From: Hugh Dickins <hughd@google.com>
To: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>, Hugh Dickins <hughd@google.com>,
Naresh Kamboju <naresh.kamboju@linaro.org>,
Chris Down <chris@chrisdown.name>,
Yafang Shao <laoar.shao@gmail.com>,
Anders Roxell <anders.roxell@linaro.org>,
"Linux F2FS DEV,
Mailing List" <linux-f2fs-devel@lists.sourceforge.net>,
linux-ext4 <linux-ext4@vger.kernel.org>,
linux-block <linux-block@vger.kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
open list <linux-kernel@vger.kernel.org>,
Linux-Next Mailing List <linux-next@vger.kernel.org>,
linux-mm <linux-mm@kvack.org>, Arnd Bergmann <arnd@arndb.de>,
Andreas Dilger <adilger.kernel@dilger.ca>,
Jaegeuk Kim <jaegeuk@kernel.org>, Theodore Ts'o <tytso@mit.edu>,
Chao Yu <chao@kernel.org>, Andrea Arcangeli <aarcange@redhat.com>,
Matthew Wilcox <willy@infradead.org>,
Chao Yu <yuchao0@huawei.com>,
lkft-triage@lists.linaro.org, Roman Gushchin <guro@fb.com>,
Cgroups <cgroups@vger.kernel.org>
Subject: Re: mm: mkfs.ext4 invoked oom-killer on i386 - pagecache_get_page
Date: Thu, 21 May 2020 13:06:28 -0700 (PDT) [thread overview]
Message-ID: <alpine.LSU.2.11.2005211250130.1158@eggly.anvils> (raw)
In-Reply-To: <20200521191746.GB815980@cmpxchg.org>
On Thu, 21 May 2020, Johannes Weiner wrote:
>
> Very much appreciate you guys tracking it down so quickly. Sorry about
> the breakage.
>
> I think mem_cgroup_disabled() checks are pretty good markers of public
> entry points to the memcg API, so I'd prefer that even if a bit more
> verbose. What do you think?
An explicit mem_cgroup_disabled() check would be fine, but I must admit,
the patch below is rather too verbose for my own taste. Your call.
>
> ---
> From cd373ec232942a9bc43ee5e7d2171352019a58fb Mon Sep 17 00:00:00 2001
> From: Hugh Dickins <hughd@google.com>
> Date: Thu, 21 May 2020 14:58:36 -0400
> Subject: [PATCH] mm: memcontrol: prepare swap controller setup for integration
> fix
>
> Fix crash with cgroup_disable=memory:
>
> > > > > + mkfs -t ext4 /dev/disk/by-id/ata-TOSHIBA_MG04ACA100N_Y8NRK0BPF6XF
> > > > > mke2fs 1.43.8 (1-Jan-2018)
> > > > > Creating filesystem with 244190646 4k blocks and 61054976 inodes
> > > > > Filesystem UUID: 3bb1a285-2cb4-44b4-b6e8-62548f3ac620
> > > > > Superblock backups stored on blocks:
> > > > > 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
> > > > > 4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
> > > > > 102400000, 214990848
> > > > > Allocating group tables: 0/7453 done
> > > > > Writing inode tables: 0/7453 done
> > > > > Creating journal (262144 blocks): [ 35.502102] BUG: kernel NULL
> > > > > pointer dereference, address: 000000c8
> > > > > [ 35.508372] #PF: supervisor read access in kernel mode
> > > > > [ 35.513506] #PF: error_code(0x0000) - not-present page
> > > > > [ 35.518638] *pde = 00000000
> > > > > [ 35.521514] Oops: 0000 [#1] SMP
> > > > > [ 35.524652] CPU: 0 PID: 145 Comm: kswapd0 Not tainted
> > > > > 5.7.0-rc6-next-20200519+ #1
> > > > > [ 35.532121] Hardware name: Supermicro SYS-5019S-ML/X11SSH-F, BIOS
> > > > > 2.2 05/23/2018
> > > > > [ 35.539507] EIP: mem_cgroup_get_nr_swap_pages+0x28/0x60
>
> do_memsw_account() used to be automatically false when the cgroup
> controller was disabled. Now that it's replaced by
> cgroup_memory_noswap, for which this isn't true, make the
> mem_cgroup_disabled() checks explicit in the swap control API.
>
> [hannes@cmpxchg.org: use mem_cgroup_disabled() in all API functions]
> Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> Debugged-by: Hugh Dickins <hughd@google.com>
> Debugged-by: Michal Hocko <mhocko@kernel.org>
> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
> ---
> mm/memcontrol.c | 47 +++++++++++++++++++++++++++++++++++++++++------
> 1 file changed, 41 insertions(+), 6 deletions(-)
I'm certainly not against a mem_cgroup_disabled() check in the only
place that's been observed to need it, as a fixup to merge into your
original patch; but this seems rather an over-reaction - and I'm a
little surprised that setting mem_cgroup_disabled() doesn't just
force cgroup_memory_noswap, saving repetitious checks elsewhere
(perhaps there's a difficulty in that, I haven't looked).
Historically, I think we've added mem_cgroup_disabled() checks
(accessing a cacheline we'd rather avoid) where they're necessary,
rather than at every "interface".
And you seem to be in a very "goto out" mood today - we all have
our "goto out" days, alternating with our "return 0" days :)
Hugh
>
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 3e000a316b59..850bca380562 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -6811,6 +6811,9 @@ void mem_cgroup_swapout(struct page *page, swp_entry_t entry)
> VM_BUG_ON_PAGE(PageLRU(page), page);
> VM_BUG_ON_PAGE(page_count(page), page);
>
> + if (mem_cgroup_disabled())
> + return;
> +
> if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
> return;
>
> @@ -6876,6 +6879,10 @@ int mem_cgroup_try_charge_swap(struct page *page, swp_entry_t entry)
> struct mem_cgroup *memcg;
> unsigned short oldid;
>
> + if (mem_cgroup_disabled())
> + return 0;
> +
> + /* Only cgroup2 has swap.max */
> if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> return 0;
>
> @@ -6920,6 +6927,9 @@ void mem_cgroup_uncharge_swap(swp_entry_t entry, unsigned int nr_pages)
> struct mem_cgroup *memcg;
> unsigned short id;
>
> + if (mem_cgroup_disabled())
> + return;
> +
> id = swap_cgroup_record(entry, 0, nr_pages);
> rcu_read_lock();
> memcg = mem_cgroup_from_id(id);
> @@ -6940,12 +6950,25 @@ long mem_cgroup_get_nr_swap_pages(struct mem_cgroup *memcg)
> {
> long nr_swap_pages = get_nr_swap_pages();
>
> - if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
> - return nr_swap_pages;
> + if (mem_cgroup_disabled())
> + goto out;
> +
> + /* Swap control disabled */
> + if (cgroup_memory_noswap)
> + goto out;
> +
> + /*
> + * Only cgroup2 has swap.max, cgroup1 does mem+sw accounting,
> + * which does not place restrictions specifically on swap.
> + */
> + if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> + goto out;
> +
> for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg))
> nr_swap_pages = min_t(long, nr_swap_pages,
> READ_ONCE(memcg->swap.max) -
> page_counter_read(&memcg->swap));
> +out:
> return nr_swap_pages;
> }
>
> @@ -6957,18 +6980,30 @@ bool mem_cgroup_swap_full(struct page *page)
>
> if (vm_swap_full())
> return true;
> - if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
> - return false;
> +
> + if (mem_cgroup_disabled())
> + goto out;
> +
> + /* Swap control disabled */
> + if (cgroup_memory_noswap)
> + goto out;
> +
> + /*
> + * Only cgroup2 has swap.max, cgroup1 does mem+sw accounting,
> + * which does not place restrictions specifically on swap.
> + */
> + if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> + goto out;
>
> memcg = page->mem_cgroup;
> if (!memcg)
> - return false;
> + goto out;
>
> for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg))
> if (page_counter_read(&memcg->swap) * 2 >=
> READ_ONCE(memcg->swap.max))
> return true;
> -
> +out:
> return false;
> }
>
> --
> 2.26.2
next prev parent reply other threads:[~2020-05-21 20:06 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-01 12:38 mm: mkfs.ext4 invoked oom-killer on i386 - pagecache_get_page Naresh Kamboju
2020-05-01 20:58 ` Andrew Morton
2020-05-18 14:10 ` Naresh Kamboju
2020-05-19 7:52 ` Michal Hocko
2020-05-19 8:11 ` Arnd Bergmann
2020-05-19 8:45 ` Michal Hocko
2020-05-20 11:56 ` Naresh Kamboju
2020-05-20 17:59 ` Naresh Kamboju
2020-05-20 19:09 ` Chris Down
2020-05-21 9:22 ` Naresh Kamboju
2020-05-21 9:35 ` Arnd Bergmann
2020-05-21 9:55 ` Michal Hocko
2020-05-21 10:41 ` Naresh Kamboju
2020-05-21 10:58 ` Michal Hocko
2020-05-21 12:24 ` Hugh Dickins
2020-05-21 12:44 ` Michal Hocko
2020-05-21 19:17 ` Johannes Weiner
2020-05-21 20:06 ` Hugh Dickins [this message]
2020-05-21 21:58 ` Johannes Weiner
2020-05-21 23:35 ` Hugh Dickins
2020-05-28 14:59 ` Michal Hocko
2020-05-21 16:34 ` Michal Hocko
2020-05-21 19:00 ` Naresh Kamboju
2020-05-21 20:53 ` Naresh Kamboju
2020-05-28 15:03 ` Michal Hocko
2020-05-28 16:17 ` Naresh Kamboju
2020-05-28 16:41 ` Chris Down
2020-05-29 1:50 ` Yafang Shao
2020-05-29 1:56 ` Chris Down
2020-05-29 9:49 ` Michal Hocko
2020-06-11 9:55 ` Michal Hocko
2020-06-12 9:43 ` Naresh Kamboju
2020-06-12 12:09 ` Michal Hocko
2020-06-17 13:37 ` Naresh Kamboju
2020-06-17 13:57 ` Chris Down
2020-06-17 14:11 ` Michal Hocko
2020-06-17 15:53 ` Naresh Kamboju
2020-06-17 16:06 ` Michal Hocko
2020-06-17 20:13 ` Naresh Kamboju
2020-06-17 21:09 ` Chris Down
2020-06-18 1:43 ` Yafang Shao
2020-06-18 12:37 ` Chris Down
2020-06-18 12:41 ` Michal Hocko
2020-06-18 12:49 ` Chris Down
2020-06-18 14:59 ` Yafang Shao
2020-06-17 13:59 ` Michal Hocko
2020-06-17 14:08 ` Chris Down
2020-05-21 2:39 ` Yafang Shao
2020-05-21 8:58 ` Naresh Kamboju
2020-05-21 9:47 ` Yafang Shao
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=alpine.LSU.2.11.2005211250130.1158@eggly.anvils \
--to=hughd@google.com \
--cc=aarcange@redhat.com \
--cc=adilger.kernel@dilger.ca \
--cc=akpm@linux-foundation.org \
--cc=anders.roxell@linaro.org \
--cc=arnd@arndb.de \
--cc=cgroups@vger.kernel.org \
--cc=chao@kernel.org \
--cc=chris@chrisdown.name \
--cc=guro@fb.com \
--cc=hannes@cmpxchg.org \
--cc=jaegeuk@kernel.org \
--cc=laoar.shao@gmail.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-f2fs-devel@lists.sourceforge.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-next@vger.kernel.org \
--cc=lkft-triage@lists.linaro.org \
--cc=mhocko@kernel.org \
--cc=naresh.kamboju@linaro.org \
--cc=tytso@mit.edu \
--cc=willy@infradead.org \
--cc=yuchao0@huawei.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).