linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Keiichi KII <k-keiichi@bx.jp.nec.com>
To: linux-kernel@vger.kernel.org
Cc: lwoodman@redhat.com, linux-mm@kvack.org, mingo@elte.hu,
	tzanussi@gmail.com, riel@redhat.com, rostedt@goodmis.org,
	akpm@linux-foundation.org, fweisbec@gmail.com,
	Munehiro Ikeda <m-ikeda@ds.jp.nec.com>,
	Atsushi Tsuji <a-tsuji@bk.jp.nec.com>
Subject: [RFC PATCH -tip 2/2 v2] add a scripts for pagecache usage per process
Date: Fri, 22 Jan 2010 19:08:57 -0500	[thread overview]
Message-ID: <4B5A3E19.6060502@bx.jp.nec.com> (raw)
In-Reply-To: <4B5A3D00.8080901@bx.jp.nec.com>

The scripts are implemented based on the trace stream scripting support.
And the scripts implement the following.
 - how many pagecaches each process has per each file
 - how many pages are cached per each file
 - how many pagecaches each process shares

To monitor pagecache usage per a process, run "pagecache-usage-record" to
record perf data for "pagecache-usage.pl" and run "pagecache-usage-report"
to display.

The below outputs show execution sample.

[file list]
        device      inode   caches
  --------------------------------
         253:0    1051413      130
         253:0    1051399        2
         253:0    1051414       44
         253:0    1051417      154

[process list]
o postmaster-2330
                            cached    added  removed      indirect
        device      inode    pages    pages    pages removed pages
  ----------------------------------------------------------------
         253:0    1051399        0        2        0             0
         253:0    1051417      154        0        0             0
         253:0    1051413      130        0        0             0
         253:0    1051414       44        0        0             0
  ----------------------------------------------------------------
  total:                       337        2        0             0

>From the output, we can know some information like:

- if "added pages" > "cached pages" on process list then
    It means repeating add/remove pagecache many times.
  => Bad case for pagecache usage

- if "added pages" <= "cached pages" on process list then
    It means no unnecessary I/O operations.
  => Good case for pagecache usage.

- if "caches" on file list > 
         sum "cached pages" per each file on process list then
    It means there are unneccessary pagecaches in the memory. 
  => Bad case for pagecache usage

Signed-off-by: Keiichi Kii <k-keiichi@bx.jp.nec.com>
Cc: Atsushi Tsuji <a-tsuji@bk.jp.nec.com>
---
 tools/perf/scripts/perl/bin/pagecache-usage-record |    7 
 tools/perf/scripts/perl/bin/pagecache-usage-report |    6 
 tools/perf/scripts/perl/pagecache-usage.pl         |  160 +++++++++++++++++++++
 3 files changed, 173 insertions(+)

Index: linux-2.6-tip/tools/perf/scripts/perl/bin/pagecache-usage-record
===================================================================
--- /dev/null
+++ linux-2.6-tip/tools/perf/scripts/perl/bin/pagecache-usage-record
@@ -0,0 +1,7 @@
+#!/bin/bash
+perf record -c 1 -f -a -M -R -e filemap:add_to_page_cache -e filemap:find_get_page -e filemap:remove_from_page_cache
+
+
+
+
+
Index: linux-2.6-tip/tools/perf/scripts/perl/bin/pagecache-usage-report
===================================================================
--- /dev/null
+++ linux-2.6-tip/tools/perf/scripts/perl/bin/pagecache-usage-report
@@ -0,0 +1,6 @@
+#!/bin/bash
+# description: pagecache usage per process
+perf trace -s ~/libexec/perf-core/scripts/perl/pagecache-usage.pl
+
+
+
Index: linux-2.6-tip/tools/perf/scripts/perl/pagecache-usage.pl
===================================================================
--- /dev/null
+++ linux-2.6-tip/tools/perf/scripts/perl/pagecache-usage.pl
@@ -0,0 +1,160 @@
+# perf trace event handlers, generated by perf trace -g perl
+# Licensed under the terms of the GNU GPL License version 2
+
+# The common_* event handler fields are the most useful fields common to
+# all events.  They don't necessarily correspond to the 'common_*' fields
+# in the format files.  Those fields not available as handler params can
+# be retrieved using Perl functions of the form common_*($context).
+# See Context.pm for the list of available functions.
+
+use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib";
+use lib "./Perf-Trace-Util/lib";
+use Perf::Trace::Core;
+use Perf::Trace::Context;
+use Perf::Trace::Util;
+use List::Util qw/sum/;
+my %files;
+my %processes;
+my %records;
+
+sub trace_end
+{
+	print_pagecache_usage_per_file();
+	print "\n";
+	print_pagecache_usage_per_process();
+	print_unhandled();
+}
+
+sub filemap::remove_from_page_cache
+{
+	my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs,
+	    $common_pid, $common_comm,
+	    $s_dev, $i_ino, $offset) = @_;
+	my $f = \%{$files{$s_dev}{$i_ino}};
+	my $r = \%{$records{$common_comm."-".$common_pid}{$f}};
+
+	delete $$f{$offset};
+	$$r{inode} = $i_ino;
+	$$r{dev} = $s_dev;
+	if (exists $$r{added}{$offset}) {
+	    $$r{removed}++;
+	} else {
+	    $$r{indirect_removed}++;
+	}
+}
+
+sub filemap::add_to_page_cache
+{
+	my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs,
+	    $common_pid, $common_comm,
+	    $s_dev, $i_ino, $offset) = @_;
+	my $f = \%{$files{$s_dev}{$i_ino}};
+	my $r = \%{$records{$common_comm."-".$common_pid}{$f}};
+
+	$$f{$offset}++;
+	$$r{added}{$offset}++;
+	$$r{inode} = $i_ino;
+	$$r{dev} = $s_dev;
+}
+
+sub filemap::find_get_page
+{
+	my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs,
+	    $common_pid, $common_comm,
+	    $s_dev, $i_ino, $offset, $page) = @_;
+	my $f = \%{$files{$s_dev}{$i_ino}};
+	my $r = \%{$records{$common_comm."-".$common_pid}{$f}};
+
+	if ($page != 0) {
+	    $$f{$offset}++;
+	    $$r{cached}++;
+	    $$r{inode} = $i_ino;
+	    $$r{dev} = $s_dev;
+	}
+}
+
+my %unhandled;
+
+sub trace_unhandled
+{
+	my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs,
+	    $common_pid, $common_comm) = @_;
+
+	$unhandled{$event_name}++;
+}
+
+sub print_unhandled
+{
+	if ((scalar keys %unhandled) == 0) {
+	    print "unhandled events nothing\n";
+	    return;
+	}
+
+	print "\nunhandled events:\n\n";
+
+	printf("%-40s  %10s\n", "event", "count");
+	printf("%-40s  %10s\n", "----------------------------------------",
+	       "-----------");
+
+	foreach my $event_name (keys %unhandled) {
+	    printf("%-40s  %10d\n", $event_name, $unhandled{$event_name});
+	}
+}
+
+sub minor
+{
+	my $dev = shift;
+	return $dev & ((1 << 20) - 1);
+}
+
+sub major
+{
+	my $dev = shift;
+	return $dev >> 20;
+}
+
+sub print_pagecache_usage_per_file
+{
+	print "[file list]\n";
+	printf("  %12s %10s %8s\n", "", "", "cached");
+	printf("  %12s %10s %8s\n", "device", "inode", "pages");
+	printf("  %s\n", '-' x 32);
+	while(my($dev, $file) = each(%files)) {
+	    while(my($inode, $r) = each(%$file)) {
+		my $count = values %$r;
+		next if $count == 0;
+		printf("  %12s %10d %8d\n",
+		       major($dev).":".minor($dev), $inode, $count);
+	    }
+	}
+}
+
+sub print_pagecache_usage_per_process
+{
+	print "[process list]\n";
+	while(my ($pid, $v) = each(%records)) {
+	    my ($sum_cached, $sum_added, $sum_removed, $sum_indirect_removed);
+
+	    print "o $pid\n";
+	    printf("  %12s %10s %8s %8s %8s %13s\n", "", "",
+		   "cached", "added", "removed", "indirect");
+	    printf("  %12s %10s %8s %8s %8s %13s\n", "device", "inode",
+		   "pages", "pages", "pages", "removed pages");
+	    printf("  %s\n", '-' x 64);
+	    while(my ($file, $r) = each(%$v)) {
+		my $added_num = List::Util::sum(values %{$$r{added}});
+		$sum_cached += $$r{cached};
+		$sum_added += $added_num;
+		$sum_removed += $$r{removed};
+		$sum_indirect_removed += $$r{indirect_removed};
+		printf("  %12s %10d %8d %8d %8d %13d\n",
+		       major($$r{dev}).":".minor($$r{dev}), $$r{inode},
+		       $$r{cached}, $added_num, $$r{removed},
+		       $$r{indirect_removed});
+	    }
+	    printf("  %s\n", '-' x 64);
+	    printf("  total: %5s %10s %8d %8d %8d %13d\n", "", "", $sum_cached,
+		   $sum_added, $sum_removed, $sum_indirect_removed);
+	    print "\n";
+	}
+}



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2010-01-23  0:11 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-23  0:04 [RFC PATCH -tip 0/2 v2] pagecache tracepoints proposal Keiichi KII
2010-01-23  0:07 ` [RFC PATCH -tip 1/2 v2] add tracepoints for pagecache Keiichi KII
2010-01-23  2:28   ` Steven Rostedt
2010-01-25 22:17     ` Keiichi KII
2010-01-23  0:08 ` Keiichi KII [this message]
2010-01-23  8:21   ` [RFC PATCH -tip 2/2 v2] add a scripts for pagecache usage per process Tom Zanussi
2010-01-25 22:16     ` Keiichi KII
2010-02-01  8:17       ` Tom Zanussi
2010-02-01 21:20         ` Keiichi KII
2010-02-23 17:54         ` Frederic Weisbecker
2010-02-23 18:13           ` Peter Zijlstra

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=4B5A3E19.6060502@bx.jp.nec.com \
    --to=k-keiichi@bx.jp.nec.com \
    --cc=a-tsuji@bk.jp.nec.com \
    --cc=akpm@linux-foundation.org \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=lwoodman@redhat.com \
    --cc=m-ikeda@ds.jp.nec.com \
    --cc=mingo@elte.hu \
    --cc=riel@redhat.com \
    --cc=rostedt@goodmis.org \
    --cc=tzanussi@gmail.com \
    /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