Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1422778AbWKEWxx (ORCPT ); Sun, 5 Nov 2006 17:53:53 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1422791AbWKEWxx (ORCPT ); Sun, 5 Nov 2006 17:53:53 -0500 Received: from ms-smtp-02.nyroc.rr.com ([24.24.2.56]:24807 "EHLO ms-smtp-02.nyroc.rr.com") by vger.kernel.org with ESMTP id S1422778AbWKEWxx (ORCPT ); Sun, 5 Nov 2006 17:53:53 -0500 Date: Sun, 5 Nov 2006 17:53:32 -0500 (EST) From: Steven Rostedt X-X-Sender: rostedt@gandalf.stny.rr.com To: Linus Torvalds cc: Oleg Nesterov , Thomas Gleixner , Andrew Morton , linux-kernel@vger.kernel.org Subject: Re: PATCH? hrtimer_wakeup: fix a theoretical race wrt rt_mutex_slowlock() In-Reply-To: Message-ID: References: <20061105193457.GA3082@oleg> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1920 Lines: 59 On Sun, 5 Nov 2006, Linus Torvalds wrote: > > That said, since "task->state" in only tested _inside_ the runqueue lock, > there is no race that I can see. Since we've gotten the runqueue lock in > order to even check task-state, the processor that _sets_ task state must > not only have done the "spin_lock()", it must also have done the > "spin_unlock()", and _that_ will not allow either the timeout or the task > state to haev leaked out from under it (because that would imply that the > critical region leaked out too). > > So I don't think the race exists anyway - the schedule() will return > immediately (because it will see TASK_RUNNING), and we'll just retry. > This whole situation is very theoretical, but I think this actually can happen *theoretically*. OK, the spin_lock doesn't do any serialization, but the unlock does. But the problem can happen before the unlock. Because of the loop. CPU 1 CPU 2 task_rq_lock() p->state = TASK_RUNNING; (from bottom of for loop) set_current_state(TASK_INTERRUPTIBLE); for (;;) { (looping) if (timeout && !timeout->task) (now CPU implements) t->task = NULL task_rq_unlock(); schedule() (with state == TASK_INTERRUPTIBLE) Again, this is very theoretical, and I don't even think that this can happen if you tried to make it. But I guess if hardware were to change in the future with the same rules that we have today with barriers, that this can be a race. -- Steve - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/