Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759191AbXELR6W (ORCPT ); Sat, 12 May 2007 13:58:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752595AbXELR6Q (ORCPT ); Sat, 12 May 2007 13:58:16 -0400 Received: from mx1.suse.de ([195.135.220.2]:42037 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751208AbXELR6P (ORCPT ); Sat, 12 May 2007 13:58:15 -0400 To: Andrew Morton Cc: Christoph Lameter , Peter Zijlstra , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Oleg Nesterov , Ingo Molnar , Thomas Gleixner , Nick Piggin Subject: Re: [PATCH 1/2] scalable rw_mutex References: <20070511131541.992688403@chello.nl> <20070511132321.895740140@chello.nl> <20070511093108.495feb70.akpm@linux-foundation.org> <20070511110522.ed459635.akpm@linux-foundation.org> From: Andi Kleen Date: 12 May 2007 20:55:28 +0200 In-Reply-To: <20070511110522.ed459635.akpm@linux-foundation.org> Message-ID: User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4085 Lines: 102 Andrew Morton writes: > On Fri, 11 May 2007 10:07:17 -0700 (PDT) > Christoph Lameter wrote: > > > On Fri, 11 May 2007, Andrew Morton wrote: > > > > > yipes. percpu_counter_sum() is expensive. > > > > Capable of triggering NMI watchdog on 4096+ processors? > > Well. That would be a millisecond per cpu which sounds improbable. And > we'd need to be calling it under local_irq_save() which we presently don't. > And nobody has reported any problems against the existing callsites. > > But it's no speed demon, that's for sure. There is one possible optimization for this I did some time ago. You don't really need to sum all over the possible map, but only all CPUs that were ever online. But this only helps on systems where the possible map is bigger than online map in the common case. But that shouldn't be the case anymore on x86 -- it just used to be. If it's true on some other architectures it might be still worth it. Old patches with an network example for reference appended. -Andi Index: linux-2.6.21-git2-net/include/linux/cpumask.h =================================================================== --- linux-2.6.21-git2-net.orig/include/linux/cpumask.h +++ linux-2.6.21-git2-net/include/linux/cpumask.h @@ -380,6 +380,7 @@ static inline void __cpus_remap(cpumask_ extern cpumask_t cpu_possible_map; extern cpumask_t cpu_online_map; extern cpumask_t cpu_present_map; +extern cpumask_t cpu_everonline_map; #if NR_CPUS > 1 #define num_online_cpus() cpus_weight(cpu_online_map) @@ -388,6 +389,7 @@ extern cpumask_t cpu_present_map; #define cpu_online(cpu) cpu_isset((cpu), cpu_online_map) #define cpu_possible(cpu) cpu_isset((cpu), cpu_possible_map) #define cpu_present(cpu) cpu_isset((cpu), cpu_present_map) +#define cpu_ever_online(cpu) cpu_isset((cpu), cpu_everonline_map) #else #define num_online_cpus() 1 #define num_possible_cpus() 1 @@ -395,6 +397,7 @@ extern cpumask_t cpu_present_map; #define cpu_online(cpu) ((cpu) == 0) #define cpu_possible(cpu) ((cpu) == 0) #define cpu_present(cpu) ((cpu) == 0) +#define cpu_ever_online(cpu) ((cpu) == 0) #endif #ifdef CONFIG_SMP @@ -409,5 +412,6 @@ int __any_online_cpu(const cpumask_t *ma #define for_each_possible_cpu(cpu) for_each_cpu_mask((cpu), cpu_possible_map) #define for_each_online_cpu(cpu) for_each_cpu_mask((cpu), cpu_online_map) #define for_each_present_cpu(cpu) for_each_cpu_mask((cpu), cpu_present_map) +#define for_each_everonline_cpu(cpu) for_each_cpu_mask((cpu), cpu_everonline_map) #endif /* __LINUX_CPUMASK_H */ Index: linux-2.6.21-git2-net/kernel/cpu.c =================================================================== --- linux-2.6.21-git2-net.orig/kernel/cpu.c +++ linux-2.6.21-git2-net/kernel/cpu.c @@ -26,6 +26,10 @@ static __cpuinitdata RAW_NOTIFIER_HEAD(c */ static int cpu_hotplug_disabled; +/* Contains any CPUs that were ever online at some point. + No guarantee they were fully initialized though */ +cpumask_t cpu_everonline_map; + #ifdef CONFIG_HOTPLUG_CPU /* Crappy recursive lock-takers in cpufreq! Complain loudly about idiots */ @@ -212,6 +216,8 @@ static int __cpuinit _cpu_up(unsigned in if (cpu_online(cpu) || !cpu_present(cpu)) return -EINVAL; + cpu_set(cpu, cpu_everonline_map); + ret = raw_notifier_call_chain(&cpu_chain, CPU_UP_PREPARE, hcpu); if (ret == NOTIFY_BAD) { printk("%s: attempt to bring up CPU %u failed\n", Index: linux-2.6.21-git2-net/net/ipv4/proc.c =================================================================== --- linux-2.6.21-git2-net.orig/net/ipv4/proc.c +++ linux-2.6.21-git2-net/net/ipv4/proc.c @@ -50,7 +50,7 @@ static int fold_prot_inuse(struct proto int res = 0; int cpu; - for_each_possible_cpu(cpu) + for_each_everonline_cpu(cpu) res += proto->stats[cpu].inuse; return res; - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/