linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Christian Brauner <brauner@kernel.org>
To: Mikulas Patocka <mpatocka@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>,
	Zdenek Kabelac <zkabelac@redhat.com>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	dm-devel@redhat.com, Jan Kara <jack@suse.cz>,
	Christoph Hellwig <hch@lst.de>,
	"Darrick J. Wong" <djwong@kernel.org>
Subject: Re: [PATCH] fix writing to the filesystem after unmount
Date: Wed, 6 Sep 2023 17:58:18 +0200	[thread overview]
Message-ID: <20230906-echtheit-dezent-6f3621821cf2@brauner> (raw)
In-Reply-To: <20230906-aufheben-hagel-9925501b7822@brauner>

On Wed, Sep 06, 2023 at 05:33:32PM +0200, Christian Brauner wrote:
> > Currently, if we freeze a filesystem with "fsfreeze" and unmount it, the 
> > mount point is removed, but the filesystem stays active and it is leaked. 
> > You can't unfreeze it with "fsfreeze --unfreeze" because the mount point 
> > is gone. (the only way how to recover it is "echo j>/proc/sysrq-trigger").
> 
> You can of course always remount and unfreeze it.
> 
> > > IOW, you'd also hang on any umount of a bind-mount. IOW, every
> > > single container making use of this filesystems via bind-mounts would
> > > hang on umount and shutdown.
> > 
> > bind-mount doesn't modify "s->s_writers.frozen", so the patch does nothing 
> > in this case. I tried unmounting bind-mounts and there was no deadlock.
> 
> With your patch what happens if you do the following?
> 
> #!/bin/sh -ex
> modprobe brd rd_size=4194304
> vgcreate vg /dev/ram0
> lvcreate -L 16M -n lv vg
> mkfs.ext4 /dev/vg/lv
> 
> mount -t ext4 /dev/vg/lv /mnt/test
> mount --bind /mnt/test /opt
> mount --make-private /opt
> 
> dmsetup suspend /dev/vg/lv
> (sleep 1; dmsetup resume /dev/vg/lv) &
> 
> umount /opt # I'd expect this to hang
> 
> md5sum /dev/vg/lv
> md5sum /dev/vg/lv
> dmsetup remove_all
> rmmod brd
> 
> > BTW. what do you think that unmount of a frozen filesystem should properly 
> > do? Fail with -EBUSY? Or, unfreeze the filesystem and unmount it? Or 
> > something else?
> 
> In my opinion we should refuse to unmount frozen filesystems and log an
> error that the filesystem is frozen. Waiting forever isn't a good idea
> in my opinion.
> 
> But this is a significant uapi change afaict so this would need to be
> hidden behind a config option, a sysctl, or it would have to be a new
> flag to umount2() MNT_UNFROZEN which would allow an administrator to use
> this flag to not unmount a frozen filesystems.

That's probably too careful. I think we could risk starting to return an
error when trying to unmount a frozen filesystem. And if that causes
regressions we could go and look at another option like MNT_UNFROZEN or
whatever.

  reply	other threads:[~2023-09-06 15:58 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-06 13:26 [PATCH] fix writing to the filesystem after unmount Mikulas Patocka
2023-09-06 14:27 ` Christian Brauner
2023-09-06 15:03   ` Mikulas Patocka
2023-09-06 15:33     ` Christian Brauner
2023-09-06 15:58       ` Christian Brauner [this message]
2023-09-06 16:01       ` Mikulas Patocka
2023-09-06 16:19         ` Christian Brauner
2023-09-06 16:52           ` Mikulas Patocka
2023-09-07  9:44             ` Jan Kara
2023-09-07 10:43               ` Christian Brauner
2023-09-07 12:04                 ` Mikulas Patocka
2023-09-08  7:32                   ` Jan Kara
2023-09-08  9:29                     ` Zdenek Kabelac
2023-09-08 10:20                       ` Jan Kara
2023-09-08 12:02                         ` Christian Brauner
2023-09-08 16:49                           ` John Stoffel
2023-09-09 11:21                         ` Christoph Hellwig
     [not found]                         ` <15c62097-d58f-4e66-bdf5-e0edb1306b2f@redhat.com>
2023-09-08 11:32                           ` Christian Brauner
2023-09-08 12:07                             ` Zdenek Kabelac
2023-09-08 12:34                               ` Christian Brauner
2023-09-12  9:10                           ` Jan Kara
2023-09-08 12:01                     ` Pavel Machek
2023-09-08 11:59               ` Pavel Machek
2023-09-06 17:10         ` Al Viro
2023-09-06 17:08     ` Al Viro
2023-09-06 15:22 ` Darrick J. Wong
2023-09-06 15:38   ` Christian Brauner

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=20230906-echtheit-dezent-6f3621821cf2@brauner \
    --to=brauner@kernel.org \
    --cc=djwong@kernel.org \
    --cc=dm-devel@redhat.com \
    --cc=hch@lst.de \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mpatocka@redhat.com \
    --cc=viro@zeniv.linux.org.uk \
    --cc=zkabelac@redhat.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).