Re: [Kernel Bug] KASAN: slab-use-after-free Read in __blkcg_rstat_flush

[Jens Axboe]

On 2/3/26 7:40 AM, Ming Lei wrote:
> On Tue, Feb 03, 2026 at 03:16:43PM +0100, Michal Koutný wrote:
>> On Tue, Feb 03, 2026 at 10:08:07PM +0800, Ming Lei <ming.lei@xxxxxxxxxx> wrote:
>>> I can't parse your question, here blkg_release() simply needs to flush
>>> all stats. Why do you talk about preventing new flush? why is it related
>>> with this UAF?
>>
>> What prevents this fix:
>>
>> --- a/block/blk-cgroup.c
>> +++ b/block/blk-cgroup.c
>> @@ -169,14 +169,6 @@ static void __blkg_release(struct rcu_head *rcu)
>>  #ifdef CONFIG_BLK_CGROUP_PUNT_BIO
>>         WARN_ON(!bio_list_empty(&blkg->async_bios));
>>  #endif
>> -       /*
>> -        * Flush all the non-empty percpu lockless lists before releasing
>> -        * us, given these stat belongs to us.
>> -        *
>> -        * blkg_stat_lock is for serializing blkg stat update
>> -        */
>> -       for_each_possible_cpu(cpu)
>> -               __blkcg_rstat_flush(blkcg, cpu);
>>
>>         /* release the blkcg and parent blkg refs this blkg has been holding */
>>         css_put(&blkg->blkcg->css);
>> @@ -195,6 +187,15 @@ static void blkg_release(struct percpu_ref *ref)
>>  {
>>         struct blkcg_gq *blkg = container_of(ref, struct blkcg_gq, refcnt);
>>
>> +       /*
>> +        * Flush all the non-empty percpu lockless lists before releasing
>> +        * us, given these stat belongs to us.
>> +        *
>> +        * blkg_stat_lock is for serializing blkg stat update
>> +        */
>> +       for_each_possible_cpu(cpu)
>> +               __blkcg_rstat_flush(blkcg, cpu);
>> +
>>         call_rcu(&blkg->rcu_head, __blkg_release);
>>  }
> 
> This one looks more clever, can you send one formal patch?

I'll drop the previous one for now - I didn't really mind it,
but if we can fix it without, then yeah agree that's better.

-- 
Jens Axboe