Re: [PATCH v3 09/12] drm/xe/xe_ras: Report core compute errors using SIGID

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Hardware error CPER emission is completely disabled due to an inverted condition check on the error signature ID.
--

commit 5c8af98618264b3c8466a350a075ee0f1ad606e7
Author: Badal Nilawar <badal.nilawar@intel.com>

    drm/xe/xe_ras: Report core compute errors using SIGID
    
    Use xe_log_* helpers to report core compute errors.

> diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
> index 3a7f80370827e..78ea5e33e244b 100644
> --- a/drivers/gpu/drm/xe/xe_ras.c
> +++ b/drivers/gpu/drm/xe/xe_ras.c
> @@ -310,21 +310,27 @@ static void ras_send_error_event(struct xe_device *xe, u8 severity, u8 component
>  	xe_drm_ras_event(xe, drm_component, drm_severity, response.value);
>  }
>  
> -static u8 handle_core_compute_errors(struct xe_ras_error_array *arr)
> +static u8 handle_core_compute_errors(struct xe_device *xe, struct xe_ras_error_array *arr)
>  {
>  	struct xe_ras_compute_error *error_info = (void *)arr->details;
> +	u8 cper_sev = ras_sev_to_cper_sev(arr->counter.common.severity);
>  	u8 uncorr_type;
>  
>  	uncorr_type = FIELD_GET(CORE_COMPUTE_UNCORR_TYPE, error_info->log_header);
>  
>  	/* Request a reset if error is global */
> -	if (uncorr_type == GLOBAL_UNCORR_ERROR)
> +	if (uncorr_type == GLOBAL_UNCORR_ERROR) {
> +		xe_log_comp(xe, cper_sev, CORE_COMPUTE, &arr->counter, sizeof(arr->counter),
> +			    "Global uncorrectable error detected\n");

[Severity: High]
This is a pre-existing issue, but will these hardware errors actually be emitted
by the CPER infrastructure?

Looking at xe_emit_hardware_error_cper() in drivers/gpu/drm/xe/xe_cper.c, which
processes these hardware SIGIDs, there appears to be an inverted condition check:

drivers/gpu/drm/xe/xe_cper.c:xe_emit_hardware_error_cper() {
	...
	if ((int)sigid >= INTEL_SIGID_GPU_XE_HARDWARE_START)
		return;
	...
}

Any time xe_emit_hardware_error_cper() is called with a valid hardware SIGID,
does this inverted check silently drop the CPER record instead of processing it?

>  		return XE_RAS_RECOVERY_ACTION_RESET;
> +	}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260906172604.2215987-14-badal.nilawar@intel.com?part=9