Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Tue, 18 Sep 2018 18:40:53 +0900
From:   Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
To:     David Howells <dhowells@redhat.com>
Cc:     Guenter Roeck <linux@roeck-us.net>, viro@zeniv.linux.org.uk,
        torvalds@linux-foundation.org, linux-fsdevel@vger.kernel.org,
        linux-kernel@vger.kernel.org, Steven Rostedt <rostedt@goodmis.org>,
        Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Subject: Re: [PATCH 14/33] vfs: Implement a filesystem superblock
 creation/configuration context [ver #11]
Message-ID: <20180918094053.GA417@jagdpanzerIV>
References: <20180911220743.GA13208@roeck-us.net>
 <20180911174641.GA15149@roeck-us.net>
 <153313703562.13253.5766498657900728120.stgit@warthog.procyon.org.uk>
 <153313714181.13253.304098108512966976.stgit@warthog.procyon.org.uk>
 <27113.1536702746@warthog.procyon.org.uk>
 <32382.1536707855@warthog.procyon.org.uk>
 <20180911235403.GA10107@roeck-us.net>
 <20180918090722.GA463@jagdpanzerIV>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180918090722.GA463@jagdpanzerIV>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On (09/18/18 18:07), Sergey Senozhatsky wrote:
> emergency_remount()
>  do_emergency_remount()
>   do_emergency_remount_callback()
>    reconfigure_super()
> 
> At fc->ops dereference:
> 
>  981         if (fc->ops->reconfigure) {
> 		^^^^^^^^^
>  982                 retval = fc->ops->reconfigure(fc);
>  983                 if (retval == 0) {
>  984                         security_sb_reconfigure(fc);
> 
> 
> So the check either better be
> 
> 	if (fc->ops && fc->ops->reconfigure)

I guess I was pretty lucky to have leading zeroes in that fc.

David, do you want to add a macro which would make `struct fs_context fc'
misuse less possible? There are 3 users right now who don't use
vfs_new_fs_context(), and none of them appear to properly set all of
`struct fs_context fc' members. This can cause problems in the future,
right?

fs/namespace.c: struct fs_context fc = {
fs/super.c:                     struct fs_context fc = {
fs/super.c:     struct fs_context fc = {

	-ss