From: Jan Kara <jack@suse.cz>
To: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: Jan Kara <jack@suse.cz>,
jack@suse.com, linux-ext4@vger.kernel.org,
linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com,
tytso@mit.edu, mhocko@suse.cz, linux-mm@kvack.org,
syzbot <syzbot+bfdded10ab7dcd7507ae@syzkaller.appspotmail.com>
Subject: Re: possible deadlock in start_this_handle (2)
Date: Mon, 15 Feb 2021 15:29:35 +0100 [thread overview]
Message-ID: <20210215142935.GB22417@quack2.suse.cz> (raw)
In-Reply-To: <aaee5d61-f988-84c3-4d16-f8b7987f3a83@i-love.sakura.ne.jp>
On Mon 15-02-21 23:06:15, Tetsuo Handa wrote:
> On 2021/02/15 21:45, Jan Kara wrote:
> > On Sat 13-02-21 23:26:37, Tetsuo Handa wrote:
> >> Excuse me, but it seems to me that nothing prevents
> >> ext4_xattr_set_handle() from reaching ext4_xattr_inode_lookup_create()
> >> without memalloc_nofs_save() when hitting ext4_get_nojournal() path.
> >> Will you explain when ext4_get_nojournal() path is executed?
> >
> > That's a good question but sadly I don't think that's it.
> > ext4_get_nojournal() is called when the filesystem is created without a
> > journal. In that case we also don't acquire jbd2_handle lockdep map. In the
> > syzbot report we can see:
>
> Since syzbot can test filesystem images, syzbot might have tested a filesystem
> image created both with and without journal within this boot.
a) I think that syzbot reboots the VM between executing different tests to
get reproducible conditions. But in theory I agree the test may have
contained one image with and one image without a journal.
*but*
b) as I wrote in the email you are replying to, the jbd2_handle key is
private per filesystem. Thus for lockdep to complain about
jbd2_handle->fs_reclaim->jbd2_handle deadlock, those jbd2_handle lockdep
maps must come from the same filesystem.
*and*
c) filesystem without journal doesn't use jbd2_handle lockdep map at all so
for such filesystems lockdep creates no dependency for jbd2_handle map.
Honza
>
> >
> > kswapd0/2246 is trying to acquire lock:
> > ffff888041a988e0 (jbd2_handle){++++}-{0:0}, at: start_this_handle+0xf81/0x1380 fs/jbd2/transaction.c:444
> >
> > but task is already holding lock:
> > ffffffff8be892c0 (fs_reclaim){+.+.}-{0:0}, at: __fs_reclaim_acquire+0x0/0x30 mm/page_alloc.c:5195
> >
> > So this filesystem has very clearly been created with a journal. Also the
> > journal lockdep tracking machinery uses:
>
> While locks held by kswapd0/2246 are fs_reclaim, shrinker_rwsem, &type->s_umount_key#38
> and jbd2_handle, isn't the dependency lockdep considers problematic is
>
> Chain exists of:
> jbd2_handle --> &ei->xattr_sem --> fs_reclaim
>
> Possible unsafe locking scenario:
>
> CPU0 CPU1
> ---- ----
> lock(fs_reclaim);
> lock(&ei->xattr_sem);
> lock(fs_reclaim);
> lock(jbd2_handle);
>
> where CPU0 is kswapd/2246 and CPU1 is the case of ext4_get_nojournal() path?
> If someone has taken jbd2_handle and &ei->xattr_sem in this order, isn't this
> dependency true?
>
> >
> > rwsem_acquire_read(&journal->j_trans_commit_map, 0, 0, _THIS_IP_);
> >
> > so a lockdep key is per-filesystem. Thus it is not possible that lockdep
> > would combine lock dependencies from two different filesystems.
> >
> > But I guess we could narrow the search for this problem by adding WARN_ONs
> > to ext4_xattr_set_handle() and ext4_xattr_inode_lookup_create() like:
> >
> > WARN_ON(ext4_handle_valid(handle) && !(current->flags & PF_MEMALLOC_NOFS));
> >
> > It would narrow down a place in which PF_MEMALLOC_NOFS flag isn't set
> > properly... At least that seems like the most plausible way forward to me.
>
> You can use CONFIG_DEBUG_AID_FOR_SYZBOT for adding such WARN_ONs on linux-next.
>
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
next prev parent reply other threads:[~2021-02-15 14:29 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <000000000000563a0205bafb7970@google.com>
2021-02-11 10:49 ` possible deadlock in start_this_handle (2) Jan Kara
2021-02-11 10:55 ` Michal Hocko
2021-02-11 11:22 ` Dmitry Vyukov
2021-02-11 11:28 ` Dmitry Vyukov
2021-02-11 12:10 ` Jan Kara
2021-02-11 12:34 ` Michal Hocko
2021-02-11 12:57 ` Matthew Wilcox
2021-02-11 13:07 ` Michal Hocko
2021-02-11 13:25 ` Matthew Wilcox
2021-02-11 14:20 ` Michal Hocko
2021-02-11 14:26 ` Matthew Wilcox
2021-02-11 16:41 ` Michal Hocko
2021-02-12 11:18 ` Tetsuo Handa
2021-02-12 12:22 ` Matthew Wilcox
2021-02-12 12:30 ` Michal Hocko
2021-02-12 12:58 ` Tetsuo Handa
2021-02-12 13:12 ` Michal Hocko
2021-02-12 13:34 ` Tetsuo Handa
2021-02-12 15:43 ` Michal Hocko
2021-02-13 10:58 ` Dmitry Vyukov
2021-02-11 13:18 ` Dmitry Vyukov
2021-02-11 11:46 ` Jan Kara
2021-02-13 14:26 ` Tetsuo Handa
2021-02-15 12:45 ` Jan Kara
2021-02-15 14:06 ` Tetsuo Handa
2021-02-15 14:29 ` Jan Kara [this message]
2021-02-19 10:15 ` Tetsuo Handa
2021-02-19 17:22 ` harshad shirwadkar
2021-03-20 10:02 ` Tetsuo Handa
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=20210215142935.GB22417@quack2.suse.cz \
--to=jack@suse.cz \
--cc=jack@suse.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.cz \
--cc=penguin-kernel@i-love.sakura.ne.jp \
--cc=syzbot+bfdded10ab7dcd7507ae@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
--cc=tytso@mit.edu \
/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).