]> asedeno.scripts.mit.edu Git - linux.git/commitdiff
drm/amdgpu: add support for BO_VAs without BO v2
authorChristian König <christian.koenig@amd.com>
Mon, 30 Jan 2017 10:01:38 +0000 (11:01 +0100)
committerAlex Deucher <alexander.deucher@amd.com>
Thu, 30 Mar 2017 03:52:55 +0000 (23:52 -0400)
For PRT support we need mappings which aren't backed by any memory.

v2: fix parameter checking

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c

index bd0d33125c18a74c354c7db2c475a7cfa230b460..f922e4356758e47b6a5b7a7b9a9365fb0f2e5700 100644 (file)
@@ -1117,7 +1117,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
        struct dma_fence *exclusive;
        int r;
 
-       if (clear) {
+       if (clear || !bo_va->bo) {
                mem = NULL;
                nodes = NULL;
                exclusive = NULL;
@@ -1134,9 +1134,15 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
                exclusive = reservation_object_get_excl(bo_va->bo->tbo.resv);
        }
 
-       flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem);
-       gtt_flags = (amdgpu_ttm_is_bound(bo_va->bo->tbo.ttm) &&
-               adev == amdgpu_ttm_adev(bo_va->bo->tbo.bdev)) ? flags : 0;
+       if (bo_va->bo) {
+               flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem);
+               gtt_flags = (amdgpu_ttm_is_bound(bo_va->bo->tbo.ttm) &&
+                       adev == amdgpu_ttm_adev(bo_va->bo->tbo.bdev)) ?
+                       flags : 0;
+       } else {
+               flags = 0x0;
+               gtt_flags = ~0x0;
+       }
 
        spin_lock(&vm->status_lock);
        if (!list_empty(&bo_va->vm_status))
@@ -1271,7 +1277,8 @@ struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
        INIT_LIST_HEAD(&bo_va->invalids);
        INIT_LIST_HEAD(&bo_va->vm_status);
 
-       list_add_tail(&bo_va->bo_list, &bo->va);
+       if (bo)
+               list_add_tail(&bo_va->bo_list, &bo->va);
 
        return bo_va;
 }
@@ -1309,7 +1316,8 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 
        /* make sure object fit at this offset */
        eaddr = saddr + size - 1;
-       if ((saddr >= eaddr) || (offset + size > amdgpu_bo_size(bo_va->bo)))
+       if (saddr >= eaddr ||
+           (bo_va->bo && offset + size > amdgpu_bo_size(bo_va->bo)))
                return -EINVAL;
 
        last_pfn = eaddr / AMDGPU_GPU_PAGE_SIZE;