linux-kernel - Re: RFC: hold i_rwsem until aio completes

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

Date:   Mon, 3 Feb 2020 18:46:41 +0100
From:   Christoph Hellwig <hch@....de>
To:     Dave Chinner <david@...morbit.com>
Cc:     Christoph Hellwig <hch@....de>, linux-xfs@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, Waiman Long <longman@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Will Deacon <will@...nel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        linux-ext4@...r.kernel.org, cluster-devel@...hat.com,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: RFC: hold i_rwsem until aio completes

On Sat, Jan 18, 2020 at 08:28:38PM +1100, Dave Chinner wrote:
> I think it's pretty gross, actually. It  makes the same mistake made
> with locking in the old direct IO code - it encodes specific lock
> operations via flags into random locations in the DIO path. This is
> a very slippery slope, and IMO it is an layering violation to encode
> specific filesystem locking smeantics into a layer that is supposed
> to be generic and completely filesystem agnostic. i.e.  this
> mechanism breaks if a filesystem moves to a different type of lock
> (e.g. range locks), and history teaches us that we'll end up making
> a horrible, unmaintainable mess to support different locking
> mechanisms and contexts.
> 
> I think that we should be moving to a model where the filesystem
> provides an unlock method in the iomap operations structure, and if
> the method is present in iomap_dio_complete() it gets called for the
> filesystem to unlock the inode at the appropriate point. This also
> allows the filesystem to provide a different method for read or
> write unlock, depending on what type of lock it held at submission.
> This gets rid of the need for the iomap code to know what type of
> lock the caller holds, too.

I'd rather avoid yet another method.  But I think with a little
tweaking we can move the unlock into the ->end_io method.