Re: [PATCH v6 12/12] drm/xe/vram: Report VRAM 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:
- [Low] xe: format string mismatch passing u64 pointer to %pa in xe_vram_probe
--

commit 0d8f1a8de9f08409f13227d2abe582e14cf48299
Author: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Subject: drm/xe/vram: Report VRAM errors using structured xe_log API

This commit replaces drm_err() and drm_info() with the structured
xe_log_err_fatal() and xe_log_info() macros using the VRAM component tag.

> diff --git a/drivers/gpu/drm/xe/xe_vram.c b/drivers/gpu/drm/xe/xe_vram.c
> index 73aef047eb386..772704dfe12fe 100644
> --- a/drivers/gpu/drm/xe/xe_vram.c
> +++ b/drivers/gpu/drm/xe/xe_vram.c

[ ... ]

> @@ -343,8 +350,9 @@ int xe_vram_probe(struct xe_device *xe)
>  			return err;
>  
>  		if (total_size > lmem_bar.io_size) {
> -			drm_info(&xe->drm, "VRAM: %pa is larger than resource %pa\n",
> -				 &total_size, &lmem_bar.io_size);
> +			xe_log_info(xe, VRAM,
> +				    "Total size %pa is larger than resource %pa\n",
> +				    &total_size, &lmem_bar.io_size);

[Severity: Low]
Does this format string mismatch cause truncation on 32-bit systems?

The %pa format specifier expects a pointer to phys_addr_t or
resource_size_t. However, total_size is declared as a u64 earlier in
xe_vram_probe().

On 32-bit systems where phys_addr_t is 32-bit, %pa will dereference the
pointer as a 32-bit integer. On little-endian systems, this prints a
truncated size, and on big-endian systems, it reads the upper 32 bits
(printing 0 for values under 4GB).

Since the GCC printf format checker does not catch custom %p pointer type
mismatches, should this log instead use %llx and pass total_size by value?

>  		}
>  
>  		remain_io_size -= min_t(u64, tile->mem.vram->actual_physical_size, remain_io_size);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261012051823.2201360-1-dnyaneshwar.bhadane@intel.com?part=12