Re: [External] Re: [PATCH v5.4.y, v4.19.y] mm: memcg/slab: fix memory leak at non-root kmem_cache destroy

[Muchun Song <songmuchun@bytedance.com>]

On Wed, Jul 15, 2020 at 11:43 PM Vlastimil Babka <vbabka@suse.cz> wrote:
>
> On 7/15/20 5:13 PM, Muchun Song wrote:
> > On Wed, Jul 15, 2020 at 7:32 PM Vlastimil Babka <vbabka@suse.cz> wrote:
> >>
> >> On 7/7/20 8:27 AM, Muchun Song wrote:
> >> > If the kmem_cache refcount is greater than one, we should not
> >> > mark the root kmem_cache as dying. If we mark the root kmem_cache
> >> > dying incorrectly, the non-root kmem_cache can never be destroyed.
> >> > It resulted in memory leak when memcg was destroyed. We can use the
> >> > following steps to reproduce.
> >> >
> >> >   1) Use kmem_cache_create() to create a new kmem_cache named A.
> >> >   2) Coincidentally, the kmem_cache A is an alias for kmem_cache B,
> >> >      so the refcount of B is just increased.
> >> >   3) Use kmem_cache_destroy() to destroy the kmem_cache A, just
> >> >      decrease the B's refcount but mark the B as dying.
> >> >   4) Create a new memory cgroup and alloc memory from the kmem_cache
> >> >      A. It leads to create a non-root kmem_cache for allocating.
> >> >   5) When destroy the memory cgroup created in the step 4), the
> >> >      non-root kmem_cache can never be destroyed.
> >> >
> >> > If we repeat steps 4) and 5), this will cause a lot of memory leak.
> >> > So only when refcount reach zero, we mark the root kmem_cache as dying.
> >> >
> >> > Fixes: 92ee383f6daa ("mm: fix race between kmem_cache destroy, create and deactivate")
> >> > Signed-off-by: Muchun Song <songmuchun@bytedance.com>
> >>
> >> CC Roman, who worked in this area recently.
> >>
> >> Also why is this marked "[PATCH v5.4.y, v4.19.y]"? Has it been fixed otherwise
> >> in 5.5+ ?
> >
> > Because the memcg slab/slub is reworked by Roman since v5.8.
>
> That rework is in mmotm, so scheduled for 5.9, AFAIK. If you mean "The new
> cgroup slab memory controller" series.

Yeah, I mean "The new cgroup slab memory controller".

>
> > Therefore, this problem exists in v5.7 and below.
>
> Even 5.7 has a stable series, so no need to list only the LTS's.
> To sum up, the patch (once reviewed) should be queued for mainline as usual,
> perhaps sent before 5.8 is final, if deemed safe enough, and with added
>
> Cc: <stable@vger.kernel.org>
>
> and the Fixes: tag you provided, the applicable stable versions will pick it.

Got it. Thanks.

>
> Vlastimil



-- 
Yours,
Muchun