Reset sequence indicates that hardware already ran into a bad state. Avoid sending unmap queue request to reset KCQ. This will also cover RAS error scenarios which need a reset to recover, hence remove the check.
Signed-off-by: Lijo Lazar <[email protected]> --- drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index 6733ff5d9628..074626650f5e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -515,15 +515,7 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id) if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) return -EINVAL; - if (!kiq_ring->sched.ready || adev->job_hang) - return 0; - /** - * This is workaround: only skip kiq_ring test - * during ras recovery in suspend stage for gfx9.4.3 - */ - if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) || - amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) && - amdgpu_ras_in_recovery(adev)) + if (!kiq_ring->sched.ready || adev->job_hang || amdgpu_in_reset(adev)) return 0; spin_lock(&kiq->ring_lock); -- 2.25.1
