From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Sun, 24 Jan 2016 09:34:56 +1100 From: Dave Chinner To: Al Viro Cc: Linus Torvalds , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org Subject: Re: [git pull] vfs.git - including i_mutex wrappers Message-ID: <20160123223456.GH6033@dastard> References: <20160123145854.GM17997@ZenIV.linux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160123145854.GM17997@ZenIV.linux.org.uk> Sender: linux-kernel-owner@vger.kernel.org List-ID: On Sat, Jan 23, 2016 at 02:58:54PM +0000, Al Viro wrote: > ->i_mutex wrappers (with small prereq in lustre), fix for too Please explain, Al? I haven't heard anything about there being i_mutex changes pending, and this commit says "over the coming cycle ->i_mutex will become rwsem". That's a complete surprise to me, and not something that should be done with no warning. What's the locking model? How are filesystems supposed to use it? Are they even allowed to use read-mode locking, and if so, what operations is it going to be safe to hold the lock in read mode? Why is this change considered valid now, when previously there's always been significant push-back to any suggestion that we should make the i_mutex a rwsem so we can do shared read-only access locking on inode operations? Cheers, Dave. -- Dave Chinner david@fromorbit.com