From: "J. Bruce Fields" Subject: Re: [PATCH 4/14] locks: add locking function that returns conflicting lock Date: Sat, 3 Feb 2007 21:02:33 -0500 Message-ID: <20070204020233.GJ7588@fieldses.org> References: <96a0abaf64b433a7e7450e7e35d0baf2c44103db.1170479265.git.bfields@citi.umich.edu> <20070203085450.GD18828@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-fsdevel@vger.kernel.org, nfs@lists.sourceforge.net, Marc Eshel To: Christoph Hellwig Return-path: In-Reply-To: <20070203085450.GD18828@infradead.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Sat, Feb 03, 2007 at 08:54:50AM +0000, Christoph Hellwig wrote: > Is there any reason we want to have both variants? I think vfs_lock_file > should simple get the last argument and we shouldn't have a separate > vfs_lock_file_conf (which btw doesn't have an exactly descriptive name). > The also allows you to kill __posix_lock_file_conf and only keep a single > posix_lock_file routines that gets the argument aswell. That sounds reasonable. > And btw, in case you ask why I demand all these addition cleanuos: > you're touching an already more than messy codebase and make it more > complex, that needs some cleanups to counterbalance :) I find locks.c really hard to read, so I'm happy for any suggestions for cleanups.... --b.