diff options
| author | Dave Airlie <airlied@redhat.com> | 2017-12-07 06:28:22 +1000 |
|---|---|---|
| committer | Dave Airlie <airlied@redhat.com> | 2017-12-07 06:28:22 +1000 |
| commit | 9c606cd4117a3c45e04a6616b1a0dbeb18eeee62 (patch) | |
| tree | aa6c1db29e1a3f687c81fa03aecd24992a76e993 /drivers/gpu/drm/amd/amdgpu | |
| parent | c5dd52f653fa74f8f4771425c6db33609ad21258 (diff) | |
| parent | 3997eea57caf542e9327df9b6bb2882a57c4c421 (diff) | |
Merge branch 'drm-next-4.16' of git://people.freedesktop.org/~agd5f/linux into drm-next
First feature request for 4.16. Highlights:
- RV and Vega header cleanups
- TTM operation context support
- 48 bit GPUVM fixes for Vega/RV
- More smatch fixes
- ECC support for vega10
- Resizeable BAR support
- Multi-display sync support in DC
- SR-IOV fixes
- Various scheduler improvements
- GPU reset fixes and vram lost tracking
- Clean up DC/powerplay interfaces
- DCN display fixes
- Various DC fixes
* 'drm-next-4.16' of git://people.freedesktop.org/~agd5f/linux: (291 commits)
drm/radeon: Use drm_fb_helper_lastclose() and _poll_changed()
drm/amdgpu: Use drm_fb_helper_lastclose() and _poll_changed()
drm/amd/display: Use drm_fb_helper_poll_changed()
drm/ttm: swap consecutive allocated pooled pages v4
drm/amdgpu: fix amdgpu_sync_resv v2
drm/ttm: swap consecutive allocated cached pages v3
drm/amd/amdgpu: set gtt size according to system memory size only
drm/amdgpu: Get rid of dep_sync as a seperate object.
drm/amdgpu: allow specifying vm_block_size for multi level PDs v2
drm/amdgpu: move validation of the VM size into the VM code
drm/amdgpu: allow non pot VM size values
drm/amdgpu: choose number of VM levels based on VM size
drm/amdgpu: unify VM size handling of Vega10 with older generation
drm/amdgpu: fix amdgpu_vm_num_entries
drm/amdgpu: fix VM PD addr shift
drm/amdgpu: correct vce4.0 fw config for SRIOV (V2)
drm/amd/display: Don't call dm_log_to_buffer directly in dc_conn_log
drm/amd/display: Add dm_logger_append_va API
drm/ttm: Use a static string instead of an array of char *
drm/amd/display: remove usage of legacy_cursor_update
...
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
81 files changed, 1633 insertions, 1247 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 0b14b5373783..5e2958a79928 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -47,6 +47,8 @@ #include <drm/amdgpu_drm.h> #include <kgd_kfd_interface.h> +#include "dm_pp_interface.h" +#include "kgd_pp_interface.h" #include "amd_shared.h" #include "amdgpu_mode.h" @@ -59,7 +61,6 @@ #include "amdgpu_sync.h" #include "amdgpu_ring.h" #include "amdgpu_vm.h" -#include "amd_powerplay.h" #include "amdgpu_dpm.h" #include "amdgpu_acp.h" #include "amdgpu_uvd.h" @@ -67,11 +68,11 @@ #include "amdgpu_vcn.h" #include "amdgpu_mn.h" #include "amdgpu_dm.h" - #include "gpu_scheduler.h" #include "amdgpu_virt.h" #include "amdgpu_gart.h" + /* * Modules parameters. */ @@ -177,6 +178,10 @@ extern int amdgpu_cik_support; #define CIK_CURSOR_WIDTH 128 #define CIK_CURSOR_HEIGHT 128 +/* GPU RESET flags */ +#define AMDGPU_RESET_INFO_VRAM_LOST (1 << 0) +#define AMDGPU_RESET_INFO_FULLRESET (1 << 1) + struct amdgpu_device; struct amdgpu_ib; struct amdgpu_cs_parser; @@ -735,6 +740,7 @@ struct amdgpu_ctx { struct amdgpu_device *adev; struct amdgpu_queue_mgr queue_mgr; unsigned reset_counter; + unsigned reset_counter_query; uint32_t vram_lost_counter; spinlock_t ring_lock; struct dma_fence **fences; @@ -743,6 +749,7 @@ struct amdgpu_ctx { enum amd_sched_priority init_priority; enum amd_sched_priority override_priority; struct mutex lock; + atomic_t guilty; }; struct amdgpu_ctx_mgr { @@ -1114,7 +1121,6 @@ struct amdgpu_job { struct amdgpu_vm *vm; struct amdgpu_ring *ring; struct amdgpu_sync sync; - struct amdgpu_sync dep_sync; struct amdgpu_sync sched_sync; struct amdgpu_ib *ibs; struct dma_fence *fence; /* the hw fence */ @@ -1405,6 +1411,7 @@ struct amdgpu_fw_vram_usage { }; int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev); +void amdgpu_fw_reserve_vram_fini(struct amdgpu_device *adev); /* * CGS @@ -1421,6 +1428,13 @@ typedef void (*amdgpu_wreg_t)(struct amdgpu_device*, uint32_t, uint32_t); typedef uint32_t (*amdgpu_block_rreg_t)(struct amdgpu_device*, uint32_t, uint32_t); typedef void (*amdgpu_block_wreg_t)(struct amdgpu_device*, uint32_t, uint32_t, uint32_t); +struct amd_powerplay { + struct cgs_device *cgs_device; + void *pp_handle; + const struct amd_ip_funcs *ip_funcs; + const struct amd_pm_funcs *pp_funcs; +}; + #define AMDGPU_RESET_MAGIC_NUM 64 struct amdgpu_device { struct device *dev; @@ -1616,9 +1630,6 @@ struct amdgpu_device { /* link all shadow bo */ struct list_head shadow_list; struct mutex shadow_list_lock; - /* link all gtt */ - spinlock_t gtt_list_lock; - struct list_head gtt_list; /* keep an lru list of rings by HW IP */ struct list_head ring_lru_list; spinlock_t ring_lru_list_lock; @@ -1629,7 +1640,8 @@ struct amdgpu_device { /* record last mm index being written through WREG32*/ unsigned long last_mm_index; - bool in_sriov_reset; + bool in_gpu_reset; + struct mutex lock_reset; }; static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev) @@ -1823,7 +1835,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_psp_check_fw_loading_status(adev, i) (adev)->firmware.funcs->check_fw_loading_status((adev), (i)) /* Common functions */ -int amdgpu_gpu_reset(struct amdgpu_device *adev); +int amdgpu_gpu_recover(struct amdgpu_device *adev, struct amdgpu_job* job); bool amdgpu_need_backup(struct amdgpu_device *adev); void amdgpu_pci_config_reset(struct amdgpu_device *adev); bool amdgpu_need_post(struct amdgpu_device *adev); @@ -1835,6 +1847,7 @@ void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain); bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo); void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base); void amdgpu_gart_location(struct amdgpu_device *adev, struct amdgpu_mc *mc); +int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev); void amdgpu_ttm_set_active_vram_size(struct amdgpu_device *adev, u64 size); int amdgpu_ttm_init(struct amdgpu_device *adev); void amdgpu_ttm_fini(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 5432af39a674..c70cda04dbfb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -85,7 +85,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) kfd2kgd = amdgpu_amdkfd_gfx_8_0_get_functions(); break; default: - dev_info(adev->dev, "kfd not supported on this ASIC\n"); + dev_dbg(adev->dev, "kfd not supported on this ASIC\n"); return; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index f450b69323fa..39f4d0df1ada 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c @@ -690,12 +690,12 @@ int amdgpu_atombios_get_clock_info(struct amdgpu_device *adev) le32_to_cpu(firmware_info->info_21.ulDefaultDispEngineClkFreq); /* set a reasonable default for DP */ if (adev->clock.default_dispclk < 53900) { - DRM_INFO("Changing default dispclk from %dMhz to 600Mhz\n", - adev->clock.default_dispclk / 100); + DRM_DEBUG("Changing default dispclk from %dMhz to 600Mhz\n", + adev->clock.default_dispclk / 100); adev->clock.default_dispclk = 60000; } else if (adev->clock.default_dispclk <= 60000) { - DRM_INFO("Changing default dispclk from %dMhz to 625Mhz\n", - adev->clock.default_dispclk / 100); + DRM_DEBUG("Changing default dispclk from %dMhz to 625Mhz\n", + adev->clock.default_dispclk / 100); adev->clock.default_dispclk = 62500; } adev->clock.dp_extclk = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c index f2b72c7c6857..85d2149b9dbe 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c @@ -948,7 +948,6 @@ static int amdgpu_cgs_get_active_displays_info(struct cgs_device *cgs_device, (amdgpu_crtc->v_border * 2); mode_info->vblank_time_us = vblank_lines * line_time_us; mode_info->refresh_rate = drm_mode_vrefresh(&amdgpu_crtc->hw_mode); - mode_info->ref_clock = adev->clock.spll.reference_freq; mode_info = NULL; } } @@ -958,7 +957,6 @@ static int amdgpu_cgs_get_active_displays_info(struct cgs_device *cgs_device, if (mode_info != NULL) { mode_info->vblank_time_us = adev->pm.pm_display_cfg.min_vblank_time; mode_info->refresh_rate = adev->pm.pm_display_cfg.vrefresh; - mode_info->ref_clock = adev->clock.spll.reference_freq; } } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 57abf7abd7a9..4cea9ab237ac 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -90,6 +90,12 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) goto free_chunk; } + /* skip guilty context job */ + if (atomic_read(&p->ctx->guilty) == 1) { + ret = -ECANCELED; + goto free_chunk; + } + mutex_lock(&p->ctx->lock); /* get chunks */ @@ -337,7 +343,7 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p, struct amdgpu_bo *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); - u64 initial_bytes_moved, bytes_moved; + struct ttm_operation_ctx ctx = { true, false }; uint32_t domain; int r; @@ -367,15 +373,13 @@ static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p, retry: amdgpu_ttm_placement_from_domain(bo, domain); - initial_bytes_moved = atomic64_read(&adev->num_bytes_moved); - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); - bytes_moved = atomic64_read(&adev->num_bytes_moved) - - initial_bytes_moved; - p->bytes_moved += bytes_moved; + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + + p->bytes_moved += ctx.bytes_moved; if (adev->mc.visible_vram_size < adev->mc.real_vram_size && bo->tbo.mem.mem_type == TTM_PL_VRAM && bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT) - p->bytes_moved_vis += bytes_moved; + p->bytes_moved_vis += ctx.bytes_moved; if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) { domain = bo->allowed_domains; @@ -390,6 +394,7 @@ static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p, struct amdgpu_bo *validated) { uint32_t domain = validated->allowed_domains; + struct ttm_operation_ctx ctx = { true, false }; int r; if (!p->evictable) @@ -431,7 +436,7 @@ static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p, bo->tbo.mem.mem_type == TTM_PL_VRAM && bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT; initial_bytes_moved = atomic64_read(&adev->num_bytes_moved); - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); bytes_moved = atomic64_read(&adev->num_bytes_moved) - initial_bytes_moved; p->bytes_moved += bytes_moved; @@ -470,6 +475,7 @@ static int amdgpu_cs_validate(void *param, struct amdgpu_bo *bo) static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p, struct list_head *validated) { + struct ttm_operation_ctx ctx = { true, false }; struct amdgpu_bo_list_entry *lobj; int r; @@ -487,8 +493,7 @@ static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p, lobj->user_pages) { amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, - false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (r) return r; amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, @@ -678,7 +683,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, if (!r && p->uf_entry.robj) { struct amdgpu_bo *uf = p->uf_entry.robj; - r = amdgpu_ttm_bind(&uf->tbo, &uf->tbo.mem); + r = amdgpu_ttm_alloc_gart(&uf->tbo); p->job->uf_addr += amdgpu_bo_gpu_offset(uf); } @@ -781,7 +786,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p) return r; r = amdgpu_sync_fence(adev, &p->job->sync, - fpriv->prt_va->last_pt_update); + fpriv->prt_va->last_pt_update, false); if (r) return r; @@ -795,7 +800,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p) return r; f = bo_va->last_pt_update; - r = amdgpu_sync_fence(adev, &p->job->sync, f); + r = amdgpu_sync_fence(adev, &p->job->sync, f, false); if (r) return r; } @@ -818,7 +823,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p) return r; f = bo_va->last_pt_update; - r = amdgpu_sync_fence(adev, &p->job->sync, f); + r = amdgpu_sync_fence(adev, &p->job->sync, f, false); if (r) return r; } @@ -829,7 +834,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p) if (r) return r; - r = amdgpu_sync_fence(adev, &p->job->sync, vm->last_update); + r = amdgpu_sync_fence(adev, &p->job->sync, vm->last_update, false); if (r) return r; @@ -865,8 +870,8 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping *m; struct amdgpu_bo *aobj = NULL; struct amdgpu_cs_chunk *chunk; + uint64_t offset, va_start; struct amdgpu_ib *ib; - uint64_t offset; uint8_t *kptr; chunk = &p->chunks[i]; @@ -876,14 +881,14 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev, if (chunk->chunk_id != AMDGPU_CHUNK_ID_IB) continue; - r = amdgpu_cs_find_mapping(p, chunk_ib->va_start, - &aobj, &m); + va_start = chunk_ib->va_start & AMDGPU_VA_HOLE_MASK; + r = amdgpu_cs_find_mapping(p, va_start, &aobj, &m); if (r) { DRM_ERROR("IB va_start is invalid\n"); return r; } - if ((chunk_ib->va_start + chunk_ib->ib_bytes) > + if ((va_start + chunk_ib->ib_bytes) > (m->last + 1) * AMDGPU_GPU_PAGE_SIZE) { DRM_ERROR("IB va_start+ib_bytes is invalid\n"); return -EINVAL; @@ -896,7 +901,7 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev, } offset = m->start * AMDGPU_GPU_PAGE_SIZE; - kptr += chunk_ib->va_start - offset; + kptr += va_start - offset; memcpy(ib->ptr, kptr, chunk_ib->ib_bytes); amdgpu_bo_kunmap(aobj); @@ -1033,8 +1038,8 @@ static int amdgpu_cs_process_fence_dep(struct amdgpu_cs_parser *p, amdgpu_ctx_put(ctx); return r; } else if (fence) { - r = amdgpu_sync_fence(p->adev, &p->job->sync, - fence); + r = amdgpu_sync_fence(p->adev, &p->job->sync, fence, + true); dma_fence_put(fence); amdgpu_ctx_put(ctx); if (r) @@ -1053,7 +1058,7 @@ static int amdgpu_syncobj_lookup_and_add_to_sync(struct amdgpu_cs_parser *p, if (r) return r; - r = amdgpu_sync_fence(p->adev, &p->job->sync, fence); + r = amdgpu_sync_fence(p->adev, &p->job->sync, fence, true); dma_fence_put(fence); return r; @@ -1194,11 +1199,10 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, job->uf_sequence = seq; amdgpu_job_free_resources(job); - amdgpu_ring_priority_get(job->ring, - amd_sched_get_job_priority(&job->base)); + amdgpu_ring_priority_get(job->ring, job->base.s_priority); trace_amdgpu_cs_ioctl(job); - amd_sched_entity_push_job(&job->base); + amd_sched_entity_push_job(&job->base, entity); ttm_eu_fence_buffer_objects(&p->ticket, &p->validated, p->fence); amdgpu_mn_unlock(p->mn); @@ -1570,6 +1574,7 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser, struct amdgpu_bo_va_mapping **map) { struct amdgpu_fpriv *fpriv = parser->filp->driver_priv; + struct ttm_operation_ctx ctx = { false, false }; struct amdgpu_vm *vm = &fpriv->vm; struct amdgpu_bo_va_mapping *mapping; int r; @@ -1590,11 +1595,10 @@ int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser, if (!((*bo)->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)) { (*bo)->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS; amdgpu_ttm_placement_from_domain(*bo, (*bo)->allowed_domains); - r = ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, false, - false); + r = ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, &ctx); if (r) return r; } - return amdgpu_ttm_bind(&(*bo)->tbo, &(*bo)->tbo.mem); + return amdgpu_ttm_alloc_gart(&(*bo)->tbo); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index c184468e2b2b..d71dc164b469 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -75,6 +75,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, } ctx->reset_counter = atomic_read(&adev->gpu_reset_counter); + ctx->reset_counter_query = ctx->reset_counter; ctx->vram_lost_counter = atomic_read(&adev->vram_lost_counter); ctx->init_priority = priority; ctx->override_priority = AMD_SCHED_PRIORITY_UNSET; @@ -90,7 +91,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, continue; r = amd_sched_entity_init(&ring->sched, &ctx->rings[i].entity, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, &ctx->guilty); if (r) goto failed; } @@ -216,11 +217,45 @@ static int amdgpu_ctx_query(struct amdgpu_device *adev, /* determine if a GPU reset has occured since the last call */ reset_counter = atomic_read(&adev->gpu_reset_counter); /* TODO: this should ideally return NO, GUILTY, or INNOCENT. */ - if (ctx->reset_counter == reset_counter) + if (ctx->reset_counter_query == reset_counter) out->state.reset_status = AMDGPU_CTX_NO_RESET; else out->state.reset_status = AMDGPU_CTX_UNKNOWN_RESET; - ctx->reset_counter = reset_counter; + ctx->reset_counter_query = reset_counter; + + mutex_unlock(&mgr->lock); + return 0; +} + +static int amdgpu_ctx_query2(struct amdgpu_device *adev, + struct amdgpu_fpriv *fpriv, uint32_t id, + union drm_amdgpu_ctx_out *out) +{ + struct amdgpu_ctx *ctx; + struct amdgpu_ctx_mgr *mgr; + + if (!fpriv) + return -EINVAL; + + mgr = &fpriv->ctx_mgr; + mutex_lock(&mgr->lock); + ctx = idr_find(&mgr->ctx_handles, id); + if (!ctx) { + mutex_unlock(&mgr->lock); + return -EINVAL; + } + + out->state.flags = 0x0; + out->state.hangs = 0x0; + + if (ctx->reset_counter != atomic_read(&adev->gpu_reset_counter)) + out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RESET; + + if (ctx->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) + out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_VRAMLOST; + + if (atomic_read(&ctx->guilty)) + out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY; mutex_unlock(&mgr->lock); return 0; @@ -257,6 +292,9 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data, case AMDGPU_CTX_OP_QUERY_STATE: r = amdgpu_ctx_query(adev, fpriv, id, &args->out); break; + case AMDGPU_CTX_OP_QUERY_STATE2: + r = amdgpu_ctx_query2(adev, fpriv, id, &args->out); + break; default: return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 3573ecdb06ee..70c9e5756b02 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -410,6 +410,9 @@ static int amdgpu_doorbell_init(struct amdgpu_device *adev) return 0; } + if (pci_resource_flags(adev->pdev, 2) & IORESOURCE_UNSET) + return -EINVAL; + /* doorbell bar mapping */ adev->doorbell.base = pci_resource_start(adev->pdev, 2); adev->doorbell.size = pci_resource_len(adev->pdev, 2); @@ -575,41 +578,13 @@ void amdgpu_wb_free(struct amdgpu_device *adev, u32 wb) * @base: base address at which to put VRAM * * Function will try to place VRAM at base address provided - * as parameter (which is so far either PCI aperture address or - * for IGP TOM base address). - * - * If there is not enough space to fit the unvisible VRAM in the 32bits - * address space then we limit the VRAM size to the aperture. - * - * Note: We don't explicitly enforce VRAM start to be aligned on VRAM size, - * this shouldn't be a problem as we are using the PCI aperture as a reference. - * Otherwise this would be needed for rv280, all r3xx, and all r4xx, but - * not IGP. - * - * Note: we use mc_vram_size as on some board we need to program the mc to - * cover the whole aperture even if VRAM size is inferior to aperture size - * Novell bug 204882 + along with lots of ubuntu ones - * - * Note: when limiting vram it's safe to overwritte real_vram_size because - * we are not in case where real_vram_size is inferior to mc_vram_size (ie - * note afected by bogus hw of Novell bug 204882 + along with lots of ubuntu - * ones) - * - * Note: IGP TOM addr should be the same as the aperture addr, we don't - * explicitly check for that though. - * - * FIXME: when reducing VRAM size align new size on power of 2. + * as parameter. */ void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base) { uint64_t limit = (uint64_t)amdgpu_vram_limit << 20; mc->vram_start = base; - if (mc->mc_vram_size > (adev->mc.mc_mask - base + 1)) { - dev_warn(adev->dev, "limiting VRAM to PCI aperture size\n"); - mc->real_vram_size = mc->aper_size; - mc->mc_vram_size = mc->aper_size; - } mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; if (limit && limit < mc->real_vram_size) mc->real_vram_size = limit; @@ -647,7 +622,10 @@ void amdgpu_gart_location(struct amdgpu_device *adev, struct amdgpu_mc *mc) dev_warn(adev->dev, "limiting GTT\n"); mc->gart_size = size_af; } - mc->gart_start = mc->vram_end + 1; + /* VCE doesn't like it when BOs cross a 4GB segment, so align + * the GART base on a 4GB boundary as well. + */ + mc->gart_start = ALIGN(mc->vram_end + 1, 0x100000000ULL); } mc->gart_end = mc->gart_start + mc->gart_size - 1; dev_info(adev->dev, "GTT: %lluM 0x%016llX - 0x%016llX\n", @@ -679,9 +657,13 @@ void amdgpu_fw_reserve_vram_fini(struct amdgpu_device *adev) */ int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev) { + struct ttm_operation_ctx ctx = { false, false }; int r = 0; - u64 gpu_addr; + int i; u64 vram_size = adev->mc.visible_vram_size; + u64 offset = adev->fw_vram_usage.start_offset; + u64 size = adev->fw_vram_usage.size; + struct amdgpu_bo *bo; adev->fw_vram_usage.va = NULL; adev->fw_vram_usage.reserved_bo = NULL; @@ -690,7 +672,7 @@ int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev) adev->fw_vram_usage.size <= vram_size) { r = amdgpu_bo_create(adev, adev->fw_vram_usage.size, - PAGE_SIZE, true, 0, + PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_VRAM, AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS, NULL, NULL, 0, &adev->fw_vram_usage.reserved_bo); @@ -700,11 +682,28 @@ int amdgpu_fw_reserve_vram_init(struct amdgpu_device *adev) r = amdgpu_bo_reserve(adev->fw_vram_usage.reserved_bo, false); if (r) goto error_reserve; + + /* remove the original mem node and create a new one at the + * request position + */ + bo = adev->fw_vram_usage.reserved_bo; + offset = ALIGN(offset, PAGE_SIZE); + for (i = 0; i < bo->placement.num_placement; ++i) { + bo->placements[i].fpfn = offset >> PAGE_SHIFT; + bo->placements[i].lpfn = (offset + size) >> PAGE_SHIFT; + } + + ttm_bo_mem_put(&bo->tbo, &bo->tbo.mem); + r = ttm_bo_mem_space(&bo->tbo, &bo->placement, + &bo->tbo.mem, &ctx); + if (r) + goto error_pin; + r = amdgpu_bo_pin_restricted(adev->fw_vram_usage.reserved_bo, AMDGPU_GEM_DOMAIN_VRAM, adev->fw_vram_usage.start_offset, (adev->fw_vram_usage.start_offset + - adev->fw_vram_usage.size), &gpu_addr); + adev->fw_vram_usage.size), NULL); if (r) goto error_pin; r = amdgpu_bo_kmap(adev->fw_vram_usage.reserved_bo, @@ -728,6 +727,75 @@ error_create: return r; } +/** + * amdgpu_device_resize_fb_bar - try to resize FB BAR + * + * @adev: amdgpu_device pointer + * + * Try to resize FB BAR to make all VRAM CPU accessible. We try very hard not + * to fail, but if any of the BARs is not accessible after the size we abort + * driver loading by returning -ENODEV. + */ +int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev) +{ + u64 space_needed = roundup_pow_of_two(adev->mc.real_vram_size); + u32 rbar_size = order_base_2(((space_needed >> 20) | 1)) - 1; + struct pci_bus *root; + struct resource *res; + unsigned i; + u16 cmd; + int r; + + /* Bypass for VF */ + if (amdgpu_sriov_vf(adev)) + return 0; + + /* Check if the root BUS has 64bit memory resources */ + root = adev->pdev->bus; + while (root->parent) + root = root->parent; + + pci_bus_for_each_resource(root, res, i) { + if (res && res->flags & IORESOURCE_MEM_64 && + res->start > 0x100000000ull) + break; + } + + /* Trying to resize is pointless without a root hub window above 4GB */ + if (!res) + return 0; + + /* Disable memory decoding while we change the BAR addresses and size */ + pci_read_config_word(adev->pdev, PCI_COMMAND, &cmd); + pci_write_config_word(adev->pdev, PCI_COMMAND, + cmd & ~PCI_COMMAND_MEMORY); + + /* Free the VRAM and doorbell BAR, we most likely need to move both. */ + amdgpu_doorbell_fini(adev); + if (adev->asic_type >= CHIP_BONAIRE) + pci_release_resource(adev->pdev, 2); + + pci_release_resource(adev->pdev, 0); + + r = pci_resize_resource(adev->pdev, 0, rbar_size); + if (r == -ENOSPC) + DRM_INFO("Not enough PCI address space for a large BAR."); + else if (r && r != -ENOTSUPP) + DRM_ERROR("Problem resizing BAR0 (%d).", r); + + pci_assign_unassigned_bus_resources(adev->pdev->bus); + + /* When the doorbell or fb BAR isn't available we have no chance of + * using the device. + */ + r = amdgpu_doorbell_init(adev); + if (r || (pci_resource_flags(adev->pdev, 0) & IORESOURCE_UNSET)) + return -ENODEV; + + pci_write_config_word(adev->pdev, PCI_COMMAND, cmd); + + return 0; +} /* * GPU helpers function. @@ -1029,7 +1097,7 @@ static int amdgpu_atombios_init(struct amdgpu_device *adev) atom_card_info->ioreg_read = cail_ioreg_read; atom_card_info->ioreg_write = cail_ioreg_write; } else { - DRM_INFO("PCI I/O BAR is not found. Using MMIO to access ATOM BIOS\n"); + DRM_DEBUG("PCI I/O BAR is not found. Using MMIO to access ATOM BIOS\n"); atom_card_info->ioreg_read = cail_reg_read; atom_card_info->ioreg_write = cail_reg_write; } @@ -1094,20 +1162,8 @@ static void amdgpu_check_block_size(struct amdgpu_device *adev) if (amdgpu_vm_block_size < 9) { dev_warn(adev->dev, "VM page table size (%d) too small\n", amdgpu_vm_block_size); - goto def_value; + amdgpu_vm_block_size = -1; } - - if (amdgpu_vm_block_size > 24 || - (amdgpu_vm_size * 1024) < (1ull << amdgpu_vm_block_size)) { - dev_warn(adev->dev, "VM page table size (%d) too large\n", - amdgpu_vm_block_size); - goto def_value; - } - - return; - -def_value: - amdgpu_vm_block_size = -1; } static void amdgpu_check_vm_size(struct amdgpu_device *adev) @@ -1116,31 +1172,11 @@ static void amdgpu_check_vm_size(struct amdgpu_device *adev) if (amdgpu_vm_size == -1) return; - if (!is_power_of_2(amdgpu_vm_size)) { - dev_warn(adev->dev, "VM size (%d) must be a power of 2\n", - amdgpu_vm_size); - goto def_value; - } - if (amdgpu_vm_size < 1) { dev_warn(adev->dev, "VM size (%d) too small, min is 1GB\n", amdgpu_vm_size); - goto def_value; + amdgpu_vm_size = -1; } - - /* - * Max GPUVM size for Cayman, SI, CI VI are 40 bits. - */ - if (amdgpu_vm_size > 1024) { - dev_warn(adev->dev, "VM size (%d) too large, max is 1TB\n", - amdgpu_vm_size); - goto def_value; - } - - return; - -def_value: - amdgpu_vm_size = -1; } /** @@ -1622,10 +1658,12 @@ static int amdgpu_early_init(struct amdgpu_device *adev) if (r) return r; + amdgpu_amdkfd_device_probe(adev); + if (amdgpu_sriov_vf(adev)) { r = amdgpu_virt_request_full_gpu(adev, true); if (r) - return r; + return -EAGAIN; } for (i = 0; i < adev->num_ip_blocks; i++) { @@ -1716,6 +1754,11 @@ static int amdgpu_init(struct amdgpu_device *adev) adev->ip_blocks[i].status.hw = true; } + amdgpu_amdkfd_device_init(adev); + + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_release_full_gpu(adev, true); + return 0; } @@ -1783,6 +1826,7 @@ static int amdgpu_fini(struct amdgpu_device *adev) { int i, r; + amdgpu_amdkfd_device_fini(adev); /* need to disable SMC first */ for (i = 0; i < adev->num_ip_blocks; i++) { if (!adev->ip_blocks[i].status.hw) @@ -1811,6 +1855,7 @@ static int amdgpu_fini(struct amdgpu_device *adev) if (!adev->ip_blocks[i].status.hw) continue; if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) { + amdgpu_free_static_csa(adev); amdgpu_wb_fini(adev); amdgpu_vram_scratch_fini(adev); } @@ -1859,7 +1904,8 @@ static int amdgpu_fini(struct amdgpu_device *adev) } if (amdgpu_sriov_vf(adev)) - amdgpu_virt_release_full_gpu(adev, false); + if (amdgpu_virt_release_full_gpu(adev, false)) + DRM_ERROR("failed to release exclusive mode on fini\n"); return 0; } @@ -2163,6 +2209,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, mutex_init(&adev->mn_lock); mutex_init(&adev->virt.vf_errors.lock); hash_init(adev->mn_hash); + mutex_init(&adev->lock_reset); amdgpu_check_arguments(adev); @@ -2179,9 +2226,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, INIT_LIST_HEAD(&adev->shadow_list); mutex_init(&adev->shadow_list_lock); - INIT_LIST_HEAD(&adev->gtt_list); - spin_lock_init(&adev->gtt_list_lock); - INIT_LIST_HEAD(&adev->ring_lru_list); spin_lock_init(&adev->ring_lru_list_lock); @@ -2267,8 +2311,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, dev_err(adev->dev, "gpu post error!\n"); goto failed; } - } else { - DRM_INFO("GPU post is not needed\n"); } if (adev->is_atom_fw) { @@ -2305,6 +2347,18 @@ int amdgpu_device_init(struct amdgpu_device *adev, r = amdgpu_init(adev); if (r) { + /* failed in exclusive mode due to timeout */ + if (amdgpu_sriov_vf(adev) && + !amdgpu_sriov_runtime(adev) && + amdgpu_virt_mmio_blocked(adev) && + !amdgpu_virt_wait_reset(adev)) { + dev_err(adev->dev, "VF exclusive mode timeout\n"); + /* Don't send request since VF is inactive. */ + adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME; + adev->virt.ops = NULL; + r = -EAGAIN; + goto failed; + } dev_err(adev->dev, "amdgpu_init failed\n"); amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_INIT_FAIL, 0, 0); amdgpu_fini(adev); @@ -2392,6 +2446,7 @@ failed: amdgpu_vf_error_trans_all(adev); if (runtime) vga_switcheroo_fini_domain_pm_ops(adev->dev); + return r; } @@ -2414,7 +2469,6 @@ void amdgpu_device_fini(struct amdgpu_device *adev) /* evict vram memory */ amdgpu_bo_evict_vram(adev); amdgpu_ib_pool_fini(adev); - amdgpu_fw_reserve_vram_fini(adev); amdgpu_fence_driver_fini(adev); amdgpu_fbdev_fini(adev); r = amdgpu_fini(adev); @@ -2819,163 +2873,172 @@ err: return r; } -/** - * amdgpu_sriov_gpu_reset - reset the asic +/* + * amdgpu_reset - reset ASIC/GPU for bare-metal or passthrough * * @adev: amdgpu device pointer - * @job: which job trigger hang + * @reset_flags: output param tells caller the reset result * - * Attempt the reset the GPU if it has hung (all asics). - * for SRIOV case. - * Returns 0 for success or an error on failure. - */ -int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, struct amdgpu_job *job) + * attempt to do soft-reset or full-reset and reinitialize Asic + * return 0 means successed otherwise failed +*/ +static int amdgpu_reset(struct amdgpu_device *adev, uint64_t* reset_flags) { - int i, j, r = 0; - int resched; - struct amdgpu_bo *bo, *tmp; - struct amdgpu_ring *ring; - struct dma_fence *fence = NULL, *next = NULL; + bool need_full_reset, vram_lost = 0; + int r; - mutex_lock(&adev->virt.lock_reset); - atomic_inc(&adev->gpu_reset_counter); - adev->in_sriov_reset = true; + need_full_reset = amdgpu_need_full_reset(adev); - /* block TTM */ - resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev); + if (!need_full_reset) { + amdgpu_pre_soft_reset(adev); + r = amdgpu_soft_reset(adev); + amdgpu_post_soft_reset(adev); + if (r || amdgpu_check_soft_reset(adev)) { + DRM_INFO("soft reset failed, will fallback to full reset!\n"); + need_full_reset = true; + } - /* we start from the ring trigger GPU hang */ - j = job ? job->ring->idx : 0; + } - /* block scheduler */ - for (i = j; i < j + AMDGPU_MAX_RINGS; ++i) { - ring = adev->rings[i % AMDGPU_MAX_RINGS]; - if (!ring || !ring->sched.thread) - continue; + if (need_full_reset) { + r = amdgpu_suspend(adev); - kthread_park(ring->sched.thread); +retry: + amdgpu_atombios_scratch_regs_save(adev); + r = amdgpu_asic_reset(adev); + amdgpu_atombios_scratch_regs_restore(adev); + /* post card */ + amdgpu_atom_asic_init(adev->mode_info.atom_context); - if (job && j != i) - continue; + if (!r) { + dev_info(adev->dev, "GPU reset succeeded, trying to resume\n"); + r = amdgpu_resume_phase1(adev); + if (r) + goto out; - /* here give the last chance to check if job removed from mirror-list - * since we already pay some time on kthread_park */ - if (job && list_empty(&job->base.node)) { - kthread_unpark(ring->sched.thread); - goto give_up_reset; + vram_lost = amdgpu_check_vram_lost(adev); + if (vram_lost) { + DRM_ERROR("VRAM is lost!\n"); + atomic_inc(&adev->vram_lost_counter); + } + + r = amdgpu_gtt_mgr_recover( + &adev->mman.bdev.man[TTM_PL_TT]); + if (r) + goto out; + + r = amdgpu_resume_phase2(adev); + if (r) + goto out; + + if (vram_lost) + amdgpu_fill_reset_magic(adev); } + } - if (amd_sched_invalidate_job(&job->base, amdgpu_job_hang_limit)) - amd_sched_job_kickout(&job->base); +out: + if (!r) { + amdgpu_irq_gpu_reset_resume_helper(adev); + r = amdgpu_ib_ring_tests(adev); + if (r) { + dev_err(adev->dev, "ib ring test failed (%d).\n", r); + r = amdgpu_suspend(adev); + need_full_reset = true; + goto retry; + } + } - /* only do job_reset on the hang ring if @job not NULL */ - amd_sched_hw_job_reset(&ring->sched); + if (reset_flags) { + if (vram_lost) + (*reset_flags) |= AMDGPU_RESET_INFO_VRAM_LOST; - /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ - amdgpu_fence_driver_force_completion_ring(ring); + if (need_full_reset) + (*reset_flags) |= AMDGPU_RESET_INFO_FULLRESET; } - /* request to take full control of GPU before re-initialization */ - if (job) - amdgpu_virt_reset_gpu(adev); - else - amdgpu_virt_request_full_gpu(adev, true); + return r; +} +/* + * amdgpu_reset_sriov - reset ASIC for SR-IOV vf + * + * @adev: amdgpu device pointer + * @reset_flags: output param tells caller the reset result + * + * do VF FLR and reinitialize Asic + * return 0 means successed otherwise failed +*/ +static int amdgpu_reset_sriov(struct amdgpu_device *adev, uint64_t *reset_flags, bool from_hypervisor) +{ + int r; + + if (from_hypervisor) + r = amdgpu_virt_request_full_gpu(adev, true); + else + r = amdgpu_virt_reset_gpu(adev); + if (r) + return r; /* Resume IP prior to SMC */ - amdgpu_sriov_reinit_early(adev); + r = amdgpu_sriov_reinit_early(adev); + if (r) + goto error; /* we need recover gart prior to run SMC/CP/SDMA resume */ - amdgpu_ttm_recover_gart(adev); + amdgpu_gtt_mgr_recover(&adev->mman.bdev.man[TTM_PL_TT]); /* now we are okay to resume SMC/CP/SDMA */ - amdgpu_sriov_reinit_late(adev); + r = amdgpu_sriov_reinit_late(adev); + if (r) + goto error; amdgpu_irq_gpu_reset_resume_helper(adev); - - if (amdgpu_ib_ring_tests(adev)) + r = amdgpu_ib_ring_tests(adev); + if (r) dev_err(adev->dev, "[GPU_RESET] ib ring test failed (%d).\n", r); +error: /* release full control of GPU after ib test */ amdgpu_virt_release_full_gpu(adev, true); - DRM_INFO("recover vram bo from shadow\n"); - - ring = adev->mman.buffer_funcs_ring; - mutex_lock(&adev->shadow_list_lock); - list_for_each_entry_safe(bo, tmp, &adev->shadow_list, shadow_list) { - next = NULL; - amdgpu_recover_vram_from_shadow(adev, ring, bo, &next); - if (fence) { - r = dma_fence_wait(fence, false); - if (r) { - WARN(r, "recovery from shadow isn't completed\n"); - break; - } - } - - dma_fence_put(fence); - fence = next; - } - mutex_unlock(&adev->shadow_list_lock); - - if (fence) { - r = dma_fence_wait(fence, false); - if (r) - WARN(r, "recovery from shadow isn't completed\n"); - } - dma_fence_put(fence); - - for (i = j; i < j + AMDGPU_MAX_RINGS; ++i) { - ring = adev->rings[i % AMDGPU_MAX_RINGS]; - if (!ring || !ring->sched.thread) - continue; - - if (job && j != i) { - kthread_unpark(ring->sched.thread); - continue; + if (reset_flags) { + if (adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) { + (*reset_flags) |= AMDGPU_RESET_INFO_VRAM_LOST; + atomic_inc(&adev->vram_lost_counter); } - amd_sched_job_recovery(&ring->sched); - kthread_unpark(ring->sched.thread); - } - - drm_helper_resume_force_mode(adev->ddev); -give_up_reset: - ttm_bo_unlock_delayed_workqueue(&adev->mman.bdev, resched); - if (r) { - /* bad news, how to tell it to userspace ? */ - dev_info(adev->dev, "GPU reset failed\n"); - } else { - dev_info(adev->dev, "GPU reset successed!\n"); + /* VF FLR or hotlink reset is always full-reset */ + (*reset_flags) |= AMDGPU_RESET_INFO_FULLRESET; } - adev->in_sriov_reset = false; - mutex_unlock(&adev->virt.lock_reset); return r; } /** - * amdgpu_gpu_reset - reset the asic + * amdgpu_gpu_recover - reset the asic and recover scheduler * * @adev: amdgpu device pointer + * @job: which job trigger hang * - * Attempt the reset the GPU if it has hung (all asics). + * Attempt to reset the GPU if it has hung (all asics). * Returns 0 for success or an error on failure. */ -int amdgpu_gpu_reset(struct amdgpu_device *adev) +int amdgpu_gpu_recover(struct amdgpu_device *adev, struct amdgpu_job *job) { struct drm_atomic_state *state = NULL; - int i, r; - int resched; - bool need_full_reset, vram_lost = false; + uint64_t reset_flags = 0; + int i, r, resched; if (!amdgpu_check_soft_reset(adev)) { DRM_INFO("No hardware hang detected. Did some blocks stall?\n"); return 0; } + dev_info(adev->dev, "GPU reset begin!\n"); + + mutex_lock(&adev->lock_reset); atomic_inc(&adev->gpu_reset_counter); + adev->in_gpu_reset = 1; /* block TTM */ resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev); @@ -2989,69 +3052,26 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev) if (!ring || !ring->sched.thread) continue; - kthread_park(ring->sched.thread); - amd_sched_hw_job_reset(&ring->sched); - } - /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ - amdgpu_fence_driver_force_completion(adev); - need_full_reset = amdgpu_need_full_reset(adev); + /* only focus on the ring hit timeout if &job not NULL */ + if (job && job->ring->idx != i) + continue; - if (!need_full_reset) { - amdgpu_pre_soft_reset(adev); - r = amdgpu_soft_reset(adev); - amdgpu_post_soft_reset(adev); - if (r || amdgpu_check_soft_reset(adev)) { - DRM_INFO("soft reset failed, will fallback to full reset!\n"); - need_full_reset = true; - } - } + kthread_park(ring->sched.thread); + amd_sched_hw_job_reset(&ring->sched, &job->base); - if (need_full_reset) { - r = amdgpu_suspend(adev); + /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ + amdgpu_fence_driver_force_completion(ring); + } -retry: - amdgpu_atombios_scratch_regs_save(adev); - r = amdgpu_asic_reset(adev); - amdgpu_atombios_scratch_regs_restore(adev); - /* post card */ - amdgpu_atom_asic_init(adev->mode_info.atom_context); + if (amdgpu_sriov_vf(adev)) + r = amdgpu_reset_sriov(adev, &reset_flags, job ? false : true); + else + r = amdgpu_reset(adev, &reset_flags); - if (!r) { - dev_info(adev->dev, "GPU reset succeeded, trying to resume\n"); - r = amdgpu_resume_phase1(adev); - if (r) - goto out; - vram_lost = amdgpu_check_vram_lost(adev); - if (vram_lost) { - DRM_ERROR("VRAM is lost!\n"); - atomic_inc(&adev->vram_lost_counter); - } - r = amdgpu_ttm_recover_gart(adev); - if (r) - goto out; - r = amdgpu_resume_phase2(adev); - if (r) - goto out; - if (vram_lost) - amdgpu_fill_reset_magic(adev); - } - } -out: if (!r) { - amdgpu_irq_gpu_reset_resume_helper(adev); - r = amdgpu_ib_ring_tests(adev); - if (r) { - dev_err(adev->dev, "ib ring test failed (%d).\n", r); - r = amdgpu_suspend(adev); - need_full_reset = true; - goto retry; - } - /** - * recovery vm page tables, since we cannot depend on VRAM is - * consistent after gpu full reset. - */ - if (need_full_reset && amdgpu_need_backup(adev)) { + if (((reset_flags & AMDGPU_RESET_INFO_FULLRESET) && !(adev->flags & AMD_IS_APU)) || + (reset_flags & AMDGPU_RESET_INFO_VRAM_LOST)) { struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring; struct amdgpu_bo *bo, *tmp; struct dma_fence *fence = NULL, *next = NULL; @@ -3080,40 +3100,56 @@ out: } dma_fence_put(fence); } + for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; if (!ring || !ring->sched.thread) continue; + /* only focus on the ring hit timeout if &job not NULL */ + if (job && job->ring->idx != i) + continue; + amd_sched_job_recovery(&ring->sched); kthread_unpark(ring->sched.thread); } } else { - dev_err(adev->dev, "asic resume failed (%d).\n", r); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - if (adev->rings[i] && adev->rings[i]->sched.thread) { - kthread_unpark(adev->rings[i]->sched.thread); - } + struct amdgpu_ring *ring = adev->rings[i]; + + if (!ring || !ring->sched.thread) + continue; + + /* only focus on the ring hit timeout if &job not NULL */ + if (job && job->ring->idx != i) + continue; + + kthread_unpark(adev->rings[i]->sched.thread); } } if (amdgpu_device_has_dc_support(adev)) { - r = drm_atomic_helper_resume(adev->ddev, state); + if (drm_atomic_helper_resume(adev->ddev, state)) + dev_info(adev->dev, "drm resume failed:%d\n", r); amdgpu_dm_display_resume(adev); - } else + } else { drm_helper_resume_force_mode(adev->ddev); + } ttm_bo_unlock_delayed_workqueue(&adev->mman.bdev, resched); + if (r) { /* bad news, how to tell it to userspace ? */ - dev_info(adev->dev, "GPU reset failed\n"); - } - else { - dev_info(adev->dev, "GPU reset successed!\n"); + dev_info(adev->dev, "GPU reset(%d) failed\n", atomic_read(&adev->gpu_reset_counter)); + amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_GPU_RESET_FAIL, 0, r); + } else { + dev_info(adev->dev, "GPU reset(%d) successed!\n",atomic_read(&adev->gpu_reset_counter)); } amdgpu_vf_error_trans_all(adev); + adev->in_gpu_reset = 0; + mutex_unlock(&adev->lock_reset); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 138beb550a58..38d47559f098 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -34,6 +34,7 @@ #include <linux/pm_runtime.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_edid.h> +#include <drm/drm_fb_helper.h> static void amdgpu_flip_callback(struct dma_fence *f, struct dma_fence_cb *cb) { @@ -556,15 +557,9 @@ amdgpu_user_framebuffer_create(struct drm_device *dev, return &amdgpu_fb->base; } -void amdgpu_output_poll_changed(struct drm_device *dev) -{ - struct amdgpu_device *adev = dev->dev_private; - amdgpu_fb_output_poll_changed(adev); -} - const struct drm_mode_config_funcs amdgpu_mode_funcs = { .fb_create = amdgpu_user_framebuffer_create, - .output_poll_changed = amdgpu_output_poll_changed + .output_poll_changed = drm_fb_helper_output_poll_changed, }; static const struct drm_prop_enum_list amdgpu_underscan_enum_list[] = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h index 3cc0ef0c055e..0bcb6c6e0ca9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h @@ -25,9 +25,7 @@ struct drm_framebuffer * amdgpu_user_framebuffer_create(struct drm_device *dev, - struct drm_file *file_priv, - const struct drm_mode_fb_cmd2 *mode_cmd); - -void amdgpu_output_poll_changed(struct drm_device *dev); + struct drm_file *file_priv, + const struct drm_mode_fb_cmd2 *mode_cmd); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h index 56caaeee6fea..a8437a3296a6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h @@ -360,6 +360,12 @@ enum amdgpu_pcie_gen { ((adev)->powerplay.pp_funcs->set_clockgating_by_smu(\ (adev)->powerplay.pp_handle, msg_id)) +#define amdgpu_dpm_notify_smu_memory_info(adev, virtual_addr_low, \ + virtual_addr_hi, mc_addr_low, mc_addr_hi, size) \ + ((adev)->powerplay.pp_funcs->notify_smu_memory_info)( \ + (adev)->powerplay.pp_handle, virtual_addr_low, \ + virtual_addr_hi, mc_addr_low, mc_addr_hi, size) + struct amdgpu_dpm { struct amdgpu_ps *ps; /* number of valid power states */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index c2f414ffb2cc..31383e004947 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -216,7 +216,7 @@ module_param_named(exp_hw_support, amdgpu_exp_hw_support, int, 0444); MODULE_PARM_DESC(dc, "Display Core driver (1 = enable, 0 = disable, -1 = auto (default))"); module_param_named(dc, amdgpu_dc, int, 0444); -MODULE_PARM_DESC(dc, "Display Core Log Level (0 = minimal (default), 1 = chatty"); +MODULE_PARM_DESC(dc_log, "Display Core Log Level (0 = minimal (default), 1 = chatty"); module_param_named(dc_log, amdgpu_dc_log, int, 0444); MODULE_PARM_DESC(sched_jobs, "the max number of jobs supported in the sw queue (default 32)"); @@ -306,7 +306,6 @@ MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled (default), 0 = disabled) module_param_named(cik_support, amdgpu_cik_support, int, 0444); #endif - static const struct pci_device_id pciidlist[] = { #ifdef CONFIG_DRM_AMDGPU_SI {0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, @@ -566,12 +565,13 @@ static int amdgpu_kick_out_firmware_fb(struct pci_dev *pdev) return 0; } + static int amdgpu_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) { struct drm_device *dev; unsigned long flags = ent->driver_data; - int ret; + int ret, retry = 0; if ((flags & AMD_EXP_HW_SUPPORT) && !amdgpu_exp_hw_support) { DRM_INFO("This hardware requires experimental hardware support.\n" @@ -604,8 +604,14 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, pci_set_drvdata(pdev, dev); +retry_init: ret = drm_dev_register(dev, ent->driver_data); - if (ret) + if (ret == -EAGAIN && ++retry <= 3) { + DRM_INFO("retry init %d\n", retry); + /* Don't request EX mode too frequently which is attacking */ + msleep(5000); + goto retry_init; + } else if (ret) goto err_pci; return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c index 90fa8e8bc6fb..ff3e9beb7d19 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c @@ -283,12 +283,6 @@ out: return ret; } -void amdgpu_fb_output_poll_changed(struct amdgpu_device *adev) -{ - if (adev->mode_info.rfbdev) - drm_fb_helper_hotplug_event(&adev->mode_info.rfbdev->helper); -} - static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfbdev) { struct amdgpu_framebuffer *rfb = &rfbdev->rfb; @@ -393,24 +387,3 @@ bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj) return true; return false; } - -void amdgpu_fbdev_restore_mode(struct amdgpu_device *adev) -{ - struct amdgpu_fbdev *afbdev; - struct drm_fb_helper *fb_helper; - int ret; - - if (!adev) - return; - - afbdev = adev->mode_info.rfbdev; - - if (!afbdev) - return; - - fb_helper = &afbdev->helper; - - ret = drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper); - if (ret) - DRM_DEBUG("failed to restore crtc mode\n"); -} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index 2fa95aef74d5..604ac03a42e4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -391,9 +391,9 @@ int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring, ring->fence_drv.irq_type = irq_type; ring->fence_drv.initialized = true; - dev_info(adev->dev, "fence driver on ring %d use gpu addr 0x%016llx, " - "cpu addr 0x%p\n", ring->idx, - ring->fence_drv.gpu_addr, ring->fence_drv.cpu_addr); + dev_dbg(adev->dev, "fence driver on ring %d use gpu addr 0x%016llx, " + "cpu addr 0x%p\n", ring->idx, + ring->fence_drv.gpu_addr, ring->fence_drv.cpu_addr); return 0; } @@ -446,7 +446,7 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring, timeout = MAX_SCHEDULE_TIMEOUT; } r = amd_sched_init(&ring->sched, &amdgpu_sched_ops, - num_hw_submission, + num_hw_submission, amdgpu_job_hang_limit, timeout, ring->name); if (r) { DRM_ERROR("Failed to create scheduler on ring %s.\n", @@ -499,7 +499,7 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev) r = amdgpu_fence_wait_empty(ring); if (r) { /* no need to trigger GPU reset as we are unloading */ - amdgpu_fence_driver_force_completion(adev); + amdgpu_fence_driver_force_completion(ring); } amdgpu_irq_put(adev, ring->fence_drv.irq_src, ring->fence_drv.irq_type); @@ -534,7 +534,7 @@ void amdgpu_fence_driver_suspend(struct amdgpu_device *adev) r = amdgpu_fence_wait_empty(ring); if (r) { /* delay GPU reset to resume */ - amdgpu_fence_driver_force_completion(adev); + amdgpu_fence_driver_force_completion(ring); } /* disable the interrupt */ @@ -571,30 +571,15 @@ void amdgpu_fence_driver_resume(struct amdgpu_device *adev) } /** - * amdgpu_fence_driver_force_completion - force all fence waiter to complete + * amdgpu_fence_driver_force_completion - force signal latest fence of ring * - * @adev: amdgpu device pointer + * @ring: fence of the ring to signal * - * In case of GPU reset failure make sure no process keep waiting on fence - * that will never complete. */ -void amdgpu_fence_driver_force_completion(struct amdgpu_device *adev) +void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring) { - int i; - - for (i = 0; i < AMDGPU_MAX_RINGS; i++) { - struct amdgpu_ring *ring = adev->rings[i]; - if (!ring || !ring->fence_drv.initialized) - continue; - - amdgpu_fence_write(ring, ring->fence_drv.sync_seq); - } -} - -void amdgpu_fence_driver_force_completion_ring(struct amdgpu_ring *ring) -{ - if (ring) - amdgpu_fence_write(ring, ring->fence_drv.sync_seq); + amdgpu_fence_write(ring, ring->fence_drv.sync_seq); + amdgpu_fence_process(ring); } /* @@ -709,25 +694,25 @@ static int amdgpu_debugfs_fence_info(struct seq_file *m, void *data) } /** - * amdgpu_debugfs_gpu_reset - manually trigger a gpu reset + * amdgpu_debugfs_gpu_recover - manually trigger a gpu reset & recover * * Manually trigger a gpu reset at the next fence wait. */ -static int amdgpu_debugfs_gpu_reset(struct seq_file *m, void *data) +static int amdgpu_debugfs_gpu_recover(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *) m->private; struct drm_device *dev = node->minor->dev; struct amdgpu_device *adev = dev->dev_private; - seq_printf(m, "gpu reset\n"); - amdgpu_gpu_reset(adev); + seq_printf(m, "gpu recover\n"); + amdgpu_gpu_recover(adev, NULL); return 0; } static const struct drm_info_list amdgpu_debugfs_fence_list[] = { {"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL}, - {"amdgpu_gpu_reset", &amdgpu_debugfs_gpu_reset, 0, NULL} + {"amdgpu_gpu_recover", &amdgpu_debugfs_gpu_recover, 0, NULL} }; static const struct drm_info_list amdgpu_debugfs_fence_list_sriov[] = { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index fe818501c520..1f51897acc5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -57,63 +57,6 @@ */ /** - * amdgpu_gart_table_ram_alloc - allocate system ram for gart page table - * - * @adev: amdgpu_device pointer - * - * Allocate system memory for GART page table - * (r1xx-r3xx, non-pcie r4xx, rs400). These asics require the - * gart table to be in system memory. - * Returns 0 for success, -ENOMEM for failure. - */ -int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev) -{ - void *ptr; - - ptr = pci_alloc_consistent(adev->pdev, adev->gart.table_size, - &adev->gart.table_addr); - if (ptr == NULL) { - return -ENOMEM; - } -#ifdef CONFIG_X86 - if (0) { - set_memory_uc((unsigned long)ptr, - adev->gart.table_size >> PAGE_SHIFT); - } -#endif - adev->gart.ptr = ptr; - memset((void *)adev->gart.ptr, 0, adev->gart.table_size); - return 0; -} - -/** - * amdgpu_gart_table_ram_free - free system ram for gart page table - * - * @adev: amdgpu_device pointer - * - * Free system memory for GART page table - * (r1xx-r3xx, non-pcie r4xx, rs400). These asics require the - * gart table to be in system memory. - */ -void amdgpu_gart_table_ram_free(struct amdgpu_device *adev) -{ - if (adev->gart.ptr == NULL) { - return; - } -#ifdef CONFIG_X86 - if (0) { - set_memory_wb((unsigned long)adev->gart.ptr, - adev->gart.table_size >> PAGE_SHIFT); - } -#endif - pci_free_consistent(adev->pdev, adev->gart.table_size, - (void *)adev->gart.ptr, - adev->gart.table_addr); - adev->gart.ptr = NULL; - adev->gart.table_addr = 0; -} - -/** * amdgpu_gart_table_vram_alloc - allocate vram for gart page table * * @adev: amdgpu_device pointer @@ -377,10 +320,8 @@ int amdgpu_gart_init(struct amdgpu_device *adev) #ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS /* Allocate pages table */ adev->gart.pages = vzalloc(sizeof(void *) * adev->gart.num_cpu_pages); - if (adev->gart.pages == NULL) { - amdgpu_gart_fini(adev); + if (adev->gart.pages == NULL) return -ENOMEM; - } #endif return 0; @@ -395,11 +336,6 @@ int amdgpu_gart_init(struct amdgpu_device *adev) */ void amdgpu_gart_fini(struct amdgpu_device *adev) { - if (adev->gart.ready) { - /* unbind pages */ - amdgpu_gart_unbind(adev, 0, adev->gart.num_cpu_pages); - } - adev->gart.ready = false; #ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS vfree(adev->gart.pages); adev->gart.pages = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h index afbe803b1a13..d4a43302c2be 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h @@ -39,7 +39,7 @@ struct amdgpu_gart_funcs; #define AMDGPU_GPU_PAGE_ALIGN(a) (((a) + AMDGPU_GPU_PAGE_MASK) & ~AMDGPU_GPU_PAGE_MASK) struct amdgpu_gart { - dma_addr_t table_addr; + u64 table_addr; struct amdgpu_bo *robj; void *ptr; unsigned num_gpu_pages; @@ -56,8 +56,6 @@ struct amdgpu_gart { const struct amdgpu_gart_funcs *gart_funcs; }; -int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev); -void amdgpu_gart_table_ram_free(struct amdgpu_device *adev); int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev); void amdgpu_gart_table_vram_free(struct amdgpu_device *adev); int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index e87eedcc0da9..eb75eb44efc6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -72,7 +72,7 @@ retry: initial_domain |= AMDGPU_GEM_DOMAIN_GTT; goto retry; } - DRM_ERROR("Failed to allocate GEM object (%ld, %d, %u, %d)\n", + DRM_DEBUG("Failed to allocate GEM object (%ld, %d, %u, %d)\n", size, initial_domain, alignment, r); } return r; @@ -282,6 +282,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) { + struct ttm_operation_ctx ctx = { true, false }; struct amdgpu_device *adev = dev->dev_private; struct drm_amdgpu_gem_userptr *args = data; struct drm_gem_object *gobj; @@ -335,7 +336,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data, goto free_pages; amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT); - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); amdgpu_bo_unreserve(bo); if (r) goto free_pages; @@ -557,14 +558,25 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, int r = 0; if (args->va_address < AMDGPU_VA_RESERVED_SIZE) { - dev_err(&dev->pdev->dev, + dev_dbg(&dev->pdev->dev, "va_address 0x%LX is in reserved area 0x%LX\n", args->va_address, AMDGPU_VA_RESERVED_SIZE); return -EINVAL; } + if (args->va_address >= AMDGPU_VA_HOLE_START && + args->va_address < AMDGPU_VA_HOLE_END) { + dev_dbg(&dev->pdev->dev, + "va_address 0x%LX is in VA hole 0x%LX-0x%LX\n", + args->va_address, AMDGPU_VA_HOLE_START, + AMDGPU_VA_HOLE_END); + return -EINVAL; + } + + args->va_address &= AMDGPU_VA_HOLE_MASK; + if ((args->flags & ~valid_flags) && (args->flags & ~prt_flags)) { - dev_err(&dev->pdev->dev, "invalid flags combination 0x%08X\n", + dev_dbg(&dev->pdev->dev, "invalid flags combination 0x%08X\n", args->flags); return -EINVAL; } @@ -576,7 +588,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, case AMDGPU_VA_OP_REPLACE: break; default: - dev_err(&dev->pdev->dev, "unsupported operation %d\n", + dev_dbg(&dev->pdev->dev, "unsupported operation %d\n", args->operation); return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c index 00e0ce10862f..e14ab34d8262 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c @@ -31,6 +31,11 @@ struct amdgpu_gtt_mgr { atomic64_t available; }; +struct amdgpu_gtt_node { + struct drm_mm_node node; + struct ttm_buffer_object *tbo; +}; + /** * amdgpu_gtt_mgr_init - init GTT manager and DRM MM * @@ -79,17 +84,17 @@ static int amdgpu_gtt_mgr_fini(struct ttm_mem_type_manager *man) } /** - * amdgpu_gtt_mgr_is_allocated - Check if mem has address space + * amdgpu_gtt_mgr_has_gart_addr - Check if mem has address space * * @mem: the mem object to check * * Check if a mem object has already address space allocated. */ -bool amdgpu_gtt_mgr_is_allocated(struct ttm_mem_reg *mem) +bool amdgpu_gtt_mgr_has_gart_addr(struct ttm_mem_reg *mem) { - struct drm_mm_node *node = mem->mm_node; + struct amdgpu_gtt_node *node = mem->mm_node; - return (node->start != AMDGPU_BO_INVALID_OFFSET); + return (node->node.start != AMDGPU_BO_INVALID_OFFSET); } /** @@ -109,12 +114,12 @@ static int amdgpu_gtt_mgr_alloc(struct ttm_mem_type_manager *man, { struct amdgpu_device *adev = amdgpu_ttm_adev(man->bdev); struct amdgpu_gtt_mgr *mgr = man->priv; - struct drm_mm_node *node = mem->mm_node; + struct amdgpu_gtt_node *node = mem->mm_node; enum drm_mm_insert_mode mode; unsigned long fpfn, lpfn; int r; - if (amdgpu_gtt_mgr_is_allocated(mem)) + if (amdgpu_gtt_mgr_has_gart_addr(mem)) return 0; if (place) @@ -132,13 +137,13 @@ static int amdgpu_gtt_mgr_alloc(struct ttm_mem_type_manager *man, mode = DRM_MM_INSERT_HIGH; spin_lock(&mgr->lock); - r = drm_mm_insert_node_in_range(&mgr->mm, node, - mem->num_pages, mem->page_alignment, 0, - fpfn, lpfn, mode); + r = drm_mm_insert_node_in_range(&mgr->mm, &node->node, mem->num_pages, + mem->page_alignment, 0, fpfn, lpfn, + mode); spin_unlock(&mgr->lock); if (!r) - mem->start = node->start; + mem->start = node->node.start; return r; } @@ -159,7 +164,7 @@ static int amdgpu_gtt_mgr_new(struct ttm_mem_type_manager *man, struct ttm_mem_reg *mem) { struct amdgpu_gtt_mgr *mgr = man->priv; - struct drm_mm_node *node; + struct amdgpu_gtt_node *node; int r; spin_lock(&mgr->lock); @@ -177,8 +182,9 @@ static int amdgpu_gtt_mgr_new(struct ttm_mem_type_manager *man, goto err_out; } - node->start = AMDGPU_BO_INVALID_OFFSET; - node->size = mem->num_pages; + node->node.start = AMDGPU_BO_INVALID_OFFSET; + node->node.size = mem->num_pages; + node->tbo = tbo; mem->mm_node = node; if (place->fpfn || place->lpfn || place->flags & TTM_PL_FLAG_TOPDOWN) { @@ -190,7 +196,7 @@ static int amdgpu_gtt_mgr_new(struct ttm_mem_type_manager *man, goto err_out; } } else { - mem->start = node->start; + mem->start = node->node.start; } return 0; @@ -214,14 +220,14 @@ static void amdgpu_gtt_mgr_del(struct ttm_mem_type_manager *man, struct ttm_mem_reg *mem) { struct amdgpu_gtt_mgr *mgr = man->priv; - struct drm_mm_node *node = mem->mm_node; + struct amdgpu_gtt_node *node = mem->mm_node; if (!node) return; spin_lock(&mgr->lock); - if (node->start != AMDGPU_BO_INVALID_OFFSET) - drm_mm_remove_node(node); + if (node->node.start != AMDGPU_BO_INVALID_OFFSET) + drm_mm_remove_node(&node->node); spin_unlock(&mgr->lock); atomic64_add(mem->num_pages, &mgr->available); @@ -244,6 +250,25 @@ uint64_t amdgpu_gtt_mgr_usage(struct ttm_mem_type_manager *man) return (result > 0 ? result : 0) * PAGE_SIZE; } +int amdgpu_gtt_mgr_recover(struct ttm_mem_type_manager *man) +{ + struct amdgpu_gtt_mgr *mgr = man->priv; + struct amdgpu_gtt_node *node; + struct drm_mm_node *mm_node; + int r = 0; + + spin_lock(&mgr->lock); + drm_mm_for_each_node(mm_node, &mgr->mm) { + node = container_of(mm_node, struct amdgpu_gtt_node, node); + r = amdgpu_ttm_recover_gart(node->tbo); + if (r) + break; + } + spin_unlock(&mgr->lock); + + return r; +} + /** * amdgpu_gtt_mgr_debug - dump VRAM table * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 659997bfff30..0cf86eb357d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -164,7 +164,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, } if (ring->funcs->emit_pipeline_sync && job && - ((tmp = amdgpu_sync_get_fence(&job->sched_sync)) || + ((tmp = amdgpu_sync_get_fence(&job->sched_sync, NULL)) || amdgpu_vm_need_pipeline_sync(ring, job))) { need_pipe_sync = true; dma_fence_put(tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index 47c5ce9807db..c340774082ea 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -88,7 +88,7 @@ static void amdgpu_irq_reset_work_func(struct work_struct *work) reset_work); if (!amdgpu_sriov_vf(adev)) - amdgpu_gpu_reset(adev); + amdgpu_gpu_recover(adev, NULL); } /* Disable *all* interrupts */ @@ -232,7 +232,7 @@ int amdgpu_irq_init(struct amdgpu_device *adev) int ret = pci_enable_msi(adev->pdev); if (!ret) { adev->irq.msi_enabled = true; - dev_info(adev->dev, "amdgpu: using MSI.\n"); + dev_dbg(adev->dev, "amdgpu: using MSI.\n"); } } @@ -262,7 +262,7 @@ int amdgpu_irq_init(struct amdgpu_device *adev) return r; } - DRM_INFO("amdgpu: irq initialized.\n"); + DRM_DEBUG("amdgpu: irq initialized.\n"); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index 0cfc68db575b..bdc210ac74f8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -37,10 +37,7 @@ static void amdgpu_job_timedout(struct amd_sched_job *s_job) atomic_read(&job->ring->fence_drv.last_seq), job->ring->fence_drv.sync_seq); - if (amdgpu_sriov_vf(job->adev)) - amdgpu_sriov_gpu_reset(job->adev, job); - else - amdgpu_gpu_reset(job->adev); + amdgpu_gpu_recover(job->adev, job); } int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, @@ -63,7 +60,6 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, (*job)->num_ibs = num_ibs; amdgpu_sync_create(&(*job)->sync); - amdgpu_sync_create(&(*job)->dep_sync); amdgpu_sync_create(&(*job)->sched_sync); (*job)->vram_lost_counter = atomic_read(&adev->vram_lost_counter); @@ -104,10 +100,9 @@ static void amdgpu_job_free_cb(struct amd_sched_job *s_job) { struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base); - amdgpu_ring_priority_put(job->ring, amd_sched_get_job_priority(s_job)); + amdgpu_ring_priority_put(job->ring, s_job->s_priority); dma_fence_put(job->fence); amdgpu_sync_free(&job->sync); - amdgpu_sync_free(&job->dep_sync); amdgpu_sync_free(&job->sched_sync); kfree(job); } @@ -118,7 +113,6 @@ void amdgpu_job_free(struct amdgpu_job *job) dma_fence_put(job->fence); amdgpu_sync_free(&job->sync); - amdgpu_sync_free(&job->dep_sync); amdgpu_sync_free(&job->sched_sync); kfree(job); } @@ -141,28 +135,29 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring, job->fence_ctx = entity->fence_context; *f = dma_fence_get(&job->base.s_fence->finished); amdgpu_job_free_resources(job); - amdgpu_ring_priority_get(job->ring, - amd_sched_get_job_priority(&job->base)); - amd_sched_entity_push_job(&job->base); + amdgpu_ring_priority_get(job->ring, job->base.s_priority); + amd_sched_entity_push_job(&job->base, entity); return 0; } -static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job) +static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job, + struct amd_sched_entity *s_entity) { struct amdgpu_job *job = to_amdgpu_job(sched_job); struct amdgpu_vm *vm = job->vm; - - struct dma_fence *fence = amdgpu_sync_get_fence(&job->dep_sync); + bool explicit = false; int r; - - if (amd_sched_dependency_optimized(fence, sched_job->s_entity)) { - r = amdgpu_sync_fence(job->adev, &job->sched_sync, fence); - if (r) - DRM_ERROR("Error adding fence to sync (%d)\n", r); + struct dma_fence *fence = amdgpu_sync_get_fence(&job->sync, &explicit); + + if (fence && explicit) { + if (amd_sched_dependency_optimized(fence, s_entity)) { + r = amdgpu_sync_fence(job->adev, &job->sched_sync, fence, false); + if (r) + DRM_ERROR("Error adding fence to sync (%d)\n", r); + } } - if (!fence) - fence = amdgpu_sync_get_fence(&job->sync); + while (fence == NULL && vm && !job->vm_id) { struct amdgpu_ring *ring = job->ring; @@ -172,7 +167,7 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job) if (r) DRM_ERROR("Error getting VM ID (%d)\n", r); - fence = amdgpu_sync_get_fence(&job->sync); + fence = amdgpu_sync_get_fence(&job->sync, NULL); } return fence; @@ -180,7 +175,7 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job) static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job) { - struct dma_fence *fence = NULL; + struct dma_fence *fence = NULL, *finished; struct amdgpu_device *adev; struct amdgpu_job *job; int r; @@ -190,15 +185,18 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job) return NULL; } job = to_amdgpu_job(sched_job); + finished = &job->base.s_fence->finished; adev = job->adev; BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL)); trace_amdgpu_sched_run_job(job); - /* skip ib schedule when vram is lost */ - if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) { - dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED); - DRM_ERROR("Skip scheduling IBs!\n"); + + if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) + dma_fence_set_error(finished, -ECANCELED);/* skip IB as well if VRAM lost */ + + if (finished->error < 0) { + DRM_INFO("Skip scheduling IBs!\n"); } else { r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job, &fence); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 720139e182a3..bd6e9a40f421 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -63,8 +63,6 @@ void amdgpu_driver_unload_kms(struct drm_device *dev) pm_runtime_forbid(dev->dev); } - amdgpu_amdkfd_device_fini(adev); - amdgpu_acpi_fini(adev); amdgpu_device_fini(adev); @@ -159,9 +157,6 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags) "Error during ACPI methods call\n"); } - amdgpu_amdkfd_device_probe(adev); - amdgpu_amdkfd_device_init(adev); - if (amdgpu_device_is_px(dev)) { pm_runtime_use_autosuspend(dev->dev); pm_runtime_set_autosuspend_delay(dev->dev, 5000); @@ -171,9 +166,6 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags) pm_runtime_put_autosuspend(dev->dev); } - if (amdgpu_sriov_vf(adev)) - amdgpu_virt_release_full_gpu(adev, true); - out: if (r) { /* balance pm_runtime_get_sync in amdgpu_driver_unload_kms */ @@ -558,6 +550,7 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file } case AMDGPU_INFO_DEV_INFO: { struct drm_amdgpu_info_device dev_info = {}; + uint64_t vm_size; dev_info.device_id = dev->pdev->device; dev_info.chip_rev = adev->rev_id; @@ -585,8 +578,17 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file dev_info.ids_flags |= AMDGPU_IDS_FLAGS_FUSION; if (amdgpu_sriov_vf(adev)) dev_info.ids_flags |= AMDGPU_IDS_FLAGS_PREEMPTION; + + vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE; dev_info.virtual_address_offset = AMDGPU_VA_RESERVED_SIZE; - dev_info.virtual_address_max = (uint64_t)adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE; + dev_info.virtual_address_max = + min(vm_size, AMDGPU_VA_HOLE_START); + + vm_size -= AMDGPU_VA_RESERVED_SIZE; + if (vm_size > AMDGPU_VA_HOLE_START) { + dev_info.high_va_offset = AMDGPU_VA_HOLE_END; + dev_info.high_va_max = AMDGPU_VA_HOLE_END | vm_size; + } dev_info.virtual_address_alignment = max((int)PAGE_SIZE, AMDGPU_GPU_PAGE_SIZE); dev_info.pte_fragment_size = (1 << adev->vm_manager.fragment_size) * AMDGPU_GPU_PAGE_SIZE; dev_info.gart_page_size = AMDGPU_GPU_PAGE_SIZE; @@ -786,9 +788,7 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file */ void amdgpu_driver_lastclose_kms(struct drm_device *dev) { - struct amdgpu_device *adev = dev->dev_private; - - amdgpu_fbdev_restore_mode(adev); + drm_fb_helper_lastclose(dev); vga_switcheroo_process_delayed_switch(); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h index ffde1e9666e8..54f06c959340 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h @@ -89,7 +89,6 @@ enum amdgpu_hpd_id { AMDGPU_HPD_4, AMDGPU_HPD_5, AMDGPU_HPD_6, - AMDGPU_HPD_LAST, AMDGPU_HPD_NONE = 0xff, }; @@ -106,7 +105,6 @@ enum amdgpu_crtc_irq { AMDGPU_CRTC_IRQ_VLINE4, AMDGPU_CRTC_IRQ_VLINE5, AMDGPU_CRTC_IRQ_VLINE6, - AMDGPU_CRTC_IRQ_LAST, AMDGPU_CRTC_IRQ_NONE = 0xff }; @@ -117,7 +115,6 @@ enum amdgpu_pageflip_irq { AMDGPU_PAGEFLIP_IRQ_D4, AMDGPU_PAGEFLIP_IRQ_D5, AMDGPU_PAGEFLIP_IRQ_D6, - AMDGPU_PAGEFLIP_IRQ_LAST, AMDGPU_PAGEFLIP_IRQ_NONE = 0xff }; @@ -661,10 +658,6 @@ void amdgpu_fbdev_fini(struct amdgpu_device *adev); void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state); int amdgpu_fbdev_total_size(struct amdgpu_device *adev); bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj); -void amdgpu_fbdev_restore_mode(struct amdgpu_device *adev); - -void amdgpu_fb_output_poll_changed(struct amdgpu_device *adev); - int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tiled); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index ea25164e7f4b..dc0a8be98043 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -281,6 +281,44 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr, *cpu_addr = NULL; } +/* Validate bo size is bit bigger then the request domain */ +static bool amdgpu_bo_validate_size(struct amdgpu_device *adev, + unsigned long size, u32 domain) +{ + struct ttm_mem_type_manager *man = NULL; + + /* + * If GTT is part of requested domains the check must succeed to + * allow fall back to GTT + */ + if (domain & AMDGPU_GEM_DOMAIN_GTT) { + man = &adev->mman.bdev.man[TTM_PL_TT]; + + if (size < (man->size << PAGE_SHIFT)) + return true; + else + goto fail; + } + + if (domain & AMDGPU_GEM_DOMAIN_VRAM) { + man = &adev->mman.bdev.man[TTM_PL_VRAM]; + + if (size < (man->size << PAGE_SHIFT)) + return true; + else + goto fail; + } + + + /* TODO add more domains checks, such as AMDGPU_GEM_DOMAIN_CPU */ + return true; + +fail: + DRM_DEBUG("BO size %lu > total memory in domain: %llu\n", size, + man->size << PAGE_SHIFT); + return false; +} + static int amdgpu_bo_do_create(struct amdgpu_device *adev, unsigned long size, int byte_align, bool kernel, u32 domain, u64 flags, @@ -289,16 +327,19 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev, uint64_t init_value, struct amdgpu_bo **bo_ptr) { + struct ttm_operation_ctx ctx = { !kernel, false }; struct amdgpu_bo *bo; enum ttm_bo_type type; unsigned long page_align; - u64 initial_bytes_moved, bytes_moved; size_t acc_size; int r; page_align = roundup(byte_align, PAGE_SIZE) >> PAGE_SHIFT; size = ALIGN(size, PAGE_SIZE); + if (!amdgpu_bo_validate_size(adev, size, domain)) + return -ENOMEM; + if (kernel) { type = ttm_bo_type_kernel; } else if (sg) { @@ -364,22 +405,19 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev, bo->tbo.bdev = &adev->mman.bdev; amdgpu_ttm_placement_from_domain(bo, domain); - initial_bytes_moved = atomic64_read(&adev->num_bytes_moved); - /* Kernel allocation are uninterruptible */ r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, type, - &bo->placement, page_align, !kernel, NULL, + &bo->placement, page_align, &ctx, NULL, acc_size, sg, resv, &amdgpu_ttm_bo_destroy); if (unlikely(r != 0)) return r; - bytes_moved = atomic64_read(&adev->num_bytes_moved) - - initial_bytes_moved; if (adev->mc.visible_vram_size < adev->mc.real_vram_size && bo->tbo.mem.mem_type == TTM_PL_VRAM && bo->tbo.mem.start < adev->mc.visible_vram_size >> PAGE_SHIFT) - amdgpu_cs_report_moved_bytes(adev, bytes_moved, bytes_moved); + amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, + ctx.bytes_moved); else - amdgpu_cs_report_moved_bytes(adev, bytes_moved, 0); + amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, 0); if (kernel) bo->tbo.priority = 1; @@ -511,6 +549,7 @@ err: int amdgpu_bo_validate(struct amdgpu_bo *bo) { + struct ttm_operation_ctx ctx = { false, false }; uint32_t domain; int r; @@ -521,7 +560,7 @@ int amdgpu_bo_validate(struct amdgpu_bo *bo) retry: amdgpu_ttm_placement_from_domain(bo, domain); - r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) { domain = bo->allowed_domains; goto retry; @@ -632,6 +671,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, u64 *gpu_addr) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + struct ttm_operation_ctx ctx = { false, false }; int r, i; if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) @@ -647,7 +687,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, if (bo->pin_count) { uint32_t mem_type = bo->tbo.mem.mem_type; - if (domain != amdgpu_mem_type_to_domain(mem_type)) + if (!(domain & amdgpu_mem_type_to_domain(mem_type))) return -EINVAL; bo->pin_count++; @@ -682,21 +722,23 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT; } - r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (unlikely(r)) { dev_err(adev->dev, "%p pin failed\n", bo); goto error; } + r = amdgpu_ttm_alloc_gart(&bo->tbo); + if (unlikely(r)) { + dev_err(adev->dev, "%p bind failed\n", bo); + goto error; + } + bo->pin_count = 1; - if (gpu_addr != NULL) { - r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem); - if (unlikely(r)) { - dev_err(adev->dev, "%p bind failed\n", bo); - goto error; - } + if (gpu_addr != NULL) *gpu_addr = amdgpu_bo_gpu_offset(bo); - } + + domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type); if (domain == AMDGPU_GEM_DOMAIN_VRAM) { adev->vram_pin_size += amdgpu_bo_size(bo); if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) @@ -717,6 +759,7 @@ int amdgpu_bo_pin(struct amdgpu_bo *bo, u32 domain, u64 *gpu_addr) int amdgpu_bo_unpin(struct amdgpu_bo *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + struct ttm_operation_ctx ctx = { false, false }; int r, i; if (!bo->pin_count) { @@ -730,7 +773,7 @@ int amdgpu_bo_unpin(struct amdgpu_bo *bo) bo->placements[i].lpfn = 0; bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT; } - r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (unlikely(r)) { dev_err(adev->dev, "%p validate failed for unpin\n", bo); goto error; @@ -779,8 +822,8 @@ int amdgpu_bo_init(struct amdgpu_device *adev) adev->mc.vram_mtrr = arch_phys_wc_add(adev->mc.aper_base, adev->mc.aper_size); DRM_INFO("Detected VRAM RAM=%lluM, BAR=%lluM\n", - adev->mc.mc_vram_size >> 20, - (unsigned long long)adev->mc.aper_size >> 20); + adev->mc.mc_vram_size >> 20, + (unsigned long long)adev->mc.aper_size >> 20); DRM_INFO("RAM width %dbits %s\n", adev->mc.vram_width, amdgpu_vram_names[adev->mc.vram_type]); return amdgpu_ttm_init(adev); @@ -902,6 +945,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo, int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev); + struct ttm_operation_ctx ctx = { false, false }; struct amdgpu_bo *abo; unsigned long offset, size; int r; @@ -935,7 +979,7 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo) abo->placement.num_busy_placement = 1; abo->placement.busy_placement = &abo->placements[1]; - r = ttm_bo_validate(bo, &abo->placement, false, false); + r = ttm_bo_validate(bo, &abo->placement, &ctx); if (unlikely(r != 0)) return r; @@ -980,7 +1024,7 @@ u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo) { WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_SYSTEM); WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_TT && - !amdgpu_ttm_is_bound(bo->tbo.ttm)); + !amdgpu_gtt_mgr_has_gart_addr(&bo->tbo.mem)); WARN_ON_ONCE(!ww_mutex_is_locked(&bo->tbo.resv->lock) && !bo->pin_count); WARN_ON_ONCE(bo->tbo.mem.start == AMDGPU_BO_INVALID_OFFSET); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 428aae048f4b..33615e2ea2e6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -187,7 +187,7 @@ static inline u64 amdgpu_bo_mmap_offset(struct amdgpu_bo *bo) static inline bool amdgpu_bo_gpu_accessible(struct amdgpu_bo *bo) { switch (bo->tbo.mem.mem_type) { - case TTM_PL_TT: return amdgpu_ttm_is_bound(bo->tbo.ttm); + case TTM_PL_TT: return amdgpu_gtt_mgr_has_gart_addr(&bo->tbo.mem); case TTM_PL_VRAM: return true; default: return false; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 6c570d4e4516..6f56ff606e43 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -30,7 +30,6 @@ #include <linux/hwmon.h> #include <linux/hwmon-sysfs.h> -#include "amd_powerplay.h" static int amdgpu_debugfs_pm_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 447d446b5015..2157d4509e84 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -264,7 +264,7 @@ static int psp_hw_start(struct psp_context *psp) struct amdgpu_device *adev = psp->adev; int ret; - if (!amdgpu_sriov_vf(adev) || !adev->in_sriov_reset) { + if (!amdgpu_sriov_vf(adev) || !adev->in_gpu_reset) { ret = psp_bootloader_load_sysdrv(psp); if (ret) return ret; @@ -334,23 +334,26 @@ static int psp_load_fw(struct amdgpu_device *adev) int ret; struct psp_context *psp = &adev->psp; + if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset != 0) + goto skip_memalloc; + psp->cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL); if (!psp->cmd) return -ENOMEM; ret = amdgpu_bo_create_kernel(adev, PSP_1_MEG, PSP_1_MEG, - AMDGPU_GEM_DOMAIN_GTT, - &psp->fw_pri_bo, - &psp->fw_pri_mc_addr, - &psp->fw_pri_buf); + AMDGPU_GEM_DOMAIN_GTT, + &psp->fw_pri_bo, + &psp->fw_pri_mc_addr, + &psp->fw_pri_buf); if (ret) goto failed; ret = amdgpu_bo_create_kernel(adev, PSP_FENCE_BUFFER_SIZE, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, - &psp->fence_buf_bo, - &psp->fence_buf_mc_addr, - &psp->fence_buf); + AMDGPU_GEM_DOMAIN_VRAM, + &psp->fence_buf_bo, + &psp->fence_buf_mc_addr, + &psp->fence_buf); if (ret) goto failed_mem2; @@ -375,6 +378,7 @@ static int psp_load_fw(struct amdgpu_device *adev) if (ret) goto failed_mem; +skip_memalloc: ret = psp_hw_start(psp); if (ret) goto failed_mem; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_queue_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_queue_mgr.c index 93d86619e802..262c1267249e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_queue_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_queue_mgr.c @@ -225,7 +225,7 @@ int amdgpu_queue_mgr_map(struct amdgpu_device *adev, /* Right now all IPs have only one instance - multiple rings. */ if (instance != 0) { - DRM_ERROR("invalid ip instance: %d\n", instance); + DRM_DEBUG("invalid ip instance: %d\n", instance); return -EINVAL; } @@ -255,13 +255,13 @@ int amdgpu_queue_mgr_map(struct amdgpu_device *adev, ip_num_rings = adev->vcn.num_enc_rings; break; default: - DRM_ERROR("unknown ip type: %d\n", hw_ip); + DRM_DEBUG("unknown ip type: %d\n", hw_ip); return -EINVAL; } if (ring >= ip_num_rings) { - DRM_ERROR("Ring index:%d exceeds maximum:%d for ip:%d\n", - ring, ip_num_rings, hw_ip); + DRM_DEBUG("Ring index:%d exceeds maximum:%d for ip:%d\n", + ring, ip_num_rings, hw_ip); return -EINVAL; } @@ -292,7 +292,7 @@ int amdgpu_queue_mgr_map(struct amdgpu_device *adev, default: *out_ring = NULL; r = -EINVAL; - DRM_ERROR("unknown HW IP type: %d\n", mapper->hw_ip); + DRM_DEBUG("unknown HW IP type: %d\n", mapper->hw_ip); } out_unlock: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index b18c2b96691f..a6b89e3932a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -79,8 +79,7 @@ struct amdgpu_fence_driver { int amdgpu_fence_driver_init(struct amdgpu_device *adev); void amdgpu_fence_driver_fini(struct amdgpu_device *adev); -void amdgpu_fence_driver_force_completion(struct amdgpu_device *adev); -void amdgpu_fence_driver_force_completion_ring(struct amdgpu_ring *ring); +void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring); int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring, unsigned num_hw_submission); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index a4bf21f8f1c1..ebe1ffbab0c1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -35,6 +35,7 @@ struct amdgpu_sync_entry { struct hlist_node node; struct dma_fence *fence; + bool explicit; }; static struct kmem_cache *amdgpu_sync_slab; @@ -141,7 +142,7 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f) * */ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, - struct dma_fence *f) + struct dma_fence *f, bool explicit) { struct amdgpu_sync_entry *e; @@ -159,6 +160,8 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, if (!e) return -ENOMEM; + e->explicit = explicit; + hash_add(sync->fences, &e->node, f->context); e->fence = dma_fence_get(f); return 0; @@ -189,10 +192,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, /* always sync to the exclusive fence */ f = reservation_object_get_excl(resv); - r = amdgpu_sync_fence(adev, sync, f); - - if (explicit_sync) - return r; + r = amdgpu_sync_fence(adev, sync, f, false); flist = reservation_object_get_list(resv); if (!flist || r) @@ -212,15 +212,15 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, (fence_owner == AMDGPU_FENCE_OWNER_VM))) continue; - /* Ignore fence from the same owner as + /* Ignore fence from the same owner and explicit one as * long as it isn't undefined. */ if (owner != AMDGPU_FENCE_OWNER_UNDEFINED && - fence_owner == owner) + (fence_owner == owner || explicit_sync)) continue; } - r = amdgpu_sync_fence(adev, sync, f); + r = amdgpu_sync_fence(adev, sync, f, false); if (r) break; } @@ -275,19 +275,21 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync, * amdgpu_sync_get_fence - get the next fence from the sync object * * @sync: sync object to use + * @explicit: true if the next fence is explicit * * Get and removes the next fence from the sync object not signaled yet. */ -struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync) +struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync, bool *explicit) { struct amdgpu_sync_entry *e; struct hlist_node *tmp; struct dma_fence *f; int i; - hash_for_each_safe(sync->fences, i, tmp, e, node) { f = e->fence; + if (explicit) + *explicit = e->explicit; hash_del(&e->node); kmem_cache_free(amdgpu_sync_slab, e); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h index 70d7e3a279a0..7aba38d5c9df 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h @@ -41,7 +41,7 @@ struct amdgpu_sync { void amdgpu_sync_create(struct amdgpu_sync *sync); int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, - struct dma_fence *f); + struct dma_fence *f, bool explicit); int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct reservation_object *resv, @@ -49,7 +49,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, bool explicit_sync); struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync, struct amdgpu_ring *ring); -struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync); +struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync, bool *explicit); int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr); void amdgpu_sync_free(struct amdgpu_sync *sync); int amdgpu_sync_init(void); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index ad5bf86ee8a3..952e0bf3bc84 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -110,7 +110,7 @@ static int amdgpu_ttm_global_init(struct amdgpu_device *adev) ring = adev->mman.buffer_funcs_ring; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL]; r = amd_sched_entity_init(&ring->sched, &adev->mman.entity, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r) { DRM_ERROR("Failed setting up TTM BO move run queue.\n"); goto error_entity; @@ -282,8 +282,7 @@ static uint64_t amdgpu_mm_node_addr(struct ttm_buffer_object *bo, { uint64_t addr = 0; - if (mem->mem_type != TTM_PL_TT || - amdgpu_gtt_mgr_is_allocated(mem)) { + if (mem->mem_type != TTM_PL_TT || amdgpu_gtt_mgr_has_gart_addr(mem)) { addr = mm_node->start << PAGE_SHIFT; addr += bo->bdev->man[mem->mem_type].gpu_offset; } @@ -369,7 +368,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, * dst to window 1 */ if (src->mem->mem_type == TTM_PL_TT && - !amdgpu_gtt_mgr_is_allocated(src->mem)) { + !amdgpu_gtt_mgr_has_gart_addr(src->mem)) { r = amdgpu_map_buffer(src->bo, src->mem, PFN_UP(cur_size + src_page_offset), src_node_start, 0, ring, @@ -383,7 +382,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, } if (dst->mem->mem_type == TTM_PL_TT && - !amdgpu_gtt_mgr_is_allocated(dst->mem)) { + !amdgpu_gtt_mgr_has_gart_addr(dst->mem)) { r = amdgpu_map_buffer(dst->bo, dst->mem, PFN_UP(cur_size + dst_page_offset), dst_node_start, 1, ring, @@ -467,9 +466,8 @@ error: return r; } -static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, +static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict, + struct ttm_operation_ctx *ctx, struct ttm_mem_reg *new_mem) { struct amdgpu_device *adev; @@ -489,8 +487,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, placements.fpfn = 0; placements.lpfn = 0; placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT; - r = ttm_bo_mem_space(bo, &placement, &tmp_mem, - interruptible, no_wait_gpu); + r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx); if (unlikely(r)) { return r; } @@ -504,19 +501,18 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, if (unlikely(r)) { goto out_cleanup; } - r = amdgpu_move_blit(bo, true, no_wait_gpu, &tmp_mem, old_mem); + r = amdgpu_move_blit(bo, true, ctx->no_wait_gpu, &tmp_mem, old_mem); if (unlikely(r)) { goto out_cleanup; } - r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, new_mem); + r = ttm_bo_move_ttm(bo, ctx->interruptible, ctx->no_wait_gpu, new_mem); out_cleanup: ttm_bo_mem_put(bo, &tmp_mem); return r; } -static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, +static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict, + struct ttm_operation_ctx *ctx, struct ttm_mem_reg *new_mem) { struct amdgpu_device *adev; @@ -536,16 +532,15 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, placements.fpfn = 0; placements.lpfn = 0; placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT; - r = ttm_bo_mem_space(bo, &placement, &tmp_mem, - interruptible, no_wait_gpu); + r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx); if (unlikely(r)) { return r; } - r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, &tmp_mem); + r = ttm_bo_move_ttm(bo, ctx->interruptible, ctx->no_wait_gpu, &tmp_mem); if (unlikely(r)) { goto out_cleanup; } - r = amdgpu_move_blit(bo, true, no_wait_gpu, new_mem, old_mem); + r = amdgpu_move_blit(bo, true, ctx->no_wait_gpu, new_mem, old_mem); if (unlikely(r)) { goto out_cleanup; } @@ -554,10 +549,9 @@ out_cleanup: return r; } -static int amdgpu_bo_move(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, - struct ttm_mem_reg *new_mem) +static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict, + struct ttm_operation_ctx *ctx, + struct ttm_mem_reg *new_mem) { struct amdgpu_device *adev; struct amdgpu_bo *abo; @@ -592,19 +586,19 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, if (old_mem->mem_type == TTM_PL_VRAM && new_mem->mem_type == TTM_PL_SYSTEM) { - r = amdgpu_move_vram_ram(bo, evict, interruptible, - no_wait_gpu, new_mem); + r = amdgpu_move_vram_ram(bo, evict, ctx, new_mem); } else if (old_mem->mem_type == TTM_PL_SYSTEM && new_mem->mem_type == TTM_PL_VRAM) { - r = amdgpu_move_ram_vram(bo, evict, interruptible, - no_wait_gpu, new_mem); + r = amdgpu_move_ram_vram(bo, evict, ctx, new_mem); } else { - r = amdgpu_move_blit(bo, evict, no_wait_gpu, new_mem, old_mem); + r = amdgpu_move_blit(bo, evict, ctx->no_wait_gpu, + new_mem, old_mem); } if (r) { memcpy: - r = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, new_mem); + r = ttm_bo_move_memcpy(bo, ctx->interruptible, + ctx->no_wait_gpu, new_mem); if (r) { return r; } @@ -690,7 +684,6 @@ struct amdgpu_ttm_tt { struct list_head guptasks; atomic_t mmu_invalidations; uint32_t last_set_pages; - struct list_head list; }; int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages) @@ -861,44 +854,35 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm, bo_mem->mem_type == AMDGPU_PL_OA) return -EINVAL; - if (!amdgpu_gtt_mgr_is_allocated(bo_mem)) + if (!amdgpu_gtt_mgr_has_gart_addr(bo_mem)) { + gtt->offset = AMDGPU_BO_INVALID_OFFSET; return 0; + } - spin_lock(>t->adev->gtt_list_lock); flags = amdgpu_ttm_tt_pte_flags(gtt->adev, ttm, bo_mem); gtt->offset = (u64)bo_mem->start << PAGE_SHIFT; r = amdgpu_gart_bind(gtt->adev, gtt->offset, ttm->num_pages, ttm->pages, gtt->ttm.dma_address, flags); - if (r) { + if (r) DRM_ERROR("failed to bind %lu pages at 0x%08llX\n", ttm->num_pages, gtt->offset); - goto error_gart_bind; - } - - list_add_tail(>t->list, >t->adev->gtt_list); -error_gart_bind: - spin_unlock(>t->adev->gtt_list_lock); return r; } -bool amdgpu_ttm_is_bound(struct ttm_tt *ttm) -{ - struct amdgpu_ttm_tt *gtt = (void *)ttm; - - return gtt && !list_empty(>t->list); -} - -int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem) +int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev); - struct ttm_tt *ttm = bo->ttm; + struct ttm_operation_ctx ctx = { false, false }; + struct amdgpu_ttm_tt *gtt = (void*)bo->ttm; struct ttm_mem_reg tmp; struct ttm_placement placement; struct ttm_place placements; + uint64_t flags; int r; - if (!ttm || amdgpu_ttm_is_bound(ttm)) + if (bo->mem.mem_type != TTM_PL_TT || + amdgpu_gtt_mgr_has_gart_addr(&bo->mem)) return 0; tmp = bo->mem; @@ -912,43 +896,44 @@ int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem) placements.flags = (bo->mem.placement & ~TTM_PL_MASK_MEM) | TTM_PL_FLAG_TT; - r = ttm_bo_mem_space(bo, &placement, &tmp, true, false); + r = ttm_bo_mem_space(bo, &placement, &tmp, &ctx); if (unlikely(r)) return r; - r = ttm_bo_move_ttm(bo, true, false, &tmp); - if (unlikely(r)) + flags = amdgpu_ttm_tt_pte_flags(adev, bo->ttm, &tmp); + gtt->offset = (u64)tmp.start << PAGE_SHIFT; + r = amdgpu_gart_bind(adev, gtt->offset, bo->ttm->num_pages, + bo->ttm->pages, gtt->ttm.dma_address, flags); + if (unlikely(r)) { ttm_bo_mem_put(bo, &tmp); - else - bo->offset = (bo->mem.start << PAGE_SHIFT) + - bo->bdev->man[bo->mem.mem_type].gpu_offset; + return r; + } - return r; + ttm_bo_mem_put(bo, &bo->mem); + bo->mem = tmp; + bo->offset = (bo->mem.start << PAGE_SHIFT) + + bo->bdev->man[bo->mem.mem_type].gpu_offset; + + return 0; } -int amdgpu_ttm_recover_gart(struct amdgpu_device *adev) +int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo) { - struct amdgpu_ttm_tt *gtt, *tmp; - struct ttm_mem_reg bo_mem; + struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev); + struct amdgpu_ttm_tt *gtt = (void *)tbo->ttm; uint64_t flags; int r; - bo_mem.mem_type = TTM_PL_TT; - spin_lock(&adev->gtt_list_lock); - list_for_each_entry_safe(gtt, tmp, &adev->gtt_list, list) { - flags = amdgpu_ttm_tt_pte_flags(gtt->adev, >t->ttm.ttm, &bo_mem); - r = amdgpu_gart_bind(adev, gtt->offset, gtt->ttm.ttm.num_pages, - gtt->ttm.ttm.pages, gtt->ttm.dma_address, - flags); - if (r) { - spin_unlock(&adev->gtt_list_lock); - DRM_ERROR("failed to bind %lu pages at 0x%08llX\n", - gtt->ttm.ttm.num_pages, gtt->offset); - return r; - } - } - spin_unlock(&adev->gtt_list_lock); - return 0; + if (!gtt) + return 0; + + flags = amdgpu_ttm_tt_pte_flags(adev, >t->ttm.ttm, &tbo->mem); + r = amdgpu_gart_bind(adev, gtt->offset, gtt->ttm.ttm.num_pages, + gtt->ttm.ttm.pages, gtt->ttm.dma_address, flags); + if (r) + DRM_ERROR("failed to bind %lu pages at 0x%08llX\n", + gtt->ttm.ttm.num_pages, gtt->offset); + return r; } static int amdgpu_ttm_backend_unbind(struct ttm_tt *ttm) @@ -959,20 +944,14 @@ static int amdgpu_ttm_backend_unbind(struct ttm_tt *ttm) if (gtt->userptr) amdgpu_ttm_tt_unpin_userptr(ttm); - if (!amdgpu_ttm_is_bound(ttm)) + if (gtt->offset == AMDGPU_BO_INVALID_OFFSET) return 0; /* unbind shouldn't be done for GDS/GWS/OA in ttm_bo_clean_mm */ - spin_lock(>t->adev->gtt_list_lock); r = amdgpu_gart_unbind(gtt->adev, gtt->offset, ttm->num_pages); - if (r) { + if (r) DRM_ERROR("failed to unbind %lu pages at 0x%08llX\n", gtt->ttm.ttm.num_pages, gtt->offset); - goto error_unbind; - } - list_del_init(>t->list); -error_unbind: - spin_unlock(>t->adev->gtt_list_lock); return r; } @@ -1009,7 +988,6 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_bo_device *bdev, kfree(gtt); return NULL; } - INIT_LIST_HEAD(>t->list); return >t->ttm.ttm; } @@ -1348,10 +1326,13 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) DRM_INFO("amdgpu: %uM of VRAM memory ready\n", (unsigned) (adev->mc.real_vram_size / (1024 * 1024))); - if (amdgpu_gtt_size == -1) - gtt_size = max((AMDGPU_DEFAULT_GTT_SIZE_MB << 20), - adev->mc.mc_vram_size); - else + if (amdgpu_gtt_size == -1) { + struct sysinfo si; + + si_meminfo(&si); + gtt_size = max(AMDGPU_DEFAULT_GTT_SIZE_MB << 20, + (uint64_t)si.totalram * si.mem_unit * 3/4); + } else gtt_size = (uint64_t)amdgpu_gtt_size << 20; r = ttm_bo_init_mm(&adev->mman.bdev, TTM_PL_TT, gtt_size >> PAGE_SHIFT); if (r) { @@ -1410,19 +1391,13 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) void amdgpu_ttm_fini(struct amdgpu_device *adev) { - int r; - if (!adev->mman.initialized) return; + amdgpu_ttm_debugfs_fini(adev); - if (adev->stolen_vga_memory) { - r = amdgpu_bo_reserve(adev->stolen_vga_memory, true); - if (r == 0) { - amdgpu_bo_unpin(adev->stolen_vga_memory); - amdgpu_bo_unreserve(adev->stolen_vga_memory); - } - amdgpu_bo_unref(&adev->stolen_vga_memory); - } + amdgpu_bo_free_kernel(&adev->stolen_vga_memory, NULL, NULL); + amdgpu_fw_reserve_vram_fini(adev); + ttm_bo_clean_mm(&adev->mman.bdev, TTM_PL_VRAM); ttm_bo_clean_mm(&adev->mman.bdev, TTM_PL_TT); if (adev->gds.mem.total_size) @@ -1432,7 +1407,6 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev) if (adev->gds.oa.total_size) ttm_bo_clean_mm(&adev->mman.bdev, AMDGPU_PL_OA); ttm_bo_device_release(&adev->mman.bdev); - amdgpu_gart_fini(adev); amdgpu_ttm_global_fini(adev); adev->mman.initialized = false; DRM_INFO("amdgpu: ttm finalized\n"); @@ -1628,7 +1602,7 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo, } if (bo->tbo.mem.mem_type == TTM_PL_TT) { - r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem); + r = amdgpu_ttm_alloc_gart(&bo->tbo); if (r) return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index abd4084982a3..4f9433e61406 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -67,8 +67,9 @@ struct amdgpu_copy_mem { extern const struct ttm_mem_type_manager_func amdgpu_gtt_mgr_func; extern const struct ttm_mem_type_manager_func amdgpu_vram_mgr_func; -bool amdgpu_gtt_mgr_is_allocated(struct ttm_mem_reg *mem); +bool amdgpu_gtt_mgr_has_gart_addr(struct ttm_mem_reg *mem); uint64_t amdgpu_gtt_mgr_usage(struct ttm_mem_type_manager *man); +int amdgpu_gtt_mgr_recover(struct ttm_mem_type_manager *man); uint64_t amdgpu_vram_mgr_usage(struct ttm_mem_type_manager *man); uint64_t amdgpu_vram_mgr_vis_usage(struct ttm_mem_type_manager *man); @@ -90,9 +91,8 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo, struct dma_fence **fence); int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma); -bool amdgpu_ttm_is_bound(struct ttm_tt *ttm); -int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem); -int amdgpu_ttm_recover_gart(struct amdgpu_device *adev); +int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo); +int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo); int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages); void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c index 65649026b836..474f88fbafce 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c @@ -359,7 +359,6 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode, int amdgpu_ucode_init_bo(struct amdgpu_device *adev) { - struct amdgpu_bo **bo = &adev->firmware.fw_buf; uint64_t fw_offset = 0; int i, err; struct amdgpu_firmware_info *ucode = NULL; @@ -370,36 +369,16 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev) return 0; } - if (!amdgpu_sriov_vf(adev) || !adev->in_sriov_reset) { - err = amdgpu_bo_create(adev, adev->firmware.fw_size, PAGE_SIZE, true, + if (!adev->in_gpu_reset) { + err = amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE, amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT, - AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS, - NULL, NULL, 0, bo); + &adev->firmware.fw_buf, + &adev->firmware.fw_buf_mc, + &adev->firmware.fw_buf_ptr); if (err) { - dev_err(adev->dev, "(%d) Firmware buffer allocate failed\n", err); + dev_err(adev->dev, "failed to create kernel buffer for firmware.fw_buf\n"); goto failed; } - - err = amdgpu_bo_reserve(*bo, false); - if (err) { - dev_err(adev->dev, "(%d) Firmware buffer reserve failed\n", err); - goto failed_reserve; - } - - err = amdgpu_bo_pin(*bo, amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT, - &adev->firmware.fw_buf_mc); - if (err) { - dev_err(adev->dev, "(%d) Firmware buffer pin failed\n", err); - goto failed_pin; - } - - err = amdgpu_bo_kmap(*bo, &adev->firmware.fw_buf_ptr); - if (err) { - dev_err(adev->dev, "(%d) Firmware buffer kmap failed\n", err); - goto failed_kmap; - } - - amdgpu_bo_unreserve(*bo); } memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size); @@ -436,12 +415,6 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev) } return 0; -failed_kmap: - amdgpu_bo_unpin(*bo); -failed_pin: - amdgpu_bo_unreserve(*bo); -failed_reserve: - amdgpu_bo_unref(bo); failed: if (err) adev->firmware.load_type = AMDGPU_FW_LOAD_DIRECT; @@ -464,8 +437,10 @@ int amdgpu_ucode_fini_bo(struct amdgpu_device *adev) ucode->kaddr = NULL; } } - amdgpu_bo_unref(&adev->firmware.fw_buf); - adev->firmware.fw_buf = NULL; + + amdgpu_bo_free_kernel(&adev->firmware.fw_buf, + &adev->firmware.fw_buf_mc, + &adev->firmware.fw_buf_ptr); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index e8bd50cf9785..2f2a9e17fdb4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -232,7 +232,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) ring = &adev->uvd.ring; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r != 0) { DRM_ERROR("Failed setting up UVD run queue.\n"); return r; @@ -408,6 +408,7 @@ static u64 amdgpu_uvd_get_addr_from_ctx(struct amdgpu_uvd_cs_ctx *ctx) */ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx) { + struct ttm_operation_ctx tctx = { false, false }; struct amdgpu_bo_va_mapping *mapping; struct amdgpu_bo *bo; uint32_t cmd; @@ -430,7 +431,7 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx) } amdgpu_uvd_force_into_uvd_segment(bo); - r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &tctx); } return r; @@ -949,6 +950,7 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx) static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, bool direct, struct dma_fence **fence) { + struct ttm_operation_ctx ctx = { true, false }; struct ttm_validate_buffer tv; struct ww_acquire_ctx ticket; struct list_head head; @@ -975,7 +977,7 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, amdgpu_uvd_force_into_uvd_segment(bo); } - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (r) goto err; @@ -1218,7 +1220,7 @@ int amdgpu_uvd_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h index 3553b92bf69a..845eea993f75 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h @@ -31,6 +31,10 @@ #define AMDGPU_UVD_SESSION_SIZE (50*1024) #define AMDGPU_UVD_FIRMWARE_OFFSET 256 +#define AMDGPU_UVD_FIRMWARE_SIZE(adev) \ + (AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(((const struct common_firmware_header *)(adev)->uvd.fw->data)->ucode_size_bytes) + \ + 8) - AMDGPU_UVD_FIRMWARE_OFFSET) + struct amdgpu_uvd { struct amdgpu_bo *vcpu_bo; void *cpu_addr; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 2918de2f39ec..ba6d846b08ff 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -176,7 +176,7 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) ring = &adev->vce.ring[0]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->vce.entity, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r != 0) { DRM_ERROR("Failed setting up VCE run queue.\n"); return r; @@ -544,6 +544,55 @@ err: } /** + * amdgpu_vce_cs_validate_bo - make sure not to cross 4GB boundary + * + * @p: parser context + * @lo: address of lower dword + * @hi: address of higher dword + * @size: minimum size + * @index: bs/fb index + * + * Make sure that no BO cross a 4GB boundary. + */ +static int amdgpu_vce_validate_bo(struct amdgpu_cs_parser *p, uint32_t ib_idx, + int lo, int hi, unsigned size, int32_t index) +{ + int64_t offset = ((uint64_t)size) * ((int64_t)index); + struct ttm_operation_ctx ctx = { false, false }; + struct amdgpu_bo_va_mapping *mapping; + unsigned i, fpfn, lpfn; + struct amdgpu_bo *bo; + uint64_t addr; + int r; + + addr = ((uint64_t)amdgpu_get_ib_value(p, ib_idx, lo)) | + ((uint64_t)amdgpu_get_ib_value(p, ib_idx, hi)) << 32; + if (index >= 0) { + addr += offset; + fpfn = PAGE_ALIGN(offset) >> PAGE_SHIFT; + lpfn = 0x100000000ULL >> PAGE_SHIFT; + } else { + fpfn = 0; + lpfn = (0x100000000ULL - PAGE_ALIGN(offset)) >> PAGE_SHIFT; + } + + r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping); + if (r) { + DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n", + addr, lo, hi, size, index); + return r; + } + + for (i = 0; i < bo->placement.num_placement; ++i) { + bo->placements[i].fpfn = max(bo->placements[i].fpfn, fpfn); + bo->placements[i].lpfn = bo->placements[i].fpfn ? + min(bo->placements[i].fpfn, lpfn) : lpfn; + } + return ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); +} + + +/** * amdgpu_vce_cs_reloc - command submission relocation * * @p: parser context @@ -648,12 +697,13 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) uint32_t allocated = 0; uint32_t tmp, handle = 0; uint32_t *size = &tmp; - int i, r = 0, idx = 0; + unsigned idx; + int i, r = 0; p->job->vm = NULL; ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo); - while (idx < ib->length_dw) { + for (idx = 0; idx < ib->length_dw;) { uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx); uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1); @@ -664,6 +714,54 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) } switch (cmd) { + case 0x00000002: /* task info */ + fb_idx = amdgpu_get_ib_value(p, ib_idx, idx + 6); + bs_idx = amdgpu_get_ib_value(p, ib_idx, idx + 7); + break; + + case 0x03000001: /* encode */ + r = amdgpu_vce_validate_bo(p, ib_idx, idx + 10, + idx + 9, 0, 0); + if (r) + goto out; + + r = amdgpu_vce_validate_bo(p, ib_idx, idx + 12, + idx + 11, 0, 0); + if (r) + goto out; + break; + + case 0x05000001: /* context buffer */ + r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3, + idx + 2, 0, 0); + if (r) + goto out; + break; + + case 0x05000004: /* video bitstream buffer */ + tmp = amdgpu_get_ib_value(p, ib_idx, idx + 4); + r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3, idx + 2, + tmp, bs_idx); + if (r) + goto out; + break; + + case 0x05000005: /* feedback buffer */ + r = amdgpu_vce_validate_bo(p, ib_idx, idx + 3, idx + 2, + 4096, fb_idx); + if (r) + goto out; + break; + } + + idx += len / 4; + } + + for (idx = 0; idx < ib->length_dw;) { + uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx); + uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1); + + switch (cmd) { case 0x00000001: /* session */ handle = amdgpu_get_ib_value(p, ib_idx, idx + 2); session_idx = amdgpu_vce_validate_handle(p, handle, @@ -954,7 +1052,7 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring) } if (i < timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed\n", @@ -999,7 +1097,7 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } error: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 041e0121590c..d7ba048c2f80 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -35,8 +35,8 @@ #include "soc15d.h" #include "soc15_common.h" -#include "vega10/soc15ip.h" -#include "raven1/VCN/vcn_1_0_offset.h" +#include "soc15ip.h" +#include "vcn/vcn_1_0_offset.h" /* 1 second timeout */ #define VCN_IDLE_TIMEOUT msecs_to_jiffies(1000) @@ -106,7 +106,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) ring = &adev->vcn.ring_dec; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->vcn.entity_dec, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r != 0) { DRM_ERROR("Failed setting up VCN dec run queue.\n"); return r; @@ -115,7 +115,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) ring = &adev->vcn.ring_enc[0]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->vcn.entity_enc, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r != 0) { DRM_ERROR("Failed setting up VCN enc run queue.\n"); return r; @@ -261,7 +261,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", @@ -274,6 +274,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, bool direct, struct dma_fence **fence) { + struct ttm_operation_ctx ctx = { true, false }; struct ttm_validate_buffer tv; struct ww_acquire_ctx ticket; struct list_head head; @@ -294,7 +295,7 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *b if (r) return r; - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (r) goto err; @@ -467,7 +468,7 @@ int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } @@ -500,7 +501,7 @@ int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed\n", @@ -643,7 +644,7 @@ int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } error: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 6738df836a70..e7dfb7b44b4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -24,6 +24,14 @@ #include "amdgpu.h" #define MAX_KIQ_REG_WAIT 100000000 /* in usecs */ +bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev) +{ + /* By now all MMIO pages except mailbox are blocked */ + /* if blocking is enabled in hypervisor. Choose the */ + /* SCRATCH_REG0 to test. */ + return RREG32_NO_KIQ(0xc040) == 0xffffffff; +} + int amdgpu_allocate_static_csa(struct amdgpu_device *adev) { int r; @@ -39,6 +47,12 @@ int amdgpu_allocate_static_csa(struct amdgpu_device *adev) return 0; } +void amdgpu_free_static_csa(struct amdgpu_device *adev) { + amdgpu_bo_free_kernel(&adev->virt.csa_obj, + &adev->virt.csa_vmid0_addr, + NULL); +} + /* * amdgpu_map_static_csa should be called during amdgpu_vm_init * it maps virtual address "AMDGPU_VA_RESERVED_SIZE - AMDGPU_CSA_SIZE" @@ -107,8 +121,6 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev) adev->enable_virtual_display = true; adev->cg_flags = 0; adev->pg_flags = 0; - - mutex_init(&adev->virt.lock_reset); } uint32_t amdgpu_virt_kiq_rreg(struct amdgpu_device *adev, uint32_t reg) @@ -228,6 +240,22 @@ int amdgpu_virt_reset_gpu(struct amdgpu_device *adev) } /** + * amdgpu_virt_wait_reset() - wait for reset gpu completed + * @amdgpu: amdgpu device. + * Wait for GPU reset completed. + * Return: Zero if reset success, otherwise will return error. + */ +int amdgpu_virt_wait_reset(struct amdgpu_device *adev) +{ + struct amdgpu_virt *virt = &adev->virt; + + if (!virt->ops || !virt->ops->wait_reset) + return -EINVAL; + + return virt->ops->wait_reset(adev); +} + +/** * amdgpu_virt_alloc_mm_table() - alloc memory for mm table * @amdgpu: amdgpu device. * MM table is used by UVD and VCE for its initialization @@ -296,7 +324,6 @@ int amdgpu_virt_fw_reserve_get_checksum(void *obj, void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev) { - uint32_t pf2vf_ver = 0; uint32_t pf2vf_size = 0; uint32_t checksum = 0; uint32_t checkval; @@ -309,9 +336,9 @@ void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev) adev->virt.fw_reserve.p_pf2vf = (struct amdgim_pf2vf_info_header *)( adev->fw_vram_usage.va + AMDGIM_DATAEXCHANGE_OFFSET); - pf2vf_ver = adev->virt.fw_reserve.p_pf2vf->version; AMDGPU_FW_VRAM_PF2VF_READ(adev, header.size, &pf2vf_size); AMDGPU_FW_VRAM_PF2VF_READ(adev, checksum, &checksum); + AMDGPU_FW_VRAM_PF2VF_READ(adev, feature_flags, &adev->virt.gim_feature); /* pf2vf message must be in 4K */ if (pf2vf_size > 0 && pf2vf_size < 4096) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index b89d37fc406f..6a83425aa9ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -55,6 +55,7 @@ struct amdgpu_virt_ops { int (*req_full_gpu)(struct amdgpu_device *adev, bool init); int (*rel_full_gpu)(struct amdgpu_device *adev, bool init); int (*reset_gpu)(struct amdgpu_device *adev); + int (*wait_reset)(struct amdgpu_device *adev); void (*trans_msg)(struct amdgpu_device *adev, u32 req, u32 data1, u32 data2, u32 data3); }; @@ -80,6 +81,8 @@ enum AMDGIM_FEATURE_FLAG { AMDGIM_FEATURE_ERROR_LOG_COLLECT = 0x1, /* GIM supports feature of loading uCodes */ AMDGIM_FEATURE_GIM_LOAD_UCODES = 0x2, + /* VRAM LOST by GIM */ + AMDGIM_FEATURE_GIM_FLR_VRAMLOST = 0x4, }; struct amdgim_pf2vf_info_header { @@ -238,7 +241,6 @@ struct amdgpu_virt { uint64_t csa_vmid0_addr; bool chained_ib_support; uint32_t reg_val_offs; - struct mutex lock_reset; struct amdgpu_irq_src ack_irq; struct amdgpu_irq_src rcv_irq; struct work_struct flr_work; @@ -246,6 +248,7 @@ struct amdgpu_virt { const struct amdgpu_virt_ops *ops; struct amdgpu_vf_error_buffer vf_errors; struct amdgpu_virt_fw_reserve fw_reserve; + uint32_t gim_feature; }; #define AMDGPU_CSA_SIZE (8 * 1024) @@ -276,16 +279,18 @@ static inline bool is_virtual_machine(void) } struct amdgpu_vm; +bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev); int amdgpu_allocate_static_csa(struct amdgpu_device *adev); int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, struct amdgpu_bo_va **bo_va); +void amdgpu_free_static_csa(struct amdgpu_device *adev); void amdgpu_virt_init_setting(struct amdgpu_device *adev); uint32_t amdgpu_virt_kiq_rreg(struct amdgpu_device *adev, uint32_t reg); void amdgpu_virt_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v); int amdgpu_virt_request_full_gpu(struct amdgpu_device *adev, bool init); int amdgpu_virt_release_full_gpu(struct amdgpu_device *adev, bool init); int amdgpu_virt_reset_gpu(struct amdgpu_device *adev); -int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, struct amdgpu_job *job); +int amdgpu_virt_wait_reset(struct amdgpu_device *adev); int amdgpu_virt_alloc_mm_table(struct amdgpu_device *adev); void amdgpu_virt_free_mm_table(struct amdgpu_device *adev); int amdgpu_virt_fw_reserve_get_checksum(void *obj, unsigned long obj_size, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index c8c26f21993c..3ecdbdfb04dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -139,6 +139,24 @@ struct amdgpu_prt_cb { }; /** + * amdgpu_vm_level_shift - return the addr shift for each level + * + * @adev: amdgpu_device pointer + * + * Returns the number of bits the pfn needs to be right shifted for a level. + */ +static unsigned amdgpu_vm_level_shift(struct amdgpu_device *adev, + unsigned level) +{ + if (level != adev->vm_manager.num_level) + return 9 * (adev->vm_manager.num_level - level - 1) + + adev->vm_manager.block_size; + else + /* For the page tables on the leaves */ + return 0; +} + +/** * amdgpu_vm_num_entries - return the number of entries in a PD/PT * * @adev: amdgpu_device pointer @@ -148,17 +166,17 @@ struct amdgpu_prt_cb { static unsigned amdgpu_vm_num_entries(struct amdgpu_device *adev, unsigned level) { + unsigned shift = amdgpu_vm_level_shift(adev, 0); + if (level == 0) /* For the root directory */ - return adev->vm_manager.max_pfn >> - (adev->vm_manager.block_size * - adev->vm_manager.num_level); - else if (level == adev->vm_manager.num_level) + return round_up(adev->vm_manager.max_pfn, 1 << shift) >> shift; + else if (level != adev->vm_manager.num_level) + /* Everything in between */ + return 512; + else /* For the page tables on the leaves */ return AMDGPU_VM_PTE_COUNT(adev); - else - /* Everything in between */ - return 1 << adev->vm_manager.block_size; } /** @@ -288,8 +306,7 @@ static int amdgpu_vm_alloc_levels(struct amdgpu_device *adev, uint64_t saddr, uint64_t eaddr, unsigned level) { - unsigned shift = (adev->vm_manager.num_level - level) * - adev->vm_manager.block_size; + unsigned shift = amdgpu_vm_level_shift(adev, level); unsigned pt_idx, from, to; int r; u64 flags; @@ -471,7 +488,7 @@ static int amdgpu_vm_grab_reserved_vmid_locked(struct amdgpu_vm *vm, id->pd_gpu_addr = 0; tmp = amdgpu_sync_peek_fence(&id->active, ring); if (tmp) { - r = amdgpu_sync_fence(adev, sync, tmp); + r = amdgpu_sync_fence(adev, sync, tmp, false); return r; } } @@ -479,7 +496,7 @@ static int amdgpu_vm_grab_reserved_vmid_locked(struct amdgpu_vm *vm, /* Good we can use this VMID. Remember this submission as * user of the VMID. */ - r = amdgpu_sync_fence(ring->adev, &id->active, fence); + r = amdgpu_sync_fence(ring->adev, &id->active, fence, false); if (r) goto out; @@ -566,7 +583,7 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, } - r = amdgpu_sync_fence(ring->adev, sync, &array->base); + r = amdgpu_sync_fence(ring->adev, sync, &array->base, false); dma_fence_put(&array->base); if (r) goto error; @@ -609,7 +626,7 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, /* Good we can use this VMID. Remember this submission as * user of the VMID. */ - r = amdgpu_sync_fence(ring->adev, &id->active, fence); + r = amdgpu_sync_fence(ring->adev, &id->active, fence, false); if (r) goto error; @@ -629,7 +646,7 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, id = idle; /* Remember this submission as user of the VMID */ - r = amdgpu_sync_fence(ring->adev, &id->active, fence); + r = amdgpu_sync_fence(ring->adev, &id->active, fence, false); if (r) goto error; @@ -1302,18 +1319,19 @@ void amdgpu_vm_get_entry(struct amdgpu_pte_update_params *p, uint64_t addr, struct amdgpu_vm_pt **entry, struct amdgpu_vm_pt **parent) { - unsigned idx, level = p->adev->vm_manager.num_level; + unsigned level = 0; *parent = NULL; *entry = &p->vm->root; while ((*entry)->entries) { - idx = addr >> (p->adev->vm_manager.block_size * level--); + unsigned idx = addr >> amdgpu_vm_level_shift(p->adev, level++); + idx %= amdgpu_bo_size((*entry)->base.bo) / 8; *parent = *entry; *entry = &(*entry)->entries[idx]; } - if (level) + if (level != p->adev->vm_manager.num_level) *entry = NULL; } @@ -1639,7 +1657,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, addr = 0; } - r = amdgpu_sync_fence(adev, &job->sync, exclusive); + r = amdgpu_sync_fence(adev, &job->sync, exclusive, false); if (r) goto error_free; @@ -2556,47 +2574,57 @@ static uint32_t amdgpu_vm_get_block_size(uint64_t vm_size) } /** - * amdgpu_vm_set_fragment_size - adjust fragment size in PTE - * - * @adev: amdgpu_device pointer - * @fragment_size_default: the default fragment size if it's set auto - */ -void amdgpu_vm_set_fragment_size(struct amdgpu_device *adev, - uint32_t fragment_size_default) -{ - if (amdgpu_vm_fragment_size == -1) - adev->vm_manager.fragment_size = fragment_size_default; - else - adev->vm_manager.fragment_size = amdgpu_vm_fragment_size; -} - -/** * amdgpu_vm_adjust_size - adjust vm size, block size and fragment size * * @adev: amdgpu_device pointer * @vm_size: the default vm size if it's set auto */ -void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint64_t vm_size, - uint32_t fragment_size_default) +void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t vm_size, + uint32_t fragment_size_default, unsigned max_level, + unsigned max_bits) { - /* adjust vm size firstly */ - if (amdgpu_vm_size == -1) - adev->vm_manager.vm_size = vm_size; - else - adev->vm_manager.vm_size = amdgpu_vm_size; + uint64_t tmp; - /* block size depends on vm size */ - if (amdgpu_vm_block_size == -1) + /* adjust vm size first */ + if (amdgpu_vm_size != -1) { + unsigned max_size = 1 << (max_bits - 30); + + vm_size = amdgpu_vm_size; + if (vm_size > max_size) { + dev_warn(adev->dev, "VM size (%d) too large, max is %u GB\n", + amdgpu_vm_size, max_size); + vm_size = max_size; + } + } + + adev->vm_manager.max_pfn = (uint64_t)vm_size << 18; + + tmp = roundup_pow_of_two(adev->vm_manager.max_pfn); + if (amdgpu_vm_block_size != -1) + tmp >>= amdgpu_vm_block_size - 9; + tmp = DIV_ROUND_UP(fls64(tmp) - 1, 9) - 1; + adev->vm_manager.num_level = min(max_level, (unsigned)tmp); + + /* block size depends on vm size and hw setup*/ + if (amdgpu_vm_block_size != -1) adev->vm_manager.block_size = - amdgpu_vm_get_block_size(adev->vm_manager.vm_size); + min((unsigned)amdgpu_vm_block_size, max_bits + - AMDGPU_GPU_PAGE_SHIFT + - 9 * adev->vm_manager.num_level); + else if (adev->vm_manager.num_level > 1) + adev->vm_manager.block_size = 9; else - adev->vm_manager.block_size = amdgpu_vm_block_size; + adev->vm_manager.block_size = amdgpu_vm_get_block_size(tmp); - amdgpu_vm_set_fragment_size(adev, fragment_size_default); + if (amdgpu_vm_fragment_size == -1) + adev->vm_manager.fragment_size = fragment_size_default; + else + adev->vm_manager.fragment_size = amdgpu_vm_fragment_size; - DRM_INFO("vm size is %llu GB, block size is %u-bit, fragment size is %u-bit\n", - adev->vm_manager.vm_size, adev->vm_manager.block_size, - adev->vm_manager.fragment_size); + DRM_INFO("vm size is %u GB, %u levels, block size is %u-bit, fragment size is %u-bit\n", + vm_size, adev->vm_manager.num_level + 1, + adev->vm_manager.block_size, + adev->vm_manager.fragment_size); } /** @@ -2637,7 +2665,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, ring = adev->vm_manager.vm_pte_rings[ring_instance]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL]; r = amd_sched_entity_init(&ring->sched, &vm->entity, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index bae77353447b..43ea131dd411 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -96,6 +96,19 @@ struct amdgpu_bo_list_entry; /* hardcode that limit for now */ #define AMDGPU_VA_RESERVED_SIZE (8ULL << 20) +/* VA hole for 48bit addresses on Vega10 */ +#define AMDGPU_VA_HOLE_START 0x0000800000000000ULL +#define AMDGPU_VA_HOLE_END 0xffff800000000000ULL + +/* + * Hardware is programmed as if the hole doesn't exists with start and end + * address values. + * + * This mask is used to remove the upper 16bits of the VA and so come up with + * the linear addr value. + */ +#define AMDGPU_VA_HOLE_MASK 0x0000ffffffffffffULL + /* max vmids dedicated for process */ #define AMDGPU_VM_MAX_RESERVED_VMID 1 @@ -221,7 +234,6 @@ struct amdgpu_vm_manager { uint64_t max_pfn; uint32_t num_level; - uint64_t vm_size; uint32_t block_size; uint32_t fragment_size; /* vram base address for page table entry */ @@ -312,10 +324,9 @@ struct amdgpu_bo_va_mapping *amdgpu_vm_bo_lookup_mapping(struct amdgpu_vm *vm, uint64_t addr); void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va); -void amdgpu_vm_set_fragment_size(struct amdgpu_device *adev, - uint32_t fragment_size_default); -void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint64_t vm_size, - uint32_t fragment_size_default); +void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint32_t vm_size, + uint32_t fragment_size_default, unsigned max_level, + unsigned max_bits); int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring, struct amdgpu_job *job); diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c index b374653bd6cf..f9b2ce9a98f3 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c @@ -65,8 +65,15 @@ static int amdgpu_atombios_i2c_process_i2c_ch(struct amdgpu_i2c_chan *chan, args.ucRegIndex = buf[0]; if (num) num--; - if (num) - memcpy(&out, &buf[1], num); + if (num) { + if (buf) { + memcpy(&out, &buf[1], num); + } else { + DRM_ERROR("hw i2c: missing buf with num > 1\n"); + r = -EINVAL; + goto done; + } + } args.lpI2CDataOut = cpu_to_le16(out); } else { if (num > ATOM_MAX_HW_I2C_READ) { diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c index 68b505c768ad..f11c0aacf19f 100644 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c @@ -4540,9 +4540,9 @@ static int ci_set_mc_special_registers(struct amdgpu_device *adev, ((temp_reg & 0xffff0000)) | ((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16); } j++; + if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) return -EINVAL; - temp_reg = RREG32(mmMC_PMG_CMD_MRS); table->mc_reg_address[j].s1 = mmMC_PMG_CMD_MRS; table->mc_reg_address[j].s0 = mmMC_SEQ_PMG_CMD_MRS_LP; @@ -4553,10 +4553,10 @@ static int ci_set_mc_special_registers(struct amdgpu_device *adev, table->mc_reg_table_entry[k].mc_data[j] |= 0x100; } j++; - if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; if (adev->mc.vram_type != AMDGPU_VRAM_TYPE_GDDR5) { + if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; table->mc_reg_address[j].s1 = mmMC_PMG_AUTO_CMD; table->mc_reg_address[j].s0 = mmMC_PMG_AUTO_CMD; for (k = 0; k < table->num_entries; k++) { @@ -4564,8 +4564,6 @@ static int ci_set_mc_special_registers(struct amdgpu_device *adev, (table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16; } j++; - if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; } break; case mmMC_SEQ_RESERVE_M: @@ -4577,8 +4575,6 @@ static int ci_set_mc_special_registers(struct amdgpu_device *adev, (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); } j++; - if (j > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; break; default: break; @@ -6625,9 +6621,9 @@ static int ci_dpm_print_clock_levels(void *handle, for (i = 0; i < pcie_table->count; i++) size += sprintf(buf + size, "%d: %s %s\n", i, - (pcie_table->dpm_levels[i].value == 0) ? "2.5GB, x1" : - (pcie_table->dpm_levels[i].value == 1) ? "5.0GB, x16" : - (pcie_table->dpm_levels[i].value == 2) ? "8.0GB, x16" : "", + (pcie_table->dpm_levels[i].value == 0) ? "2.5GT/s, x1" : + (pcie_table->dpm_levels[i].value == 1) ? "5.0GT/s, x16" : + (pcie_table->dpm_levels[i].value == 2) ? "8.0GT/s, x16" : "", (i == now) ? "*" : ""); break; default: diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index a296f7bbe57c..8ba056a2a5da 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -757,72 +757,72 @@ static void cik_init_golden_registers(struct amdgpu_device *adev) case CHIP_BONAIRE: amdgpu_program_register_sequence(adev, bonaire_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(bonaire_mgcg_cgcg_init)); + ARRAY_SIZE(bonaire_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, bonaire_golden_registers, - (const u32)ARRAY_SIZE(bonaire_golden_registers)); + ARRAY_SIZE(bonaire_golden_registers)); amdgpu_program_register_sequence(adev, bonaire_golden_common_registers, - (const u32)ARRAY_SIZE(bonaire_golden_common_registers)); + ARRAY_SIZE(bonaire_golden_common_registers)); amdgpu_program_register_sequence(adev, bonaire_golden_spm_registers, - (const u32)ARRAY_SIZE(bonaire_golden_spm_registers)); + ARRAY_SIZE(bonaire_golden_spm_registers)); break; case CHIP_KABINI: amdgpu_program_register_sequence(adev, kalindi_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(kalindi_mgcg_cgcg_init)); + ARRAY_SIZE(kalindi_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, kalindi_golden_registers, - (const u32)ARRAY_SIZE(kalindi_golden_registers)); + ARRAY_SIZE(kalindi_golden_registers)); amdgpu_program_register_sequence(adev, kalindi_golden_common_registers, - (const u32)ARRAY_SIZE(kalindi_golden_common_registers)); + ARRAY_SIZE(kalindi_golden_common_registers)); amdgpu_program_register_sequence(adev, kalindi_golden_spm_registers, - (const u32)ARRAY_SIZE(kalindi_golden_spm_registers)); + ARRAY_SIZE(kalindi_golden_spm_registers)); break; case CHIP_MULLINS: amdgpu_program_register_sequence(adev, kalindi_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(kalindi_mgcg_cgcg_init)); + ARRAY_SIZE(kalindi_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, godavari_golden_registers, - (const u32)ARRAY_SIZE(godavari_golden_registers)); + ARRAY_SIZE(godavari_golden_registers)); amdgpu_program_register_sequence(adev, kalindi_golden_common_registers, - (const u32)ARRAY_SIZE(kalindi_golden_common_registers)); + ARRAY_SIZE(kalindi_golden_common_registers)); amdgpu_program_register_sequence(adev, kalindi_golden_spm_registers, - (const u32)ARRAY_SIZE(kalindi_golden_spm_registers)); + ARRAY_SIZE(kalindi_golden_spm_registers)); break; case CHIP_KAVERI: amdgpu_program_register_sequence(adev, spectre_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(spectre_mgcg_cgcg_init)); + ARRAY_SIZE(spectre_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, spectre_golden_registers, - (const u32)ARRAY_SIZE(spectre_golden_registers)); + ARRAY_SIZE(spectre_golden_registers)); amdgpu_program_register_sequence(adev, spectre_golden_common_registers, - (const u32)ARRAY_SIZE(spectre_golden_common_registers)); + ARRAY_SIZE(spectre_golden_common_registers)); amdgpu_program_register_sequence(adev, spectre_golden_spm_registers, - (const u32)ARRAY_SIZE(spectre_golden_spm_registers)); + ARRAY_SIZE(spectre_golden_spm_registers)); break; case CHIP_HAWAII: amdgpu_program_register_sequence(adev, hawaii_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(hawaii_mgcg_cgcg_init)); + ARRAY_SIZE(hawaii_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, hawaii_golden_registers, - (const u32)ARRAY_SIZE(hawaii_golden_registers)); + ARRAY_SIZE(hawaii_golden_registers)); amdgpu_program_register_sequence(adev, hawaii_golden_common_registers, - (const u32)ARRAY_SIZE(hawaii_golden_common_registers)); + ARRAY_SIZE(hawaii_golden_common_registers)); amdgpu_program_register_sequence(adev, hawaii_golden_spm_registers, - (const u32)ARRAY_SIZE(hawaii_golden_spm_registers)); + ARRAY_SIZE(hawaii_golden_spm_registers)); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 60cecd117705..ed26dcbc4f79 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -657,7 +657,7 @@ static int cik_sdma_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", ring->idx, tmp); @@ -724,7 +724,7 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 4e519dc42916..a397111c2ced 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -147,18 +147,18 @@ static void dce_v10_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_FIJI: amdgpu_program_register_sequence(adev, fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(fiji_mgcg_cgcg_init)); + ARRAY_SIZE(fiji_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_fiji_a10, - (const u32)ARRAY_SIZE(golden_settings_fiji_a10)); + ARRAY_SIZE(golden_settings_fiji_a10)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tonga_mgcg_cgcg_init)); + ARRAY_SIZE(tonga_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_tonga_a11, - (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); + ARRAY_SIZE(golden_settings_tonga_a11)); break; default: break; @@ -2773,7 +2773,6 @@ static int dce_v10_0_early_init(void *handle) adev->audio_endpt_wreg = &dce_v10_0_audio_endpt_wreg; dce_v10_0_set_display_funcs(adev); - dce_v10_0_set_irq_funcs(adev); adev->mode_info.num_crtc = dce_v10_0_get_num_crtc(adev); @@ -2788,6 +2787,8 @@ static int dce_v10_0_early_init(void *handle) return -EINVAL; } + dce_v10_0_set_irq_funcs(adev); + return 0; } @@ -3635,13 +3636,16 @@ static const struct amdgpu_irq_src_funcs dce_v10_0_hpd_irq_funcs = { static void dce_v10_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST; + if (adev->mode_info.num_crtc > 0) + adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_VLINE1 + adev->mode_info.num_crtc; + else + adev->crtc_irq.num_types = 0; adev->crtc_irq.funcs = &dce_v10_0_crtc_irq_funcs; - adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST; + adev->pageflip_irq.num_types = adev->mode_info.num_crtc; adev->pageflip_irq.funcs = &dce_v10_0_pageflip_irq_funcs; - adev->hpd_irq.num_types = AMDGPU_HPD_LAST; + adev->hpd_irq.num_types = adev->mode_info.num_hpd; adev->hpd_irq.funcs = &dce_v10_0_hpd_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 11edc75edaa9..67e670989e81 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -156,26 +156,26 @@ static void dce_v11_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, cz_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(cz_mgcg_cgcg_init)); + ARRAY_SIZE(cz_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, cz_golden_settings_a11, - (const u32)ARRAY_SIZE(cz_golden_settings_a11)); + ARRAY_SIZE(cz_golden_settings_a11)); break; case CHIP_STONEY: amdgpu_program_register_sequence(adev, stoney_golden_settings_a11, - (const u32)ARRAY_SIZE(stoney_golden_settings_a11)); + ARRAY_SIZE(stoney_golden_settings_a11)); break; case CHIP_POLARIS11: case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, polaris11_golden_settings_a11, - (const u32)ARRAY_SIZE(polaris11_golden_settings_a11)); + ARRAY_SIZE(polaris11_golden_settings_a11)); break; case CHIP_POLARIS10: amdgpu_program_register_sequence(adev, polaris10_golden_settings_a11, - (const u32)ARRAY_SIZE(polaris10_golden_settings_a11)); + ARRAY_SIZE(polaris10_golden_settings_a11)); break; default: break; @@ -2876,7 +2876,6 @@ static int dce_v11_0_early_init(void *handle) adev->audio_endpt_wreg = &dce_v11_0_audio_endpt_wreg; dce_v11_0_set_display_funcs(adev); - dce_v11_0_set_irq_funcs(adev); adev->mode_info.num_crtc = dce_v11_0_get_num_crtc(adev); @@ -2903,6 +2902,8 @@ static int dce_v11_0_early_init(void *handle) return -EINVAL; } + dce_v11_0_set_irq_funcs(adev); + return 0; } @@ -3759,13 +3760,16 @@ static const struct amdgpu_irq_src_funcs dce_v11_0_hpd_irq_funcs = { static void dce_v11_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST; + if (adev->mode_info.num_crtc > 0) + adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_VLINE1 + adev->mode_info.num_crtc; + else + adev->crtc_irq.num_types = 0; adev->crtc_irq.funcs = &dce_v11_0_crtc_irq_funcs; - adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST; + adev->pageflip_irq.num_types = adev->mode_info.num_crtc; adev->pageflip_irq.funcs = &dce_v11_0_pageflip_irq_funcs; - adev->hpd_irq.num_types = AMDGPU_HPD_LAST; + adev->hpd_irq.num_types = adev->mode_info.num_hpd; adev->hpd_irq.funcs = &dce_v11_0_hpd_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c index a51e35f824a1..bd2c4f727df6 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c @@ -2639,7 +2639,6 @@ static int dce_v6_0_early_init(void *handle) adev->audio_endpt_wreg = &dce_v6_0_audio_endpt_wreg; dce_v6_0_set_display_funcs(adev); - dce_v6_0_set_irq_funcs(adev); adev->mode_info.num_crtc = dce_v6_0_get_num_crtc(adev); @@ -2658,6 +2657,8 @@ static int dce_v6_0_early_init(void *handle) return -EINVAL; } + dce_v6_0_set_irq_funcs(adev); + return 0; } @@ -3441,13 +3442,16 @@ static const struct amdgpu_irq_src_funcs dce_v6_0_hpd_irq_funcs = { static void dce_v6_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST; + if (adev->mode_info.num_crtc > 0) + adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_VLINE1 + adev->mode_info.num_crtc; + else + adev->crtc_irq.num_types = 0; adev->crtc_irq.funcs = &dce_v6_0_crtc_irq_funcs; - adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST; + adev->pageflip_irq.num_types = adev->mode_info.num_crtc; adev->pageflip_irq.funcs = &dce_v6_0_pageflip_irq_funcs; - adev->hpd_irq.num_types = AMDGPU_HPD_LAST; + adev->hpd_irq.num_types = adev->mode_info.num_hpd; adev->hpd_irq.funcs = &dce_v6_0_hpd_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 9cf14b8b2db9..c008dc030687 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2664,7 +2664,6 @@ static int dce_v8_0_early_init(void *handle) adev->audio_endpt_wreg = &dce_v8_0_audio_endpt_wreg; dce_v8_0_set_display_funcs(adev); - dce_v8_0_set_irq_funcs(adev); adev->mode_info.num_crtc = dce_v8_0_get_num_crtc(adev); @@ -2688,6 +2687,8 @@ static int dce_v8_0_early_init(void *handle) return -EINVAL; } + dce_v8_0_set_irq_funcs(adev); + return 0; } @@ -3525,13 +3526,16 @@ static const struct amdgpu_irq_src_funcs dce_v8_0_hpd_irq_funcs = { static void dce_v8_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST; + if (adev->mode_info.num_crtc > 0) + adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_VLINE1 + adev->mode_info.num_crtc; + else + adev->crtc_irq.num_types = 0; adev->crtc_irq.funcs = &dce_v8_0_crtc_irq_funcs; - adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST; + adev->pageflip_irq.num_types = adev->mode_info.num_crtc; adev->pageflip_irq.funcs = &dce_v8_0_pageflip_irq_funcs; - adev->hpd_irq.num_types = AMDGPU_HPD_LAST; + adev->hpd_irq.num_types = adev->mode_info.num_hpd; adev->hpd_irq.funcs = &dce_v8_0_hpd_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c index a8829af120c1..120dd3b26fc2 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c @@ -44,6 +44,9 @@ static void dce_virtual_set_display_funcs(struct amdgpu_device *adev); static void dce_virtual_set_irq_funcs(struct amdgpu_device *adev); static int dce_virtual_connector_encoder_init(struct amdgpu_device *adev, int index); +static void dce_virtual_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev, + int crtc, + enum amdgpu_interrupt_state state); /** * dce_virtual_vblank_wait - vblank wait asic callback. @@ -437,6 +440,8 @@ static int dce_virtual_sw_fini(void *handle) drm_kms_helper_poll_fini(adev->ddev); drm_mode_config_cleanup(adev->ddev); + /* clear crtcs pointer to avoid dce irq finish routine access freed data */ + memset(adev->mode_info.crtcs, 0, sizeof(adev->mode_info.crtcs[0]) * AMDGPU_MAX_CRTCS); adev->mode_info.mode_config_initialized = false; return 0; } @@ -489,6 +494,13 @@ static int dce_virtual_hw_init(void *handle) static int dce_virtual_hw_fini(void *handle) { + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i = 0; + + for (i = 0; i<adev->mode_info.num_crtc; i++) + if (adev->mode_info.crtcs[i]) + dce_virtual_set_crtc_vblank_interrupt_state(adev, i, AMDGPU_IRQ_STATE_DISABLE); + return 0; } @@ -723,7 +735,7 @@ static void dce_virtual_set_crtc_vblank_interrupt_state(struct amdgpu_device *ad int crtc, enum amdgpu_interrupt_state state) { - if (crtc >= adev->mode_info.num_crtc) { + if (crtc >= adev->mode_info.num_crtc || !adev->mode_info.crtcs[crtc]) { DRM_DEBUG("invalid crtc %d\n", crtc); return; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index dbbe986f90f2..edef17d93527 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -1798,7 +1798,7 @@ static int gfx_v6_0_ring_test_ring(struct amdgpu_ring *ring) DRM_UDELAY(1); } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (scratch(0x%04X)=0x%08X)\n", ring->idx, scratch, tmp); @@ -1951,7 +1951,7 @@ static int gfx_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", @@ -2962,25 +2962,7 @@ static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1)); buffer[count++] = cpu_to_le32(mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); - - switch (adev->asic_type) { - case CHIP_TAHITI: - case CHIP_PITCAIRN: - buffer[count++] = cpu_to_le32(0x2a00126a); - break; - case CHIP_VERDE: - buffer[count++] = cpu_to_le32(0x0000124a); - break; - case CHIP_OLAND: - buffer[count++] = cpu_to_le32(0x00000082); - break; - case CHIP_HAINAN: - buffer[count++] = cpu_to_le32(0x00000000); - break; - default: - buffer[count++] = cpu_to_le32(0x00000000); - break; - } + buffer[count++] = cpu_to_le32(adev->gfx.config.rb_config[0][0].raster_config); buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0)); buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 419ba0ce7ee5..83d94c23aa78 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -2085,7 +2085,7 @@ static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring) DRM_UDELAY(1); } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (scratch(0x%04X)=0x%08X)\n", ring->idx, scratch, tmp); @@ -2365,7 +2365,7 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", @@ -2551,29 +2551,8 @@ static int gfx_v7_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); amdgpu_ring_write(ring, mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); - switch (adev->asic_type) { - case CHIP_BONAIRE: - amdgpu_ring_write(ring, 0x16000012); - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_KAVERI: - amdgpu_ring_write(ring, 0x00000000); /* XXX */ - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_KABINI: - case CHIP_MULLINS: - amdgpu_ring_write(ring, 0x00000000); /* XXX */ - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_HAWAII: - amdgpu_ring_write(ring, 0x3a00161a); - amdgpu_ring_write(ring, 0x0000002e); - break; - default: - amdgpu_ring_write(ring, 0x00000000); - amdgpu_ring_write(ring, 0x00000000); - break; - } + amdgpu_ring_write(ring, adev->gfx.config.rb_config[0][0].raster_config); + amdgpu_ring_write(ring, adev->gfx.config.rb_config[0][0].raster_config_1); amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); amdgpu_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 9ecdf621a74a..d02493cf9175 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -681,53 +681,53 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_TOPAZ: amdgpu_program_register_sequence(adev, iceland_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(iceland_mgcg_cgcg_init)); + ARRAY_SIZE(iceland_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_iceland_a11, - (const u32)ARRAY_SIZE(golden_settings_iceland_a11)); + ARRAY_SIZE(golden_settings_iceland_a11)); amdgpu_program_register_sequence(adev, iceland_golden_common_all, - (const u32)ARRAY_SIZE(iceland_golden_common_all)); + ARRAY_SIZE(iceland_golden_common_all)); break; case CHIP_FIJI: amdgpu_program_register_sequence(adev, fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(fiji_mgcg_cgcg_init)); + ARRAY_SIZE(fiji_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_fiji_a10, - (const u32)ARRAY_SIZE(golden_settings_fiji_a10)); + ARRAY_SIZE(golden_settings_fiji_a10)); amdgpu_program_register_sequence(adev, fiji_golden_common_all, - (const u32)ARRAY_SIZE(fiji_golden_common_all)); + ARRAY_SIZE(fiji_golden_common_all)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tonga_mgcg_cgcg_init)); + ARRAY_SIZE(tonga_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_tonga_a11, - (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); + ARRAY_SIZE(golden_settings_tonga_a11)); amdgpu_program_register_sequence(adev, tonga_golden_common_all, - (const u32)ARRAY_SIZE(tonga_golden_common_all)); + ARRAY_SIZE(tonga_golden_common_all)); break; case CHIP_POLARIS11: case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); + ARRAY_SIZE(golden_settings_polaris11_a11)); amdgpu_program_register_sequence(adev, polaris11_golden_common_all, - (const u32)ARRAY_SIZE(polaris11_golden_common_all)); + ARRAY_SIZE(polaris11_golden_common_all)); break; case CHIP_POLARIS10: amdgpu_program_register_sequence(adev, golden_settings_polaris10_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris10_a11)); + ARRAY_SIZE(golden_settings_polaris10_a11)); amdgpu_program_register_sequence(adev, polaris10_golden_common_all, - (const u32)ARRAY_SIZE(polaris10_golden_common_all)); + ARRAY_SIZE(polaris10_golden_common_all)); WREG32_SMC(ixCG_ACLK_CNTL, 0x0000001C); if (adev->pdev->revision == 0xc7 && ((adev->pdev->subsystem_device == 0xb37 && adev->pdev->subsystem_vendor == 0x1002) || @@ -740,24 +740,24 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, cz_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(cz_mgcg_cgcg_init)); + ARRAY_SIZE(cz_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, cz_golden_settings_a11, - (const u32)ARRAY_SIZE(cz_golden_settings_a11)); + ARRAY_SIZE(cz_golden_settings_a11)); amdgpu_program_register_sequence(adev, cz_golden_common_all, - (const u32)ARRAY_SIZE(cz_golden_common_all)); + ARRAY_SIZE(cz_golden_common_all)); break; case CHIP_STONEY: amdgpu_program_register_sequence(adev, stoney_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(stoney_mgcg_cgcg_init)); + ARRAY_SIZE(stoney_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, stoney_golden_settings_a11, - (const u32)ARRAY_SIZE(stoney_golden_settings_a11)); + ARRAY_SIZE(stoney_golden_settings_a11)); amdgpu_program_register_sequence(adev, stoney_golden_common_all, - (const u32)ARRAY_SIZE(stoney_golden_common_all)); + ARRAY_SIZE(stoney_golden_common_all)); break; default: break; @@ -804,7 +804,7 @@ static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring) DRM_UDELAY(1); } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (scratch(0x%04X)=0x%08X)\n", @@ -856,7 +856,7 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", @@ -2114,7 +2114,6 @@ static int gfx_v8_0_sw_fini(void *handle) amdgpu_gfx_compute_mqd_sw_fini(adev); amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring, &adev->gfx.kiq.irq); amdgpu_gfx_kiq_fini(adev); - amdgpu_bo_free_kernel(&adev->virt.csa_obj, &adev->virt.csa_vmid0_addr, NULL); gfx_v8_0_mec_fini(adev); gfx_v8_0_rlc_fini(adev); @@ -3851,6 +3850,14 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) break; udelay(1); } + if (k == adev->usec_timeout) { + gfx_v8_0_select_se_sh(adev, 0xffffffff, + 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); + DRM_INFO("Timeout wait for RLC serdes %u,%u\n", + i, j); + return; + } } } gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); @@ -4305,37 +4312,8 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); amdgpu_ring_write(ring, mmPA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START); - switch (adev->asic_type) { - case CHIP_TONGA: - case CHIP_POLARIS10: - amdgpu_ring_write(ring, 0x16000012); - amdgpu_ring_write(ring, 0x0000002A); - break; - case CHIP_POLARIS11: - case CHIP_POLARIS12: - amdgpu_ring_write(ring, 0x16000012); - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_FIJI: - amdgpu_ring_write(ring, 0x3a00161a); - amdgpu_ring_write(ring, 0x0000002e); - break; - case CHIP_CARRIZO: - amdgpu_ring_write(ring, 0x00000002); - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_TOPAZ: - amdgpu_ring_write(ring, adev->gfx.config.num_rbs == 1 ? - 0x00000000 : 0x00000002); - amdgpu_ring_write(ring, 0x00000000); - break; - case CHIP_STONEY: - amdgpu_ring_write(ring, 0x00000000); - amdgpu_ring_write(ring, 0x00000000); - break; - default: - BUG(); - } + amdgpu_ring_write(ring, adev->gfx.config.rb_config[0][0].raster_config); + amdgpu_ring_write(ring, adev->gfx.config.rb_config[0][0].raster_config_1); amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); amdgpu_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE); @@ -4816,7 +4794,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring) gfx_v8_0_kiq_setting(ring); - if (adev->in_sriov_reset) { /* for GPU_RESET case */ + if (adev->in_gpu_reset) { /* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation)); @@ -4853,7 +4831,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring) struct vi_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!adev->in_sriov_reset && !adev->gfx.in_suspend) { + if (!adev->in_gpu_reset && !adev->gfx.in_suspend) { memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation)); ((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; ((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; @@ -4865,13 +4843,10 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring) if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation)); - } else if (adev->in_sriov_reset) { /* for GPU_RESET case */ + } else if (adev->in_gpu_reset) { /* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation)); - /* reset ring buffer */ - ring->wptr = 0; - amdgpu_ring_clear_ring(ring); } else { amdgpu_ring_clear_ring(ring); } @@ -4946,6 +4921,13 @@ static int gfx_v8_0_kiq_resume(struct amdgpu_device *adev) /* Test KCQs */ for (i = 0; i < adev->gfx.num_compute_rings; i++) { ring = &adev->gfx.compute_ring[i]; + if (adev->in_gpu_reset) { + /* move reset ring buffer to here to workaround + * compute ring test failed + */ + ring->wptr = 0; + amdgpu_ring_clear_ring(ring); + } ring->ready = true; r = amdgpu_ring_test_ring(ring); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index da43813d67a4..6c5289ae67be 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -28,11 +28,11 @@ #include "soc15.h" #include "soc15d.h" -#include "vega10/soc15ip.h" -#include "vega10/GC/gc_9_0_offset.h" -#include "vega10/GC/gc_9_0_sh_mask.h" -#include "vega10/vega10_enum.h" -#include "vega10/HDP/hdp_4_0_offset.h" +#include "soc15ip.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" +#include "vega10_enum.h" +#include "hdp/hdp_4_0_offset.h" #include "soc15_common.h" #include "clearstate_gfx9.h" @@ -232,18 +232,18 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_VEGA10: amdgpu_program_register_sequence(adev, golden_settings_gc_9_0, - (const u32)ARRAY_SIZE(golden_settings_gc_9_0)); + ARRAY_SIZE(golden_settings_gc_9_0)); amdgpu_program_register_sequence(adev, golden_settings_gc_9_0_vg10, - (const u32)ARRAY_SIZE(golden_settings_gc_9_0_vg10)); + ARRAY_SIZE(golden_settings_gc_9_0_vg10)); break; case CHIP_RAVEN: amdgpu_program_register_sequence(adev, golden_settings_gc_9_1, - (const u32)ARRAY_SIZE(golden_settings_gc_9_1)); + ARRAY_SIZE(golden_settings_gc_9_1)); amdgpu_program_register_sequence(adev, golden_settings_gc_9_1_rv1, - (const u32)ARRAY_SIZE(golden_settings_gc_9_1_rv1)); + ARRAY_SIZE(golden_settings_gc_9_1_rv1)); break; default: break; @@ -327,7 +327,7 @@ static int gfx_v9_0_ring_test_ring(struct amdgpu_ring *ring) DRM_UDELAY(1); } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (scratch(0x%04X)=0x%08X)\n", @@ -379,7 +379,7 @@ static int gfx_v9_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", @@ -1464,7 +1464,6 @@ static int gfx_v9_0_sw_fini(void *handle) amdgpu_gfx_compute_mqd_sw_fini(adev); amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring, &adev->gfx.kiq.irq); amdgpu_gfx_kiq_fini(adev); - amdgpu_bo_free_kernel(&adev->virt.csa_obj, &adev->virt.csa_vmid0_addr, NULL); gfx_v9_0_mec_fini(adev); gfx_v9_0_ngg_fini(adev); @@ -1645,6 +1644,14 @@ static void gfx_v9_0_wait_for_rlc_serdes(struct amdgpu_device *adev) break; udelay(1); } + if (k == adev->usec_timeout) { + gfx_v9_0_select_se_sh(adev, 0xffffffff, + 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); + DRM_INFO("Timeout wait for RLC serdes %u,%u\n", + i, j); + return; + } } } gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); @@ -2749,7 +2756,7 @@ static int gfx_v9_0_kiq_init_queue(struct amdgpu_ring *ring) gfx_v9_0_kiq_setting(ring); - if (adev->in_sriov_reset) { /* for GPU_RESET case */ + if (adev->in_gpu_reset) { /* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation)); @@ -2787,7 +2794,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring) struct v9_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!adev->in_sriov_reset && !adev->gfx.in_suspend) { + if (!adev->in_gpu_reset && !adev->gfx.in_suspend) { memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation)); ((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF; ((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF; @@ -2799,7 +2806,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring) if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct v9_mqd_allocation)); - } else if (adev->in_sriov_reset) { /* for GPU_RESET case */ + } else if (adev->in_gpu_reset) { /* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct v9_mqd_allocation)); diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c index c17996e18086..f1effadfbaa6 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c @@ -23,11 +23,11 @@ #include "amdgpu.h" #include "gfxhub_v1_0.h" -#include "vega10/soc15ip.h" -#include "vega10/GC/gc_9_0_offset.h" -#include "vega10/GC/gc_9_0_sh_mask.h" -#include "vega10/GC/gc_9_0_default.h" -#include "vega10/vega10_enum.h" +#include "soc15ip.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" +#include "gc/gc_9_0_default.h" +#include "vega10_enum.h" #include "soc15_common.h" diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c index f4603a7c8ef3..468281f10e8d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c @@ -222,11 +222,6 @@ static void gmc_v6_0_vram_gtt_location(struct amdgpu_device *adev, u64 base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF; base <<= 24; - if (mc->mc_vram_size > 0xFFC0000000ULL) { - dev_warn(adev->dev, "limiting VRAM\n"); - mc->real_vram_size = 0xFFC0000000ULL; - mc->mc_vram_size = 0xFFC0000000ULL; - } amdgpu_vram_location(adev, &adev->mc, base); amdgpu_gart_location(adev, mc); } @@ -283,6 +278,7 @@ static int gmc_v6_0_mc_init(struct amdgpu_device *adev) u32 tmp; int chansize, numchan; + int r; tmp = RREG32(mmMC_ARB_RAMCFG); if (tmp & (1 << 11)) { @@ -324,12 +320,17 @@ static int gmc_v6_0_mc_init(struct amdgpu_device *adev) break; } adev->mc.vram_width = numchan * chansize; - /* Could aper size report 0 ? */ - adev->mc.aper_base = pci_resource_start(adev->pdev, 0); - adev->mc.aper_size = pci_resource_len(adev->pdev, 0); /* size in MB on si */ adev->mc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; adev->mc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; + + if (!(adev->flags & AMD_IS_APU)) { + r = amdgpu_device_resize_fb_bar(adev); + if (r) + return r; + } + adev->mc.aper_base = pci_resource_start(adev->pdev, 0); + adev->mc.aper_size = pci_resource_len(adev->pdev, 0); adev->mc.visible_vram_size = adev->mc.aper_size; /* set the gart size */ @@ -831,8 +832,7 @@ static int gmc_v6_0_sw_init(void *handle) if (r) return r; - amdgpu_vm_adjust_size(adev, 64, 9); - adev->vm_manager.max_pfn = adev->vm_manager.vm_size << 18; + amdgpu_vm_adjust_size(adev, 64, 9, 1, 40); adev->mc.mc_mask = 0xffffffffffULL; @@ -877,7 +877,6 @@ static int gmc_v6_0_sw_init(void *handle) * amdkfd will use VMIDs 8-15 */ adev->vm_manager.id_mgr[0].num_ids = AMDGPU_NUM_OF_VMIDS; - adev->vm_manager.num_level = 1; amdgpu_vm_manager_init(adev); /* base offset of vram pages */ @@ -897,9 +896,9 @@ static int gmc_v6_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_gem_force_release(adev); amdgpu_vm_manager_fini(adev); gmc_v6_0_gart_fini(adev); - amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); release_firmware(adev->mc.fw); adev->mc.fw = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index b0528ca9207b..68a85051f4b7 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -69,10 +69,10 @@ static void gmc_v7_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_TOPAZ: amdgpu_program_register_sequence(adev, iceland_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(iceland_mgcg_cgcg_init)); + ARRAY_SIZE(iceland_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_iceland_a11, - (const u32)ARRAY_SIZE(golden_settings_iceland_a11)); + ARRAY_SIZE(golden_settings_iceland_a11)); break; default: break; @@ -240,12 +240,6 @@ static void gmc_v7_0_vram_gtt_location(struct amdgpu_device *adev, u64 base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF; base <<= 24; - if (mc->mc_vram_size > 0xFFC0000000ULL) { - /* leave room for at least 1024M GTT */ - dev_warn(adev->dev, "limiting VRAM\n"); - mc->real_vram_size = 0xFFC0000000ULL; - mc->mc_vram_size = 0xFFC0000000ULL; - } amdgpu_vram_location(adev, &adev->mc, base); amdgpu_gart_location(adev, mc); } @@ -322,6 +316,8 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev) */ static int gmc_v7_0_mc_init(struct amdgpu_device *adev) { + int r; + adev->mc.vram_width = amdgpu_atombios_get_vram_width(adev); if (!adev->mc.vram_width) { u32 tmp; @@ -367,13 +363,18 @@ static int gmc_v7_0_mc_init(struct amdgpu_device *adev) } adev->mc.vram_width = numchan * chansize; } - /* Could aper size report 0 ? */ - adev->mc.aper_base = pci_resource_start(adev->pdev, 0); - adev->mc.aper_size = pci_resource_len(adev->pdev, 0); /* size in MB on si */ adev->mc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; adev->mc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; + if (!(adev->flags & AMD_IS_APU)) { + r = amdgpu_device_resize_fb_bar(adev); + if (r) + return r; + } + adev->mc.aper_base = pci_resource_start(adev->pdev, 0); + adev->mc.aper_size = pci_resource_len(adev->pdev, 0); + #ifdef CONFIG_X86_64 if (adev->flags & AMD_IS_APU) { adev->mc.aper_base = ((u64)RREG32(mmMC_VM_FB_OFFSET)) << 22; @@ -970,8 +971,7 @@ static int gmc_v7_0_sw_init(void *handle) * Currently set to 4GB ((1 << 20) 4k pages). * Max GPUVM size for cayman and SI is 40 bits. */ - amdgpu_vm_adjust_size(adev, 64, 9); - adev->vm_manager.max_pfn = adev->vm_manager.vm_size << 18; + amdgpu_vm_adjust_size(adev, 64, 9, 1, 40); /* Set the internal MC address mask * This is the max address of the GPU's @@ -1026,7 +1026,6 @@ static int gmc_v7_0_sw_init(void *handle) * amdkfd will use VMIDs 8-15 */ adev->vm_manager.id_mgr[0].num_ids = AMDGPU_NUM_OF_VMIDS; - adev->vm_manager.num_level = 1; amdgpu_vm_manager_init(adev); /* base offset of vram pages */ @@ -1046,9 +1045,9 @@ static int gmc_v7_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_gem_force_release(adev); amdgpu_vm_manager_fini(adev); gmc_v7_0_gart_fini(adev); - amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); release_firmware(adev->mc.fw); adev->mc.fw = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index f368cfe2f585..46ec97e70e5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -122,42 +122,42 @@ static void gmc_v8_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_FIJI: amdgpu_program_register_sequence(adev, fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(fiji_mgcg_cgcg_init)); + ARRAY_SIZE(fiji_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_fiji_a10, - (const u32)ARRAY_SIZE(golden_settings_fiji_a10)); + ARRAY_SIZE(golden_settings_fiji_a10)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tonga_mgcg_cgcg_init)); + ARRAY_SIZE(tonga_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_tonga_a11, - (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); + ARRAY_SIZE(golden_settings_tonga_a11)); break; case CHIP_POLARIS11: case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); + ARRAY_SIZE(golden_settings_polaris11_a11)); break; case CHIP_POLARIS10: amdgpu_program_register_sequence(adev, golden_settings_polaris10_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris10_a11)); + ARRAY_SIZE(golden_settings_polaris10_a11)); break; case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, cz_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(cz_mgcg_cgcg_init)); + ARRAY_SIZE(cz_mgcg_cgcg_init)); break; case CHIP_STONEY: amdgpu_program_register_sequence(adev, stoney_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(stoney_mgcg_cgcg_init)); + ARRAY_SIZE(stoney_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_stoney_common, - (const u32)ARRAY_SIZE(golden_settings_stoney_common)); + ARRAY_SIZE(golden_settings_stoney_common)); break; default: break; @@ -405,12 +405,6 @@ static void gmc_v8_0_vram_gtt_location(struct amdgpu_device *adev, base = RREG32(mmMC_VM_FB_LOCATION) & 0xFFFF; base <<= 24; - if (mc->mc_vram_size > 0xFFC0000000ULL) { - /* leave room for at least 1024M GTT */ - dev_warn(adev->dev, "limiting VRAM\n"); - mc->real_vram_size = 0xFFC0000000ULL; - mc->mc_vram_size = 0xFFC0000000ULL; - } amdgpu_vram_location(adev, &adev->mc, base); amdgpu_gart_location(adev, mc); } @@ -498,6 +492,8 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev) */ static int gmc_v8_0_mc_init(struct amdgpu_device *adev) { + int r; + adev->mc.vram_width = amdgpu_atombios_get_vram_width(adev); if (!adev->mc.vram_width) { u32 tmp; @@ -543,13 +539,18 @@ static int gmc_v8_0_mc_init(struct amdgpu_device *adev) } adev->mc.vram_width = numchan * chansize; } - /* Could aper size report 0 ? */ - adev->mc.aper_base = pci_resource_start(adev->pdev, 0); - adev->mc.aper_size = pci_resource_len(adev->pdev, 0); /* size in MB on si */ adev->mc.mc_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; adev->mc.real_vram_size = RREG32(mmCONFIG_MEMSIZE) * 1024ULL * 1024ULL; + if (!(adev->flags & AMD_IS_APU)) { + r = amdgpu_device_resize_fb_bar(adev); + if (r) + return r; + } + adev->mc.aper_base = pci_resource_start(adev->pdev, 0); + adev->mc.aper_size = pci_resource_len(adev->pdev, 0); + #ifdef CONFIG_X86_64 if (adev->flags & AMD_IS_APU) { adev->mc.aper_base = ((u64)RREG32(mmMC_VM_FB_OFFSET)) << 22; @@ -1067,8 +1068,7 @@ static int gmc_v8_0_sw_init(void *handle) * Currently set to 4GB ((1 << 20) 4k pages). * Max GPUVM size for cayman and SI is 40 bits. */ - amdgpu_vm_adjust_size(adev, 64, 9); - adev->vm_manager.max_pfn = adev->vm_manager.vm_size << 18; + amdgpu_vm_adjust_size(adev, 64, 9, 1, 40); /* Set the internal MC address mask * This is the max address of the GPU's @@ -1123,7 +1123,6 @@ static int gmc_v8_0_sw_init(void *handle) * amdkfd will use VMIDs 8-15 */ adev->vm_manager.id_mgr[0].num_ids = AMDGPU_NUM_OF_VMIDS; - adev->vm_manager.num_level = 1; amdgpu_vm_manager_init(adev); /* base offset of vram pages */ @@ -1143,9 +1142,9 @@ static int gmc_v8_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_gem_force_release(adev); amdgpu_vm_manager_fini(adev); gmc_v8_0_gart_fini(adev); - amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); release_firmware(adev->mc.fw); adev->mc.fw = NULL; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index c8f1aebeac7a..cc972153d401 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -25,17 +25,18 @@ #include "gmc_v9_0.h" #include "amdgpu_atomfirmware.h" -#include "vega10/soc15ip.h" -#include "vega10/HDP/hdp_4_0_offset.h" -#include "vega10/HDP/hdp_4_0_sh_mask.h" -#include "vega10/GC/gc_9_0_sh_mask.h" -#include "vega10/DC/dce_12_0_offset.h" -#include "vega10/DC/dce_12_0_sh_mask.h" -#include "vega10/vega10_enum.h" -#include "vega10/MMHUB/mmhub_1_0_offset.h" -#include "vega10/ATHUB/athub_1_0_offset.h" +#include "soc15ip.h" +#include "hdp/hdp_4_0_offset.h" +#include "hdp/hdp_4_0_sh_mask.h" +#include "gc/gc_9_0_sh_mask.h" +#include "dce/dce_12_0_offset.h" +#include "dce/dce_12_0_sh_mask.h" +#include "vega10_enum.h" +#include "mmhub/mmhub_1_0_offset.h" +#include "athub/athub_1_0_offset.h" #include "soc15_common.h" +#include "umc/umc_6_0_sh_mask.h" #include "nbio_v6_1.h" #include "nbio_v7_0.h" @@ -85,6 +86,121 @@ static const u32 golden_settings_athub_1_0_0[] = SOC15_REG_OFFSET(ATHUB, 0, mmRPB_ARB_CNTL2), 0x00ff00ff, 0x00080008 }; +/* Ecc related register addresses, (BASE + reg offset) */ +/* Universal Memory Controller caps (may be fused). */ +/* UMCCH:UmcLocalCap */ +#define UMCLOCALCAPS_ADDR0 (0x00014306 + 0x00000000) +#define UMCLOCALCAPS_ADDR1 (0x00014306 + 0x00000800) +#define UMCLOCALCAPS_ADDR2 (0x00014306 + 0x00001000) +#define UMCLOCALCAPS_ADDR3 (0x00014306 + 0x00001800) +#define UMCLOCALCAPS_ADDR4 (0x00054306 + 0x00000000) +#define UMCLOCALCAPS_ADDR5 (0x00054306 + 0x00000800) +#define UMCLOCALCAPS_ADDR6 (0x00054306 + 0x00001000) +#define UMCLOCALCAPS_ADDR7 (0x00054306 + 0x00001800) +#define UMCLOCALCAPS_ADDR8 (0x00094306 + 0x00000000) +#define UMCLOCALCAPS_ADDR9 (0x00094306 + 0x00000800) +#define UMCLOCALCAPS_ADDR10 (0x00094306 + 0x00001000) +#define UMCLOCALCAPS_ADDR11 (0x00094306 + 0x00001800) +#define UMCLOCALCAPS_ADDR12 (0x000d4306 + 0x00000000) +#define UMCLOCALCAPS_ADDR13 (0x000d4306 + 0x00000800) +#define UMCLOCALCAPS_ADDR14 (0x000d4306 + 0x00001000) +#define UMCLOCALCAPS_ADDR15 (0x000d4306 + 0x00001800) + +/* Universal Memory Controller Channel config. */ +/* UMCCH:UMC_CONFIG */ +#define UMCCH_UMC_CONFIG_ADDR0 (0x00014040 + 0x00000000) +#define UMCCH_UMC_CONFIG_ADDR1 (0x00014040 + 0x00000800) +#define UMCCH_UMC_CONFIG_ADDR2 (0x00014040 + 0x00001000) +#define UMCCH_UMC_CONFIG_ADDR3 (0x00014040 + 0x00001800) +#define UMCCH_UMC_CONFIG_ADDR4 (0x00054040 + 0x00000000) +#define UMCCH_UMC_CONFIG_ADDR5 (0x00054040 + 0x00000800) +#define UMCCH_UMC_CONFIG_ADDR6 (0x00054040 + 0x00001000) +#define UMCCH_UMC_CONFIG_ADDR7 (0x00054040 + 0x00001800) +#define UMCCH_UMC_CONFIG_ADDR8 (0x00094040 + 0x00000000) +#define UMCCH_UMC_CONFIG_ADDR9 (0x00094040 + 0x00000800) +#define UMCCH_UMC_CONFIG_ADDR10 (0x00094040 + 0x00001000) +#define UMCCH_UMC_CONFIG_ADDR11 (0x00094040 + 0x00001800) +#define UMCCH_UMC_CONFIG_ADDR12 (0x000d4040 + 0x00000000) +#define UMCCH_UMC_CONFIG_ADDR13 (0x000d4040 + 0x00000800) +#define UMCCH_UMC_CONFIG_ADDR14 (0x000d4040 + 0x00001000) +#define UMCCH_UMC_CONFIG_ADDR15 (0x000d4040 + 0x00001800) + +/* Universal Memory Controller Channel Ecc config. */ +/* UMCCH:EccCtrl */ +#define UMCCH_ECCCTRL_ADDR0 (0x00014053 + 0x00000000) +#define UMCCH_ECCCTRL_ADDR1 (0x00014053 + 0x00000800) +#define UMCCH_ECCCTRL_ADDR2 (0x00014053 + 0x00001000) +#define UMCCH_ECCCTRL_ADDR3 (0x00014053 + 0x00001800) +#define UMCCH_ECCCTRL_ADDR4 (0x00054053 + 0x00000000) +#define UMCCH_ECCCTRL_ADDR5 (0x00054053 + 0x00000800) +#define UMCCH_ECCCTRL_ADDR6 (0x00054053 + 0x00001000) +#define UMCCH_ECCCTRL_ADDR7 (0x00054053 + 0x00001800) +#define UMCCH_ECCCTRL_ADDR8 (0x00094053 + 0x00000000) +#define UMCCH_ECCCTRL_ADDR9 (0x00094053 + 0x00000800) +#define UMCCH_ECCCTRL_ADDR10 (0x00094053 + 0x00001000) +#define UMCCH_ECCCTRL_ADDR11 (0x00094053 + 0x00001800) +#define UMCCH_ECCCTRL_ADDR12 (0x000d4053 + 0x00000000) +#define UMCCH_ECCCTRL_ADDR13 (0x000d4053 + 0x00000800) +#define UMCCH_ECCCTRL_ADDR14 (0x000d4053 + 0x00001000) +#define UMCCH_ECCCTRL_ADDR15 (0x000d4053 + 0x00001800) + +static const uint32_t ecc_umclocalcap_addrs[] = { + UMCLOCALCAPS_ADDR0, + UMCLOCALCAPS_ADDR1, + UMCLOCALCAPS_ADDR2, + UMCLOCALCAPS_ADDR3, + UMCLOCALCAPS_ADDR4, + UMCLOCALCAPS_ADDR5, + UMCLOCALCAPS_ADDR6, + UMCLOCALCAPS_ADDR7, + UMCLOCALCAPS_ADDR8, + UMCLOCALCAPS_ADDR9, + UMCLOCALCAPS_ADDR10, + UMCLOCALCAPS_ADDR11, + UMCLOCALCAPS_ADDR12, + UMCLOCALCAPS_ADDR13, + UMCLOCALCAPS_ADDR14, + UMCLOCALCAPS_ADDR15, +}; + +static const uint32_t ecc_umcch_umc_config_addrs[] = { + UMCCH_UMC_CONFIG_ADDR0, + UMCCH_UMC_CONFIG_ADDR1, + UMCCH_UMC_CONFIG_ADDR2, + UMCCH_UMC_CONFIG_ADDR3, + UMCCH_UMC_CONFIG_ADDR4, + UMCCH_UMC_CONFIG_ADDR5, + UMCCH_UMC_CONFIG_ADDR6, + UMCCH_UMC_CONFIG_ADDR7, + UMCCH_UMC_CONFIG_ADDR8, + UMCCH_UMC_CONFIG_ADDR9, + UMCCH_UMC_CONFIG_ADDR10, + UMCCH_UMC_CONFIG_ADDR11, + UMCCH_UMC_CONFIG_ADDR12, + UMCCH_UMC_CONFIG_ADDR13, + UMCCH_UMC_CONFIG_ADDR14, + UMCCH_UMC_CONFIG_ADDR15, +}; + +static const uint32_t ecc_umcch_eccctrl_addrs[] = { + UMCCH_ECCCTRL_ADDR0, + UMCCH_ECCCTRL_ADDR1, + UMCCH_ECCCTRL_ADDR2, + UMCCH_ECCCTRL_ADDR3, + UMCCH_ECCCTRL_ADDR4, + UMCCH_ECCCTRL_ADDR5, + UMCCH_ECCCTRL_ADDR6, + UMCCH_ECCCTRL_ADDR7, + UMCCH_ECCCTRL_ADDR8, + UMCCH_ECCCTRL_ADDR9, + UMCCH_ECCCTRL_ADDR10, + UMCCH_ECCCTRL_ADDR11, + UMCCH_ECCCTRL_ADDR12, + UMCCH_ECCCTRL_ADDR13, + UMCCH_ECCCTRL_ADDR14, + UMCCH_ECCCTRL_ADDR15, +}; + static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type, @@ -389,6 +505,85 @@ static int gmc_v9_0_early_init(void *handle) return 0; } +static int gmc_v9_0_ecc_available(struct amdgpu_device *adev) +{ + uint32_t reg_val; + uint32_t reg_addr; + uint32_t field_val; + size_t i; + uint32_t fv2; + size_t lost_sheep; + + DRM_DEBUG("ecc: gmc_v9_0_ecc_available()\n"); + + lost_sheep = 0; + for (i = 0; i < ARRAY_SIZE(ecc_umclocalcap_addrs); ++i) { + reg_addr = ecc_umclocalcap_addrs[i]; + DRM_DEBUG("ecc: " + "UMCCH_UmcLocalCap[%zu]: reg_addr: 0x%08x\n", + i, reg_addr); + reg_val = RREG32(reg_addr); + field_val = REG_GET_FIELD(reg_val, UMCCH0_0_UmcLocalCap, + EccDis); + DRM_DEBUG("ecc: " + "reg_val: 0x%08x, " + "EccDis: 0x%08x, ", + reg_val, field_val); + if (field_val) { + DRM_ERROR("ecc: UmcLocalCap:EccDis is set.\n"); + ++lost_sheep; + } + } + + for (i = 0; i < ARRAY_SIZE(ecc_umcch_umc_config_addrs); ++i) { + reg_addr = ecc_umcch_umc_config_addrs[i]; + DRM_DEBUG("ecc: " + "UMCCH0_0_UMC_CONFIG[%zu]: reg_addr: 0x%08x", + i, reg_addr); + reg_val = RREG32(reg_addr); + field_val = REG_GET_FIELD(reg_val, UMCCH0_0_UMC_CONFIG, + DramReady); + DRM_DEBUG("ecc: " + "reg_val: 0x%08x, " + "DramReady: 0x%08x\n", + reg_val, field_val); + + if (!field_val) { + DRM_ERROR("ecc: UMC_CONFIG:DramReady is not set.\n"); + ++lost_sheep; + } + } + + for (i = 0; i < ARRAY_SIZE(ecc_umcch_eccctrl_addrs); ++i) { + reg_addr = ecc_umcch_eccctrl_addrs[i]; + DRM_DEBUG("ecc: " + "UMCCH_EccCtrl[%zu]: reg_addr: 0x%08x, ", + i, reg_addr); + reg_val = RREG32(reg_addr); + field_val = REG_GET_FIELD(reg_val, UMCCH0_0_EccCtrl, + WrEccEn); + fv2 = REG_GET_FIELD(reg_val, UMCCH0_0_EccCtrl, + RdEccEn); + DRM_DEBUG("ecc: " + "reg_val: 0x%08x, " + "WrEccEn: 0x%08x, " + "RdEccEn: 0x%08x\n", + reg_val, field_val, fv2); + + if (!field_val) { + DRM_DEBUG("ecc: WrEccEn is not set\n"); + ++lost_sheep; + } + if (!fv2) { + DRM_DEBUG("ecc: RdEccEn is not set\n"); + ++lost_sheep; + } + } + + DRM_DEBUG("ecc: lost_sheep: %zu\n", lost_sheep); + return lost_sheep == 0; +} + static int gmc_v9_0_late_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -403,6 +598,7 @@ static int gmc_v9_0_late_init(void *handle) */ unsigned vm_inv_eng[AMDGPU_MAX_VMHUBS] = { 4, 4 }; unsigned i; + int r; for(i = 0; i < adev->num_rings; ++i) { struct amdgpu_ring *ring = adev->rings[i]; @@ -418,6 +614,16 @@ static int gmc_v9_0_late_init(void *handle) for(i = 0; i < AMDGPU_MAX_VMHUBS; ++i) BUG_ON(vm_inv_eng[i] > 16); + r = gmc_v9_0_ecc_available(adev); + if (r == 1) { + DRM_INFO("ECC is active.\n"); + } else if (r == 0) { + DRM_INFO("ECC is not present.\n"); + } else { + DRM_ERROR("gmc_v9_0_ecc_available() failed. r: %d\n", r); + return r; + } + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0); } @@ -449,6 +655,7 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev) { u32 tmp; int chansize, numchan; + int r; adev->mc.vram_width = amdgpu_atomfirmware_get_vram_width(adev); if (!adev->mc.vram_width) { @@ -491,17 +698,22 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev) adev->mc.vram_width = numchan * chansize; } - /* Could aper size report 0 ? */ - adev->mc.aper_base = pci_resource_start(adev->pdev, 0); - adev->mc.aper_size = pci_resource_len(adev->pdev, 0); /* size in MB on si */ adev->mc.mc_vram_size = ((adev->flags & AMD_IS_APU) ? nbio_v7_0_get_memsize(adev) : nbio_v6_1_get_memsize(adev)) * 1024ULL * 1024ULL; adev->mc.real_vram_size = adev->mc.mc_vram_size; - adev->mc.visible_vram_size = adev->mc.aper_size; + + if (!(adev->flags & AMD_IS_APU)) { + r = amdgpu_device_resize_fb_bar(adev); + if (r) + return r; + } + adev->mc.aper_base = pci_resource_start(adev->pdev, 0); + adev->mc.aper_size = pci_resource_len(adev->pdev, 0); /* In case the PCI BAR is larger than the actual amount of vram */ + adev->mc.visible_vram_size = adev->mc.aper_size; if (adev->mc.visible_vram_size > adev->mc.real_vram_size) adev->mc.visible_vram_size = adev->mc.real_vram_size; @@ -557,16 +769,11 @@ static int gmc_v9_0_sw_init(void *handle) switch (adev->asic_type) { case CHIP_RAVEN: adev->mc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN; - if (adev->rev_id == 0x0 || adev->rev_id == 0x1) { - adev->vm_manager.vm_size = 1U << 18; - adev->vm_manager.block_size = 9; - adev->vm_manager.num_level = 3; - amdgpu_vm_set_fragment_size(adev, 9); - } else { + if (adev->rev_id == 0x0 || adev->rev_id == 0x1) + amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); + else /* vm_size is 64GB for legacy 2-level page support */ - amdgpu_vm_adjust_size(adev, 64, 9); - adev->vm_manager.num_level = 1; - } + amdgpu_vm_adjust_size(adev, 64, 9, 1, 48); break; case CHIP_VEGA10: /* XXX Don't know how to get VRAM type yet. */ @@ -576,20 +783,12 @@ static int gmc_v9_0_sw_init(void *handle) * vm size is 256TB (48bit), maximum size of Vega10, * block size 512 (9bit) */ - adev->vm_manager.vm_size = 1U << 18; - adev->vm_manager.block_size = 9; - adev->vm_manager.num_level = 3; - amdgpu_vm_set_fragment_size(adev, 9); + amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); break; default: break; } - DRM_INFO("vm size is %llu GB, block size is %u-bit,fragment size is %u-bit\n", - adev->vm_manager.vm_size, - adev->vm_manager.block_size, - adev->vm_manager.fragment_size); - /* This interrupt is VMC page fault.*/ r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_VMC, 0, &adev->mc.vm_fault); @@ -599,8 +798,6 @@ static int gmc_v9_0_sw_init(void *handle) if (r) return r; - adev->vm_manager.max_pfn = adev->vm_manager.vm_size << 18; - /* Set the internal MC address mask * This is the max address of the GPU's * internal address space. @@ -660,7 +857,7 @@ static int gmc_v9_0_sw_init(void *handle) } /** - * gmc_v8_0_gart_fini - vm fini callback + * gmc_v9_0_gart_fini - vm fini callback * * @adev: amdgpu_device pointer * @@ -676,9 +873,9 @@ static int gmc_v9_0_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_gem_force_release(adev); amdgpu_vm_manager_fini(adev); gmc_v9_0_gart_fini(adev); - amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); return 0; @@ -690,15 +887,15 @@ static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_VEGA10: amdgpu_program_register_sequence(adev, golden_settings_mmhub_1_0_0, - (const u32)ARRAY_SIZE(golden_settings_mmhub_1_0_0)); + ARRAY_SIZE(golden_settings_mmhub_1_0_0)); amdgpu_program_register_sequence(adev, golden_settings_athub_1_0_0, - (const u32)ARRAY_SIZE(golden_settings_athub_1_0_0)); + ARRAY_SIZE(golden_settings_athub_1_0_0)); break; case CHIP_RAVEN: amdgpu_program_register_sequence(adev, golden_settings_athub_1_0_0, - (const u32)ARRAY_SIZE(golden_settings_athub_1_0_0)); + ARRAY_SIZE(golden_settings_athub_1_0_0)); break; default: break; @@ -718,7 +915,7 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev) amdgpu_program_register_sequence(adev, golden_settings_vega10_hdp, - (const u32)ARRAY_SIZE(golden_settings_vega10_hdp)); + ARRAY_SIZE(golden_settings_vega10_hdp)); if (adev->gart.robj == NULL) { dev_err(adev->dev, "No VRAM object for PCIE GART.\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c index cc21c4bdec27..bd160d8700e0 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c @@ -23,14 +23,13 @@ #include "amdgpu.h" #include "mmhub_v1_0.h" -#include "vega10/soc15ip.h" -#include "vega10/MMHUB/mmhub_1_0_offset.h" -#include "vega10/MMHUB/mmhub_1_0_sh_mask.h" -#include "vega10/MMHUB/mmhub_1_0_default.h" -#include "vega10/ATHUB/athub_1_0_offset.h" -#include "vega10/ATHUB/athub_1_0_sh_mask.h" -#include "vega10/ATHUB/athub_1_0_default.h" -#include "vega10/vega10_enum.h" +#include "soc15ip.h" +#include "mmhub/mmhub_1_0_offset.h" +#include "mmhub/mmhub_1_0_sh_mask.h" +#include "mmhub/mmhub_1_0_default.h" +#include "athub/athub_1_0_offset.h" +#include "athub/athub_1_0_sh_mask.h" +#include "vega10_enum.h" #include "soc15_common.h" diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index b4906d2f30d3..ad9054e3903c 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -22,11 +22,11 @@ */ #include "amdgpu.h" -#include "vega10/soc15ip.h" -#include "vega10/NBIO/nbio_6_1_offset.h" -#include "vega10/NBIO/nbio_6_1_sh_mask.h" -#include "vega10/GC/gc_9_0_offset.h" -#include "vega10/GC/gc_9_0_sh_mask.h" +#include "soc15ip.h" +#include "nbio/nbio_6_1_offset.h" +#include "nbio/nbio_6_1_sh_mask.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" #include "soc15.h" #include "vega10_ih.h" #include "soc15_common.h" @@ -254,7 +254,7 @@ static void xgpu_ai_mailbox_flr_work(struct work_struct *work) } /* Trigger recovery due to world switch failure */ - amdgpu_sriov_gpu_reset(adev, NULL); + amdgpu_gpu_recover(adev, NULL); } static int xgpu_ai_set_mailbox_rcv_irq(struct amdgpu_device *adev, @@ -282,9 +282,17 @@ static int xgpu_ai_mailbox_rcv_irq(struct amdgpu_device *adev, /* see what event we get */ r = xgpu_ai_mailbox_rcv_msg(adev, IDH_FLR_NOTIFICATION); - /* only handle FLR_NOTIFY now */ - if (!r) - schedule_work(&adev->virt.flr_work); + /* sometimes the interrupt is delayed to inject to VM, so under such case + * the IDH_FLR_NOTIFICATION is overwritten by VF FLR from GIM side, thus + * above recieve message could be failed, we should schedule the flr_work + * anyway + */ + if (r) { + DRM_ERROR("FLR_NOTIFICATION is missed\n"); + xgpu_ai_mailbox_send_ack(adev); + } + + schedule_work(&adev->virt.flr_work); } return 0; @@ -353,5 +361,6 @@ const struct amdgpu_virt_ops xgpu_ai_virt_ops = { .req_full_gpu = xgpu_ai_request_full_gpu_access, .rel_full_gpu = xgpu_ai_release_full_gpu_access, .reset_gpu = xgpu_ai_request_reset, + .wait_reset = NULL, .trans_msg = xgpu_ai_mailbox_trans_msg, }; diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c index c25a831f94ec..df52824c0cd4 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c @@ -281,29 +281,29 @@ void xgpu_vi_init_golden_registers(struct amdgpu_device *adev) case CHIP_FIJI: amdgpu_program_register_sequence(adev, xgpu_fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_fiji_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, xgpu_fiji_golden_settings_a10, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_fiji_golden_settings_a10)); amdgpu_program_register_sequence(adev, xgpu_fiji_golden_common_all, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_fiji_golden_common_all)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, xgpu_tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_tonga_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, xgpu_tonga_golden_settings_a11, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_tonga_golden_settings_a11)); amdgpu_program_register_sequence(adev, xgpu_tonga_golden_common_all, - (const u32)ARRAY_SIZE( + ARRAY_SIZE( xgpu_tonga_golden_common_all)); break; default: @@ -446,8 +446,10 @@ static int xgpu_vi_send_access_requests(struct amdgpu_device *adev, request == IDH_REQ_GPU_FINI_ACCESS || request == IDH_REQ_GPU_RESET_ACCESS) { r = xgpu_vi_poll_msg(adev, IDH_READY_TO_ACCESS_GPU); - if (r) - pr_err("Doesn't get ack from pf, continue\n"); + if (r) { + pr_err("Doesn't get ack from pf, give up\n"); + return r; + } } return 0; @@ -458,6 +460,11 @@ static int xgpu_vi_request_reset(struct amdgpu_device *adev) return xgpu_vi_send_access_requests(adev, IDH_REQ_GPU_RESET_ACCESS); } +static int xgpu_vi_wait_reset_cmpl(struct amdgpu_device *adev) +{ + return xgpu_vi_poll_msg(adev, IDH_FLR_NOTIFICATION_CMPL); +} + static int xgpu_vi_request_full_gpu_access(struct amdgpu_device *adev, bool init) { @@ -514,7 +521,7 @@ static void xgpu_vi_mailbox_flr_work(struct work_struct *work) } /* Trigger recovery due to world switch failure */ - amdgpu_sriov_gpu_reset(adev, NULL); + amdgpu_gpu_recover(adev, NULL); } static int xgpu_vi_set_mailbox_rcv_irq(struct amdgpu_device *adev, @@ -613,5 +620,6 @@ const struct amdgpu_virt_ops xgpu_vi_virt_ops = { .req_full_gpu = xgpu_vi_request_full_gpu_access, .rel_full_gpu = xgpu_vi_release_full_gpu_access, .reset_gpu = xgpu_vi_request_reset, + .wait_reset = xgpu_vi_wait_reset_cmpl, .trans_msg = NULL, /* Does not need to trans VF errors to host. */ }; diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c index 904a1bab9b9f..76db711097c7 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c @@ -24,11 +24,11 @@ #include "amdgpu_atombios.h" #include "nbio_v6_1.h" -#include "vega10/soc15ip.h" -#include "vega10/NBIO/nbio_6_1_default.h" -#include "vega10/NBIO/nbio_6_1_offset.h" -#include "vega10/NBIO/nbio_6_1_sh_mask.h" -#include "vega10/vega10_enum.h" +#include "soc15ip.h" +#include "nbio/nbio_6_1_default.h" +#include "nbio/nbio_6_1_offset.h" +#include "nbio/nbio_6_1_sh_mask.h" +#include "vega10_enum.h" #define smnCPM_CONTROL 0x11180460 #define smnPCIE_CNTL2 0x11180070 diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c index f802b973410a..1fb77174e02c 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c @@ -24,11 +24,11 @@ #include "amdgpu_atombios.h" #include "nbio_v7_0.h" -#include "vega10/soc15ip.h" -#include "raven1/NBIO/nbio_7_0_default.h" -#include "raven1/NBIO/nbio_7_0_offset.h" -#include "raven1/NBIO/nbio_7_0_sh_mask.h" -#include "vega10/vega10_enum.h" +#include "soc15ip.h" +#include "nbio/nbio_7_0_default.h" +#include "nbio/nbio_7_0_offset.h" +#include "nbio/nbio_7_0_sh_mask.h" +#include "vega10_enum.h" #define smnNBIF_MGCG_CTRL_LCLK 0x1013a05c diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c index 4e20d91d5d50..78fe3f2917a0 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c @@ -30,10 +30,10 @@ #include "soc15_common.h" #include "psp_v10_0.h" -#include "vega10/soc15ip.h" -#include "raven1/MP/mp_10_0_offset.h" -#include "raven1/GC/gc_9_1_offset.h" -#include "raven1/SDMA0/sdma0_4_1_offset.h" +#include "soc15ip.h" +#include "mp/mp_10_0_offset.h" +#include "gc/gc_9_1_offset.h" +#include "sdma0/sdma0_4_1_offset.h" MODULE_FIRMWARE("amdgpu/raven_asd.bin"); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c index c7bcfe8e286c..e75a23d858ef 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c @@ -31,12 +31,12 @@ #include "soc15_common.h" #include "psp_v3_1.h" -#include "vega10/soc15ip.h" -#include "vega10/MP/mp_9_0_offset.h" -#include "vega10/MP/mp_9_0_sh_mask.h" -#include "vega10/GC/gc_9_0_offset.h" -#include "vega10/SDMA0/sdma0_4_0_offset.h" -#include "vega10/NBIO/nbio_6_1_offset.h" +#include "soc15ip.h" +#include "mp/mp_9_0_offset.h" +#include "mp/mp_9_0_sh_mask.h" +#include "gc/gc_9_0_offset.h" +#include "sdma0/sdma0_4_0_offset.h" +#include "nbio/nbio_6_1_offset.h" MODULE_FIRMWARE("amdgpu/vega10_sos.bin"); MODULE_FIRMWARE("amdgpu/vega10_asd.bin"); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index 67f375bfe452..121e628e7cdb 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -95,10 +95,10 @@ static void sdma_v2_4_init_golden_registers(struct amdgpu_device *adev) case CHIP_TOPAZ: amdgpu_program_register_sequence(adev, iceland_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(iceland_mgcg_cgcg_init)); + ARRAY_SIZE(iceland_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_iceland_a11, - (const u32)ARRAY_SIZE(golden_settings_iceland_a11)); + ARRAY_SIZE(golden_settings_iceland_a11)); break; default: break; @@ -633,7 +633,7 @@ static int sdma_v2_4_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", ring->idx, tmp); @@ -704,7 +704,7 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 6d06f8eb659f..c8c93f9dac21 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -194,45 +194,45 @@ static void sdma_v3_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_FIJI: amdgpu_program_register_sequence(adev, fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(fiji_mgcg_cgcg_init)); + ARRAY_SIZE(fiji_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_fiji_a10, - (const u32)ARRAY_SIZE(golden_settings_fiji_a10)); + ARRAY_SIZE(golden_settings_fiji_a10)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tonga_mgcg_cgcg_init)); + ARRAY_SIZE(tonga_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, golden_settings_tonga_a11, - (const u32)ARRAY_SIZE(golden_settings_tonga_a11)); + ARRAY_SIZE(golden_settings_tonga_a11)); break; case CHIP_POLARIS11: case CHIP_POLARIS12: amdgpu_program_register_sequence(adev, golden_settings_polaris11_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris11_a11)); + ARRAY_SIZE(golden_settings_polaris11_a11)); break; case CHIP_POLARIS10: amdgpu_program_register_sequence(adev, golden_settings_polaris10_a11, - (const u32)ARRAY_SIZE(golden_settings_polaris10_a11)); + ARRAY_SIZE(golden_settings_polaris10_a11)); break; case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, cz_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(cz_mgcg_cgcg_init)); + ARRAY_SIZE(cz_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, cz_golden_settings_a11, - (const u32)ARRAY_SIZE(cz_golden_settings_a11)); + ARRAY_SIZE(cz_golden_settings_a11)); break; case CHIP_STONEY: amdgpu_program_register_sequence(adev, stoney_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(stoney_mgcg_cgcg_init)); + ARRAY_SIZE(stoney_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, stoney_golden_settings_a11, - (const u32)ARRAY_SIZE(stoney_golden_settings_a11)); + ARRAY_SIZE(stoney_golden_settings_a11)); break; default: break; @@ -893,7 +893,7 @@ static int sdma_v3_0_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", ring->idx, tmp); @@ -964,7 +964,7 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 46009db3d195..4c55f21e37a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -27,15 +27,15 @@ #include "amdgpu_ucode.h" #include "amdgpu_trace.h" -#include "vega10/soc15ip.h" -#include "vega10/SDMA0/sdma0_4_0_offset.h" -#include "vega10/SDMA0/sdma0_4_0_sh_mask.h" -#include "vega10/SDMA1/sdma1_4_0_offset.h" -#include "vega10/SDMA1/sdma1_4_0_sh_mask.h" -#include "vega10/MMHUB/mmhub_1_0_offset.h" -#include "vega10/MMHUB/mmhub_1_0_sh_mask.h" -#include "vega10/HDP/hdp_4_0_offset.h" -#include "raven1/SDMA0/sdma0_4_1_default.h" +#include "soc15ip.h" +#include "sdma0/sdma0_4_0_offset.h" +#include "sdma0/sdma0_4_0_sh_mask.h" +#include "sdma1/sdma1_4_0_offset.h" +#include "sdma1/sdma1_4_0_sh_mask.h" +#include "mmhub/mmhub_1_0_offset.h" +#include "mmhub/mmhub_1_0_sh_mask.h" +#include "hdp/hdp_4_0_offset.h" +#include "sdma0/sdma0_4_1_default.h" #include "soc15_common.h" #include "soc15.h" @@ -132,18 +132,18 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev) case CHIP_VEGA10: amdgpu_program_register_sequence(adev, golden_settings_sdma_4, - (const u32)ARRAY_SIZE(golden_settings_sdma_4)); + ARRAY_SIZE(golden_settings_sdma_4)); amdgpu_program_register_sequence(adev, golden_settings_sdma_vg10, - (const u32)ARRAY_SIZE(golden_settings_sdma_vg10)); + ARRAY_SIZE(golden_settings_sdma_vg10)); break; case CHIP_RAVEN: amdgpu_program_register_sequence(adev, golden_settings_sdma_4_1, - (const u32)ARRAY_SIZE(golden_settings_sdma_4_1)); + ARRAY_SIZE(golden_settings_sdma_4_1)); amdgpu_program_register_sequence(adev, golden_settings_sdma_rv1, - (const u32)ARRAY_SIZE(golden_settings_sdma_rv1)); + ARRAY_SIZE(golden_settings_sdma_rv1)); break; default: break; @@ -919,7 +919,7 @@ static int sdma_v4_0_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", ring->idx, tmp); @@ -990,7 +990,7 @@ static int sdma_v4_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c index 8284d5dbfc30..49eef3090f08 100644 --- a/drivers/gpu/drm/amd/amdgpu/si.c +++ b/drivers/gpu/drm/amd/amdgpu/si.c @@ -1392,63 +1392,63 @@ static void si_init_golden_registers(struct amdgpu_device *adev) case CHIP_TAHITI: amdgpu_program_register_sequence(adev, tahiti_golden_registers, - (const u32)ARRAY_SIZE(tahiti_golden_registers)); + ARRAY_SIZE(tahiti_golden_registers)); amdgpu_program_register_sequence(adev, tahiti_golden_rlc_registers, - (const u32)ARRAY_SIZE(tahiti_golden_rlc_registers)); + ARRAY_SIZE(tahiti_golden_rlc_registers)); amdgpu_program_register_sequence(adev, tahiti_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tahiti_mgcg_cgcg_init)); + ARRAY_SIZE(tahiti_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, tahiti_golden_registers2, - (const u32)ARRAY_SIZE(tahiti_golden_registers2)); + ARRAY_SIZE(tahiti_golden_registers2)); break; case CHIP_PITCAIRN: amdgpu_program_register_sequence(adev, pitcairn_golden_registers, - (const u32)ARRAY_SIZE(pitcairn_golden_registers)); + ARRAY_SIZE(pitcairn_golden_registers)); amdgpu_program_register_sequence(adev, pitcairn_golden_rlc_registers, - (const u32)ARRAY_SIZE(pitcairn_golden_rlc_registers)); + ARRAY_SIZE(pitcairn_golden_rlc_registers)); amdgpu_program_register_sequence(adev, pitcairn_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(pitcairn_mgcg_cgcg_init)); + ARRAY_SIZE(pitcairn_mgcg_cgcg_init)); break; case CHIP_VERDE: amdgpu_program_register_sequence(adev, verde_golden_registers, - (const u32)ARRAY_SIZE(verde_golden_registers)); + ARRAY_SIZE(verde_golden_registers)); amdgpu_program_register_sequence(adev, verde_golden_rlc_registers, - (const u32)ARRAY_SIZE(verde_golden_rlc_registers)); + ARRAY_SIZE(verde_golden_rlc_registers)); amdgpu_program_register_sequence(adev, verde_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(verde_mgcg_cgcg_init)); + ARRAY_SIZE(verde_mgcg_cgcg_init)); amdgpu_program_register_sequence(adev, verde_pg_init, - (const u32)ARRAY_SIZE(verde_pg_init)); + ARRAY_SIZE(verde_pg_init)); break; case CHIP_OLAND: amdgpu_program_register_sequence(adev, oland_golden_registers, - (const u32)ARRAY_SIZE(oland_golden_registers)); + ARRAY_SIZE(oland_golden_registers)); amdgpu_program_register_sequence(adev, oland_golden_rlc_registers, - (const u32)ARRAY_SIZE(oland_golden_rlc_registers)); + ARRAY_SIZE(oland_golden_rlc_registers)); amdgpu_program_register_sequence(adev, oland_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(oland_mgcg_cgcg_init)); + ARRAY_SIZE(oland_mgcg_cgcg_init)); break; case CHIP_HAINAN: amdgpu_program_register_sequence(adev, hainan_golden_registers, - (const u32)ARRAY_SIZE(hainan_golden_registers)); + ARRAY_SIZE(hainan_golden_registers)); amdgpu_program_register_sequence(adev, hainan_golden_registers2, - (const u32)ARRAY_SIZE(hainan_golden_registers2)); + ARRAY_SIZE(hainan_golden_registers2)); amdgpu_program_register_sequence(adev, hainan_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(hainan_mgcg_cgcg_init)); + ARRAY_SIZE(hainan_mgcg_cgcg_init)); break; diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index 3fa2fbf8c9a1..ee469a906cd3 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -252,7 +252,7 @@ static int si_dma_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i); + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", ring->idx, tmp); @@ -317,7 +317,7 @@ static int si_dma_ring_test_ib(struct amdgpu_ring *ring, long timeout) } tmp = le32_to_cpu(adev->wb.wb[index]); if (tmp == 0xDEADBEEF) { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); diff --git a/drivers/gpu/drm/amd/amdgpu/si_dpm.c b/drivers/gpu/drm/amd/amdgpu/si_dpm.c index 51fd0c9a20a5..299cb3161b2c 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dpm.c @@ -5845,9 +5845,9 @@ static int si_set_mc_special_registers(struct amdgpu_device *adev, ((temp_reg & 0xffff0000)) | ((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16); j++; + if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) return -EINVAL; - temp_reg = RREG32(MC_PMG_CMD_MRS); table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS; table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP; @@ -5859,18 +5859,16 @@ static int si_set_mc_special_registers(struct amdgpu_device *adev, table->mc_reg_table_entry[k].mc_data[j] |= 0x100; } j++; - if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; if (adev->mc.vram_type != AMDGPU_VRAM_TYPE_GDDR5) { + if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) + return -EINVAL; table->mc_reg_address[j].s1 = MC_PMG_AUTO_CMD; table->mc_reg_address[j].s0 = MC_PMG_AUTO_CMD; for (k = 0; k < table->num_entries; k++) table->mc_reg_table_entry[k].mc_data[j] = (table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16; j++; - if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; } break; case MC_SEQ_RESERVE_M: @@ -5882,8 +5880,6 @@ static int si_set_mc_special_registers(struct amdgpu_device *adev, (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); j++; - if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 4e67fe1e7955..f134ca0c093c 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -34,18 +34,18 @@ #include "atom.h" #include "amd_pcie.h" -#include "vega10/soc15ip.h" -#include "vega10/UVD/uvd_7_0_offset.h" -#include "vega10/GC/gc_9_0_offset.h" -#include "vega10/GC/gc_9_0_sh_mask.h" -#include "vega10/SDMA0/sdma0_4_0_offset.h" -#include "vega10/SDMA1/sdma1_4_0_offset.h" -#include "vega10/HDP/hdp_4_0_offset.h" -#include "vega10/HDP/hdp_4_0_sh_mask.h" -#include "vega10/MP/mp_9_0_offset.h" -#include "vega10/MP/mp_9_0_sh_mask.h" -#include "vega10/SMUIO/smuio_9_0_offset.h" -#include "vega10/SMUIO/smuio_9_0_sh_mask.h" +#include "soc15ip.h" +#include "uvd/uvd_7_0_offset.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" +#include "sdma0/sdma0_4_0_offset.h" +#include "sdma1/sdma1_4_0_offset.h" +#include "hdp/hdp_4_0_offset.h" +#include "hdp/hdp_4_0_sh_mask.h" +#include "mp/mp_9_0_offset.h" +#include "mp/mp_9_0_sh_mask.h" +#include "smuio/smuio_9_0_offset.h" +#include "smuio/smuio_9_0_sh_mask.h" #include "soc15.h" #include "soc15_common.h" @@ -265,12 +265,12 @@ static void soc15_init_golden_registers(struct amdgpu_device *adev) case CHIP_VEGA10: amdgpu_program_register_sequence(adev, vega10_golden_init, - (const u32)ARRAY_SIZE(vega10_golden_init)); + ARRAY_SIZE(vega10_golden_init)); break; case CHIP_RAVEN: amdgpu_program_register_sequence(adev, raven_golden_init, - (const u32)ARRAY_SIZE(raven_golden_init)); + ARRAY_SIZE(raven_golden_init)); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 8ab0f78794a5..b13ae34be1c2 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -521,7 +521,7 @@ static int uvd_v4_2_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", @@ -563,7 +563,7 @@ static void uvd_v4_2_mc_resume(struct amdgpu_device *adev) /* programm the VCPU memory controller bits 0-27 */ addr = (adev->uvd.gpu_addr + AMDGPU_UVD_FIRMWARE_OFFSET) >> 3; - size = AMDGPU_GPU_PAGE_ALIGN(adev->uvd.fw->size + 4) >> 3; + size = AMDGPU_UVD_FIRMWARE_SIZE(adev) >> 3; WREG32(mmUVD_VCPU_CACHE_OFFSET0, addr); WREG32(mmUVD_VCPU_CACHE_SIZE0, size); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index bb6d46e168a3..a4b0f1d842b7 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -258,7 +258,7 @@ static void uvd_v5_0_mc_resume(struct amdgpu_device *adev) upper_32_bits(adev->uvd.gpu_addr)); offset = AMDGPU_UVD_FIRMWARE_OFFSET; - size = AMDGPU_GPU_PAGE_ALIGN(adev->uvd.fw->size + 4); + size = AMDGPU_UVD_FIRMWARE_SIZE(adev); WREG32(mmUVD_VCPU_CACHE_OFFSET0, offset >> 3); WREG32(mmUVD_VCPU_CACHE_SIZE0, size); @@ -536,7 +536,7 @@ static int uvd_v5_0_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index 920910ac8663..0e8b887cf03e 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -184,7 +184,7 @@ static int uvd_v6_0_enc_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed\n", @@ -360,7 +360,7 @@ static int uvd_v6_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } error: @@ -416,7 +416,7 @@ static int uvd_v6_0_sw_init(void *handle) ring = &adev->uvd.ring_enc[0]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity_enc, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r) { DRM_ERROR("Failed setting up UVD ENC run queue.\n"); return r; @@ -603,7 +603,7 @@ static void uvd_v6_0_mc_resume(struct amdgpu_device *adev) upper_32_bits(adev->uvd.gpu_addr)); offset = AMDGPU_UVD_FIRMWARE_OFFSET; - size = AMDGPU_GPU_PAGE_ALIGN(adev->uvd.fw->size + 4); + size = AMDGPU_UVD_FIRMWARE_SIZE(adev); WREG32(mmUVD_VCPU_CACHE_OFFSET0, offset >> 3); WREG32(mmUVD_VCPU_CACHE_SIZE0, size); @@ -1008,7 +1008,7 @@ static int uvd_v6_0_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index 6634545060fd..660fa41dc877 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -29,16 +29,16 @@ #include "soc15_common.h" #include "mmsch_v1_0.h" -#include "vega10/soc15ip.h" -#include "vega10/UVD/uvd_7_0_offset.h" -#include "vega10/UVD/uvd_7_0_sh_mask.h" -#include "vega10/VCE/vce_4_0_offset.h" -#include "vega10/VCE/vce_4_0_default.h" -#include "vega10/VCE/vce_4_0_sh_mask.h" -#include "vega10/NBIF/nbif_6_1_offset.h" -#include "vega10/HDP/hdp_4_0_offset.h" -#include "vega10/MMHUB/mmhub_1_0_offset.h" -#include "vega10/MMHUB/mmhub_1_0_sh_mask.h" +#include "soc15ip.h" +#include "uvd/uvd_7_0_offset.h" +#include "uvd/uvd_7_0_sh_mask.h" +#include "vce/vce_4_0_offset.h" +#include "vce/vce_4_0_default.h" +#include "vce/vce_4_0_sh_mask.h" +#include "nbif/nbif_6_1_offset.h" +#include "hdp/hdp_4_0_offset.h" +#include "mmhub/mmhub_1_0_offset.h" +#include "mmhub/mmhub_1_0_sh_mask.h" static void uvd_v7_0_set_ring_funcs(struct amdgpu_device *adev); static void uvd_v7_0_set_enc_ring_funcs(struct amdgpu_device *adev); @@ -184,7 +184,7 @@ static int uvd_v7_0_enc_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed\n", @@ -359,7 +359,7 @@ static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout) } else if (r < 0) { DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_DEBUG("ib test on ring %d succeeded\n", ring->idx); r = 0; } error: @@ -418,7 +418,7 @@ static int uvd_v7_0_sw_init(void *handle) ring = &adev->uvd.ring_enc[0]; rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity_enc, - rq, amdgpu_sched_jobs); + rq, amdgpu_sched_jobs, NULL); if (r) { DRM_ERROR("Failed setting up UVD ENC run queue.\n"); return r; @@ -616,7 +616,7 @@ static int uvd_v7_0_resume(void *handle) */ static void uvd_v7_0_mc_resume(struct amdgpu_device *adev) { - uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->uvd.fw->size + 4); + uint32_t size = AMDGPU_UVD_FIRMWARE_SIZE(adev); uint32_t offset; if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { @@ -1192,7 +1192,7 @@ static int uvd_v7_0_ring_test_ring(struct amdgpu_ring *ring) } if (i < adev->usec_timeout) { - DRM_INFO("ring test on %d succeeded in %d usecs\n", + DRM_DEBUG("ring test on %d succeeded in %d usecs\n", ring->idx, i); } else { DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n", diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c index 75745544600a..f2f713650074 100644..100755 --- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c @@ -32,12 +32,12 @@ #include "soc15_common.h" #include "mmsch_v1_0.h" -#include "vega10/soc15ip.h" -#include "vega10/VCE/vce_4_0_offset.h" -#include "vega10/VCE/vce_4_0_default.h" -#include "vega10/VCE/vce_4_0_sh_mask.h" -#include "vega10/MMHUB/mmhub_1_0_offset.h" -#include "vega10/MMHUB/mmhub_1_0_sh_mask.h" +#include "soc15ip.h" +#include "vce/vce_4_0_offset.h" +#include "vce/vce_4_0_default.h" +#include "vce/vce_4_0_sh_mask.h" +#include "mmhub/mmhub_1_0_offset.h" +#include "mmhub/mmhub_1_0_sh_mask.h" #define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK 0x02 @@ -243,37 +243,49 @@ static int vce_v4_0_sriov_start(struct amdgpu_device *adev) MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VM_CTRL), 0); if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR0), - adev->firmware.ucode[AMDGPU_UCODE_ID_VCE].mc_addr >> 8); - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR1), - adev->firmware.ucode[AMDGPU_UCODE_ID_VCE].mc_addr >> 8); - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR2), + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_40BIT_BAR0), adev->firmware.ucode[AMDGPU_UCODE_ID_VCE].mc_addr >> 8); + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_64BIT_BAR0), + (adev->firmware.ucode[AMDGPU_UCODE_ID_VCE].mc_addr >> 40) & 0xff); } else { - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR0), + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_40BIT_BAR0), adev->vce.gpu_addr >> 8); - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR1), + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_64BIT_BAR0), + (adev->vce.gpu_addr >> 40) & 0xff); + } + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_40BIT_BAR1), adev->vce.gpu_addr >> 8); - MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_VCPU_CACHE_40BIT_BAR2), + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_64BIT_BAR1), + (adev->vce.gpu_addr >> 40) & 0xff); + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_40BIT_BAR2), adev->vce.gpu_addr >> 8); - } + MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, + mmVCE_LMI_VCPU_CACHE_64BIT_BAR2), + (adev->vce.gpu_addr >> 40) & 0xff); offset = AMDGPU_VCE_FIRMWARE_OFFSET; size = VCE_V4_0_FW_SIZE; MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_OFFSET0), - offset & 0x7FFFFFFF); + offset & ~0x0f000000); MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_SIZE0), size); - offset += size; + offset = (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) ? offset + size : 0; size = VCE_V4_0_STACK_SIZE; MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_OFFSET1), - offset & 0x7FFFFFFF); + (offset & ~0x0f000000) | (1 << 24)); MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_SIZE1), size); offset += size; size = VCE_V4_0_DATA_SIZE; MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_OFFSET2), - offset & 0x7FFFFFFF); + (offset & ~0x0f000000) | (2 << 24)); MMSCH_V1_0_INSERT_DIRECT_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_VCPU_CACHE_SIZE2), size); MMSCH_V1_0_INSERT_DIRECT_RD_MOD_WT(SOC15_REG_OFFSET(VCE, 0, mmVCE_LMI_CTRL2), ~0x100, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 0450ac5ba6b6..e4673f792545 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -28,12 +28,12 @@ #include "soc15d.h" #include "soc15_common.h" -#include "vega10/soc15ip.h" -#include "raven1/VCN/vcn_1_0_offset.h" -#include "raven1/VCN/vcn_1_0_sh_mask.h" -#include "vega10/HDP/hdp_4_0_offset.h" -#include "raven1/MMHUB/mmhub_9_1_offset.h" -#include "raven1/MMHUB/mmhub_9_1_sh_mask.h" +#include "soc15ip.h" +#include "vcn/vcn_1_0_offset.h" +#include "vcn/vcn_1_0_sh_mask.h" +#include "hdp/hdp_4_0_offset.h" +#include "mmhub/mmhub_9_1_offset.h" +#include "mmhub/mmhub_9_1_sh_mask.h" static int vcn_v1_0_start(struct amdgpu_device *adev); static int vcn_v1_0_stop(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index 697325737ba8..ca778cd4e6e8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -26,9 +26,9 @@ #include "soc15.h" -#include "vega10/soc15ip.h" -#include "vega10/OSSSYS/osssys_4_0_offset.h" -#include "vega10/OSSSYS/osssys_4_0_sh_mask.h" +#include "soc15ip.h" +#include "oss/osssys_4_0_offset.h" +#include "oss/osssys_4_0_sh_mask.h" #include "soc15_common.h" #include "vega10_ih.h" @@ -46,11 +46,11 @@ static void vega10_ih_set_interrupt_funcs(struct amdgpu_device *adev); */ static void vega10_ih_enable_interrupts(struct amdgpu_device *adev) { - u32 ih_rb_cntl = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL)); + u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 1); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 1); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL), ih_rb_cntl); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl); adev->irq.ih.enabled = true; } @@ -63,14 +63,14 @@ static void vega10_ih_enable_interrupts(struct amdgpu_device *adev) */ static void vega10_ih_disable_interrupts(struct amdgpu_device *adev) { - u32 ih_rb_cntl = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL)); + u32 ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_ENABLE, 0); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 0); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL), ih_rb_cntl); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl); /* set rptr, wptr to 0 */ - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR), 0); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR), 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0); adev->irq.ih.enabled = false; adev->irq.ih.rptr = 0; } @@ -102,15 +102,15 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev) else nbio_v6_1_ih_control(adev); - ih_rb_cntl = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL)); + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL); /* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/ if (adev->irq.ih.use_bus_addr) { - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE), adev->irq.ih.rb_dma_addr >> 8); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI), ((u64)adev->irq.ih.rb_dma_addr >> 40) & 0xff); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, adev->irq.ih.rb_dma_addr >> 8); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, ((u64)adev->irq.ih.rb_dma_addr >> 40) & 0xff); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SPACE, 1); } else { - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE), adev->irq.ih.gpu_addr >> 8); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_BASE_HI), (adev->irq.ih.gpu_addr >> 40) & 0xff); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, adev->irq.ih.gpu_addr >> 8); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, (adev->irq.ih.gpu_addr >> 40) & 0xff); ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SPACE, 4); } rb_bufsz = order_base_2(adev->irq.ih.ring_size / 4); @@ -126,21 +126,21 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev) if (adev->irq.msi_enabled) ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RPTR_REARM, 1); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL), ih_rb_cntl); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl); /* set the writeback address whether it's enabled or not */ if (adev->irq.ih.use_bus_addr) wptr_off = adev->irq.ih.rb_dma_addr + (adev->irq.ih.wptr_offs * 4); else wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO), lower_32_bits(wptr_off)); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI), upper_32_bits(wptr_off) & 0xFF); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); /* set rptr, wptr to 0 */ - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR), 0); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR), 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0); - ih_doorbell_rtpr = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR)); + ih_doorbell_rtpr = RREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR); if (adev->irq.ih.use_doorbell) { ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, IH_DOORBELL_RPTR, OFFSET, adev->irq.ih.doorbell_index); @@ -150,20 +150,20 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev) ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, IH_DOORBELL_RPTR, ENABLE, 0); } - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_DOORBELL_RPTR), ih_doorbell_rtpr); + WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR, ih_doorbell_rtpr); if (adev->flags & AMD_IS_APU) nbio_v7_0_ih_doorbell_range(adev, adev->irq.ih.use_doorbell, adev->irq.ih.doorbell_index); else nbio_v6_1_ih_doorbell_range(adev, adev->irq.ih.use_doorbell, adev->irq.ih.doorbell_index); - tmp = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL)); + tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL); tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL, CLIENT18_IS_STORM_CLIENT, 1); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL), tmp); + WREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL, tmp); - tmp = RREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_INT_FLOOD_CNTL)); + tmp = RREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL); tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1); - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_INT_FLOOD_CNTL), tmp); + WREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL, tmp); pci_set_master(adev->pdev); @@ -367,7 +367,7 @@ static void vega10_ih_set_rptr(struct amdgpu_device *adev) adev->wb.wb[adev->irq.ih.rptr_offs] = adev->irq.ih.rptr; WDOORBELL32(adev->irq.ih.doorbell_index, adev->irq.ih.rptr); } else { - WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_RPTR), adev->irq.ih.rptr); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, adev->irq.ih.rptr); } } diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 3a4c2fa7e36d..bb8ca9489546 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -284,27 +284,27 @@ static void vi_init_golden_registers(struct amdgpu_device *adev) case CHIP_TOPAZ: amdgpu_program_register_sequence(adev, iceland_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(iceland_mgcg_cgcg_init)); + ARRAY_SIZE(iceland_mgcg_cgcg_init)); break; case CHIP_FIJI: amdgpu_program_register_sequence(adev, fiji_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(fiji_mgcg_cgcg_init)); + ARRAY_SIZE(fiji_mgcg_cgcg_init)); break; case CHIP_TONGA: amdgpu_program_register_sequence(adev, tonga_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(tonga_mgcg_cgcg_init)); + ARRAY_SIZE(tonga_mgcg_cgcg_init)); break; case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, cz_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(cz_mgcg_cgcg_init)); + ARRAY_SIZE(cz_mgcg_cgcg_init)); break; case CHIP_STONEY: amdgpu_program_register_sequence(adev, stoney_mgcg_cgcg_init, - (const u32)ARRAY_SIZE(stoney_mgcg_cgcg_init)); + ARRAY_SIZE(stoney_mgcg_cgcg_init)); break; case CHIP_POLARIS11: case CHIP_POLARIS10: |