Received: by 2002:a25:c205:0:0:0:0:0 with SMTP id s5csp533993ybf; Wed, 26 Feb 2020 18:00:44 -0800 (PST) X-Google-Smtp-Source: APXvYqxH0TjetWwDUizgz5/QljcCHTlZk3c5NSNX0BSzrI1HPtBkPwIsTpqChu1ZCmBbJBG5gIxw X-Received: by 2002:aca:44d7:: with SMTP id r206mr1592291oia.33.1582768844473; Wed, 26 Feb 2020 18:00:44 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1582768844; cv=none; d=google.com; s=arc-20160816; b=DhcMH9PLJ2vvebz7H63npLzGKsn/OiMPSjQ9tXhULX3CumaXENWBP/0x44beYhpFsE oa7g5oyWhMyUCakGHiqMOfnP7r0i9818zL/rmE/w2O9VAe4efltP5mooGktKYEgnXzDr qSMyaQ50sgYeoy/o9CUu87BMquMqDGSlnnXb6KEagTG0cCJpJF87wS9Pz0BgMganmDFN M/aY9dAMlk0crvTTdyrrzuj1UlDMf9pJtUhhnut2IFuItj0DrX7CEHf7Tu5UnApsbdjD A9oBmby1JXK4Swdo1Kmmk2iA/ylrTHnbub8xqME3zrgLNYNOv4J4tGXBGV9fw/cnC3Hs ZKzg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:mime-version:message-id:date:subject:cc :to:from; bh=/UUA1phVq3Oansy4FsJ47zCVLrgoLYVZhBg6ur1wrBc=; b=xi2xUGiFQaAoYGs35vunayQdk7As8idMRLEcPgvWIlGfYVXKBiFVEbCLnFgC7a07m9 aCdRusAeGDZhDqMzxxqKoEltBhIbqRpZvW2vBIWnMZkDxHvfy8plcD1ow9W0rEMm5Fl0 P6QUNaIwS4HkxE/okz0or5j6KmpKGp07UtDPTisU+yrg93nrPtpMByqRjZ4ulVoS4GDg pmyPrifcjUba3ueyblAMnIhTsaTarxkVvNuVecpF98tVlY9fhnwYXLpSMftyB0Nz9OWG fJZnsozrc6mOp5SG5c6jlKSWc3FfV1nK3FVAN97XtKo0RIbt8iy/A59vggiWJ9arIfue AhbQ== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Return-Path: Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67]) by mx.google.com with ESMTP id a143si493329oii.179.2020.02.26.18.00.32; Wed, 26 Feb 2020 18:00:44 -0800 (PST) Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67; Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728244AbgB0B66 (ORCPT + 99 others); Wed, 26 Feb 2020 20:58:58 -0500 Received: from szxga05-in.huawei.com ([45.249.212.191]:10700 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1728177AbgB0B66 (ORCPT ); Wed, 26 Feb 2020 20:58:58 -0500 Received: from DGGEMS402-HUB.china.huawei.com (unknown [172.30.72.58]) by Forcepoint Email with ESMTP id 9AEA941909B378DF80C2; Thu, 27 Feb 2020 09:58:47 +0800 (CST) Received: from localhost.localdomain (10.69.192.56) by DGGEMS402-HUB.china.huawei.com (10.3.19.202) with Microsoft SMTP Server id 14.3.439.0; Thu, 27 Feb 2020 09:58:41 +0800 From: Shaokun Zhang To: , CC: yuqi jin , Rusty Russell , Andrew Morton , Juergen Gross , Paul Burton , Michal Hocko , "Michael Ellerman" , Mike Rapoport , "Anshuman Khandual" , Shaokun Zhang Subject: [PATCH v5] lib: optimize cpumask_local_spread() Date: Thu, 27 Feb 2020 09:58:08 +0800 Message-ID: <1582768688-2314-1-git-send-email-zhangshaokun@hisilicon.com> X-Mailer: git-send-email 2.7.4 MIME-Version: 1.0 Content-Type: text/plain X-Originating-IP: [10.69.192.56] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: yuqi jin In multi-processor and NUMA system, I/O driver will find cpu cores that which shall be bound IRQ. When cpu cores in the local numa have been used, it is better to find the node closest to the local numa node for performance, instead of choosing any online cpu immediately. On Huawei Kunpeng 920 server, there are 4 NUMA node(0 - 3) in the 2-cpu system(0 - 1). The topology of this server is followed: available: 4 nodes (0-3) node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 node 0 size: 63379 MB node 0 free: 61899 MB node 1 cpus: 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 node 1 size: 64509 MB node 1 free: 63942 MB node 2 cpus: 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 node 2 size: 64509 MB node 2 free: 63056 MB node 3 cpus: 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 node 3 size: 63997 MB node 3 free: 63420 MB node distances: node 0 1 2 3 0: 10 16 32 33 1: 16 10 25 32 2: 32 25 10 16 3: 33 32 16 10 We perform PS (parameter server) business test, the behavior of the service is that the client initiates a request through the network card, the server responds to the request after calculation. When two PS processes run on node2 and node3 separately and the network card is located on 'node2' which is in cpu1, the performance of node2 (26W QPS) and node3 (22W QPS) is different. It is better that the NIC queues are bound to the cpu1 cores in turn, then XPS will also be properly initialized, while cpumask_local_spread only considers the local node. When the number of NIC queues exceeds the number of cores in the local node, it returns to the online core directly. So when PS runs on node3 sending a calculated request, the performance is not as good as the node2. The IRQ from 369-392 will be bound from NUMA node0 to NUMA node3 with this patch, before the patch: Euler:/sys/bus/pci # cat /proc/irq/369/smp_affinity_list 0 Euler:/sys/bus/pci # cat /proc/irq/370/smp_affinity_list 1 ... Euler:/sys/bus/pci # cat /proc/irq/391/smp_affinity_list 22 Euler:/sys/bus/pci # cat /proc/irq/392/smp_affinity_list 23 After the patch: Euler:/sys/bus/pci # cat /proc/irq/369/smp_affinity_list 72 Euler:/sys/bus/pci # cat /proc/irq/370/smp_affinity_list 73 ... Euler:/sys/bus/pci # cat /proc/irq/391/smp_affinity_list 94 Euler:/sys/bus/pci # cat /proc/irq/392/smp_affinity_list 95 So the performance of the node3 is the same as node2 that is 26W QPS when the network card is still in 'node2' with the patch. It is considered that the NIC and other I/O devices shall initialize the interrupt binding, if the cores of the local node are used up, it is reasonable to return the node closest to it. Let's optimize it and find the nearest node through NUMA distance for the non-local NUMA nodes. Cc: Rusty Russell Cc: Andrew Morton Cc: Juergen Gross Cc: Paul Burton Cc: Michal Hocko Cc: Michael Ellerman Cc: Mike Rapoport Cc: Anshuman Khandual Signed-off-by: yuqi jin Signed-off-by: Shaokun Zhang --- ChangeLog from v4: 1. Rebase to 5.6-rc3 ChangeLog from v3: 1. Make spread_lock local to cpumask_local_spread(); 2. Add more descriptions on the affinities change in log; ChangeLog from v2: 1. Change the variables as static and use spinlock to protect; 2. Give more explantation on test and performance; lib/cpumask.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 90 insertions(+), 12 deletions(-) diff --git a/lib/cpumask.c b/lib/cpumask.c index 0cb672eb107c..f7394ba36116 100644 --- a/lib/cpumask.c +++ b/lib/cpumask.c @@ -6,6 +6,7 @@ #include #include #include +#include /** * cpumask_next - get the next cpu in a cpumask @@ -192,18 +193,39 @@ void __init free_bootmem_cpumask_var(cpumask_var_t mask) } #endif -/** - * cpumask_local_spread - select the i'th cpu with local numa cpu's first - * @i: index number - * @node: local numa_node - * - * This function selects an online CPU according to a numa aware policy; - * local cpus are returned first, followed by non-local ones, then it - * wraps around. - * - * It's not very efficient, but useful for setup. - */ -unsigned int cpumask_local_spread(unsigned int i, int node) +static void calc_node_distance(int *node_dist, int node) +{ + int i; + + for (i = 0; i < nr_node_ids; i++) + node_dist[i] = node_distance(node, i); +} + +static int find_nearest_node(int *node_dist, bool *used) +{ + int i, min_dist = node_dist[0], node_id = -1; + + /* Choose the first unused node to compare */ + for (i = 0; i < nr_node_ids; i++) { + if (used[i] == 0) { + min_dist = node_dist[i]; + node_id = i; + break; + } + } + + /* Compare and return the nearest node */ + for (i = 0; i < nr_node_ids; i++) { + if (node_dist[i] < min_dist && used[i] == 0) { + min_dist = node_dist[i]; + node_id = i; + } + } + + return node_id; +} + +static unsigned int __cpumask_local_spread(unsigned int i, int node) { int cpu; @@ -231,4 +253,60 @@ unsigned int cpumask_local_spread(unsigned int i, int node) } BUG(); } + +/** + * cpumask_local_spread - select the i'th cpu with local numa cpu's first + * @i: index number + * @node: local numa_node + * + * This function selects an online CPU according to a numa aware policy; + * local cpus are returned first, followed by the nearest non-local ones, + * then it wraps around. + * + * It's not very efficient, but useful for setup. + */ +unsigned int cpumask_local_spread(unsigned int i, int node) +{ + static DEFINE_SPINLOCK(spread_lock); + static int node_dist[MAX_NUMNODES]; + static bool used[MAX_NUMNODES]; + unsigned long flags; + int cpu, j, id; + + /* Wrap: we always want a cpu. */ + i %= num_online_cpus(); + + if (node == NUMA_NO_NODE) { + for_each_cpu(cpu, cpu_online_mask) + if (i-- == 0) + return cpu; + } else { + if (nr_node_ids > MAX_NUMNODES) + return __cpumask_local_spread(i, node); + + spin_lock_irqsave(&spread_lock, flags); + memset(used, 0, nr_node_ids * sizeof(bool)); + calc_node_distance(node_dist, node); + for (j = 0; j < nr_node_ids; j++) { + id = find_nearest_node(node_dist, used); + if (id < 0) + break; + + for_each_cpu_and(cpu, cpumask_of_node(id), + cpu_online_mask) + if (i-- == 0) { + spin_unlock_irqrestore(&spread_lock, + flags); + return cpu; + } + used[id] = 1; + } + spin_unlock_irqrestore(&spread_lock, flags); + + for_each_cpu(cpu, cpu_online_mask) + if (i-- == 0) + return cpu; + } + BUG(); +} EXPORT_SYMBOL(cpumask_local_spread); -- 2.7.4