Re: [PATCH v2] drm/xe/pci_error: Allow debugging PCI errors

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Unconditional device wedging and disconnection on recoverable non-fatal AER errors when debugging mode is enabled.
--

--- Patch [1]: [PATCH v2] drm/xe/pci_error: Allow debugging PCI errors ---
commit 26ac63dc9350283e2a915daf342c236caf34f680
Author: Raag Jadav <[email protected]>

drm/xe/pci_error: Allow debugging PCI errors

This commit updates the PCI error handling path to disconnect the device
and prevent a bus reset when the XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET
debug mode is enabled. This preserves the device state so it is available
to the user for debugging.

> diff --git a/drivers/gpu/drm/xe/xe_pci_error.c b/drivers/gpu/drm/xe/xe_pci_error.c
> index e41af2ac7f23b..98aa26275ad66 100644
> --- a/drivers/gpu/drm/xe/xe_pci_error.c
> +++ b/drivers/gpu/drm/xe/xe_pci_error.c
> @@ -83,6 +83,12 @@ static pci_ers_result_t xe_pci_error_mmio_enabled(struct pci_dev *pdev)
>  	xe_info(xe, "PCI error: MMIO enabled\n");
>  	action = xe_ras_process_errors(xe);
>  
> +	/* User wants to debug the error, prevent reset */
> +	if (xe->wedged.mode == XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET) {
> +		xe_device_declare_wedged(xe);
> +		return PCI_ERS_RESULT_DISCONNECT;
> +	}

[Severity: Medium]
Does this code unnecessarily wedge the device on recoverable errors? 

When XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET is enabled, any non-fatal AER 
that reaches xe_pci_error_mmio_enabled() will unconditionally wedge the 
device here. This happens even if xe_ras_process_errors(xe) successfully 
handles the error and returns XE_RAS_RECOVERY_ACTION_RECOVERED in the 
action variable.

Since the commit message specifically mentions preventing bus resets for 
debugging, should this disconnect and wedge logic only be triggered if 
action == XE_RAS_RECOVERY_ACTION_RESET?

> +
>  	return ras_action_to_pci_result(pdev, action);
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1