From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH v5 05/10] blk-mq: Unregister debugfs attributes earlier To: Bart Van Assche , "osandov@osandov.com" References: <20170425203745.19946-1-bart.vanassche@sandisk.com> <20170425203745.19946-6-bart.vanassche@sandisk.com> <20170425213018.GB6361@vader> <1493159087.2628.28.camel@sandisk.com> Cc: "linux-block@vger.kernel.org" From: Jens Axboe Message-ID: Date: Tue, 25 Apr 2017 15:29:39 -0700 MIME-Version: 1.0 In-Reply-To: <1493159087.2628.28.camel@sandisk.com> Content-Type: text/plain; charset=windows-1252 List-ID: On 04/25/2017 03:24 PM, Bart Van Assche wrote: > On Tue, 2017-04-25 at 14:30 -0700, Omar Sandoval wrote: >> On Tue, Apr 25, 2017 at 01:37:40PM -0700, Bart Van Assche wrote: >>> One of the debugfs attributes allows to run a queue. Since running >>> a queue after a queue has entered the "dead" state is not allowed >>> and triggers a use-after-free, unregister the debugfs attributes >>> before a queue reaches the "dead" state. >> >> Still not happy with this commit message. I'd prefer: >> >> We currently call blk_mq_free_queue() from blk_cleanup_queue() before we >> unregister the debugfs attributes for that queue in blk_release_queue(). >> This leaves a window open during which accessing most of the mq debugfs >> attributes would cause a use-after-free. Additionally, the "state" >> attribute allows running the queue, which we should not do after the >> queue has entered the "dead" state. Fix both of these cases by >> unregistering the debugfs attributes before this. > > Hello Omar, > > That's a very verbose description. How about this? > > Unregister the debugfs attributes before freeing of request queue > resources starts to avoid that a use-after-free can be triggered > through one of the debugfs attributes. Personally I find Omar's commit message much cleaner to read, and more easily understandable. We really don't need to be laconic in commit messages. -- Jens Axboe