Re: [PATCH v2 02/13] mm/slub: skip handle_failed_objexts_alloc() with profiling disabled

[Vlastimil Babka (SUSE)]

On 7/21/26 07:51, Harry Yoo wrote:
> 
> 
> On 7/20/26 11:16 PM, Vlastimil Babka (SUSE) wrote:
>> The function might get called with memory allocation profiling disabled,
>> when the obj_ext array is allocated for objcg pointers only. The
>> handling is however unnecessary in that case, so skip it.
>> 
>> This would otherwise become a real bug later, as pointed out by sashiko.
>> For now it's just an optimization.
>> 
>> Link: https://sashiko.dev/#/patchset/20260715-b4-objext_split-v1-0-9a49c4ccf4c3@xxxxxxxxxx?part=10
>> Signed-off-by: Vlastimil Babka (SUSE) <vbabka@xxxxxxxxxx>
>> ---
>>  mm/slub.c | 7 ++++---
>>  1 file changed, 4 insertions(+), 3 deletions(-)
>> 
>> diff --git a/mm/slub.c b/mm/slub.c
>> index 76acb78f2655..95fa6fbad11a 100644
>> --- a/mm/slub.c
>> +++ b/mm/slub.c
>> @@ -2101,15 +2101,16 @@ static inline bool mark_failed_objexts_alloc(struct slab *slab)
>>  static inline void handle_failed_objexts_alloc(unsigned long obj_exts,
>>  			struct slabobj_ext *vec, unsigned int objects)
>>  {
>> +	if (!mem_alloc_profiling_enabled())
>> +		return;
> 
> This looks racy.

Indeed, not sure how I convinced myself it's ok.

> Can we instead do this later in the series depending on
> slab_obj_ext_has_codetag_key's value?

Yeah.

>>  	/*
>>  	 * If vector previously failed to allocate then we have live
>>  	 * objects with no tag reference. Mark all references in this
>>  	 * vector as empty to avoid warnings later on.
>>  	 */
>>  	if (obj_exts == OBJEXTS_ALLOC_FAIL) {
>> -		unsigned int i;
>> -
>> -		for (i = 0; i < objects; i++)
>> +		for (unsigned int i = 0; i < objects; i++)
>>  			set_codetag_empty(&vec[i].ref);
>>  	}
>>  }
>> 
>