From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754645Ab3LJQ6F (ORCPT ); Tue, 10 Dec 2013 11:58:05 -0500 Received: from merlin.infradead.org ([205.233.59.134]:55333 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752892Ab3LJQ6B (ORCPT ); Tue, 10 Dec 2013 11:58:01 -0500 Date: Tue, 10 Dec 2013 17:57:36 +0100 From: Peter Zijlstra To: Davidlohr Bueso Cc: linux-kernel@vger.kernel.org, mingo@kernel.org, dvhart@linux.intel.com, tglx@linutronix.de, paulmck@linux.vnet.ibm.com, efault@gmx.de, jeffm@suse.com, torvalds@linux-foundation.org, scott.norton@hp.com, tom.vaden@hp.com, aswin@hp.com, Waiman.Long@hp.com, jason.low2@hp.com Subject: Re: [PATCH v2 4/4] futex: Avoid taking hb lock if nothing to wakeup Message-ID: <20131210165736.GN12849@twins.programming.kicks-ass.net> References: <1386063927-6545-1-git-send-email-davidlohr@hp.com> <1386063927-6545-5-git-send-email-davidlohr@hp.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1386063927-6545-5-git-send-email-davidlohr@hp.com> User-Agent: Mutt/1.5.21 (2012-12-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Dec 03, 2013 at 01:45:27AM -0800, Davidlohr Bueso wrote: > --- a/kernel/futex.c > +++ b/kernel/futex.c > @@ -82,10 +82,12 @@ > * The waker side modifies the user space value of the futex and calls > * futex_wake(). It computes the hash bucket and acquires the hash > * bucket lock. Then it looks for waiters on that futex in the hash > - * bucket and wakes them. > - * > - * Note that the spin_lock serializes waiters and wakers, so that the > - * following scenario is avoided: > + * bucket and wakes them. Why not let this be the start of a new paragraph? > In scenarios where wakeups are called and no > + * tasks are blocked on a futex, taking the hb spinlock can be avoided > + * and simply return. In order for this optimization to work, ordering > + * guarantees must exist so that the waiter being added to the list is > + * acknowledged when the list is concurrently being checked by the waker, > + * avoiding scenarios like the following: > * > * CPU 0 CPU 1 > * val = *futex; > @@ -106,24 +108,40 @@ > * This would cause the waiter on CPU 0 to wait forever because it > * missed the transition of the user space value from val to newval > * and the waker did not find the waiter in the hash bucket queue. > + * The correct serialization ensures that a waiter either observes > + * the changed user space value before blocking or is woken by a > + * concurrent waker: > * > * CPU 0 CPU 1 > * val = *futex; > * sys_futex(WAIT, futex, val); > * futex_wait(futex, val); > + * > + * mb(); <-- paired with ------ > + * | > + * lock(hash_bucket(futex)); | > + * | > + * uval = *futex; | > + * | *futex = newval; > + * | sys_futex(WAKE, futex); > + * | futex_wake(futex); > + * | > + * --------> mb(); > * if (uval == val) > + * queue(); > * unlock(hash_bucket(futex)); > + * schedule(); if (!queue_empty()) > + * lock(hash_bucket(futex)); > + * wake_waiters(futex); > + * unlock(hash_bucket(futex)); > + * > + * The length of the list is tracked with atomic ops (hb->waiters), > + * providing the necessary memory barriers for the waiters. For the > + * waker side, however, we rely on get_futex_key_refs(), using either > + * ihold() or the atomic_inc(), for shared futexes. The former provides > + * a full mb on all architectures. For architectures that do not have an > + * implicit barrier in atomic_inc/dec, we explicitly add it - please > + * refer to futex_get_mm() and hb_waiters_inc/dec(). > */ This comment actually confuses me :/ It isn't at all explained what purpose the memory barriers serve.