Re: [PATCH 4/4] mm/slub: free KFENCE objects in slab_free_hook()

[Chengming Zhou <chengming.zhou@linux.dev>]

On 2023/12/6 17:58, Vlastimil Babka wrote:
> On 12/5/23 14:27, Chengming Zhou wrote:
>> On 2023/12/5 03:34, Vlastimil Babka wrote:
>>> When freeing an object that was allocated from KFENCE, we do that in the
>>> slowpath __slab_free(), relying on the fact that KFENCE "slab" cannot be
>>> the cpu slab, so the fastpath has to fallback to the slowpath.
>>>
>>> This optimization doesn't help much though, because is_kfence_address()
>>> is checked earlier anyway during the free hook processing or detached
>>> freelist building. Thus we can simplify the code by making the
>>> slab_free_hook() free the KFENCE object immediately, similarly to KASAN
>>> quarantine.
>>>
>>> In slab_free_hook() we can place kfence_free() above init processing, as
>>> callers have been making sure to set init to false for KFENCE objects.
>>> This simplifies slab_free(). This places it also above kasan_slab_free()
>>> which is ok as that skips KFENCE objects anyway.
>>>
>>> While at it also determine the init value in slab_free_freelist_hook()
>>> outside of the loop.
>>>
>>> This change will also make introducing per cpu array caches easier.
>>>
>>> Tested-by: Marco Elver <elver@google.com>
>>> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
>>> ---
>>>  mm/slub.c | 22 ++++++++++------------
>>>  1 file changed, 10 insertions(+), 12 deletions(-)
>>>
>>> diff --git a/mm/slub.c b/mm/slub.c
>>> index ed2fa92e914c..e38c2b712f6c 100644
>>> --- a/mm/slub.c
>>> +++ b/mm/slub.c
>>> @@ -2039,7 +2039,7 @@ static inline void memcg_slab_free_hook(struct kmem_cache *s, struct slab *slab,
>>>   * production configuration these hooks all should produce no code at all.
>>>   *
>>>   * Returns true if freeing of the object can proceed, false if its reuse
>>> - * was delayed by KASAN quarantine.
>>> + * was delayed by KASAN quarantine, or it was returned to KFENCE.
>>>   */
>>>  static __always_inline
>>>  bool slab_free_hook(struct kmem_cache *s, void *x, bool init)
>>> @@ -2057,6 +2057,9 @@ bool slab_free_hook(struct kmem_cache *s, void *x, bool init)
>>>  		__kcsan_check_access(x, s->object_size,
>>>  				     KCSAN_ACCESS_WRITE | KCSAN_ACCESS_ASSERT);
>>>  
>>> +	if (kfence_free(kasan_reset_tag(x)))
>>
>> I'm wondering if "kasan_reset_tag()" is needed here?
> 
> I think so, because AFAICS the is_kfence_address() check in kfence_free()
> could be a false negative otherwise. In fact now I even question some of the

Ok.

> other is_kfence_address() checks in mm/slub.c, mainly
> build_detached_freelist() which starts from pointers coming directly from
> slab users. Insight from KASAN/KFENCE folks appreciated :)
> 
I know very little about KASAN/KFENCE, looking forward to their insight. :)

Just saw a check in __kasan_slab_alloc():

	if (is_kfence_address(object))
		return (void *)object;

So thought it seems that a kfence object would be skipped by KASAN.

Thanks!