From: Christoph Hellwig <hch@lst.de>
To: Christian Brauner <brauner@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>,
syzbot <syzbot+2faac0423fdc9692822b@syzkaller.appspotmail.com>,
jack@suse.cz, linux-fsdevel@vger.kernel.org,
linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com,
viro@zeniv.linux.org.uk
Subject: Re: [syzbot] [fs?] KASAN: slab-use-after-free Read in test_bdev_super_fc
Date: Fri, 4 Aug 2023 16:02:01 +0200 [thread overview]
Message-ID: <20230804140201.GA27600@lst.de> (raw)
In-Reply-To: <20230804-abstieg-behilflich-eda2ce9c2c0f@brauner>
On Fri, Aug 04, 2023 at 03:20:45PM +0200, Christian Brauner wrote:
> On Fri, Aug 04, 2023 at 12:14:08PM +0200, Christoph Hellwig wrote:
> > FYI, I can reproduce this trivially locally, but even after spending a
> > significant time with the trace I'm still puzzled at what is going
> > on. I've started trying to make sense of the lockdep report about
> > returning to userspace with s_umount held, originall locked in
> > get_tree_bdev and am still missing how it could happen.
>
> So in the old scheme:
>
> s = alloc_super()
> -> down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);
>
> and assume you're not finding an old one immediately afterwards you'd
>
> -> spin_lock(&sb_lock)
>
> static int set_bdev_super(struct super_block *s, void *data)
> {
> s->s_bdev = data;
> s->s_dev = s->s_bdev->bd_dev;
> s->s_bdi = bdi_get(s->s_bdev->bd_disk->bdi);
>
> if (bdev_stable_writes(s->s_bdev))
> s->s_iflags |= SB_I_STABLE_WRITES;
> return 0;
> }
>
> -> spin_unlock(&sb_lock)
>
> in the new scheme you're doing:
>
> s = alloc_super()
> -> down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);
>
> and assume you're not finding an old one immediately afterwards you'd
>
> up_write(&s->s_umount);
>
> error = setup_bdev_super(s, fc->sb_flags, fc);
> -> spin_lock(&sb_lock);
> sb->s_bdev = bdev;
> sb->s_bdi = bdi_get(bdev->bd_disk->bdi);
> if (bdev_stable_writes(bdev))
> sb->s_iflags |= SB_I_STABLE_WRITES;
> -> spin_unlock(&sb_lock);
>
> down_write(&s->s_umount);
>
> Which looks like the lock ordering here is changed?
Yes, that none only should be safe, but more importantly should not
lead to a return to userspace with s_umount held.
Anyway, debugging a regression in mainline right now so I'm taking a
break from this one.
next prev parent reply other threads:[~2023-08-04 14:02 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-03 22:14 [syzbot] [fs?] KASAN: slab-use-after-free Read in test_bdev_super_fc syzbot
2023-08-04 10:14 ` Christoph Hellwig
2023-08-04 13:20 ` Christian Brauner
2023-08-04 14:02 ` Christoph Hellwig [this message]
2023-08-04 14:36 ` Christian Brauner
2023-08-04 14:43 ` Christoph Hellwig
2023-08-04 14:49 ` Christian Brauner
2023-08-04 15:29 ` Christian Brauner
2023-08-05 8:57 ` Christoph Hellwig
[not found] ` <20230805084316.1699-1-hdanton@sina.com>
2023-08-05 8:48 ` Christoph Hellwig
[not found] <20230804125406.1583-1-hdanton@sina.com>
2023-08-04 19:13 ` syzbot
[not found] <20230804233428.1642-1-hdanton@sina.com>
2023-08-05 0:11 ` syzbot
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=20230804140201.GA27600@lst.de \
--to=hch@lst.de \
--cc=brauner@kernel.org \
--cc=jack@suse.cz \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=syzbot+2faac0423fdc9692822b@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
--cc=viro@zeniv.linux.org.uk \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).