From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: linux-tip-commits@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, hpa@zytor.com, mingo@redhat.com,
a.p.zijlstra@chello.nl, tglx@linutronix.de, mingo@elte.hu
Subject: [tip:perfcounters/core] perf_counter: provide major/minor page fault software events
Date: Fri, 13 Mar 2009 13:07:19 GMT [thread overview]
Message-ID: <tip-07904cb8292bad6c020a1a3de499a58969da1624@git.kernel.org> (raw)
In-Reply-To: <20090313112301.770243471@chello.nl>
Commit-ID: 07904cb8292bad6c020a1a3de499a58969da1624
Gitweb: http://git.kernel.org/tip/07904cb8292bad6c020a1a3de499a58969da1624
Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
AuthorDate: Fri, 13 Mar 2009 12:21:34 +0100
Commit: Ingo Molnar <mingo@elte.hu>
CommitDate: Fri, 13 Mar 2009 14:02:45 +0100
perf_counter: provide major/minor page fault software events
Provide separate sw counters for major and minor page faults.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20090313112301.770243471@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
arch/powerpc/mm/fault.c | 5 ++++-
arch/x86/mm/fault.c | 7 +++++--
include/linux/perf_counter.h | 4 +++-
kernel/perf_counter.c | 22 +++++++++-------------
4 files changed, 21 insertions(+), 17 deletions(-)
diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
index b044dc7..de37a3a 100644
--- a/arch/powerpc/mm/fault.c
+++ b/arch/powerpc/mm/fault.c
@@ -324,6 +324,7 @@ good_area:
}
if (ret & VM_FAULT_MAJOR) {
current->maj_flt++;
+ perf_swcounter_event(PERF_COUNT_PAGE_FAULTS_MAJ, 1, 0, regs);
#ifdef CONFIG_PPC_SMLPAR
if (firmware_has_feature(FW_FEATURE_CMO)) {
preempt_disable();
@@ -331,8 +332,10 @@ good_area:
preempt_enable();
}
#endif
- } else
+ } else {
current->min_flt++;
+ perf_swcounter_event(PERF_COUNT_PAGE_FAULTS_MIN, 1, 0, regs);
+ }
up_read(&mm->mmap_sem);
return 0;
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index c872575..f2d3324 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -1140,10 +1140,13 @@ good_area:
return;
}
- if (fault & VM_FAULT_MAJOR)
+ if (fault & VM_FAULT_MAJOR) {
tsk->maj_flt++;
- else
+ perf_swcounter_event(PERF_COUNT_PAGE_FAULTS_MAJ, 1, 0, regs);
+ } else {
tsk->min_flt++;
+ perf_swcounter_event(PERF_COUNT_PAGE_FAULTS_MIN, 1, 0, regs);
+ }
check_v8086_mode(regs, address, tsk);
diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h
index 3fefc3b..4b14a8e 100644
--- a/include/linux/perf_counter.h
+++ b/include/linux/perf_counter.h
@@ -49,8 +49,10 @@ enum hw_event_types {
PERF_COUNT_PAGE_FAULTS = -3,
PERF_COUNT_CONTEXT_SWITCHES = -4,
PERF_COUNT_CPU_MIGRATIONS = -5,
+ PERF_COUNT_PAGE_FAULTS_MIN = -6,
+ PERF_COUNT_PAGE_FAULTS_MAJ = -7,
- PERF_SW_EVENTS_MIN = -6,
+ PERF_SW_EVENTS_MIN = -8,
};
/*
diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c
index 1773c5d..68950a3 100644
--- a/kernel/perf_counter.c
+++ b/kernel/perf_counter.c
@@ -1503,6 +1503,12 @@ static void perf_swcounter_disable(struct perf_counter *counter)
perf_swcounter_update(counter);
}
+static const struct hw_perf_counter_ops perf_ops_generic = {
+ .enable = perf_swcounter_enable,
+ .disable = perf_swcounter_disable,
+ .read = perf_swcounter_read,
+};
+
/*
* Software counter: cpu wall time clock
*/
@@ -1604,16 +1610,6 @@ static const struct hw_perf_counter_ops perf_ops_task_clock = {
};
/*
- * Software counter: page faults
- */
-
-static const struct hw_perf_counter_ops perf_ops_page_faults = {
- .enable = perf_swcounter_enable,
- .disable = perf_swcounter_disable,
- .read = perf_swcounter_read,
-};
-
-/*
* Software counter: context switches
*/
@@ -1753,9 +1749,9 @@ sw_perf_counter_init(struct perf_counter *counter)
hw_ops = &perf_ops_cpu_clock;
break;
case PERF_COUNT_PAGE_FAULTS:
- if (!(counter->hw_event.exclude_user ||
- counter->hw_event.exclude_kernel))
- hw_ops = &perf_ops_page_faults;
+ case PERF_COUNT_PAGE_FAULTS_MIN:
+ case PERF_COUNT_PAGE_FAULTS_MAJ:
+ hw_ops = &perf_ops_generic;
break;
case PERF_COUNT_CONTEXT_SWITCHES:
if (!counter->hw_event.exclude_kernel)
next prev parent reply other threads:[~2009-03-13 13:09 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-13 11:21 [PATCH 00/11] generic software counters -v2 Peter Zijlstra
2009-03-13 11:21 ` [PATCH 01/11] sched: remove extra call overhead for schedule() Peter Zijlstra
2009-03-13 13:00 ` [tip:core/locking] " Peter Zijlstra
2009-04-20 19:00 ` [tip:sched/core] " tip-bot for Peter Zijlstra
2009-03-13 11:21 ` [PATCH 02/11] hrtimer: fix rq->lock inversion (again) Peter Zijlstra
2009-03-13 13:00 ` [tip:core/locking] " Peter Zijlstra
2009-03-13 13:27 ` Peter Zijlstra
2009-03-13 14:57 ` Peter Zijlstra
2009-03-31 12:57 ` Peter Zijlstra
2009-04-02 19:45 ` [tip:timers/urgent] " Peter Zijlstra
2009-03-13 11:21 ` [PATCH 03/11] perf_counter: x86: fix 32bit irq_period assumption Peter Zijlstra
2009-03-13 13:00 ` [tip:perfcounters/core] perf_counter: x86: fix 32-bit " Peter Zijlstra
2009-03-13 13:06 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 04/11] perf_counter: use list_move_tail Peter Zijlstra
2009-03-13 13:00 ` [tip:perfcounters/core] perf_counter: use list_move_tail() Peter Zijlstra
2009-03-13 13:06 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 05/11] perf_counter: add comment to barrier Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:06 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 06/11] perf_counter: x86: use ULL postfix for 64bit constants Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:06 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 07/11] perf_counter: software counter event infrastructure Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:07 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 08/11] perf_counter: provide pagefault software events Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:07 ` Peter Zijlstra
2009-03-13 11:21 ` [PATCH 09/11] perf_counter: provide major/minor page fault " Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:07 ` Peter Zijlstra [this message]
2009-03-13 11:21 ` [PATCH 10/11] perf_counter: hrtimer based sampling for software time events Peter Zijlstra
2009-03-13 13:01 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:07 ` Peter Zijlstra
2009-03-13 15:43 ` [PATCH 10.5/11] perf_counter: fix hrtimer sampling Peter Zijlstra
2009-03-13 16:09 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 11:21 ` [PATCH 11/11] perf_counter: add an event_list Peter Zijlstra
2009-03-13 13:02 ` [tip:perfcounters/core] " Peter Zijlstra
2009-03-13 13:07 ` Peter Zijlstra
2009-03-13 13:07 ` [PATCH 00/11] generic software counters -v2 Ingo Molnar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=tip-07904cb8292bad6c020a1a3de499a58969da1624@git.kernel.org \
--to=a.p.zijlstra@chello.nl \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-tip-commits@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=mingo@redhat.com \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox