From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: + lib-optimize-cpumask_local_spread.patch added to -mm tree Date: Wed, 26 Feb 2020 19:50:46 -0800 Message-ID: <20200227035046.QC-oYV8zN%akpm@linux-foundation.org> References: <20200203173311.6269a8be06a05e5a4aa08a93@linux-foundation.org> Reply-To: linux-kernel@vger.kernel.org Return-path: Received: from mail.kernel.org ([198.145.29.99]:53278 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728221AbgB0Dus (ORCPT ); Wed, 26 Feb 2020 22:50:48 -0500 In-Reply-To: <20200203173311.6269a8be06a05e5a4aa08a93@linux-foundation.org> Sender: mm-commits-owner@vger.kernel.org List-Id: mm-commits@vger.kernel.org To: anshuman.khandual@arm.com, jgross@suse.com, jinyuqi@huawei.com, mgorman@techsingularity.net, mhocko@suse.com, mingo@elte.hu, mm-commits@vger.kernel.org, mpe@ellerman.id.au, paul.burton@mips.com, peterz@infradead.org, rppt@linux.ibm.com, rusty@rustcorp.com.au, zhangshaokun@hisilicon.com The patch titled Subject: lib: optimize cpumask_local_spread() has been added to the -mm tree. Its filename is lib-optimize-cpumask_local_spread.patch This patch should soon appear at http://ozlabs.org/~akpm/mmots/broken-out/lib-optimize-cpumask_local_spread.patch and later at http://ozlabs.org/~akpm/mmotm/broken-out/lib-optimize-cpumask_local_spread.patch Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next and is updated there every 3-4 working days ------------------------------------------------------ From: yuqi jin Subject: lib: optimize cpumask_local_spread() 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. Link: http://lkml.kernel.org/r/1582768688-2314-1-git-send-email-zhangshaokun@hisilicon.com Signed-off-by: yuqi jin Signed-off-by: Shaokun Zhang Cc: Rusty Russell Cc: Juergen Gross Cc: Paul Burton Cc: Michal Hocko Cc: Michael Ellerman Cc: Mike Rapoport Cc: Anshuman Khandual Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Mel Gorman Signed-off-by: Andrew Morton --- lib/cpumask.c | 102 ++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 90 insertions(+), 12 deletions(-) --- a/lib/cpumask.c~lib-optimize-cpumask_local_spread +++ a/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(cpu } #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(unsign } 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); _ Patches currently in -mm which might be from jinyuqi@huawei.com are lib-optimize-cpumask_local_spread.patch