From: Ming Lei <ming.lei@redhat.com>
To: Luis Chamberlain <mcgrof@kernel.org>
Cc: tj@kernel.org, gregkh@linuxfoundation.org,
akpm@linux-foundation.org, minchan@kernel.org, jeyu@kernel.org,
shuah@kernel.org, bvanassche@acm.org, dan.j.williams@intel.com,
joe@perches.com, tglx@linutronix.de, keescook@chromium.org,
rostedt@goodmis.org, linux-spdx@vger.kernel.org,
linux-doc@vger.kernel.org, linux-block@vger.kernel.org,
linux-fsdevel@vger.kernel.org, linux-kselftest@vger.kernel.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v8 09/12] sysfs: fix deadlock race with module removal
Date: Tue, 12 Oct 2021 08:20:46 +0800 [thread overview]
Message-ID: <YWTU3kTlJKONyFjZ@T590> (raw)
In-Reply-To: <YWSr2trabEJflzlj@bombadil.infradead.org>
On Mon, Oct 11, 2021 at 02:25:46PM -0700, Luis Chamberlain wrote:
> On Tue, Oct 05, 2021 at 05:24:18PM +0800, Ming Lei wrote:
> > On Mon, Sep 27, 2021 at 09:38:02AM -0700, Luis Chamberlain wrote:
> > > When driver sysfs attributes use a lock also used on module removal we
> > > can race to deadlock. This happens when for instance a sysfs file on
> > > a driver is used, then at the same time we have module removal call
> > > trigger. The module removal call code holds a lock, and then the
> > > driver's sysfs file entry waits for the same lock. While holding the
> > > lock the module removal tries to remove the sysfs entries, but these
> > > cannot be removed yet as one is waiting for a lock. This won't complete
> > > as the lock is already held. Likewise module removal cannot complete,
> > > and so we deadlock.
> > >
> > > This can now be easily reproducible with our sysfs selftest as follows:
> > >
> > > ./tools/testing/selftests/sysfs/sysfs.sh -t 0027
> > >
> > > This uses a local driver lock. Test 0028 can also be used, that uses
> > > the rtnl_lock():
> > >
> > > ./tools/testing/selftests/sysfs/sysfs.sh -t 0028
> > >
> > > To fix this we extend the struct kernfs_node with a module reference
> > > and use the try_module_get() after kernfs_get_active() is called. As
> > > documented in the prior patch, we now know that once kernfs_get_active()
> > > is called the module is implicitly guarded to exist and cannot be removed.
> > > This is because the module is the one in charge of removing the same
> > > sysfs file it created, and removal of sysfs files on module exit will wait
> > > until they don't have any active references. By using a try_module_get()
> > > after kernfs_get_active() we yield to let module removal trump calls to
> > > process a sysfs operation, while also preventing module removal if a sysfs
> > > operation is in already progress. This prevents the deadlock.
> > >
> > > This deadlock was first reported with the zram driver, however the live
> >
> > Looks not see the lock pattern you mentioned in zram driver, can you
> > share the related zram code?
>
> I recommend to not look at the zram driver, instead look at the
> test_sysfs driver as that abstracts the issue more clearly and uses
Looks test_sysfs isn't in linus tree, where can I find it? Also please
update your commit log about this wrong info if it can't be applied on
zram.
> two different locks as an example. The point is that if on module
> removal *any* lock is used which is *also* used on the sysfs file
> created by the module, you can deadlock.
>
> > > And this can lead to this condition:
> > >
> > > CPU A CPU B
> > > foo_store()
> > > foo_exit()
> > > mutex_lock(&foo)
> > > mutex_lock(&foo)
> > > del_gendisk(some_struct->disk);
> > > device_del()
> > > device_remove_groups()
> >
> > I guess the deadlock exists if foo_exit() is called anywhere. If yes,
> > look the issue may not be related with removing module directly, right?
>
> No, the reason this can deadlock is that the module exit routine will
> patiently wait for the sysfs / kernfs files to be stop being used,
Can you share the code which waits for the sysfs / kernfs files to be
stop being used? And why does it make a difference in case of being
called from module_exit()?
Thanks,
Ming
next prev parent reply other threads:[~2021-10-12 0:21 UTC|newest]
Thread overview: 94+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-27 16:37 [PATCH v8 00/12] syfs: generic deadlock fix with module removal Luis Chamberlain
2021-09-27 16:37 ` [PATCH v8 01/12] LICENSES: Add the copyleft-next-0.3.1 license Luis Chamberlain
[not found] ` <202110050907.35FBD2A1@keescook>
[not found] ` <YWR2ZrtzChamY1y4@bombadil.infradead.org>
2021-10-11 17:57 ` Kees Cook
2021-09-27 16:37 ` [PATCH v8 02/12] testing: use the copyleft-next-0.3.1 SPDX tag Luis Chamberlain
2021-10-05 16:11 ` Kees Cook
2021-09-27 16:37 ` [PATCH v8 03/12] selftests: add tests_sysfs module Luis Chamberlain
2021-10-05 14:16 ` Greg KH
2021-10-05 16:57 ` Tim.Bird
2021-10-11 17:40 ` Luis Chamberlain
2021-10-11 17:38 ` Luis Chamberlain
2021-10-07 14:23 ` Miroslav Benes
2021-10-11 19:11 ` Luis Chamberlain
[not found] ` <202110050912.3DF681ED@keescook>
2021-10-11 19:03 ` Luis Chamberlain
2021-09-27 16:37 ` [PATCH v8 04/12] kernfs: add initial failure injection support Luis Chamberlain
2021-10-05 19:47 ` Kees Cook
2021-10-11 20:44 ` Luis Chamberlain
2021-09-27 16:37 ` [PATCH v8 05/12] test_sysfs: add support to use kernfs failure injection Luis Chamberlain
2021-10-05 19:51 ` Kees Cook
2021-10-11 20:56 ` Luis Chamberlain
2021-09-27 16:37 ` [PATCH v8 06/12] kernel/module: add documentation for try_module_get() Luis Chamberlain
2021-10-05 19:58 ` Kees Cook
2021-10-11 21:16 ` Luis Chamberlain
2021-09-27 16:38 ` [PATCH v8 07/12] fs/kernfs/symlink.c: replace S_IRWXUGO with 0777 on kernfs_create_link() Luis Chamberlain
2021-10-05 19:59 ` Kees Cook
2021-09-27 16:38 ` [PATCH v8 08/12] fs/sysfs/dir.c: replace S_IRWXU|S_IRUGO|S_IXUGO with 0755 sysfs_create_dir_ns() Luis Chamberlain
2021-10-05 16:05 ` Kees Cook
2021-09-27 16:38 ` [PATCH v8 09/12] sysfs: fix deadlock race with module removal Luis Chamberlain
2021-10-05 9:24 ` Ming Lei
2021-10-11 21:25 ` Luis Chamberlain
2021-10-12 0:20 ` Ming Lei [this message]
2021-10-12 21:18 ` Luis Chamberlain
2021-10-13 1:07 ` Ming Lei
2021-10-13 12:35 ` Luis Chamberlain
2021-10-13 15:04 ` Ming Lei
2021-10-13 21:16 ` Luis Chamberlain
2021-10-05 20:50 ` Kees Cook
2021-10-11 22:26 ` Luis Chamberlain
2021-10-13 12:41 ` Luis Chamberlain
2021-09-27 16:38 ` [PATCH v8 10/12] test_sysfs: enable deadlock tests by default Luis Chamberlain
2021-09-27 16:38 ` [PATCH v8 11/12] zram: fix crashes with cpu hotplug multistate Luis Chamberlain
2021-10-05 20:55 ` Kees Cook
2021-10-11 18:27 ` Luis Chamberlain
2021-10-14 1:55 ` Ming Lei
2021-10-14 2:11 ` Ming Lei
2021-10-14 20:24 ` Luis Chamberlain
2021-10-14 23:52 ` Ming Lei
2021-10-15 0:22 ` Luis Chamberlain
2021-10-15 8:36 ` Ming Lei
2021-10-15 8:52 ` Greg KH
2021-10-15 17:31 ` Luis Chamberlain
2021-10-16 11:28 ` Ming Lei
2021-10-18 19:32 ` Luis Chamberlain
2021-10-19 2:34 ` Ming Lei
2021-10-19 6:23 ` Miroslav Benes
2021-10-19 9:23 ` Ming Lei
2021-10-20 6:43 ` Miroslav Benes
2021-10-20 7:49 ` Ming Lei
2021-10-20 8:19 ` Miroslav Benes
2021-10-20 8:28 ` Greg KH
2021-10-25 9:58 ` Miroslav Benes
2021-10-20 10:09 ` Ming Lei
2021-10-26 8:48 ` Petr Mladek
2021-10-26 15:37 ` Ming Lei
2021-10-26 17:01 ` Luis Chamberlain
2021-10-27 11:57 ` Miroslav Benes
2021-10-27 14:27 ` Luis Chamberlain
2021-11-02 15:24 ` Petr Mladek
2021-11-02 16:25 ` Luis Chamberlain
2021-11-03 0:01 ` Ming Lei
2021-11-03 12:44 ` Luis Chamberlain
2021-10-27 11:42 ` Miroslav Benes
2021-11-02 14:15 ` Petr Mladek
2021-11-02 14:51 ` Petr Mladek
2021-11-02 15:17 ` Ming Lei
2021-11-02 14:56 ` Ming Lei
2021-10-19 15:28 ` Luis Chamberlain
2021-10-19 16:29 ` Ming Lei
2021-10-19 19:36 ` Luis Chamberlain
2021-10-20 1:15 ` Ming Lei
2021-10-20 15:48 ` Luis Chamberlain
2021-10-21 0:39 ` Ming Lei
2021-10-21 17:18 ` Luis Chamberlain
2021-10-22 0:05 ` Ming Lei
2021-10-19 15:50 ` Luis Chamberlain
2021-10-19 16:25 ` Greg KH
2021-10-19 16:30 ` Luis Chamberlain
2021-10-19 17:28 ` Greg KH
2021-10-19 19:46 ` Luis Chamberlain
2021-10-19 16:39 ` Ming Lei
2021-10-19 19:38 ` Luis Chamberlain
2021-10-20 0:55 ` Ming Lei
2021-09-27 16:38 ` [PATCH v8 12/12] zram: use ATTRIBUTE_GROUPS to fix sysfs deadlock module removal Luis Chamberlain
2021-10-05 20:57 ` Kees Cook
2021-10-11 18:28 ` Luis Chamberlain
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=YWTU3kTlJKONyFjZ@T590 \
--to=ming.lei@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=bvanassche@acm.org \
--cc=dan.j.williams@intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=jeyu@kernel.org \
--cc=joe@perches.com \
--cc=keescook@chromium.org \
--cc=linux-block@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-spdx@vger.kernel.org \
--cc=mcgrof@kernel.org \
--cc=minchan@kernel.org \
--cc=rostedt@goodmis.org \
--cc=shuah@kernel.org \
--cc=tglx@linutronix.de \
--cc=tj@kernel.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).