From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754056AbdKGM64 (ORCPT ); Tue, 7 Nov 2017 07:58:56 -0500 Received: from merlin.infradead.org ([205.233.59.134]:60394 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751764AbdKGM6z (ORCPT ); Tue, 7 Nov 2017 07:58:55 -0500 Date: Tue, 7 Nov 2017 13:58:43 +0100 From: Peter Zijlstra To: Waiman Long Cc: Ingo Molnar , linux-kernel@vger.kernel.org, Paolo Bonzini , Juergen Gross , Radim =?utf-8?B?S3LEjW3DocWZ?= , Boris Ostrovsky Subject: Re: [PATCH] locking/pvqspinlock: Hybrid PV queued/unfair locks Message-ID: <20171107125843.GM3165@worktop.lehotels.local> References: <1509723331-10469-1-git-send-email-longman@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1509723331-10469-1-git-send-email-longman@redhat.com> User-Agent: Mutt/1.5.22.1 (2013-10-16) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Nov 03, 2017 at 11:35:31AM -0400, Waiman Long wrote: > Currently, all the lock waiters entering the slowpath will do one > lock stealing attempt to acquire the lock. That helps performance, > especially in VMs with over-committed vCPUs. However, the current > pvqspinlocks still don't perform as good as unfair locks in many cases. > On the other hands, unfair locks do have the problem of lock starvation > that pvqspinlocks don't have. > > This patch combines the best attributes of an unfair lock and a > pvqspinlock into a hybrid lock with 2 modes - queued mode & unfair > mode. A lock waiter goes into the unfair mode when there are waiters > in the wait queue but the pending bit isn't set. Otherwise, it will > go into the queued mode waiting in the queue for its turn. > > On a 2-socket 36-core E5-2699 v3 system (HT off), a kernel build > (make -j) was done in a VM with unpinned vCPUs 3 times with the > best time selected and is the number of vCPUs available. The build > times of the original pvqspinlock, hybrid pvqspinlock and unfair lock > with various number of vCPUs are as follows: > > vCPUs pvqlock hybrid pvqlock unfair lock > ----- ------- -------------- ----------- > 30 342.1s 329.1s 329.1s > 36 314.1s 305.3s 307.3s > 45 345.0s 302.1s 306.6s > 54 365.4s 308.6s 307.8s > 72 358.9s 293.6s 303.9s > 108 343.0s 285.9s 304.2s > > The hybrid pvqspinlock performs better or comparable to the unfair > lock. > > By turning on QUEUED_LOCK_STAT, the table below showed the number > of lock acquisitions in unfair mode and queue mode after a kernel > build with various number of vCPUs. > > vCPUs queued mode unfair mode > ----- ----------- ----------- > 30 9,130,518 294,954 > 36 10,856,614 386,809 > 45 8,467,264 11,475,373 > 54 6,409,987 19,670,855 > 72 4,782,063 25,712,180 > > It can be seen that as the VM became more and more over-committed, > the ratio of locks acquired in unfair mode increases. This is all > done automatically to get the best overall performance as possible. > > The table below shows the kernel build times on a smaller 2-socket > 16-core 32-thread E5-2620 v4 system. > > vCPUs pvqlock hybrid pvqlock unfair lock > ----- ------- -------------- ----------- > 16 436.8s 433.4s 435.6s > 36 366.2s 364.8s 364.5s > 48 423.6s 376.3s 370.2s > 64 433.1s 376.6s 376.8s > > Again, the performance of the hybrid pvqspinlock was comparable to > that of the unfair lock. You forgot to test a starvation case. And you also forgot to describe how they cannot happen. I really cannot remember how all this is supposed to work.