Re: [PATCH v4 01/12] drm/xe/cper: Add CPER reporting entry point for xe_log

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] drm/xe: Silent telemetry loss when hardware SIGIDs lack exact RAS data payload
--

commit a8ad6417f3a91a193b5e9bf972e9a0e60ff46e5a
Author: Badal Nilawar <badal.nilawar@intel.com>

drm/xe/cper: Add CPER reporting entry point for xe_log

Introduce xe_cper_emit() as the single entry point for emitting
CPER records from xe_log. Hardware SIGIDs are routed to a hardware
error helper, and software SIGIDs are left as TODO.

> diff --git a/drivers/gpu/drm/xe/xe_cper.c b/drivers/gpu/drm/xe/xe_cper.c
> new file mode 100644
> index 0000000000000..c5caa0e8f496a
> --- /dev/null
> +++ b/drivers/gpu/drm/xe/xe_cper.c
[ ... ]
> +void xe_cper_emit(struct xe_device *xe, int cper_sev, enum xe_sigid sigid,
> +		  u32 component, u32 location, const void *data, size_t len,
> +		  struct va_format *vaf)
> +{
> +	if (!xe)
> +		return;
> +
> +	if (is_hw_sigid(sigid)) {
> +		if (!IS_ERR_OR_NULL(data) && len == sizeof(struct xe_ras_error_class))
> +			xe_cper_emit_hw_err(xe, cper_sev, sigid, data);
> +		return;

[Severity: Medium]
If a hardware SIGID log entry is missing the expected RAS data payload (e.g.,
the data pointer is ERR_PTR, NULL, or the length is unexpected), does this
unconditionally return and drop the telemetry?

Since there is no fallback to emit a software CPER log or print a warning,
it looks like this could result in a silent loss of hardware error data.

> +	}
> +
> +	/* TODO software CPER */
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261009104602.1075405-14-badal.nilawar@intel.com?part=1