Re: [PATCH i-g-t v2 06/11] lib/xe: Assert that cache is available for xe_wa()

[Matt Atwood <matthew.s.atwood@intel.com>]

On Wed, Sep 23, 2026 at 06:41:13PM -0300, Gustavo Sousa wrote:
> While we do print a warning if xe_wa() is called without the necessary
> data cached in xe_device, it would make any test that relies on it to
> be unreliable in that scenario.
> 
> Make sure that we cause the test to fail instead by adding an
> igt_assert() to check that the wa_cache is not NULL.
> 
> Suggested-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Atwood <matthew.s.atwood@intel.com>
> Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
> ---
>  lib/xe/xe_wa.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/lib/xe/xe_wa.c b/lib/xe/xe_wa.c
> index 05af64151e7c..5f09efd2373a 100644
> --- a/lib/xe/xe_wa.c
> +++ b/lib/xe/xe_wa.c
> @@ -178,14 +178,12 @@ int xe_wa_test_debugfs_dump_has_wa(char *dump, const char *wa)
>   * @wa: Name of the workaround to be checked.
>   *
>   * Return a boolean indicating whether the workaround is enabled.
> - * On error, returns false and a warning is printed.
>   */
>  bool xe_wa(int fd, const char *wa)
>  {
>  	struct xe_wa_cache *wa_cache = xe_device_get(fd)->wa_cache;
>  
> -	if (igt_warn_on(!wa_cache))
> -		return false;
> +	igt_assert(wa_cache);
>  
>  	for (int i = 0; i < wa_cache->num_dumps; i++)
>  		if (debugfs_dump_has_wa(wa_cache->dumps[i], wa))
> 
> -- 
> 2.55.0
>