Return-Path: Date: Tue, 18 Feb 2014 10:30:56 -0500 From: Tejun Heo To: Peter Hurley Cc: Ben Hutchings , linux-kernel@vger.kernel.org, Stefan Richter , stable@vger.kernel.org Subject: Re: [PATCH] workqueue: Document exceptions to work item non-reentrancy guarantee Message-ID: <20140218153056.GC30043@htj.dyndns.org> References: <1392493119-9277-1-git-send-email-peter@hurleysoftware.com> <1392687834.10088.17.camel@deadeye.wl.decadent.org.uk> <5302C58E.1070903@hurleysoftware.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <5302C58E.1070903@hurleysoftware.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: On Mon, Feb 17, 2014 at 09:29:34PM -0500, Peter Hurley wrote: > >It never would have occurred to me that you could safely change the > >function for a work item that is already scheduled or running. > >Especially given that PREPARE_WORK() is just a simple assignment (i.e. > >no serialisation). > > process_one_work() has an established order that safely allows for > resetting the work function and scheduling the work, and further > guaranteeing that the new work function will run. > > Further, existing memory barriers ensure that > 1. The new work function is visible on all cpus before testing if > the work is already pending. > 2. The new work function is stored as the worker's current function > before the work is marked as not pending. > > If this wasn't possible, then single-threaded workqueues could > not be used for multiple functions without flushing work. > > I wonder if the floppy driver is broken too. Ugh... I'd just rather remove PREPARE_WORK altogether. It's a pretty dumb thing to do anyway. I'll look into it. Thanks. -- tejun