Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
From: Roman Gushchin <guro@fb.com>
To: Shakeel Butt <shakeelb@google.com>
Cc: Richard Palethorpe <rpalethorpe@suse.com>,
	Linux MM <linux-mm@kvack.org>,
	LKML <linux-kernel@vger.kernel.org>,
	LTP List <ltp@lists.linux.it>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Christoph Lameter <cl@linux.com>,
	Michal Hocko <mhocko@kernel.org>, Tejun Heo <tj@kernel.org>,
	Vlastimil Babka <vbabka@suse.cz>
Subject: Re: [PATCH v4] mm: memcg/slab: Stop reparented obj_cgroups from charging root
Date: Thu, 22 Oct 2020 10:25:08 -0700
Message-ID: <20201022172508.GF300658@carbon.dhcp.thefacebook.com> (raw)
In-Reply-To: <CALvZod4u79DSwyM=Kg8g5tR1L5eomHgUkaJmp6s9D3jC7OnN0A@mail.gmail.com>

On Thu, Oct 22, 2020 at 09:37:02AM -0700, Shakeel Butt wrote:
> On Thu, Oct 22, 2020 at 5:29 AM Richard Palethorpe <rpalethorpe@suse.com> wrote:
> >
> > When use_hierarchy=1, SLAB objects which outlive their descendant
> > memcg are moved to their parent memcg where they may be uncharged
> > because charges are made recursively from leaf to root nodes.
> >
> > However when use_hierarchy=0, they are reparented directly to root and
> > charging is not made recursively. Therefor uncharging will result in a
> > counter underflow on the root memcg, but no other ancestors.
> >
> > To prevent this, we check whether we are about to uncharge the root
> > memcg and whether use_hierarchy=0. If this is the case then we skip
> > uncharging. The root memcg does not have its own objcg, so any objcg
> > which is uncharging to it must have been reparented.
> >
> > Note that on the default hierarchy (CGroupV2 now) root always has
> > use_hierarchy=1. So this only effects CGroupV1.
> >
> > The warning can be, unreliably, reproduced with the LTP test
> > madvise06 if the entire patch series
> > https://lore.kernel.org/linux-mm/20200623174037.3951353-1-guro@fb.com/
> > is present. Although the listed commit in 'fixes' appears to introduce
> > the bug, I can not reproduce it with just that commit and bisecting
> > runs into other bugs.
> >
> > [   12.029417] WARNING: CPU: 2 PID: 21 at mm/page_counter.c:57 page_counter_uncharge (mm/page_counter.c:57 mm/page_counter.c:50 mm/page_counter.c:156)
> > [   12.029539] Modules linked in:
> > [   12.029611] CPU: 2 PID: 21 Comm: ksoftirqd/2 Not tainted 5.9.0-rc7-22-default #76
> > [   12.029729] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-48-gd9c812d-rebuilt.opensuse.org 04/01/2014
> > [   12.029908] RIP: 0010:page_counter_uncharge (mm/page_counter.c:57 mm/page_counter.c:50 mm/page_counter.c:156)
> > [ 12.029991] Code: 0f c1 45 00 4c 29 e0 48 89 ef 48 89 c3 48 89 c6 e8 2a fe ff ff 48 85 db 78 10 48 8b 6d 28 48 85 ed 75 d8 5b 5d 41 5c 41 5d c3 <0f> 0b eb ec 90 e8 db 47 36 27 48 8b 17 48 39 d6 72 41 41 54 49 89
> > [   12.030258] RSP: 0018:ffffa5d8000efd08 EFLAGS: 00010086
> > [   12.030344] RAX: ffffffffffffffff RBX: ffffffffffffffff RCX: 0000000000000009
> > [   12.030455] RDX: 000000000000000b RSI: ffffffffffffffff RDI: ffff8ef8c7d2b248
> > [   12.030561] RBP: ffff8ef8c7d2b248 R08: ffff8ef8c78b19c8 R09: 0000000000000001
> > [   12.030672] R10: 0000000000000000 R11: ffff8ef8c780e0d0 R12: 0000000000000001
> > [   12.030784] R13: ffffffffffffffff R14: ffff8ef9478b19c8 R15: 0000000000000000
> > [   12.030895] FS:  0000000000000000(0000) GS:ffff8ef8fbc80000(0000) knlGS:0000000000000000
> > [   12.031017] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [   12.031104] CR2: 00007f72c0af93ec CR3: 000000005c40a000 CR4: 00000000000006e0
> > [   12.031209] Call Trace:
> > [   12.031267] __memcg_kmem_uncharge (mm/memcontrol.c:3022)
> > [   12.031470] drain_obj_stock (./include/linux/rcupdate.h:689 mm/memcontrol.c:3114)
> > [   12.031594] refill_obj_stock (mm/memcontrol.c:3166)
> > [   12.031733] ? rcu_do_batch (kernel/rcu/tree.c:2438)
> > [   12.032075] memcg_slab_free_hook (./include/linux/mm.h:1294 ./include/linux/mm.h:1441 mm/slab.h:368 mm/slab.h:348)
> > [   12.032339] kmem_cache_free (mm/slub.c:3107 mm/slub.c:3143 mm/slub.c:3158)
> > [   12.032464] rcu_do_batch (kernel/rcu/tree.c:2438)
> > [   12.032567] rcu_core (kernel/rcu/tree_plugin.h:2122 kernel/rcu/tree_plugin.h:2157 kernel/rcu/tree.c:2661)
> > [   12.032664] __do_softirq (./arch/x86/include/asm/jump_label.h:25 ./include/linux/jump_label.h:200 ./include/trace/events/irq.h:142 kernel/softirq.c:299)
> > [   12.032766] run_ksoftirqd (./arch/x86/include/asm/irqflags.h:54 ./arch/x86/include/asm/irqflags.h:94 kernel/softirq.c:653 kernel/softirq.c:644)
> > [   12.032852] smpboot_thread_fn (kernel/smpboot.c:165)
> > [   12.032940] ? smpboot_register_percpu_thread (kernel/smpboot.c:108)
> > [   12.033059] kthread (kernel/kthread.c:292)
> > [   12.033148] ? __kthread_bind_mask (kernel/kthread.c:245)
> > [   12.033269] ret_from_fork (arch/x86/entry/entry_64.S:300)
> > [   12.033357] ---[ end trace 961dbfc01c109d1f ]---
> >
> > [    9.841552] ------------[ cut here ]------------
> > [    9.841788] WARNING: CPU: 0 PID: 12 at mm/page_counter.c:57 page_counter_uncharge (mm/page_counter.c:57 mm/page_counter.c:50 mm/page_counter.c:156)
> > [    9.841982] Modules linked in:
> > [    9.842072] CPU: 0 PID: 12 Comm: kworker/0:1 Not tainted 5.9.0-rc7-22-default #77
> > [    9.842266] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-48-gd9c812d-rebuilt.opensuse.org 04/01/2014
> > [    9.842571] Workqueue: events drain_local_stock
> > [    9.842750] RIP: 0010:page_counter_uncharge (mm/page_counter.c:57 mm/page_counter.c:50 mm/page_counter.c:156)
> > [ 9.842894] Code: 0f c1 45 00 4c 29 e0 48 89 ef 48 89 c3 48 89 c6 e8 2a fe ff ff 48 85 db 78 10 48 8b 6d 28 48 85 ed 75 d8 5b 5d 41 5c 41 5d c3 <0f> 0b eb ec 90 e8 4b f9 88 2a 48 8b 17 48 39 d6 72 41 41 54 49 89
> > [    9.843438] RSP: 0018:ffffb1c18006be28 EFLAGS: 00010086
> > [    9.843585] RAX: ffffffffffffffff RBX: ffffffffffffffff RCX: ffff94803bc2cae0
> > [    9.843806] RDX: 0000000000000001 RSI: ffffffffffffffff RDI: ffff948007d2b248
> > [    9.844026] RBP: ffff948007d2b248 R08: ffff948007c58eb0 R09: ffff948007da05ac
> > [    9.844248] R10: 0000000000000018 R11: 0000000000000018 R12: 0000000000000001
> > [    9.844477] R13: ffffffffffffffff R14: 0000000000000000 R15: ffff94803bc2cac0
> > [    9.844696] FS:  0000000000000000(0000) GS:ffff94803bc00000(0000) knlGS:0000000000000000
> > [    9.844915] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [    9.845096] CR2: 00007f0579ee0384 CR3: 000000002cc0a000 CR4: 00000000000006f0
> > [    9.845319] Call Trace:
> > [    9.845429] __memcg_kmem_uncharge (mm/memcontrol.c:3022)
> > [    9.845582] drain_obj_stock (./include/linux/rcupdate.h:689 mm/memcontrol.c:3114)
> > [    9.845684] drain_local_stock (mm/memcontrol.c:2255)
> > [    9.845789] process_one_work (./arch/x86/include/asm/jump_label.h:25 ./include/linux/jump_label.h:200 ./include/trace/events/workqueue.h:108 kernel/workqueue.c:2274)
> > [    9.845898] worker_thread (./include/linux/list.h:282 kernel/workqueue.c:2416)
> > [    9.846034] ? process_one_work (kernel/workqueue.c:2358)
> > [    9.846162] kthread (kernel/kthread.c:292)
> > [    9.846271] ? __kthread_bind_mask (kernel/kthread.c:245)
> > [    9.846420] ret_from_fork (arch/x86/entry/entry_64.S:300)
> > [    9.846531] ---[ end trace 8b5647c1eba9d18a ]---
> >
> > Reported-by: ltp@lists.linux.it
> > Signed-off-by: Richard Palethorpe <rpalethorpe@suse.com>
> > Acked-by: Roman Gushchin <guro@fb.com>
> > Cc: Johannes Weiner <hannes@cmpxchg.org>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Cc: Shakeel Butt <shakeelb@google.com>
> > Cc: Christoph Lameter <cl@linux.com>
> > Cc: Michal Hocko <mhocko@kernel.org>
> > Cc: Tejun Heo <tj@kernel.org>
> > Cc: Vlastimil Babka <vbabka@suse.cz>
> > Fixes: bf4f059954dc ("mm: memcg/slab: obj_cgroup API")
> 
> This looks fine to me but I would prefer Roman's approach of charging
> to root and for use_hierarchy=0, linking them to root.

+1

I think it's safer because it has no effect on use_hierarchy == true case,
and this is what we should really care about. But the latest Richard's version
looks correct to me, so I'm fine with using it for stable too.

> 
> Since bf4f059954dc ("mm: memcg/slab: obj_cgroup API") is in 5.9, I
> think we can take this patch for 5.9 and 5.10 but keep Roman's cleanup
> for 5.11.
> 
> What does everyone think?

I think we should use the link to the root approach both for stable backports
and for 5.11+, to keep them in sync. The cleanup (always charging the root cgroup)
is not directly related to this problem, and we can keep it for 5.11+ only.

Thanks!


  reply index

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-14 19:07 [RFC PATCH] " Richard Palethorpe
2020-10-14 20:08 ` Roman Gushchin
2020-10-16  5:40   ` Richard Palethorpe
2020-10-16  9:47 ` Michal Koutný
2020-10-16 10:41   ` Richard Palethorpe
2020-10-16 15:05     ` Richard Palethorpe
2020-10-16 17:26       ` Michal Koutný
2020-10-16 14:53   ` Johannes Weiner
2020-10-16 17:02     ` Roman Gushchin
2020-10-16 17:15     ` Michal Koutný
2020-10-19  8:45       ` Richard Palethorpe
2020-10-19  9:58         ` [PATCH v3] " Richard Palethorpe
2020-10-19 16:58           ` Shakeel Butt
2020-10-20  5:52             ` Richard Palethorpe
2020-10-20 13:49               ` Richard Palethorpe
2020-10-20 16:56                 ` Shakeel Butt
2020-10-21 20:32                   ` Roman Gushchin
2020-10-20 17:24               ` Michal Koutný
2020-10-22  7:04                 ` Richard Palethorpe
2020-10-22 12:28                   ` [PATCH v4] " Richard Palethorpe
2020-10-22 16:37                     ` Shakeel Butt
2020-10-22 17:25                       ` Roman Gushchin [this message]
2020-10-22 23:59                         ` Shakeel Butt
2020-10-23  0:40                           ` Roman Gushchin
2020-10-23 15:44                             ` Johannes Weiner
2020-10-23 16:41                             ` Shakeel Butt
2020-10-26  7:32                             ` Richard Palethorpe
2020-10-26 23:14                               ` Roman Gushchin
2020-10-19 22:28       ` [RFC PATCH] " Roman Gushchin
2020-10-20  6:04         ` Richard Palethorpe
2020-10-20 12:02           ` Richard Palethorpe
2020-10-20 14:48         ` Richard Palethorpe
2020-10-20 16:27         ` Michal Koutný
2020-10-20 17:07           ` Roman Gushchin
2020-10-20 18:18             ` Johannes Weiner
2020-10-21 19:33               ` Roman Gushchin
2020-10-23 16:30                 ` Johannes Weiner
2020-11-10  1:27                   ` Roman Gushchin
2020-11-10 15:11                     ` Shakeel Butt
2020-11-10 19:13                       ` Roman Gushchin
2020-11-20 17:46                       ` Michal Koutný
2020-11-03 13:22                 ` Michal Hocko
2020-11-03 21:30                   ` Roman Gushchin
2020-10-20 16:55         ` Shakeel Butt
2020-10-20 17:17           ` Roman Gushchin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20201022172508.GF300658@carbon.dhcp.thefacebook.com \
    --to=guro@fb.com \
    --cc=akpm@linux-foundation.org \
    --cc=cl@linux.com \
    --cc=hannes@cmpxchg.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=ltp@lists.linux.it \
    --cc=mhocko@kernel.org \
    --cc=rpalethorpe@suse.com \
    --cc=shakeelb@google.com \
    --cc=tj@kernel.org \
    --cc=vbabka@suse.cz \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-mm Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-mm/0 linux-mm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-mm linux-mm/ https://lore.kernel.org/linux-mm \
		linux-mm@kvack.org
	public-inbox-index linux-mm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kvack.linux-mm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git