From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tao Ma Subject: Re: [URGENT PATCH] ext4: fix potential deadlock in ext4_evict_inode() Date: Fri, 26 Aug 2011 17:10:35 +0800 Message-ID: <4E57630B.7000202@tao.ma> References: <20110826073507.GZ3162@dastard> <20110826084403.GA3162@dastard> <20110826085057.GA13311@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Dave Chinner , Theodore Ts'o , Jiaying Zhang , linux-ext4@vger.kernel.org To: Christoph Hellwig Return-path: Received: from oproxy8-pub.bluehost.com ([69.89.22.20]:54125 "HELO oproxy8-pub.bluehost.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1754566Ab1HZJKk (ORCPT ); Fri, 26 Aug 2011 05:10:40 -0400 In-Reply-To: <20110826085057.GA13311@infradead.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: Hi Christoph, On 08/26/2011 04:50 PM, Christoph Hellwig wrote: > On Fri, Aug 26, 2011 at 06:44:03PM +1000, Dave Chinner wrote: >> While the lockdep report is false positive, I agree that your >> change is the right fix to make - the IO completions are already >> queued on the workqueue, so they don't need to be flushed to get >> them to complete. All that needs to be done is call >> ext4_ioend_wait() for them to complete, and that gets rid of the >> i_mutex altogether. (*) > > The really correct fix is to stop using the ioend wait like I did > in my pending series for XFS. The only thing preventing ext4 from > doing that is that it marks pages uptodate from irq context, > before finishing the ioends. This has a lot of nasty implications > as the the page might be marked uptodate without actually having > that state on disk. yeah, the DIO read on a buffer write file issue is also another side effect of it.[1] So could you please talk a little more about how xfs handle it now? [1] http://www.spinics.net/lists/linux-ext4/msg27139.html Thanks Tao