Re: Possible mem cgroup bug in kernels between 4.18.0 and 5.3-rc1.

From: Masoud Sharbiani <msharbiani@apple.com>
To: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Michal Hocko <mhocko@kernel.org>,
	Greg KH <gregkh@linuxfoundation.org>,
	hannes@cmpxchg.org, vdavydov.dev@gmail.com, linux-mm@kvack.org,
	cgroups@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: Possible mem cgroup bug in kernels between 4.18.0 and 5.3-rc1.
Date: Sat, 03 Aug 2019 10:41:40 -0700	[thread overview]
Message-ID: <77568CFC-280C-4C0F-85FC-92F1212BC6FC@apple.com> (raw)
In-Reply-To: <d7efccf4-7f07-10da-077d-a58dafbf627e@I-love.SAKURA.ne.jp>

> On Aug 3, 2019, at 8:51 AM, Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> wrote:
> 
> Masoud, will you try this patch?

Gladly.
It looks like it is working (and OOMing properly).

> 
> By the way, is /sys/fs/cgroup/memory/leaker/memory.usage_in_bytes remains non-zero
> despite /sys/fs/cgroup/memory/leaker/tasks became empty due to memcg OOM killer expected?
> Deleting big-data-file.bin after memcg OOM killer reduces some, but still remains
> non-zero.

Yes. I had not noticed that:

[ 1114.190477] oom_reaper: reaped process 1942 (leaker), now anon-rss:0kB, file-
rss:0kB, shmem-rss:0kB
./test-script.sh: line 16:  1942 Killed                  ./leaker -p 10240 -c 100000

[root@localhost laleaker]# cat  /sys/fs/cgroup/memory/leaker/memory.usage_in_bytes
3194880
[root@localhost laleaker]# cat  /sys/fs/cgroup/memory/leaker/memory.limit_in_bytes
536870912
[root@localhost laleaker]# rm -f big-data-file.bin
[root@localhost laleaker]# cat  /sys/fs/cgroup/memory/leaker/memory.usage_in_bytes
2838528

Thanks!
Masoud

PS: Tried hand-back-porting it to 4.19-y and it didn’t work. I think there are other patches between 4.19.0 and 5.3 that could be necessary…

> 
> ----------------------------------------
> From 2f92c70f390f42185c6e2abb8dda98b1b7d02fa9 Mon Sep 17 00:00:00 2001
> From: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
> Date: Sun, 4 Aug 2019 00:41:30 +0900
> Subject: [PATCH] memcg, oom: don't require __GFP_FS when invoking memcg OOM killer
> 
> Masoud Sharbiani noticed that commit 29ef680ae7c21110 ("memcg, oom: move
> out_of_memory back to the charge path") broke memcg OOM called from
> __xfs_filemap_fault() path. It turned out that try_chage() is retrying
> forever without making forward progress because mem_cgroup_oom(GFP_NOFS)
> cannot invoke the OOM killer due to commit 3da88fb3bacfaa33 ("mm, oom:
> move GFP_NOFS check to out_of_memory"). Regarding memcg OOM, we need to
> bypass GFP_NOFS check in order to guarantee forward progress.
> 
> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
> Reported-by: Masoud Sharbiani <msharbiani@apple.com>
> Bisected-by: Masoud Sharbiani <msharbiani@apple.com>
> Fixes: 29ef680ae7c21110 ("memcg, oom: move out_of_memory back to the charge path")
> ---
> mm/oom_kill.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index eda2e2a..26804ab 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -1068,9 +1068,10 @@ bool out_of_memory(struct oom_control *oc)
> 	 * The OOM killer does not compensate for IO-less reclaim.
> 	 * pagefault_out_of_memory lost its gfp context so we have to
> 	 * make sure exclude 0 mask - all other users should have at least
> -	 * ___GFP_DIRECT_RECLAIM to get here.
> +	 * ___GFP_DIRECT_RECLAIM to get here. But mem_cgroup_oom() has to
> +	 * invoke the OOM killer even if it is a GFP_NOFS allocation.
> 	 */
> -	if (oc->gfp_mask && !(oc->gfp_mask & __GFP_FS))
> +	if (oc->gfp_mask && !(oc->gfp_mask & __GFP_FS) && !is_memcg_oom(oc))
> 		return true;
> 
> 	/*
> -- 
> 1.8.3.1
> 
> 