linux-kernel - Re: Found the commit that causes the OOMs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <28c262360906280636l93130ffk14086314e2a6dcb7@mail.gmail.com>
Date:	Sun, 28 Jun 2009 22:36:49 +0900
From:	Minchan Kim <minchan.kim@...il.com>
To:	Wu Fengguang <fengguang.wu@...el.com>
Cc:	Johannes Weiner <hannes@...xchg.org>,
	David Howells <dhowells@...hat.com>,
	"riel@...hat.com" <riel@...hat.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Christoph Lameter <cl@...ux-foundation.org>,
	KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>,
	"peterz@...radead.org" <peterz@...radead.org>,
	"tytso@....edu" <tytso@....edu>,
	"linux-mm@...ck.org" <linux-mm@...ck.org>,
	"elladan@...imo.com" <elladan@...imo.com>,
	"npiggin@...e.de" <npiggin@...e.de>,
	"Barnes, Jesse" <jesse.barnes@...el.com>
Subject: Re: Found the commit that causes the OOMs

On Sun, Jun 28, 2009 at 10:30 PM, Minchan Kim<minchan.kim@...il.com> wrote:
> HI, Wu.
>
> On Sun, Jun 28, 2009 at 8:32 PM, Wu Fengguang<fengguang.wu@...el.com> wrote:
>> On Sat, Jun 27, 2009 at 08:54:12PM +0800, Johannes Weiner wrote:
>>> On Sat, Jun 27, 2009 at 08:12:49AM +0100, David Howells wrote:
>>> >
>>> > I've managed to bisect things to find the commit that causes the OOMs.  It's:
>>> >
>>> >     commit 69c854817566db82c362797b4a6521d0b00fe1d8
>>> >     Author: MinChan Kim <minchan.kim@...il.com>
>>> >     Date:   Tue Jun 16 15:32:44 2009 -0700
>>> >
>>> >         vmscan: prevent shrinking of active anon lru list in case of no swap space V3
>>> >
>>> >         shrink_zone() can deactivate active anon pages even if we don't have a
>>> >         swap device.  Many embedded products don't have a swap device.  So the
>>> >         deactivation of anon pages is unnecessary.
>>> >
>>> >         This patch prevents unnecessary deactivation of anon lru pages.  But, it
>>> >         don't prevent aging of anon pages to swap out.
>>> >
>>> >         Signed-off-by: Minchan Kim <minchan.kim@...il.com>
>>> >         Acked-by: KOSAKI Motohiro <kosaki.motohiro@...fujitsu.com>
>>> >         Cc: Johannes Weiner <hannes@...xchg.org>
>>> >         Acked-by: Rik van Riel <riel@...hat.com>
>>> >         Signed-off-by: Andrew Morton <akpm@...ux-foundation.org>
>>> >         Signed-off-by: Linus Torvalds <torvalds@...ux-foundation.org>
>>> >
>>> > This exhibits the problem.  The previous commit:
>>> >
>>> >     commit 35282a2de4e5e4e173ab61aa9d7015886021a821
>>> >     Author: Brice Goglin <Brice.Goglin@...-lyon.org>
>>> >     Date:   Tue Jun 16 15:32:43 2009 -0700
>>> >
>>> >         migration: only migrate_prep() once per move_pages()
>>> >
>>> > survives 16 iterations of the LTP syscall testsuite without exhibiting the
>>> > problem.
>>>
>>> Here is the patch in question:
>>>
>>> diff --git a/mm/vmscan.c b/mm/vmscan.c
>>> index 7592d8e..879d034 100644
>>> --- a/mm/vmscan.c
>>> +++ b/mm/vmscan.c
>>> @@ -1570,7 +1570,7 @@ static void shrink_zone(int priority, struct zone *zone,
>>>        * Even if we did not try to evict anon pages at all, we want to
>>>        * rebalance the anon lru active/inactive ratio.
>>>        */
>>> -     if (inactive_anon_is_low(zone, sc))
>>> +     if (inactive_anon_is_low(zone, sc) && nr_swap_pages > 0)
>>>               shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0);
>>>
>>>       throttle_vm_writeout(sc->gfp_mask);
>>>
>>> When this was discussed, I think we missed that nr_swap_pages can
>>> actually get zero on swap systems as well and this should have been
>>> total_swap_pages - otherwise we also stop balancing the two anon lists
>>> when swap is _full_ which was not the intention of this change at all.
>>
>> Exactly. In Jesse's OOM case, the swap is exhausted.
>> total_swap_pages is the better choice in this situation.
>>
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426766] Active_anon:290797 active_file:28 inactive_anon:97034
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426767]  inactive_file:61 unevictable:11322 dirty:0 writeback:0 unstable:0
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426768]  free:3341 slab:13776 mapped:5880 pagetables:6851 bounce:0
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426772] DMA free:7776kB min:40kB low:48kB high:60kB active_anon:556kB inactive_anon:524kB
>> +active_file:16kB inactive_file:0kB unevictable:0kB present:15340kB pages_scanned:30 all_unreclaimable? no
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426775] lowmem_reserve[]: 0 1935 1935 1935
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426781] DMA32 free:5588kB min:5608kB low:7008kB high:8412kB active_anon:1162632kB
>> +inactive_anon:387612kB active_file:96kB inactive_file:256kB unevictable:45288kB present:1982128kB pages_scanned:980
>> +all_unreclaimable? no
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426784] lowmem_reserve[]: 0 0 0 0
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426787] DMA: 64*4kB 77*8kB 45*16kB 18*32kB 4*64kB 2*128kB 2*256kB 3*512kB 1*1024kB
>> +1*2048kB 0*4096kB = 7800kB
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426796] DMA32: 871*4kB 149*8kB 1*16kB 2*32kB 1*64kB 0*128kB 1*256kB 1*512kB 0*1024kB
>> +0*2048kB 0*4096kB = 5588kB
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426804] 151250 total pagecache pages
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426806] 18973 pages in swap cache
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426808] Swap cache stats: add 610640, delete 591667, find 144356/181468
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426810] Free swap  = 0kB
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.426811] Total swap = 979956kB
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.434828] 507136 pages RAM
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.434831] 23325 pages reserved
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.434832] 190892 pages shared
>> Jun 18 07:44:53 jbarnes-g45 kernel: [64377.434833] 248816 pages non-shared
>>
>>
>> In David's OOM case, there are two symptoms:
>> 1) 70000 unaccounted/leaked pages as found by Andrew
>>   (plus rather big number of PG_buddy and pagetable pages)
>> 2) almost zero active_file/inactive_file; small inactive_anon;
>>   many slab and active_anon pages.
>>
>> In the situation of (2), the slab cache is _under_ scanned. So David
>> got OOM when vmscan should have squeezed some free pages from the slab
>> cache. Which is one important side effect of MinChan's patch?
>
> My patch's side effect is (2).
>
> My guessing is following as.
>
> 1. The number of page scanned in shrink_slab is increased in shrink_page_list.
> And it is doubled for mapped page or swapcache.
> 2. shrink_page_list is called by shrink_inactive_list
> 3. shrink_inactive_list is called by shrink_list
>
> Look at the shrink_list.
> If inactive lru list is low, it always call shrink_active_list not
> shrink_inactive_list in case of anon.

I missed most important point.
My patch's side effect is that it keeps inactive anon's lru low.
So I think it is caused by my patch's side effect.

> It means it doesn't increased sc->nr_scanned.
> Then shrink_slab can't shrink enough slab pages.
> So, David OOM have a lot of slab pages and active anon pages.
>
> Does it make sense ?
> If it make sense, we have to change shrink_slab's pressure method.
> What do you think ?
>
>
> --
> Kinds regards,
> Minchan Kim
>



-- 
Kinds regards,
Minchan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/