From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756257AbZCJOXq (ORCPT ); Tue, 10 Mar 2009 10:23:46 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1755910AbZCJOWa (ORCPT ); Tue, 10 Mar 2009 10:22:30 -0400 Received: from bombadil.infradead.org ([18.85.46.34]:48795 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755770AbZCJOW3 (ORCPT ); Tue, 10 Mar 2009 10:22:29 -0400 Message-Id: <20090310142015.761644317@chello.nl> References: <20090310141913.961861111@chello.nl> User-Agent: quilt/0.46-1 Date: Tue, 10 Mar 2009 15:19:15 +0100 From: Peter Zijlstra To: mingo@elte.hu, paulus@samba.org Cc: linux-kernel@vger.kernel.org, Peter Zijlstra Subject: [PATCH 2/4] perf_counter: provide pagefault software events Content-Disposition: inline; filename=perf_swcounter_pgfault.patch X-Bad-Reply: References but no 'Re:' in Subject. Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org We use the generic software counter infrastructure to provide page fault events. Signed-off-by: Peter Zijlstra --- arch/powerpc/mm/fault.c | 3 ++ arch/x86/mm/fault.c | 3 ++ kernel/perf_counter.c | 53 ++---------------------------------------------- 3 files changed, 9 insertions(+), 50 deletions(-) Index: linux-2.6/kernel/perf_counter.c =================================================================== --- linux-2.6.orig/kernel/perf_counter.c +++ linux-2.6/kernel/perf_counter.c @@ -1601,57 +1601,10 @@ static const struct hw_perf_counter_ops * Software counter: page faults */ -#ifdef CONFIG_VM_EVENT_COUNTERS -#define cpu_page_faults() __get_cpu_var(vm_event_states).event[PGFAULT] -#else -#define cpu_page_faults() 0 -#endif - -static u64 get_page_faults(struct perf_counter *counter) -{ - struct task_struct *curr = counter->ctx->task; - - if (curr) - return curr->maj_flt + curr->min_flt; - return cpu_page_faults(); -} - -static void page_faults_perf_counter_update(struct perf_counter *counter) -{ - u64 prev, now; - s64 delta; - - prev = atomic64_read(&counter->hw.prev_count); - now = get_page_faults(counter); - - atomic64_set(&counter->hw.prev_count, now); - - delta = now - prev; - - atomic64_add(delta, &counter->count); -} - -static void page_faults_perf_counter_read(struct perf_counter *counter) -{ - page_faults_perf_counter_update(counter); -} - -static int page_faults_perf_counter_enable(struct perf_counter *counter) -{ - if (counter->prev_state <= PERF_COUNTER_STATE_OFF) - atomic64_set(&counter->hw.prev_count, get_page_faults(counter)); - return 0; -} - -static void page_faults_perf_counter_disable(struct perf_counter *counter) -{ - page_faults_perf_counter_update(counter); -} - static const struct hw_perf_counter_ops perf_ops_page_faults = { - .enable = page_faults_perf_counter_enable, - .disable = page_faults_perf_counter_disable, - .read = page_faults_perf_counter_read, + .enable = perf_swcounter_enable, + .disable = perf_swcounter_disable, + .read = perf_swcounter_read, }; /* Index: linux-2.6/arch/powerpc/mm/fault.c =================================================================== --- linux-2.6.orig/arch/powerpc/mm/fault.c +++ linux-2.6/arch/powerpc/mm/fault.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -170,6 +171,8 @@ int __kprobes do_page_fault(struct pt_re die("Weird page fault", regs, SIGSEGV); } + perf_swcount_event(PERF_COUNT_PAGE_FAULTS, 1, regs); + /* When running in the kernel we expect faults to occur only to * addresses in user space. All other faults represent errors in the * kernel and should generate an OOPS. Unfortunately, in the case of an Index: linux-2.6/arch/x86/mm/fault.c =================================================================== --- linux-2.6.orig/arch/x86/mm/fault.c +++ linux-2.6/arch/x86/mm/fault.c @@ -27,6 +27,7 @@ #include #include #include +#include #include @@ -1056,6 +1057,8 @@ do_page_fault(struct pt_regs *regs, unsi if (unlikely(error_code & PF_RSVD)) pgtable_bad(regs, error_code, address); + perf_swcounter_event(PERF_COUNT_PAGE_FAULTS, 1, regs); + /* * If we're in an interrupt, have no user context or are running * in an atomic region then we must not take the fault: --