From: Peter Zijlstra Subject: Re: reiserfs locking (v2) Date: Fri, 02 Jul 2010 16:38:33 +0200 Message-ID: <1278081513.1917.260.camel@laptop> References: <20100702093451.GA3973@swordfish.minsk.epam.com> <20100702131248.GA5324@nowhere> <1278078267.1917.254.camel@laptop> <20100702143419.GC5324@nowhere> Mime-Version: 1.0 Content-Transfer-Encoding: 8BIT Return-path: In-Reply-To: <20100702143419.GC5324@nowhere> Sender: reiserfs-devel-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="us-ascii" To: Frederic Weisbecker Cc: Sergey Senozhatsky , Jan Kara , Christoph Hellwig , Andrew Morton , reiserfs-devel@vger.kernel.org, linux-kernel@vger.kernel.org, Chris Mason , Jeff Mahoney On Fri, 2010-07-02 at 16:34 +0200, Frederic Weisbecker wrote: > vfs_readdir() locks the directory before calling the fs, and none > of the nested classes would be in its good right there, as the > directory is the current inode to work on, not a parent nor a > child or so. Nah, what I was referring to is sticking the i_mutex into a different class on inode creation, but apparently we already do, see unlock_new_inode(). So its not a pure directory thing..