1 files changed, 41 insertions, 54 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index d0d0ea565e3d..d8f1335bc68f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -55,8 +55,8 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p,
 	bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj));
 	p->uf_entry.priority = 0;
 	p->uf_entry.tv.bo = &bo->tbo;
-	/* One for TTM and one for the CS job */
-	p->uf_entry.tv.num_shared = 2;
+	/* One for TTM and two for the CS job */
+	p->uf_entry.tv.num_shared = 3;
 
 	drm_gem_object_put(gobj);
 
@@ -116,7 +116,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 	int ret;
 
 	if (cs->in.num_chunks == 0)
-		return 0;
+		return -EINVAL;
 
 	chunk_array = kvmalloc_array(cs->in.num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (!chunk_array)
@@ -519,6 +519,8 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 			return r;
 	}
 
+	mutex_lock(&p->bo_list->bo_list_mutex);
+
 	/* One for TTM and one for the CS job */
 	amdgpu_bo_list_for_each_entry(e, p->bo_list)
 		e->tv.num_shared = 2;
@@ -545,14 +547,15 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 					GFP_KERNEL | __GFP_ZERO);
 		if (!e->user_pages) {
 			DRM_ERROR("kvmalloc_array failure\n");
-			return -ENOMEM;
+			r = -ENOMEM;
+			goto out_free_user_pages;
 		}
 
 		r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages);
 		if (r) {
 			kvfree(e->user_pages);
 			e->user_pages = NULL;
-			return r;
+			goto out_free_user_pages;
 		}
 
 		for (i = 0; i < bo->tbo.ttm->num_pages; i++) {
@@ -569,21 +572,13 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS)
 			DRM_ERROR("ttm_eu_reserve_buffers failed.\n");
-		goto out;
+		goto out_free_user_pages;
 	}
 
 	amdgpu_bo_list_for_each_entry(e, p->bo_list) {
 		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
 
 		e->bo_va = amdgpu_vm_bo_find(vm, bo);
-
-		if (bo->tbo.base.dma_buf && !amdgpu_bo_explicit_sync(bo)) {
-			e->chain = dma_fence_chain_alloc();
-			if (!e->chain) {
-				r = -ENOMEM;
-				goto error_validate;
-			}
-		}
 	}
 
 	/* Move fence waiting after getting reservation lock of
@@ -644,14 +639,22 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	}
 
 error_validate:
+	if (r)
+		ttm_eu_backoff_reservation(&p->ticket, &p->validated);
+
+out_free_user_pages:
 	if (r) {
-		amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-			dma_fence_chain_free(e->chain);
-			e->chain = NULL;
+		amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
+			struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
+
+			if (!e->user_pages)
+				continue;
+			amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+			kvfree(e->user_pages);
+			e->user_pages = NULL;
 		}
-		ttm_eu_backoff_reservation(&p->ticket, &p->validated);
+		mutex_unlock(&p->bo_list->bo_list_mutex);
 	}
-out:
 	return r;
 }
 
@@ -691,15 +694,9 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error,
 	unsigned i;
 
 	if (error && backoff) {
-		struct amdgpu_bo_list_entry *e;
-
-		amdgpu_bo_list_for_each_entry(e, parser->bo_list) {
-			dma_fence_chain_free(e->chain);
-			e->chain = NULL;
-		}
-
 		ttm_eu_backoff_reservation(&parser->ticket,
 					   &parser->validated);
+		mutex_unlock(&parser->bo_list->bo_list_mutex);
 	}
 
 	for (i = 0; i < parser->num_post_deps; i++) {
@@ -809,22 +806,22 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_vm_bo_update(adev, fpriv->prt_va, false, NULL);
+	r = amdgpu_vm_bo_update(adev, fpriv->prt_va, false);
 	if (r)
 		return r;
 
-	r = amdgpu_sync_vm_fence(&p->job->sync, fpriv->prt_va->last_pt_update);
+	r = amdgpu_sync_fence(&p->job->sync, fpriv->prt_va->last_pt_update);
 	if (r)
 		return r;
 
 	if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
 		bo_va = fpriv->csa_va;
 		BUG_ON(!bo_va);
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 		if (r)
 			return r;
 
-		r = amdgpu_sync_vm_fence(&p->job->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->job->sync, bo_va->last_pt_update);
 		if (r)
 			return r;
 	}
@@ -839,13 +836,17 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 		if (bo_va == NULL)
 			continue;
 
-		r = amdgpu_vm_bo_update(adev, bo_va, false, NULL);
-		if (r)
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
+		if (r) {
+			mutex_unlock(&p->bo_list->bo_list_mutex);
 			return r;
+		}
 
-		r = amdgpu_sync_vm_fence(&p->job->sync, bo_va->last_pt_update);
-		if (r)
+		r = amdgpu_sync_fence(&p->job->sync, bo_va->last_pt_update);
+		if (r) {
+			mutex_unlock(&p->bo_list->bo_list_mutex);
 			return r;
+		}
 	}
 
 	r = amdgpu_vm_handle_moved(adev, vm);
@@ -856,7 +857,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_vm_fence(&p->job->sync, vm->last_update);
+	r = amdgpu_sync_fence(&p->job->sync, vm->last_update);
 	if (r)
 		return r;
 
@@ -1260,7 +1261,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 
 	p->fence = dma_fence_get(&job->base.s_fence->finished);
 
-	amdgpu_ctx_add_fence(p->ctx, entity, p->fence, &seq);
+	seq = amdgpu_ctx_add_fence(p->ctx, entity, p->fence);
 	amdgpu_cs_post_dependencies(p);
 
 	if ((job->preamble_status & AMDGPU_PREAMBLE_IB_PRESENT) &&
@@ -1280,27 +1281,13 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 
 	amdgpu_vm_move_to_lru_tail(p->adev, &fpriv->vm);
 
-	amdgpu_bo_list_for_each_entry(e, p->bo_list) {
-		struct dma_resv *resv = e->tv.bo->base.resv;
-		struct dma_fence_chain *chain = e->chain;
-
-		if (!chain)
-			continue;
-
-		/*
-		 * Work around dma_resv shortcomings by wrapping up the
-		 * submission in a dma_fence_chain and add it as exclusive
-		 * fence.
-		 */
-		dma_fence_chain_init(chain, dma_resv_excl_fence(resv),
-				     dma_fence_get(p->fence), 1);
-
-		rcu_assign_pointer(resv->fence_excl, &chain->base);
-		e->chain = NULL;
-	}
+	/* Make sure all BOs are remembered as writers */
+	amdgpu_bo_list_for_each_entry(e, p->bo_list)
+		e->tv.num_shared = 0;
 
 	ttm_eu_fence_buffer_objects(&p->ticket, &p->validated, p->fence);
 	mutex_unlock(&p->adev->notifier_lock);
+	mutex_unlock(&p->bo_list->bo_list_mutex);
 
 	return 0;