linux-kernel - Re: WARNING in kill_block

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Wed, 11 Apr 2018 01:59:38 +0100
From:   Al Viro <viro@...IV.linux.org.uk>
To:     Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
Cc:     Michal Hocko <mhocko@...e.com>,
        syzbot <syzbot+5a170e19c963a2e0df79@...kaller.appspotmail.com>,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        syzkaller-bugs@...glegroups.com, linux-mm <linux-mm@...ck.org>,
        Dmitry Vyukov <dvyukov@...gle.com>
Subject: Re: WARNING in kill_block_super

On Wed, Apr 04, 2018 at 07:53:07PM +0900, Tetsuo Handa wrote:
> Al and Michal, are you OK with this patch?

First of all, it does *NOT* fix the problems with careless ->kill_sb().
The fuse-blk case is the only real rationale so far.  Said that,

> @@ -166,6 +166,7 @@ static void destroy_unused_super(struct super_block *s)
>  	security_sb_free(s);
>  	put_user_ns(s->s_user_ns);
>  	kfree(s->s_subtype);
> +	kfree(s->s_shrink.nr_deferred);

is probably better done with an inlined helper (fs/super.c has no business knowing
about ->nr_deferred name, and there probably will be other users of that
preallocation of yours).  And the same helper would be better off zeroing the
pointer, same as unregister_shrinker() does.


> -int register_shrinker(struct shrinker *shrinker)
> +int prepare_shrinker(struct shrinker *shrinker)

preallocate_shrinker(), perhaps?

> +int register_shrinker(struct shrinker *shrinker)
> +{
> +	int err = prepare_shrinker(shrinker);
> +
> +	if (err)
> +		return err;
> +	register_shrinker_prepared(shrinker);

	if (!err)
		register_....;
	return err;

would be better, IMO.