On 9/26/24 14:54, Vlastimil Babka wrote: > On 9/25/24 14:56, Hyeonggon Yoo wrote: >> >> I don't think this was a noise :) IMO some people want to see WARNING >> during testing to catch errors, >> but not for the slub_kunit test case. I think a proper approach here >> would be suppressing >> warnings while running slub_kunit test cases, but print WARNING when >> it is not running slub_kunit test cases. >> >> That would require some work changing the slub error reporting logic >> to print WARNING on certain errors. >> Any opinions, Vlastimil? > > Yes, we should suppress the existing warning on kmem_cache_destroy() in Done here: https://lore.kernel.org/all/20240930-b4-slub-kunit-fix-v1-0-32ca9dbbbc11@xxxxxxx/ > kunit test context, and separately we can change pr_err() to WARN() as long > as they are still suppressed in kunit test context. Can be done later separately as it would be out of the scope for a rcX fix. >> Thanks, >> Hyeonggon >
