On 10/28/25 14:09, Pierre-Eric Pelloux-Prayer wrote: > BO's reservation object must be locked before using > amdgpu_vm_bo_update_shared otherwise dma_resv_assert_held will > complain in amdgpu_vm_update_shared. > > Signed-off-by: Pierre-Eric Pelloux-Prayer <[email protected]>
Reviewed-by: Christian König <[email protected]> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c > index ff98c87b2e0b..830412f04b6c 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c > @@ -81,13 +81,20 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf, > struct drm_gem_object *obj = dmabuf->priv; > struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); > struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); > + int r; > > if (!amdgpu_dmabuf_is_xgmi_accessible(attach_adev, bo) && > pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0) > attach->peer2peer = false; > > + r = dma_resv_lock(bo->tbo.base.resv, NULL); > + if (r) > + return r; > + > amdgpu_vm_bo_update_shared(bo); > > + dma_resv_unlock(bo->tbo.base.resv); > + > return 0; > } >
