From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754397AbZIAM1M (ORCPT ); Tue, 1 Sep 2009 08:27:12 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754375AbZIAM1M (ORCPT ); Tue, 1 Sep 2009 08:27:12 -0400 Received: from mx1.redhat.com ([209.132.183.28]:38963 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754346AbZIAM1L (ORCPT ); Tue, 1 Sep 2009 08:27:11 -0400 Date: Tue, 1 Sep 2009 14:23:25 +0200 From: Oleg Nesterov To: Andrew Morton , Roland McGrath Cc: KAMEZAWA Hiroyuki , linux-kernel@vger.kernel.org Subject: [PATCH -mm 1/2] do_wait-wakeup-optimization: fix child_wait_callback()->eligible_child() usage Message-ID: <20090901122325.GB20989@redhat.com> References: <20090827144453.25f1161b.kamezawa.hiroyu@jp.fujitsu.com> <20090827160532.d6386722.kamezawa.hiroyu@jp.fujitsu.com> <20090827093441.GA3451@redhat.com> <20090827184303.500ac1f0.kamezawa.hiroyu@jp.fujitsu.com> <20090827100846.GA6462@redhat.com> <20090827193133.b7eed4a9.kamezawa.hiroyu@jp.fujitsu.com> <20090827105209.GA8469@redhat.com> <20090828171724.GA17445@redhat.com> <20090828191624.0F5BC45B02@magilla.sf.frob.com> <20090901122240.GA20989@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090901122240.GA20989@redhat.com> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org (on top of do_wait-wakeup-optimization-child_wait_callback-check-__wnothread-case.patch, fixes do_wait-wakeup-optimization-change-__wake_up_parent-to-use-filtered-wakeup.patch) child_wait_callback()->eligible_child() is not right, we can miss the wakeup if the task was detached before __wake_up_parent() and the caller of do_wait() didn't use __WALL. Move ->wo_pid checks from eligible_child() to the new helper, eligible_pid(), and change child_wait_callback() to use it instead of eligible_child(). Note: actually I think it would be better to fix the __WCLONE check in eligible_child(), it doesn't look exactly right. But it is not clear what is the supposed behaviour, and any change is user-visible. Reported-by: KAMEZAWA Hiroyuki Signed-off-by: Oleg Nesterov --- kernel/exit.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) --- WAIT/kernel/exit.c~2_WAKE_PARENT_ELIGIBLE_FIX 2009-09-01 12:49:27.000000000 +0200 +++ WAIT/kernel/exit.c 2009-09-01 12:59:23.000000000 +0200 @@ -1106,13 +1106,16 @@ static struct pid *task_pid_type(struct return pid; } -static int eligible_child(struct wait_opts *wo, struct task_struct *p) +static inline int eligible_pid(struct wait_opts *wo, struct task_struct *p) { - if (wo->wo_type < PIDTYPE_MAX) { - if (task_pid_type(p, wo->wo_type) != wo->wo_pid) - return 0; - } + return wo->wo_type == PIDTYPE_MAX || + task_pid_type(p, wo->wo_type) == wo->wo_pid; +} +static int eligible_child(struct wait_opts *wo, struct task_struct *p) +{ + if (!eligible_pid(wo, p)) + return 0; /* Wait for all children (clone and not) if __WALL is set; * otherwise, wait for clone children *only* if __WCLONE is * set; otherwise, wait for non-clone children *only*. (Note: @@ -1564,7 +1567,7 @@ static int child_wait_callback(wait_queu child_wait); struct task_struct *p = key; - if (!eligible_child(wo, p)) + if (!eligible_pid(wo, p)) return 0; if ((wo->wo_flags & __WNOTHREAD) && wait->private != p->parent)