diff options
Diffstat (limited to 'drivers/gpu/drm')
102 files changed, 701 insertions, 414 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 2a0df4cabb99..6f5b4a0e0a34 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1290,6 +1290,7 @@ struct amdgpu_device { bool debug_disable_gpu_ring_reset; bool debug_vm_userptr; bool debug_disable_ce_logs; + bool debug_enable_ce_cs; /* Protection for the following isolation structure */ struct mutex enforce_isolation_mutex; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 83020963dfde..a2ca9acf8c4e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -2329,10 +2329,9 @@ void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_mem *mem) int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev, struct kfd_vm_fault_info *mem) { - if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) { + if (atomic_read_acquire(&adev->gmc.vm_fault_info_updated) == 1) { *mem = *adev->gmc.vm_fault_info; - mb(); /* make sure read happened */ - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); } return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c index ef996493115f..425a3e564360 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT /* * Copyright 2025 Advanced Micro Devices, Inc. * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h index bcb97d245673..353421807387 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.h @@ -1,4 +1,4 @@ -/* SPDX-License-Identifier: GPL-2.0 */ +/* SPDX-License-Identifier: MIT */ /* * Copyright 2025 Advanced Micro Devices, Inc. * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 9cd7741d2254..2f6a96af7fb1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -364,6 +364,12 @@ static int amdgpu_cs_p2_ib(struct amdgpu_cs_parser *p, if (p->uf_bo && ring->funcs->no_user_fence) return -EINVAL; + if (!p->adev->debug_enable_ce_cs && + chunk_ib->flags & AMDGPU_IB_FLAG_CE) { + dev_err_ratelimited(p->adev->dev, "CE CS is blocked, use debug=0x400 to override\n"); + return -EINVAL; + } + if (chunk_ib->ip_type == AMDGPU_HW_IP_GFX && chunk_ib->flags & AMDGPU_IB_FLAG_PREEMPT) { if (chunk_ib->flags & AMDGPU_IB_FLAG_CE) @@ -702,7 +708,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev, */ const s64 us_upper_bound = 200000; - if (!adev->mm_stats.log2_max_MBps) { + if ((!adev->mm_stats.log2_max_MBps) || !ttm_resource_manager_used(&adev->mman.vram_mgr.manager)) { *max_bytes = 0; *max_vis_bytes = 0; return; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 7a899fb4de29..3d032c4e2dce 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -1882,6 +1882,13 @@ static bool amdgpu_device_pcie_dynamic_switching_supported(struct amdgpu_device static bool amdgpu_device_aspm_support_quirk(struct amdgpu_device *adev) { + /* Enabling ASPM causes randoms hangs on Tahiti and Oland on Zen4. + * It's unclear if this is a platform-specific or GPU-specific issue. + * Disable ASPM on SI for the time being. + */ + if (adev->family == AMDGPU_FAMILY_SI) + return true; + #if IS_ENABLED(CONFIG_X86) struct cpuinfo_x86 *c = &cpu_data(0); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index 73401f0aeb34..dd7b2b796427 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -1033,7 +1033,9 @@ static uint8_t amdgpu_discovery_get_harvest_info(struct amdgpu_device *adev, /* Until a uniform way is figured, get mask based on hwid */ switch (hw_id) { case VCN_HWID: - harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; + /* VCN vs UVD+VCE */ + if (!amdgpu_ip_version(adev, VCE_HWIP, 0)) + harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; break; case DMU_HWID: if (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK) @@ -2565,7 +2567,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 0); @@ -2592,7 +2596,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 1); @@ -2619,8 +2625,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 1; + adev->sdma.sdma_mask = 1; adev->vcn.num_vcn_inst = 1; adev->gmc.num_umc = 2; + adev->gfx.xcc_mask = 1; if (adev->apu_flags & AMD_APU_IS_RAVEN2) { adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 2, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 2, 0); @@ -2665,7 +2673,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega20_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 0); @@ -2693,8 +2703,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); arct_reg_base_init(adev); adev->sdma.num_instances = 8; + adev->sdma.sdma_mask = 0xff; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 1); @@ -2726,8 +2738,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); aldebaran_reg_base_init(adev); adev->sdma.num_instances = 5; + adev->sdma.sdma_mask = 0x1f; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 4, 0); @@ -2762,6 +2776,8 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) } else { cyan_skillfish_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(2, 0, 3); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(2, 0, 3); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(5, 0, 1); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index bff25ef3e2d0..61268aa82df4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -144,7 +144,8 @@ enum AMDGPU_DEBUG_MASK { AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6), AMDGPU_DEBUG_SMU_POOL = BIT(7), AMDGPU_DEBUG_VM_USERPTR = BIT(8), - AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9) + AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9), + AMDGPU_DEBUG_ENABLE_CE_CS = BIT(10) }; unsigned int amdgpu_vram_limit = UINT_MAX; @@ -2289,6 +2290,11 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev) pr_info("debug: disable kernel logs of correctable errors\n"); adev->debug_disable_ce_logs = true; } + + if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_CE_CS) { + pr_info("debug: allowing command submission to CE engine\n"); + adev->debug_enable_ce_cs = true; + } } static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index fd8cca241da6..18a7829122d2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -758,11 +758,42 @@ void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring) * @fence: fence of the ring to signal * */ -void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *fence) +void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af) { - dma_fence_set_error(&fence->base, -ETIME); - amdgpu_fence_write(fence->ring, fence->seq); - amdgpu_fence_process(fence->ring); + struct dma_fence *unprocessed; + struct dma_fence __rcu **ptr; + struct amdgpu_fence *fence; + struct amdgpu_ring *ring = af->ring; + unsigned long flags; + u32 seq, last_seq; + + last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask; + seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask; + + /* mark all fences from the guilty context with an error */ + spin_lock_irqsave(&ring->fence_drv.lock, flags); + do { + last_seq++; + last_seq &= ring->fence_drv.num_fences_mask; + + ptr = &ring->fence_drv.fences[last_seq]; + rcu_read_lock(); + unprocessed = rcu_dereference(*ptr); + + if (unprocessed && !dma_fence_is_signaled_locked(unprocessed)) { + fence = container_of(unprocessed, struct amdgpu_fence, base); + + if (fence == af) + dma_fence_set_error(&fence->base, -ETIME); + else if (fence->context == af->context) + dma_fence_set_error(&fence->base, -ECANCELED); + } + rcu_read_unlock(); + } while (last_seq != seq); + spin_unlock_irqrestore(&ring->fence_drv.lock, flags); + /* signal the guilty fence */ + amdgpu_fence_write(ring, af->seq); + amdgpu_fence_process(ring); } void amdgpu_fence_save_wptr(struct dma_fence *fence) @@ -790,14 +821,19 @@ void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring, struct dma_fence *unprocessed; struct dma_fence __rcu **ptr; struct amdgpu_fence *fence; - u64 wptr, i, seqno; + u64 wptr; + u32 seq, last_seq; - seqno = amdgpu_fence_read(ring); + last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask; + seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask; wptr = ring->fence_drv.signalled_wptr; ring->ring_backup_entries_to_copy = 0; - for (i = seqno + 1; i <= ring->fence_drv.sync_seq; ++i) { - ptr = &ring->fence_drv.fences[i & ring->fence_drv.num_fences_mask]; + do { + last_seq++; + last_seq &= ring->fence_drv.num_fences_mask; + + ptr = &ring->fence_drv.fences[last_seq]; rcu_read_lock(); unprocessed = rcu_dereference(*ptr); @@ -813,7 +849,7 @@ void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring, wptr = fence->wptr; } rcu_read_unlock(); - } + } while (last_seq != seq); } /* diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index 6b7d66b6d4cc..63ee6ba6a931 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -371,7 +371,7 @@ static int amdgpu_debugfs_jpeg_sched_mask_set(void *data, u64 val) for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { ring = &adev->jpeg.inst[i].ring_dec[j]; - if (val & (BIT_ULL(1) << ((i * adev->jpeg.num_jpeg_rings) + j))) + if (val & (BIT_ULL((i * adev->jpeg.num_jpeg_rings) + j))) ring->sched.ready = true; else ring->sched.ready = false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index a9327472c651..b3e6b3fcdf2c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -758,7 +758,8 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) ui64 = atomic64_read(&adev->num_vram_cpu_page_faults); return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0; case AMDGPU_INFO_VRAM_USAGE: - ui64 = ttm_resource_manager_usage(&adev->mman.vram_mgr.manager); + ui64 = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) : 0; return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0; case AMDGPU_INFO_VIS_VRAM_USAGE: ui64 = amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr); @@ -804,8 +805,8 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) mem.vram.usable_heap_size = adev->gmc.real_vram_size - atomic64_read(&adev->vram_pin_size) - AMDGPU_VM_RESERVED_VRAM; - mem.vram.heap_usage = - ttm_resource_manager_usage(vram_man); + mem.vram.heap_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(vram_man) : 0; mem.vram.max_allocation = mem.vram.usable_heap_size * 3 / 4; mem.cpu_accessible_vram.total_heap_size = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 5bf9be073cdd..4883adcfbb4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -409,7 +409,7 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, return -EINVAL; /* Clear the doorbell array before detection */ - memset(adev->mes.hung_queue_db_array_cpu_addr, 0, + memset(adev->mes.hung_queue_db_array_cpu_addr, AMDGPU_MES_INVALID_DB_OFFSET, adev->mes.hung_queue_db_array_size * sizeof(u32)); input.queue_type = queue_type; input.detect_only = detect_only; @@ -420,12 +420,17 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, dev_err(adev->dev, "failed to detect and reset\n"); } else { *hung_db_num = 0; - for (i = 0; i < adev->mes.hung_queue_db_array_size; i++) { + for (i = 0; i < adev->mes.hung_queue_hqd_info_offset; i++) { if (db_array[i] != AMDGPU_MES_INVALID_DB_OFFSET) { hung_db_array[i] = db_array[i]; *hung_db_num += 1; } } + + /* + * TODO: return HQD info for MES scheduled user compute queue reset cases + * stored in hung_db_array hqd info offset to full array size + */ } return r; @@ -686,14 +691,11 @@ out: bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev) { uint32_t mes_rev = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK; - bool is_supported = false; - - if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && - amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && - mes_rev >= 0x63) - is_supported = true; - return is_supported; + return ((amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && + amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && + mes_rev >= 0x63) || + amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0)); } /* Fix me -- node_id is used to identify the correct MES instances in the future */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h index 6b506fc72f58..97c137c90f97 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h @@ -149,6 +149,7 @@ struct amdgpu_mes { void *resource_1_addr[AMDGPU_MAX_MES_PIPES]; int hung_queue_db_array_size; + int hung_queue_hqd_info_offset; struct amdgpu_bo *hung_queue_db_array_gpu_obj; uint64_t hung_queue_db_array_gpu_addr; void *hung_queue_db_array_cpu_addr; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 8f6ce948c684..5ec5c3ff22bb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -811,7 +811,7 @@ int amdgpu_ring_reset_helper_end(struct amdgpu_ring *ring, if (r) return r; - /* signal the fence of the bad job */ + /* signal the guilty fence and set an error on all fences from the context */ if (guilty_fence) amdgpu_fence_driver_guilty_force_completion(guilty_fence); /* Re-emit the non-guilty commands */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index b6b649179776..4b46e3c26ff3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -155,7 +155,7 @@ extern const struct drm_sched_backend_ops amdgpu_sched_ops; void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring); void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error); void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring); -void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *fence); +void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af); void amdgpu_fence_save_wptr(struct dma_fence *fence); int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 3328ab63376b..f96beb96c75c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -598,8 +598,8 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev) vf2pf_info->driver_cert = 0; vf2pf_info->os_info.all = 0; - vf2pf_info->fb_usage = - ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) >> 20; + vf2pf_info->fb_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) >> 20 : 0; vf2pf_info->fb_vis_usage = amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr) >> 20; vf2pf_info->fb_size = adev->gmc.real_vram_size >> 20; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c index 474bfe36c0c2..aa78c2ee9e21 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c @@ -322,6 +322,26 @@ static int vpe_early_init(struct amdgpu_ip_block *ip_block) return 0; } +static bool vpe_need_dpm0_at_power_down(struct amdgpu_device *adev) +{ + switch (amdgpu_ip_version(adev, VPE_HWIP, 0)) { + case IP_VERSION(6, 1, 1): + return adev->pm.fw_version < 0x0a640500; + default: + return false; + } +} + +static int vpe_get_dpm_level(struct amdgpu_device *adev) +{ + struct amdgpu_vpe *vpe = &adev->vpe; + + if (!adev->pm.dpm_enabled) + return 0; + + return RREG32(vpe_get_reg_offset(vpe, 0, vpe->regs.dpm_request_lv)); +} + static void vpe_idle_work_handler(struct work_struct *work) { struct amdgpu_device *adev = @@ -329,11 +349,17 @@ static void vpe_idle_work_handler(struct work_struct *work) unsigned int fences = 0; fences += amdgpu_fence_count_emitted(&adev->vpe.ring); + if (fences) + goto reschedule; - if (fences == 0) - amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE, AMD_PG_STATE_GATE); - else - schedule_delayed_work(&adev->vpe.idle_work, VPE_IDLE_TIMEOUT); + if (vpe_need_dpm0_at_power_down(adev) && vpe_get_dpm_level(adev) != 0) + goto reschedule; + + amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VPE, AMD_PG_STATE_GATE); + return; + +reschedule: + schedule_delayed_work(&adev->vpe.idle_work, VPE_IDLE_TIMEOUT); } static int vpe_common_init(struct amdgpu_vpe *vpe) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index a5adb2ed9b3c..9d934c07fa6b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -234,6 +234,9 @@ static umode_t amdgpu_vram_attrs_is_visible(struct kobject *kobj, !adev->gmc.vram_vendor) return 0; + if (!ttm_resource_manager_used(&adev->mman.vram_mgr.manager)) + return 0; + return attr->mode; } diff --git a/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c b/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c index 96616a865aac..ed1e25661706 100644 --- a/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c +++ b/drivers/gpu/drm/amd/amdgpu/cyan_skillfish_reg_init.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0 +// SPDX-License-Identifier: MIT /* * Copyright 2018 Advanced Micro Devices, Inc. * diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 66c47c466532..d61eb9f187c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -5862,8 +5862,6 @@ static void gfx_v11_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; - BUG_ON(ib->flags & AMDGPU_IB_FLAG_CE); - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); control |= ib->length_dw | (vmid << 24); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 710ec9c34e43..93fde0f9af87 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -4419,8 +4419,6 @@ static void gfx_v12_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; - BUG_ON(ib->flags & AMDGPU_IB_FLAG_CE); - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); control |= ib->length_dw | (vmid << 24); diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 93d7ccb7d013..0e5e54d0a9a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -1068,7 +1068,7 @@ static int gmc_v7_0_sw_init(struct amdgpu_ip_block *ip_block) GFP_KERNEL); if (!adev->gmc.vm_fault_info) return -ENOMEM; - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); return 0; } @@ -1290,7 +1290,7 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev, vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID); if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid) - && !atomic_read(&adev->gmc.vm_fault_info_updated)) { + && !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) { struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info; u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, @@ -1306,8 +1306,7 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev, info->prot_read = protections & 0x8 ? true : false; info->prot_write = protections & 0x10 ? true : false; info->prot_exec = protections & 0x20 ? true : false; - mb(); - atomic_set(&adev->gmc.vm_fault_info_updated, 1); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 1); } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index c5e2a2c41e06..e1509480dfc2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1183,7 +1183,7 @@ static int gmc_v8_0_sw_init(struct amdgpu_ip_block *ip_block) GFP_KERNEL); if (!adev->gmc.vm_fault_info) return -ENOMEM; - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); return 0; } @@ -1478,7 +1478,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID); if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid) - && !atomic_read(&adev->gmc.vm_fault_info_updated)) { + && !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) { struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info; u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, @@ -1494,8 +1494,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, info->prot_read = protections & 0x8 ? true : false; info->prot_write = protections & 0x10 ? true : false; info->prot_exec = protections & 0x20 ? true : false; - mb(); - atomic_set(&adev->gmc.vm_fault_info_updated, 1); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 1); } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c index 2db9b2c63693..1cd9eaeef38f 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c @@ -208,10 +208,10 @@ static int mes_userq_detect_and_reset(struct amdgpu_device *adev, struct amdgpu_userq_mgr *uqm, *tmp; unsigned int hung_db_num = 0; int queue_id, r, i; - u32 db_array[4]; + u32 db_array[8]; - if (db_array_size > 4) { - dev_err(adev->dev, "DB array size (%d vs 4) too small\n", + if (db_array_size > 8) { + dev_err(adev->dev, "DB array size (%d vs 8) too small\n", db_array_size); return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index e82188431f79..da575bb1377f 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -66,7 +66,8 @@ static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev); #define GFX_MES_DRAM_SIZE 0x80000 #define MES11_HW_RESOURCE_1_SIZE (128 * AMDGPU_GPU_PAGE_SIZE) -#define MES11_HUNG_DB_OFFSET_ARRAY_SIZE 4 +#define MES11_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset, [4:7] = hqd info */ +#define MES11_HUNG_HQD_INFO_OFFSET 4 static void mes_v11_0_ring_set_wptr(struct amdgpu_ring *ring) { @@ -1720,8 +1721,9 @@ static int mes_v11_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe, r; - adev->mes.hung_queue_db_array_size = - MES11_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_db_array_size = MES11_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_hqd_info_offset = MES11_HUNG_HQD_INFO_OFFSET; + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE) continue; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index aff06f06aeee..7f3512d9de07 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -47,7 +47,8 @@ static int mes_v12_0_kiq_hw_fini(struct amdgpu_device *adev); #define MES_EOP_SIZE 2048 -#define MES12_HUNG_DB_OFFSET_ARRAY_SIZE 4 +#define MES12_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset [4:7] hqd info */ +#define MES12_HUNG_HQD_INFO_OFFSET 4 static void mes_v12_0_ring_set_wptr(struct amdgpu_ring *ring) { @@ -228,7 +229,12 @@ static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes, pipe, x_pkt->header.opcode); r = amdgpu_fence_wait_polling(ring, seq, timeout); - if (r < 1 || !*status_ptr) { + + /* + * status_ptr[31:0] == 0 (fail) or status_ptr[63:0] == 1 (success). + * If status_ptr[31:0] == 0 then status_ptr[63:32] will have debug error information. + */ + if (r < 1 || !(lower_32_bits(*status_ptr))) { if (misc_op_str) dev_err(adev->dev, "MES(%d) failed to respond to msg=%s (%s)\n", @@ -1899,8 +1905,9 @@ static int mes_v12_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe, r; - adev->mes.hung_queue_db_array_size = - MES12_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_db_array_size = MES12_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_hqd_info_offset = MES12_HUNG_HQD_INFO_OFFSET; + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { r = amdgpu_mes_init_microcode(adev, pipe); if (r) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index 6c5c7c1bf5ed..6e7bc983fc0b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -1209,6 +1209,15 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, pr_debug_ratelimited("Evicting process pid %d queues\n", pdd->process->lead_thread->pid); + if (dqm->dev->kfd->shared_resources.enable_mes) { + pdd->last_evict_timestamp = get_jiffies_64(); + retval = suspend_all_queues_mes(dqm); + if (retval) { + dev_err(dev, "Suspending all queues failed"); + goto out; + } + } + /* Mark all queues as evicted. Deactivate all active queues on * the qpd. */ @@ -1221,23 +1230,27 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, decrement_queue_count(dqm, qpd, q); if (dqm->dev->kfd->shared_resources.enable_mes) { - int err; - - err = remove_queue_mes(dqm, q, qpd); - if (err) { + retval = remove_queue_mes(dqm, q, qpd); + if (retval) { dev_err(dev, "Failed to evict queue %d\n", q->properties.queue_id); - retval = err; + goto out; } } } - pdd->last_evict_timestamp = get_jiffies_64(); - if (!dqm->dev->kfd->shared_resources.enable_mes) + + if (!dqm->dev->kfd->shared_resources.enable_mes) { + pdd->last_evict_timestamp = get_jiffies_64(); retval = execute_queues_cpsch(dqm, qpd->is_debug ? KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES : KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); + } else { + retval = resume_all_queues_mes(dqm); + if (retval) + dev_err(dev, "Resuming all queues failed"); + } out: dqm_unlock(dqm); @@ -3098,61 +3111,17 @@ out: return ret; } -static int kfd_dqm_evict_pasid_mes(struct device_queue_manager *dqm, - struct qcm_process_device *qpd) -{ - struct device *dev = dqm->dev->adev->dev; - int ret = 0; - - /* Check if process is already evicted */ - dqm_lock(dqm); - if (qpd->evicted) { - /* Increment the evicted count to make sure the - * process stays evicted before its terminated. - */ - qpd->evicted++; - dqm_unlock(dqm); - goto out; - } - dqm_unlock(dqm); - - ret = suspend_all_queues_mes(dqm); - if (ret) { - dev_err(dev, "Suspending all queues failed"); - goto out; - } - - ret = dqm->ops.evict_process_queues(dqm, qpd); - if (ret) { - dev_err(dev, "Evicting process queues failed"); - goto out; - } - - ret = resume_all_queues_mes(dqm); - if (ret) - dev_err(dev, "Resuming all queues failed"); - -out: - return ret; -} - int kfd_evict_process_device(struct kfd_process_device *pdd) { struct device_queue_manager *dqm; struct kfd_process *p; - int ret = 0; p = pdd->process; dqm = pdd->dev->dqm; WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid); - if (dqm->dev->kfd->shared_resources.enable_mes) - ret = kfd_dqm_evict_pasid_mes(dqm, &pdd->qpd); - else - ret = dqm->ops.evict_process_queues(dqm, &pdd->qpd); - - return ret; + return dqm->ops.evict_process_queues(dqm, &pdd->qpd); } int reserve_debug_trap_vmid(struct device_queue_manager *dqm, 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 0d03e324d5b9..bfa3199591b6 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -551,13 +551,13 @@ static void schedule_dc_vmin_vmax(struct amdgpu_device *adev, struct dc_stream_state *stream, struct dc_crtc_timing_adjust *adjust) { - struct vupdate_offload_work *offload_work = kzalloc(sizeof(*offload_work), GFP_KERNEL); + struct vupdate_offload_work *offload_work = kzalloc(sizeof(*offload_work), GFP_NOWAIT); if (!offload_work) { drm_dbg_driver(adev_to_drm(adev), "Failed to allocate vupdate_offload_work\n"); return; } - struct dc_crtc_timing_adjust *adjust_copy = kzalloc(sizeof(*adjust_copy), GFP_KERNEL); + struct dc_crtc_timing_adjust *adjust_copy = kzalloc(sizeof(*adjust_copy), GFP_NOWAIT); if (!adjust_copy) { drm_dbg_driver(adev_to_drm(adev), "Failed to allocate adjust_copy\n"); kfree(offload_work); @@ -2085,8 +2085,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) dc_hardware_init(adev->dm.dc); - adev->dm.restore_backlight = true; - adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev); if (!adev->dm.hpd_rx_offload_wq) { drm_err(adev_to_drm(adev), "failed to create hpd rx offload workqueue.\n"); @@ -3442,7 +3440,6 @@ static int dm_resume(struct amdgpu_ip_block *ip_block) dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D0); dc_resume(dm->dc); - adev->dm.restore_backlight = true; amdgpu_dm_irq_resume_early(adev); @@ -9969,6 +9966,7 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, bool mode_set_reset_required = false; u32 i; struct dc_commit_streams_params params = {dc_state->streams, dc_state->stream_count}; + bool set_backlight_level = false; /* Disable writeback */ for_each_old_connector_in_state(state, connector, old_con_state, i) { @@ -10088,6 +10086,7 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, acrtc->hw_mode = new_crtc_state->mode; crtc->hwmode = new_crtc_state->mode; mode_set_reset_required = true; + set_backlight_level = true; } else if (modereset_required(new_crtc_state)) { drm_dbg_atomic(dev, "Atomic commit: RESET. crtc id %d:[%p]\n", @@ -10144,16 +10143,13 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, * to fix a flicker issue. * It will cause the dm->actual_brightness is not the current panel brightness * level. (the dm->brightness is the correct panel level) - * So we set the backlight level with dm->brightness value after initial - * set mode. Use restore_backlight flag to avoid setting backlight level - * for every subsequent mode set. + * So we set the backlight level with dm->brightness value after set mode */ - if (dm->restore_backlight) { + if (set_backlight_level) { for (i = 0; i < dm->num_of_edps; i++) { if (dm->backlight_dev[i]) amdgpu_dm_backlight_set_level(dm, i, dm->brightness[i]); } - dm->restore_backlight = false; } } 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 009f206226f0..db75e991ac7b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -631,13 +631,6 @@ struct amdgpu_display_manager { u32 actual_brightness[AMDGPU_DM_MAX_NUM_EDP]; /** - * @restore_backlight: - * - * Flag to indicate whether to restore backlight after modeset. - */ - bool restore_backlight; - - /** * @aux_hpd_discon_quirk: * * quirk for hpd discon while aux is on-going. diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c index 1ec9d03ad747..38f9ea313dcb 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c @@ -248,6 +248,8 @@ static void amdgpu_dm_crtc_vblank_control_worker(struct work_struct *work) struct vblank_control_work *vblank_work = container_of(work, struct vblank_control_work, work); struct amdgpu_display_manager *dm = vblank_work->dm; + struct amdgpu_device *adev = drm_to_adev(dm->ddev); + int r; mutex_lock(&dm->dc_lock); @@ -277,7 +279,16 @@ static void amdgpu_dm_crtc_vblank_control_worker(struct work_struct *work) if (dm->active_vblank_irq_count == 0) { dc_post_update_surfaces_to_stream(dm->dc); + + r = amdgpu_dpm_pause_power_profile(adev, true); + if (r) + dev_warn(adev->dev, "failed to set default power profile mode\n"); + dc_allow_idle_optimizations(dm->dc, true); + + r = amdgpu_dpm_pause_power_profile(adev, false); + if (r) + dev_warn(adev->dev, "failed to restore the power profile mode\n"); } mutex_unlock(&dm->dc_lock); @@ -297,8 +308,12 @@ static inline int amdgpu_dm_crtc_set_vblank(struct drm_crtc *crtc, bool enable) int irq_type; int rc = 0; - if (acrtc->otg_inst == -1) - goto skip; + if (enable && !acrtc->base.enabled) { + drm_dbg_vbl(crtc->dev, + "Reject vblank enable on unconfigured CRTC %d (enabled=%d)\n", + acrtc->crtc_id, acrtc->base.enabled); + return -EINVAL; + } irq_type = amdgpu_display_crtc_idx_to_irq_type(adev, acrtc->crtc_id); @@ -383,7 +398,7 @@ static inline int amdgpu_dm_crtc_set_vblank(struct drm_crtc *crtc, bool enable) return rc; } #endif -skip: + if (amdgpu_in_reset(adev)) return 0; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index fe100e4c9801..cc21337a182f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -83,6 +83,7 @@ static void apply_edid_quirks(struct drm_device *dev, struct edid *edid, struct edid_caps->panel_patch.remove_sink_ext_caps = true; break; case drm_edid_encode_panel_id('S', 'D', 'C', 0x4154): + case drm_edid_encode_panel_id('S', 'D', 'C', 0x4171): drm_dbg_driver(dev, "Disabling VSC on monitor with panel id %X\n", panel_id); edid_caps->panel_patch.disable_colorimetry = true; break; diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c index 09be2a90cc79..4f569cd8a5d6 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c @@ -578,9 +578,6 @@ static void dpp3_power_on_blnd_lut( dpp_base->ctx->dc->optimized_required = true; dpp_base->deferred_reg_writes.bits.disable_blnd_lut = true; } - } else { - REG_SET(CM_MEM_PWR_CTRL, 0, - BLNDGAM_MEM_PWR_FORCE, power_on == true ? 0 : 1); } } diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c index 7c276c319086..ce3d0b45fb4c 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c @@ -200,6 +200,9 @@ void dcn401_init_hw(struct dc *dc) */ struct dc_link *link = dc->links[i]; + if (link->ep_type != DISPLAY_ENDPOINT_PHY) + continue; + link->link_enc->funcs->hw_init(link->link_enc); /* Check for enabled DIG to identify enabled display */ diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h b/drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h index 41c76ba9ba56..62a39204fe0b 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h @@ -44,7 +44,13 @@ */ #define MAX_PIPES 6 #define MAX_PHANTOM_PIPES (MAX_PIPES / 2) -#define MAX_LINKS (MAX_PIPES * 2 +2) + +#define MAX_DPIA 6 +#define MAX_CONNECTOR 6 +#define MAX_VIRTUAL_LINKS 4 + +#define MAX_LINKS (MAX_DPIA + MAX_CONNECTOR + MAX_VIRTUAL_LINKS) + #define MAX_DIG_LINK_ENCODERS 7 #define MAX_DWB_PIPES 1 #define MAX_HPO_DP2_ENCODERS 4 diff --git a/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c b/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c index 9e33bf937a69..2676ae9f6fe8 100644 --- a/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c +++ b/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c @@ -78,6 +78,7 @@ static void dp_retrain_link_dp_test(struct dc_link *link, struct audio_output audio_output[MAX_PIPES]; struct dc_stream_state *streams_on_link[MAX_PIPES]; int num_streams_on_link = 0; + struct dc *dc = (struct dc *)link->dc; needs_divider_update = (link->dc->link_srv->dp_get_encoding_format(link_setting) != link->dc->link_srv->dp_get_encoding_format((const struct dc_link_settings *) &link->cur_link_settings)); @@ -150,7 +151,7 @@ static void dp_retrain_link_dp_test(struct dc_link *link, if (streams_on_link[i] && streams_on_link[i]->link && streams_on_link[i]->link == link) { stream_update.stream = streams_on_link[i]; stream_update.dpms_off = &dpms_off; - dc_update_planes_and_stream(state->clk_mgr->ctx->dc, NULL, 0, streams_on_link[i], &stream_update); + dc_update_planes_and_stream(dc, NULL, 0, streams_on_link[i], &stream_update); } } } diff --git a/drivers/gpu/drm/amd/include/amd_cper.h b/drivers/gpu/drm/amd/include/amd_cper.h index 086869264425..a252ee4c7874 100644 --- a/drivers/gpu/drm/amd/include/amd_cper.h +++ b/drivers/gpu/drm/amd/include/amd_cper.h @@ -1,4 +1,4 @@ -/* SPDX-License-Identifier: GPL-2.0 */ +/* SPDX-License-Identifier: MIT */ /* * Copyright 2025 Advanced Micro Devices, Inc. * diff --git a/drivers/gpu/drm/amd/include/ivsrcid/vcn/irqsrcs_vcn_5_0.h b/drivers/gpu/drm/amd/include/ivsrcid/vcn/irqsrcs_vcn_5_0.h index 64b553e7de1a..e7fdcee22a71 100644 --- a/drivers/gpu/drm/amd/include/ivsrcid/vcn/irqsrcs_vcn_5_0.h +++ b/drivers/gpu/drm/amd/include/ivsrcid/vcn/irqsrcs_vcn_5_0.h @@ -1,4 +1,4 @@ -/* SPDX-License-Identifier: GPL-2.0 */ +/* SPDX-License-Identifier: MIT */ /* * Copyright 2024 Advanced Micro Devices, Inc. All rights reserved. diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c index cf9932e68055..3a9522c17fee 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c @@ -3500,6 +3500,11 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, * for these GPUs to calculate bandwidth requirements. */ if (high_pixelclock_count) { + /* Work around flickering lines at the bottom edge + * of the screen when using a single 4K 60Hz monitor. + */ + disable_mclk_switching = true; + /* On Oland, we observe some flickering when two 4K 60Hz * displays are connected, possibly because voltage is too low. * Raise the voltage by requiring a higher SCLK. diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c index 8da882c51856..9b28c0728269 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c @@ -5444,8 +5444,7 @@ static int smu7_get_thermal_temperature_range(struct pp_hwmgr *hwmgr, thermal_data->max = table_info->cac_dtp_table->usSoftwareShutdownTemp * PP_TEMPERATURE_UNITS_PER_CENTIGRADES; else if (hwmgr->pp_table_version == PP_TABLE_V0) - thermal_data->max = data->thermal_temp_setting.temperature_shutdown * - PP_TEMPERATURE_UNITS_PER_CENTIGRADES; + thermal_data->max = data->thermal_temp_setting.temperature_shutdown; thermal_data->sw_ctf_threshold = thermal_data->max; diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/fiji_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/fiji_smumgr.c index d2dbd90bb427..0a876c840c79 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/fiji_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/fiji_smumgr.c @@ -2024,7 +2024,7 @@ static int fiji_init_smc_table(struct pp_hwmgr *hwmgr) table->VoltageResponseTime = 0; table->PhaseResponseTime = 0; table->MemoryThermThrottleEnable = 1; - table->PCIeBootLinkLevel = 0; /* 0:Gen1 1:Gen2 2:Gen3*/ + table->PCIeBootLinkLevel = (uint8_t) (data->dpm_table.pcie_speed_table.count); table->PCIeGenInterval = 1; table->VRConfig = 0; diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/iceland_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/iceland_smumgr.c index 1f50f1e74c48..aa3ae9b115c4 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/iceland_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/iceland_smumgr.c @@ -2028,7 +2028,7 @@ static int iceland_init_smc_table(struct pp_hwmgr *hwmgr) table->VoltageResponseTime = 0; table->PhaseResponseTime = 0; table->MemoryThermThrottleEnable = 1; - table->PCIeBootLinkLevel = 0; + table->PCIeBootLinkLevel = (uint8_t) (data->dpm_table.pcie_speed_table.count); table->PCIeGenInterval = 1; result = iceland_populate_smc_svi2_config(hwmgr, table); diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c index f532f7c69259..a8961a8f5c42 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c @@ -969,7 +969,7 @@ int smu_cmn_update_table(struct smu_context *smu, table_index); uint32_t table_size; int ret = 0; - if (!table_data || table_id >= SMU_TABLE_COUNT || table_id < 0) + if (!table_data || table_index >= SMU_TABLE_COUNT || table_id < 0) return -EINVAL; table_size = smu_table->tables[table_index].size; diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h index c15aef014f69..d41bd876167c 100644 --- a/drivers/gpu/drm/ast/ast_drv.h +++ b/drivers/gpu/drm/ast/ast_drv.h @@ -282,13 +282,13 @@ static inline void __ast_write8_i(void __iomem *addr, u32 reg, u8 index, u8 val) __ast_write8(addr, reg + 1, val); } -static inline void __ast_write8_i_masked(void __iomem *addr, u32 reg, u8 index, u8 read_mask, +static inline void __ast_write8_i_masked(void __iomem *addr, u32 reg, u8 index, u8 preserve_mask, u8 val) { - u8 tmp = __ast_read8_i_masked(addr, reg, index, read_mask); + u8 tmp = __ast_read8_i_masked(addr, reg, index, preserve_mask); - tmp |= val; - __ast_write8_i(addr, reg, index, tmp); + val &= ~preserve_mask; + __ast_write8_i(addr, reg, index, tmp | val); } static inline u32 ast_read32(struct ast_device *ast, u32 reg) diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c index b4e8edc7c767..30b011ed0a05 100644 --- a/drivers/gpu/drm/ast/ast_mode.c +++ b/drivers/gpu/drm/ast/ast_mode.c @@ -836,22 +836,24 @@ ast_crtc_helper_atomic_flush(struct drm_crtc *crtc, static void ast_crtc_helper_atomic_enable(struct drm_crtc *crtc, struct drm_atomic_state *state) { struct ast_device *ast = to_ast_device(crtc->dev); + u8 vgacr17 = 0x00; + u8 vgacrb6 = 0xff; - ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, 0x00); - ast_set_index_reg_mask(ast, AST_IO_VGASRI, 0x01, 0xdf, 0x00); + vgacr17 |= AST_IO_VGACR17_SYNC_ENABLE; + vgacrb6 &= ~(AST_IO_VGACRB6_VSYNC_OFF | AST_IO_VGACRB6_HSYNC_OFF); + + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0x17, 0x7f, vgacr17); + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, vgacrb6); } static void ast_crtc_helper_atomic_disable(struct drm_crtc *crtc, struct drm_atomic_state *state) { struct drm_crtc_state *old_crtc_state = drm_atomic_get_old_crtc_state(state, crtc); struct ast_device *ast = to_ast_device(crtc->dev); - u8 vgacrb6; + u8 vgacr17 = 0xff; - ast_set_index_reg_mask(ast, AST_IO_VGASRI, 0x01, 0xdf, AST_IO_VGASR1_SD); - - vgacrb6 = AST_IO_VGACRB6_VSYNC_OFF | - AST_IO_VGACRB6_HSYNC_OFF; - ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, vgacrb6); + vgacr17 &= ~AST_IO_VGACR17_SYNC_ENABLE; + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0x17, 0x7f, vgacr17); /* * HW cursors require the underlying primary plane and CRTC to diff --git a/drivers/gpu/drm/ast/ast_reg.h b/drivers/gpu/drm/ast/ast_reg.h index e15adaf3a80e..30578e3b07e4 100644 --- a/drivers/gpu/drm/ast/ast_reg.h +++ b/drivers/gpu/drm/ast/ast_reg.h @@ -29,6 +29,7 @@ #define AST_IO_VGAGRI (0x4E) #define AST_IO_VGACRI (0x54) +#define AST_IO_VGACR17_SYNC_ENABLE BIT(7) /* called "Hardware reset" in docs */ #define AST_IO_VGACR80_PASSWORD (0xa8) #define AST_IO_VGACR99_VGAMEM_RSRV_MASK GENMASK(1, 0) #define AST_IO_VGACRA1_VGAIO_DISABLED BIT(1) diff --git a/drivers/gpu/drm/bridge/lontium-lt9211.c b/drivers/gpu/drm/bridge/lontium-lt9211.c index 399fa7eebd49..03fc8fd10f20 100644 --- a/drivers/gpu/drm/bridge/lontium-lt9211.c +++ b/drivers/gpu/drm/bridge/lontium-lt9211.c @@ -121,8 +121,7 @@ static int lt9211_read_chipid(struct lt9211 *ctx) } /* Test for known Chip ID. */ - if (chipid[0] != REG_CHIPID0_VALUE || chipid[1] != REG_CHIPID1_VALUE || - chipid[2] != REG_CHIPID2_VALUE) { + if (chipid[0] != REG_CHIPID0_VALUE || chipid[1] != REG_CHIPID1_VALUE) { dev_err(ctx->dev, "Unknown Chip ID: 0x%02x 0x%02x 0x%02x\n", chipid[0], chipid[1], chipid[2]); return -EINVAL; diff --git a/drivers/gpu/drm/ci/gitlab-ci.yml b/drivers/gpu/drm/ci/gitlab-ci.yml index d502d146b177..56638814bb28 100644 --- a/drivers/gpu/drm/ci/gitlab-ci.yml +++ b/drivers/gpu/drm/ci/gitlab-ci.yml @@ -280,7 +280,7 @@ sanity: GIT_STRATEGY: none script: # ci-fairy check-commits --junit-xml=check-commits.xml - - ci-fairy check-merge-request --require-allow-collaboration --junit-xml=check-merge-request.xml + # - ci-fairy check-merge-request --require-allow-collaboration --junit-xml=check-merge-request.xml - | set -eu image_tags=( diff --git a/drivers/gpu/drm/drm_draw.c b/drivers/gpu/drm/drm_draw.c index 9dc0408fbbea..5b956229c82f 100644 --- a/drivers/gpu/drm/drm_draw.c +++ b/drivers/gpu/drm/drm_draw.c @@ -127,7 +127,7 @@ EXPORT_SYMBOL(drm_draw_fill16); void drm_draw_fill24(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, - u16 color) + u32 color) { unsigned int y, x; diff --git a/drivers/gpu/drm/drm_draw_internal.h b/drivers/gpu/drm/drm_draw_internal.h index f121ee7339dc..20cb404e23ea 100644 --- a/drivers/gpu/drm/drm_draw_internal.h +++ b/drivers/gpu/drm/drm_draw_internal.h @@ -47,7 +47,7 @@ void drm_draw_fill16(struct iosys_map *dmap, unsigned int dpitch, void drm_draw_fill24(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, - u16 color); + u32 color); void drm_draw_fill32(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, diff --git a/drivers/gpu/drm/drm_gem_atomic_helper.c b/drivers/gpu/drm/drm_gem_atomic_helper.c index ebf305fb24f0..6fb55601252f 100644 --- a/drivers/gpu/drm/drm_gem_atomic_helper.c +++ b/drivers/gpu/drm/drm_gem_atomic_helper.c @@ -310,8 +310,12 @@ EXPORT_SYMBOL(drm_gem_destroy_shadow_plane_state); void __drm_gem_reset_shadow_plane(struct drm_plane *plane, struct drm_shadow_plane_state *shadow_plane_state) { - __drm_atomic_helper_plane_reset(plane, &shadow_plane_state->base); - drm_format_conv_state_init(&shadow_plane_state->fmtcnv_state); + if (shadow_plane_state) { + __drm_atomic_helper_plane_reset(plane, &shadow_plane_state->base); + drm_format_conv_state_init(&shadow_plane_state->fmtcnv_state); + } else { + __drm_atomic_helper_plane_reset(plane, NULL); + } } EXPORT_SYMBOL(__drm_gem_reset_shadow_plane); diff --git a/drivers/gpu/drm/drm_panic.c b/drivers/gpu/drm/drm_panic.c index 1d6312fa1429..d4b6ea42db0f 100644 --- a/drivers/gpu/drm/drm_panic.c +++ b/drivers/gpu/drm/drm_panic.c @@ -174,6 +174,33 @@ static void drm_panic_write_pixel24(void *vaddr, unsigned int offset, u32 color) *p = color & 0xff; } +/* + * Special case if the pixel crosses page boundaries + */ +static void drm_panic_write_pixel24_xpage(void *vaddr, struct page *next_page, + unsigned int offset, u32 color) +{ + u8 *vaddr2; + u8 *p = vaddr + offset; + + vaddr2 = kmap_local_page_try_from_panic(next_page); + + *p++ = color & 0xff; + color >>= 8; + + if (offset == PAGE_SIZE - 1) + p = vaddr2; + + *p++ = color & 0xff; + color >>= 8; + + if (offset == PAGE_SIZE - 2) + p = vaddr2; + + *p = color & 0xff; + kunmap_local(vaddr2); +} + static void drm_panic_write_pixel32(void *vaddr, unsigned int offset, u32 color) { u32 *p = vaddr + offset; @@ -231,7 +258,14 @@ static void drm_panic_blit_page(struct page **pages, unsigned int dpitch, page = new_page; vaddr = kmap_local_page_try_from_panic(pages[page]); } - if (vaddr) + if (!vaddr) + continue; + + // Special case for 24bit, as a pixel might cross page boundaries + if (cpp == 3 && offset + 3 > PAGE_SIZE) + drm_panic_write_pixel24_xpage(vaddr, pages[page + 1], + offset, fg32); + else drm_panic_write_pixel(vaddr, offset, fg32, cpp); } } @@ -321,7 +355,15 @@ static void drm_panic_fill_page(struct page **pages, unsigned int dpitch, page = new_page; vaddr = kmap_local_page_try_from_panic(pages[page]); } - drm_panic_write_pixel(vaddr, offset, color, cpp); + if (!vaddr) + continue; + + // Special case for 24bit, as a pixel might cross page boundaries + if (cpp == 3 && offset + 3 > PAGE_SIZE) + drm_panic_write_pixel24_xpage(vaddr, pages[page + 1], + offset, color); + else + drm_panic_write_pixel(vaddr, offset, color, cpp); } } if (vaddr) @@ -429,6 +471,9 @@ static void drm_panic_logo_rect(struct drm_rect *rect, const struct font_desc *f static void drm_panic_logo_draw(struct drm_scanout_buffer *sb, struct drm_rect *rect, const struct font_desc *font, u32 fg_color) { + if (rect->x2 > sb->width || rect->y2 > sb->height) + return; + if (logo_mono) drm_panic_blit(sb, rect, logo_mono->data, DIV_ROUND_UP(drm_rect_width(rect), 8), 1, fg_color); @@ -477,7 +522,7 @@ static int draw_line_with_wrap(struct drm_scanout_buffer *sb, const struct font_ struct drm_panic_line *line, int yoffset, u32 fg_color) { int chars_per_row = sb->width / font->width; - struct drm_rect r_txt = DRM_RECT_INIT(0, yoffset, sb->width, sb->height); + struct drm_rect r_txt = DRM_RECT_INIT(0, yoffset, sb->width, font->height); struct drm_panic_line line_wrap; if (line->len > chars_per_row) { @@ -520,7 +565,7 @@ static void draw_panic_static_kmsg(struct drm_scanout_buffer *sb) struct drm_panic_line line; int yoffset; - if (!font) + if (!font || font->width > sb->width) return; yoffset = sb->height - font->height - (sb->height % font->height) / 2; @@ -733,7 +778,10 @@ static int _draw_panic_static_qr_code(struct drm_scanout_buffer *sb) pr_debug("QR width %d and scale %d\n", qr_width, scale); r_qr_canvas = DRM_RECT_INIT(0, 0, qr_canvas_width * scale, qr_canvas_width * scale); - v_margin = (sb->height - drm_rect_height(&r_qr_canvas) - drm_rect_height(&r_msg)) / 5; + v_margin = sb->height - drm_rect_height(&r_qr_canvas) - drm_rect_height(&r_msg); + if (v_margin < 0) + return -ENOSPC; + v_margin /= 5; drm_rect_translate(&r_qr_canvas, (sb->width - r_qr_canvas.x2) / 2, 2 * v_margin); r_qr = DRM_RECT_INIT(r_qr_canvas.x1 + QR_MARGIN * scale, r_qr_canvas.y1 + QR_MARGIN * scale, @@ -746,7 +794,7 @@ static int _draw_panic_static_qr_code(struct drm_scanout_buffer *sb) /* Fill with the background color, and draw text on top */ drm_panic_fill(sb, &r_screen, bg_color); - if (!drm_rect_overlap(&r_logo, &r_msg) && !drm_rect_overlap(&r_logo, &r_qr)) + if (!drm_rect_overlap(&r_logo, &r_msg) && !drm_rect_overlap(&r_logo, &r_qr_canvas)) drm_panic_logo_draw(sb, &r_logo, font, fg_color); draw_txt_rectangle(sb, font, panic_msg, panic_msg_lines, true, &r_msg, fg_color); diff --git a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c index b13a17276d07..88385dc3b30d 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_buffer.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_buffer.c @@ -347,7 +347,7 @@ void etnaviv_buffer_queue(struct etnaviv_gpu *gpu, u32 exec_state, u32 link_target, link_dwords; bool switch_context = gpu->exec_state != exec_state; bool switch_mmu_context = gpu->mmu_context != mmu_context; - unsigned int new_flush_seq = READ_ONCE(gpu->mmu_context->flush_seq); + unsigned int new_flush_seq = READ_ONCE(mmu_context->flush_seq); bool need_flush = switch_mmu_context || gpu->flush_seq != new_flush_seq; bool has_blt = !!(gpu->identity.minor_features5 & chipMinorFeatures5_BLT_ENGINE); diff --git a/drivers/gpu/drm/i915/display/intel_dmc.c b/drivers/gpu/drm/i915/display/intel_dmc.c index 77a0199f9ea5..4a4cace1f879 100644 --- a/drivers/gpu/drm/i915/display/intel_dmc.c +++ b/drivers/gpu/drm/i915/display/intel_dmc.c @@ -546,6 +546,36 @@ static bool is_event_handler(struct intel_display *display, REG_FIELD_GET(DMC_EVT_CTL_EVENT_ID_MASK, data) == event_id; } +static bool fixup_dmc_evt(struct intel_display *display, + enum intel_dmc_id dmc_id, + i915_reg_t reg_ctl, u32 *data_ctl, + i915_reg_t reg_htp, u32 *data_htp) +{ + if (!is_dmc_evt_ctl_reg(display, dmc_id, reg_ctl)) + return false; + + if (!is_dmc_evt_htp_reg(display, dmc_id, reg_htp)) + return false; + + /* make sure reg_ctl and reg_htp are for the same event */ + if (i915_mmio_reg_offset(reg_ctl) - i915_mmio_reg_offset(DMC_EVT_CTL(display, dmc_id, 0)) != + i915_mmio_reg_offset(reg_htp) - i915_mmio_reg_offset(DMC_EVT_HTP(display, dmc_id, 0))) + return false; + + /* + * On ADL-S the HRR event handler is not restored after DC6. + * Clear it to zero from the beginning to avoid mismatches later. + */ + if (display->platform.alderlake_s && dmc_id == DMC_FW_MAIN && + is_event_handler(display, dmc_id, MAINDMC_EVENT_VBLANK_A, reg_ctl, *data_ctl)) { + *data_ctl = 0; + *data_htp = 0; + return true; + } + + return false; +} + static bool disable_dmc_evt(struct intel_display *display, enum intel_dmc_id dmc_id, i915_reg_t reg, u32 data) @@ -1064,9 +1094,32 @@ static u32 parse_dmc_fw_header(struct intel_dmc *dmc, for (i = 0; i < mmio_count; i++) { dmc_info->mmioaddr[i] = _MMIO(mmioaddr[i]); dmc_info->mmiodata[i] = mmiodata[i]; + } + + for (i = 0; i < mmio_count - 1; i++) { + u32 orig_mmiodata[2] = { + dmc_info->mmiodata[i], + dmc_info->mmiodata[i+1], + }; + + if (!fixup_dmc_evt(display, dmc_id, + dmc_info->mmioaddr[i], &dmc_info->mmiodata[i], + dmc_info->mmioaddr[i+1], &dmc_info->mmiodata[i+1])) + continue; + + drm_dbg_kms(display->drm, + " mmio[%d]: 0x%x = 0x%x->0x%x (EVT_CTL)\n", + i, i915_mmio_reg_offset(dmc_info->mmioaddr[i]), + orig_mmiodata[0], dmc_info->mmiodata[i]); + drm_dbg_kms(display->drm, + " mmio[%d]: 0x%x = 0x%x->0x%x (EVT_HTP)\n", + i+1, i915_mmio_reg_offset(dmc_info->mmioaddr[i+1]), + orig_mmiodata[1], dmc_info->mmiodata[i+1]); + } + for (i = 0; i < mmio_count; i++) { drm_dbg_kms(display->drm, " mmio[%d]: 0x%x = 0x%x%s%s\n", - i, mmioaddr[i], mmiodata[i], + i, i915_mmio_reg_offset(dmc_info->mmioaddr[i]), dmc_info->mmiodata[i], is_dmc_evt_ctl_reg(display, dmc_id, dmc_info->mmioaddr[i]) ? " (EVT_CTL)" : is_dmc_evt_htp_reg(display, dmc_id, dmc_info->mmioaddr[i]) ? " (EVT_HTP)" : "", disable_dmc_evt(display, dmc_id, dmc_info->mmioaddr[i], diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/drm/i915/display/intel_fb.c index 22a4a1575d22..c48384e58ea1 100644 --- a/drivers/gpu/drm/i915/display/intel_fb.c +++ b/drivers/gpu/drm/i915/display/intel_fb.c @@ -2113,10 +2113,11 @@ static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb) if (intel_fb_uses_dpt(fb)) intel_dpt_destroy(intel_fb->dpt_vm); - intel_frontbuffer_put(intel_fb->frontbuffer); - intel_fb_bo_framebuffer_fini(intel_fb_bo(fb)); + intel_frontbuffer_put(intel_fb->frontbuffer); + + kfree(intel_fb->panic); kfree(intel_fb); } @@ -2215,19 +2216,27 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, struct intel_display *display = to_intel_display(obj->dev); struct drm_framebuffer *fb = &intel_fb->base; u32 max_stride; - int ret = -EINVAL; + int ret; int i; - ret = intel_fb_bo_framebuffer_init(fb, obj, mode_cmd); - if (ret) - return ret; + intel_fb->panic = intel_panic_alloc(); + if (!intel_fb->panic) + return -ENOMEM; + /* + * intel_frontbuffer_get() must be done before + * intel_fb_bo_framebuffer_init() to avoid set_tiling vs. addfb race. + */ intel_fb->frontbuffer = intel_frontbuffer_get(obj); if (!intel_fb->frontbuffer) { ret = -ENOMEM; - goto err; + goto err_free_panic; } + ret = intel_fb_bo_framebuffer_init(fb, obj, mode_cmd); + if (ret) + goto err_frontbuffer_put; + ret = -EINVAL; if (!drm_any_plane_has_format(display->drm, mode_cmd->pixel_format, @@ -2235,7 +2244,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "unsupported pixel format %p4cc / modifier 0x%llx\n", &mode_cmd->pixel_format, mode_cmd->modifier[0]); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } max_stride = intel_fb_max_stride(display, mode_cmd->pixel_format, @@ -2246,7 +2255,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, mode_cmd->modifier[0] != DRM_FORMAT_MOD_LINEAR ? "tiled" : "linear", mode_cmd->pitches[0], max_stride); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } /* FIXME need to adjust LINOFF/TILEOFF accordingly. */ @@ -2254,7 +2263,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "plane 0 offset (0x%08x) must be 0\n", mode_cmd->offsets[0]); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } drm_helper_mode_fill_fb_struct(display->drm, fb, info, mode_cmd); @@ -2264,7 +2273,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, if (mode_cmd->handles[i] != mode_cmd->handles[0]) { drm_dbg_kms(display->drm, "bad plane %d handle\n", i); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } stride_alignment = intel_fb_stride_alignment(fb, i); @@ -2272,7 +2281,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "plane %d pitch (%d) must be at least %u byte aligned\n", i, fb->pitches[i], stride_alignment); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } if (intel_fb_is_gen12_ccs_aux_plane(fb, i)) { @@ -2282,7 +2291,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "ccs aux plane %d pitch (%d) must be %d\n", i, fb->pitches[i], ccs_aux_stride); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } } @@ -2291,7 +2300,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, ret = intel_fill_fb_info(display, intel_fb); if (ret) - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; if (intel_fb_uses_dpt(fb)) { struct i915_address_space *vm; @@ -2317,10 +2326,13 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, err_free_dpt: if (intel_fb_uses_dpt(fb)) intel_dpt_destroy(intel_fb->dpt_vm); +err_bo_framebuffer_fini: + intel_fb_bo_framebuffer_fini(obj); err_frontbuffer_put: intel_frontbuffer_put(intel_fb->frontbuffer); -err: - intel_fb_bo_framebuffer_fini(obj); +err_free_panic: + kfree(intel_fb->panic); + return ret; } @@ -2347,20 +2359,11 @@ intel_user_framebuffer_create(struct drm_device *dev, struct intel_framebuffer *intel_framebuffer_alloc(void) { struct intel_framebuffer *intel_fb; - struct intel_panic *panic; intel_fb = kzalloc(sizeof(*intel_fb), GFP_KERNEL); if (!intel_fb) return NULL; - panic = intel_panic_alloc(); - if (!panic) { - kfree(intel_fb); - return NULL; - } - - intel_fb->panic = panic; - return intel_fb; } diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c index 43be5377ddc1..73ed28ac9573 100644 --- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c @@ -270,6 +270,8 @@ static void frontbuffer_release(struct kref *ref) spin_unlock(&display->fb_tracking.lock); i915_active_fini(&front->write); + + drm_gem_object_put(obj); kfree_rcu(front, rcu); } @@ -287,6 +289,8 @@ intel_frontbuffer_get(struct drm_gem_object *obj) if (!front) return NULL; + drm_gem_object_get(obj); + front->obj = obj; kref_init(&front->ref); atomic_set(&front->bits, 0); @@ -299,8 +303,12 @@ intel_frontbuffer_get(struct drm_gem_object *obj) spin_lock(&display->fb_tracking.lock); cur = intel_bo_set_frontbuffer(obj, front); spin_unlock(&display->fb_tracking.lock); - if (cur != front) + + if (cur != front) { + drm_gem_object_put(obj); kfree(front); + } + return cur; } diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 01bf304c705f..10eb93a34cf2 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -3402,6 +3402,7 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) struct intel_display *display = to_intel_display(intel_dp); if (DISPLAY_VER(display) < 20 && intel_dp->psr.psr2_sel_fetch_enabled) { + /* Selective fetch prior LNL */ if (intel_dp->psr.psr2_sel_fetch_cff_enabled) { /* can we turn CFF off? */ if (intel_dp->psr.busy_frontbuffer_bits == 0) @@ -3420,12 +3421,19 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) intel_psr_configure_full_frame_update(intel_dp); intel_psr_force_update(intel_dp); + } else if (!intel_dp->psr.psr2_sel_fetch_enabled) { + /* + * PSR1 on all platforms + * PSR2 HW tracking + * Panel Replay Full frame update + */ + intel_psr_force_update(intel_dp); } else { + /* Selective update LNL onwards */ intel_psr_exit(intel_dp); } - if ((!intel_dp->psr.psr2_sel_fetch_enabled || DISPLAY_VER(display) >= 20) && - !intel_dp->psr.busy_frontbuffer_bits) + if (!intel_dp->psr.active && !intel_dp->psr.busy_frontbuffer_bits) queue_work(display->wq.unordered, &intel_dp->psr.work); } diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h b/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h index b6dc3d1b9bb1..b682969e3a29 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h @@ -89,12 +89,10 @@ i915_gem_object_set_frontbuffer(struct drm_i915_gem_object *obj, if (!front) { RCU_INIT_POINTER(obj->frontbuffer, NULL); - drm_gem_object_put(intel_bo_to_drm_bo(obj)); } else if (rcu_access_pointer(obj->frontbuffer)) { cur = rcu_dereference_protected(obj->frontbuffer, true); kref_get(&cur->ref); } else { - drm_gem_object_get(intel_bo_to_drm_bo(obj)); rcu_assign_pointer(obj->frontbuffer, front); } diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c index 3e7e5badcc2b..2c651ec024ef 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c @@ -1325,9 +1325,16 @@ static int ct_receive(struct intel_guc_ct *ct) static void ct_try_receive_message(struct intel_guc_ct *ct) { + struct intel_guc *guc = ct_to_guc(ct); int ret; - if (GEM_WARN_ON(!ct->enabled)) + if (!ct->enabled) { + GEM_WARN_ON(!guc_to_gt(guc)->uc.reset_in_progress); + return; + } + + /* When interrupt disabled, message handling is not expected */ + if (!guc->interrupts.enabled) return; ret = ct_receive(ct); diff --git a/drivers/gpu/drm/imx/ipuv3/parallel-display.c b/drivers/gpu/drm/imx/ipuv3/parallel-display.c index 6d8325c76697..7fc6af703307 100644 --- a/drivers/gpu/drm/imx/ipuv3/parallel-display.c +++ b/drivers/gpu/drm/imx/ipuv3/parallel-display.c @@ -25,19 +25,18 @@ struct imx_parallel_display_encoder { struct drm_encoder encoder; - struct drm_bridge bridge; - struct imx_parallel_display *pd; }; struct imx_parallel_display { struct device *dev; u32 bus_format; struct drm_bridge *next_bridge; + struct drm_bridge bridge; }; static inline struct imx_parallel_display *bridge_to_imxpd(struct drm_bridge *b) { - return container_of(b, struct imx_parallel_display_encoder, bridge)->pd; + return container_of(b, struct imx_parallel_display, bridge); } static const u32 imx_pd_bus_fmts[] = { @@ -195,15 +194,13 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data) if (IS_ERR(imxpd_encoder)) return PTR_ERR(imxpd_encoder); - imxpd_encoder->pd = imxpd; encoder = &imxpd_encoder->encoder; - bridge = &imxpd_encoder->bridge; + bridge = &imxpd->bridge; ret = imx_drm_encoder_parse_of(drm, encoder, imxpd->dev->of_node); if (ret) return ret; - bridge->funcs = &imx_pd_bridge_funcs; drm_bridge_attach(encoder, bridge, NULL, DRM_BRIDGE_ATTACH_NO_CONNECTOR); connector = drm_bridge_connector_init(drm, encoder); @@ -228,9 +225,10 @@ static int imx_pd_probe(struct platform_device *pdev) u32 bus_format = 0; const char *fmt; - imxpd = devm_kzalloc(dev, sizeof(*imxpd), GFP_KERNEL); - if (!imxpd) - return -ENOMEM; + imxpd = devm_drm_bridge_alloc(dev, struct imx_parallel_display, bridge, + &imx_pd_bridge_funcs); + if (IS_ERR(imxpd)) + return PTR_ERR(imxpd); /* port@1 is the output port */ imxpd->next_bridge = devm_drm_of_get_bridge(dev, np, 1, 0); @@ -258,6 +256,8 @@ static int imx_pd_probe(struct platform_device *pdev) platform_set_drvdata(pdev, imxpd); + devm_drm_bridge_add(dev, &imxpd->bridge); + return component_add(dev, &imx_pd_ops); } diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c index eb5537f0ac90..31ff2922758a 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c @@ -686,10 +686,6 @@ err_free: for (i = 0; i < private->data->mmsys_dev_num; i++) private->all_drm_private[i]->drm = NULL; err_put_dev: - for (i = 0; i < private->data->mmsys_dev_num; i++) { - /* For device_find_child in mtk_drm_get_all_priv() */ - put_device(private->all_drm_private[i]->dev); - } put_device(private->mutex_dev); return ret; } @@ -697,18 +693,12 @@ err_put_dev: static void mtk_drm_unbind(struct device *dev) { struct mtk_drm_private *private = dev_get_drvdata(dev); - int i; /* for multi mmsys dev, unregister drm dev in mmsys master */ if (private->drm_master) { drm_dev_unregister(private->drm); mtk_drm_kms_deinit(private->drm); drm_dev_put(private->drm); - - for (i = 0; i < private->data->mmsys_dev_num; i++) { - /* For device_find_child in mtk_drm_get_all_priv() */ - put_device(private->all_drm_private[i]->dev); - } put_device(private->mutex_dev); } private->mtk_drm_bound = false; diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index fc62fef2fed8..4e6dc16e4a4c 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -780,6 +780,9 @@ static bool fw_block_mem(struct a6xx_gmu_bo *bo, const struct block_header *blk) return true; } +#define NEXT_BLK(blk) \ + ((const struct block_header *)((const char *)(blk) + sizeof(*(blk)) + (blk)->size)) + static int a6xx_gmu_fw_load(struct a6xx_gmu *gmu) { struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); @@ -811,7 +814,7 @@ static int a6xx_gmu_fw_load(struct a6xx_gmu *gmu) for (blk = (const struct block_header *) fw_image->data; (const u8*) blk < fw_image->data + fw_image->size; - blk = (const struct block_header *) &blk->data[blk->size >> 2]) { + blk = NEXT_BLK(blk)) { if (blk->size == 0) continue; diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index afaa3cfefd35..4b5a4edd0702 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -348,13 +348,6 @@ int adreno_fault_handler(struct msm_gpu *gpu, unsigned long iova, int flags, return 0; } -static bool -adreno_smmu_has_prr(struct msm_gpu *gpu) -{ - struct adreno_smmu_priv *adreno_smmu = dev_get_drvdata(&gpu->pdev->dev); - return adreno_smmu && adreno_smmu->set_prr_addr; -} - int adreno_get_param(struct msm_gpu *gpu, struct msm_context *ctx, uint32_t param, uint64_t *value, uint32_t *len) { diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c index 4b970a59deaf..2f8156051d9b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c @@ -1545,6 +1545,9 @@ static enum drm_mode_status dpu_crtc_mode_valid(struct drm_crtc *crtc, adjusted_mode_clk = dpu_core_perf_adjusted_mode_clk(mode->clock, dpu_kms->perf.perf_cfg); + if (dpu_kms->catalog->caps->has_3d_merge) + adjusted_mode_clk /= 2; + /* * The given mode, adjusted for the perf clock factor, should not exceed * the max core clock rate diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c index 6641455c4ec6..9f8d1bba9139 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c @@ -267,8 +267,8 @@ static const u32 wb2_formats_rgb_yuv[] = { .base = 0x200, .len = 0xa0,}, \ .csc_blk = {.name = "csc", \ .base = 0x320, .len = 0x100,}, \ - .format_list = plane_formats_yuv, \ - .num_formats = ARRAY_SIZE(plane_formats_yuv), \ + .format_list = plane_formats, \ + .num_formats = ARRAY_SIZE(plane_formats), \ .rotation_cfg = NULL, \ } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c index f54cf0faa1c7..905524ceeb1f 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c @@ -500,13 +500,15 @@ static void _dpu_plane_setup_pixel_ext(struct dpu_hw_scaler3_cfg *scale_cfg, int i; for (i = 0; i < DPU_MAX_PLANES; i++) { + uint32_t w = src_w, h = src_h; + if (i == DPU_SSPP_COMP_1_2 || i == DPU_SSPP_COMP_2) { - src_w /= chroma_subsmpl_h; - src_h /= chroma_subsmpl_v; + w /= chroma_subsmpl_h; + h /= chroma_subsmpl_v; } - pixel_ext->num_ext_pxls_top[i] = src_h; - pixel_ext->num_ext_pxls_left[i] = src_w; + pixel_ext->num_ext_pxls_top[i] = h; + pixel_ext->num_ext_pxls_left[i] = w; } } @@ -740,7 +742,7 @@ static int dpu_plane_atomic_check_pipe(struct dpu_plane *pdpu, * We already have verified scaling against platform limitations. * Now check if the SSPP supports scaling at all. */ - if (!sblk->scaler_blk.len && + if (!(sblk->scaler_blk.len && pipe->sspp->ops.setup_scaler) && ((drm_rect_width(&new_plane_state->src) >> 16 != drm_rect_width(&new_plane_state->dst)) || (drm_rect_height(&new_plane_state->src) >> 16 != @@ -1278,7 +1280,7 @@ int dpu_assign_plane_resources(struct dpu_global_state *global_state, state, plane_state, prev_adjacent_plane_state); if (ret) - break; + return ret; prev_adjacent_plane_state = plane_state; } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c index 2c77c74fac0f..d9c3b0a1d091 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c @@ -842,7 +842,7 @@ struct dpu_hw_sspp *dpu_rm_reserve_sspp(struct dpu_rm *rm, if (!reqs->scale && !reqs->yuv) hw_sspp = dpu_rm_try_sspp(rm, global_state, crtc, reqs, SSPP_TYPE_DMA); - if (!hw_sspp && reqs->scale) + if (!hw_sspp && !reqs->yuv) hw_sspp = dpu_rm_try_sspp(rm, global_state, crtc, reqs, SSPP_TYPE_RGB); if (!hw_sspp) hw_sspp = dpu_rm_try_sspp(rm, global_state, crtc, reqs, SSPP_TYPE_VIG); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_writeback.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_writeback.c index cd73468e369a..7545c0293efb 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_writeback.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_writeback.c @@ -72,6 +72,9 @@ static int dpu_wb_conn_atomic_check(struct drm_connector *connector, DPU_ERROR("invalid fb w=%d, maxlinewidth=%u\n", fb->width, dpu_wb_conn->maxlinewidth); return -EINVAL; + } else if (fb->modifier != DRM_FORMAT_MOD_LINEAR) { + DPU_ERROR("unsupported fb modifier:%#llx\n", fb->modifier); + return -EINVAL; } return drm_atomic_helper_check_wb_connector_state(conn_state->connector, conn_state->state); diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h index e391505fdaf0..3cbf08231492 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h @@ -109,7 +109,6 @@ struct msm_dsi_phy { struct msm_dsi_dphy_timing timing; const struct msm_dsi_phy_cfg *cfg; void *tuning_cfg; - void *pll_data; enum msm_dsi_phy_usecase usecase; bool regulator_ldo_mode; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c index 32f06edd21a9..c5e1d2016bcc 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c @@ -426,11 +426,8 @@ static void dsi_pll_enable_pll_bias(struct dsi_pll_7nm *pll) u32 data; spin_lock_irqsave(&pll->pll_enable_lock, flags); - if (pll->pll_enable_cnt++) { - spin_unlock_irqrestore(&pll->pll_enable_lock, flags); - WARN_ON(pll->pll_enable_cnt == INT_MAX); - return; - } + pll->pll_enable_cnt++; + WARN_ON(pll->pll_enable_cnt == INT_MAX); data = readl(pll->phy->base + REG_DSI_7nm_PHY_CMN_CTRL_0); data |= DSI_7nm_PHY_CMN_CTRL_0_PLL_SHUTDOWNB; @@ -876,7 +873,6 @@ static int dsi_pll_7nm_init(struct msm_dsi_phy *phy) spin_lock_init(&pll_7nm->pll_enable_lock); pll_7nm->phy = phy; - phy->pll_data = pll_7nm; ret = pll_7nm_register(pll_7nm, phy->provided_clocks->hws); if (ret) { @@ -965,10 +961,8 @@ static int dsi_7nm_phy_enable(struct msm_dsi_phy *phy, u32 const delay_us = 5; u32 const timeout_us = 1000; struct msm_dsi_dphy_timing *timing = &phy->timing; - struct dsi_pll_7nm *pll = phy->pll_data; void __iomem *base = phy->base; bool less_than_1500_mhz; - unsigned long flags; u32 vreg_ctrl_0, vreg_ctrl_1, lane_ctrl0; u32 glbl_pemph_ctrl_0; u32 glbl_str_swi_cal_sel_ctrl, glbl_hstx_str_ctrl_0; @@ -1090,13 +1084,10 @@ static int dsi_7nm_phy_enable(struct msm_dsi_phy *phy, glbl_rescode_bot_ctrl = 0x3c; } - spin_lock_irqsave(&pll->pll_enable_lock, flags); - pll->pll_enable_cnt = 1; /* de-assert digital and pll power down */ data = DSI_7nm_PHY_CMN_CTRL_0_DIGTOP_PWRDN_B | DSI_7nm_PHY_CMN_CTRL_0_PLL_SHUTDOWNB; writel(data, base + REG_DSI_7nm_PHY_CMN_CTRL_0); - spin_unlock_irqrestore(&pll->pll_enable_lock, flags); /* Assert PLL core reset */ writel(0x00, base + REG_DSI_7nm_PHY_CMN_PLL_CNTRL); @@ -1209,9 +1200,7 @@ static bool dsi_7nm_set_continuous_clock(struct msm_dsi_phy *phy, bool enable) static void dsi_7nm_phy_disable(struct msm_dsi_phy *phy) { - struct dsi_pll_7nm *pll = phy->pll_data; void __iomem *base = phy->base; - unsigned long flags; u32 data; DBG(""); @@ -1238,11 +1227,8 @@ static void dsi_7nm_phy_disable(struct msm_dsi_phy *phy) writel(data, base + REG_DSI_7nm_PHY_CMN_CTRL_0); writel(0, base + REG_DSI_7nm_PHY_CMN_LANE_CTRL0); - spin_lock_irqsave(&pll->pll_enable_lock, flags); - pll->pll_enable_cnt = 0; /* Turn off all PHY blocks */ writel(0x00, base + REG_DSI_7nm_PHY_CMN_CTRL_0); - spin_unlock_irqrestore(&pll->pll_enable_lock, flags); /* make sure phy is turned off */ wmb(); diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c index 07d8cdd6bb2e..9f7fbe577abb 100644 --- a/drivers/gpu/drm/msm/msm_gem.c +++ b/drivers/gpu/drm/msm/msm_gem.c @@ -1120,12 +1120,16 @@ static void msm_gem_free_object(struct drm_gem_object *obj) put_pages(obj); } - if (obj->resv != &obj->_resv) { + /* + * In error paths, we could end up here before msm_gem_new_handle() + * has changed obj->resv to point to the shared resv. In this case, + * we don't want to drop a ref to the shared r_obj that we haven't + * taken yet. + */ + if ((msm_obj->flags & MSM_BO_NO_SHARE) && (obj->resv != &obj->_resv)) { struct drm_gem_object *r_obj = container_of(obj->resv, struct drm_gem_object, _resv); - WARN_ON(!(msm_obj->flags & MSM_BO_NO_SHARE)); - /* Drop reference we hold to shared resv obj: */ drm_gem_object_put(r_obj); } diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index 3ab3b27134f9..75d9f3574370 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -414,6 +414,11 @@ static void submit_attach_object_fences(struct msm_gem_submit *submit) submit->user_fence, DMA_RESV_USAGE_BOOKKEEP, DMA_RESV_USAGE_BOOKKEEP); + + last_fence = vm->last_fence; + vm->last_fence = dma_fence_unwrap_merge(submit->user_fence, last_fence); + dma_fence_put(last_fence); + return; } @@ -427,10 +432,6 @@ static void submit_attach_object_fences(struct msm_gem_submit *submit) dma_resv_add_fence(obj->resv, submit->user_fence, DMA_RESV_USAGE_READ); } - - last_fence = vm->last_fence; - vm->last_fence = dma_fence_unwrap_merge(submit->user_fence, last_fence); - dma_fence_put(last_fence); } static int submit_bo(struct msm_gem_submit *submit, uint32_t idx, diff --git a/drivers/gpu/drm/msm/msm_gem_vma.c b/drivers/gpu/drm/msm/msm_gem_vma.c index 8316af1723c2..89a95977f41e 100644 --- a/drivers/gpu/drm/msm/msm_gem_vma.c +++ b/drivers/gpu/drm/msm/msm_gem_vma.c @@ -971,6 +971,7 @@ static int lookup_op(struct msm_vm_bind_job *job, const struct drm_msm_vm_bind_op *op) { struct drm_device *dev = job->vm->drm; + struct msm_drm_private *priv = dev->dev_private; int i = job->nr_ops++; int ret = 0; @@ -1017,6 +1018,11 @@ lookup_op(struct msm_vm_bind_job *job, const struct drm_msm_vm_bind_op *op) break; } + if ((op->op == MSM_VM_BIND_OP_MAP_NULL) && + !adreno_smmu_has_prr(priv->gpu)) { + ret = UERR(EINVAL, dev, "PRR not supported\n"); + } + return ret; } @@ -1421,7 +1427,7 @@ msm_ioctl_vm_bind(struct drm_device *dev, void *data, struct drm_file *file) * Maybe we could allow just UNMAP ops? OTOH userspace should just * immediately close the device file and all will be torn down. */ - if (to_msm_vm(ctx->vm)->unusable) + if (to_msm_vm(msm_context_vm(dev, ctx))->unusable) return UERR(EPIPE, dev, "context is unusable"); /* diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h index a597f2bee30b..2894fc118485 100644 --- a/drivers/gpu/drm/msm/msm_gpu.h +++ b/drivers/gpu/drm/msm/msm_gpu.h @@ -299,6 +299,17 @@ static inline struct msm_gpu *dev_to_gpu(struct device *dev) return container_of(adreno_smmu, struct msm_gpu, adreno_smmu); } +static inline bool +adreno_smmu_has_prr(struct msm_gpu *gpu) +{ + struct adreno_smmu_priv *adreno_smmu = dev_get_drvdata(&gpu->pdev->dev); + + if (!adreno_smmu) + return false; + + return adreno_smmu && adreno_smmu->set_prr_addr; +} + /* It turns out that all targets use the same ringbuffer size */ #define MSM_GPU_RINGBUFFER_SZ SZ_32K #define MSM_GPU_RINGBUFFER_BLKSIZE 32 diff --git a/drivers/gpu/drm/msm/msm_iommu.c b/drivers/gpu/drm/msm/msm_iommu.c index 0e18619f96cb..a188617653e8 100644 --- a/drivers/gpu/drm/msm/msm_iommu.c +++ b/drivers/gpu/drm/msm/msm_iommu.c @@ -338,6 +338,8 @@ msm_iommu_pagetable_prealloc_allocate(struct msm_mmu *mmu, struct msm_mmu_preall ret = kmem_cache_alloc_bulk(pt_cache, GFP_KERNEL, p->count, p->pages); if (ret != p->count) { + kfree(p->pages); + p->pages = NULL; p->count = ret; return -ENOMEM; } @@ -351,6 +353,9 @@ msm_iommu_pagetable_prealloc_cleanup(struct msm_mmu *mmu, struct msm_mmu_preallo struct kmem_cache *pt_cache = get_pt_cache(mmu); uint32_t remaining_pt_count = p->count - p->ptr; + if (!p->pages) + return; + if (p->count > 0) trace_msm_mmu_prealloc_cleanup(p->count, remaining_pt_count); diff --git a/drivers/gpu/drm/nouveau/nouveau_sched.c b/drivers/gpu/drm/nouveau/nouveau_sched.c index e60f7892f5ce..a7bf539e5d86 100644 --- a/drivers/gpu/drm/nouveau/nouveau_sched.c +++ b/drivers/gpu/drm/nouveau/nouveau_sched.c @@ -482,6 +482,17 @@ nouveau_sched_create(struct nouveau_sched **psched, struct nouveau_drm *drm, return 0; } +static bool +nouveau_sched_job_list_empty(struct nouveau_sched *sched) +{ + bool empty; + + spin_lock(&sched->job.list.lock); + empty = list_empty(&sched->job.list.head); + spin_unlock(&sched->job.list.lock); + + return empty; +} static void nouveau_sched_fini(struct nouveau_sched *sched) @@ -489,8 +500,7 @@ nouveau_sched_fini(struct nouveau_sched *sched) struct drm_gpu_scheduler *drm_sched = &sched->base; struct drm_sched_entity *entity = &sched->entity; - rmb(); /* for list_empty to work without lock */ - wait_event(sched->job.wq, list_empty(&sched->job.list.head)); + wait_event(sched->job.wq, nouveau_sched_job_list_empty(sched)); drm_sched_entity_fini(entity); drm_sched_fini(drm_sched); diff --git a/drivers/gpu/drm/panel/panel-kingdisplay-kd097d04.c b/drivers/gpu/drm/panel/panel-kingdisplay-kd097d04.c index 2fc7b0779b37..893af9b16756 100644 --- a/drivers/gpu/drm/panel/panel-kingdisplay-kd097d04.c +++ b/drivers/gpu/drm/panel/panel-kingdisplay-kd097d04.c @@ -359,7 +359,7 @@ static int kingdisplay_panel_probe(struct mipi_dsi_device *dsi) dsi->lanes = 4; dsi->format = MIPI_DSI_FMT_RGB888; dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_BURST | - MIPI_DSI_MODE_LPM; + MIPI_DSI_MODE_LPM | MIPI_DSI_MODE_NO_EOT_PACKET; kingdisplay = devm_drm_panel_alloc(&dsi->dev, __typeof(*kingdisplay), base, &kingdisplay_panel_funcs, diff --git a/drivers/gpu/drm/panel/panel-sitronix-st7789v.c b/drivers/gpu/drm/panel/panel-sitronix-st7789v.c index 04d91929eedd..d5f821d6b23c 100644 --- a/drivers/gpu/drm/panel/panel-sitronix-st7789v.c +++ b/drivers/gpu/drm/panel/panel-sitronix-st7789v.c @@ -249,6 +249,11 @@ static const struct drm_display_mode default_mode = { .flags = DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC, }; +/* + * The mode data for this panel has been reverse engineered without access + * to the panel datasheet / manual. Using DRM_MODE_FLAG_PHSYNC like all + * other panels results in garbage data on the display. + */ static const struct drm_display_mode t28cp45tn89_mode = { .clock = 6008, .hdisplay = 240, @@ -261,7 +266,7 @@ static const struct drm_display_mode t28cp45tn89_mode = { .vtotal = 320 + 8 + 4 + 4, .width_mm = 43, .height_mm = 57, - .flags = DRM_MODE_FLAG_PVSYNC | DRM_MODE_FLAG_NVSYNC, + .flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC, }; static const struct drm_display_mode et028013dma_mode = { diff --git a/drivers/gpu/drm/panthor/panthor_fw.c b/drivers/gpu/drm/panthor/panthor_fw.c index 9bf06e55eaee..df767e82148a 100644 --- a/drivers/gpu/drm/panthor/panthor_fw.c +++ b/drivers/gpu/drm/panthor/panthor_fw.c @@ -1099,6 +1099,7 @@ void panthor_fw_pre_reset(struct panthor_device *ptdev, bool on_hang) } panthor_job_irq_suspend(&ptdev->fw->irq); + panthor_fw_stop(ptdev); } /** diff --git a/drivers/gpu/drm/panthor/panthor_mmu.c b/drivers/gpu/drm/panthor/panthor_mmu.c index 6dec4354e378..7870e7dbaa5d 100644 --- a/drivers/gpu/drm/panthor/panthor_mmu.c +++ b/drivers/gpu/drm/panthor/panthor_mmu.c @@ -1175,10 +1175,14 @@ panthor_vm_op_ctx_prealloc_vmas(struct panthor_vm_op_ctx *op_ctx) break; case DRM_PANTHOR_VM_BIND_OP_TYPE_UNMAP: - /* Partial unmaps might trigger a remap with either a prev or a next VA, - * but not both. + /* Two VMAs can be needed for an unmap, as an unmap can happen + * in the middle of a drm_gpuva, requiring a remap with both + * prev & next VA. Or an unmap can span more than one drm_gpuva + * where the first and last ones are covered partially, requring + * a remap for the first with a prev VA and remap for the last + * with a next VA. */ - vma_count = 1; + vma_count = 2; break; default: diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 88e821d67af7..9c8907bc61d9 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -314,17 +314,17 @@ static int radeon_pci_probe(struct pci_dev *pdev, ret = pci_enable_device(pdev); if (ret) - goto err_free; + return ret; pci_set_drvdata(pdev, ddev); ret = radeon_driver_load_kms(ddev, flags); if (ret) - goto err_agp; + goto err; ret = drm_dev_register(ddev, flags); if (ret) - goto err_agp; + goto err; if (rdev->mc.real_vram_size <= (8 * 1024 * 1024)) format = drm_format_info(DRM_FORMAT_C8); @@ -337,30 +337,14 @@ static int radeon_pci_probe(struct pci_dev *pdev, return 0; -err_agp: +err: pci_disable_device(pdev); -err_free: - drm_dev_put(ddev); return ret; } static void -radeon_pci_remove(struct pci_dev *pdev) -{ - struct drm_device *dev = pci_get_drvdata(pdev); - - drm_put_dev(dev); -} - -static void radeon_pci_shutdown(struct pci_dev *pdev) { - /* if we are running in a VM, make sure the device - * torn down properly on reboot/shutdown - */ - if (radeon_device_is_virtual()) - radeon_pci_remove(pdev); - #if defined(CONFIG_PPC64) || defined(CONFIG_MACH_LOONGSON64) /* * Some adapters need to be suspended before a @@ -613,7 +597,6 @@ static struct pci_driver radeon_kms_pci_driver = { .name = DRIVER_NAME, .id_table = pciidlist, .probe = radeon_pci_probe, - .remove = radeon_pci_remove, .shutdown = radeon_pci_shutdown, .driver.pm = &radeon_pm_ops, }; diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 645e33bf7947..ba1446acd703 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -84,7 +84,6 @@ void radeon_driver_unload_kms(struct drm_device *dev) rdev->agp = NULL; done_free: - kfree(rdev); dev->dev_private = NULL; } diff --git a/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c index 7b613997bb50..727cdf768161 100644 --- a/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c +++ b/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c @@ -361,7 +361,7 @@ static void dw_hdmi_rk3228_setup_hpd(struct dw_hdmi *dw_hdmi, void *data) regmap_write(hdmi->regmap, RK3228_GRF_SOC_CON2, FIELD_PREP_WM16(RK3228_HDMI_SDAIN_MSK, 1) | - FIELD_PREP_WM16(RK3328_HDMI_SCLIN_MSK, 1)); + FIELD_PREP_WM16(RK3228_HDMI_SCLIN_MSK, 1)); } static enum drm_connector_status diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c index b50927a824b4..7ec7bea5e38e 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c @@ -1031,7 +1031,7 @@ static int vop2_plane_atomic_check(struct drm_plane *plane, return format; if (drm_rect_width(src) >> 16 < 4 || drm_rect_height(src) >> 16 < 4 || - drm_rect_width(dest) < 4 || drm_rect_width(dest) < 4) { + drm_rect_width(dest) < 4 || drm_rect_height(dest) < 4) { drm_err(vop2->drm, "Invalid size: %dx%d->%dx%d, min size is 4x4\n", drm_rect_width(src) >> 16, drm_rect_height(src) >> 16, drm_rect_width(dest), drm_rect_height(dest)); diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c index 5a4697f636f2..c8e949f4a568 100644 --- a/drivers/gpu/drm/scheduler/sched_entity.c +++ b/drivers/gpu/drm/scheduler/sched_entity.c @@ -70,6 +70,7 @@ int drm_sched_entity_init(struct drm_sched_entity *entity, entity->guilty = guilty; entity->num_sched_list = num_sched_list; entity->priority = priority; + entity->last_user = current->group_leader; /* * It's perfectly valid to initialize an entity without having a valid * scheduler attached. It's just not valid to use the scheduler before it @@ -302,7 +303,7 @@ long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout) /* For a killed process disallow further enqueueing of jobs. */ last_user = cmpxchg(&entity->last_user, current->group_leader, NULL); - if ((!last_user || last_user == current->group_leader) && + if (last_user == current->group_leader && (current->flags & PF_EXITING) && (current->exit_code == SIGKILL)) drm_sched_entity_kill(entity); @@ -552,10 +553,11 @@ void drm_sched_entity_select_rq(struct drm_sched_entity *entity) drm_sched_rq_remove_entity(entity->rq, entity); entity->rq = rq; } - spin_unlock(&entity->lock); if (entity->num_sched_list == 1) entity->sched_list = NULL; + + spin_unlock(&entity->lock); } /** diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 46119aacb809..c39f0245e3a9 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -965,13 +965,14 @@ int drm_sched_job_add_resv_dependencies(struct drm_sched_job *job, dma_resv_assert_held(resv); dma_resv_for_each_fence(&cursor, resv, usage, fence) { - /* Make sure to grab an additional ref on the added fence */ - dma_fence_get(fence); - ret = drm_sched_job_add_dependency(job, fence); - if (ret) { - dma_fence_put(fence); + /* + * As drm_sched_job_add_dependency always consumes the fence + * reference (even when it fails), and dma_resv_for_each_fence + * is not obtaining one, we need to grab one before calling. + */ + ret = drm_sched_job_add_dependency(job, dma_fence_get(fence)); + if (ret) return ret; - } } return 0; } diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h index 06cb6b02ec64..51f2a03847f9 100644 --- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h +++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h @@ -342,6 +342,7 @@ #define POWERGATE_ENABLE XE_REG(0xa210) #define RENDER_POWERGATE_ENABLE REG_BIT(0) #define MEDIA_POWERGATE_ENABLE REG_BIT(1) +#define MEDIA_SAMPLERS_POWERGATE_ENABLE REG_BIT(2) #define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n)) #define VDN_MFXVDENC_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n)) diff --git a/drivers/gpu/drm/xe/tests/xe_pci.c b/drivers/gpu/drm/xe/tests/xe_pci.c index 69e2840c7ef0..663a79ec960d 100644 --- a/drivers/gpu/drm/xe/tests/xe_pci.c +++ b/drivers/gpu/drm/xe/tests/xe_pci.c @@ -66,6 +66,7 @@ KUNIT_ARRAY_PARAM(platform, cases, xe_pci_fake_data_desc); /** * xe_pci_fake_data_gen_params - Generate struct xe_pci_fake_data parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -242,6 +243,7 @@ KUNIT_ARRAY_PARAM(pci_id, pciidlist, xe_pci_id_kunit_desc); /** * xe_pci_graphics_ip_gen_param - Generate graphics struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -266,6 +268,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_graphics_ip_gen_param); /** * xe_pci_media_ip_gen_param - Generate media struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -290,6 +293,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_media_ip_gen_param); /** * xe_pci_id_gen_param - Generate struct pci_device_id parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -376,6 +380,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_fake_device_init); /** * xe_pci_live_device_gen_param - Helper to iterate Xe devices as KUnit parameters + * @test: test context object * @prev: the previously returned value, or NULL for the first iteration * @desc: the buffer for a parameter name * diff --git a/drivers/gpu/drm/xe/xe_bo_evict.c b/drivers/gpu/drm/xe/xe_bo_evict.c index d5dbc51e8612..bc5b4c5fab81 100644 --- a/drivers/gpu/drm/xe/xe_bo_evict.c +++ b/drivers/gpu/drm/xe/xe_bo_evict.c @@ -182,7 +182,6 @@ int xe_bo_evict_all(struct xe_device *xe) static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) { - struct xe_device *xe = xe_bo_device(bo); int ret; ret = xe_bo_restore_pinned(bo); @@ -201,13 +200,6 @@ static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) } } - /* - * We expect validate to trigger a move VRAM and our move code - * should setup the iosys map. - */ - xe_assert(xe, !(bo->flags & XE_BO_FLAG_PINNED_LATE_RESTORE) || - !iosys_map_is_null(&bo->vmap)); - return 0; } diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index 2883b39c9b37..34d33965eac2 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -1070,7 +1070,7 @@ void xe_device_l2_flush(struct xe_device *xe) spin_lock(>->global_invl_lock); xe_mmio_write32(>->mmio, XE2_GLOBAL_INVAL, 0x1); - if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 500, NULL, true)) + if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 1000, NULL, true)) xe_gt_err_once(gt, "Global invalidation timeout\n"); spin_unlock(>->global_invl_lock); diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c index 7fdd0a97a628..5edc0cad47e2 100644 --- a/drivers/gpu/drm/xe/xe_ggtt.c +++ b/drivers/gpu/drm/xe/xe_ggtt.c @@ -292,6 +292,9 @@ int xe_ggtt_init_early(struct xe_ggtt *ggtt) ggtt->pt_ops = &xelp_pt_ops; ggtt->wq = alloc_workqueue("xe-ggtt-wq", 0, WQ_MEM_RECLAIM); + if (!ggtt->wq) + return -ENOMEM; + __xe_ggtt_init_early(ggtt, xe_wopcm_size(xe)); err = drmm_add_action_or_reset(&xe->drm, ggtt_fini_early, ggtt); diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index 3e0ad7e5b5df..6d3db5e55d98 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -813,12 +813,16 @@ static int gt_reset(struct xe_gt *gt) unsigned int fw_ref; int err; - if (xe_device_wedged(gt_to_xe(gt))) - return -ECANCELED; + if (xe_device_wedged(gt_to_xe(gt))) { + err = -ECANCELED; + goto err_pm_put; + } /* We only support GT resets with GuC submission */ - if (!xe_device_uc_enabled(gt_to_xe(gt))) - return -ENODEV; + if (!xe_device_uc_enabled(gt_to_xe(gt))) { + err = -ENODEV; + goto err_pm_put; + } xe_gt_info(gt, "reset started\n"); @@ -826,8 +830,6 @@ static int gt_reset(struct xe_gt *gt) if (!err) xe_gt_warn(gt, "reset block failed to get lifted"); - xe_pm_runtime_get(gt_to_xe(gt)); - if (xe_fault_inject_gt_reset()) { err = -ECANCELED; goto err_fail; @@ -874,6 +876,7 @@ err_fail: xe_gt_err(gt, "reset failed (%pe)\n", ERR_PTR(err)); xe_device_declare_wedged(gt_to_xe(gt)); +err_pm_put: xe_pm_runtime_put(gt_to_xe(gt)); return err; @@ -895,7 +898,9 @@ void xe_gt_reset_async(struct xe_gt *gt) return; xe_gt_info(gt, "reset queued\n"); - queue_work(gt->ordered_wq, >->reset.worker); + xe_pm_runtime_get_noresume(gt_to_xe(gt)); + if (!queue_work(gt->ordered_wq, >->reset.worker)) + xe_pm_runtime_put(gt_to_xe(gt)); } void xe_gt_suspend_prepare(struct xe_gt *gt) diff --git a/drivers/gpu/drm/xe/xe_gt_idle.c b/drivers/gpu/drm/xe/xe_gt_idle.c index f8950a52d0a4..bdc9d9877ec4 100644 --- a/drivers/gpu/drm/xe/xe_gt_idle.c +++ b/drivers/gpu/drm/xe/xe_gt_idle.c @@ -124,6 +124,9 @@ void xe_gt_idle_enable_pg(struct xe_gt *gt) if (xe_gt_is_main_type(gt)) gtidle->powergate_enable |= RENDER_POWERGATE_ENABLE; + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + gtidle->powergate_enable |= MEDIA_SAMPLERS_POWERGATE_ENABLE; + if (xe->info.platform != XE_DG1) { for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) { if ((gt->info.engine_mask & BIT(i))) @@ -246,6 +249,11 @@ int xe_gt_idle_pg_print(struct xe_gt *gt, struct drm_printer *p) drm_printf(p, "Media Slice%d Power Gate Status: %s\n", n, str_up_down(pg_status & media_slices[n].status_bit)); } + + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + drm_printf(p, "Media Samplers Power Gating Enabled: %s\n", + str_yes_no(pg_enabled & MEDIA_SAMPLERS_POWERGATE_ENABLE)); + return 0; } diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 53024eb5670b..94ed8159496f 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -44,6 +44,7 @@ #include "xe_ring_ops_types.h" #include "xe_sched_job.h" #include "xe_trace.h" +#include "xe_uc_fw.h" #include "xe_vm.h" static struct xe_guc * @@ -1489,7 +1490,17 @@ static void __guc_exec_queue_process_msg_cleanup(struct xe_sched_msg *msg) xe_gt_assert(guc_to_gt(guc), !(q->flags & EXEC_QUEUE_FLAG_PERMANENT)); trace_xe_exec_queue_cleanup_entity(q); - if (exec_queue_registered(q)) + /* + * Expected state transitions for cleanup: + * - If the exec queue is registered and GuC firmware is running, we must first + * disable scheduling and deregister the queue to ensure proper teardown and + * resource release in the GuC, then destroy the exec queue on driver side. + * - If the GuC is already stopped (e.g., during driver unload or GPU reset), + * we cannot expect a response for the deregister request. In this case, + * it is safe to directly destroy the exec queue on driver side, as the GuC + * will not process further requests and all resources must be cleaned up locally. + */ + if (exec_queue_registered(q) && xe_uc_fw_is_running(&guc->fw)) disable_scheduling_deregister(guc, q); else __guc_exec_queue_destroy(guc, q); diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index 1d667fa36cf3..a36ce7dce8cc 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -434,7 +434,7 @@ int xe_migrate_init(struct xe_migrate *m) err = xe_migrate_lock_prepare_vm(tile, m, vm); if (err) - return err; + goto err_out; if (xe->info.has_usm) { struct xe_hw_engine *hwe = xe_gt_hw_engine(primary_gt, @@ -2113,7 +2113,9 @@ int xe_migrate_access_memory(struct xe_migrate *m, struct xe_bo *bo, if (current_bytes & ~PAGE_MASK) { int pitch = 4; - current_bytes = min_t(int, current_bytes, S16_MAX * pitch); + current_bytes = min_t(int, current_bytes, + round_down(S16_MAX * pitch, + XE_CACHELINE_BYTES)); } __fence = xe_migrate_vram(m, current_bytes, diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c index be91343829dd..9a6df79fc5b6 100644 --- a/drivers/gpu/drm/xe/xe_pci.c +++ b/drivers/gpu/drm/xe/xe_pci.c @@ -867,6 +867,8 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (err) return err; + xe_vram_resize_bar(xe); + err = xe_device_probe_early(xe); /* * In Boot Survivability mode, no drm card is exposed and driver diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c index a1c88f9a6c76..07f96bda638a 100644 --- a/drivers/gpu/drm/xe/xe_pt.c +++ b/drivers/gpu/drm/xe/xe_pt.c @@ -2022,7 +2022,7 @@ static int op_prepare(struct xe_vm *vm, case DRM_GPUVA_OP_MAP: if ((!op->map.immediate && xe_vm_in_fault_mode(vm) && !op->map.invalidate_on_bind) || - op->map.is_cpu_addr_mirror) + (op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) break; err = bind_op_prepare(vm, tile, pt_update_ops, op->map.vma, @@ -2252,7 +2252,7 @@ static void op_commit(struct xe_vm *vm, switch (op->base.op) { case DRM_GPUVA_OP_MAP: if ((!op->map.immediate && xe_vm_in_fault_mode(vm)) || - op->map.is_cpu_addr_mirror) + (op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) break; bind_op_commit(vm, tile, pt_update_ops, op->map.vma, fence, diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c index 7e2db71ff34e..129e7818565c 100644 --- a/drivers/gpu/drm/xe/xe_svm.c +++ b/drivers/gpu/drm/xe/xe_svm.c @@ -302,6 +302,11 @@ static int xe_svm_range_set_default_attr(struct xe_vm *vm, u64 range_start, u64 if (!vma) return -EINVAL; + if (!(vma->gpuva.flags & XE_VMA_MADV_AUTORESET)) { + drm_dbg(&vm->xe->drm, "Skipping madvise reset for vma.\n"); + return 0; + } + if (xe_vma_has_default_mem_attrs(vma)) return 0; @@ -1034,6 +1039,9 @@ retry: if (err) return err; + dpagemap = xe_vma_resolve_pagemap(vma, tile); + if (!dpagemap && !ctx.devmem_only) + ctx.device_private_page_owner = NULL; range = xe_svm_range_find_or_insert(vm, fault_addr, vma, &ctx); if (IS_ERR(range)) @@ -1054,7 +1062,6 @@ retry: range_debug(range, "PAGE FAULT"); - dpagemap = xe_vma_resolve_pagemap(vma, tile); if (--migrate_try_count >= 0 && xe_svm_range_needs_migrate_to_vram(range, vma, !!dpagemap || ctx.devmem_only)) { ktime_t migrate_start = xe_svm_stats_ktime_get(); @@ -1073,7 +1080,17 @@ retry: drm_dbg(&vm->xe->drm, "VRAM allocation failed, falling back to retrying fault, asid=%u, errno=%pe\n", vm->usm.asid, ERR_PTR(err)); - goto retry; + + /* + * In the devmem-only case, mixed mappings may + * be found. The get_pages function will fix + * these up to a single location, allowing the + * page fault handler to make forward progress. + */ + if (ctx.devmem_only) + goto get_pages; + else + goto retry; } else { drm_err(&vm->xe->drm, "VRAM allocation failed, retry count exceeded, asid=%u, errno=%pe\n", @@ -1083,6 +1100,7 @@ retry: } } +get_pages: get_pages_start = xe_svm_stats_ktime_get(); range_debug(range, "GET PAGES"); diff --git a/drivers/gpu/drm/xe/xe_validation.h b/drivers/gpu/drm/xe/xe_validation.h index fec331d791e7..b2d09c596714 100644 --- a/drivers/gpu/drm/xe/xe_validation.h +++ b/drivers/gpu/drm/xe/xe_validation.h @@ -166,10 +166,10 @@ xe_validation_device_init(struct xe_validation_device *val) */ DEFINE_CLASS(xe_validation, struct xe_validation_ctx *, if (_T) xe_validation_ctx_fini(_T);, - ({_ret = xe_validation_ctx_init(_ctx, _val, _exec, _flags); - _ret ? NULL : _ctx; }), + ({*_ret = xe_validation_ctx_init(_ctx, _val, _exec, _flags); + *_ret ? NULL : _ctx; }), struct xe_validation_ctx *_ctx, struct xe_validation_device *_val, - struct drm_exec *_exec, const struct xe_val_flags _flags, int _ret); + struct drm_exec *_exec, const struct xe_val_flags _flags, int *_ret); static inline void *class_xe_validation_lock_ptr(class_xe_validation_t *_T) {return *_T; } #define class_xe_validation_is_conditional true @@ -186,7 +186,7 @@ static inline void *class_xe_validation_lock_ptr(class_xe_validation_t *_T) * exhaustive eviction. */ #define xe_validation_guard(_ctx, _val, _exec, _flags, _ret) \ - scoped_guard(xe_validation, _ctx, _val, _exec, _flags, _ret) \ + scoped_guard(xe_validation, _ctx, _val, _exec, _flags, &_ret) \ drm_exec_until_all_locked(_exec) #endif diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 027e6ce648c5..63c65e3d207b 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -616,6 +616,13 @@ static void xe_vma_ops_incr_pt_update_ops(struct xe_vma_ops *vops, u8 tile_mask, vops->pt_update_ops[i].num_ops += inc_val; } +#define XE_VMA_CREATE_MASK ( \ + XE_VMA_READ_ONLY | \ + XE_VMA_DUMPABLE | \ + XE_VMA_SYSTEM_ALLOCATOR | \ + DRM_GPUVA_SPARSE | \ + XE_VMA_MADV_AUTORESET) + static void xe_vm_populate_rebind(struct xe_vma_op *op, struct xe_vma *vma, u8 tile_mask) { @@ -628,8 +635,7 @@ static void xe_vm_populate_rebind(struct xe_vma_op *op, struct xe_vma *vma, op->base.map.gem.offset = vma->gpuva.gem.offset; op->map.vma = vma; op->map.immediate = true; - op->map.dumpable = vma->gpuva.flags & XE_VMA_DUMPABLE; - op->map.is_null = xe_vma_is_null(vma); + op->map.vma_flags = vma->gpuva.flags & XE_VMA_CREATE_MASK; } static int xe_vm_ops_add_rebind(struct xe_vma_ops *vops, struct xe_vma *vma, @@ -932,11 +938,6 @@ static void xe_vma_free(struct xe_vma *vma) kfree(vma); } -#define VMA_CREATE_FLAG_READ_ONLY BIT(0) -#define VMA_CREATE_FLAG_IS_NULL BIT(1) -#define VMA_CREATE_FLAG_DUMPABLE BIT(2) -#define VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR BIT(3) - static struct xe_vma *xe_vma_create(struct xe_vm *vm, struct xe_bo *bo, u64 bo_offset_or_userptr, @@ -947,11 +948,8 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, struct xe_vma *vma; struct xe_tile *tile; u8 id; - bool read_only = (flags & VMA_CREATE_FLAG_READ_ONLY); - bool is_null = (flags & VMA_CREATE_FLAG_IS_NULL); - bool dumpable = (flags & VMA_CREATE_FLAG_DUMPABLE); - bool is_cpu_addr_mirror = - (flags & VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR); + bool is_null = (flags & DRM_GPUVA_SPARSE); + bool is_cpu_addr_mirror = (flags & XE_VMA_SYSTEM_ALLOCATOR); xe_assert(vm->xe, start < end); xe_assert(vm->xe, end < vm->size); @@ -972,10 +970,6 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, if (!vma) return ERR_PTR(-ENOMEM); - if (is_cpu_addr_mirror) - vma->gpuva.flags |= XE_VMA_SYSTEM_ALLOCATOR; - if (is_null) - vma->gpuva.flags |= DRM_GPUVA_SPARSE; if (bo) vma->gpuva.gem.obj = &bo->ttm.base; } @@ -986,10 +980,7 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, vma->gpuva.vm = &vm->gpuvm; vma->gpuva.va.addr = start; vma->gpuva.va.range = end - start + 1; - if (read_only) - vma->gpuva.flags |= XE_VMA_READ_ONLY; - if (dumpable) - vma->gpuva.flags |= XE_VMA_DUMPABLE; + vma->gpuva.flags = flags; for_each_tile(tile, vm->xe, id) vma->tile_mask |= 0x1 << id; @@ -2272,12 +2263,16 @@ vm_bind_ioctl_ops_create(struct xe_vm *vm, struct xe_vma_ops *vops, if (__op->op == DRM_GPUVA_OP_MAP) { op->map.immediate = flags & DRM_XE_VM_BIND_FLAG_IMMEDIATE; - op->map.read_only = - flags & DRM_XE_VM_BIND_FLAG_READONLY; - op->map.is_null = flags & DRM_XE_VM_BIND_FLAG_NULL; - op->map.is_cpu_addr_mirror = flags & - DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR; - op->map.dumpable = flags & DRM_XE_VM_BIND_FLAG_DUMPABLE; + if (flags & DRM_XE_VM_BIND_FLAG_READONLY) + op->map.vma_flags |= XE_VMA_READ_ONLY; + if (flags & DRM_XE_VM_BIND_FLAG_NULL) + op->map.vma_flags |= DRM_GPUVA_SPARSE; + if (flags & DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR) + op->map.vma_flags |= XE_VMA_SYSTEM_ALLOCATOR; + if (flags & DRM_XE_VM_BIND_FLAG_DUMPABLE) + op->map.vma_flags |= XE_VMA_DUMPABLE; + if (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) + op->map.vma_flags |= XE_VMA_MADV_AUTORESET; op->map.pat_index = pat_index; op->map.invalidate_on_bind = __xe_vm_needs_clear_scratch_pages(vm, flags); @@ -2590,14 +2585,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, .pat_index = op->map.pat_index, }; - flags |= op->map.read_only ? - VMA_CREATE_FLAG_READ_ONLY : 0; - flags |= op->map.is_null ? - VMA_CREATE_FLAG_IS_NULL : 0; - flags |= op->map.dumpable ? - VMA_CREATE_FLAG_DUMPABLE : 0; - flags |= op->map.is_cpu_addr_mirror ? - VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0; + flags |= op->map.vma_flags & XE_VMA_CREATE_MASK; vma = new_vma(vm, &op->base.map, &default_attr, flags); @@ -2606,7 +2594,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, op->map.vma = vma; if (((op->map.immediate || !xe_vm_in_fault_mode(vm)) && - !op->map.is_cpu_addr_mirror) || + !(op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) || op->map.invalidate_on_bind) xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask, 1); @@ -2637,18 +2625,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, op->remap.start = xe_vma_start(old); op->remap.range = xe_vma_size(old); - flags |= op->base.remap.unmap->va->flags & - XE_VMA_READ_ONLY ? - VMA_CREATE_FLAG_READ_ONLY : 0; - flags |= op->base.remap.unmap->va->flags & - DRM_GPUVA_SPARSE ? - VMA_CREATE_FLAG_IS_NULL : 0; - flags |= op->base.remap.unmap->va->flags & - XE_VMA_DUMPABLE ? - VMA_CREATE_FLAG_DUMPABLE : 0; - flags |= xe_vma_is_cpu_addr_mirror(old) ? - VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0; - + flags |= op->base.remap.unmap->va->flags & XE_VMA_CREATE_MASK; if (op->base.remap.prev) { vma = new_vma(vm, op->base.remap.prev, &old->attr, flags); @@ -2832,7 +2809,7 @@ static void vm_bind_ioctl_ops_unwind(struct xe_vm *vm, } static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, - bool validate) + bool res_evict, bool validate) { struct xe_bo *bo = xe_vma_bo(vma); struct xe_vm *vm = xe_vma_vm(vma); @@ -2843,7 +2820,8 @@ static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, err = drm_exec_lock_obj(exec, &bo->ttm.base); if (!err && validate) err = xe_bo_validate(bo, vm, - !xe_vm_in_preempt_fence_mode(vm), exec); + !xe_vm_in_preempt_fence_mode(vm) && + res_evict, exec); } return err; @@ -2913,14 +2891,23 @@ static int prefetch_ranges(struct xe_vm *vm, struct xe_vma_op *op) } static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, - struct xe_vma_op *op) + struct xe_vma_ops *vops, struct xe_vma_op *op) { int err = 0; + bool res_evict; + + /* + * We only allow evicting a BO within the VM if it is not part of an + * array of binds, as an array of binds can evict another BO within the + * bind. + */ + res_evict = !(vops->flags & XE_VMA_OPS_ARRAY_OF_BINDS); switch (op->base.op) { case DRM_GPUVA_OP_MAP: if (!op->map.invalidate_on_bind) err = vma_lock_and_validate(exec, op->map.vma, + res_evict, !xe_vm_in_fault_mode(vm) || op->map.immediate); break; @@ -2931,11 +2918,13 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.remap.unmap->va), - false); + res_evict, false); if (!err && op->remap.prev) - err = vma_lock_and_validate(exec, op->remap.prev, true); + err = vma_lock_and_validate(exec, op->remap.prev, + res_evict, true); if (!err && op->remap.next) - err = vma_lock_and_validate(exec, op->remap.next, true); + err = vma_lock_and_validate(exec, op->remap.next, + res_evict, true); break; case DRM_GPUVA_OP_UNMAP: err = check_ufence(gpuva_to_vma(op->base.unmap.va)); @@ -2944,7 +2933,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.unmap.va), - false); + res_evict, false); break; case DRM_GPUVA_OP_PREFETCH: { @@ -2959,7 +2948,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.prefetch.va), - false); + res_evict, false); if (!err && !xe_vma_has_no_bo(vma)) err = xe_bo_migrate(xe_vma_bo(vma), region_to_mem_type[region], @@ -3005,7 +2994,7 @@ static int vm_bind_ioctl_ops_lock_and_prep(struct drm_exec *exec, return err; list_for_each_entry(op, &vops->list, link) { - err = op_lock_and_prep(exec, vm, op); + err = op_lock_and_prep(exec, vm, vops, op); if (err) return err; } @@ -3267,7 +3256,8 @@ ALLOW_ERROR_INJECTION(vm_bind_ioctl_ops_execute, ERRNO); DRM_XE_VM_BIND_FLAG_NULL | \ DRM_XE_VM_BIND_FLAG_DUMPABLE | \ DRM_XE_VM_BIND_FLAG_CHECK_PXP | \ - DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR) + DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR | \ + DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) #ifdef TEST_VM_OPS_ERROR #define SUPPORTED_FLAGS (SUPPORTED_FLAGS_STUB | FORCE_OP_ERROR) @@ -3382,7 +3372,9 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm, XE_IOCTL_DBG(xe, (prefetch_region != DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC && !(BIT(prefetch_region) & xe->info.mem_region_mask))) || XE_IOCTL_DBG(xe, obj && - op == DRM_XE_VM_BIND_OP_UNMAP)) { + op == DRM_XE_VM_BIND_OP_UNMAP) || + XE_IOCTL_DBG(xe, (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) && + (!is_cpu_addr_mirror || op != DRM_XE_VM_BIND_OP_MAP))) { err = -EINVAL; goto free_bind_ops; } @@ -3638,6 +3630,8 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) } xe_vma_ops_init(&vops, vm, q, syncs, num_syncs); + if (args->num_binds > 1) + vops.flags |= XE_VMA_OPS_ARRAY_OF_BINDS; for (i = 0; i < args->num_binds; ++i) { u64 range = bind_ops[i].range; u64 addr = bind_ops[i].addr; @@ -4198,7 +4192,7 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, struct xe_vma_ops vops; struct drm_gpuva_ops *ops = NULL; struct drm_gpuva_op *__op; - bool is_cpu_addr_mirror = false; + unsigned int vma_flags = 0; bool remap_op = false; struct xe_vma_mem_attr tmp_attr; u16 default_pat; @@ -4228,15 +4222,17 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, vma = gpuva_to_vma(op->base.unmap.va); XE_WARN_ON(!xe_vma_has_default_mem_attrs(vma)); default_pat = vma->attr.default_pat_index; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_REMAP) { vma = gpuva_to_vma(op->base.remap.unmap->va); default_pat = vma->attr.default_pat_index; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_MAP) { - op->map.is_cpu_addr_mirror = true; + op->map.vma_flags |= vma_flags & XE_VMA_CREATE_MASK; op->map.pat_index = default_pat; } } else { @@ -4245,11 +4241,7 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, xe_assert(vm->xe, !remap_op); xe_assert(vm->xe, xe_vma_has_no_bo(vma)); remap_op = true; - - if (xe_vma_is_cpu_addr_mirror(vma)) - is_cpu_addr_mirror = true; - else - is_cpu_addr_mirror = false; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_MAP) { @@ -4258,10 +4250,10 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, /* * In case of madvise ops DRM_GPUVA_OP_MAP is * always after DRM_GPUVA_OP_REMAP, so ensure - * we assign op->map.is_cpu_addr_mirror true - * if REMAP is for xe_vma_is_cpu_addr_mirror vma + * to propagate the flags from the vma we're + * unmapping. */ - op->map.is_cpu_addr_mirror = is_cpu_addr_mirror; + op->map.vma_flags |= vma_flags & XE_VMA_CREATE_MASK; } } print_op(vm->xe, __op); diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index da39940501d8..d6e2a0fdd4b3 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -46,6 +46,7 @@ struct xe_vm_pgtable_update_op; #define XE_VMA_PTE_COMPACT (DRM_GPUVA_USERBITS << 7) #define XE_VMA_DUMPABLE (DRM_GPUVA_USERBITS << 8) #define XE_VMA_SYSTEM_ALLOCATOR (DRM_GPUVA_USERBITS << 9) +#define XE_VMA_MADV_AUTORESET (DRM_GPUVA_USERBITS << 10) /** * struct xe_vma_mem_attr - memory attributes associated with vma @@ -345,17 +346,10 @@ struct xe_vm { struct xe_vma_op_map { /** @vma: VMA to map */ struct xe_vma *vma; + unsigned int vma_flags; /** @immediate: Immediate bind */ bool immediate; /** @read_only: Read only */ - bool read_only; - /** @is_null: is NULL binding */ - bool is_null; - /** @is_cpu_addr_mirror: is CPU address mirror binding */ - bool is_cpu_addr_mirror; - /** @dumpable: whether BO is dumped on GPU hang */ - bool dumpable; - /** @invalidate: invalidate the VMA before bind */ bool invalidate_on_bind; /** @pat_index: The pat index to use for this operation. */ u16 pat_index; @@ -476,6 +470,7 @@ struct xe_vma_ops { /** @flag: signify the properties within xe_vma_ops*/ #define XE_VMA_OPS_FLAG_HAS_SVM_PREFETCH BIT(0) #define XE_VMA_OPS_FLAG_MADVISE BIT(1) +#define XE_VMA_OPS_ARRAY_OF_BINDS BIT(2) u32 flags; #ifdef TEST_VM_OPS_ERROR /** @inject_error: inject error to test error handling */ diff --git a/drivers/gpu/drm/xe/xe_vram.c b/drivers/gpu/drm/xe/xe_vram.c index b44ebf50fedb..652df7a5f4f6 100644 --- a/drivers/gpu/drm/xe/xe_vram.c +++ b/drivers/gpu/drm/xe/xe_vram.c @@ -26,15 +26,35 @@ #define BAR_SIZE_SHIFT 20 -static void -_resize_bar(struct xe_device *xe, int resno, resource_size_t size) +/* + * Release all the BARs that could influence/block LMEMBAR resizing, i.e. + * assigned IORESOURCE_MEM_64 BARs + */ +static void release_bars(struct pci_dev *pdev) +{ + struct resource *res; + int i; + + pci_dev_for_each_resource(pdev, res, i) { + /* Resource already un-assigned, do not reset it */ + if (!res->parent) + continue; + + /* No need to release unrelated BARs */ + if (!(res->flags & IORESOURCE_MEM_64)) + continue; + + pci_release_resource(pdev, i); + } +} + +static void resize_bar(struct xe_device *xe, int resno, resource_size_t size) { struct pci_dev *pdev = to_pci_dev(xe->drm.dev); int bar_size = pci_rebar_bytes_to_size(size); int ret; - if (pci_resource_len(pdev, resno)) - pci_release_resource(pdev, resno); + release_bars(pdev); ret = pci_resize_resource(pdev, resno, bar_size); if (ret) { @@ -50,7 +70,7 @@ _resize_bar(struct xe_device *xe, int resno, resource_size_t size) * if force_vram_bar_size is set, attempt to set to the requested size * else set to maximum possible size */ -static void resize_vram_bar(struct xe_device *xe) +void xe_vram_resize_bar(struct xe_device *xe) { int force_vram_bar_size = xe_modparam.force_vram_bar_size; struct pci_dev *pdev = to_pci_dev(xe->drm.dev); @@ -119,7 +139,7 @@ static void resize_vram_bar(struct xe_device *xe) pci_read_config_dword(pdev, PCI_COMMAND, &pci_cmd); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd & ~PCI_COMMAND_MEMORY); - _resize_bar(xe, LMEM_BAR, rebar_size); + resize_bar(xe, LMEM_BAR, rebar_size); pci_assign_unassigned_bus_resources(pdev->bus); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd); @@ -148,8 +168,6 @@ static int determine_lmem_bar_size(struct xe_device *xe, struct xe_vram_region * return -ENXIO; } - resize_vram_bar(xe); - lmem_bar->io_start = pci_resource_start(pdev, LMEM_BAR); lmem_bar->io_size = pci_resource_len(pdev, LMEM_BAR); if (!lmem_bar->io_size) diff --git a/drivers/gpu/drm/xe/xe_vram.h b/drivers/gpu/drm/xe/xe_vram.h index 72860f714fc6..13505cfb184d 100644 --- a/drivers/gpu/drm/xe/xe_vram.h +++ b/drivers/gpu/drm/xe/xe_vram.h @@ -11,6 +11,7 @@ struct xe_device; struct xe_vram_region; +void xe_vram_resize_bar(struct xe_device *xe); int xe_vram_probe(struct xe_device *xe); struct xe_vram_region *xe_vram_region_alloc(struct xe_device *xe, u8 id, u32 placement); |