From: Al Viro <viro@ZenIV.linux.org.uk>
To: Alan Jenkins <alan.christopher.jenkins@gmail.com>
Cc: David Howells <dhowells@redhat.com>,
torvalds@linux-foundation.org, ebiederm@xmission.com,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
mszeredi@redhat.com
Subject: Re: [PATCH 03/34] teach move_mount(2) to work with OPEN_TREE_CLONE [ver #12]
Date: Sat, 20 Oct 2018 12:48:26 +0100 [thread overview]
Message-ID: <20181020114826.GM32577@ZenIV.linux.org.uk> (raw)
In-Reply-To: <209e8c35-d26e-0a29-84d7-b8b1d0ecbebc@gmail.com>
On Sat, Oct 20, 2018 at 12:06:32PM +0100, Alan Jenkins wrote:
> You posted an analysis of a GPF, where you showed the reference count was
> clearly one less than it should have been. You narrowed this down to a step
> where you connected an unmounted mount (MNT_UMOUNT) to a mounted mount. So
> your analysis is consistent with the comment in disconnect_mount(), which
> says 1) you're not allowed to do that, 2) the reason is because of different
> reference-counting rules. AFAICT, the GPF you analyzed would be prevented
> by the fix in do_move_mount(), checking for MNT_UMOUNT.
Not just refcounting; it's that fs_pin is really intended to have ->kill()
triggered only once. If you look at the pin_kill() (which is where the
livelock happened) you'll see what's going on - anyone hitting it between
the first call and freeing of the object will be sleeping until ->kill()
from the first call gets through pin_remove(), at which point they bugger
off (being very careful with accessing the sucker to avoid use-after-free).
MNT_UMOUNT means that there's no way back.
> pre-date MNT_UMOUNT. I *think* the added check in dissolve_on_fput() makes
> things right, but I don't understand enough to be sure.
That, plus making sure that do_move_mount() grabs a reference in case
of successfully attaching a tree. I hate passing bool argument, BTW -
better just do mnt_add_count() either before attach_recursive_mnt()
and decrement on failure, or, better yet, just do it on success. Note
that namespace_sem is held, so the damn thing *can't* disappear under
us - nobody will be able to detach it until we drop namespace_lock.
> diff --git a/fs/namespace.c b/fs/namespace.c
> index 4dfe7e23b7ee..e8d61d5f581d 100644
> --- a/fs/namespace.c
> +++ b/fs/namespace.c
> @@ -1763,7 +1763,7 @@ void dissolve_on_fput(struct vfsmount *mnt)
> {
> namespace_lock();
> lock_mount_hash();
> - if (!real_mount(mnt)->mnt_ns) {
> + if (!real_mount(mnt)->mnt_ns && !(mnt->mnt_flags & MNT_UMOUNT)) {
> mntget(mnt);
> umount_tree(real_mount(mnt), UMOUNT_CONNECTED);
> }
> @@ -2469,7 +2469,7 @@ static int do_move_mount(struct path *old_path, struct path *new_path)
> if (old->mnt_ns && !attached)
> goto out1;
>
> - if (old->mnt.mnt_flags & MNT_LOCKED)
> + if (old->mnt.mnt_flags & (MNT_LOCKED | MNT_UMOUNT))
> goto out1;
>
> if (old_path->dentry != old_path->mnt->mnt_root)
next prev parent reply other threads:[~2018-10-20 11:48 UTC|newest]
Thread overview: 98+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-21 16:30 [PATCH 00/34] VFS: Introduce filesystem context [ver #12] David Howells
2018-09-21 16:30 ` [PATCH 01/34] vfs: syscall: Add open_tree(2) to reference or clone a mount " David Howells
2018-10-21 16:41 ` Eric W. Biederman
2018-09-21 16:30 ` [PATCH 02/34] vfs: syscall: Add move_mount(2) to move mounts around " David Howells
2018-09-21 16:30 ` [PATCH 03/34] teach move_mount(2) to work with OPEN_TREE_CLONE " David Howells
2018-10-05 18:24 ` Alan Jenkins
2018-10-07 10:48 ` Alan Jenkins
2018-10-07 19:20 ` Alan Jenkins
2018-10-10 12:36 ` David Howells
2018-10-12 14:22 ` Alan Jenkins
2018-10-12 14:54 ` David Howells
2018-10-12 14:57 ` Alan Jenkins
2018-10-11 9:17 ` David Howells
2018-10-11 11:48 ` Alan Jenkins
2018-10-11 13:10 ` David Howells
2018-10-11 12:14 ` David Howells
2018-10-11 12:23 ` Alan Jenkins
2018-10-11 15:33 ` David Howells
2018-10-11 18:38 ` Eric W. Biederman
2018-10-11 20:17 ` David Howells
2018-10-13 6:06 ` Al Viro
2018-10-17 17:45 ` Alan Jenkins
2018-10-18 20:09 ` David Howells
2018-10-18 20:58 ` David Howells
2018-10-19 11:57 ` David Howells
2018-10-19 13:37 ` David Howells
2018-10-19 17:35 ` Alan Jenkins
2018-10-19 21:35 ` David Howells
2018-10-19 21:40 ` David Howells
2018-10-19 22:36 ` David Howells
2018-10-20 5:25 ` Al Viro
2018-10-20 11:06 ` Alan Jenkins
2018-10-20 11:48 ` Al Viro [this message]
2018-10-20 12:26 ` Al Viro
2018-10-21 0:40 ` David Howells
2018-10-10 11:56 ` David Howells
2018-10-10 12:31 ` David Howells
2018-10-10 12:39 ` Alan Jenkins
2018-10-10 12:50 ` David Howells
2018-10-10 13:02 ` David Howells
2018-10-10 13:06 ` Alan Jenkins
2018-10-21 16:57 ` Eric W. Biederman
2018-10-23 11:19 ` Alan Jenkins
2018-10-23 16:22 ` Al Viro
2018-09-21 16:30 ` [PATCH 04/34] vfs: Suppress MS_* flag defs within the kernel unless explicitly enabled " David Howells
2018-09-21 16:30 ` [PATCH 05/34] vfs: Introduce the basic header for the new mount API's filesystem context " David Howells
2018-09-21 16:30 ` [PATCH 06/34] vfs: Introduce logging functions " David Howells
2018-09-21 16:31 ` [PATCH 07/34] vfs: Add configuration parser helpers " David Howells
2019-03-14 7:46 ` Geert Uytterhoeven
2019-03-14 10:27 ` David Howells
2019-03-14 10:49 ` Geert Uytterhoeven
2018-09-21 16:31 ` [PATCH 08/34] vfs: Add LSM hooks for the new mount API " David Howells
2018-09-21 16:31 ` [PATCH 09/34] vfs: Put security flags into the fs_context struct " David Howells
2018-09-21 16:31 ` [PATCH 10/34] selinux: Implement the new mount API LSM hooks " David Howells
2018-09-21 16:31 ` [PATCH 11/34] smack: Implement filesystem context security " David Howells
2018-09-21 16:31 ` [PATCH 12/34] apparmor: Implement security hooks for the new mount API " David Howells
2018-09-21 16:31 ` [PATCH 13/34] tomoyo: " David Howells
2018-09-21 16:32 ` [PATCH 14/34] vfs: Separate changing mount flags full remount " David Howells
2018-09-21 16:32 ` [PATCH 15/34] vfs: Implement a filesystem superblock creation/configuration context " David Howells
2018-09-21 16:32 ` [PATCH 16/34] vfs: Remove unused code after filesystem context changes " David Howells
2018-09-21 16:32 ` [PATCH 17/34] procfs: Move proc_fill_super() to fs/proc/root.c " David Howells
2018-09-21 16:32 ` [PATCH 18/34] proc: Add fs_context support to procfs " David Howells
2018-09-21 16:32 ` [PATCH 19/34] ipc: Convert mqueue fs to fs_context " David Howells
2018-09-21 16:32 ` [PATCH 20/34] cpuset: Use " David Howells
2018-09-21 16:33 ` [PATCH 21/34] kernfs, sysfs, cgroup, intel_rdt: Support " David Howells
2018-11-19 4:23 ` Andrei Vagin
2018-12-06 17:08 ` Andrei Vagin
2018-09-21 16:33 ` [PATCH 22/34] hugetlbfs: Convert to " David Howells
2018-09-21 16:33 ` [PATCH 23/34] vfs: Remove kern_mount_data() " David Howells
2018-09-21 16:33 ` [PATCH 24/34] vfs: Provide documentation for new mount API " David Howells
2018-09-21 16:33 ` [PATCH 25/34] Make anon_inodes unconditional " David Howells
2018-09-21 16:33 ` [PATCH 26/34] vfs: syscall: Add fsopen() to prepare for superblock creation " David Howells
2018-09-21 16:33 ` [PATCH 27/34] vfs: Implement logging through fs_context " David Howells
2018-09-21 16:33 ` [PATCH 28/34] vfs: Add some logging to the core users of the fs_context log " David Howells
2018-09-21 16:34 ` [PATCH 29/34] vfs: syscall: Add fsconfig() for configuring and managing a context " David Howells
2018-09-21 16:34 ` [PATCH 30/34] vfs: syscall: Add fsmount() to create a mount for a superblock " David Howells
2018-09-21 16:34 ` [PATCH 31/34] vfs: syscall: Add fspick() to select a superblock for reconfiguration " David Howells
2018-10-12 14:49 ` Alan Jenkins
2018-10-13 6:11 ` Al Viro
2018-10-13 9:45 ` Alan Jenkins
2018-10-13 23:04 ` Andy Lutomirski
2018-10-17 13:15 ` David Howells
2018-10-17 13:20 ` David Howells
2018-10-17 14:31 ` Alan Jenkins
2018-10-17 14:35 ` Eric W. Biederman
2018-10-17 14:55 ` Alan Jenkins
2018-10-17 15:24 ` David Howells
2018-10-17 15:38 ` Eric W. Biederman
2018-10-17 15:45 ` David Howells
2018-10-17 17:41 ` Alan Jenkins
2018-10-17 21:20 ` David Howells
2018-10-17 22:13 ` Alan Jenkins
2018-09-21 16:34 ` [PATCH 32/34] afs: Add fs_context support " David Howells
2018-09-21 16:34 ` [PATCH 33/34] afs: Use fs_context to pass parameters over automount " David Howells
2018-09-21 16:34 ` [PATCH 34/34] vfs: Add a sample program for the new mount API " David Howells
2018-12-17 14:12 ` Anders Roxell
2018-12-20 16:36 ` David Howells
2018-10-04 18:37 ` [PATCH 00/34] VFS: Introduce filesystem context " Eric W. Biederman
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=20181020114826.GM32577@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=alan.christopher.jenkins@gmail.com \
--cc=dhowells@redhat.com \
--cc=ebiederm@xmission.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mszeredi@redhat.com \
--cc=torvalds@linux-foundation.org \
/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).