On 1/16/26 17:20, Alex Deucher wrote: > It was leftover from when the driver supported drm sched > resubmit. That was dropped long ago, so drop this as well. > Leaving this in place also causes userspace to treat > the context as innocent. Removing this fixes reset behavior. > > Signed-off-by: Alex Deucher <[email protected]>
Looks like you didn't saw my earlier reply. This won't work because we unfortunately still need this for the AMDGPU_CTX_QUERY2_FLAGS_GUILTY. We first need to fix that one before we can drop that here. Regards, Christian. > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > index ed7f13752f462..70faf914b4f0b 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > @@ -5817,9 +5817,6 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device > *adev, > > amdgpu_fence_driver_isr_toggle(adev, false); > > - if (job && job->vm) > - drm_sched_increase_karma(&job->base); > - > r = amdgpu_reset_prepare_hwcontext(adev, reset_context); > /* If reset handler not implemented, continue; otherwise return */ > if (r == -EOPNOTSUPP)
