On Tue, Aug 30, 2022 at 11:24 AM Yifan Zhang <yifan1.zhang@xxxxxxx> wrote:
>
> there is only one SDMA engine in SDMA 6.0.1, the sdma_hqd_mask has to be
> zeroed for the 2nd engine, otherwise MES scheduler will consider 2nd
> engine exists and map/unmap SDMA queues to the non-existent engine.
>
> Signed-off-by: Yifan Zhang <yifan1.zhang@xxxxxxx>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
> index fe82b8b19a4e..f876fd62ab81 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
> @@ -185,6 +185,14 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
> adev->mes.sdma_hqd_mask[i] = 0xfc;
> }
>
> + /* there is only one SDMA engine in SDMA 6.0.1, the sdma_hqd_mask
> + * has to be zeroed for the 2nd engine, otherwise MES scheduler will
> + * consider 2nd engine exists and map/unmap SDMA queues to the non-existent
> + * engine.
> + */
> + if (adev->ip_versions[SDMA0_HWIP][0] == IP_VERSION(6, 0, 1))
How about checking for the number of instances rather than a specific
IP version?
E.g.,
if (adev->sdma.num_instances == 1)
This makes things more future proof for newer SDMA 6.x IPs.
Alex
> + adev->mes.sdma_hqd_mask[AMDGPU_MES_MAX_SDMA_PIPES - 1] = 0;
> +
> r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs);
> if (r) {
> dev_err(adev->dev,
> --
> 2.37.1
>
