Am 2023-04-04 um 05:56 schrieb Shane Xiao:
For DMA-MAP userptr on other GPUs, the dma address array has been populated in amdgpu_ttm_backend_bind.
OK. I think "has been populated" should be "will be populated", because amdgpu_ttm_backend_bind happens as a callback from the ttm_bo_validate call below. With that fixed in the patch description, the patch is
Reviewed-by: Felix Kuehling <felix.kuehl...@amd.com>
Remove the redundant call from the driver. Signed-off-by: Shane Xiao <shane.x...@amd.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index bcb0a7b32703..94ee8f638c12 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -488,9 +488,6 @@ kfd_mem_dmamap_userptr(struct kgd_mem *mem, if (unlikely(ret)) goto release_sg;- drm_prime_sg_to_dma_addr_array(ttm->sg, ttm->dma_address,- ttm->num_pages); - amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_GTT); ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); if (ret)
