From mboxrd@z Thu Jan 1 00:00:00 1970 From: Waiman Long Subject: Re: [PATCH v9 0/5] Add NUMA-awareness to qspinlock Date: Fri, 24 Jan 2020 20:59:28 -0500 Message-ID: <02defadb-217d-7803-88a1-ec72a37eda28@redhat.com> References: <20200115035920.54451-1-alex.kogan@oracle.com> <20200124222434.GA7196@paulmck-ThinkPad-P72> <6AAE7FC6-F5DE-4067-8BC4-77F27948CD09@oracle.com> <20200125005713.GZ2935@paulmck-ThinkPad-P72> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from us-smtp-1.mimecast.com ([205.139.110.61]:27070 "EHLO us-smtp-delivery-1.mimecast.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S2387608AbgAYB7i (ORCPT ); Fri, 24 Jan 2020 20:59:38 -0500 In-Reply-To: <20200125005713.GZ2935@paulmck-ThinkPad-P72> Content-Language: en-US Sender: linux-arch-owner@vger.kernel.org List-ID: To: paulmck@kernel.org, Alex Kogan Cc: linux@armlinux.org.uk, Peter Zijlstra , Ingo Molnar , Will Deacon , Arnd Bergmann , linux-arch@vger.kernel.org, linux-arm-kernel , linux-kernel@vger.kernel.org, tglx@linutronix.de, bp@alien8.de, hpa@zytor.com, x86@kernel.org, guohanjun@huawei.com, jglauber@marvell.com, dave.dice@oracle.com, steven.sistare@oracle.com, daniel.m.jordan@oracle.com On 1/24/20 7:57 PM, Paul E. McKenney wrote: > On Fri, Jan 24, 2020 at 06:39:02PM -0500, Alex Kogan wrote: >> Hi, Paul. >> >> Thanks for running those experiments! >> >>> On Jan 24, 2020, at 5:24 PM, Paul E. McKenney wr= ote: >>> >>> On Tue, Jan 14, 2020 at 10:59:15PM -0500, Alex Kogan wrote: >>>> Minor changes from v8 based on feedback from Longman: >>>> ----------------------------------------------------- >>>> >>>> - Add __init to cna_configure_spin_lock_slowpath(). >>>> >>>> - Fix the comment for cna_scan_main_queue(). >>>> >>>> - Change the type of intra_node_handoff_threshold to unsigned int. >>>> >>>> >>>> Summary >>>> ------- >>>> >>>> Lock throughput can be increased by handing a lock to a waiter on th= e >>>> same NUMA node as the lock holder, provided care is taken to avoid >>>> starvation of waiters on other NUMA nodes. This patch introduces CNA >>>> (compact NUMA-aware lock) as the slow path for qspinlock. It is >>>> enabled through a configuration option (NUMA_AWARE_SPINLOCKS). >>>> >>>> CNA is a NUMA-aware version of the MCS lock. Spinning threads are >>>> organized in two queues, a main queue for threads running on the sam= e >>>> node as the current lock holder, and a secondary queue for threads >>>> running on other nodes. Threads store the ID of the node on which >>>> they are running in their queue nodes. After acquiring the MCS lock = and >>>> before acquiring the spinlock, the lock holder scans the main queue >>>> looking for a thread running on the same node (pre-scan). If found (= call >>>> it thread T), all threads in the main queue between the current lock >>>> holder and T are moved to the end of the secondary queue. If such T >>>> is not found, we make another scan of the main queue after acquiring= =20 >>>> the spinlock when unlocking the MCS lock (post-scan), starting at th= e >>>> node where pre-scan stopped. If both scans fail to find such T, the >>>> MCS lock is passed to the first thread in the secondary queue. If th= e >>>> secondary queue is empty, the MCS lock is passed to the next thread = in the >>>> main queue. To avoid starvation of threads in the secondary queue, t= hose >>>> threads are moved back to the head of the main queue after a certain >>>> number of intra-node lock hand-offs. >>>> >>>> More details are available at https://urldefense.proofpoint.com/v2/u= rl?u=3Dhttps-3A__arxiv.org_abs_1810.05600&d=3DDwIBAg&c=3DRoP1YumCXCgaWHvl= ZYR8PZh8Bv7qIrMUB65eapI_JnE&r=3DHvhk3F4omdCk-GE1PTOm3Kn0A7ApWOZ2aZLTuVxFK= 4k&m=3D1KUGGZYTHnQ25fgRFppdNvpJfI0rOO_Usdu18RDu_14&s=3DF12nhHutwnPNt_TQ2E= LER0DhtsHlEI9EiW1nDPhm5-Y&e=3D . >>>> >>>> The series applies on top of v5.5.0-rc6, commit b3a987b026. >>>> Performance numbers are available in previous revisions >>>> of the series. >>>> >>>> Further comments are welcome and appreciated. >>> I ran this on a large system with a version of locktorture that was >>> modified to print out the maximum and minimum per-CPU lock-acquisitio= n >>> counts, and with CPU hotplug disabled. I also modified the LOCK01 an= d >>> LOCK04 scenarios to use 220 hardware threads. >>> >>> Here is what the test ended up with at the end of a one-hour run: >>> >>> LOCK01 (exclusive): >>> Writes: Total: 1241107333 Max/Min: 9206962/60902 ??? Fail: 0 >>> >>> LOCK04 (rwlock): >>> Writes: Total: 232991963 Max/Min: 2631574/74582 ??? Fail: 0 >>> Reads : Total: 216935386 Max/Min: 2735939/28665 ??? Fail: 0 >>> >>> The "???" strings are printed because the ratio of maximum to minimum= exceeds >>> a factor of two. >> Is this what you expect / have seen with the existing qspinlock? >> >>> I also ran 30-minute runs on my laptop, which has 12 hardware threads= : >>> >>> LOCK01 (exclusive): >>> Writes: Total: 3992072782 Max/Min: 259368782/97231961 ??? Fail: 0 >>> >>> LOCK04 (rwlock): >>> Writes: Total: 131063892 Max/Min: 13136206/5876157 ??? Fail: 0 >>> Reads : Total: 144876801 Max/Min: 19999535/4873442 ??? Fail: 0 >> I assume the system above is multi-socket, but your laptop is probably= not? >> >> If that=E2=80=99s the case, CNA should not be enabled on your laptop (= grep >> kernel logs for "Enabling CNA spinlock=E2=80=9D to be sure). >> >>> These also exceed the factor-of-two cutoff, but not as dramatically. >>> The readers for the reader-writer lock fared worst, with a 4-to-1 rat= io. >>> >>> These tests did run within guest OSes. >> So I really wonder if CNA was enabled here, or whether this is what yo= u get >> with paravirt qspinlock. >> >>> Is that configuration out of >>> scope for this locking algorithm? In addition (as might well also ha= ve >>> been the case for the locktorture runs in your paper), these tests ru= n >>> a pair of stress-test tasks for each hardware thread. >>> >>> Is this expected behavior? >> The results do appear skewed a bit too much, but it would be helpful t= o know >> what qspinlock we are looking at, and how they compare to the existing= qspinlock, >> in case it is indeed CNA. > You called it! I will play with QEMU's -numa argument to see if I can = get > CNA to run for me. Please accept my apologies for the false alarm. > > Thanx, Paul > CNA is not currently supported in a VM guest simply because the numa information is not reliable. You will have to run it on baremetal to test it. Sorry for that. Regards, Longman