6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/usertu104.c @@ -0,0 +1,45 @@ +/* + * Copyright 2018 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. + */ +#include "user.h" + +static int +tu104_fifo_user_map(struct nvkm_object *object, void *argv, u32 argc, + enum nvkm_object_map *type, u64 *addr, u64 *size) +{ + struct nvkm_device *device = object->engine->subdev.device; + *addr = 0xbb0000 + device->func->resource_addr(device, 0); + *size = 0x010000; + *type = NVKM_OBJECT_MAP_IO; + return 0; +} + +static const struct nvkm_object_func +tu104_fifo_user = { + .map = tu104_fifo_user_map, +}; + +int +tu104_fifo_user_new(const struct nvkm_oclass *oclass, void *argv, u32 argc, + struct nvkm_object **pobject) +{ + return nvkm_object_new_(&tu104_fifo_user, oclass, argv, argc, pobject); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/falcon/base.c b/drivers/gpu/drm/nouveau/nvkm/falcon/base.c index 14be41f24155..427340153640 100644 --- a/drivers/gpu/drm/nouveau/nvkm/falcon/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/falcon/base.c @@ -197,7 +197,7 @@ nvkm_falcon_ctor(const struct nvkm_falcon_func *func, case NVKM_SUBDEV_PMU: debug_reg = 0xc08; break; - case NVKM_ENGINE_NVDEC: + case NVKM_ENGINE_NVDEC0: debug_reg = 0xd00; break; case NVKM_ENGINE_SEC2: diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild index e5830453813d..ab0282dc0736 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/Kbuild @@ -5,3 +5,4 @@ nvkm-y += nvkm/subdev/bar/gf100.o nvkm-y += nvkm/subdev/bar/gk20a.o nvkm-y += nvkm/subdev/bar/gm107.o nvkm-y += nvkm/subdev/bar/gm20b.o +nvkm-y += nvkm/subdev/bar/tu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c index 243f0a5c8a62..209a6a40834a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/base.c @@ -36,6 +36,16 @@ nvkm_bar_bar1_vmm(struct nvkm_device *device) return device->bar->func->bar1.vmm(device->bar); } +void +nvkm_bar_bar1_reset(struct nvkm_device *device) +{ + struct nvkm_bar *bar = device->bar; + if (bar) { + bar->func->bar1.init(bar); + bar->func->bar1.wait(bar); + } +} + struct nvkm_vmm * nvkm_bar_bar2_vmm(struct nvkm_device *device) { @@ -49,6 +59,16 @@ nvkm_bar_bar2_vmm(struct nvkm_device *device) } void +nvkm_bar_bar2_reset(struct nvkm_device *device) +{ + struct nvkm_bar *bar = device->bar; + if (bar && bar->bar2) { + bar->func->bar2.init(bar); + bar->func->bar2.wait(bar); + } +} + +void nvkm_bar_bar2_fini(struct nvkm_device *device) { struct nvkm_bar *bar = device->bar; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu104.c new file mode 100644 index 000000000000..ecaead156e9b --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/tu104.c @@ -0,0 +1,98 @@ +/* + * Copyright 2018 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. + */ +#include "gf100.h" + +#include <core/memory.h> +#include <subdev/timer.h> + +static void +tu104_bar_bar2_wait(struct nvkm_bar *bar) +{ + struct nvkm_device *device = bar->subdev.device; + nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0xb80f50) & 0x0000000c)) + break; + ); +} + +static void +tu104_bar_bar2_fini(struct nvkm_bar *bar) +{ + nvkm_mask(bar->subdev.device, 0xb80f48, 0x80000000, 0x00000000); +} + +static void +tu104_bar_bar2_init(struct nvkm_bar *base) +{ + struct nvkm_device *device = base->subdev.device; + struct gf100_bar *bar = gf100_bar(base); + u32 addr = nvkm_memory_addr(bar->bar[0].inst) >> 12; + if (bar->bar2_halve) + addr |= 0x40000000; + nvkm_wr32(device, 0xb80f48, 0x80000000 | addr); +} + +static void +tu104_bar_bar1_wait(struct nvkm_bar *bar) +{ + struct nvkm_device *device = bar->subdev.device; + nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0xb80f50) & 0x00000003)) + break; + ); +} + +static void +tu104_bar_bar1_fini(struct nvkm_bar *bar) +{ + nvkm_mask(bar->subdev.device, 0xb80f40, 0x80000000, 0x00000000); +} + +static void +tu104_bar_bar1_init(struct nvkm_bar *base) +{ + struct nvkm_device *device = base->subdev.device; + struct gf100_bar *bar = gf100_bar(base); + const u32 addr = nvkm_memory_addr(bar->bar[1].inst) >> 12; + nvkm_wr32(device, 0xb80f40, 0x80000000 | addr); +} + +static const struct nvkm_bar_func +tu104_bar = { + .dtor = gf100_bar_dtor, + .oneinit = gf100_bar_oneinit, + .bar1.init = tu104_bar_bar1_init, + .bar1.fini = tu104_bar_bar1_fini, + .bar1.wait = tu104_bar_bar1_wait, + .bar1.vmm = gf100_bar_bar1_vmm, + .bar2.init = tu104_bar_bar2_init, + .bar2.fini = tu104_bar_bar2_fini, + .bar2.wait = tu104_bar_bar2_wait, + .bar2.vmm = gf100_bar_bar2_vmm, + .flush = g84_bar_flush, +}; + +int +tu104_bar_new(struct nvkm_device *device, int index, struct nvkm_bar **pbar) +{ + return gf100_bar_new_(&tu104_bar, device, index, pbar); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild index 50a436926484..3ef505a5c01b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/Kbuild @@ -13,3 +13,4 @@ nvkm-y += nvkm/subdev/devinit/gf100.o nvkm-y += nvkm/subdev/devinit/gm107.o nvkm-y += nvkm/subdev/devinit/gm200.o nvkm-y += nvkm/subdev/devinit/gv100.o +nvkm-y += nvkm/subdev/devinit/tu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c index 17235e940ca9..59940dacc2ba 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/gm200.c @@ -105,6 +105,15 @@ pmu_load(struct nv50_devinit *init, u8 type, bool post, return pmu_exec(init, pmu.init_addr_pmu), 0; } +void +gm200_devinit_preos(struct nv50_devinit *init, bool post) +{ + /* Optional: Execute PRE_OS application on PMU, which should at + * least take care of fans until a full PMU has been loaded. + */ + pmu_load(init, 0x01, post, NULL, NULL); +} + int gm200_devinit_post(struct nvkm_devinit *base, bool post) { @@ -156,10 +165,7 @@ gm200_devinit_post(struct nvkm_devinit *base, bool post) return -ETIMEDOUT; } - /* Optional: Execute PRE_OS application on PMU, which should at - * least take care of fans until a full PMU has been loaded. - */ - pmu_load(init, 0x01, post, NULL, NULL); + gm200_devinit_preos(init, post); return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h index 9b9f0dc1e192..72d130bb7f7c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/nv50.h @@ -26,4 +26,5 @@ void gf100_devinit_preinit(struct nvkm_devinit *); u64 gm107_devinit_disable(struct nvkm_devinit *); int gm200_devinit_post(struct nvkm_devinit *, bool); +void gm200_devinit_preos(struct nv50_devinit *, bool); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu104.c new file mode 100644 index 000000000000..aae87b3fc429 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/devinit/tu104.c @@ -0,0 +1,89 @@ +/* + * Copyright 2018 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. + */ +#include "nv50.h" + +#include <subdev/bios.h> +#include <subdev/bios/pll.h> +#include <subdev/clk/pll.h> + +static int +tu104_devinit_pll_set(struct nvkm_devinit *init, u32 type, u32 freq) +{ + struct nvkm_subdev *subdev = &init->subdev; + struct nvkm_device *device = subdev->device; + struct nvbios_pll info; + int head = type - PLL_VPLL0; + int N, fN, M, P; + int ret; + + ret = nvbios_pll_parse(device->bios, type, &info); + if (ret) + return ret; + + ret = gt215_pll_calc(subdev, &info, freq, &N, &fN, &M, &P); + if (ret < 0) + return ret; + + switch (info.type) { + case PLL_VPLL0: + case PLL_VPLL1: + case PLL_VPLL2: + case PLL_VPLL3: + nvkm_wr32(device, 0x00ef10 + (head * 0x40), fN << 16); + nvkm_wr32(device, 0x00ef04 + (head * 0x40), (P << 16) | + (N << 8) | + (M << 0)); + /*XXX*/ + nvkm_wr32(device, 0x00ef0c + (head * 0x40), 0x00000900); + nvkm_wr32(device, 0x00ef00 + (head * 0x40), 0x02000014); + break; + default: + nvkm_warn(subdev, "%08x/%dKhz unimplemented\n", type, freq); + ret = -EINVAL; + break; + } + + return ret; +} + +static int +tu104_devinit_post(struct nvkm_devinit *base, bool post) +{ + struct nv50_devinit *init = nv50_devinit(base); + gm200_devinit_preos(init, post); + return 0; +} + +static const struct nvkm_devinit_func +tu104_devinit = { + .init = nv50_devinit_init, + .post = tu104_devinit_post, + .pll_set = tu104_devinit_pll_set, + .disable = gm107_devinit_disable, +}; + +int +tu104_devinit_new(struct nvkm_device *device, int index, + struct nvkm_devinit **pinit) +{ + return nv50_devinit_new_(&tu104_devinit, device, index, pinit); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/Kbuild index 45bb46fb0929..794eb1745b2f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/Kbuild @@ -1,3 +1,4 @@ nvkm-y += nvkm/subdev/fault/base.o nvkm-y += nvkm/subdev/fault/gp100.o nvkm-y += nvkm/subdev/fault/gv100.o +nvkm-y += nvkm/subdev/fault/tu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c index 16ad91c91a7b..4ba1e21e8fda 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/base.c @@ -23,21 +23,19 @@ #include <core/memory.h> #include <core/notify.h> -#include <subdev/bar.h> -#include <subdev/mmu.h> static void nvkm_fault_ntfy_fini(struct nvkm_event *event, int type, int index) { struct nvkm_fault *fault = container_of(event, typeof(*fault), event); - fault->func->buffer.fini(fault->buffer[index]); + fault->func->buffer.intr(fault->buffer[index], false); } static void nvkm_fault_ntfy_init(struct nvkm_event *event, int type, int index) { struct nvkm_fault *fault = container_of(event, typeof(*fault), event); - fault->func->buffer.init(fault->buffer[index]); + fault->func->buffer.intr(fault->buffer[index], true); } static int @@ -91,7 +89,6 @@ nvkm_fault_oneinit_buffer(struct nvkm_fault *fault, int id) { struct nvkm_subdev *subdev = &fault->subdev; struct nvkm_device *device = subdev->device; - struct nvkm_vmm *bar2 = nvkm_bar_bar2_vmm(device); struct nvkm_fault_buffer *buffer; int ret; @@ -99,7 +96,7 @@ nvkm_fault_oneinit_buffer(struct nvkm_fault *fault, int id) return -ENOMEM; buffer->fault = fault; buffer->id = id; - buffer->entries = fault->func->buffer.entries(buffer); + fault->func->buffer.info(buffer); fault->buffer[id] = buffer; nvkm_debug(subdev, "buffer %d: %d entries\n", id, buffer->entries); @@ -110,12 +107,12 @@ nvkm_fault_oneinit_buffer(struct nvkm_fault *fault, int id) if (ret) return ret; - ret = nvkm_vmm_get(bar2, 12, nvkm_memory_size(buffer->mem), - &buffer->vma); - if (ret) - return ret; + /* Pin fault buffer in BAR2. */ + buffer->addr = nvkm_memory_bar2(buffer->mem); + if (buffer->addr == ~0ULL) + return -EFAULT; - return nvkm_memory_map(buffer->mem, 0, bar2, buffer->vma, NULL, 0); + return 0; } static int @@ -146,7 +143,6 @@ nvkm_fault_oneinit(struct nvkm_subdev *subdev) static void * nvkm_fault_dtor(struct nvkm_subdev *subdev) { - struct nvkm_vmm *bar2 = nvkm_bar_bar2_vmm(subdev->device); struct nvkm_fault *fault = nvkm_fault(subdev); int i; @@ -154,7 +150,6 @@ nvkm_fault_dtor(struct nvkm_subdev *subdev) for (i = 0; i < fault->buffer_nr; i++) { if (fault->buffer[i]) { - nvkm_vmm_put(bar2, &fault->buffer[i]->vma); nvkm_memory_unref(&fault->buffer[i]->mem); kfree(fault->buffer[i]); } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c index 5e71db2e8d75..8fb96fe614f9 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gp100.c @@ -21,7 +21,14 @@ */ #include "priv.h" -#include <subdev/mmu.h> +#include <subdev/mc.h> + +static void +gp100_fault_buffer_intr(struct nvkm_fault_buffer *buffer, bool enable) +{ + struct nvkm_device *device = buffer->fault->subdev.device; + nvkm_mc_intr_mask(device, NVKM_SUBDEV_FAULT, enable); +} static void gp100_fault_buffer_fini(struct nvkm_fault_buffer *buffer) @@ -34,15 +41,17 @@ static void gp100_fault_buffer_init(struct nvkm_fault_buffer *buffer) { struct nvkm_device *device = buffer->fault->subdev.device; - nvkm_wr32(device, 0x002a74, upper_32_bits(buffer->vma->addr)); - nvkm_wr32(device, 0x002a70, lower_32_bits(buffer->vma->addr)); + nvkm_wr32(device, 0x002a74, upper_32_bits(buffer->addr)); + nvkm_wr32(device, 0x002a70, lower_32_bits(buffer->addr)); nvkm_mask(device, 0x002a70, 0x00000001, 0x00000001); } -static u32 -gp100_fault_buffer_entries(struct nvkm_fault_buffer *buffer) +static void +gp100_fault_buffer_info(struct nvkm_fault_buffer *buffer) { - return nvkm_rd32(buffer->fault->subdev.device, 0x002a78); + buffer->entries = nvkm_rd32(buffer->fault->subdev.device, 0x002a78); + buffer->get = 0x002a7c; + buffer->put = 0x002a80; } static void @@ -56,9 +65,10 @@ gp100_fault = { .intr = gp100_fault_intr, .buffer.nr = 1, .buffer.entry_size = 32, - .buffer.entries = gp100_fault_buffer_entries, + .buffer.info = gp100_fault_buffer_info, .buffer.init = gp100_fault_buffer_init, .buffer.fini = gp100_fault_buffer_fini, + .buffer.intr = gp100_fault_buffer_intr, }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c index 3cd610d7deb5..6fc54e17c935 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/gv100.c @@ -30,9 +30,8 @@ gv100_fault_buffer_process(struct nvkm_fault_buffer *buffer) { struct nvkm_device *device = buffer->fault->subdev.device; struct nvkm_memory *mem = buffer->mem; - const u32 foff = buffer->id * 0x14; - u32 get = nvkm_rd32(device, 0x100e2c + foff); - u32 put = nvkm_rd32(device, 0x100e30 + foff); + u32 get = nvkm_rd32(device, buffer->get); + u32 put = nvkm_rd32(device, buffer->put); if (put == get) return; @@ -51,7 +50,7 @@ gv100_fault_buffer_process(struct nvkm_fault_buffer *buffer) if (++get == buffer->entries) get = 0; - nvkm_wr32(device, 0x100e2c + foff, get); + nvkm_wr32(device, buffer->get, get); info.addr = ((u64)addrhi << 32) | addrlo; info.inst = ((u64)insthi << 32) | instlo; @@ -70,13 +69,21 @@ gv100_fault_buffer_process(struct nvkm_fault_buffer *buffer) } static void -gv100_fault_buffer_fini(struct nvkm_fault_buffer *buffer) +gv100_fault_buffer_intr(struct nvkm_fault_buffer *buffer, bool enable) { struct nvkm_device *device = buffer->fault->subdev.device; const u32 intr = buffer->id ? 0x08000000 : 0x20000000; - const u32 foff = buffer->id * 0x14; + if (enable) + nvkm_mask(device, 0x100a2c, intr, intr); + else + nvkm_mask(device, 0x100a34, intr, intr); +} - nvkm_mask(device, 0x100a34, intr, intr); +static void +gv100_fault_buffer_fini(struct nvkm_fault_buffer *buffer) +{ + struct nvkm_device *device = buffer->fault->subdev.device; + const u32 foff = buffer->id * 0x14; nvkm_mask(device, 0x100e34 + foff, 0x80000000, 0x00000000); } @@ -84,23 +91,25 @@ static void gv100_fault_buffer_init(struct nvkm_fault_buffer *buffer) { struct nvkm_device *device = buffer->fault->subdev.device; - const u32 intr = buffer->id ? 0x08000000 : 0x20000000; const u32 foff = buffer->id * 0x14; nvkm_mask(device, 0x100e34 + foff, 0xc0000000, 0x40000000); - nvkm_wr32(device, 0x100e28 + foff, upper_32_bits(buffer->vma->addr)); - nvkm_wr32(device, 0x100e24 + foff, lower_32_bits(buffer->vma->addr)); + nvkm_wr32(device, 0x100e28 + foff, upper_32_bits(buffer->addr)); + nvkm_wr32(device, 0x100e24 + foff, lower_32_bits(buffer->addr)); nvkm_mask(device, 0x100e34 + foff, 0x80000000, 0x80000000); - nvkm_mask(device, 0x100a2c, intr, intr); } -static u32 -gv100_fault_buffer_entries(struct nvkm_fault_buffer *buffer) +static void +gv100_fault_buffer_info(struct nvkm_fault_buffer *buffer) { struct nvkm_device *device = buffer->fault->subdev.device; const u32 foff = buffer->id * 0x14; + nvkm_mask(device, 0x100e34 + foff, 0x40000000, 0x40000000); - return nvkm_rd32(device, 0x100e34 + foff) & 0x000fffff; + + buffer->entries = nvkm_rd32(device, 0x100e34 + foff) & 0x000fffff; + buffer->get = 0x100e2c + foff; + buffer->put = 0x100e30 + foff; } static int @@ -166,6 +175,8 @@ static void gv100_fault_fini(struct nvkm_fault *fault) { nvkm_notify_put(&fault->nrpfb); + if (fault->buffer[0]) + fault->func->buffer.fini(fault->buffer[0]); nvkm_mask(fault->subdev.device, 0x100a34, 0x80000000, 0x80000000); } @@ -173,14 +184,15 @@ static void gv100_fault_init(struct nvkm_fault *fault) { nvkm_mask(fault->subdev.device, 0x100a2c, 0x80000000, 0x80000000); + fault->func->buffer.init(fault->buffer[0]); nvkm_notify_get(&fault->nrpfb); } -static int +int gv100_fault_oneinit(struct nvkm_fault *fault) { return nvkm_notify_init(&fault->buffer[0]->object, &fault->event, - gv100_fault_ntfy_nrpfb, false, NULL, 0, 0, + gv100_fault_ntfy_nrpfb, true, NULL, 0, 0, &fault->nrpfb); } @@ -192,9 +204,10 @@ gv100_fault = { .intr = gv100_fault_intr, .buffer.nr = 2, .buffer.entry_size = 32, - .buffer.entries = gv100_fault_buffer_entries, + .buffer.info = gv100_fault_buffer_info, .buffer.init = gv100_fault_buffer_init, .buffer.fini = gv100_fault_buffer_fini, + .buffer.intr = gv100_fault_buffer_intr, }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h index e4d2f5234fd1..8ca8b2876dad 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/priv.h @@ -12,8 +12,10 @@ struct nvkm_fault_buffer { struct nvkm_fault *fault; int id; int entries; + u32 get; + u32 put; struct nvkm_memory *mem; - struct nvkm_vma *vma; + u64 addr; }; int nvkm_fault_new_(const struct nvkm_fault_func *, struct nvkm_device *, @@ -27,9 +29,12 @@ struct nvkm_fault_func { struct { int nr; u32 entry_size; - u32 (*entries)(struct nvkm_fault_buffer *); + void (*info)(struct nvkm_fault_buffer *); void (*init)(struct nvkm_fault_buffer *); void (*fini)(struct nvkm_fault_buffer *); + void (*intr)(struct nvkm_fault_buffer *, bool enable); } buffer; }; + +int gv100_fault_oneinit(struct nvkm_fault *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu104.c new file mode 100644 index 000000000000..9c8a3adf99d7 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fault/tu104.c @@ -0,0 +1,167 @@ +/* + * Copyright 2018 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. + */ +#include "priv.h" + +#include <core/memory.h> +#include <subdev/mmu.h> +#include <engine/fifo.h> + +#include <nvif/class.h> + +static void +tu104_fault_buffer_intr(struct nvkm_fault_buffer *buffer, bool enable) +{ + /*XXX: Earlier versions of RM touched the old regs on Turing, + * which don't appear to actually work anymore, but newer + * versions of RM don't appear to touch anything at all.. + */ +} + +static void +tu104_fault_buffer_fini(struct nvkm_fault_buffer *buffer) +{ + struct nvkm_device *device = buffer->fault->subdev.device; + const u32 foff = buffer->id * 0x20; + nvkm_mask(device, 0xb83010 + foff, 0x80000000, 0x00000000); +} + +static void +tu104_fault_buffer_init(struct nvkm_fault_buffer *buffer) +{ + struct nvkm_device *device = buffer->fault->subdev.device; + const u32 foff = buffer->id * 0x20; + + nvkm_mask(device, 0xb83010 + foff, 0xc0000000, 0x40000000); + nvkm_wr32(device, 0xb83004 + foff, upper_32_bits(buffer->addr)); + nvkm_wr32(device, 0xb83000 + foff, lower_32_bits(buffer->addr)); + nvkm_mask(device, 0xb83010 + foff, 0x80000000, 0x80000000); +} + +static void +tu104_fault_buffer_info(struct nvkm_fault_buffer *buffer) +{ + struct nvkm_device *device = buffer->fault->subdev.device; + const u32 foff = buffer->id * 0x20; + + nvkm_mask(device, 0xb83010 + foff, 0x40000000, 0x40000000); + + buffer->entries = nvkm_rd32(device, 0xb83010 + foff) & 0x000fffff; + buffer->get = 0xb83008 + foff; + buffer->put = 0xb8300c + foff; +} + +static void +tu104_fault_intr_fault(struct nvkm_fault *fault) +{ + struct nvkm_subdev *subdev = &fault->subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_fault_data info; + const u32 addrlo = nvkm_rd32(device, 0xb83080); + const u32 addrhi = nvkm_rd32(device, 0xb83084); + const u32 info0 = nvkm_rd32(device, 0xb83088); + const u32 insthi = nvkm_rd32(device, 0xb8308c); + const u32 info1 = nvkm_rd32(device, 0xb83090); + + info.addr = ((u64)addrhi << 32) | addrlo; + info.inst = ((u64)insthi << 32) | (info0 & 0xfffff000); + info.time = 0; + info.engine = (info0 & 0x000000ff); + info.valid = (info1 & 0x80000000) >> 31; + info.gpc = (info1 & 0x1f000000) >> 24; + info.hub = (info1 & 0x00100000) >> 20; + info.access = (info1 & 0x000f0000) >> 16; + info.client = (info1 & 0x00007f00) >> 8; + info.reason = (info1 & 0x0000001f); + + nvkm_fifo_fault(device->fifo, &info); +} + +static void +tu104_fault_intr(struct nvkm_fault *fault) +{ + struct nvkm_subdev *subdev = &fault->subdev; + struct nvkm_device *device = subdev->device; + u32 stat = nvkm_rd32(device, 0xb83094); + + if (stat & 0x80000000) { + tu104_fault_intr_fault(fault); + nvkm_wr32(device, 0xb83094, 0x80000000); + stat &= ~0x80000000; + } + + if (stat & 0x00000200) { + if (fault->buffer[0]) { + nvkm_event_send(&fault->event, 1, 0, NULL, 0); + stat &= ~0x00000200; + } + } + + /*XXX: guess, can't confirm until we get fw... */ + if (stat & 0x00000100) { + if (fault->buffer[1]) { + nvkm_event_send(&fault->event, 1, 1, NULL, 0); + stat &= ~0x00000100; + } + } + + if (stat) { + nvkm_debug(subdev, "intr %08x\n", stat); + } +} + +static void +tu104_fault_fini(struct nvkm_fault *fault) +{ + nvkm_notify_put(&fault->nrpfb); + if (fault->buffer[0]) + fault->func->buffer.fini(fault->buffer[0]); + /*XXX: disable priv faults */ +} + +static void +tu104_fault_init(struct nvkm_fault *fault) +{ + /*XXX: enable priv faults */ + fault->func->buffer.init(fault->buffer[0]); + nvkm_notify_get(&fault->nrpfb); +} + +static const struct nvkm_fault_func +tu104_fault = { + .oneinit = gv100_fault_oneinit, + .init = tu104_fault_init, + .fini = tu104_fault_fini, + .intr = tu104_fault_intr, + .buffer.nr = 2, + .buffer.entry_size = 32, + .buffer.info = tu104_fault_buffer_info, + .buffer.init = tu104_fault_buffer_init, + .buffer.fini = tu104_fault_buffer_fini, + .buffer.intr = tu104_fault_buffer_intr, +}; + +int +tu104_fault_new(struct nvkm_device *device, int index, + struct nvkm_fault **pfault) +{ + return nvkm_fault_new_(&tu104_fault, device, index, pfault); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c index 434d2fc5bb1c..b2bb5a3ccb02 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c @@ -68,10 +68,13 @@ nvkm_fb_bios_memtype(struct nvkm_bios *bios) if (nvbios_M0203Em(bios, ramcfg, &ver, &hdr, &M0203E)) { switch (M0203E.type) { - case M0203E_TYPE_DDR2 : return NVKM_RAM_TYPE_DDR2; - case M0203E_TYPE_DDR3 : return NVKM_RAM_TYPE_DDR3; - case M0203E_TYPE_GDDR3: return NVKM_RAM_TYPE_GDDR3; - case M0203E_TYPE_GDDR5: return NVKM_RAM_TYPE_GDDR5; + case M0203E_TYPE_DDR2 : return NVKM_RAM_TYPE_DDR2; + case M0203E_TYPE_DDR3 : return NVKM_RAM_TYPE_DDR3; + case M0203E_TYPE_GDDR3 : return NVKM_RAM_TYPE_GDDR3; + case M0203E_TYPE_GDDR5 : return NVKM_RAM_TYPE_GDDR5; + case M0203E_TYPE_GDDR5X: return NVKM_RAM_TYPE_GDDR5X; + case M0203E_TYPE_GDDR6 : return NVKM_RAM_TYPE_GDDR6; + case M0203E_TYPE_HBM2 : return NVKM_RAM_TYPE_HBM2; default: nvkm_warn(subdev, "M0203E type %02x\n", M0203E.type); return NVKM_RAM_TYPE_UNKNOWN; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c index 24c7bd505731..b11867f682cb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.c @@ -184,6 +184,9 @@ nvkm_ram_ctor(const struct nvkm_ram_func *func, struct nvkm_fb *fb, [NVKM_RAM_TYPE_GDDR3 ] = "GDDR3", [NVKM_RAM_TYPE_GDDR4 ] = "GDDR4", [NVKM_RAM_TYPE_GDDR5 ] = "GDDR5", + [NVKM_RAM_TYPE_GDDR5X ] = "GDDR5X", + [NVKM_RAM_TYPE_GDDR6 ] = "GDDR6", + [NVKM_RAM_TYPE_HBM2 ] = "HBM2", }; struct nvkm_subdev *subdev = &fb->subdev; int ret; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c index db48a1daca0c..02c4eb28cef4 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c @@ -288,6 +288,19 @@ nv50_instobj_addr(struct nvkm_memory *memory) return nvkm_memory_addr(nv50_instobj(memory)->ram); } +static u64 +nv50_instobj_bar2(struct nvkm_memory *memory) +{ + struct nv50_instobj *iobj = nv50_instobj(memory); + u64 addr = ~0ULL; + if (nv50_instobj_acquire(&iobj->base.memory)) { + iobj->lru.next = NULL; /* Exclude from eviction. */ + addr = iobj->bar->addr; + } + nv50_instobj_release(&iobj->base.memory); + return addr; +} + static enum nvkm_memory_target nv50_instobj_target(struct nvkm_memory *memory) { @@ -325,8 +338,9 @@ static const struct nvkm_memory_func nv50_instobj_func = { .dtor = nv50_instobj_dtor, .target = nv50_instobj_target, - .size = nv50_instobj_size, + .bar2 = nv50_instobj_bar2, .addr = nv50_instobj_addr, + .size = nv50_instobj_size, .boot = nv50_instobj_boot, .acquire = nv50_instobj_acquire, .release = nv50_instobj_release, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild index 2befbe36dc28..f3b06329c338 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild @@ -12,3 +12,4 @@ nvkm-y += nvkm/subdev/mc/gk104.o nvkm-y += nvkm/subdev/mc/gk20a.o nvkm-y += nvkm/subdev/mc/gp100.o nvkm-y += nvkm/subdev/mc/gp10b.o +nvkm-y += nvkm/subdev/mc/tu104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c index 09f669ac6630..0e57ab2a709f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c @@ -108,6 +108,9 @@ nvkm_mc_intr(struct nvkm_device *device, bool *handled) if (stat) nvkm_error(&mc->subdev, "intr %08x\n", stat); *handled = intr != 0; + + if (mc->func->intr_hack) + mc->func->intr_hack(mc, handled); } static u32 diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h index d9e3691d45b7..eb91a4cf452b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h @@ -26,6 +26,7 @@ struct nvkm_mc_func { void (*intr_mask)(struct nvkm_mc *, u32 mask, u32 stat); /* retrieve pending interrupt mask (NV_PMC_INTR) */ u32 (*intr_stat)(struct nvkm_mc *); + void (*intr_hack)(struct nvkm_mc *, bool *handled); const struct nvkm_mc_map *reset; void (*unk260)(struct nvkm_mc *, u32); }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu104.c new file mode 100644 index 000000000000..b7165bd18999 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/tu104.c @@ -0,0 +1,55 @@ +/* + * Copyright 2018 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. + */ +#include "priv.h" + +static void +tu104_mc_intr_hack(struct nvkm_mc *mc, bool *handled) +{ + struct nvkm_device *device = mc->subdev.device; + u32 stat = nvkm_rd32(device, 0xb81010); + if (stat & 0x00000050) { + struct nvkm_subdev *subdev = + nvkm_device_subdev(device, NVKM_SUBDEV_FAULT); + nvkm_wr32(device, 0xb81010, stat & 0x00000050); + if (subdev) + nvkm_subdev_intr(subdev); + *handled = true; + } +} + +static const struct nvkm_mc_func +tu104_mc = { + .init = nv50_mc_init, + .intr = gp100_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, + .intr_hack = tu104_mc_intr_hack, + .reset = gk104_mc_reset, +}; + +int +tu104_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc) +{ + return gp100_mc_new_(&tu104_mc, device, index, pmc); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild index 58a24e3a0598..8966180b36cc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/Kbuild @@ -13,6 +13,7 @@ nvkm-y += nvkm/subdev/mmu/gm20b.o nvkm-y += nvkm/subdev/mmu/gp100.o nvkm-y += nvkm/subdev/mmu/gp10b.o nvkm-y += nvkm/subdev/mmu/gv100.o +nvkm-y += nvkm/subdev/mmu/tu104.o nvkm-y += nvkm/subdev/mmu/mem.o nvkm-y += nvkm/subdev/mmu/memnv04.o @@ -33,6 +34,7 @@ nvkm-y += nvkm/subdev/mmu/vmmgm20b.o nvkm-y += nvkm/subdev/mmu/vmmgp100.o nvkm-y += nvkm/subdev/mmu/vmmgp10b.o nvkm-y += nvkm/subdev/mmu/vmmgv100.o +nvkm-y += nvkm/subdev/mmu/vmmtu104.o nvkm-y += nvkm/subdev/mmu/umem.o nvkm-y += nvkm/subdev/mmu/ummu.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu104.c new file mode 100644 index 000000000000..8e6f4096170d --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/tu104.c @@ -0,0 +1,43 @@ +/* + * Copyright 2018 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. + */ +#include "mem.h" +#include "vmm.h" + +#include <core/option.h> + +#include <nvif/class.h> + +static const struct nvkm_mmu_func +tu104_mmu = { + .dma_bits = 47, + .mmu = {{ -1, -1, NVIF_CLASS_MMU_GF100}}, + .mem = {{ -1, 0, NVIF_CLASS_MEM_GF100}, gf100_mem_new, gf100_mem_map }, + .vmm = {{ -1, 0, NVIF_CLASS_VMM_GP100}, tu104_vmm_new }, + .kind = gm200_mmu_kind, + .kind_sys = true, +}; + +int +tu104_mmu_new(struct nvkm_device *device, int index, struct nvkm_mmu **pmmu) +{ + return nvkm_mmu_new_(&tu104_mmu, device, index, pmmu); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c index 37b201b95f15..6889076097ec 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c @@ -134,23 +134,10 @@ nvkm_uvmm_mthd_map(struct nvkm_uvmm *uvmm, void *argv, u32 argc) goto fail; } - if (vma->addr != addr) { - const u64 tail = vma->size + vma->addr - addr; - if (ret = -ENOMEM, !(vma = nvkm_vma_tail(vma, tail))) - goto fail; - vma->part = true; - nvkm_vmm_node_insert(vmm, vma); - } - - if (vma->size != size) { - const u64 tail = vma->size - size; - struct nvkm_vma *tmp; - if (ret = -ENOMEM, !(tmp = nvkm_vma_tail(vma, tail))) { - nvkm_vmm_unmap_region(vmm, vma); - goto fail; - } - tmp->part = true; - nvkm_vmm_node_insert(vmm, tmp); + vma = nvkm_vmm_node_split(vmm, vma, addr, size); + if (!vma) { + ret = -ENOMEM; + goto fail; } } vma->busy = true; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c index 7459def78d50..6b87fff014b3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c @@ -767,6 +767,20 @@ nvkm_vma_tail(struct nvkm_vma *vma, u64 tail) return new; } +static inline void +nvkm_vmm_free_remove(struct nvkm_vmm *vmm, struct nvkm_vma *vma) +{ + rb_erase(&vma->tree, &vmm->free); +} + +static inline void +nvkm_vmm_free_delete(struct nvkm_vmm *vmm, struct nvkm_vma *vma) +{ + nvkm_vmm_free_remove(vmm, vma); + list_del(&vma->head); + kfree(vma); +} + static void nvkm_vmm_free_insert(struct nvkm_vmm *vmm, struct nvkm_vma *vma) { @@ -795,7 +809,21 @@ nvkm_vmm_free_insert(struct nvkm_vmm *vmm, struct nvkm_vma *vma) rb_insert_color(&vma->tree, &vmm->free); } -void +static inline void +nvkm_vmm_node_remove(struct nvkm_vmm *vmm, struct nvkm_vma *vma) +{ + rb_erase(&vma->tree, &vmm->root); +} + +static inline void +nvkm_vmm_node_delete(struct nvkm_vmm *vmm, struct nvkm_vma *vma) +{ + nvkm_vmm_node_remove(vmm, vma); + list_del(&vma->head); + kfree(vma); +} + +static void nvkm_vmm_node_insert(struct nvkm_vmm *vmm, struct nvkm_vma *vma) { struct rb_node **ptr = &vmm->root.rb_node; @@ -834,6 +862,78 @@ nvkm_vmm_node_search(struct nvkm_vmm *vmm, u64 addr) return NULL; } +#define node(root, dir) (((root)->head.dir == &vmm->list) ? NULL : \ + list_entry((root)->head.dir, struct nvkm_vma, head)) + +static struct nvkm_vma * +nvkm_vmm_node_merge(struct nvkm_vmm *vmm, struct nvkm_vma *prev, + struct nvkm_vma *vma, struct nvkm_vma *next, u64 size) +{ + if (next) { + if (vma->size == size) { + vma->size += next->size; + nvkm_vmm_node_delete(vmm, next); + if (prev) { + prev->size += vma->size; + nvkm_vmm_node_delete(vmm, vma); + return prev; + } + return vma; + } + BUG_ON(prev); + + nvkm_vmm_node_remove(vmm, next); + vma->size -= size; + next->addr -= size; + next->size += size; + nvkm_vmm_node_insert(vmm, next); + return next; + } + + if (prev) { + if (vma->size != size) { + nvkm_vmm_node_remove(vmm, vma); + prev->size += size; + vma->addr += size; + vma->size -= size; + nvkm_vmm_node_insert(vmm, vma); + } else { + prev->size += vma->size; + nvkm_vmm_node_delete(vmm, vma); + } + return prev; + } + + return vma; +} + +struct nvkm_vma * +nvkm_vmm_node_split(struct nvkm_vmm *vmm, + struct nvkm_vma *vma, u64 addr, u64 size) +{ + struct nvkm_vma *prev = NULL; + + if (vma->addr != addr) { + prev = vma; + if (!(vma = nvkm_vma_tail(vma, vma->size + vma->addr - addr))) + return NULL; + vma->part = true; + nvkm_vmm_node_insert(vmm, vma); + } + + if (vma->size != size) { + struct nvkm_vma *tmp; + if (!(tmp = nvkm_vma_tail(vma, vma->size - size))) { + nvkm_vmm_node_merge(vmm, prev, vma, NULL, vma->size); + return NULL; + } + tmp->part = true; + nvkm_vmm_node_insert(vmm, tmp); + } + + return vma; +} + static void nvkm_vmm_dtor(struct nvkm_vmm *vmm) { @@ -954,37 +1054,20 @@ nvkm_vmm_new_(const struct nvkm_vmm_func *func, struct nvkm_mmu *mmu, return nvkm_vmm_ctor(func, mmu, hdr, addr, size, key, name, *pvmm); } -#define node(root, dir) ((root)->head.dir == &vmm->list) ? NULL : \ - list_entry((root)->head.dir, struct nvkm_vma, head) - void nvkm_vmm_unmap_region(struct nvkm_vmm *vmm, struct nvkm_vma *vma) { - struct nvkm_vma *next; + struct nvkm_vma *next = node(vma, next); + struct nvkm_vma *prev = NULL; nvkm_memory_tags_put(vma->memory, vmm->mmu->subdev.device, &vma->tags); nvkm_memory_unref(&vma->memory); - if (vma->part) { - struct nvkm_vma *prev = node(vma, prev); - if (!prev->memory) { - prev->size += vma->size; - rb_erase(&vma->tree, &vmm->root); - list_del(&vma->head); - kfree(vma); - vma = prev; - } - } - - next = node(vma, next); - if (next && next->part) { - if (!next->memory) { - vma->size += next->size; - rb_erase(&next->tree, &vmm->root); - list_del(&next->head); - kfree(next); - } - } + if (!vma->part || ((prev = node(vma, prev)), prev->memory)) + prev = NULL; + if (!next->part || next->memory) + next = NULL; + nvkm_vmm_node_merge(vmm, prev, vma, next, vma->size); } void @@ -1163,18 +1246,14 @@ nvkm_vmm_put_region(struct nvkm_vmm *vmm, struct nvkm_vma *vma) struct nvkm_vma *prev, *next; if ((prev = node(vma, prev)) && !prev->used) { - rb_erase(&prev->tree, &vmm->free); - list_del(&prev->head); vma->addr = prev->addr; vma->size += prev->size; - kfree(prev); + nvkm_vmm_free_delete(vmm, prev); } if ((next = node(vma, next)) && !next->used) { - rb_erase(&next->tree, &vmm->free); - list_del(&next->head); vma->size += next->size; - kfree(next); + nvkm_vmm_free_delete(vmm, next); } nvkm_vmm_free_insert(vmm, vma); @@ -1250,7 +1329,7 @@ nvkm_vmm_put_locked(struct nvkm_vmm *vmm, struct nvkm_vma *vma) } /* Remove VMA from the list of allocated nodes. */ - rb_erase(&vma->tree, &vmm->root); + nvkm_vmm_node_remove(vmm, vma); /* Merge VMA back into the free list. */ vma->page = NVKM_VMA_PAGE_NONE; @@ -1357,7 +1436,7 @@ nvkm_vmm_get_locked(struct nvkm_vmm *vmm, bool getref, bool mapref, bool sparse, tail = ALIGN_DOWN(tail, vmm->func->page_block); if (addr <= tail && tail - addr >= size) { - rb_erase(&this->tree, &vmm->free); + nvkm_vmm_free_remove(vmm, this); vma = this; break; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h index 1a3b0a3724ca..42ad326521a3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h @@ -157,6 +157,8 @@ int nvkm_vmm_ctor(const struct nvkm_vmm_func *, struct nvkm_mmu *, u32 pd_header, u64 addr, u64 size, struct lock_class_key *, const char *name, struct nvkm_vmm *); struct nvkm_vma *nvkm_vmm_node_search(struct nvkm_vmm *, u64 addr); +struct nvkm_vma *nvkm_vmm_node_split(struct nvkm_vmm *, struct nvkm_vma *, + u64 addr, u64 size); int nvkm_vmm_get_locked(struct nvkm_vmm *, bool getref, bool mapref, bool sparse, u8 page, u8 align, u64 size, struct nvkm_vma **pvma); @@ -165,7 +167,6 @@ void nvkm_vmm_unmap_locked(struct nvkm_vmm *, struct nvkm_vma *); void nvkm_vmm_unmap_region(struct nvkm_vmm *vmm, struct nvkm_vma *vma); struct nvkm_vma *nvkm_vma_tail(struct nvkm_vma *, u64 tail); -void nvkm_vmm_node_insert(struct nvkm_vmm *, struct nvkm_vma *); int nv04_vmm_new_(const struct nvkm_vmm_func *, struct nvkm_mmu *, u32, u64, u64, void *, u32, struct lock_class_key *, @@ -200,6 +201,8 @@ int gp100_vmm_join(struct nvkm_vmm *, struct nvkm_memory *); int gp100_vmm_valid(struct nvkm_vmm *, void *, u32, struct nvkm_vmm_map *); void gp100_vmm_flush(struct nvkm_vmm *, int); +int gv100_vmm_join(struct nvkm_vmm *, struct nvkm_memory *); + int nv04_vmm_new(struct nvkm_mmu *, u64, u64, void *, u32, struct lock_class_key *, const char *, struct nvkm_vmm **); int nv41_vmm_new(struct nvkm_mmu *, u64, u64, void *, u32, @@ -239,6 +242,9 @@ int gp10b_vmm_new(struct nvkm_mmu *, u64, u64, void *, u32, int gv100_vmm_new(struct nvkm_mmu *, u64, u64, void *, u32, struct lock_class_key *, const char *, struct nvkm_vmm **); +int tu104_vmm_new(struct nvkm_mmu *, u64, u64, void *, u32, + struct lock_class_key *, const char *, + struct nvkm_vmm **); #define VMM_PRINT(l,v,p,f,a...) do { \ struct nvkm_vmm *_vmm = (v); \ diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu104.c new file mode 100644 index 000000000000..adaadd92110f --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu104.c @@ -0,0 +1,77 @@ +/* + * Copyright 2018 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. + */ +#include "vmm.h" + +#include <subdev/timer.h> + +static void +tu104_vmm_flush(struct nvkm_vmm *vmm, int depth) +{ + struct nvkm_subdev *subdev = &vmm->mmu->subdev; + struct nvkm_device *device = subdev->device; + u32 type = depth << 24; /*XXX: not confirmed */ + + type = 0x00000001; /* PAGE_ALL */ + if (atomic_read(&vmm->engref[NVKM_SUBDEV_BAR])) + type |= 0x00000004; /* HUB_ONLY */ + + mutex_lock(&subdev->mutex); + + nvkm_wr32(device, 0xb830a0, vmm->pd->pt[0]->addr >> 8); + nvkm_wr32(device, 0xb830a4, 0x00000000); + nvkm_wr32(device, 0x100e68, 0x00000000); + nvkm_wr32(device, 0xb830b0, 0x80000000 | type); + + nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0xb830b0) & 0x80000000)) + break; + ); + + mutex_unlock(&subdev->mutex); +} + +static const struct nvkm_vmm_func +tu104_vmm = { + .join = gv100_vmm_join, + .part = gf100_vmm_part, + .aper = gf100_vmm_aper, + .valid = gp100_vmm_valid, + .flush = tu104_vmm_flush, + .page = { + { 47, &gp100_vmm_desc_16[4], NVKM_VMM_PAGE_Sxxx }, + { 38, &gp100_vmm_desc_16[3], NVKM_VMM_PAGE_Sxxx }, + { 29, &gp100_vmm_desc_16[2], NVKM_VMM_PAGE_Sxxx }, + { 21, &gp100_vmm_desc_16[1], NVKM_VMM_PAGE_SVxC }, + { 16, &gp100_vmm_desc_16[0], NVKM_VMM_PAGE_SVxC }, + { 12, &gp100_vmm_desc_12[0], NVKM_VMM_PAGE_SVHx }, + {} + } +}; + +int +tu104_vmm_new(struct nvkm_mmu *mmu, u64 addr, u64 size, + void *argv, u32 argc, struct lock_class_key *key, + const char *name, struct nvkm_vmm **pvmm) +{ + return nv04_vmm_new_(&tu104_vmm, mmu, 0, addr, size, + argv, argc, key, name, pvmm); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gp102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gp102.c index 1f7a3c1a7f50..84a2f243ed9b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gp102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gp102.c @@ -59,10 +59,10 @@ gp102_run_secure_scrub(struct nvkm_secboot *sb) nvkm_debug(subdev, "running VPR scrubber binary on NVDEC...\n"); - engine = nvkm_engine_ref(&device->nvdec->engine); + engine = nvkm_engine_ref(&device->nvdec[0]->engine); if (IS_ERR(engine)) return PTR_ERR(engine); - falcon = device->nvdec->falcon; + falcon = device->nvdec[0]->falcon; nvkm_falcon_get(falcon, &sb->subdev); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c index 3695cde669f8..07914e36939e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/therm/base.c @@ -132,11 +132,12 @@ nvkm_therm_update(struct nvkm_therm *therm, int mode) duty = nvkm_therm_update_linear(therm); break; case NVBIOS_THERM_FAN_OTHER: - if (therm->cstate) + if (therm->cstate) { duty = therm->cstate; - else + poll = false; + } else { duty = nvkm_therm_update_linear_fallback(therm); - poll = false; + } break; } immd = false; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c index 36de23d12ae4..dd922033628c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c @@ -23,6 +23,42 @@ */ #include "priv.h" +s64 +nvkm_timer_wait_test(struct nvkm_timer_wait *wait) +{ + struct nvkm_subdev *subdev = &wait->tmr->subdev; + u64 time = nvkm_timer_read(wait->tmr); + + if (wait->reads == 0) { + wait->time0 = time; + wait->time1 = time; + } + + if (wait->time1 == time) { + if (wait->reads++ == 16) { + nvkm_fatal(subdev, "stalled at %016llx\n", time); + return -ETIMEDOUT; + } + } else { + wait->time1 = time; + wait->reads = 1; + } + + if (wait->time1 - wait->time0 > wait->limit) + return -ETIMEDOUT; + + return wait->time1 - wait->time0; +} + +void +nvkm_timer_wait_init(struct nvkm_device *device, u64 nsec, + struct nvkm_timer_wait *wait) +{ + wait->tmr = device->timer; + wait->limit = nsec; + wait->reads = 0; +} + u64 nvkm_timer_read(struct nvkm_timer *tmr) { diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c index 4f1f3e890650..39081eadfd84 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c @@ -86,7 +86,7 @@ gk104_top_oneinit(struct nvkm_top *top) case 0x0000000d: A_(SEC2 ); break; case 0x0000000e: B_(NVENC ); break; case 0x0000000f: A_(NVENC1); break; - case 0x00000010: A_(NVDEC ); break; + case 0x00000010: B_(NVDEC ); break; case 0x00000013: B_(CE ); break; break; default: diff --git a/drivers/gpu/drm/omapdrm/displays/panel-dpi.c b/drivers/gpu/drm/omapdrm/displays/panel-dpi.c index 1f8161b041be..465120809eb3 100644 --- a/drivers/gpu/drm/omapdrm/displays/panel-dpi.c +++ b/drivers/gpu/drm/omapdrm/displays/panel-dpi.c @@ -177,6 +177,7 @@ static int panel_dpi_probe(struct platform_device *pdev) dssdev->type = OMAP_DISPLAY_TYPE_DPI; dssdev->owner = THIS_MODULE; dssdev->of_ports = BIT(0); + drm_bus_flags_from_videomode(&ddata->vm, &dssdev->bus_flags); omapdss_display_init(dssdev); omapdss_device_register(dssdev); diff --git a/drivers/gpu/drm/omapdrm/dss/dsi.c b/drivers/gpu/drm/omapdrm/dss/dsi.c index 0a485c5b982e..00a9c2ab9e6c 100644 --- a/drivers/gpu/drm/omapdrm/dss/dsi.c +++ b/drivers/gpu/drm/omapdrm/dss/dsi.c @@ -5418,9 +5418,15 @@ static int dsi_probe(struct platform_device *pdev) dsi->num_lanes_supported = 3; } + r = of_platform_populate(dev->of_node, NULL, NULL, dev); + if (r) { + DSSERR("Failed to populate DSI child devices: %d\n", r); + goto err_pm_disable; + } + r = dsi_init_output(dsi); if (r) - goto err_pm_disable; + goto err_of_depopulate; r = dsi_probe_of(dsi); if (r) { @@ -5428,22 +5434,16 @@ static int dsi_probe(struct platform_device *pdev) goto err_uninit_output; } - r = of_platform_populate(dev->of_node, NULL, NULL, dev); - if (r) { - DSSERR("Failed to populate DSI child devices: %d\n", r); - goto err_uninit_output; - } - r = component_add(&pdev->dev, &dsi_component_ops); if (r) - goto err_of_depopulate; + goto err_uninit_output; return 0; -err_of_depopulate: - of_platform_depopulate(dev); err_uninit_output: dsi_uninit_output(dsi); +err_of_depopulate: + of_platform_depopulate(dev); err_pm_disable: pm_runtime_disable(dev); return r; diff --git a/drivers/gpu/drm/omapdrm/dss/omapdss.h b/drivers/gpu/drm/omapdrm/dss/omapdss.h index 1f698a95a94a..33e15cb77efa 100644 --- a/drivers/gpu/drm/omapdrm/dss/omapdss.h +++ b/drivers/gpu/drm/omapdrm/dss/omapdss.h @@ -432,7 +432,7 @@ struct omap_dss_device { const struct omap_dss_driver *driver; const struct omap_dss_device_ops *ops; unsigned long ops_flags; - unsigned long bus_flags; + u32 bus_flags; /* helper variable for driver suspend/resume */ bool activate_after_resume; diff --git a/drivers/gpu/drm/omapdrm/omap_encoder.c b/drivers/gpu/drm/omapdrm/omap_encoder.c index 452e625f6ce3..933ebc9f9faa 100644 --- a/drivers/gpu/drm/omapdrm/omap_encoder.c +++ b/drivers/gpu/drm/omapdrm/omap_encoder.c @@ -52,17 +52,44 @@ static const struct drm_encoder_funcs omap_encoder_funcs = { .destroy = omap_encoder_destroy, }; +static void omap_encoder_hdmi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct omap_encoder *omap_encoder = to_omap_encoder(encoder); + struct omap_dss_device *dssdev = omap_encoder->output; + struct drm_connector *connector; + bool hdmi_mode; + + hdmi_mode = false; + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + hdmi_mode = omap_connector_get_hdmi_mode(connector); + break; + } + } + + if (dssdev->ops->hdmi.set_hdmi_mode) + dssdev->ops->hdmi.set_hdmi_mode(dssdev, hdmi_mode); + + if (hdmi_mode && dssdev->ops->hdmi.set_infoframe) { + struct hdmi_avi_infoframe avi; + int r; + + r = drm_hdmi_avi_infoframe_from_display_mode(&avi, adjusted_mode, + false); + if (r == 0) + dssdev->ops->hdmi.set_infoframe(dssdev, &avi); + } +} + static void omap_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { - struct drm_device *dev = encoder->dev; struct omap_encoder *omap_encoder = to_omap_encoder(encoder); - struct drm_connector *connector; struct omap_dss_device *dssdev; struct videomode vm = { 0 }; - bool hdmi_mode; - int r; drm_display_mode_to_videomode(adjusted_mode, &vm); @@ -112,27 +139,8 @@ static void omap_encoder_mode_set(struct drm_encoder *encoder, } /* Set the HDMI mode and HDMI infoframe if applicable. */ - hdmi_mode = false; - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - if (connector->encoder == encoder) { - hdmi_mode = omap_connector_get_hdmi_mode(connector); - break; - } - } - - dssdev = omap_encoder->output; - - if (dssdev->ops->hdmi.set_hdmi_mode) - dssdev->ops->hdmi.set_hdmi_mode(dssdev, hdmi_mode); - - if (hdmi_mode && dssdev->ops->hdmi.set_infoframe) { - struct hdmi_avi_infoframe avi; - - r = drm_hdmi_avi_infoframe_from_display_mode(&avi, adjusted_mode, - false); - if (r == 0) - dssdev->ops->hdmi.set_infoframe(dssdev, &avi); - } + if (omap_encoder->output->output_type == OMAP_DISPLAY_TYPE_HDMI) + omap_encoder_hdmi_mode_set(encoder, adjusted_mode); } static void omap_encoder_disable(struct drm_encoder *encoder) diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c index 13c8a662f9b4..ccb090f3ab30 100644 --- a/drivers/gpu/drm/qxl/qxl_drv.c +++ b/drivers/gpu/drm/qxl/qxl_drv.c @@ -250,14 +250,10 @@ static struct drm_driver qxl_driver = { #if defined(CONFIG_DEBUG_FS) .debugfs_init = qxl_debugfs_init, #endif - .prime_handle_to_fd = drm_gem_prime_handle_to_fd, - .prime_fd_to_handle = drm_gem_prime_fd_to_handle, .gem_prime_export = drm_gem_prime_export, .gem_prime_import = drm_gem_prime_import, .gem_prime_pin = qxl_gem_prime_pin, .gem_prime_unpin = qxl_gem_prime_unpin, - .gem_prime_get_sg_table = qxl_gem_prime_get_sg_table, - .gem_prime_import_sg_table = qxl_gem_prime_import_sg_table, .gem_prime_vmap = qxl_gem_prime_vmap, .gem_prime_vunmap = qxl_gem_prime_vunmap, .gem_prime_mmap = qxl_gem_prime_mmap, diff --git a/drivers/gpu/drm/qxl/qxl_ioctl.c b/drivers/gpu/drm/qxl/qxl_ioctl.c index 6e828158bcb0..d410e2925162 100644 --- a/drivers/gpu/drm/qxl/qxl_ioctl.c +++ b/drivers/gpu/drm/qxl/qxl_ioctl.c @@ -163,8 +163,7 @@ static int qxl_process_single_command(struct qxl_device *qdev, if (cmd->command_size > PAGE_SIZE - sizeof(union qxl_release_info)) return -EINVAL; - if (!access_ok(VERIFY_READ, - u64_to_user_ptr(cmd->command), + if (!access_ok(u64_to_user_ptr(cmd->command), cmd->command_size)) return -EFAULT; diff --git a/drivers/gpu/drm/qxl/qxl_prime.c b/drivers/gpu/drm/qxl/qxl_prime.c index a55dece118b2..df65d3c1a7b8 100644 --- a/drivers/gpu/drm/qxl/qxl_prime.c +++ b/drivers/gpu/drm/qxl/qxl_prime.c @@ -38,20 +38,6 @@ void qxl_gem_prime_unpin(struct drm_gem_object *obj) WARN_ONCE(1, "not implemented"); } -struct sg_table *qxl_gem_prime_get_sg_table(struct drm_gem_object *obj) -{ - WARN_ONCE(1, "not implemented"); - return ERR_PTR(-ENOSYS); -} - -struct drm_gem_object *qxl_gem_prime_import_sg_table( - struct drm_device *dev, struct dma_buf_attachment *attach, - struct sg_table *table) -{ - WARN_ONCE(1, "not implemented"); - return ERR_PTR(-ENOSYS); -} - void *qxl_gem_prime_vmap(struct drm_gem_object *obj) { WARN_ONCE(1, "not implemented"); diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c index 0a693fede05e..30f85f0130cb 100644 --- a/drivers/gpu/drm/qxl/qxl_release.c +++ b/drivers/gpu/drm/qxl/qxl_release.c @@ -217,7 +217,7 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo) qxl_bo_ref(bo); entry->tv.bo = &bo->tbo; - entry->tv.shared = false; + entry->tv.num_shared = 0; list_add_tail(&entry->tv.head, &release->bos); return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 1ae31dbc61c6..f43305329939 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -178,7 +178,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) } p->relocs[i].tv.bo = &p->relocs[i].robj->tbo; - p->relocs[i].tv.shared = !r->write_domain; + p->relocs[i].tv.num_shared = !r->write_domain; radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head, priority); @@ -253,7 +253,7 @@ static int radeon_cs_sync_rings(struct radeon_cs_parser *p) resv = reloc->robj->tbo.resv; r = radeon_sync_resv(p->rdev, &p->ib.sync, resv, - reloc->tv.shared); + reloc->tv.num_shared); if (r) return r; } diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index 27d8e7dd2d06..44617dec8183 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -552,7 +552,7 @@ static void radeon_gem_va_update_vm(struct radeon_device *rdev, INIT_LIST_HEAD(&list); tv.bo = &bo_va->bo->tbo; - tv.shared = true; + tv.num_shared = 1; list_add(&tv.head, &list); vm_bos = radeon_vm_get_bos(rdev, bo_va->vm, &list); diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c index f8b35df44c60..b3019505065a 100644 --- a/drivers/gpu/drm/radeon/radeon_mn.c +++ b/drivers/gpu/drm/radeon/radeon_mn.c @@ -119,40 +119,38 @@ static void radeon_mn_release(struct mmu_notifier *mn, * unmap them by move them into system domain again. */ static int radeon_mn_invalidate_range_start(struct mmu_notifier *mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end, - bool blockable) + const struct mmu_notifier_range *range) { struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn); struct ttm_operation_ctx ctx = { false, false }; struct interval_tree_node *it; + unsigned long end; int ret = 0; /* notification is exclusive, but interval is inclusive */ - end -= 1; + end = range->end - 1; /* TODO we should be able to split locking for interval tree and * the tear down. */ - if (blockable) + if (range->blockable) mutex_lock(&rmn->lock); else if (!mutex_trylock(&rmn->lock)) return -EAGAIN; - it = interval_tree_iter_first(&rmn->objects, start, end); + it = interval_tree_iter_first(&rmn->objects, range->start, end); while (it) { struct radeon_mn_node *node; struct radeon_bo *bo; long r; - if (!blockable) { + if (!range->blockable) { ret = -EAGAIN; goto out_unlock; } node = container_of(it, struct radeon_mn_node, it); - it = interval_tree_iter_next(it, start, end); + it = interval_tree_iter_next(it, range->start, end); list_for_each_entry(bo, &node->bos, mn_list) { diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index a3d2ca07a058..0d374211661c 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -142,7 +142,7 @@ struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, list[0].preferred_domains = RADEON_GEM_DOMAIN_VRAM; list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM; list[0].tv.bo = &vm->page_directory->tbo; - list[0].tv.shared = true; + list[0].tv.num_shared = 1; list[0].tiling_flags = 0; list_add(&list[0].tv.head, head); @@ -154,7 +154,7 @@ struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, list[idx].preferred_domains = RADEON_GEM_DOMAIN_VRAM; list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM; list[idx].tv.bo = &list[idx].robj->tbo; - list[idx].tv.shared = true; + list[idx].tv.num_shared = 1; list[idx].tiling_flags = 0; list_add(&list[idx++].tv.head, head); } diff --git a/drivers/gpu/drm/rcar-du/rcar_du_group.c b/drivers/gpu/drm/rcar-du/rcar_du_group.c index d85f0a1c1581..cebf313c6e1f 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_group.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_group.c @@ -202,10 +202,25 @@ void rcar_du_group_put(struct rcar_du_group *rgrp) static void __rcar_du_group_start_stop(struct rcar_du_group *rgrp, bool start) { - struct rcar_du_crtc *rcrtc = &rgrp->dev->crtcs[rgrp->index * 2]; + struct rcar_du_device *rcdu = rgrp->dev; + + /* + * Group start/stop is controlled by the DRES and DEN bits of DSYSR0 + * for the first group and DSYSR2 for the second group. On most DU + * instances, this maps to the first CRTC of the group, and we can just + * use rcar_du_crtc_dsysr_clr_set() to access the correct DSYSR. On + * M3-N, however, DU2 doesn't exist, but DSYSR2 does. We thus need to + * access the register directly using group read/write. + */ + if (rcdu->info->channels_mask & BIT(rgrp->index * 2)) { + struct rcar_du_crtc *rcrtc = &rgrp->dev->crtcs[rgrp->index * 2]; - rcar_du_crtc_dsysr_clr_set(rcrtc, DSYSR_DRES | DSYSR_DEN, - start ? DSYSR_DEN : DSYSR_DRES); + rcar_du_crtc_dsysr_clr_set(rcrtc, DSYSR_DRES | DSYSR_DEN, + start ? DSYSR_DEN : DSYSR_DRES); + } else { + rcar_du_group_write(rgrp, DSYSR, + start ? DSYSR_DEN : DSYSR_DRES); + } } void rcar_du_group_start_stop(struct rcar_du_group *rgrp, bool start) diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c index 37f9a3b651ab..be6c2573039a 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c @@ -448,11 +448,6 @@ static int rockchip_drm_platform_remove(struct platform_device *pdev) return 0; } -static void rockchip_drm_platform_shutdown(struct platform_device *pdev) -{ - rockchip_drm_platform_remove(pdev); -} - static const struct of_device_id rockchip_drm_dt_ids[] = { { .compatible = "rockchip,display-subsystem", }, { /* sentinel */ }, @@ -462,7 +457,6 @@ MODULE_DEVICE_TABLE(of, rockchip_drm_dt_ids); static struct platform_driver rockchip_drm_platform_driver = { .probe = rockchip_drm_platform_probe, .remove = rockchip_drm_platform_remove, - .shutdown = rockchip_drm_platform_shutdown, .driver = { .name = "rockchip-drm", .of_match_table = rockchip_drm_dt_ids, diff --git a/drivers/gpu/drm/rockchip/rockchip_rgb.c b/drivers/gpu/drm/rockchip/rockchip_rgb.c index 96ac1458a59c..37f93022a106 100644 --- a/drivers/gpu/drm/rockchip/rockchip_rgb.c +++ b/drivers/gpu/drm/rockchip/rockchip_rgb.c @@ -113,8 +113,10 @@ struct rockchip_rgb *rockchip_rgb_init(struct device *dev, child_count++; ret = drm_of_find_panel_or_bridge(dev->of_node, 0, endpoint_id, &panel, &bridge); - if (!ret) + if (!ret) { + of_node_put(endpoint); break; + } } of_node_put(port); diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 9d4cd196037a..dbb69063b3d5 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -211,6 +211,62 @@ void drm_sched_fault(struct drm_gpu_scheduler *sched) } EXPORT_SYMBOL(drm_sched_fault); +/** + * drm_sched_suspend_timeout - Suspend scheduler job timeout + * + * @sched: scheduler instance for which to suspend the timeout + * + * Suspend the delayed work timeout for the scheduler. This is done by + * modifying the delayed work timeout to an arbitrary large value, + * MAX_SCHEDULE_TIMEOUT in this case. Note that this function can be + * called from an IRQ context. + * + * Returns the timeout remaining + * + */ +unsigned long drm_sched_suspend_timeout(struct drm_gpu_scheduler *sched) +{ + unsigned long sched_timeout, now = jiffies; + + sched_timeout = sched->work_tdr.timer.expires; + + /* + * Modify the timeout to an arbitrarily large value. This also prevents + * the timeout to be restarted when new submissions arrive + */ + if (mod_delayed_work(system_wq, &sched->work_tdr, MAX_SCHEDULE_TIMEOUT) + && time_after(sched_timeout, now)) + return sched_timeout - now; + else + return sched->timeout; +} +EXPORT_SYMBOL(drm_sched_suspend_timeout); + +/** + * drm_sched_resume_timeout - Resume scheduler job timeout + * + * @sched: scheduler instance for which to resume the timeout + * @remaining: remaining timeout + * + * Resume the delayed work timeout for the scheduler. Note that + * this function can be called from an IRQ context. + */ +void drm_sched_resume_timeout(struct drm_gpu_scheduler *sched, + unsigned long remaining) +{ + unsigned long flags; + + spin_lock_irqsave(&sched->job_list_lock, flags); + + if (list_empty(&sched->ring_mirror_list)) + cancel_delayed_work(&sched->work_tdr); + else + mod_delayed_work(system_wq, &sched->work_tdr, remaining); + + spin_unlock_irqrestore(&sched->job_list_lock, flags); +} +EXPORT_SYMBOL(drm_sched_resume_timeout); + /* job_finish is called after hw fence signaled */ static void drm_sched_job_finish(struct work_struct *work) @@ -218,6 +274,7 @@ static void drm_sched_job_finish(struct work_struct *work) struct drm_sched_job *s_job = container_of(work, struct drm_sched_job, finish_work); struct drm_gpu_scheduler *sched = s_job->sched; + unsigned long flags; /* * Canceling the timeout without removing our job from the ring mirror @@ -228,12 +285,12 @@ static void drm_sched_job_finish(struct work_struct *work) */ cancel_delayed_work_sync(&sched->work_tdr); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); /* remove job from ring_mirror_list */ list_del_init(&s_job->node); /* queue TDR for next job */ drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); sched->ops->free_job(s_job); } @@ -249,20 +306,22 @@ static void drm_sched_job_finish_cb(struct dma_fence *f, static void drm_sched_job_begin(struct drm_sched_job *s_job) { struct drm_gpu_scheduler *sched = s_job->sched; + unsigned long flags; dma_fence_add_callback(&s_job->s_fence->finished, &s_job->finish_cb, drm_sched_job_finish_cb); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_add_tail(&s_job->node, &sched->ring_mirror_list); drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } static void drm_sched_job_timedout(struct work_struct *work) { struct drm_gpu_scheduler *sched; struct drm_sched_job *job; + unsigned long flags; sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work); job = list_first_entry_or_null(&sched->ring_mirror_list, @@ -271,9 +330,9 @@ static void drm_sched_job_timedout(struct work_struct *work) if (job) job->sched->ops->timedout_job(job); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } /** @@ -287,9 +346,10 @@ void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_jo { struct drm_sched_job *s_job; struct drm_sched_entity *entity, *tmp; + unsigned long flags; int i; - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_reverse(s_job, &sched->ring_mirror_list, node) { if (s_job->s_fence->parent && dma_fence_remove_callback(s_job->s_fence->parent, @@ -299,7 +359,7 @@ void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_jo atomic_dec(&sched->hw_rq_count); } } - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); if (bad && bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) { atomic_inc(&bad->karma); @@ -337,9 +397,10 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) { struct drm_sched_job *s_job, *tmp; bool found_guilty = false; + unsigned long flags; int r; - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { struct drm_sched_fence *s_fence = s_job->s_fence; struct dma_fence *fence; @@ -353,7 +414,7 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) if (found_guilty && s_job->s_fence->scheduled.context == guilty_context) dma_fence_set_error(&s_fence->finished, -ECANCELED); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); fence = sched->ops->run_job(s_job); atomic_inc(&sched->hw_rq_count); @@ -372,10 +433,10 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) drm_sched_expel_job_unlocked(s_job); drm_sched_process_job(NULL, &s_fence->cb); } - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); } drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } EXPORT_SYMBOL(drm_sched_job_recovery); @@ -612,7 +673,7 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, long timeout, const char *name) { - int i; + int i, ret; sched->ops = ops; sched->hw_submission_limit = hw_submission; sched->name = name; @@ -633,8 +694,10 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, /* Each scheduler will run on a seperate kernel thread */ sched->thread = kthread_run(drm_sched_main, sched, sched->name); if (IS_ERR(sched->thread)) { + ret = PTR_ERR(sched->thread); + sched->thread = NULL; DRM_ERROR("Failed to create scheduler for %s.\n", name); - return PTR_ERR(sched->thread); + return ret; } sched->ready = true; diff --git a/drivers/gpu/drm/selftests/test-drm_damage_helper.c b/drivers/gpu/drm/selftests/test-drm_damage_helper.c index a2f753205a3e..9d2bcdf8bc29 100644 --- a/drivers/gpu/drm/selftests/test-drm_damage_helper.c +++ b/drivers/gpu/drm/selftests/test-drm_damage_helper.c @@ -53,7 +53,7 @@ static bool check_damage_clip(struct drm_plane_state *state, struct drm_rect *r, int src_y2 = (state->src.y2 >> 16) + !!(state->src.y2 & 0xFFFF); if (x1 >= x2 || y1 >= y2) { - pr_err("Cannot have damage clip with no dimention.\n"); + pr_err("Cannot have damage clip with no dimension.\n"); return false; } diff --git a/drivers/gpu/drm/sun4i/sun4i_backend.c b/drivers/gpu/drm/sun4i/sun4i_backend.c index 9e9255ee59cd..a021bab11a4f 100644 --- a/drivers/gpu/drm/sun4i/sun4i_backend.c +++ b/drivers/gpu/drm/sun4i/sun4i_backend.c @@ -786,17 +786,18 @@ static struct sun4i_frontend *sun4i_backend_find_frontend(struct sun4i_drv *drv, remote = of_graph_get_remote_port_parent(ep); if (!remote) continue; + of_node_put(remote); /* does this node match any registered engines? */ list_for_each_entry(frontend, &drv->frontend_list, list) { if (remote == frontend->node) { - of_node_put(remote); of_node_put(port); + of_node_put(ep); return frontend; } } } - + of_node_put(port); return ERR_PTR(-EINVAL); } diff --git a/drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c b/drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c index 061d2e0d9011..416da5376701 100644 --- a/drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c +++ b/drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c @@ -92,6 +92,8 @@ static void sun4i_hdmi_disable(struct drm_encoder *encoder) val = readl(hdmi->base + SUN4I_HDMI_VID_CTRL_REG); val &= ~SUN4I_HDMI_VID_CTRL_ENABLE; writel(val, hdmi->base + SUN4I_HDMI_VID_CTRL_REG); + + clk_disable_unprepare(hdmi->tmds_clk); } static void sun4i_hdmi_enable(struct drm_encoder *encoder) @@ -102,6 +104,8 @@ static void sun4i_hdmi_enable(struct drm_encoder *encoder) DRM_DEBUG_DRIVER("Enabling the HDMI Output\n"); + clk_prepare_enable(hdmi->tmds_clk); + sun4i_hdmi_setup_avi_infoframes(hdmi, mode); val |= SUN4I_HDMI_PKT_CTRL_TYPE(0, SUN4I_HDMI_PKT_AVI); val |= SUN4I_HDMI_PKT_CTRL_TYPE(1, SUN4I_HDMI_PKT_END); diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c index f80e82e16475..607a6ea17ecc 100644 --- a/drivers/gpu/drm/tegra/dc.c +++ b/drivers/gpu/drm/tegra/dc.c @@ -1978,6 +1978,23 @@ static irqreturn_t tegra_dc_irq(int irq, void *data) return IRQ_HANDLED; } +static bool tegra_dc_has_window_groups(struct tegra_dc *dc) +{ + unsigned int i; + + if (!dc->soc->wgrps) + return true; + + for (i = 0; i < dc->soc->num_wgrps; i++) { + const struct tegra_windowgroup_soc *wgrp = &dc->soc->wgrps[i]; + + if (wgrp->dc == dc->pipe && wgrp->num_windows > 0) + return true; + } + + return false; +} + static int tegra_dc_init(struct host1x_client *client) { struct drm_device *drm = dev_get_drvdata(client->parent); @@ -1993,22 +2010,8 @@ static int tegra_dc_init(struct host1x_client *client) * assign a primary plane to them, which in turn will cause KMS to * crash. */ - if (dc->soc->wgrps) { - bool has_wgrps = false; - unsigned int i; - - for (i = 0; i < dc->soc->num_wgrps; i++) { - const struct tegra_windowgroup_soc *wgrp = &dc->soc->wgrps[i]; - - if (wgrp->dc == dc->pipe && wgrp->num_windows > 0) { - has_wgrps = true; - break; - } - } - - if (!has_wgrps) - return 0; - } + if (!tegra_dc_has_window_groups(dc)) + return 0; dc->syncpt = host1x_syncpt_request(client, flags); if (!dc->syncpt) @@ -2094,6 +2097,9 @@ static int tegra_dc_exit(struct host1x_client *client) struct tegra_dc *dc = host1x_client_to_dc(client); int err; + if (!tegra_dc_has_window_groups(dc)) + return 0; + devm_free_irq(dc->dev, dc->irq, dc); err = tegra_dc_rgb_exit(dc); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 65ea4988b332..4b70ce664c41 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1274,6 +1274,7 @@ static const struct of_device_id host1x_drm_subdevs[] = { { .compatible = "nvidia,tegra194-display", }, { .compatible = "nvidia,tegra194-dc", }, { .compatible = "nvidia,tegra194-sor", }, + { .compatible = "nvidia,tegra194-vic", }, { /* sentinel */ } }; diff --git a/drivers/gpu/drm/tegra/falcon.c b/drivers/gpu/drm/tegra/falcon.c index f685e72949d1..352d05feabb0 100644 --- a/drivers/gpu/drm/tegra/falcon.c +++ b/drivers/gpu/drm/tegra/falcon.c @@ -141,9 +141,9 @@ int falcon_load_firmware(struct falcon *falcon) /* allocate iova space for the firmware */ falcon->firmware.vaddr = falcon->ops->alloc(falcon, firmware->size, &falcon->firmware.paddr); - if (!falcon->firmware.vaddr) { - dev_err(falcon->dev, "dma memory mapping failed\n"); - return -ENOMEM; + if (IS_ERR(falcon->firmware.vaddr)) { + dev_err(falcon->dev, "DMA memory mapping failed\n"); + return PTR_ERR(falcon->firmware.vaddr); } /* copy firmware image into local area. this also ensures endianness */ @@ -197,11 +197,19 @@ void falcon_exit(struct falcon *falcon) int falcon_boot(struct falcon *falcon) { unsigned long offset; + u32 value; int err; if (!falcon->firmware.vaddr) return -EINVAL; + err = readl_poll_timeout(falcon->regs + FALCON_DMACTL, value, + (value & (FALCON_DMACTL_IMEM_SCRUBBING | + FALCON_DMACTL_DMEM_SCRUBBING)) == 0, + 10, 10000); + if (err < 0) + return err; + falcon_writel(falcon, 0, FALCON_DMACTL); /* setup the address of the binary data so Falcon can access it later */ diff --git a/drivers/gpu/drm/tegra/hub.c b/drivers/gpu/drm/tegra/hub.c index 6112d9042979..922a48d5a483 100644 --- a/drivers/gpu/drm/tegra/hub.c +++ b/drivers/gpu/drm/tegra/hub.c @@ -742,7 +742,9 @@ static const struct host1x_client_ops tegra_display_hub_ops = { static int tegra_display_hub_probe(struct platform_device *pdev) { + struct device_node *child = NULL; struct tegra_display_hub *hub; + struct clk *clk; unsigned int i; int err; @@ -801,6 +803,34 @@ static int tegra_display_hub_probe(struct platform_device *pdev) return err; } + hub->num_heads = of_get_child_count(pdev->dev.of_node); + + hub->clk_heads = devm_kcalloc(&pdev->dev, hub->num_heads, sizeof(clk), + GFP_KERNEL); + if (!hub->clk_heads) + return -ENOMEM; + + for (i = 0; i < hub->num_heads; i++) { + child = of_get_next_child(pdev->dev.of_node, child); + if (!child) { + dev_err(&pdev->dev, "failed to find node for head %u\n", + i); + return -ENODEV; + } + + clk = devm_get_clk_from_child(&pdev->dev, child, "dc"); + if (IS_ERR(clk)) { + dev_err(&pdev->dev, "failed to get clock for head %u\n", + i); + of_node_put(child); + return PTR_ERR(clk); + } + + hub->clk_heads[i] = clk; + } + + of_node_put(child); + /* XXX: enable clock across reset? */ err = reset_control_assert(hub->rst); if (err < 0) @@ -840,12 +870,16 @@ static int tegra_display_hub_remove(struct platform_device *pdev) static int __maybe_unused tegra_display_hub_suspend(struct device *dev) { struct tegra_display_hub *hub = dev_get_drvdata(dev); + unsigned int i = hub->num_heads; int err; err = reset_control_assert(hub->rst); if (err < 0) return err; + while (i--) + clk_disable_unprepare(hub->clk_heads[i]); + clk_disable_unprepare(hub->clk_hub); clk_disable_unprepare(hub->clk_dsc); clk_disable_unprepare(hub->clk_disp); @@ -856,6 +890,7 @@ static int __maybe_unused tegra_display_hub_suspend(struct device *dev) static int __maybe_unused tegra_display_hub_resume(struct device *dev) { struct tegra_display_hub *hub = dev_get_drvdata(dev); + unsigned int i; int err; err = clk_prepare_enable(hub->clk_disp); @@ -870,13 +905,22 @@ static int __maybe_unused tegra_display_hub_resume(struct device *dev) if (err < 0) goto disable_dsc; + for (i = 0; i < hub->num_heads; i++) { + err = clk_prepare_enable(hub->clk_heads[i]); + if (err < 0) + goto disable_heads; + } + err = reset_control_deassert(hub->rst); if (err < 0) - goto disable_hub; + goto disable_heads; return 0; -disable_hub: +disable_heads: + while (i--) + clk_disable_unprepare(hub->clk_heads[i]); + clk_disable_unprepare(hub->clk_hub); disable_dsc: clk_disable_unprepare(hub->clk_dsc); diff --git a/drivers/gpu/drm/tegra/hub.h b/drivers/gpu/drm/tegra/hub.h index 6696a85fc1f2..479087c0705a 100644 --- a/drivers/gpu/drm/tegra/hub.h +++ b/drivers/gpu/drm/tegra/hub.h @@ -49,6 +49,9 @@ struct tegra_display_hub { struct clk *clk_hub; struct reset_control *rst; + unsigned int num_heads; + struct clk **clk_heads; + const struct tegra_display_hub_soc *soc; struct tegra_windowgroup *wgrps; }; diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c index b129da2e5afd..ef8692b7075a 100644 --- a/drivers/gpu/drm/tegra/sor.c +++ b/drivers/gpu/drm/tegra/sor.c @@ -19,6 +19,8 @@ #include <soc/tegra/pmc.h> +#include <sound/hda_verbs.h> + #include <drm/drm_atomic_helper.h> #include <drm/drm_dp_helper.h> #include <drm/drm_panel.h> @@ -29,14 +31,6 @@ #include "sor.h" #include "trace.h" -/* - * XXX Remove this after the commit adding it to soc/tegra/pmc.h has been - * merged. Having this around after the commit is merged should be safe since - * the preprocessor will effectively replace all occurrences and therefore no - * duplicate will be defined. - */ -#define TEGRA_IO_PAD_HDMI_DP0 26 - #define SOR_REKEY 0x38 struct tegra_sor_hdmi_settings { @@ -407,6 +401,7 @@ struct tegra_sor { const struct tegra_sor_soc *soc; void __iomem *regs; unsigned int index; + unsigned int irq; struct reset_control *rst; struct clk *clk_parent; @@ -433,6 +428,11 @@ struct tegra_sor { struct delayed_work scdc; bool scdc_enabled; + + struct { + unsigned int sample_rate; + unsigned int channels; + } audio; }; struct tegra_sor_state { @@ -2139,6 +2139,144 @@ tegra_sor_hdmi_setup_avi_infoframe(struct tegra_sor *sor, return 0; } +static void tegra_sor_write_eld(struct tegra_sor *sor) +{ + size_t length = drm_eld_size(sor->output.connector.eld), i; + + for (i = 0; i < length; i++) + tegra_sor_writel(sor, i << 8 | sor->output.connector.eld[i], + SOR_AUDIO_HDA_ELD_BUFWR); + + /* + * 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 < 96; i++) + tegra_sor_writel(sor, i << 8 | 0, SOR_AUDIO_HDA_ELD_BUFWR); +} + +static void tegra_sor_audio_prepare(struct tegra_sor *sor) +{ + u32 value; + + tegra_sor_write_eld(sor); + + value = SOR_AUDIO_HDA_PRESENSE_ELDV | SOR_AUDIO_HDA_PRESENSE_PD; + tegra_sor_writel(sor, value, SOR_AUDIO_HDA_PRESENSE); +} + +static void tegra_sor_audio_unprepare(struct tegra_sor *sor) +{ + tegra_sor_writel(sor, 0, SOR_AUDIO_HDA_PRESENSE); +} + +static int tegra_sor_hdmi_enable_audio_infoframe(struct tegra_sor *sor) +{ + u8 buffer[HDMI_INFOFRAME_SIZE(AUDIO)]; + struct hdmi_audio_infoframe frame; + u32 value; + int err; + + err = hdmi_audio_infoframe_init(&frame); + if (err < 0) { + dev_err(sor->dev, "failed to setup audio infoframe: %d\n", err); + return err; + } + + frame.channels = sor->audio.channels; + + err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer)); + if (err < 0) { + dev_err(sor->dev, "failed to pack audio infoframe: %d\n", err); + return err; + } + + tegra_sor_hdmi_write_infopack(sor, buffer, err); + + value = tegra_sor_readl(sor, SOR_HDMI_AUDIO_INFOFRAME_CTRL); + value |= INFOFRAME_CTRL_CHECKSUM_ENABLE; + value |= INFOFRAME_CTRL_ENABLE; + tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_INFOFRAME_CTRL); + + return 0; +} + +static void tegra_sor_hdmi_audio_enable(struct tegra_sor *sor) +{ + u32 value; + + value = tegra_sor_readl(sor, SOR_AUDIO_CNTRL); + + /* select HDA audio input */ + value &= ~SOR_AUDIO_CNTRL_SOURCE_SELECT(SOURCE_SELECT_MASK); + value |= SOR_AUDIO_CNTRL_SOURCE_SELECT(SOURCE_SELECT_HDA); + + /* inject null samples */ + if (sor->audio.channels != 2) + value &= ~SOR_AUDIO_CNTRL_INJECT_NULLSMPL; + else + value |= SOR_AUDIO_CNTRL_INJECT_NULLSMPL; + + value |= SOR_AUDIO_CNTRL_AFIFO_FLUSH; + + tegra_sor_writel(sor, value, SOR_AUDIO_CNTRL); + + /* enable advertising HBR capability */ + tegra_sor_writel(sor, SOR_AUDIO_SPARE_HBR_ENABLE, SOR_AUDIO_SPARE); + + tegra_sor_writel(sor, 0, SOR_HDMI_ACR_CTRL); + + value = SOR_HDMI_SPARE_ACR_PRIORITY_HIGH | + SOR_HDMI_SPARE_CTS_RESET(1) | + SOR_HDMI_SPARE_HW_CTS_ENABLE; + tegra_sor_writel(sor, value, SOR_HDMI_SPARE); + + /* enable HW CTS */ + value = SOR_HDMI_ACR_SUBPACK_LOW_SB1(0); + tegra_sor_writel(sor, value, SOR_HDMI_ACR_0441_SUBPACK_LOW); + + /* allow packet to be sent */ + value = SOR_HDMI_ACR_SUBPACK_HIGH_ENABLE; + tegra_sor_writel(sor, value, SOR_HDMI_ACR_0441_SUBPACK_HIGH); + + /* reset N counter and enable lookup */ + value = SOR_HDMI_AUDIO_N_RESET | SOR_HDMI_AUDIO_N_LOOKUP; + tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_N); + + value = (24000 * 4096) / (128 * sor->audio.sample_rate / 1000); + tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0320); + tegra_sor_writel(sor, 4096, SOR_AUDIO_NVAL_0320); + + tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_0441); + tegra_sor_writel(sor, 4704, SOR_AUDIO_NVAL_0441); + + tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_0882); + tegra_sor_writel(sor, 9408, SOR_AUDIO_NVAL_0882); + + tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_1764); + tegra_sor_writel(sor, 18816, SOR_AUDIO_NVAL_1764); + + value = (24000 * 6144) / (128 * sor->audio.sample_rate / 1000); + tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0480); + tegra_sor_writel(sor, 6144, SOR_AUDIO_NVAL_0480); + + value = (24000 * 12288) / (128 * sor->audio.sample_rate / 1000); + tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0960); + tegra_sor_writel(sor, 12288, SOR_AUDIO_NVAL_0960); + + value = (24000 * 24576) / (128 * sor->audio.sample_rate / 1000); + tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_1920); + tegra_sor_writel(sor, 24576, SOR_AUDIO_NVAL_1920); + + value = tegra_sor_readl(sor, SOR_HDMI_AUDIO_N); + value &= ~SOR_HDMI_AUDIO_N_RESET; + tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_N); + + tegra_sor_hdmi_enable_audio_infoframe(sor); +} + static void tegra_sor_hdmi_disable_audio_infoframe(struct tegra_sor *sor) { u32 value; @@ -2148,6 +2286,11 @@ static void tegra_sor_hdmi_disable_audio_infoframe(struct tegra_sor *sor) tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_INFOFRAME_CTRL); } +static void tegra_sor_hdmi_audio_disable(struct tegra_sor *sor) +{ + tegra_sor_hdmi_disable_audio_infoframe(sor); +} + static struct tegra_sor_hdmi_settings * tegra_sor_hdmi_find_settings(struct tegra_sor *sor, unsigned long frequency) { @@ -2243,6 +2386,7 @@ static void tegra_sor_hdmi_disable(struct drm_encoder *encoder) u32 value; int err; + tegra_sor_audio_unprepare(sor); tegra_sor_hdmi_scdc_stop(sor); err = tegra_sor_detach(sor); @@ -2651,6 +2795,7 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) dev_err(sor->dev, "failed to wakeup SOR: %d\n", err); tegra_sor_hdmi_scdc_start(sor); + tegra_sor_audio_prepare(sor); } static const struct drm_encoder_helper_funcs tegra_sor_hdmi_helpers = { @@ -2666,6 +2811,7 @@ static int tegra_sor_init(struct host1x_client *client) struct tegra_sor *sor = host1x_client_to_sor(client); int connector = DRM_MODE_CONNECTOR_Unknown; int encoder = DRM_MODE_ENCODER_NONE; + u32 value; int err; if (!sor->aux) { @@ -2759,6 +2905,15 @@ static int tegra_sor_init(struct host1x_client *client) if (err < 0) return err; + /* + * Enable and unmask the HDA codec SCRATCH0 register interrupt. This + * is used for interoperability between the HDA codec driver and the + * HDMI/DP driver. + */ + value = SOR_INT_CODEC_SCRATCH1 | SOR_INT_CODEC_SCRATCH0; + tegra_sor_writel(sor, value, SOR_INT_ENABLE); + tegra_sor_writel(sor, value, SOR_INT_MASK); + return 0; } @@ -2767,6 +2922,9 @@ static int tegra_sor_exit(struct host1x_client *client) struct tegra_sor *sor = host1x_client_to_sor(client); int err; + tegra_sor_writel(sor, 0, SOR_INT_MASK); + tegra_sor_writel(sor, 0, SOR_INT_ENABLE); + tegra_output_exit(&sor->output); if (sor->aux) { @@ -3037,6 +3195,54 @@ static int tegra_sor_parse_dt(struct tegra_sor *sor) return 0; } +static void tegra_hda_parse_format(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_sor_irq(int irq, void *data) +{ + struct tegra_sor *sor = data; + u32 value; + + value = tegra_sor_readl(sor, SOR_INT_STATUS); + tegra_sor_writel(sor, value, SOR_INT_STATUS); + + if (value & SOR_INT_CODEC_SCRATCH0) { + value = tegra_sor_readl(sor, SOR_AUDIO_HDA_CODEC_SCRATCH0); + + if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) { + unsigned int format, sample_rate, channels; + + format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK; + + tegra_hda_parse_format(format, &sample_rate, &channels); + + sor->audio.sample_rate = sample_rate; + sor->audio.channels = channels; + + tegra_sor_hdmi_audio_enable(sor); + } else { + tegra_sor_hdmi_audio_disable(sor); + } + } + + return IRQ_HANDLED; +} + static int tegra_sor_probe(struct platform_device *pdev) { struct device_node *np; @@ -3119,14 +3325,38 @@ static int tegra_sor_probe(struct platform_device *pdev) 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); + err = platform_get_irq(pdev, 0); + if (err < 0) { + dev_err(&pdev->dev, "failed to get IRQ: %d\n", err); + goto remove; + } + + sor->irq = err; + + err = devm_request_irq(sor->dev, sor->irq, tegra_sor_irq, 0, + dev_name(sor->dev), sor); + if (err < 0) { + dev_err(&pdev->dev, "failed to request IRQ: %d\n", err); + goto remove; + } + + sor->rst = devm_reset_control_get(&pdev->dev, "sor"); + if (IS_ERR(sor->rst)) { + err = PTR_ERR(sor->rst); + + if (err != -EBUSY || WARN_ON(!pdev->dev.pm_domain)) { dev_err(&pdev->dev, "failed to get reset control: %d\n", err); goto remove; } + + /* + * At this point, the reset control is most likely being used + * by the generic power domain implementation. With any luck + * the power domain will have taken care of resetting the SOR + * and we don't have to do anything. + */ + sor->rst = NULL; } sor->clk = devm_clk_get(&pdev->dev, NULL); diff --git a/drivers/gpu/drm/tegra/sor.h b/drivers/gpu/drm/tegra/sor.h index fb0854d92a27..13f7e68bec42 100644 --- a/drivers/gpu/drm/tegra/sor.h +++ b/drivers/gpu/drm/tegra/sor.h @@ -364,12 +364,28 @@ #define INFOFRAME_HEADER_VERSION(x) (((x) & 0xff) << 8) #define INFOFRAME_HEADER_TYPE(x) (((x) & 0xff) << 0) +#define SOR_HDMI_ACR_CTRL 0xb1 + +#define SOR_HDMI_ACR_0320_SUBPACK_LOW 0xb2 +#define SOR_HDMI_ACR_SUBPACK_LOW_SB1(x) (((x) & 0xff) << 24) + +#define SOR_HDMI_ACR_0320_SUBPACK_HIGH 0xb3 +#define SOR_HDMI_ACR_SUBPACK_HIGH_ENABLE (1 << 31) + +#define SOR_HDMI_ACR_0441_SUBPACK_LOW 0xb4 +#define SOR_HDMI_ACR_0441_SUBPACK_HIGH 0xb5 + #define SOR_HDMI_CTRL 0xc0 #define SOR_HDMI_CTRL_ENABLE (1 << 30) #define SOR_HDMI_CTRL_MAX_AC_PACKET(x) (((x) & 0x1f) << 16) #define SOR_HDMI_CTRL_AUDIO_LAYOUT (1 << 10) #define SOR_HDMI_CTRL_REKEY(x) (((x) & 0x7f) << 0) +#define SOR_HDMI_SPARE 0xcb +#define SOR_HDMI_SPARE_ACR_PRIORITY_HIGH (1 << 31) +#define SOR_HDMI_SPARE_CTS_RESET(x) (((x) & 0x7) << 16) +#define SOR_HDMI_SPARE_HW_CTS_ENABLE (1 << 0) + #define SOR_REFCLK 0xe6 #define SOR_REFCLK_DIV_INT(x) ((((x) >> 2) & 0xff) << 8) #define SOR_REFCLK_DIV_FRAC(x) (((x) & 0x3) << 6) @@ -378,10 +394,62 @@ #define SOR_INPUT_CONTROL_ARM_VIDEO_RANGE_LIMITED (1 << 1) #define SOR_INPUT_CONTROL_HDMI_SRC_SELECT(x) (((x) & 0x1) << 0) +#define SOR_AUDIO_CNTRL 0xfc +#define SOR_AUDIO_CNTRL_INJECT_NULLSMPL (1 << 29) +#define SOR_AUDIO_CNTRL_SOURCE_SELECT(x) (((x) & 0x3) << 20) +#define SOURCE_SELECT_MASK 0x3 +#define SOURCE_SELECT_HDA 0x2 +#define SOURCE_SELECT_SPDIF 0x1 +#define SOURCE_SELECT_AUTO 0x0 +#define SOR_AUDIO_CNTRL_AFIFO_FLUSH (1 << 12) + +#define SOR_AUDIO_SPARE 0xfe +#define SOR_AUDIO_SPARE_HBR_ENABLE (1 << 27) + +#define SOR_AUDIO_NVAL_0320 0xff +#define SOR_AUDIO_NVAL_0441 0x100 +#define SOR_AUDIO_NVAL_0882 0x101 +#define SOR_AUDIO_NVAL_1764 0x102 +#define SOR_AUDIO_NVAL_0480 0x103 +#define SOR_AUDIO_NVAL_0960 0x104 +#define SOR_AUDIO_NVAL_1920 0x105 + +#define SOR_AUDIO_HDA_CODEC_SCRATCH0 0x10a +#define SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID (1 << 30) +#define SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK 0xffff + +#define SOR_AUDIO_HDA_ELD_BUFWR 0x10c +#define SOR_AUDIO_HDA_ELD_BUFWR_INDEX(x) (((x) & 0xff) << 8) +#define SOR_AUDIO_HDA_ELD_BUFWR_DATA(x) (((x) & 0xff) << 0) + +#define SOR_AUDIO_HDA_PRESENSE 0x10d +#define SOR_AUDIO_HDA_PRESENSE_ELDV (1 << 1) +#define SOR_AUDIO_HDA_PRESENSE_PD (1 << 0) + +#define SOR_AUDIO_AVAL_0320 0x10f +#define SOR_AUDIO_AVAL_0441 0x110 +#define SOR_AUDIO_AVAL_0882 0x111 +#define SOR_AUDIO_AVAL_1764 0x112 +#define SOR_AUDIO_AVAL_0480 0x113 +#define SOR_AUDIO_AVAL_0960 0x114 +#define SOR_AUDIO_AVAL_1920 0x115 + +#define SOR_INT_STATUS 0x11c +#define SOR_INT_CODEC_CP_REQUEST (1 << 2) +#define SOR_INT_CODEC_SCRATCH1 (1 << 1) +#define SOR_INT_CODEC_SCRATCH0 (1 << 0) + +#define SOR_INT_MASK 0x11d +#define SOR_INT_ENABLE 0x11e + #define SOR_HDMI_VSI_INFOFRAME_CTRL 0x123 #define SOR_HDMI_VSI_INFOFRAME_STATUS 0x124 #define SOR_HDMI_VSI_INFOFRAME_HEADER 0x125 +#define SOR_HDMI_AUDIO_N 0x13c +#define SOR_HDMI_AUDIO_N_LOOKUP (1 << 28) +#define SOR_HDMI_AUDIO_N_RESET (1 << 20) + #define SOR_HDMI2_CTRL 0x13e #define SOR_HDMI2_CTRL_CLOCK_MODE_DIV_BY_4 (1 << 1) #define SOR_HDMI2_CTRL_SCRAMBLE (1 << 0) diff --git a/drivers/gpu/drm/tegra/vic.c b/drivers/gpu/drm/tegra/vic.c index 9f657a63b0bb..d47983deb1cf 100644 --- a/drivers/gpu/drm/tegra/vic.c +++ b/drivers/gpu/drm/tegra/vic.c @@ -38,6 +38,7 @@ struct vic { struct iommu_domain *domain; struct device *dev; struct clk *clk; + struct reset_control *rst; /* Platform configuration */ const struct vic_config *config; @@ -56,13 +57,37 @@ static void vic_writel(struct vic *vic, u32 value, unsigned int offset) static int vic_runtime_resume(struct device *dev) { struct vic *vic = dev_get_drvdata(dev); + int err; + + err = clk_prepare_enable(vic->clk); + if (err < 0) + return err; + + usleep_range(10, 20); + + err = reset_control_deassert(vic->rst); + if (err < 0) + goto disable; + + usleep_range(10, 20); + + return 0; - return clk_prepare_enable(vic->clk); +disable: + clk_disable_unprepare(vic->clk); + return err; } static int vic_runtime_suspend(struct device *dev) { struct vic *vic = dev_get_drvdata(dev); + int err; + + err = reset_control_assert(vic->rst); + if (err < 0) + return err; + + usleep_range(2000, 4000); clk_disable_unprepare(vic->clk); @@ -282,10 +307,18 @@ static const struct vic_config vic_t186_config = { .version = 0x18, }; +#define NVIDIA_TEGRA_194_VIC_FIRMWARE "nvidia/tegra194/vic.bin" + +static const struct vic_config vic_t194_config = { + .firmware = NVIDIA_TEGRA_194_VIC_FIRMWARE, + .version = 0x19, +}; + static const struct of_device_id vic_match[] = { { .compatible = "nvidia,tegra124-vic", .data = &vic_t124_config }, { .compatible = "nvidia,tegra210-vic", .data = &vic_t210_config }, { .compatible = "nvidia,tegra186-vic", .data = &vic_t186_config }, + { .compatible = "nvidia,tegra194-vic", .data = &vic_t194_config }, { }, }; @@ -323,6 +356,14 @@ static int vic_probe(struct platform_device *pdev) return PTR_ERR(vic->clk); } + if (!dev->pm_domain) { + vic->rst = devm_reset_control_get(dev, "vic"); + if (IS_ERR(vic->rst)) { + dev_err(&pdev->dev, "failed to get reset\n"); + return PTR_ERR(vic->rst); + } + } + vic->falcon.dev = dev; vic->falcon.regs = vic->regs; vic->falcon.ops = &vic_falcon_ops; @@ -418,3 +459,6 @@ MODULE_FIRMWARE(NVIDIA_TEGRA_210_VIC_FIRMWARE); #if IS_ENABLED(CONFIG_ARCH_TEGRA_186_SOC) MODULE_FIRMWARE(NVIDIA_TEGRA_186_VIC_FIRMWARE); #endif +#if IS_ENABLED(CONFIG_ARCH_TEGRA_194_SOC) +MODULE_FIRMWARE(NVIDIA_TEGRA_194_VIC_FIRMWARE); +#endif diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index d87935bf8e30..0ec08394e17a 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -77,38 +77,39 @@ static inline int ttm_mem_type_from_place(const struct ttm_place *place, return 0; } -static void ttm_mem_type_debug(struct ttm_bo_device *bdev, int mem_type) +static void ttm_mem_type_debug(struct ttm_bo_device *bdev, struct drm_printer *p, + int mem_type) { struct ttm_mem_type_manager *man = &bdev->man[mem_type]; - struct drm_printer p = drm_debug_printer(TTM_PFX); - pr_err(" has_type: %d\n", man->has_type); - pr_err(" use_type: %d\n", man->use_type); - pr_err(" flags: 0x%08X\n", man->flags); - pr_err(" gpu_offset: 0x%08llX\n", man->gpu_offset); - pr_err(" size: %llu\n", man->size); - pr_err(" available_caching: 0x%08X\n", man->available_caching); - pr_err(" default_caching: 0x%08X\n", man->default_caching); + drm_printf(p, " has_type: %d\n", man->has_type); + drm_printf(p, " use_type: %d\n", man->use_type); + drm_printf(p, " flags: 0x%08X\n", man->flags); + drm_printf(p, " gpu_offset: 0x%08llX\n", man->gpu_offset); + drm_printf(p, " size: %llu\n", man->size); + drm_printf(p, " available_caching: 0x%08X\n", man->available_caching); + drm_printf(p, " default_caching: 0x%08X\n", man->default_caching); if (mem_type != TTM_PL_SYSTEM) - (*man->func->debug)(man, &p); + (*man->func->debug)(man, p); } static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo, struct ttm_placement *placement) { + struct drm_printer p = drm_debug_printer(TTM_PFX); int i, ret, mem_type; - pr_err("No space for %p (%lu pages, %luK, %luM)\n", - bo, bo->mem.num_pages, bo->mem.size >> 10, - bo->mem.size >> 20); + drm_printf(&p, "No space for %p (%lu pages, %luK, %luM)\n", + bo, bo->mem.num_pages, bo->mem.size >> 10, + bo->mem.size >> 20); for (i = 0; i < placement->num_placement; i++) { ret = ttm_mem_type_from_place(&placement->placement[i], &mem_type); if (ret) return; - pr_err(" placement[%d]=0x%08X (%d)\n", - i, placement->placement[i].flags, mem_type); - ttm_mem_type_debug(bo->bdev, mem_type); + drm_printf(&p, " placement[%d]=0x%08X (%d)\n", + i, placement->placement[i].flags, mem_type); + ttm_mem_type_debug(bo->bdev, &p, mem_type); } } diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c index efa005a1c1b7..93860346c426 100644 --- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c +++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c @@ -126,10 +126,11 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, } if (!ret) { - if (!entry->shared) + if (!entry->num_shared) continue; - ret = reservation_object_reserve_shared(bo->resv, 1); + ret = reservation_object_reserve_shared(bo->resv, + entry->num_shared); if (!ret) continue; } @@ -150,8 +151,9 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, } } - if (!ret && entry->shared) - ret = reservation_object_reserve_shared(bo->resv, 1); + if (!ret && entry->num_shared) + ret = reservation_object_reserve_shared(bo->resv, + entry->num_shared); if (unlikely(ret != 0)) { if (ret == -EINTR) @@ -199,7 +201,7 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket, list_for_each_entry(entry, list, head) { bo = entry->bo; - if (entry->shared) + if (entry->num_shared) reservation_object_add_shared_fence(bo->resv, fence); else reservation_object_add_excl_fence(bo->resv, fence); diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c index f7f32a885af7..2d1aaca49105 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.c +++ b/drivers/gpu/drm/virtio/virtgpu_drv.c @@ -127,14 +127,10 @@ static struct drm_driver driver = { #if defined(CONFIG_DEBUG_FS) .debugfs_init = virtio_gpu_debugfs_init, #endif - .prime_handle_to_fd = drm_gem_prime_handle_to_fd, - .prime_fd_to_handle = drm_gem_prime_fd_to_handle, .gem_prime_export = drm_gem_prime_export, .gem_prime_import = drm_gem_prime_import, .gem_prime_pin = virtgpu_gem_prime_pin, .gem_prime_unpin = virtgpu_gem_prime_unpin, - .gem_prime_get_sg_table = virtgpu_gem_prime_get_sg_table, - .gem_prime_import_sg_table = virtgpu_gem_prime_import_sg_table, .gem_prime_vmap = virtgpu_gem_prime_vmap, .gem_prime_vunmap = virtgpu_gem_prime_vunmap, .gem_prime_mmap = virtgpu_gem_prime_mmap, diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index 1deb41d42ea4..0c15000f926e 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -372,10 +372,6 @@ int virtio_gpu_object_wait(struct virtio_gpu_object *bo, bool no_wait); /* virtgpu_prime.c */ int virtgpu_gem_prime_pin(struct drm_gem_object *obj); void virtgpu_gem_prime_unpin(struct drm_gem_object *obj); -struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj); -struct drm_gem_object *virtgpu_gem_prime_import_sg_table( - struct drm_device *dev, struct dma_buf_attachment *attach, - struct sg_table *sgt); void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj); void virtgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr); int virtgpu_gem_prime_mmap(struct drm_gem_object *obj, diff --git a/drivers/gpu/drm/virtio/virtgpu_prime.c b/drivers/gpu/drm/virtio/virtgpu_prime.c index 86ce0ae93f59..c59ec34c80a5 100644 --- a/drivers/gpu/drm/virtio/virtgpu_prime.c +++ b/drivers/gpu/drm/virtio/virtgpu_prime.c @@ -39,20 +39,6 @@ void virtgpu_gem_prime_unpin(struct drm_gem_object *obj) WARN_ONCE(1, "not implemented"); } -struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj) -{ - WARN_ONCE(1, "not implemented"); - return ERR_PTR(-ENODEV); -} - -struct drm_gem_object *virtgpu_gem_prime_import_sg_table( - struct drm_device *dev, struct dma_buf_attachment *attach, - struct sg_table *table) -{ - WARN_ONCE(1, "not implemented"); - return ERR_PTR(-ENODEV); -} - void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj) { struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 9fd8b4e75a8c..25afb1d594e3 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -34,7 +34,7 @@ #include <drm/ttm/ttm_placement.h> #include <drm/ttm/ttm_bo_driver.h> #include <drm/ttm/ttm_module.h> -#include <linux/dma_remapping.h> +#include <linux/intel-iommu.h> #define VMWGFX_DRIVER_DESC "Linux drm driver for VMware graphics devices" #define VMWGFX_CHIP_SVGAII 0 @@ -49,6 +49,8 @@ #define VMWGFX_REPO "In Tree" +#define VMWGFX_VALIDATION_MEM_GRAN (16*PAGE_SIZE) + /** * Fully encoded drm commands. Might move to vmw_drm.h @@ -581,7 +583,7 @@ static int vmw_dma_select_mode(struct vmw_private *dev_priv) dev_priv->map_mode = vmw_dma_map_populate; - if (dma_ops->sync_single_for_cpu) + if (dma_ops && dma_ops->sync_single_for_cpu) dev_priv->map_mode = vmw_dma_alloc_coherent; #ifdef CONFIG_SWIOTLB if (swiotlb_nr_tbl() == 0) @@ -911,7 +913,7 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) spin_unlock(&dev_priv->cap_lock); } - + vmw_validation_mem_init_ttm(dev_priv, VMWGFX_VALIDATION_MEM_GRAN); ret = vmw_kms_init(dev_priv); if (unlikely(ret != 0)) goto out_no_kms; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index d7f6cb9331de..cd607ba9c2fe 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -593,6 +593,9 @@ struct vmw_private { struct vmw_cmdbuf_man *cman; DECLARE_BITMAP(irqthread_pending, VMW_IRQTHREAD_MAX); + + /* Validation memory reservation */ + struct vmw_validation_mem vvm; }; static inline struct vmw_surface *vmw_res_to_srf(struct vmw_resource *res) @@ -831,6 +834,8 @@ extern int vmw_fifo_flush(struct vmw_private *dev_priv, extern int vmw_mmap(struct file *filp, struct vm_area_struct *vma); +extern void vmw_validation_mem_init_ttm(struct vmw_private *dev_priv, + size_t gran); /** * TTM buffer object driver - vmwgfx_ttm_buffer.c */ diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index 260650bb5560..f2d13a72c05d 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -3835,6 +3835,8 @@ int vmw_execbuf_process(struct drm_file *file_priv, struct sync_file *sync_file = NULL; DECLARE_VAL_CONTEXT(val_ctx, &sw_context->res_ht, 1); + vmw_validation_set_val_mem(&val_ctx, &dev_priv->vvm); + if (flags & DRM_VMW_EXECBUF_FLAG_EXPORT_FENCE_FD) { out_fence_fd = get_unused_fd_flags(O_CLOEXEC); if (out_fence_fd < 0) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 8a029bade32a..3025bfc001a1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -85,7 +85,7 @@ static void vmw_resource_release(struct kref *kref) struct ttm_validate_buffer val_buf; val_buf.bo = bo; - val_buf.shared = false; + val_buf.num_shared = 0; res->func->unbind(res, false, &val_buf); } res->backup_dirty = false; @@ -462,7 +462,7 @@ vmw_resource_check_buffer(struct ww_acquire_ctx *ticket, INIT_LIST_HEAD(&val_list); val_buf->bo = ttm_bo_reference(&res->backup->base); - val_buf->shared = false; + val_buf->num_shared = 0; list_add_tail(&val_buf->head, &val_list); ret = ttm_eu_reserve_buffers(ticket, &val_list, interruptible, NULL); if (unlikely(ret != 0)) @@ -565,7 +565,7 @@ static int vmw_resource_do_evict(struct ww_acquire_ctx *ticket, BUG_ON(!func->may_evict); val_buf.bo = NULL; - val_buf.shared = false; + val_buf.num_shared = 0; ret = vmw_resource_check_buffer(ticket, res, interruptible, &val_buf); if (unlikely(ret != 0)) return ret; @@ -614,7 +614,7 @@ int vmw_resource_validate(struct vmw_resource *res, bool intr) return 0; val_buf.bo = NULL; - val_buf.shared = false; + val_buf.num_shared = 0; if (res->backup) val_buf.bo = &res->backup->base; do { @@ -685,7 +685,7 @@ void vmw_resource_unbind_list(struct vmw_buffer_object *vbo) struct vmw_resource *res, *next; struct ttm_validate_buffer val_buf = { .bo = &vbo->base, - .shared = false + .num_shared = 0 }; lockdep_assert_held(&vbo->base.resv->lock.base); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c index 154eb09aa91e..e6d75e377dd8 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c @@ -42,3 +42,39 @@ int vmw_mmap(struct file *filp, struct vm_area_struct *vma) dev_priv = vmw_priv(file_priv->minor->dev); return ttm_bo_mmap(filp, vma, &dev_priv->bdev); } + +/* struct vmw_validation_mem callback */ +static int vmw_vmt_reserve(struct vmw_validation_mem *m, size_t size) +{ + static struct ttm_operation_ctx ctx = {.interruptible = false, + .no_wait_gpu = false}; + struct vmw_private *dev_priv = container_of(m, struct vmw_private, vvm); + + return ttm_mem_global_alloc(vmw_mem_glob(dev_priv), size, &ctx); +} + +/* struct vmw_validation_mem callback */ +static void vmw_vmt_unreserve(struct vmw_validation_mem *m, size_t size) +{ + struct vmw_private *dev_priv = container_of(m, struct vmw_private, vvm); + + return ttm_mem_global_free(vmw_mem_glob(dev_priv), size); +} + +/** + * vmw_validation_mem_init_ttm - Interface the validation memory tracker + * to ttm. + * @dev_priv: Pointer to struct vmw_private. The reason we choose a vmw private + * rather than a struct vmw_validation_mem is to make sure assumption in the + * callbacks that struct vmw_private derives from struct vmw_validation_mem + * holds true. + * @gran: The recommended allocation granularity + */ +void vmw_validation_mem_init_ttm(struct vmw_private *dev_priv, size_t gran) +{ + struct vmw_validation_mem *vvm = &dev_priv->vvm; + + vvm->reserve_mem = vmw_vmt_reserve; + vvm->unreserve_mem = vmw_vmt_unreserve; + vvm->gran = gran; +} diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c index 184025fa938e..b3f547fc5d3d 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c @@ -104,11 +104,25 @@ void *vmw_validation_mem_alloc(struct vmw_validation_context *ctx, return NULL; if (ctx->mem_size_left < size) { - struct page *page = alloc_page(GFP_KERNEL | __GFP_ZERO); + struct page *page; + if (ctx->vm && ctx->vm_size_left < PAGE_SIZE) { + int ret = ctx->vm->reserve_mem(ctx->vm, ctx->vm->gran); + + if (ret) + return NULL; + + ctx->vm_size_left += ctx->vm->gran; + ctx->total_mem += ctx->vm->gran; + } + + page = alloc_page(GFP_KERNEL | __GFP_ZERO); if (!page) return NULL; + if (ctx->vm) + ctx->vm_size_left -= PAGE_SIZE; + list_add_tail(&page->lru, &ctx->page_list); ctx->page_address = page_address(page); ctx->mem_size_left = PAGE_SIZE; @@ -138,6 +152,11 @@ static void vmw_validation_mem_free(struct vmw_validation_context *ctx) } ctx->mem_size_left = 0; + if (ctx->vm && ctx->total_mem) { + ctx->vm->unreserve_mem(ctx->vm, ctx->total_mem); + ctx->total_mem = 0; + ctx->vm_size_left = 0; + } } /** @@ -266,7 +285,7 @@ int vmw_validation_add_bo(struct vmw_validation_context *ctx, val_buf->bo = ttm_bo_get_unless_zero(&vbo->base); if (!val_buf->bo) return -ESRCH; - val_buf->shared = false; + val_buf->num_shared = 0; list_add_tail(&val_buf->head, &ctx->bo_list); bo_node->as_mob = as_mob; bo_node->cpu_blit = cpu_blit; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h index b57e3292c386..3b396fea40d7 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h @@ -34,6 +34,21 @@ #include <drm/ttm/ttm_execbuf_util.h> /** + * struct vmw_validation_mem - Custom interface to provide memory reservations + * for the validation code. + * @reserve_mem: Callback to reserve memory + * @unreserve_mem: Callback to unreserve memory + * @gran: Reservation granularity. Contains a hint how much memory should + * be reserved in each call to @reserve_mem(). A slow implementation may want + * reservation to be done in large batches. + */ +struct vmw_validation_mem { + int (*reserve_mem)(struct vmw_validation_mem *m, size_t size); + void (*unreserve_mem)(struct vmw_validation_mem *m, size_t size); + size_t gran; +}; + +/** * struct vmw_validation_context - Per command submission validation context * @ht: Hash table used to find resource- or buffer object duplicates * @resource_list: List head for resource validation metadata @@ -47,6 +62,10 @@ * buffer objects * @mem_size_left: Free memory left in the last page in @page_list * @page_address: Kernel virtual address of the last page in @page_list + * @vm: A pointer to the memory reservation interface or NULL if no + * memory reservation is needed. + * @vm_size_left: Amount of reserved memory that so far has not been allocated. + * @total_mem: Amount of reserved memory. */ struct vmw_validation_context { struct drm_open_hash *ht; @@ -59,6 +78,9 @@ struct vmw_validation_context { unsigned int merge_dups; unsigned int mem_size_left; u8 *page_address; + struct vmw_validation_mem *vm; + size_t vm_size_left; + size_t total_mem; }; struct vmw_buffer_object; @@ -102,6 +124,21 @@ vmw_validation_has_bos(struct vmw_validation_context *ctx) } /** + * vmw_validation_set_val_mem - Register a validation mem object for + * validation memory reservation + * @ctx: The validation context + * @vm: Pointer to a struct vmw_validation_mem + * + * Must be set before the first attempt to allocate validation memory. + */ +static inline void +vmw_validation_set_val_mem(struct vmw_validation_context *ctx, + struct vmw_validation_mem *vm) +{ + ctx->vm = vm; +} + +/** * vmw_validation_set_ht - Register a hash table for duplicate finding * @ctx: The validation context * @ht: Pointer to a hash table to use for duplicate finding diff --git a/drivers/gpu/drm/xen/Kconfig b/drivers/gpu/drm/xen/Kconfig index 4cca160782ab..f969d486855d 100644 --- a/drivers/gpu/drm/xen/Kconfig +++ b/drivers/gpu/drm/xen/Kconfig @@ -12,6 +12,7 @@ config DRM_XEN_FRONTEND select DRM_KMS_HELPER select VIDEOMODE_HELPERS select XEN_XENBUS_FRONTEND + select XEN_FRONT_PGDIR_SHBUF help Choose this option if you want to enable a para-virtualized frontend DRM/KMS driver for Xen guest OSes. diff --git a/drivers/gpu/drm/xen/Makefile b/drivers/gpu/drm/xen/Makefile index 712afff5ffc3..825905f67faa 100644 --- a/drivers/gpu/drm/xen/Makefile +++ b/drivers/gpu/drm/xen/Makefile @@ -4,7 +4,6 @@ drm_xen_front-objs := xen_drm_front.o \ xen_drm_front_kms.o \ xen_drm_front_conn.o \ xen_drm_front_evtchnl.o \ - xen_drm_front_shbuf.o \ xen_drm_front_cfg.o \ xen_drm_front_gem.o diff --git a/drivers/gpu/drm/xen/xen_drm_front.c b/drivers/gpu/drm/xen/xen_drm_front.c index 6b6d5ab82ec3..4d3d36fc3a5d 100644 --- a/drivers/gpu/drm/xen/xen_drm_front.c +++ b/drivers/gpu/drm/xen/xen_drm_front.c @@ -19,6 +19,7 @@ #include <xen/xen.h> #include <xen/xenbus.h> +#include <xen/xen-front-pgdir-shbuf.h> #include <xen/interface/io/displif.h> #include "xen_drm_front.h" @@ -26,28 +27,20 @@ #include "xen_drm_front_evtchnl.h" #include "xen_drm_front_gem.h" #include "xen_drm_front_kms.h" -#include "xen_drm_front_shbuf.h" struct xen_drm_front_dbuf { struct list_head list; u64 dbuf_cookie; u64 fb_cookie; - struct xen_drm_front_shbuf *shbuf; + + struct xen_front_pgdir_shbuf shbuf; }; -static int dbuf_add_to_list(struct xen_drm_front_info *front_info, - struct xen_drm_front_shbuf *shbuf, u64 dbuf_cookie) +static void dbuf_add_to_list(struct xen_drm_front_info *front_info, + struct xen_drm_front_dbuf *dbuf, u64 dbuf_cookie) { - struct xen_drm_front_dbuf *dbuf; - - dbuf = kzalloc(sizeof(*dbuf), GFP_KERNEL); - if (!dbuf) - return -ENOMEM; - dbuf->dbuf_cookie = dbuf_cookie; - dbuf->shbuf = shbuf; list_add(&dbuf->list, &front_info->dbuf_list); - return 0; } static struct xen_drm_front_dbuf *dbuf_get(struct list_head *dbuf_list, @@ -62,15 +55,6 @@ static struct xen_drm_front_dbuf *dbuf_get(struct list_head *dbuf_list, return NULL; } -static void dbuf_flush_fb(struct list_head *dbuf_list, u64 fb_cookie) -{ - struct xen_drm_front_dbuf *buf, *q; - - list_for_each_entry_safe(buf, q, dbuf_list, list) - if (buf->fb_cookie == fb_cookie) - xen_drm_front_shbuf_flush(buf->shbuf); -} - static void dbuf_free(struct list_head *dbuf_list, u64 dbuf_cookie) { struct xen_drm_front_dbuf *buf, *q; @@ -78,8 +62,8 @@ static void dbuf_free(struct list_head *dbuf_list, u64 dbuf_cookie) list_for_each_entry_safe(buf, q, dbuf_list, list) if (buf->dbuf_cookie == dbuf_cookie) { list_del(&buf->list); - xen_drm_front_shbuf_unmap(buf->shbuf); - xen_drm_front_shbuf_free(buf->shbuf); + xen_front_pgdir_shbuf_unmap(&buf->shbuf); + xen_front_pgdir_shbuf_free(&buf->shbuf); kfree(buf); break; } @@ -91,8 +75,8 @@ static void dbuf_free_all(struct list_head *dbuf_list) list_for_each_entry_safe(buf, q, dbuf_list, list) { list_del(&buf->list); - xen_drm_front_shbuf_unmap(buf->shbuf); - xen_drm_front_shbuf_free(buf->shbuf); + xen_front_pgdir_shbuf_unmap(&buf->shbuf); + xen_front_pgdir_shbuf_free(&buf->shbuf); kfree(buf); } } @@ -171,9 +155,9 @@ int xen_drm_front_dbuf_create(struct xen_drm_front_info *front_info, u32 bpp, u64 size, struct page **pages) { struct xen_drm_front_evtchnl *evtchnl; - struct xen_drm_front_shbuf *shbuf; + struct xen_drm_front_dbuf *dbuf; struct xendispl_req *req; - struct xen_drm_front_shbuf_cfg buf_cfg; + struct xen_front_pgdir_shbuf_cfg buf_cfg; unsigned long flags; int ret; @@ -181,28 +165,29 @@ int xen_drm_front_dbuf_create(struct xen_drm_front_info *front_info, if (unlikely(!evtchnl)) return -EIO; + dbuf = kzalloc(sizeof(*dbuf), GFP_KERNEL); + if (!dbuf) + return -ENOMEM; + + dbuf_add_to_list(front_info, dbuf, dbuf_cookie); + memset(&buf_cfg, 0, sizeof(buf_cfg)); buf_cfg.xb_dev = front_info->xb_dev; + buf_cfg.num_pages = DIV_ROUND_UP(size, PAGE_SIZE); buf_cfg.pages = pages; - buf_cfg.size = size; + buf_cfg.pgdir = &dbuf->shbuf; buf_cfg.be_alloc = front_info->cfg.be_alloc; - shbuf = xen_drm_front_shbuf_alloc(&buf_cfg); - if (IS_ERR(shbuf)) - return PTR_ERR(shbuf); - - ret = dbuf_add_to_list(front_info, shbuf, dbuf_cookie); - if (ret < 0) { - xen_drm_front_shbuf_free(shbuf); - return ret; - } + ret = xen_front_pgdir_shbuf_alloc(&buf_cfg); + if (ret < 0) + goto fail_shbuf_alloc; mutex_lock(&evtchnl->u.req.req_io_lock); spin_lock_irqsave(&front_info->io_lock, flags); req = be_prepare_req(evtchnl, XENDISPL_OP_DBUF_CREATE); req->op.dbuf_create.gref_directory = - xen_drm_front_shbuf_get_dir_start(shbuf); + xen_front_pgdir_shbuf_get_dir_start(&dbuf->shbuf); req->op.dbuf_create.buffer_sz = size; req->op.dbuf_create.dbuf_cookie = dbuf_cookie; req->op.dbuf_create.width = width; @@ -221,7 +206,7 @@ int xen_drm_front_dbuf_create(struct xen_drm_front_info *front_info, if (ret < 0) goto fail; - ret = xen_drm_front_shbuf_map(shbuf); + ret = xen_front_pgdir_shbuf_map(&dbuf->shbuf); if (ret < 0) goto fail; @@ -230,6 +215,7 @@ int xen_drm_front_dbuf_create(struct xen_drm_front_info *front_info, fail: mutex_unlock(&evtchnl->u.req.req_io_lock); +fail_shbuf_alloc: dbuf_free(&front_info->dbuf_list, dbuf_cookie); return ret; } @@ -358,7 +344,6 @@ int xen_drm_front_page_flip(struct xen_drm_front_info *front_info, if (unlikely(conn_idx >= front_info->num_evt_pairs)) return -EINVAL; - dbuf_flush_fb(&front_info->dbuf_list, fb_cookie); evtchnl = &front_info->evt_pairs[conn_idx].req; mutex_lock(&evtchnl->u.req.req_io_lock); diff --git a/drivers/gpu/drm/xen/xen_drm_front_gem.c b/drivers/gpu/drm/xen/xen_drm_front_gem.c index 47ff019d3aef..28bc501af450 100644 --- a/drivers/gpu/drm/xen/xen_drm_front_gem.c +++ b/drivers/gpu/drm/xen/xen_drm_front_gem.c @@ -22,7 +22,6 @@ #include <xen/balloon.h> #include "xen_drm_front.h" -#include "xen_drm_front_shbuf.h" struct xen_gem_object { struct drm_gem_object base; diff --git a/drivers/gpu/drm/xen/xen_drm_front_shbuf.c b/drivers/gpu/drm/xen/xen_drm_front_shbuf.c deleted file mode 100644 index d333b67cc1a0..000000000000 --- a/drivers/gpu/drm/xen/xen_drm_front_shbuf.c +++ /dev/null @@ -1,414 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 OR MIT - -/* - * Xen para-virtual DRM device - * - * Copyright (C) 2016-2018 EPAM Systems Inc. - * - * Author: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com> - */ - -#include <drm/drmP.h> - -#if defined(CONFIG_X86) -#include <drm/drm_cache.h> -#endif -#include <linux/errno.h> -#include <linux/mm.h> - -#include <asm/xen/hypervisor.h> -#include <xen/balloon.h> -#include <xen/xen.h> -#include <xen/xenbus.h> -#include <xen/interface/io/ring.h> -#include <xen/interface/io/displif.h> - -#include "xen_drm_front.h" -#include "xen_drm_front_shbuf.h" - -struct xen_drm_front_shbuf_ops { - /* - * Calculate number of grefs required to handle this buffer, - * e.g. if grefs are required for page directory only or the buffer - * pages as well. - */ - void (*calc_num_grefs)(struct xen_drm_front_shbuf *buf); - /* Fill page directory according to para-virtual display protocol. */ - void (*fill_page_dir)(struct xen_drm_front_shbuf *buf); - /* Claim grant references for the pages of the buffer. */ - int (*grant_refs_for_buffer)(struct xen_drm_front_shbuf *buf, - grant_ref_t *priv_gref_head, int gref_idx); - /* Map grant references of the buffer. */ - int (*map)(struct xen_drm_front_shbuf *buf); - /* Unmap grant references of the buffer. */ - int (*unmap)(struct xen_drm_front_shbuf *buf); -}; - -grant_ref_t xen_drm_front_shbuf_get_dir_start(struct xen_drm_front_shbuf *buf) -{ - if (!buf->grefs) - return GRANT_INVALID_REF; - - return buf->grefs[0]; -} - -int xen_drm_front_shbuf_map(struct xen_drm_front_shbuf *buf) -{ - if (buf->ops->map) - return buf->ops->map(buf); - - /* no need to map own grant references */ - return 0; -} - -int xen_drm_front_shbuf_unmap(struct xen_drm_front_shbuf *buf) -{ - if (buf->ops->unmap) - return buf->ops->unmap(buf); - - /* no need to unmap own grant references */ - return 0; -} - -void xen_drm_front_shbuf_flush(struct xen_drm_front_shbuf *buf) -{ -#if defined(CONFIG_X86) - drm_clflush_pages(buf->pages, buf->num_pages); -#endif -} - -void xen_drm_front_shbuf_free(struct xen_drm_front_shbuf *buf) -{ - if (buf->grefs) { - int i; - - for (i = 0; i < buf->num_grefs; i++) - if (buf->grefs[i] != GRANT_INVALID_REF) - gnttab_end_foreign_access(buf->grefs[i], - 0, 0UL); - } - kfree(buf->grefs); - kfree(buf->directory); - kfree(buf); -} - -/* - * number of grefs a page can hold with respect to the - * struct xendispl_page_directory header - */ -#define XEN_DRM_NUM_GREFS_PER_PAGE ((PAGE_SIZE - \ - offsetof(struct xendispl_page_directory, gref)) / \ - sizeof(grant_ref_t)) - -static int get_num_pages_dir(struct xen_drm_front_shbuf *buf) -{ - /* number of pages the page directory consumes itself */ - return DIV_ROUND_UP(buf->num_pages, XEN_DRM_NUM_GREFS_PER_PAGE); -} - -static void backend_calc_num_grefs(struct xen_drm_front_shbuf *buf) -{ - /* only for pages the page directory consumes itself */ - buf->num_grefs = get_num_pages_dir(buf); -} - -static void guest_calc_num_grefs(struct xen_drm_front_shbuf *buf) -{ - /* - * number of pages the page directory consumes itself - * plus grefs for the buffer pages - */ - buf->num_grefs = get_num_pages_dir(buf) + buf->num_pages; -} - -#define xen_page_to_vaddr(page) \ - ((uintptr_t)pfn_to_kaddr(page_to_xen_pfn(page))) - -static int backend_unmap(struct xen_drm_front_shbuf *buf) -{ - struct gnttab_unmap_grant_ref *unmap_ops; - int i, ret; - - if (!buf->pages || !buf->backend_map_handles || !buf->grefs) - return 0; - - unmap_ops = kcalloc(buf->num_pages, sizeof(*unmap_ops), - GFP_KERNEL); - if (!unmap_ops) { - DRM_ERROR("Failed to get memory while unmapping\n"); - return -ENOMEM; - } - - for (i = 0; i < buf->num_pages; i++) { - phys_addr_t addr; - - addr = xen_page_to_vaddr(buf->pages[i]); - gnttab_set_unmap_op(&unmap_ops[i], addr, GNTMAP_host_map, - buf->backend_map_handles[i]); - } - - ret = gnttab_unmap_refs(unmap_ops, NULL, buf->pages, - buf->num_pages); - - for (i = 0; i < buf->num_pages; i++) { - if (unlikely(unmap_ops[i].status != GNTST_okay)) - DRM_ERROR("Failed to unmap page %d: %d\n", - i, unmap_ops[i].status); - } - - if (ret) - DRM_ERROR("Failed to unmap grant references, ret %d", ret); - - kfree(unmap_ops); - kfree(buf->backend_map_handles); - buf->backend_map_handles = NULL; - return ret; -} - -static int backend_map(struct xen_drm_front_shbuf *buf) -{ - struct gnttab_map_grant_ref *map_ops = NULL; - unsigned char *ptr; - int ret, cur_gref, cur_dir_page, cur_page, grefs_left; - - map_ops = kcalloc(buf->num_pages, sizeof(*map_ops), GFP_KERNEL); - if (!map_ops) - return -ENOMEM; - - buf->backend_map_handles = kcalloc(buf->num_pages, - sizeof(*buf->backend_map_handles), - GFP_KERNEL); - if (!buf->backend_map_handles) { - kfree(map_ops); - return -ENOMEM; - } - - /* - * read page directory to get grefs from the backend: for external - * buffer we only allocate buf->grefs for the page directory, - * so buf->num_grefs has number of pages in the page directory itself - */ - ptr = buf->directory; - grefs_left = buf->num_pages; - cur_page = 0; - for (cur_dir_page = 0; cur_dir_page < buf->num_grefs; cur_dir_page++) { - struct xendispl_page_directory *page_dir = - (struct xendispl_page_directory *)ptr; - int to_copy = XEN_DRM_NUM_GREFS_PER_PAGE; - - if (to_copy > grefs_left) - to_copy = grefs_left; - - for (cur_gref = 0; cur_gref < to_copy; cur_gref++) { - phys_addr_t addr; - - addr = xen_page_to_vaddr(buf->pages[cur_page]); - gnttab_set_map_op(&map_ops[cur_page], addr, - GNTMAP_host_map, - page_dir->gref[cur_gref], - buf->xb_dev->otherend_id); - cur_page++; - } - - grefs_left -= to_copy; - ptr += PAGE_SIZE; - } - ret = gnttab_map_refs(map_ops, NULL, buf->pages, buf->num_pages); - - /* save handles even if error, so we can unmap */ - for (cur_page = 0; cur_page < buf->num_pages; cur_page++) { - buf->backend_map_handles[cur_page] = map_ops[cur_page].handle; - if (unlikely(map_ops[cur_page].status != GNTST_okay)) - DRM_ERROR("Failed to map page %d: %d\n", - cur_page, map_ops[cur_page].status); - } - - if (ret) { - DRM_ERROR("Failed to map grant references, ret %d", ret); - backend_unmap(buf); - } - - kfree(map_ops); - return ret; -} - -static void backend_fill_page_dir(struct xen_drm_front_shbuf *buf) -{ - struct xendispl_page_directory *page_dir; - unsigned char *ptr; - int i, num_pages_dir; - - ptr = buf->directory; - num_pages_dir = get_num_pages_dir(buf); - - /* fill only grefs for the page directory itself */ - for (i = 0; i < num_pages_dir - 1; i++) { - page_dir = (struct xendispl_page_directory *)ptr; - - page_dir->gref_dir_next_page = buf->grefs[i + 1]; - ptr += PAGE_SIZE; - } - /* last page must say there is no more pages */ - page_dir = (struct xendispl_page_directory *)ptr; - page_dir->gref_dir_next_page = GRANT_INVALID_REF; -} - -static void guest_fill_page_dir(struct xen_drm_front_shbuf *buf) -{ - unsigned char *ptr; - int cur_gref, grefs_left, to_copy, i, num_pages_dir; - - ptr = buf->directory; - num_pages_dir = get_num_pages_dir(buf); - - /* - * while copying, skip grefs at start, they are for pages - * granted for the page directory itself - */ - cur_gref = num_pages_dir; - grefs_left = buf->num_pages; - for (i = 0; i < num_pages_dir; i++) { - struct xendispl_page_directory *page_dir = - (struct xendispl_page_directory *)ptr; - - if (grefs_left <= XEN_DRM_NUM_GREFS_PER_PAGE) { - to_copy = grefs_left; - page_dir->gref_dir_next_page = GRANT_INVALID_REF; - } else { - to_copy = XEN_DRM_NUM_GREFS_PER_PAGE; - page_dir->gref_dir_next_page = buf->grefs[i + 1]; - } - memcpy(&page_dir->gref, &buf->grefs[cur_gref], - to_copy * sizeof(grant_ref_t)); - ptr += PAGE_SIZE; - grefs_left -= to_copy; - cur_gref += to_copy; - } -} - -static int guest_grant_refs_for_buffer(struct xen_drm_front_shbuf *buf, - grant_ref_t *priv_gref_head, - int gref_idx) -{ - int i, cur_ref, otherend_id; - - otherend_id = buf->xb_dev->otherend_id; - for (i = 0; i < buf->num_pages; i++) { - cur_ref = gnttab_claim_grant_reference(priv_gref_head); - if (cur_ref < 0) - return cur_ref; - - gnttab_grant_foreign_access_ref(cur_ref, otherend_id, - xen_page_to_gfn(buf->pages[i]), - 0); - buf->grefs[gref_idx++] = cur_ref; - } - return 0; -} - -static int grant_references(struct xen_drm_front_shbuf *buf) -{ - grant_ref_t priv_gref_head; - int ret, i, j, cur_ref; - int otherend_id, num_pages_dir; - - ret = gnttab_alloc_grant_references(buf->num_grefs, &priv_gref_head); - if (ret < 0) { - DRM_ERROR("Cannot allocate grant references\n"); - return ret; - } - - otherend_id = buf->xb_dev->otherend_id; - j = 0; - num_pages_dir = get_num_pages_dir(buf); - for (i = 0; i < num_pages_dir; i++) { - unsigned long frame; - - cur_ref = gnttab_claim_grant_reference(&priv_gref_head); - if (cur_ref < 0) - return cur_ref; - - frame = xen_page_to_gfn(virt_to_page(buf->directory + - PAGE_SIZE * i)); - gnttab_grant_foreign_access_ref(cur_ref, otherend_id, frame, 0); - buf->grefs[j++] = cur_ref; - } - - if (buf->ops->grant_refs_for_buffer) { - ret = buf->ops->grant_refs_for_buffer(buf, &priv_gref_head, j); - if (ret) - return ret; - } - - gnttab_free_grant_references(priv_gref_head); - return 0; -} - -static int alloc_storage(struct xen_drm_front_shbuf *buf) -{ - buf->grefs = kcalloc(buf->num_grefs, sizeof(*buf->grefs), GFP_KERNEL); - if (!buf->grefs) - return -ENOMEM; - - buf->directory = kcalloc(get_num_pages_dir(buf), PAGE_SIZE, GFP_KERNEL); - if (!buf->directory) - return -ENOMEM; - - return 0; -} - -/* - * For be allocated buffers we don't need grant_refs_for_buffer as those - * grant references are allocated at backend side - */ -static const struct xen_drm_front_shbuf_ops backend_ops = { - .calc_num_grefs = backend_calc_num_grefs, - .fill_page_dir = backend_fill_page_dir, - .map = backend_map, - .unmap = backend_unmap -}; - -/* For locally granted references we do not need to map/unmap the references */ -static const struct xen_drm_front_shbuf_ops local_ops = { - .calc_num_grefs = guest_calc_num_grefs, - .fill_page_dir = guest_fill_page_dir, - .grant_refs_for_buffer = guest_grant_refs_for_buffer, -}; - -struct xen_drm_front_shbuf * -xen_drm_front_shbuf_alloc(struct xen_drm_front_shbuf_cfg *cfg) -{ - struct xen_drm_front_shbuf *buf; - int ret; - - buf = kzalloc(sizeof(*buf), GFP_KERNEL); - if (!buf) - return ERR_PTR(-ENOMEM); - - if (cfg->be_alloc) - buf->ops = &backend_ops; - else - buf->ops = &local_ops; - - buf->xb_dev = cfg->xb_dev; - buf->num_pages = DIV_ROUND_UP(cfg->size, PAGE_SIZE); - buf->pages = cfg->pages; - - buf->ops->calc_num_grefs(buf); - - ret = alloc_storage(buf); - if (ret) - goto fail; - - ret = grant_references(buf); - if (ret) - goto fail; - - buf->ops->fill_page_dir(buf); - - return buf; - -fail: - xen_drm_front_shbuf_free(buf); - return ERR_PTR(ret); -} diff --git a/drivers/gpu/drm/xen/xen_drm_front_shbuf.h b/drivers/gpu/drm/xen/xen_drm_front_shbuf.h deleted file mode 100644 index 7545c692539e..000000000000 --- a/drivers/gpu/drm/xen/xen_drm_front_shbuf.h +++ /dev/null @@ -1,64 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 OR MIT */ - -/* - * Xen para-virtual DRM device - * - * Copyright (C) 2016-2018 EPAM Systems Inc. - * - * Author: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com> - */ - -#ifndef __XEN_DRM_FRONT_SHBUF_H_ -#define __XEN_DRM_FRONT_SHBUF_H_ - -#include <linux/kernel.h> -#include <linux/scatterlist.h> - -#include <xen/grant_table.h> - -struct xen_drm_front_shbuf { - /* - * number of references granted for the backend use: - * - for allocated/imported dma-buf's this holds number of grant - * references for the page directory and pages of the buffer - * - for the buffer provided by the backend this holds number of - * grant references for the page directory as grant references for - * the buffer will be provided by the backend - */ - int num_grefs; - grant_ref_t *grefs; - unsigned char *directory; - - int num_pages; - struct page **pages; - - struct xenbus_device *xb_dev; - - /* these are the ops used internally depending on be_alloc mode */ - const struct xen_drm_front_shbuf_ops *ops; - - /* Xen map handles for the buffer allocated by the backend */ - grant_handle_t *backend_map_handles; -}; - -struct xen_drm_front_shbuf_cfg { - struct xenbus_device *xb_dev; - size_t size; - struct page **pages; - bool be_alloc; -}; - -struct xen_drm_front_shbuf * -xen_drm_front_shbuf_alloc(struct xen_drm_front_shbuf_cfg *cfg); - -grant_ref_t xen_drm_front_shbuf_get_dir_start(struct xen_drm_front_shbuf *buf); - -int xen_drm_front_shbuf_map(struct xen_drm_front_shbuf *buf); - -int xen_drm_front_shbuf_unmap(struct xen_drm_front_shbuf *buf); - -void xen_drm_front_shbuf_flush(struct xen_drm_front_shbuf *buf); - -void xen_drm_front_shbuf_free(struct xen_drm_front_shbuf *buf); - -#endif /* __XEN_DRM_FRONT_SHBUF_H_ */ |
