lkml.org 
[lkml]   [2018]   [Aug]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: WARNING in try_charge
On Mon, Aug 6, 2018 at 7:44 PM, Michal Hocko <mhocko@kernel.org> wrote:
> On Mon 06-08-18 08:42:02, syzbot wrote:
>> Hello,
>>
>> syzbot has tested the proposed patch but the reproducer still triggered
>> crash:
>> WARNING in try_charge
>>
>> Killed process 6410 (syz-executor5) total-vm:37708kB, anon-rss:2128kB,
>> file-rss:0kB, shmem-rss:0kB
>> oom_reaper: reaped process 6410 (syz-executor5), now anon-rss:0kB,
>> file-rss:0kB, shmem-rss:0kB
>> task=syz-executor5 pid=6410 invoked memcg oom killer. oom_victim=1
>
> Thank you. This is useful. The full oom picture is this
> : [ 65.363983] task=syz-executor5 pid=6415 invoked memcg oom killer. oom_victim=0
> [...]
> : [ 65.920355] Task in /ile0 killed as a result of limit of /ile0
> : [ 65.926389] memory: usage 0kB, limit 0kB, failcnt 20
> : [ 65.931518] memory+swap: usage 0kB, limit 9007199254740988kB, failcnt 0
> : [ 65.938296] kmem: usage 0kB, limit 9007199254740988kB, failcnt 0
> : [ 65.944467] Memory cgroup stats for /ile0: cache:0KB rss:0KB rss_huge:0KB shmem:0KB mapped_file:0KB dirty:0KB writeback:0KB swap:0KB inactive_anon:0KB active_anon:0KB inactive_file:0KB active_file:0KB unevictable:0KB
> : [ 65.963878] Tasks state (memory values in pages):
> : [ 65.968743] [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name
> : [ 65.977615] [ 6410] 0 6410 9427 532 61440 0 0 syz-executor5
> : [ 65.986647] Memory cgroup out of memory: Kill process 6410 (syz-executor5) score 547000 or sacrifice child
> : [ 65.996474] Killed process 6410 (syz-executor5) total-vm:37708kB, anon-rss:2128kB, file-rss:0kB, shmem-rss:0kB
> : [ 66.007471] oom_reaper: reaped process 6410 (syz-executor5), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB
> : [ 66.017652] task=syz-executor5 pid=6410 invoked memcg oom killer. oom_victim=1
> : [ 66.025137] ------------[ cut here ]------------
> : [ 66.029927] Memory cgroup charge failed because of no reclaimable memory! This looks like a misconfiguration or a kernel bug.
> : [ 66.030061] WARNING: CPU: 1 PID: 6410 at mm/memcontrol.c:1707 try_charge+0x734/0x1680
>
> So we have only a single task in the memcg and it is this task which
> triggers the OOM. It gets killed and oom_reaped. This means that
> out_of_memory should return with true and so we should retry and force
> the charge as I've already mentioned. For some reason this task has
> triggered the oom killer path again and then we haven't found any
> eligible task and resulted in the warning. This shouldn't happen.
>
> I will stare to the code some more to see how the heck we get there
> without passing
> if (unlikely(tsk_is_oom_victim(current) ||
> fatal_signal_pending(current) ||
> current->flags & PF_EXITING))
> goto force;


This is the syzkaller reproducer if it will give you any hints:

mkdir(&(0x7f0000000340)='./file0\x00', 0x0)
mount(&(0x7f00000000c0)='./file0//ile0\x00',
&(0x7f0000000080)='./file0\x00', &(0x7f0000000200)='cgroup2\x00', 0x0,
0x0)
r0 = open(&(0x7f0000000040)='./file0//ile0\x00', 0x0, 0x0)
r1 = openat$cgroup_procs(r0, &(0x7f00000001c0)='cgroup.procs\x00', 0x2, 0x0)
write$cgroup_pid(r1, &(0x7f0000000100), 0x12)
syz_mount_image$xfs(&(0x7f0000000280)='xfs\x00',
&(0x7f00000002c0)='./file0//ile0\x00', 0x0, 0x0, &(0x7f0000000740),
0x0, &(0x7f0000000800))
getsockopt$sock_cred(r0, 0x1, 0x11, &(0x7f0000000180), &(0x7f0000000240)=0xc)
r2 = open(&(0x7f0000000040)='./file0//ile0\x00', 0x0, 0x0)
write$FUSE_STATFS(r2, &(0x7f0000000280)={0x60, 0xfffffffffffffffe,
0x7, {{0x401, 0x7, 0x7d, 0x1c, 0xd8, 0xb7, 0x2, 0x3}}}, 0x60)
r3 = openat$cgroup_int(r2, &(0x7f00000001c0)='memory.max\x00', 0x2, 0x0)
write$cgroup_int(r3, &(0x7f00000000c0), 0x12)

\
 
 \ /
  Last update: 2018-08-06 19:51    [W:0.101 / U:2.956 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site