From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752751AbZBDHOZ (ORCPT ); Wed, 4 Feb 2009 02:14:25 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751277AbZBDHOQ (ORCPT ); Wed, 4 Feb 2009 02:14:16 -0500 Received: from verein.lst.de ([213.95.11.210]:52745 "EHLO verein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751239AbZBDHOQ (ORCPT ); Wed, 4 Feb 2009 02:14:16 -0500 Date: Wed, 4 Feb 2009 08:13:20 +0100 From: Christoph Hellwig To: Jonathan Corbet Cc: Andrew Morton , mpm@selenic.com, dada1@cosmosbay.com, linux-kernel@vger.kernel.org, andi@firstfloor.org, oleg@redhat.com, viro@ZenIV.linux.org.uk, davidel@xmailserver.org, davem@davemloft.net, hch@lst.de, linux-api@vger.kernel.org, alan@lxorguk.ukuu.org.uk Subject: Re: [PATCH 2/4] Convert epoll to a bitlock Message-ID: <20090204071320.GA19348@lst.de> References: <1233598811-6871-1-git-send-email-corbet@lwn.net> <1233598811-6871-3-git-send-email-corbet@lwn.net> <20090203133942.2ecec281.akpm@linux-foundation.org> <4988BD4E.8080206@cosmosbay.com> <20090203140543.6e915f97.akpm@linux-foundation.org> <1233699722.3243.127.camel@calx> <20090203153740.363d0a04@bike.lwn.net> <20090203145346.8df40277.akpm@linux-foundation.org> <20090203161931.4054a25e@bike.lwn.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090203161931.4054a25e@bike.lwn.net> User-Agent: Mutt/1.3.28i X-Spam-Score: 0 () Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Feb 03, 2009 at 04:19:31PM -0700, Jonathan Corbet wrote: > 1) Use i_lock to protect accesses to f_flags. This would enable some > BKL usage to be removed, but would not fix fasync. What about just turning f_ep_lock into f_lock and using it? > 2) Move responsibility for the FASYNC bit into ->fasync(), with > fasync_helper() doing it in almost all situations. The remaining > BKL usage would then go away. > > 3) The same optional fasync() return values cleanup. These two sound like a good thing to do no matter what the final locking looks like. I think they should be moved to the front of the patch series and queued up no matter what.