Re: [PATCH v4 0/4] make unused huge shrinker memcg aware

[Qi Zheng <qi.zheng@linux.dev>]

Hi Andrew,

On 8/28/26 6:50 AM, Andrew Morton wrote:
> On Mon, 17 Aug 2026 17:03:24 +0800 Qi Zheng <qi.zheng@linux.dev> wrote:
> 
>> Changes in v4:
>> Changes in v3:
>> Changes in v2:
> 
> Thanks for the diligent versioning info.  fyi, it is conventional to
> maintain this below the --- separator.  It's not really the most
> important part of the [0/N]!

Got it.

> 
>>
>> The shmem unused huge shrinker maintains a per-superblock list of inodes
>> whose tail huge folio extends beyond i_size.  Because this list is not
>> memcg aware, reclaim triggered by memcg A can scan inodes across the
>> entire superblock and split huge folios charged to unrelated memcg B,
>> causing unexpected impact on it.
>>
>> In the worst case, memcg A has no reclaimable shmem at all, making the
>> reclaim entirely useless and incurring unnecessary latency.  We observed
>> this in production, where page lock contention during split caused
>> multi-hundred-millisecond stalls:
> 
> Ugh.  That's the most important part!
> 
>>    tid 11340 comm scanner locked a page for 182264 us! kstack:
>>            unlock_page+1
>>            split_huge_page_to_list+3135
>>            shmem_unused_huge_shrink+767
>>            super_cache_scan+329
>>            do_shrink_slab+291
>>            shrink_slab+533
>>            shrink_node+400
>>            do_try_to_free_pages+206
>>            try_to_free_mem_cgroup_pages+262
>>            try_charge_memcg+591
>>            mem_cgroup_charge+136
>>            __handle_mm_fault+2431
>>            handle_mm_fault+194
>>            do_user_addr_fault+462
>>            __do_page_fault+176
>>            do_page_fault+48
>>            page_fault+62
>>
>> Usama's recent patch [1] prevents the shmem unused shrinker from being
>> invoked during memcg-level reclaim altogether, but this is overly
>> conservative: we can do better by reclaiming only the shmem charged to
>> the reclaiming memcg.
>>
>> This series converts the shrinker list to a memcg-aware list_lru, so
>> that non-root memcg reclaim walks only candidates charged to the
>> reclaiming memcg.  Global reclaim, root memcg reclaim and shmem quota
>> reclaim retain their existing global semantics.
>>
>> To avoid pinning a dying memcg through a long-lived CSS reference, each
>> inode stores an obj_cgroup reference instead of a mem_cgroup reference.
>> The list_lru add/delete paths resolve the current memcg from the objcg
>> under RCU, staying consistent with list_lru's own memcg migration on
>> offline.
> 
> Sashiko said a few things and they look disturbing-if-true:
> 
> 	https://sashiko.dev/#/patchset/cover.1786955972.git.zhengqi.arch@bytedance.com
> 
> (Apologies if this has already been considered - we don't have ways of
> tracking all this (yet, I hope) apart from personal memory and personal
> memorys are quite fried at present)

As both Usama and I have pointed out [1][2], [PATCH v4 1/4] is the part
that got dropped during the merge. The complete patch [3] was actually
reviewed a while ago.

[1].https://lore.kernel.org/all/20260810101954.822260-1-usama.arif@linux.dev/
[2]. 
https://lore.kernel.org/all/9c7efd5f-f8d3-4926-acb4-34c326ffb1c3@linux.dev/
[3]. 
https://lore.kernel.org/all/20260715103516.2410175-1-usama.arif@linux.dev/

Thanks,
Qi

>