On 3/18/2025 9:19 AM, Ahmad Rehman wrote:
> For SRIOV, skip the SDMA queue reset and return
> error. The engine/queue reset failure will trigger
> FLR in the sequence.
>
> Signed-off-by: Ahmad Rehman <ahmad.reh...@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
> b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
> index fd34dc138081..5b2d4ca92c17 100644
> --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
> +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
> @@ -1666,6 +1666,11 @@ static int sdma_v4_4_2_reset_queue(struct amdgpu_ring
> *ring, unsigned int vmid)
> {
> struct amdgpu_device *adev = ring->adev;
> u32 id = GET_INST(SDMA0, ring->me);
> +
> + /* Return error for SRIOV will result in FLR */
> + if (amdgpu_sriov_vf(adev))
> + return -EINVAL;
Prefer something like below -
if (!(adev->sdma.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE))
return -EOPNOTSUPP;
Then keep the supported_reset mask updated for VF in
sdma_v4_4_2_update_reset_mask() => if (amdgpu_sriov_vf(adev)) return; //
do nothing.
Thanks,
Lijo
> +
> return amdgpu_sdma_reset_engine(adev, id, true);
> }
>
