Re: [PATCH 11/12] drm/xe/lmtt: Report invalidation errors using structured xe_log API

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Incorrect error code passed to structured logging and redundant error formatting.
--

--- Patch [11]: [PATCH 11/12] drm/xe/lmtt: Report invalidation errors using structured xe_log API ---
commit 923aca6a5a5574864c744c6654055fb24cc4ad4d
Author: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>

drm/xe/lmtt: Report invalidation errors using structured xe_log API

This commit replaces xe_tile_sriov_err() with xe_log_err() using the LMTT
component tag for LMTT and MERT invalidation failures. It attempts to
preserve the SR-IOV prefix in the log messages.

> diff --git a/drivers/gpu/drm/xe/xe_lmtt.c b/drivers/gpu/drm/xe/xe_lmtt.c
> --- a/drivers/gpu/drm/xe/xe_lmtt.c
> +++ b/drivers/gpu/drm/xe/xe_lmtt.c
[ ... ]
> @@ -302,14 +303,16 @@ void xe_lmtt_invalidate_hw(struct xe_lmtt *lmtt)
>  
>  	err = lmtt_invalidate_hw(lmtt);
>  	if (err)
> -		xe_tile_sriov_err(tile, "LMTT invalidation failed (%pe)",
> -				  ERR_PTR(err));
> +		xe_log_err(tile, LMTT, -EIO,  "%s invalidation failed (%pe)",
> +			   xe_sriov_printk_prefix(xe),
> +			   ERR_PTR(err));

[Severity: Medium]
Since err holds the actual error code returned by lmtt_invalidate_hw(),
should this pass err to xe_log_err instead of hardcoding -EIO?

Additionally, does explicitly including (%pe) in the format string here
result in redundant and conflicting error codes in the log output, since
xe_log_err() automatically records the passed error code?

>  
>  	if (xe_device_has_mert(xe) && xe_tile_is_root(tile)) {
>  		err = xe_mert_invalidate_lmtt(xe);
>  		if (err)
> -			xe_tile_sriov_err(tile, "MERT LMTT invalidation failed (%pe)",
> -					  ERR_PTR(err));
> +			xe_log_err(tile, LMTT, -EIO,  "%s MERT invalidation failed (%pe)",
> +				   xe_sriov_printk_prefix(xe),
> +				   ERR_PTR(err));

[Severity: Medium]
Similar to the above, xe_mert_invalidate_lmtt() can return -ETIMEDOUT. Does
hardcoding -EIO here result in incorrect structured telemetry logs?

Also, would removing the (%pe) simplify the format string and avoid
redundant logging?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260819191442.1957986-1-dnyaneshwar.bhadane@intel.com?part=11