On Thu, Jan 9, 2025 at 3:58 AM Kenneth Feng <kenneth.feng@xxxxxxx> wrote:
>
> Disable gfxoff with the compute workload on gfx12. This is a
> workaround for the opencl test failure.
>
> Signed-off-by: Kenneth Feng <kenneth.feng@xxxxxxx>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> index 2e5732dfd425..e4cf25c47509 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> @@ -715,8 +715,10 @@ int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev,
> void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle)
> {
> enum amd_powergating_state state = idle ? AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE;
> - if (IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 11 &&
> - ((adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK) <= 64)) {
> + if ((IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 11 &&
> + ((adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK) <= 64)) ||
> + (IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 12 &&
> + ((adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK) > 0x6C))) {
If this is just a specific FW version that is problematic, we should
just check the particular version. If this is a general problem, just
check the IP version until the issue is fixed and then we can update
the check for a specific fw version, etc.
Alex
> pr_debug("GFXOFF is %s\n", idle ? "enabled" : "disabled");
> amdgpu_gfx_off_ctrl(adev, idle);
> } else if ((IP_VERSION_MAJ(amdgpu_ip_version(adev, GC_HWIP, 0)) == 9) &&
> --
> 2.34.1
>
