static int amdgpu_ctx_init(struct amdgpu_device *adev, enum drm_sched_priority priority, struct drm_file *filp, struct amdgpu_ctx *ctx) { unsigned i, j; int r; if (priority < 0 || priority >= DRM_SCHED_PRIORITY_MAX) return -EINVAL; r = amdgpu_ctx_priority_permit(filp, priority); if (r) return r; memset(ctx, 0, sizeof(*ctx)); ctx->adev = adev; kref_init(&ctx->refcount); spin_lock_init(&ctx->ring_lock); ctx->fences = kcalloc(amdgpu_sched_jobs * AMDGPU_MAX_RINGS, sizeof(struct dma_fence*), GFP_KERNEL); if (!ctx->fences) return -ENOMEM; mutex_init(&ctx->lock); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { ctx->rings[i].sequence = 1; ctx->rings[i].fences = &ctx->fences[amdgpu_sched_jobs * i]; } 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 = DRM_SCHED_PRIORITY_UNSET; /* create context entity for each ring */ for (i = 0; i < adev->num_rings; i++) { struct amdgpu_ring *ring = adev->rings[i]; struct drm_sched_rq *rq; rq = &ring->sched.sched_rq[priority]; if (ring == &adev->gfx.kiq.ring) continue; r = drm_sched_entity_init(&ring->sched, &ctx->rings[i].entity, rq, &ctx->guilty); if (r) goto failed; } r = amdgpu_queue_mgr_init(adev, &ctx->queue_mgr); if (r) goto failed; return 0; failed: for (j = 0; j < i; j++) drm_sched_entity_fini(&adev->rings[j]->sched, &ctx->rings[j].entity); kfree(ctx->fences); ctx->fences = NULL; return r; }
static int amdgpu_ctx_init(struct amdgpu_device *adev, enum drm_sched_priority priority, struct drm_file *filp, struct amdgpu_ctx *ctx) { unsigned num_entities = amdgput_ctx_total_num_entities(); unsigned i, j; int r; if (priority < 0 || priority >= DRM_SCHED_PRIORITY_MAX) return -EINVAL; r = amdgpu_ctx_priority_permit(filp, priority); if (r) return r; memset(ctx, 0, sizeof(*ctx)); ctx->adev = adev; ctx->fences = kcalloc(amdgpu_sched_jobs * num_entities, sizeof(struct dma_fence*), GFP_KERNEL); if (!ctx->fences) return -ENOMEM; ctx->entities[0] = kcalloc(num_entities, sizeof(struct amdgpu_ctx_entity), GFP_KERNEL); if (!ctx->entities[0]) { r = -ENOMEM; goto error_free_fences; } for (i = 0; i < num_entities; ++i) { struct amdgpu_ctx_entity *entity = &ctx->entities[0][i]; entity->sequence = 1; entity->fences = &ctx->fences[amdgpu_sched_jobs * i]; } for (i = 1; i < AMDGPU_HW_IP_NUM; ++i) ctx->entities[i] = ctx->entities[i - 1] + amdgpu_ctx_num_entities[i - 1]; kref_init(&ctx->refcount); spin_lock_init(&ctx->ring_lock); mutex_init(&ctx->lock); 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 = DRM_SCHED_PRIORITY_UNSET; for (i = 0; i < AMDGPU_HW_IP_NUM; ++i) { struct amdgpu_ring *rings[AMDGPU_MAX_RINGS]; struct drm_sched_rq *rqs[AMDGPU_MAX_RINGS]; unsigned num_rings; unsigned num_rqs = 0; switch (i) { case AMDGPU_HW_IP_GFX: rings[0] = &adev->gfx.gfx_ring[0]; num_rings = 1; break; case AMDGPU_HW_IP_COMPUTE: for (j = 0; j < adev->gfx.num_compute_rings; ++j) rings[j] = &adev->gfx.compute_ring[j]; num_rings = adev->gfx.num_compute_rings; break; case AMDGPU_HW_IP_DMA: for (j = 0; j < adev->sdma.num_instances; ++j) rings[j] = &adev->sdma.instance[j].ring; num_rings = adev->sdma.num_instances; break; case AMDGPU_HW_IP_UVD: rings[0] = &adev->uvd.inst[0].ring; num_rings = 1; break; case AMDGPU_HW_IP_VCE: rings[0] = &adev->vce.ring[0]; num_rings = 1; break; case AMDGPU_HW_IP_UVD_ENC: rings[0] = &adev->uvd.inst[0].ring_enc[0]; num_rings = 1; break; case AMDGPU_HW_IP_VCN_DEC: rings[0] = &adev->vcn.ring_dec; num_rings = 1; break; case AMDGPU_HW_IP_VCN_ENC: rings[0] = &adev->vcn.ring_enc[0]; num_rings = 1; break; case AMDGPU_HW_IP_VCN_JPEG: rings[0] = &adev->vcn.ring_jpeg; num_rings = 1; break; } for (j = 0; j < num_rings; ++j) { if (!rings[j]->adev) continue; rqs[num_rqs++] = &rings[j]->sched.sched_rq[priority]; } for (j = 0; j < amdgpu_ctx_num_entities[i]; ++j) r = drm_sched_entity_init(&ctx->entities[i][j].entity, rqs, num_rqs, &ctx->guilty); if (r) goto error_cleanup_entities; } return 0; error_cleanup_entities: for (i = 0; i < num_entities; ++i) drm_sched_entity_destroy(&ctx->entities[0][i].entity); kfree(ctx->entities[0]); error_free_fences: kfree(ctx->fences); ctx->fences = NULL; return r; }