Return-Path: Received: from mail-px0-f174.google.com ([209.85.212.174]:51470 "EHLO mail-px0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751674Ab0HTNXU (ORCPT ); Fri, 20 Aug 2010 09:23:20 -0400 Date: Fri, 20 Aug 2010 21:23:09 +0800 From: Wu Fengguang To: Trond Myklebust Cc: Jeff Layton , Christoph Hellwig , linux-nfs@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-mm@kvack.org Subject: Re: why are WB_SYNC_NONE COMMITs being done with FLUSH_SYNC set ? Message-ID: <20100820132309.GB20126@localhost> References: <20100819101525.076831ad@barsoom.rdu.redhat.com> <20100819143710.GA4752@infradead.org> <1282229905.6199.19.camel@heimdal.trondhjem.org> <20100819151618.5f769dc9@tlielax.poochiereds.net> <1282246999.7799.66.camel@heimdal.trondhjem.org> Content-Type: text/plain; charset=us-ascii In-Reply-To: <1282246999.7799.66.camel@heimdal.trondhjem.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 > > Here's a lightly tested patch that turns the check for the two flags > > into a check for WB_SYNC_NONE. It seems to do the right thing, but I > > don't have a clear testcase for it. Does this look reasonable? > > Looks fine to me. I'll queue it up for the post-2.6.36 merge window... Trond, I just created a patch that removes the wbc->nonblocking definition and all its references except NFS. So there will be merge dependencies. What should we do? To push both patches to Andrew's -mm tree? Thanks, Fengguang