From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail191.messagelabs.com (mail191.messagelabs.com [216.82.242.19]) by kanga.kvack.org (Postfix) with SMTP id E2BC36B007B for ; Sat, 18 Sep 2010 11:21:30 -0400 (EDT) Received: by pzk26 with SMTP id 26so1091303pzk.14 for ; Sat, 18 Sep 2010 08:21:29 -0700 (PDT) Date: Sun, 19 Sep 2010 00:21:20 +0900 From: Minchan Kim Subject: Re: cgroup oom regression introduced by 6a5ce1b94e1e5979f8db579f77d6e08a5f44c13b Message-ID: <20100918152120.GA21343@barrios-desktop> References: <1296415999.1298271284814035815.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> <290491919.1298351284814354705.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <290491919.1298351284814354705.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> Sender: owner-linux-mm@kvack.org To: caiqian@redhat.com Cc: linux-mm , KOSAKI Motohiro , "M. Vefa Bicakci" , Johannes Weiner , Rik van Riel , stable@kernel.org, akpm@linux-foundation.org List-ID: On Sat, Sep 18, 2010 at 08:52:34AM -0400, caiqian@redhat.com wrote: > This test hung the kernel without triggering oom. > # mount -t cgroup -o memory none /cgroup/memory/ > # mkdir /cgroup/memory/A > # echo $$ >/cgroup/memory/A/tasks > # echo 4096M >/cgroup/memory/A/memory.limit_in_bytes > # echo 4096M >/cgroup/memory/A/memory.memsw.limit_in_bytes > # use malloc to allocate more than 4G memory. > > Sometimes, this had been thrown out of console, > localhost.localdomain login: INFO: task sm1:5065 blocked for more than 120 seconds. > "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. > sm1 D 00000000fffca130 0 5065 5051 0x00000080 > ffff880c5f419c38 0000000000000086 ffff880c5f419bc8 ffffffff81034ca8 > ffff880100000000 0000000000015440 ffff880c608ab4e0 0000000000015440 > ffff880c608aba40 ffff880c5f419fd8 ffff880c608aba48 ffff880c5f419fd8 > Call Trace: > [] ? pvclock_clocksource_read+0x58/0xd0 > [] ? sync_page+0x0/0x50 > [] io_schedule+0x73/0xc0 > [] sync_page+0x3d/0x50 > [] __wait_on_bit_lock+0x5a/0xc0 > [] __lock_page+0x67/0x70 > [] ? wake_bit_function+0x0/0x50 > [] ? find_get_page+0x1e/0xa0 > [] filemap_fault+0x33c/0x450 > [] __do_fault+0x54/0x550 > [] ? __mem_cgroup_commit_charge+0x5a/0xa0 > [] handle_mm_fault+0x1c2/0xc70 > [] do_page_fault+0x11c/0x320 > [] page_fault+0x25/0x30 > > Reverted the following commit from mmotm tree made the problem go away. > commit 6a5ce1b94e1e5979f8db579f77d6e08a5f44c13b > Author: Minchan Kim > Date: Thu Sep 16 01:17:26 2010 +0200 > > M. Vefa Bicakci reported 2.6.35 kernel hang up when hibernation on his > 32bit 3GB mem machine. > (https://bugzilla.kernel.org/show_bug.cgi?id=16771). Also he bisected > the regression to > > commit bb21c7ce18eff8e6e7877ca1d06c6db719376e3c > Author: KOSAKI Motohiro > Date: Fri Jun 4 14:15:05 2010 -0700 > > vmscan: fix do_try_to_free_pages() return value when priority==0 reclaim failure > > At first impression, this seemed very strange because the above commit > only chenged function return value and hibernate_preallocate_memory() > ignore return value of shrink_all_memory(). But it's related. > > Now, page allocation from hibernation code may enter infinite loop if the > system has highmem. The reasons are that vmscan don't care enough OOM > case when oom_killer_disabled. > > The problem sequence is following as. > > 1. hibernation > 2. oom_disable > 3. alloc_pages > 4. do_try_to_free_pages > if (scanning_global_lru(sc) && !all_unreclaimable) > return 1; > > If kswapd is not freozen, it would set zone->all_unreclaimable to 1 and > then shrink_zones maybe return true(ie, all_unreclaimable is true). So at > last, alloc_pages could go to _nopage_. If it is, it should have no > problem. > > This patch adds all_unreclaimable check to protect in direct reclaim path, > too. It can care of hibernation OOM case and help bailout > all_unreclaimable case slightly. > > Signed-off-by: KOSAKI Motohiro > Signed-off-by: Minchan Kim > Reported-by: M. Vefa Bicakci > Acked-by: Rafael J. Wysocki > Reviewed-by: Johannes Weiner > Cc: Rik van Riel > Cc: > Signed-off-by: Andrew Morton > > diff --git a/mm/vmscan.c b/mm/vmscan.c > index 225a759..f56a8c3 100644 > --- a/mm/vmscan.c > +++ b/mm/vmscan.c > @@ -1804,12 +1804,11 @@ static void shrink_zone(int priority, struct zone *zone, > * If a zone is deemed to be full of pinned pages then just give it a light > * scan then give up on it. > */ > -static bool shrink_zones(int priority, struct zonelist *zonelist, > +static void shrink_zones(int priority, struct zonelist *zonelist, > struct scan_control *sc) > { > struct zoneref *z; > struct zone *zone; > - bool all_unreclaimable = true; > > for_each_zone_zonelist_nodemask(zone, z, zonelist, > gfp_zone(sc->gfp_mask), sc->nodemask) { > @@ -1827,8 +1826,36 @@ static bool shrink_zones(int priority, struct zonelist *zonelist, > } > > shrink_zone(priority, zone, sc); > - all_unreclaimable = false; > } > +} > + > +static inline bool zone_reclaimable(struct zone *zone) > +{ > + return zone->pages_scanned < zone_reclaimable_pages(zone) * 6; > +} > + > +static inline bool all_unreclaimable(struct zonelist *zonelist, > + struct scan_control *sc) > +{ > + struct zoneref *z; > + struct zone *zone; > + bool all_unreclaimable = true; > + > + if (!scanning_global_lru(sc)) > + return false; > + > + for_each_zone_zonelist_nodemask(zone, z, zonelist, > + gfp_zone(sc->gfp_mask), sc->nodemask) { > + if (!populated_zone(zone)) > + continue; > + if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) > + continue; > + if (zone_reclaimable(zone)) { > + all_unreclaimable = false; > + break; > + } > + } > + > return all_unreclaimable; > } > > @@ -1852,7 +1879,6 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, > struct scan_control *sc) > { > int priority; > - bool all_unreclaimable; > unsigned long total_scanned = 0; > struct reclaim_state *reclaim_state = current->reclaim_state; > struct zoneref *z; > @@ -1869,7 +1895,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, > sc->nr_scanned = 0; > if (!priority) > disable_swap_token(); > - all_unreclaimable = shrink_zones(priority, zonelist, sc); > + shrink_zones(priority, zonelist, sc); > /* > * Don't shrink slabs when reclaiming memory from > * over limit cgroups > @@ -1931,7 +1957,7 @@ out: > return sc->nr_reclaimed; > > /* top priority shrink_zones still had more to do? don't OOM, then */ > - if (scanning_global_lru(sc) && !all_unreclaimable) > + if (!all_unreclaimable(zonelist, sc)) > return 1; > > return 0; > @@ -2197,8 +2223,7 @@ loop_again: > total_scanned += sc.nr_scanned; > if (zone->all_unreclaimable) > continue; > - if (nr_slab == 0 && > - zone->pages_scanned >= (zone_reclaimable_pages(zone) * 6)) > + if (nr_slab == 0 && !zone_reclaimable(zone)) > zone->all_unreclaimable = 1; > /* > * If we've done a decent amount of scanning and Thanks for the reporting. Could you test below patch?