From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tim Chen Subject: Re: vfsmount lock issues on very large ppc64 box Date: Mon, 18 Jul 2011 09:41:16 -0700 Message-ID: <1311007276.2707.37.camel@schen9-DESK> References: <20110717105027.53cc3ca4@kryten> <20110717010427.GC5359@parisc-linux.org> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Anton Blanchard , npiggin@kernel.dk, viro@zeniv.linux.org.uk, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, Andi Kleen , Alexander Viro To: Matthew Wilcox Return-path: In-Reply-To: <20110717010427.GC5359@parisc-linux.org> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Sat, 2011-07-16 at 19:04 -0600, Matthew Wilcox wrote: > On Sun, Jul 17, 2011 at 10:50:27AM +1000, Anton Blanchard wrote: > > Looking closer, all of these calls are in pipefs and sockfs. > > Since we never mount either filesystem they never get a long term > > reference and we always end up in the very slow write brlock path > > that takes a lock for each online CPU. > > > > Here is a quick hack that takes a long term reference on pipefs > > and sockfs which fixes the problem. Any thoughts on how we should > > fix it properly? > > I know Tim and Andi have been looking into this ... I forget what their > fix was though. > I've sent out a trial patch a while ago on this problem, but didn't get a response back. Nick or Al, can you take another look at this patch now that other folks are also running into this issue? http://marc.info/?l=linux-fsdevel&m=130273200502778&w=2 Thanks. Tim