From owner-svn-src-all@freebsd.org Wed Aug 26 07:24:00 2015 Return-Path: Delivered-To: svn-src-all@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 835339C3400; Wed, 26 Aug 2015 07:24:00 +0000 (UTC) (envelope-from hps@selasky.org) Received: from mail.turbocat.net (mail.turbocat.net [IPv6:2a01:4f8:d16:4514::2]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 4916ED89; Wed, 26 Aug 2015 07:24:00 +0000 (UTC) (envelope-from hps@selasky.org) Received: from laptop015.home.selasky.org (cm-176.74.213.204.customer.telag.net [176.74.213.204]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.turbocat.net (Postfix) with ESMTPSA id 614801FE023; Wed, 26 Aug 2015 09:23:58 +0200 (CEST) Subject: Re: svn commit: r286880 - head/sys/kern To: Julien Charbon , src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org, John Baldwin References: <201508181015.t7IAFAex055889@repo.freebsd.org> From: Hans Petter Selasky Message-ID: <55DD69E5.4090904@selasky.org> Date: Wed, 26 Aug 2015 09:25:25 +0200 User-Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:38.0) Gecko/20100101 Thunderbird/38.1.0 MIME-Version: 1.0 In-Reply-To: <201508181015.t7IAFAex055889@repo.freebsd.org> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 26 Aug 2015 07:24:00 -0000 On 08/18/15 12:15, Julien Charbon wrote: > Author: jch > Date: Tue Aug 18 10:15:09 2015 > New Revision: 286880 > URL: https://svnweb.freebsd.org/changeset/base/286880 > > Log: > callout_stop() should return 0 (fail) when the callout is currently > being serviced and indeed unstoppable. > > A scenario to reproduce this case is: > > - the callout is being serviced and at same time, > - callout_reset() is called on this callout that sets > the CALLOUT_PENDING flag and at same time, > - callout_stop() is called on this callout and returns 1 (success) > even if the callout is indeed currently running and unstoppable. > > This issue was caught up while making r284245 (D2763) workaround, and > was discussed at BSDCan 2015. Once applied the r284245 workaround > is not needed anymore and will be reverted. > > Differential Revision: https://reviews.freebsd.org/D3078 > Reviewed by: jhb > Sponsored by: Verisign, Inc. > > Modified: > head/sys/kern/kern_timeout.c > > Modified: head/sys/kern/kern_timeout.c > ============================================================================== > --- head/sys/kern/kern_timeout.c Tue Aug 18 10:07:03 2015 (r286879) > +++ head/sys/kern/kern_timeout.c Tue Aug 18 10:15:09 2015 (r286880) > @@ -1150,7 +1150,7 @@ _callout_stop_safe(struct callout *c, in > struct callout_cpu *cc, *old_cc; > struct lock_class *class; > int direct, sq_locked, use_lock; > - int not_on_a_list; > + int not_on_a_list, not_running; > > if (safe) > WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, c->c_lock, > @@ -1378,8 +1378,15 @@ again: > } > } > callout_cc_del(c, cc); > + > + /* > + * If we are asked to stop a callout which is currently in progress > + * and indeed impossible to stop then return 0. > + */ > + not_running = !(cc_exec_curr(cc, direct) == c); > + > CC_UNLOCK(cc); > - return (1); > + return (not_running); > } > > void > > Hi, I think this patch is incomplete and can break the return value for non-MPSAFE callouts. I think the correct statement should check the value of "use_lock" too: not_running = !(cc_exec_curr(cc, direct) == c && use_lock == 0); Because if the callback process waits for lock "c_lock" in the callback process then we have above "cc_exec_curr(cc, direct) == c" is satisfied too, and non-MPSAFE callouts are always cancelable, via cc_exec_cancel(cc, direct) = true; > class->lc_lock(c_lock, lock_status); > /* > * The callout may have been cancelled > * while we switched locks. > */ > if (cc_exec_cancel(cc, direct)) { > class->lc_unlock(c_lock); > goto skip; > } > /* The callout cannot be stopped now. */ > cc_exec_cancel(cc, direct) = true; --HPS