Re: [PATCH v10 1/9] drm/xe/gt: Use GT ordered workqueue for wedging

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] drm/xe: gt_wedge_worker assertion fails if device is suspended
--

commit ad400bddb33c2efaf9ed050aad03a83745b08aa2
Author: Raag Jadav <raag.jadav@intel.com>

drm/xe/gt: Use GT ordered workqueue for wedging

This commit introduces a GT wedge worker to handle execution queue
teardown asynchronously, avoiding scheduler disruption while jobs are
in-flight. It also adds a for_each_initialized_gt() macro to prevent
races if the device is wedged during the probe sequence.

> diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
> index bc9bf6f0b5a79..27f32627a1f9b 100644
> --- a/drivers/gpu/drm/xe/xe_gt.c
> +++ b/drivers/gpu/drm/xe/xe_gt.c
[ ... ]
> @@ -1204,6 +1209,15 @@ struct xe_hw_engine *xe_gt_any_hw_engine(struct xe_gt *gt)
>  	return NULL;
>  }
>  
> +static void gt_wedge_worker(struct work_struct *w)
> +{
> +	struct xe_gt *gt = container_of(w, typeof(*gt), wedge.worker);
> +
> +	xe_device_assert_mem_access(gt_to_xe(gt));

[Severity: High]
Does this assertion fail if the device is suspended?

If a hardware fault or PCIe AER injection triggers wedging while the
device is in a runtime-suspended state (like D3hot), the error handler
calls xe_device_declare_wedged() in drivers/gpu/drm/xe/xe_device.c.

That function uses xe_pm_runtime_get_noresume():

	if (!atomic_xchg(&xe->wedged.flag, 1)) {
		xe->needs_flr_on_fini = true;
		xe_pm_runtime_get_noresume(xe);

This increments the PM usage count but leaves the device suspended.

When this asynchronous gt_wedge_worker executes, it unconditionally hits
xe_device_assert_mem_access(). Since the device is still
runtime-suspended, this assertion will fail, potentially causing a kernel
warning that disrupts the recovery process.

> +	xe_uc_declare_wedged(&gt->uc);
> +	xe_tlb_inval_reset(&gt->tlb_inval);
> +}
> +

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261016072623.457392-1-raag.jadav@intel.com?part=1