Re: Found the commit that causes the OOMs

[KOSAKI Motohiro]

2009/6/30 Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>:
> On Mon, 29 Jun 2009 13:43:55 +0100 David Howells <dhowells@xxxxxxxxxx> wrote:
>
>> KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx> wrote:
>>
>> > David, Can you please try to following patch? it was posted to LKML
>> > about 1-2 week ago.
>> >
>> > Subject "[BUGFIX][PATCH] fix lumpy reclaim lru handiling at
>> > isolate_lru_pages v2"
>>
>> It is already committed, but I ran a test on the latest Linus kernel anyway:
>>
>> msgctl11 invoked oom-killer: gfp_mask=0xd0, order=1, oom_adj=0
>> msgctl11 cpuset=/ mems_allowed=0
>> Pid: 20366, comm: msgctl11 Not tainted 2.6.31-rc1-cachefs #144
>> Call Trace:
>>  [<ffffffff810718d2>] ? oom_kill_process.clone.0+0xa9/0x245
>>  [<ffffffff81071b99>] ? __out_of_memory+0x12b/0x142
>>  [<ffffffff81071c1a>] ? out_of_memory+0x6a/0x94
>>  [<ffffffff810742e4>] ? __alloc_pages_nodemask+0x42e/0x51d
>>  [<ffffffff81031416>] ? copy_process+0x95/0x114f
>>  [<ffffffff8107443c>] ? __get_free_pages+0x12/0x4f
>>  [<ffffffff81031439>] ? copy_process+0xb8/0x114f
>>  [<ffffffff8108192e>] ? handle_mm_fault+0x5dd/0x62f
>>  [<ffffffff8103260f>] ? do_fork+0x13f/0x2ba
>>  [<ffffffff81022c22>] ? do_page_fault+0x1f8/0x20d
>>  [<ffffffff8100b0d3>] ? stub_clone+0x13/0x20
>>  [<ffffffff8100ad6b>] ? system_call_fastpath+0x16/0x1b
>> Mem-Info:
>> DMA per-cpu:
>> CPU    0: hi:    0, btch:   1 usd:   0
>> CPU    1: hi:    0, btch:   1 usd:   0
>> DMA32 per-cpu:
>> CPU    0: hi:  186, btch:  31 usd: 159
>> CPU    1: hi:  186, btch:  31 usd:   2
>> Active_anon:70477 active_file:1 inactive_anon:4514
>>  inactive_file:7 unevictable:0 dirty:0 writeback:0 unstable:0
>>  free:1954 slab:42078 mapped:237 pagetables:57791 bounce:0
>
> ~170k pages unreclaimable and ~70k pages unaccounted for.
>
> This does not look like a reclaim problem?

OK. we need learn testcase more.

[read test program source code... ]

this program makes `cat /proc/sys/kernel/msgmni` * 10 processes.
At least, one process creation need one userland stack page (i.e. one anon)
+ one kernel stack page (i.e. one unaccount page) + one pagetable page.

In my 1GB box environment,  default msgmni is 11969.
Oh well, the system physical ram (255744) is less than needed pages (11969 * 3).

In addition, those processes call msgsnd(lrand48() % 99) 1000 times.
msgsnd makes one kmalloc. it mean kernel makes tons random size slab heap and
it become very fragment.

Ummm, I think  this test don't gurantee success on 1GB box.


note: I use distro kernel (Fedora11: kernel-2.6.29+ ).


>> DMA free:3932kB min:60kB low:72kB high:88kB active_anon:236kB inactive_anon:0kB active_file:4kB inactive_file:4kB unevictable:0kB present:15364kB pages_scanned:0 all_unreclaimable? no
>> lowmem_reserve[]: 0 968 968 968
>> DMA32 free:3884kB min:3948kB low:4932kB high:5920kB active_anon:281672kB inactive_anon:18056kB active_file:0kB inactive_file:24kB unevictable:0kB present:992032kB pages_scanned:6 all_unreclaimable? no
>> lowmem_reserve[]: 0 0 0 0
>> DMA: 180*4kB 36*8kB 3*16kB 0*32kB 1*64kB 0*128kB 1*256kB 1*512kB 0*1024kB 1*2048kB 0*4096kB = 3936kB
>> DMA32: 491*4kB 0*8kB 0*16kB 0*32kB 0*64kB 1*128kB 1*256kB 1*512kB 1*1024kB 0*2048kB 0*4096kB = 3884kB
>> 1808 total pagecache pages
>> 0 pages in swap cache
>> Swap cache stats: add 0, delete 0, find 0/0
>> Free swap  = 0kB
>> Total swap = 0kB
>> 255744 pages RAM
>> 5589 pages reserved
>> 249340 pages shared
>> 219039 pages non-shared
>> Out of memory: kill process 11471 (msgctl11) score 112393 or a child
>> Killed process 12318 (msgctl11)
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@xxxxxxxxxx  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/