On 2/26/25 09:12, Hyesoo Yu wrote:
> If a slab object is corrupted or an error occurs in its internal
> value, continuing after restoration may cause other side effects.
> At this point, it is difficult to debug because the problem occurred
> in the past. It is useful to use WARN() to catch errors at the point
> of issue because WARN() could trigger panic for system debugging when
> panic_on_warn is enabled. WARN() is added where to detect the error
> on slab_err and object_err.
>
> It makes sense to only do the WARN() after printing the logs. slab_err
> is splited to __slab_err that calls the WARN() and it is called after
> printing logs.
>
> Changes in v4:
> - Remove WARN() in kmem_cache_destroy to remove redundant warning.
>
> Changes in v3:
> - move the WARN from slab_fix to slab_err, object_err and check_obj to
> use WARN on all error reporting paths.
>
> Changes in v2:
> - Replace direct calling with BUG_ON with the use of WARN in slab_fix.
>
> Signed-off-by: Hyesoo Yu <hyesoo.yu@xxxxxxxxxxx>
As Harry said. I'll remove that locally.
> ---
> mm/slab_common.c | 3 ---
> mm/slub.c | 31 +++++++++++++++++++------------
> 2 files changed, 19 insertions(+), 15 deletions(-)
>
> diff --git a/mm/slab_common.c b/mm/slab_common.c
> index 477fa471da18..d13f4ffe252b 100644
> --- a/mm/slab_common.c
> +++ b/mm/slab_common.c
> @@ -517,9 +517,6 @@ void kmem_cache_destroy(struct kmem_cache *s)
> kasan_cache_shutdown(s);
>
> err = __kmem_cache_shutdown(s);
> - if (!slab_in_kunit_test())
> - WARN(err, "%s %s: Slab cache still has objects when called from %pS",
> - __func__, s->name, (void *)_RET_IP_);
I think I'll keep this one, because the more detailed warning via
list_slab_objects() is only enabled with CONFIG_SLUB_DEBUG.
If it's not enabled, the kmem_cache_destroy() failure rather should not be
silent.
So slab_in_kunit_test() would also stay.
> } else {
> - list_slab_objects(s, slab,
> - "Objects remaining in %s on __kmem_cache_shutdown()");
> + list_slab_objects(s, slab);
I tried to extract slab_bug() and __slab_err() from list_slab_objects() but
they were also only available with CONFIG_SLUB_DEBUG.
Perhaps we can improve that, but as a follow-up cleanup so we don't hold
this up further.
> }
> }
> spin_unlock_irq(&n->list_lock);
