2010-02-03 03:53:14 +00:00
|
|
|
#define _FILE_OFFSET_BITS 64
|
|
|
|
|
2011-10-26 14:41:38 +00:00
|
|
|
#include "util.h"
|
2009-06-25 15:05:54 +00:00
|
|
|
#include <sys/types.h>
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
#include <byteswap.h>
|
2009-06-25 15:05:54 +00:00
|
|
|
#include <unistd.h>
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
2009-11-11 03:51:03 +00:00
|
|
|
#include <linux/list.h>
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
#include <linux/kernel.h>
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
#include <sys/utsname.h>
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-01-11 22:56:53 +00:00
|
|
|
#include "evlist.h"
|
2011-03-10 14:15:54 +00:00
|
|
|
#include "evsel.h"
|
2009-06-25 15:05:54 +00:00
|
|
|
#include "header.h"
|
2009-10-06 21:36:47 +00:00
|
|
|
#include "../perf.h"
|
|
|
|
#include "trace-event.h"
|
2009-12-13 21:50:25 +00:00
|
|
|
#include "session.h"
|
2009-11-11 03:51:03 +00:00
|
|
|
#include "symbol.h"
|
2009-11-11 03:51:05 +00:00
|
|
|
#include "debug.h"
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
#include "cpumap.h"
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2010-06-17 09:39:01 +00:00
|
|
|
static bool no_buildid_cache = false;
|
|
|
|
|
2009-09-12 05:52:51 +00:00
|
|
|
static int event_count;
|
|
|
|
static struct perf_trace_event_type *events;
|
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
static u32 header_argc;
|
|
|
|
static const char **header_argv;
|
|
|
|
|
2009-12-29 00:48:33 +00:00
|
|
|
int perf_header__push_event(u64 id, const char *name)
|
2009-09-12 05:52:51 +00:00
|
|
|
{
|
2009-09-19 11:36:30 +00:00
|
|
|
if (strlen(name) > MAX_EVENT_NAME)
|
2009-10-21 19:34:06 +00:00
|
|
|
pr_warning("Event %s will be truncated\n", name);
|
2009-09-12 05:52:51 +00:00
|
|
|
|
|
|
|
if (!events) {
|
|
|
|
events = malloc(sizeof(struct perf_trace_event_type));
|
2009-12-29 00:48:33 +00:00
|
|
|
if (events == NULL)
|
|
|
|
return -ENOMEM;
|
2009-09-12 05:52:51 +00:00
|
|
|
} else {
|
2009-12-29 00:48:33 +00:00
|
|
|
struct perf_trace_event_type *nevents;
|
|
|
|
|
|
|
|
nevents = realloc(events, (event_count + 1) * sizeof(*events));
|
|
|
|
if (nevents == NULL)
|
|
|
|
return -ENOMEM;
|
|
|
|
events = nevents;
|
2009-09-12 05:52:51 +00:00
|
|
|
}
|
|
|
|
memset(&events[event_count], 0, sizeof(struct perf_trace_event_type));
|
|
|
|
events[event_count].event_id = id;
|
2009-09-19 11:36:30 +00:00
|
|
|
strncpy(events[event_count].name, name, MAX_EVENT_NAME - 1);
|
2009-09-12 05:52:51 +00:00
|
|
|
event_count++;
|
2009-12-29 00:48:33 +00:00
|
|
|
return 0;
|
2009-09-12 05:52:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
char *perf_header__find_event(u64 id)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0 ; i < event_count; i++) {
|
|
|
|
if (events[i].event_id == id)
|
|
|
|
return events[i].name;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2009-06-25 15:05:54 +00:00
|
|
|
static const char *__perf_magic = "PERFFILE";
|
|
|
|
|
|
|
|
#define PERF_MAGIC (*(u64 *)__perf_magic)
|
|
|
|
|
|
|
|
struct perf_file_attr {
|
perf: Do the big rename: Performance Counters -> Performance Events
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 10:02:48 +00:00
|
|
|
struct perf_event_attr attr;
|
2009-06-25 15:05:54 +00:00
|
|
|
struct perf_file_section ids;
|
|
|
|
};
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
void perf_header__set_feat(struct perf_header *header, int feat)
|
perf symbols: Use the buildids if present
With this change 'perf record' will intercept PERF_RECORD_MMAP
calls, creating a linked list of DSOs, then when the session
finishes, it will traverse this list and read the buildids,
stashing them at the end of the file and will set up a new
feature bit in the header bitmask.
'perf report' will then notice this feature and populate the
'dsos' list and set the build ids.
When reading the symtabs it will refuse to load from a file that
doesn't have the same build id. This improves the
reliability of the profiler output, as symbols and profiling
data is more guaranteed to match.
Example:
[root@doppio ~]# perf report | head
/home/acme/bin/perf with build id b1ea544ac3746e7538972548a09aadecc5753868 not found, continuing without symbols
# Samples: 2621434559
#
# Overhead Command Shared Object Symbol
# ........ ............... ............................. ......
#
7.91% init [kernel] [k] read_hpet
7.64% init [kernel] [k] mwait_idle_with_hints
7.60% swapper [kernel] [k] read_hpet
7.60% swapper [kernel] [k] mwait_idle_with_hints
3.65% init [kernel] [k] 0xffffffffa02339d9
[root@doppio ~]#
In this case the 'perf' binary was an older one, vanished,
so its symbols probably wouldn't match or would cause subtly
different (and misleading) output.
Next patches will support the kernel as well, reading the build
id notes for it and the modules from /sys.
Another patch should also introduce a new plumbing command:
'perf list-buildids'
that will then be used in porcelain that is distro specific to
fetch -debuginfo packages where such buildids are present. This
will in turn allow for one to run 'perf record' in one machine
and 'perf report' in another.
Future work on having the buildid sent directly from the kernel
in the PERF_RECORD_MMAP event is needed to close races, as the
DSO can be changed during a 'perf record' session, but this
patch at least helps with non-corner cases and current/older
kernels.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Frank Ch. Eigler <fche@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: K. Prasad <prasad@linux.vnet.ibm.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <1257367843-26224-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-04 20:50:43 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
set_bit(feat, header->adds_features);
|
perf symbols: Use the buildids if present
With this change 'perf record' will intercept PERF_RECORD_MMAP
calls, creating a linked list of DSOs, then when the session
finishes, it will traverse this list and read the buildids,
stashing them at the end of the file and will set up a new
feature bit in the header bitmask.
'perf report' will then notice this feature and populate the
'dsos' list and set the build ids.
When reading the symtabs it will refuse to load from a file that
doesn't have the same build id. This improves the
reliability of the profiler output, as symbols and profiling
data is more guaranteed to match.
Example:
[root@doppio ~]# perf report | head
/home/acme/bin/perf with build id b1ea544ac3746e7538972548a09aadecc5753868 not found, continuing without symbols
# Samples: 2621434559
#
# Overhead Command Shared Object Symbol
# ........ ............... ............................. ......
#
7.91% init [kernel] [k] read_hpet
7.64% init [kernel] [k] mwait_idle_with_hints
7.60% swapper [kernel] [k] read_hpet
7.60% swapper [kernel] [k] mwait_idle_with_hints
3.65% init [kernel] [k] 0xffffffffa02339d9
[root@doppio ~]#
In this case the 'perf' binary was an older one, vanished,
so its symbols probably wouldn't match or would cause subtly
different (and misleading) output.
Next patches will support the kernel as well, reading the build
id notes for it and the modules from /sys.
Another patch should also introduce a new plumbing command:
'perf list-buildids'
that will then be used in porcelain that is distro specific to
fetch -debuginfo packages where such buildids are present. This
will in turn allow for one to run 'perf record' in one machine
and 'perf report' in another.
Future work on having the buildid sent directly from the kernel
in the PERF_RECORD_MMAP event is needed to close races, as the
DSO can be changed during a 'perf record' session, but this
patch at least helps with non-corner cases and current/older
kernels.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Frank Ch. Eigler <fche@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: K. Prasad <prasad@linux.vnet.ibm.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <1257367843-26224-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-04 20:50:43 +00:00
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
void perf_header__clear_feat(struct perf_header *header, int feat)
|
2010-11-26 21:39:15 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
clear_bit(feat, header->adds_features);
|
2010-11-26 21:39:15 +00:00
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
bool perf_header__has_feat(const struct perf_header *header, int feat)
|
perf symbols: Use the buildids if present
With this change 'perf record' will intercept PERF_RECORD_MMAP
calls, creating a linked list of DSOs, then when the session
finishes, it will traverse this list and read the buildids,
stashing them at the end of the file and will set up a new
feature bit in the header bitmask.
'perf report' will then notice this feature and populate the
'dsos' list and set the build ids.
When reading the symtabs it will refuse to load from a file that
doesn't have the same build id. This improves the
reliability of the profiler output, as symbols and profiling
data is more guaranteed to match.
Example:
[root@doppio ~]# perf report | head
/home/acme/bin/perf with build id b1ea544ac3746e7538972548a09aadecc5753868 not found, continuing without symbols
# Samples: 2621434559
#
# Overhead Command Shared Object Symbol
# ........ ............... ............................. ......
#
7.91% init [kernel] [k] read_hpet
7.64% init [kernel] [k] mwait_idle_with_hints
7.60% swapper [kernel] [k] read_hpet
7.60% swapper [kernel] [k] mwait_idle_with_hints
3.65% init [kernel] [k] 0xffffffffa02339d9
[root@doppio ~]#
In this case the 'perf' binary was an older one, vanished,
so its symbols probably wouldn't match or would cause subtly
different (and misleading) output.
Next patches will support the kernel as well, reading the build
id notes for it and the modules from /sys.
Another patch should also introduce a new plumbing command:
'perf list-buildids'
that will then be used in porcelain that is distro specific to
fetch -debuginfo packages where such buildids are present. This
will in turn allow for one to run 'perf record' in one machine
and 'perf report' in another.
Future work on having the buildid sent directly from the kernel
in the PERF_RECORD_MMAP event is needed to close races, as the
DSO can be changed during a 'perf record' session, but this
patch at least helps with non-corner cases and current/older
kernels.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Frank Ch. Eigler <fche@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: K. Prasad <prasad@linux.vnet.ibm.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <1257367843-26224-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-04 20:50:43 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
return test_bit(feat, header->adds_features);
|
perf symbols: Use the buildids if present
With this change 'perf record' will intercept PERF_RECORD_MMAP
calls, creating a linked list of DSOs, then when the session
finishes, it will traverse this list and read the buildids,
stashing them at the end of the file and will set up a new
feature bit in the header bitmask.
'perf report' will then notice this feature and populate the
'dsos' list and set the build ids.
When reading the symtabs it will refuse to load from a file that
doesn't have the same build id. This improves the
reliability of the profiler output, as symbols and profiling
data is more guaranteed to match.
Example:
[root@doppio ~]# perf report | head
/home/acme/bin/perf with build id b1ea544ac3746e7538972548a09aadecc5753868 not found, continuing without symbols
# Samples: 2621434559
#
# Overhead Command Shared Object Symbol
# ........ ............... ............................. ......
#
7.91% init [kernel] [k] read_hpet
7.64% init [kernel] [k] mwait_idle_with_hints
7.60% swapper [kernel] [k] read_hpet
7.60% swapper [kernel] [k] mwait_idle_with_hints
3.65% init [kernel] [k] 0xffffffffa02339d9
[root@doppio ~]#
In this case the 'perf' binary was an older one, vanished,
so its symbols probably wouldn't match or would cause subtly
different (and misleading) output.
Next patches will support the kernel as well, reading the build
id notes for it and the modules from /sys.
Another patch should also introduce a new plumbing command:
'perf list-buildids'
that will then be used in porcelain that is distro specific to
fetch -debuginfo packages where such buildids are present. This
will in turn allow for one to run 'perf record' in one machine
and 'perf report' in another.
Future work on having the buildid sent directly from the kernel
in the PERF_RECORD_MMAP event is needed to close races, as the
DSO can be changed during a 'perf record' session, but this
patch at least helps with non-corner cases and current/older
kernels.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Frank Ch. Eigler <fche@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: K. Prasad <prasad@linux.vnet.ibm.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
LKML-Reference: <1257367843-26224-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-04 20:50:43 +00:00
|
|
|
}
|
|
|
|
|
2009-11-17 03:18:12 +00:00
|
|
|
static int do_write(int fd, const void *buf, size_t size)
|
2009-06-25 15:05:54 +00:00
|
|
|
{
|
|
|
|
while (size) {
|
|
|
|
int ret = write(fd, buf, size);
|
|
|
|
|
|
|
|
if (ret < 0)
|
2009-11-19 16:55:56 +00:00
|
|
|
return -errno;
|
2009-06-25 15:05:54 +00:00
|
|
|
|
|
|
|
size -= ret;
|
|
|
|
buf += ret;
|
|
|
|
}
|
2009-11-17 03:18:12 +00:00
|
|
|
|
|
|
|
return 0;
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
|
|
|
|
2010-01-04 18:19:28 +00:00
|
|
|
#define NAME_ALIGN 64
|
|
|
|
|
|
|
|
static int write_padded(int fd, const void *bf, size_t count,
|
|
|
|
size_t count_aligned)
|
|
|
|
{
|
|
|
|
static const char zero_buf[NAME_ALIGN];
|
|
|
|
int err = do_write(fd, bf, count);
|
|
|
|
|
|
|
|
if (!err)
|
|
|
|
err = do_write(fd, zero_buf, count_aligned - count);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
static int do_write_string(int fd, const char *str)
|
|
|
|
{
|
|
|
|
u32 len, olen;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
olen = strlen(str) + 1;
|
|
|
|
len = ALIGN(olen, NAME_ALIGN);
|
|
|
|
|
|
|
|
/* write len, incl. \0 */
|
|
|
|
ret = do_write(fd, &len, sizeof(len));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
return write_padded(fd, str, olen, len);
|
|
|
|
}
|
|
|
|
|
|
|
|
static char *do_read_string(int fd, struct perf_header *ph)
|
|
|
|
{
|
|
|
|
ssize_t sz, ret;
|
|
|
|
u32 len;
|
|
|
|
char *buf;
|
|
|
|
|
|
|
|
sz = read(fd, &len, sizeof(len));
|
|
|
|
if (sz < (ssize_t)sizeof(len))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
len = bswap_32(len);
|
|
|
|
|
|
|
|
buf = malloc(len);
|
|
|
|
if (!buf)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
ret = read(fd, buf, len);
|
|
|
|
if (ret == (ssize_t)len) {
|
|
|
|
/*
|
|
|
|
* strings are padded by zeroes
|
|
|
|
* thus the actual strlen of buf
|
|
|
|
* may be less than len
|
|
|
|
*/
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
free(buf);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
perf_header__set_cmdline(int argc, const char **argv)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
header_argc = (u32)argc;
|
|
|
|
|
|
|
|
/* do not include NULL termination */
|
|
|
|
header_argv = calloc(argc, sizeof(char *));
|
|
|
|
if (!header_argv)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* must copy argv contents because it gets moved
|
|
|
|
* around during option parsing
|
|
|
|
*/
|
|
|
|
for (i = 0; i < argc ; i++)
|
|
|
|
header_argv[i] = argv[i];
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-12-07 09:02:53 +00:00
|
|
|
#define dsos__for_each_with_build_id(pos, head) \
|
|
|
|
list_for_each_entry(pos, head, node) \
|
|
|
|
if (!pos->has_build_id) \
|
|
|
|
continue; \
|
|
|
|
else
|
|
|
|
|
|
|
|
static int __dsos__write_buildid_table(struct list_head *head, pid_t pid,
|
|
|
|
u16 misc, int fd)
|
|
|
|
{
|
|
|
|
struct dso *pos;
|
|
|
|
|
|
|
|
dsos__for_each_with_build_id(pos, head) {
|
|
|
|
int err;
|
|
|
|
struct build_id_event b;
|
|
|
|
size_t len;
|
|
|
|
|
|
|
|
if (!pos->hit)
|
|
|
|
continue;
|
|
|
|
len = pos->long_name_len + 1;
|
|
|
|
len = ALIGN(len, NAME_ALIGN);
|
|
|
|
memset(&b, 0, sizeof(b));
|
|
|
|
memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id));
|
|
|
|
b.pid = pid;
|
|
|
|
b.header.misc = misc;
|
|
|
|
b.header.size = sizeof(b) + len;
|
|
|
|
err = do_write(fd, &b, sizeof(b));
|
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
err = write_padded(fd, pos->long_name,
|
|
|
|
pos->long_name_len + 1, len);
|
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int machine__write_buildid_table(struct machine *machine, int fd)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
u16 kmisc = PERF_RECORD_MISC_KERNEL,
|
|
|
|
umisc = PERF_RECORD_MISC_USER;
|
|
|
|
|
|
|
|
if (!machine__is_host(machine)) {
|
|
|
|
kmisc = PERF_RECORD_MISC_GUEST_KERNEL;
|
|
|
|
umisc = PERF_RECORD_MISC_GUEST_USER;
|
|
|
|
}
|
|
|
|
|
|
|
|
err = __dsos__write_buildid_table(&machine->kernel_dsos, machine->pid,
|
|
|
|
kmisc, fd);
|
|
|
|
if (err == 0)
|
|
|
|
err = __dsos__write_buildid_table(&machine->user_dsos,
|
|
|
|
machine->pid, umisc, fd);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dsos__write_buildid_table(struct perf_header *header, int fd)
|
|
|
|
{
|
|
|
|
struct perf_session *session = container_of(header,
|
|
|
|
struct perf_session, header);
|
|
|
|
struct rb_node *nd;
|
|
|
|
int err = machine__write_buildid_table(&session->host_machine, fd);
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
|
|
|
|
struct machine *pos = rb_entry(nd, struct machine, rb_node);
|
|
|
|
err = machine__write_buildid_table(pos, fd);
|
|
|
|
if (err)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
int build_id_cache__add_s(const char *sbuild_id, const char *debugdir,
|
|
|
|
const char *name, bool is_kallsyms)
|
|
|
|
{
|
|
|
|
const size_t size = PATH_MAX;
|
|
|
|
char *realname, *filename = zalloc(size),
|
|
|
|
*linkname = zalloc(size), *targetname;
|
|
|
|
int len, err = -1;
|
|
|
|
|
|
|
|
if (is_kallsyms) {
|
|
|
|
if (symbol_conf.kptr_restrict) {
|
|
|
|
pr_debug("Not caching a kptr_restrict'ed /proc/kallsyms\n");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
realname = (char *)name;
|
|
|
|
} else
|
|
|
|
realname = realpath(name, NULL);
|
|
|
|
|
|
|
|
if (realname == NULL || filename == NULL || linkname == NULL)
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
len = snprintf(filename, size, "%s%s%s",
|
|
|
|
debugdir, is_kallsyms ? "/" : "", realname);
|
|
|
|
if (mkdir_p(filename, 0755))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
snprintf(filename + len, sizeof(filename) - len, "/%s", sbuild_id);
|
|
|
|
|
|
|
|
if (access(filename, F_OK)) {
|
|
|
|
if (is_kallsyms) {
|
|
|
|
if (copyfile("/proc/kallsyms", filename))
|
|
|
|
goto out_free;
|
|
|
|
} else if (link(realname, filename) && copyfile(name, filename))
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
len = snprintf(linkname, size, "%s/.build-id/%.2s",
|
|
|
|
debugdir, sbuild_id);
|
|
|
|
|
|
|
|
if (access(linkname, X_OK) && mkdir_p(linkname, 0755))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
snprintf(linkname + len, size - len, "/%s", sbuild_id + 2);
|
|
|
|
targetname = filename + strlen(debugdir) - 5;
|
|
|
|
memcpy(targetname, "../..", 5);
|
|
|
|
|
|
|
|
if (symlink(targetname, linkname) == 0)
|
|
|
|
err = 0;
|
|
|
|
out_free:
|
|
|
|
if (!is_kallsyms)
|
|
|
|
free(realname);
|
|
|
|
free(filename);
|
|
|
|
free(linkname);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int build_id_cache__add_b(const u8 *build_id, size_t build_id_size,
|
|
|
|
const char *name, const char *debugdir,
|
|
|
|
bool is_kallsyms)
|
|
|
|
{
|
|
|
|
char sbuild_id[BUILD_ID_SIZE * 2 + 1];
|
|
|
|
|
|
|
|
build_id__sprintf(build_id, build_id_size, sbuild_id);
|
|
|
|
|
|
|
|
return build_id_cache__add_s(sbuild_id, debugdir, name, is_kallsyms);
|
|
|
|
}
|
|
|
|
|
|
|
|
int build_id_cache__remove_s(const char *sbuild_id, const char *debugdir)
|
|
|
|
{
|
|
|
|
const size_t size = PATH_MAX;
|
|
|
|
char *filename = zalloc(size),
|
|
|
|
*linkname = zalloc(size);
|
|
|
|
int err = -1;
|
|
|
|
|
|
|
|
if (filename == NULL || linkname == NULL)
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
snprintf(linkname, size, "%s/.build-id/%.2s/%s",
|
|
|
|
debugdir, sbuild_id, sbuild_id + 2);
|
|
|
|
|
|
|
|
if (access(linkname, F_OK))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
if (readlink(linkname, filename, size - 1) < 0)
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
if (unlink(linkname))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Since the link is relative, we must make it absolute:
|
|
|
|
*/
|
|
|
|
snprintf(linkname, size, "%s/.build-id/%.2s/%s",
|
|
|
|
debugdir, sbuild_id, filename);
|
|
|
|
|
|
|
|
if (unlink(linkname))
|
|
|
|
goto out_free;
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
out_free:
|
|
|
|
free(filename);
|
|
|
|
free(linkname);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dso__cache_build_id(struct dso *dso, const char *debugdir)
|
|
|
|
{
|
|
|
|
bool is_kallsyms = dso->kernel && dso->long_name[0] != '/';
|
|
|
|
|
|
|
|
return build_id_cache__add_b(dso->build_id, sizeof(dso->build_id),
|
|
|
|
dso->long_name, debugdir, is_kallsyms);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir)
|
|
|
|
{
|
|
|
|
struct dso *pos;
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
dsos__for_each_with_build_id(pos, head)
|
|
|
|
if (dso__cache_build_id(pos, debugdir))
|
|
|
|
err = -1;
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int machine__cache_build_ids(struct machine *machine, const char *debugdir)
|
|
|
|
{
|
|
|
|
int ret = __dsos__cache_build_ids(&machine->kernel_dsos, debugdir);
|
|
|
|
ret |= __dsos__cache_build_ids(&machine->user_dsos, debugdir);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int perf_session__cache_build_ids(struct perf_session *session)
|
|
|
|
{
|
|
|
|
struct rb_node *nd;
|
|
|
|
int ret;
|
|
|
|
char debugdir[PATH_MAX];
|
|
|
|
|
|
|
|
snprintf(debugdir, sizeof(debugdir), "%s", buildid_dir);
|
|
|
|
|
|
|
|
if (mkdir(debugdir, 0755) != 0 && errno != EEXIST)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
ret = machine__cache_build_ids(&session->host_machine, debugdir);
|
|
|
|
|
|
|
|
for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
|
|
|
|
struct machine *pos = rb_entry(nd, struct machine, rb_node);
|
|
|
|
ret |= machine__cache_build_ids(pos, debugdir);
|
|
|
|
}
|
|
|
|
return ret ? -1 : 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool machine__read_build_ids(struct machine *machine, bool with_hits)
|
|
|
|
{
|
|
|
|
bool ret = __dsos__read_build_ids(&machine->kernel_dsos, with_hits);
|
|
|
|
ret |= __dsos__read_build_ids(&machine->user_dsos, with_hits);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool perf_session__read_build_ids(struct perf_session *session, bool with_hits)
|
|
|
|
{
|
|
|
|
struct rb_node *nd;
|
|
|
|
bool ret = machine__read_build_ids(&session->host_machine, with_hits);
|
|
|
|
|
|
|
|
for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
|
|
|
|
struct machine *pos = rb_entry(nd, struct machine, rb_node);
|
|
|
|
ret |= machine__read_build_ids(pos, with_hits);
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
static int write_trace_info(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist)
|
|
|
|
{
|
|
|
|
return read_tracing_data(fd, &evlist->entries);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int write_build_id(int fd, struct perf_header *h,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
struct perf_session *session;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
session = container_of(h, struct perf_session, header);
|
|
|
|
|
2011-12-07 09:02:55 +00:00
|
|
|
if (!perf_session__read_build_ids(session, true))
|
|
|
|
return -1;
|
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = dsos__write_buildid_table(h, fd);
|
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write buildid table\n");
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
if (!no_buildid_cache)
|
|
|
|
perf_session__cache_build_ids(session);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_hostname(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
struct utsname uts;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = uname(&uts);
|
|
|
|
if (ret < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
return do_write_string(fd, uts.nodename);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_osrelease(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
struct utsname uts;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = uname(&uts);
|
|
|
|
if (ret < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
return do_write_string(fd, uts.release);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_arch(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
struct utsname uts;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = uname(&uts);
|
|
|
|
if (ret < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
return do_write_string(fd, uts.machine);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_version(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
return do_write_string(fd, perf_version_string);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_cpudesc(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
#ifndef CPUINFO_PROC
|
|
|
|
#define CPUINFO_PROC NULL
|
|
|
|
#endif
|
|
|
|
FILE *file;
|
|
|
|
char *buf = NULL;
|
|
|
|
char *s, *p;
|
|
|
|
const char *search = CPUINFO_PROC;
|
|
|
|
size_t len = 0;
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
if (!search)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
file = fopen("/proc/cpuinfo", "r");
|
|
|
|
if (!file)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
while (getline(&buf, &len, file) > 0) {
|
|
|
|
ret = strncmp(buf, search, strlen(search));
|
|
|
|
if (!ret)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ret)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
s = buf;
|
|
|
|
|
|
|
|
p = strchr(buf, ':');
|
|
|
|
if (p && *(p+1) == ' ' && *(p+2))
|
|
|
|
s = p + 2;
|
|
|
|
p = strchr(s, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
|
|
|
|
/* squash extra space characters (branding string) */
|
|
|
|
p = s;
|
|
|
|
while (*p) {
|
|
|
|
if (isspace(*p)) {
|
|
|
|
char *r = p + 1;
|
|
|
|
char *q = r;
|
|
|
|
*p = ' ';
|
|
|
|
while (*q && isspace(*q))
|
|
|
|
q++;
|
|
|
|
if (q != (p+1))
|
|
|
|
while ((*r++ = *q++));
|
|
|
|
}
|
|
|
|
p++;
|
|
|
|
}
|
|
|
|
ret = do_write_string(fd, s);
|
|
|
|
done:
|
|
|
|
free(buf);
|
|
|
|
fclose(file);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_nrcpus(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
long nr;
|
|
|
|
u32 nrc, nra;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
nr = sysconf(_SC_NPROCESSORS_CONF);
|
|
|
|
if (nr < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
nrc = (u32)(nr & UINT_MAX);
|
|
|
|
|
|
|
|
nr = sysconf(_SC_NPROCESSORS_ONLN);
|
|
|
|
if (nr < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
nra = (u32)(nr & UINT_MAX);
|
|
|
|
|
|
|
|
ret = do_write(fd, &nrc, sizeof(nrc));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
return do_write(fd, &nra, sizeof(nra));
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_event_desc(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist)
|
|
|
|
{
|
|
|
|
struct perf_evsel *attr;
|
|
|
|
u32 nre = 0, nri, sz;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
list_for_each_entry(attr, &evlist->entries, node)
|
|
|
|
nre++;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* write number of events
|
|
|
|
*/
|
|
|
|
ret = do_write(fd, &nre, sizeof(nre));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* size of perf_event_attr struct
|
|
|
|
*/
|
|
|
|
sz = (u32)sizeof(attr->attr);
|
|
|
|
ret = do_write(fd, &sz, sizeof(sz));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
list_for_each_entry(attr, &evlist->entries, node) {
|
|
|
|
|
|
|
|
ret = do_write(fd, &attr->attr, sz);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
/*
|
|
|
|
* write number of unique id per event
|
|
|
|
* there is one id per instance of an event
|
|
|
|
*
|
|
|
|
* copy into an nri to be independent of the
|
|
|
|
* type of ids,
|
|
|
|
*/
|
|
|
|
nri = attr->ids;
|
|
|
|
ret = do_write(fd, &nri, sizeof(nri));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* write event string as passed on cmdline
|
|
|
|
*/
|
2011-11-28 09:03:29 +00:00
|
|
|
ret = do_write_string(fd, event_name(attr));
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
/*
|
|
|
|
* write unique ids for this event
|
|
|
|
*/
|
|
|
|
ret = do_write(fd, attr->id, attr->ids * sizeof(u64));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_cmdline(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
char buf[MAXPATHLEN];
|
|
|
|
char proc[32];
|
|
|
|
u32 i, n;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* actual atual path to perf binary
|
|
|
|
*/
|
|
|
|
sprintf(proc, "/proc/%d/exe", getpid());
|
|
|
|
ret = readlink(proc, buf, sizeof(buf));
|
|
|
|
if (ret <= 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
/* readlink() does not add null termination */
|
|
|
|
buf[ret] = '\0';
|
|
|
|
|
|
|
|
/* account for binary path */
|
|
|
|
n = header_argc + 1;
|
|
|
|
|
|
|
|
ret = do_write(fd, &n, sizeof(n));
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret = do_write_string(fd, buf);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
for (i = 0 ; i < header_argc; i++) {
|
|
|
|
ret = do_write_string(fd, header_argv[i]);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define CORE_SIB_FMT \
|
|
|
|
"/sys/devices/system/cpu/cpu%d/topology/core_siblings_list"
|
|
|
|
#define THRD_SIB_FMT \
|
|
|
|
"/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list"
|
|
|
|
|
|
|
|
struct cpu_topo {
|
|
|
|
u32 core_sib;
|
|
|
|
u32 thread_sib;
|
|
|
|
char **core_siblings;
|
|
|
|
char **thread_siblings;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int build_cpu_topo(struct cpu_topo *tp, int cpu)
|
|
|
|
{
|
|
|
|
FILE *fp;
|
|
|
|
char filename[MAXPATHLEN];
|
|
|
|
char *buf = NULL, *p;
|
|
|
|
size_t len = 0;
|
|
|
|
u32 i = 0;
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
sprintf(filename, CORE_SIB_FMT, cpu);
|
|
|
|
fp = fopen(filename, "r");
|
|
|
|
if (!fp)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
if (getline(&buf, &len, fp) <= 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
fclose(fp);
|
|
|
|
|
|
|
|
p = strchr(buf, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
|
|
|
|
for (i = 0; i < tp->core_sib; i++) {
|
|
|
|
if (!strcmp(buf, tp->core_siblings[i]))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (i == tp->core_sib) {
|
|
|
|
tp->core_siblings[i] = buf;
|
|
|
|
tp->core_sib++;
|
|
|
|
buf = NULL;
|
|
|
|
len = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
sprintf(filename, THRD_SIB_FMT, cpu);
|
|
|
|
fp = fopen(filename, "r");
|
|
|
|
if (!fp)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
if (getline(&buf, &len, fp) <= 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
p = strchr(buf, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
|
|
|
|
for (i = 0; i < tp->thread_sib; i++) {
|
|
|
|
if (!strcmp(buf, tp->thread_siblings[i]))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (i == tp->thread_sib) {
|
|
|
|
tp->thread_siblings[i] = buf;
|
|
|
|
tp->thread_sib++;
|
|
|
|
buf = NULL;
|
|
|
|
}
|
|
|
|
ret = 0;
|
|
|
|
done:
|
|
|
|
if(fp)
|
|
|
|
fclose(fp);
|
|
|
|
free(buf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void free_cpu_topo(struct cpu_topo *tp)
|
|
|
|
{
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
if (!tp)
|
|
|
|
return;
|
|
|
|
|
|
|
|
for (i = 0 ; i < tp->core_sib; i++)
|
|
|
|
free(tp->core_siblings[i]);
|
|
|
|
|
|
|
|
for (i = 0 ; i < tp->thread_sib; i++)
|
|
|
|
free(tp->thread_siblings[i]);
|
|
|
|
|
|
|
|
free(tp);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct cpu_topo *build_cpu_topology(void)
|
|
|
|
{
|
|
|
|
struct cpu_topo *tp;
|
|
|
|
void *addr;
|
|
|
|
u32 nr, i;
|
|
|
|
size_t sz;
|
|
|
|
long ncpus;
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
ncpus = sysconf(_SC_NPROCESSORS_CONF);
|
|
|
|
if (ncpus < 0)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
nr = (u32)(ncpus & UINT_MAX);
|
|
|
|
|
|
|
|
sz = nr * sizeof(char *);
|
|
|
|
|
|
|
|
addr = calloc(1, sizeof(*tp) + 2 * sz);
|
|
|
|
if (!addr)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
tp = addr;
|
|
|
|
|
|
|
|
addr += sizeof(*tp);
|
|
|
|
tp->core_siblings = addr;
|
|
|
|
addr += sz;
|
|
|
|
tp->thread_siblings = addr;
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
ret = build_cpu_topo(tp, i);
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (ret) {
|
|
|
|
free_cpu_topo(tp);
|
|
|
|
tp = NULL;
|
|
|
|
}
|
|
|
|
return tp;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_cpu_topology(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
struct cpu_topo *tp;
|
|
|
|
u32 i;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
tp = build_cpu_topology();
|
|
|
|
if (!tp)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
ret = do_write(fd, &tp->core_sib, sizeof(tp->core_sib));
|
|
|
|
if (ret < 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
for (i = 0; i < tp->core_sib; i++) {
|
|
|
|
ret = do_write_string(fd, tp->core_siblings[i]);
|
|
|
|
if (ret < 0)
|
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
ret = do_write(fd, &tp->thread_sib, sizeof(tp->thread_sib));
|
|
|
|
if (ret < 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
for (i = 0; i < tp->thread_sib; i++) {
|
|
|
|
ret = do_write_string(fd, tp->thread_siblings[i]);
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
done:
|
|
|
|
free_cpu_topo(tp);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
static int write_total_mem(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
char *buf = NULL;
|
|
|
|
FILE *fp;
|
|
|
|
size_t len = 0;
|
|
|
|
int ret = -1, n;
|
|
|
|
uint64_t mem;
|
|
|
|
|
|
|
|
fp = fopen("/proc/meminfo", "r");
|
|
|
|
if (!fp)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
while (getline(&buf, &len, fp) > 0) {
|
|
|
|
ret = strncmp(buf, "MemTotal:", 9);
|
|
|
|
if (!ret)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (!ret) {
|
|
|
|
n = sscanf(buf, "%*s %"PRIu64, &mem);
|
|
|
|
if (n == 1)
|
|
|
|
ret = do_write(fd, &mem, sizeof(mem));
|
|
|
|
}
|
|
|
|
free(buf);
|
|
|
|
fclose(fp);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_topo_node(int fd, int node)
|
|
|
|
{
|
|
|
|
char str[MAXPATHLEN];
|
|
|
|
char field[32];
|
|
|
|
char *buf = NULL, *p;
|
|
|
|
size_t len = 0;
|
|
|
|
FILE *fp;
|
|
|
|
u64 mem_total, mem_free, mem;
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
sprintf(str, "/sys/devices/system/node/node%d/meminfo", node);
|
|
|
|
fp = fopen(str, "r");
|
|
|
|
if (!fp)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
while (getline(&buf, &len, fp) > 0) {
|
|
|
|
/* skip over invalid lines */
|
|
|
|
if (!strchr(buf, ':'))
|
|
|
|
continue;
|
|
|
|
if (sscanf(buf, "%*s %*d %s %"PRIu64, field, &mem) != 2)
|
|
|
|
goto done;
|
|
|
|
if (!strcmp(field, "MemTotal:"))
|
|
|
|
mem_total = mem;
|
|
|
|
if (!strcmp(field, "MemFree:"))
|
|
|
|
mem_free = mem;
|
|
|
|
}
|
|
|
|
|
|
|
|
fclose(fp);
|
|
|
|
|
|
|
|
ret = do_write(fd, &mem_total, sizeof(u64));
|
|
|
|
if (ret)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
ret = do_write(fd, &mem_free, sizeof(u64));
|
|
|
|
if (ret)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
ret = -1;
|
|
|
|
sprintf(str, "/sys/devices/system/node/node%d/cpulist", node);
|
|
|
|
|
|
|
|
fp = fopen(str, "r");
|
|
|
|
if (!fp)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
if (getline(&buf, &len, fp) <= 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
p = strchr(buf, '\n');
|
|
|
|
if (p)
|
|
|
|
*p = '\0';
|
|
|
|
|
|
|
|
ret = do_write_string(fd, buf);
|
|
|
|
done:
|
|
|
|
free(buf);
|
|
|
|
fclose(fp);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_numa_topology(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
char *buf = NULL;
|
|
|
|
size_t len = 0;
|
|
|
|
FILE *fp;
|
|
|
|
struct cpu_map *node_map = NULL;
|
|
|
|
char *c;
|
|
|
|
u32 nr, i, j;
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
fp = fopen("/sys/devices/system/node/online", "r");
|
|
|
|
if (!fp)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
if (getline(&buf, &len, fp) <= 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
c = strchr(buf, '\n');
|
|
|
|
if (c)
|
|
|
|
*c = '\0';
|
|
|
|
|
|
|
|
node_map = cpu_map__new(buf);
|
|
|
|
if (!node_map)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
nr = (u32)node_map->nr;
|
|
|
|
|
|
|
|
ret = do_write(fd, &nr, sizeof(nr));
|
|
|
|
if (ret < 0)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
j = (u32)node_map->map[i];
|
|
|
|
ret = do_write(fd, &j, sizeof(j));
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
|
|
|
|
ret = write_topo_node(fd, i);
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
done:
|
|
|
|
free(buf);
|
|
|
|
fclose(fp);
|
|
|
|
free(node_map);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* default get_cpuid(): nothing gets recorded
|
|
|
|
* actual implementation must be in arch/$(ARCH)/util/header.c
|
|
|
|
*/
|
|
|
|
int __attribute__((weak)) get_cpuid(char *buffer __used, size_t sz __used)
|
|
|
|
{
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int write_cpuid(int fd, struct perf_header *h __used,
|
|
|
|
struct perf_evlist *evlist __used)
|
|
|
|
{
|
|
|
|
char buffer[64];
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = get_cpuid(buffer, sizeof(buffer));
|
|
|
|
if (!ret)
|
|
|
|
goto write_it;
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
write_it:
|
|
|
|
return do_write_string(fd, buffer);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_hostname(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# hostname : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_osrelease(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# os release : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_arch(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# arch : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_cpudesc(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# cpudesc : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_nrcpus(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
ssize_t ret;
|
|
|
|
u32 nr;
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
nr = -1; /* interpreted as error */
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
fprintf(fp, "# nrcpus online : %u\n", nr);
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
nr = -1; /* interpreted as error */
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
fprintf(fp, "# nrcpus avail : %u\n", nr);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_version(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# perf version : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_cmdline(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
ssize_t ret;
|
|
|
|
char *str;
|
|
|
|
u32 nr, i;
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
fprintf(fp, "# cmdline : ");
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "%s ", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
fputc('\n', fp);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_cpu_topology(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
ssize_t ret;
|
|
|
|
u32 nr, i;
|
|
|
|
char *str;
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# sibling cores : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# sibling threads : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_event_desc(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
struct perf_event_attr attr;
|
|
|
|
uint64_t id;
|
|
|
|
void *buf = NULL;
|
|
|
|
char *str;
|
|
|
|
u32 nre, sz, nr, i, j, msz;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* number of events */
|
|
|
|
ret = read(fd, &nre, sizeof(nre));
|
|
|
|
if (ret != (ssize_t)sizeof(nre))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nre = bswap_32(nre);
|
|
|
|
|
|
|
|
ret = read(fd, &sz, sizeof(sz));
|
|
|
|
if (ret != (ssize_t)sizeof(sz))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
sz = bswap_32(sz);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ensure it is at least to our ABI rev
|
|
|
|
*/
|
|
|
|
if (sz < (u32)sizeof(attr))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
memset(&attr, 0, sizeof(attr));
|
|
|
|
|
|
|
|
/* read entire region to sync up to next field */
|
|
|
|
buf = malloc(sz);
|
|
|
|
if (!buf)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
msz = sizeof(attr);
|
|
|
|
if (sz < msz)
|
|
|
|
msz = sz;
|
|
|
|
|
|
|
|
for (i = 0 ; i < nre; i++) {
|
|
|
|
|
|
|
|
ret = read(fd, buf, sz);
|
|
|
|
if (ret != (ssize_t)sz)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
perf_event__attr_swap(buf);
|
|
|
|
|
|
|
|
memcpy(&attr, buf, msz);
|
|
|
|
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# event : name = %s, ", str);
|
|
|
|
free(str);
|
|
|
|
|
|
|
|
fprintf(fp, "type = %d, config = 0x%"PRIx64
|
|
|
|
", config1 = 0x%"PRIx64", config2 = 0x%"PRIx64,
|
|
|
|
attr.type,
|
|
|
|
(u64)attr.config,
|
|
|
|
(u64)attr.config1,
|
|
|
|
(u64)attr.config2);
|
|
|
|
|
|
|
|
fprintf(fp, ", excl_usr = %d, excl_kern = %d",
|
|
|
|
attr.exclude_user,
|
|
|
|
attr.exclude_kernel);
|
|
|
|
|
|
|
|
if (nr)
|
|
|
|
fprintf(fp, ", id = {");
|
|
|
|
|
|
|
|
for (j = 0 ; j < nr; j++) {
|
|
|
|
ret = read(fd, &id, sizeof(id));
|
|
|
|
if (ret != (ssize_t)sizeof(id))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (ph->needs_swap)
|
|
|
|
id = bswap_64(id);
|
|
|
|
|
|
|
|
if (j)
|
|
|
|
fputc(',', fp);
|
|
|
|
|
|
|
|
fprintf(fp, " %"PRIu64, id);
|
|
|
|
}
|
|
|
|
if (nr && j == nr)
|
|
|
|
fprintf(fp, " }");
|
|
|
|
fputc('\n', fp);
|
|
|
|
}
|
|
|
|
free(buf);
|
|
|
|
return;
|
|
|
|
error:
|
|
|
|
fprintf(fp, "# event desc: not available or unable to read\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_total_mem(struct perf_header *h __used, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
uint64_t mem;
|
|
|
|
ssize_t ret;
|
|
|
|
|
|
|
|
ret = read(fd, &mem, sizeof(mem));
|
|
|
|
if (ret != sizeof(mem))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (h->needs_swap)
|
|
|
|
mem = bswap_64(mem);
|
|
|
|
|
|
|
|
fprintf(fp, "# total memory : %"PRIu64" kB\n", mem);
|
|
|
|
return;
|
|
|
|
error:
|
|
|
|
fprintf(fp, "# total memory : unknown\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_numa_topology(struct perf_header *h __used, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
ssize_t ret;
|
|
|
|
u32 nr, c, i;
|
|
|
|
char *str;
|
|
|
|
uint64_t mem_total, mem_free;
|
|
|
|
|
|
|
|
/* nr nodes */
|
|
|
|
ret = read(fd, &nr, sizeof(nr));
|
|
|
|
if (ret != (ssize_t)sizeof(nr))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (h->needs_swap)
|
|
|
|
nr = bswap_32(nr);
|
|
|
|
|
|
|
|
for (i = 0; i < nr; i++) {
|
|
|
|
|
|
|
|
/* node number */
|
|
|
|
ret = read(fd, &c, sizeof(c));
|
|
|
|
if (ret != (ssize_t)sizeof(c))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (h->needs_swap)
|
|
|
|
c = bswap_32(c);
|
|
|
|
|
|
|
|
ret = read(fd, &mem_total, sizeof(u64));
|
|
|
|
if (ret != sizeof(u64))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
ret = read(fd, &mem_free, sizeof(u64));
|
|
|
|
if (ret != sizeof(u64))
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
if (h->needs_swap) {
|
|
|
|
mem_total = bswap_64(mem_total);
|
|
|
|
mem_free = bswap_64(mem_free);
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf(fp, "# node%u meminfo : total = %"PRIu64" kB,"
|
|
|
|
" free = %"PRIu64" kB\n",
|
|
|
|
c,
|
|
|
|
mem_total,
|
|
|
|
mem_free);
|
|
|
|
|
|
|
|
str = do_read_string(fd, h);
|
|
|
|
fprintf(fp, "# node%u cpu list : %s\n", c, str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
error:
|
|
|
|
fprintf(fp, "# numa topology : not available\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void print_cpuid(struct perf_header *ph, int fd, FILE *fp)
|
|
|
|
{
|
|
|
|
char *str = do_read_string(fd, ph);
|
|
|
|
fprintf(fp, "# cpuid : %s\n", str);
|
|
|
|
free(str);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct feature_ops {
|
|
|
|
int (*write)(int fd, struct perf_header *h, struct perf_evlist *evlist);
|
|
|
|
void (*print)(struct perf_header *h, int fd, FILE *fp);
|
|
|
|
const char *name;
|
|
|
|
bool full_only;
|
|
|
|
};
|
|
|
|
|
2011-12-07 09:02:56 +00:00
|
|
|
#define FEAT_OPA(n, func) \
|
|
|
|
[n] = { .name = #n, .write = write_##func, .print = print_##func }
|
|
|
|
#define FEAT_OPF(n, func) \
|
|
|
|
[n] = { .name = #n, .write = write_##func, .print = print_##func, .full_only = true }
|
|
|
|
|
|
|
|
/* feature_ops not implemented: */
|
|
|
|
#define print_trace_info NULL
|
|
|
|
#define print_build_id NULL
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
|
|
|
|
static const struct feature_ops feat_ops[HEADER_LAST_FEATURE] = {
|
2011-12-07 09:02:56 +00:00
|
|
|
FEAT_OPA(HEADER_TRACE_INFO, trace_info),
|
|
|
|
FEAT_OPA(HEADER_BUILD_ID, build_id),
|
|
|
|
FEAT_OPA(HEADER_HOSTNAME, hostname),
|
|
|
|
FEAT_OPA(HEADER_OSRELEASE, osrelease),
|
|
|
|
FEAT_OPA(HEADER_VERSION, version),
|
|
|
|
FEAT_OPA(HEADER_ARCH, arch),
|
|
|
|
FEAT_OPA(HEADER_NRCPUS, nrcpus),
|
|
|
|
FEAT_OPA(HEADER_CPUDESC, cpudesc),
|
|
|
|
FEAT_OPA(HEADER_CPUID, cpuid),
|
|
|
|
FEAT_OPA(HEADER_TOTAL_MEM, total_mem),
|
|
|
|
FEAT_OPA(HEADER_EVENT_DESC, event_desc),
|
|
|
|
FEAT_OPA(HEADER_CMDLINE, cmdline),
|
|
|
|
FEAT_OPF(HEADER_CPU_TOPOLOGY, cpu_topology),
|
|
|
|
FEAT_OPF(HEADER_NUMA_TOPOLOGY, numa_topology),
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
struct header_print_data {
|
|
|
|
FILE *fp;
|
|
|
|
bool full; /* extended list of headers */
|
|
|
|
};
|
|
|
|
|
|
|
|
static int perf_file_section__fprintf_info(struct perf_file_section *section,
|
|
|
|
struct perf_header *ph,
|
|
|
|
int feat, int fd, void *data)
|
|
|
|
{
|
|
|
|
struct header_print_data *hd = data;
|
|
|
|
|
|
|
|
if (lseek(fd, section->offset, SEEK_SET) == (off_t)-1) {
|
|
|
|
pr_debug("Failed to lseek to %" PRIu64 " offset for feature "
|
|
|
|
"%d, continuing...\n", section->offset, feat);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
if (feat < HEADER_TRACE_INFO || feat >= HEADER_LAST_FEATURE) {
|
|
|
|
pr_warning("unknown feature %d\n", feat);
|
2011-12-07 09:02:51 +00:00
|
|
|
return 0;
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
}
|
|
|
|
if (!feat_ops[feat].print)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (!feat_ops[feat].full_only || hd->full)
|
|
|
|
feat_ops[feat].print(ph, fd, hd->fp);
|
|
|
|
else
|
|
|
|
fprintf(hd->fp, "# %s info available, use -I to display\n",
|
|
|
|
feat_ops[feat].name);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int perf_header__fprintf_info(struct perf_session *session, FILE *fp, bool full)
|
|
|
|
{
|
|
|
|
struct header_print_data hd;
|
|
|
|
struct perf_header *header = &session->header;
|
|
|
|
int fd = session->fd;
|
|
|
|
hd.fp = fp;
|
|
|
|
hd.full = full;
|
|
|
|
|
|
|
|
perf_header__process_sections(header, fd, &hd,
|
|
|
|
perf_file_section__fprintf_info);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int do_write_feat(int fd, struct perf_header *h, int type,
|
|
|
|
struct perf_file_section **p,
|
|
|
|
struct perf_evlist *evlist)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (perf_header__has_feat(h, type)) {
|
|
|
|
|
|
|
|
(*p)->offset = lseek(fd, 0, SEEK_CUR);
|
|
|
|
|
|
|
|
err = feat_ops[type].write(fd, h, evlist);
|
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write feature %d\n", type);
|
|
|
|
|
|
|
|
/* undo anything written */
|
|
|
|
lseek(fd, (*p)->offset, SEEK_SET);
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
(*p)->size = lseek(fd, 0, SEEK_CUR) - (*p)->offset;
|
|
|
|
(*p)++;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
static int perf_header__adds_write(struct perf_header *header,
|
2011-01-11 22:56:53 +00:00
|
|
|
struct perf_evlist *evlist, int fd)
|
2009-10-17 15:12:34 +00:00
|
|
|
{
|
2009-11-11 03:51:07 +00:00
|
|
|
int nr_sections;
|
2010-05-19 16:41:23 +00:00
|
|
|
struct perf_session *session;
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
struct perf_file_section *feat_sec, *p;
|
2009-11-11 03:51:07 +00:00
|
|
|
int sec_size;
|
|
|
|
u64 sec_start;
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
int err;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
session = container_of(header, struct perf_session, header);
|
2010-11-26 21:39:15 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
|
2009-11-11 03:51:07 +00:00
|
|
|
if (!nr_sections)
|
2009-11-19 16:55:56 +00:00
|
|
|
return 0;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
feat_sec = p = calloc(sizeof(*feat_sec), nr_sections);
|
2009-11-19 16:55:56 +00:00
|
|
|
if (feat_sec == NULL)
|
|
|
|
return -ENOMEM;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
|
|
|
sec_size = sizeof(*feat_sec) * nr_sections;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
sec_start = header->data_offset + header->data_size;
|
2010-02-04 08:46:42 +00:00
|
|
|
lseek(fd, sec_start + sec_size, SEEK_SET);
|
2009-10-17 15:12:34 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = do_write_feat(fd, header, HEADER_TRACE_INFO, &p, evlist);
|
|
|
|
if (err)
|
2011-12-07 09:02:55 +00:00
|
|
|
perf_header__clear_feat(header, HEADER_TRACE_INFO);
|
2009-11-11 03:51:07 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = do_write_feat(fd, header, HEADER_BUILD_ID, &p, evlist);
|
2011-12-07 09:02:55 +00:00
|
|
|
if (err)
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
perf_header__clear_feat(header, HEADER_BUILD_ID);
|
2009-11-11 03:51:03 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = do_write_feat(fd, header, HEADER_HOSTNAME, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_HOSTNAME);
|
2009-11-11 03:51:07 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = do_write_feat(fd, header, HEADER_OSRELEASE, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_OSRELEASE);
|
2009-11-11 03:51:07 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = do_write_feat(fd, header, HEADER_VERSION, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_VERSION);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_ARCH, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_ARCH);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_NRCPUS, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_NRCPUS);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_CPUDESC, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_CPUDESC);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_CPUID, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_CPUID);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_TOTAL_MEM, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_TOTAL_MEM);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_CMDLINE, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_CMDLINE);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_EVENT_DESC, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_EVENT_DESC);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_CPU_TOPOLOGY, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_CPU_TOPOLOGY);
|
|
|
|
|
|
|
|
err = do_write_feat(fd, header, HEADER_NUMA_TOPOLOGY, &p, evlist);
|
|
|
|
if (err)
|
|
|
|
perf_header__clear_feat(header, HEADER_NUMA_TOPOLOGY);
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2010-02-04 08:46:42 +00:00
|
|
|
lseek(fd, sec_start, SEEK_SET);
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
/*
|
|
|
|
* may write more than needed due to dropped feature, but
|
|
|
|
* this is okay, reader will skip the mising entries
|
|
|
|
*/
|
2009-11-19 16:55:56 +00:00
|
|
|
err = do_write(fd, feat_sec, sec_size);
|
|
|
|
if (err < 0)
|
|
|
|
pr_debug("failed to write feature section\n");
|
2009-11-11 03:51:07 +00:00
|
|
|
free(feat_sec);
|
2009-11-19 16:55:56 +00:00
|
|
|
return err;
|
2009-11-11 03:51:07 +00:00
|
|
|
}
|
2009-10-17 15:12:34 +00:00
|
|
|
|
2010-04-02 04:59:15 +00:00
|
|
|
int perf_header__write_pipe(int fd)
|
|
|
|
{
|
|
|
|
struct perf_pipe_file_header f_header;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
f_header = (struct perf_pipe_file_header){
|
|
|
|
.magic = PERF_MAGIC,
|
|
|
|
.size = sizeof(f_header),
|
|
|
|
};
|
|
|
|
|
|
|
|
err = do_write(fd, &f_header, sizeof(f_header));
|
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write perf pipe header\n");
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
int perf_session__write_header(struct perf_session *session,
|
|
|
|
struct perf_evlist *evlist,
|
|
|
|
int fd, bool at_exit)
|
2009-06-25 15:05:54 +00:00
|
|
|
{
|
|
|
|
struct perf_file_header f_header;
|
|
|
|
struct perf_file_attr f_attr;
|
2011-03-09 11:13:19 +00:00
|
|
|
struct perf_header *header = &session->header;
|
2011-03-10 14:15:54 +00:00
|
|
|
struct perf_evsel *attr, *pair = NULL;
|
|
|
|
int err;
|
2009-06-25 15:05:54 +00:00
|
|
|
|
|
|
|
lseek(fd, sizeof(f_header), SEEK_SET);
|
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
if (session->evlist != evlist)
|
|
|
|
pair = list_entry(session->evlist->entries.next, struct perf_evsel, node);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
list_for_each_entry(attr, &evlist->entries, node) {
|
2009-06-25 15:05:54 +00:00
|
|
|
attr->id_offset = lseek(fd, 0, SEEK_CUR);
|
2009-11-19 16:55:56 +00:00
|
|
|
err = do_write(fd, attr->id, attr->ids * sizeof(u64));
|
|
|
|
if (err < 0) {
|
2011-03-10 14:15:54 +00:00
|
|
|
out_err_write:
|
2009-11-19 16:55:56 +00:00
|
|
|
pr_debug("failed to write perf header\n");
|
|
|
|
return err;
|
|
|
|
}
|
2011-03-10 14:15:54 +00:00
|
|
|
if (session->evlist != evlist) {
|
|
|
|
err = do_write(fd, pair->id, pair->ids * sizeof(u64));
|
|
|
|
if (err < 0)
|
|
|
|
goto out_err_write;
|
|
|
|
attr->ids += pair->ids;
|
|
|
|
pair = list_entry(pair->node.next, struct perf_evsel, node);
|
|
|
|
}
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
header->attr_offset = lseek(fd, 0, SEEK_CUR);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
list_for_each_entry(attr, &evlist->entries, node) {
|
2009-06-25 15:05:54 +00:00
|
|
|
f_attr = (struct perf_file_attr){
|
|
|
|
.attr = attr->attr,
|
|
|
|
.ids = {
|
|
|
|
.offset = attr->id_offset,
|
|
|
|
.size = attr->ids * sizeof(u64),
|
|
|
|
}
|
|
|
|
};
|
2009-11-19 16:55:56 +00:00
|
|
|
err = do_write(fd, &f_attr, sizeof(f_attr));
|
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write perf header attribute\n");
|
|
|
|
return err;
|
|
|
|
}
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
header->event_offset = lseek(fd, 0, SEEK_CUR);
|
|
|
|
header->event_size = event_count * sizeof(struct perf_trace_event_type);
|
2009-11-19 16:55:56 +00:00
|
|
|
if (events) {
|
2011-03-09 11:13:19 +00:00
|
|
|
err = do_write(fd, events, header->event_size);
|
2009-11-19 16:55:56 +00:00
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write perf header events\n");
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
}
|
2009-09-12 05:52:51 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
header->data_offset = lseek(fd, 0, SEEK_CUR);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2009-11-19 16:55:56 +00:00
|
|
|
if (at_exit) {
|
2011-03-09 11:13:19 +00:00
|
|
|
err = perf_header__adds_write(header, evlist, fd);
|
2009-11-19 16:55:56 +00:00
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
}
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2009-06-25 15:05:54 +00:00
|
|
|
f_header = (struct perf_file_header){
|
|
|
|
.magic = PERF_MAGIC,
|
|
|
|
.size = sizeof(f_header),
|
|
|
|
.attr_size = sizeof(f_attr),
|
|
|
|
.attrs = {
|
2011-03-09 11:13:19 +00:00
|
|
|
.offset = header->attr_offset,
|
2011-03-10 14:15:54 +00:00
|
|
|
.size = evlist->nr_entries * sizeof(f_attr),
|
2009-06-25 15:05:54 +00:00
|
|
|
},
|
|
|
|
.data = {
|
2011-03-09 11:13:19 +00:00
|
|
|
.offset = header->data_offset,
|
|
|
|
.size = header->data_size,
|
2009-06-25 15:05:54 +00:00
|
|
|
},
|
2009-09-12 05:52:51 +00:00
|
|
|
.event_types = {
|
2011-03-09 11:13:19 +00:00
|
|
|
.offset = header->event_offset,
|
|
|
|
.size = header->event_size,
|
2009-09-12 05:52:51 +00:00
|
|
|
},
|
2009-06-25 15:05:54 +00:00
|
|
|
};
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
memcpy(&f_header.adds_features, &header->adds_features, sizeof(header->adds_features));
|
2009-10-17 15:12:34 +00:00
|
|
|
|
2009-06-25 15:05:54 +00:00
|
|
|
lseek(fd, 0, SEEK_SET);
|
2009-11-19 16:55:56 +00:00
|
|
|
err = do_write(fd, &f_header, sizeof(f_header));
|
|
|
|
if (err < 0) {
|
|
|
|
pr_debug("failed to write perf header\n");
|
|
|
|
return err;
|
|
|
|
}
|
2011-03-09 11:13:19 +00:00
|
|
|
lseek(fd, header->data_offset + header->data_size, SEEK_SET);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
header->frozen = 1;
|
2009-11-19 16:55:56 +00:00
|
|
|
return 0;
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
static int perf_header__getbuffer64(struct perf_header *header,
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
int fd, void *buf, size_t size)
|
|
|
|
{
|
2011-01-03 18:50:55 +00:00
|
|
|
if (readn(fd, buf, size) <= 0)
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
return -1;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->needs_swap)
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
mem_bswap_64(buf, size);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
int perf_header__process_sections(struct perf_header *header, int fd,
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
void *data,
|
2011-03-09 11:13:19 +00:00
|
|
|
int (*process)(struct perf_file_section *section,
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
struct perf_header *ph,
|
|
|
|
int feat, int fd, void *data))
|
2009-10-17 15:12:34 +00:00
|
|
|
{
|
2009-11-11 03:51:07 +00:00
|
|
|
struct perf_file_section *feat_sec;
|
|
|
|
int nr_sections;
|
|
|
|
int sec_size;
|
|
|
|
int idx = 0;
|
2009-12-29 00:48:32 +00:00
|
|
|
int err = -1, feat = 1;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
|
2009-11-11 03:51:07 +00:00
|
|
|
if (!nr_sections)
|
2009-11-16 18:32:43 +00:00
|
|
|
return 0;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
|
|
|
feat_sec = calloc(sizeof(*feat_sec), nr_sections);
|
|
|
|
if (!feat_sec)
|
2009-11-16 18:32:43 +00:00
|
|
|
return -1;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
|
|
|
sec_size = sizeof(*feat_sec) * nr_sections;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
lseek(fd, header->data_offset + header->data_size, SEEK_SET);
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_header__getbuffer64(header, fd, feat_sec, sec_size))
|
2009-12-29 00:48:32 +00:00
|
|
|
goto out_free;
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2009-12-29 00:48:32 +00:00
|
|
|
err = 0;
|
2009-11-16 18:32:43 +00:00
|
|
|
while (idx < nr_sections && feat < HEADER_LAST_FEATURE) {
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_header__has_feat(header, feat)) {
|
2009-11-16 18:32:43 +00:00
|
|
|
struct perf_file_section *sec = &feat_sec[idx++];
|
2009-10-17 15:12:34 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
err = process(sec, header, feat, fd, data);
|
2009-11-16 18:32:43 +00:00
|
|
|
if (err < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
++feat;
|
2009-10-17 15:12:34 +00:00
|
|
|
}
|
2009-12-29 00:48:32 +00:00
|
|
|
out_free:
|
2009-11-16 18:32:43 +00:00
|
|
|
free(feat_sec);
|
|
|
|
return err;
|
2009-12-29 00:48:32 +00:00
|
|
|
}
|
2009-11-11 03:51:05 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
int perf_file_header__read(struct perf_file_header *header,
|
2009-11-16 18:32:43 +00:00
|
|
|
struct perf_header *ph, int fd)
|
|
|
|
{
|
|
|
|
lseek(fd, 0, SEEK_SET);
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (readn(fd, header, sizeof(*header)) <= 0 ||
|
|
|
|
memcmp(&header->magic, __perf_magic, sizeof(header->magic)))
|
2009-11-16 18:32:43 +00:00
|
|
|
return -1;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->attr_size != sizeof(struct perf_file_attr)) {
|
|
|
|
u64 attr_size = bswap_64(header->attr_size);
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
|
|
|
|
if (attr_size != sizeof(struct perf_file_attr))
|
|
|
|
return -1;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
mem_bswap_64(header, offsetof(struct perf_file_header,
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
adds_features));
|
|
|
|
ph->needs_swap = true;
|
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->size != sizeof(*header)) {
|
2009-11-16 18:32:43 +00:00
|
|
|
/* Support the previous format */
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->size == offsetof(typeof(*header), adds_features))
|
|
|
|
bitmap_zero(header->adds_features, HEADER_FEAT_BITS);
|
2009-11-16 18:32:43 +00:00
|
|
|
else
|
|
|
|
return -1;
|
2011-10-18 23:34:01 +00:00
|
|
|
} else if (ph->needs_swap) {
|
|
|
|
unsigned int i;
|
|
|
|
/*
|
|
|
|
* feature bitmap is declared as an array of unsigned longs --
|
|
|
|
* not good since its size can differ between the host that
|
|
|
|
* generated the data file and the host analyzing the file.
|
|
|
|
*
|
|
|
|
* We need to handle endianness, but we don't know the size of
|
|
|
|
* the unsigned long where the file was generated. Take a best
|
|
|
|
* guess at determining it: try 64-bit swap first (ie., file
|
|
|
|
* created on a 64-bit host), and check if the hostname feature
|
|
|
|
* bit is set (this feature bit is forced on as of fbe96f2).
|
|
|
|
* If the bit is not, undo the 64-bit swap and try a 32-bit
|
|
|
|
* swap. If the hostname bit is still not set (e.g., older data
|
|
|
|
* file), punt and fallback to the original behavior --
|
|
|
|
* clearing all feature bits and setting buildid.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < BITS_TO_LONGS(HEADER_FEAT_BITS); ++i)
|
|
|
|
header->adds_features[i] = bswap_64(header->adds_features[i]);
|
|
|
|
|
|
|
|
if (!test_bit(HEADER_HOSTNAME, header->adds_features)) {
|
|
|
|
for (i = 0; i < BITS_TO_LONGS(HEADER_FEAT_BITS); ++i) {
|
|
|
|
header->adds_features[i] = bswap_64(header->adds_features[i]);
|
|
|
|
header->adds_features[i] = bswap_32(header->adds_features[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!test_bit(HEADER_HOSTNAME, header->adds_features)) {
|
|
|
|
bitmap_zero(header->adds_features, HEADER_FEAT_BITS);
|
|
|
|
set_bit(HEADER_BUILD_ID, header->adds_features);
|
|
|
|
}
|
2009-11-11 03:51:05 +00:00
|
|
|
}
|
2009-11-11 03:51:07 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
memcpy(&ph->adds_features, &header->adds_features,
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
sizeof(ph->adds_features));
|
2009-11-16 18:32:43 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
ph->event_offset = header->event_types.offset;
|
|
|
|
ph->event_size = header->event_types.size;
|
|
|
|
ph->data_offset = header->data.offset;
|
|
|
|
ph->data_size = header->data.size;
|
2009-11-16 18:32:43 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2010-04-19 05:32:50 +00:00
|
|
|
static int __event_process_build_id(struct build_id_event *bev,
|
|
|
|
char *filename,
|
|
|
|
struct perf_session *session)
|
|
|
|
{
|
|
|
|
int err = -1;
|
|
|
|
struct list_head *head;
|
2010-04-28 00:17:50 +00:00
|
|
|
struct machine *machine;
|
2010-04-19 05:32:50 +00:00
|
|
|
u16 misc;
|
|
|
|
struct dso *dso;
|
|
|
|
enum dso_kernel_type dso_type;
|
|
|
|
|
2010-04-28 00:17:50 +00:00
|
|
|
machine = perf_session__findnew_machine(session, bev->pid);
|
|
|
|
if (!machine)
|
2010-04-19 05:32:50 +00:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
misc = bev->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
|
|
|
|
|
|
|
|
switch (misc) {
|
|
|
|
case PERF_RECORD_MISC_KERNEL:
|
|
|
|
dso_type = DSO_TYPE_KERNEL;
|
2010-04-28 00:17:50 +00:00
|
|
|
head = &machine->kernel_dsos;
|
2010-04-19 05:32:50 +00:00
|
|
|
break;
|
|
|
|
case PERF_RECORD_MISC_GUEST_KERNEL:
|
|
|
|
dso_type = DSO_TYPE_GUEST_KERNEL;
|
2010-04-28 00:17:50 +00:00
|
|
|
head = &machine->kernel_dsos;
|
2010-04-19 05:32:50 +00:00
|
|
|
break;
|
|
|
|
case PERF_RECORD_MISC_USER:
|
|
|
|
case PERF_RECORD_MISC_GUEST_USER:
|
|
|
|
dso_type = DSO_TYPE_USER;
|
2010-04-28 00:17:50 +00:00
|
|
|
head = &machine->user_dsos;
|
2010-04-19 05:32:50 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
dso = __dsos__findnew(head, filename);
|
|
|
|
if (dso != NULL) {
|
2010-05-01 06:41:20 +00:00
|
|
|
char sbuild_id[BUILD_ID_SIZE * 2 + 1];
|
|
|
|
|
2010-04-19 05:32:50 +00:00
|
|
|
dso__set_build_id(dso, &bev->build_id);
|
2010-05-01 06:41:20 +00:00
|
|
|
|
|
|
|
if (filename[0] == '[')
|
|
|
|
dso->kernel = dso_type;
|
|
|
|
|
|
|
|
build_id__sprintf(dso->build_id, sizeof(dso->build_id),
|
|
|
|
sbuild_id);
|
|
|
|
pr_debug("build id event received for %s: %s\n",
|
|
|
|
dso->long_name, sbuild_id);
|
|
|
|
}
|
2010-04-19 05:32:50 +00:00
|
|
|
|
|
|
|
err = 0;
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-03-22 16:36:45 +00:00
|
|
|
static int perf_header__read_build_ids_abi_quirk(struct perf_header *header,
|
|
|
|
int input, u64 offset, u64 size)
|
|
|
|
{
|
|
|
|
struct perf_session *session = container_of(header, struct perf_session, header);
|
|
|
|
struct {
|
|
|
|
struct perf_event_header header;
|
|
|
|
u8 build_id[ALIGN(BUILD_ID_SIZE, sizeof(u64))];
|
|
|
|
char filename[0];
|
|
|
|
} old_bev;
|
|
|
|
struct build_id_event bev;
|
|
|
|
char filename[PATH_MAX];
|
|
|
|
u64 limit = offset + size;
|
|
|
|
|
|
|
|
while (offset < limit) {
|
|
|
|
ssize_t len;
|
|
|
|
|
|
|
|
if (read(input, &old_bev, sizeof(old_bev)) != sizeof(old_bev))
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
if (header->needs_swap)
|
|
|
|
perf_event_header__bswap(&old_bev.header);
|
|
|
|
|
|
|
|
len = old_bev.header.size - sizeof(old_bev);
|
|
|
|
if (read(input, filename, len) != len)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
bev.header = old_bev.header;
|
2011-06-01 19:43:46 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* As the pid is the missing value, we need to fill
|
|
|
|
* it properly. The header.misc value give us nice hint.
|
|
|
|
*/
|
|
|
|
bev.pid = HOST_KERNEL_ID;
|
|
|
|
if (bev.header.misc == PERF_RECORD_MISC_GUEST_USER ||
|
|
|
|
bev.header.misc == PERF_RECORD_MISC_GUEST_KERNEL)
|
|
|
|
bev.pid = DEFAULT_GUEST_KERNEL_ID;
|
|
|
|
|
2011-03-22 16:36:45 +00:00
|
|
|
memcpy(bev.build_id, old_bev.build_id, sizeof(bev.build_id));
|
|
|
|
__event_process_build_id(&bev, filename, session);
|
|
|
|
|
|
|
|
offset += bev.header.size;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
static int perf_header__read_build_ids(struct perf_header *header,
|
|
|
|
int input, u64 offset, u64 size)
|
2010-04-19 05:32:50 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
struct perf_session *session = container_of(header, struct perf_session, header);
|
2010-04-19 05:32:50 +00:00
|
|
|
struct build_id_event bev;
|
|
|
|
char filename[PATH_MAX];
|
2011-03-22 16:36:45 +00:00
|
|
|
u64 limit = offset + size, orig_offset = offset;
|
2010-04-19 05:32:50 +00:00
|
|
|
int err = -1;
|
|
|
|
|
|
|
|
while (offset < limit) {
|
|
|
|
ssize_t len;
|
|
|
|
|
|
|
|
if (read(input, &bev, sizeof(bev)) != sizeof(bev))
|
|
|
|
goto out;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->needs_swap)
|
2010-04-19 05:32:50 +00:00
|
|
|
perf_event_header__bswap(&bev.header);
|
|
|
|
|
|
|
|
len = bev.header.size - sizeof(bev);
|
|
|
|
if (read(input, filename, len) != len)
|
|
|
|
goto out;
|
2011-03-22 16:36:45 +00:00
|
|
|
/*
|
|
|
|
* The a1645ce1 changeset:
|
|
|
|
*
|
|
|
|
* "perf: 'perf kvm' tool for monitoring guest performance from host"
|
|
|
|
*
|
|
|
|
* Added a field to struct build_id_event that broke the file
|
|
|
|
* format.
|
|
|
|
*
|
|
|
|
* Since the kernel build-id is the first entry, process the
|
|
|
|
* table using the old format if the well known
|
|
|
|
* '[kernel.kallsyms]' string for the kernel build-id has the
|
|
|
|
* first 4 characters chopped off (where the pid_t sits).
|
|
|
|
*/
|
|
|
|
if (memcmp(filename, "nel.kallsyms]", 13) == 0) {
|
|
|
|
if (lseek(input, orig_offset, SEEK_SET) == (off_t)-1)
|
|
|
|
return -1;
|
|
|
|
return perf_header__read_build_ids_abi_quirk(header, input, offset, size);
|
|
|
|
}
|
2010-04-19 05:32:50 +00:00
|
|
|
|
|
|
|
__event_process_build_id(&bev, filename, session);
|
|
|
|
|
|
|
|
offset += bev.header.size;
|
|
|
|
}
|
|
|
|
err = 0;
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
static int perf_file_section__process(struct perf_file_section *section,
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
struct perf_header *ph,
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
int feat, int fd, void *data __used)
|
2009-11-16 18:32:43 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
if (lseek(fd, section->offset, SEEK_SET) == (off_t)-1) {
|
2011-01-22 22:37:02 +00:00
|
|
|
pr_debug("Failed to lseek to %" PRIu64 " offset for feature "
|
2011-03-09 11:13:19 +00:00
|
|
|
"%d, continuing...\n", section->offset, feat);
|
2009-11-16 18:32:43 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (feat) {
|
|
|
|
case HEADER_TRACE_INFO:
|
2010-05-01 06:41:20 +00:00
|
|
|
trace_report(fd, false);
|
2009-11-16 18:32:43 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
case HEADER_BUILD_ID:
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_header__read_build_ids(ph, fd, section->offset, section->size))
|
2009-11-16 18:32:43 +00:00
|
|
|
pr_debug("Failed to read buildids, continuing...\n");
|
|
|
|
break;
|
perf script: Fix unknown feature comment
"perf script -v" emits:
unknown feature 3, continuing...
unknown feature 4, continuing...
unknown feature 5, continuing...
unknown feature 6, continuing...
unknown feature 7, continuing...
unknown feature 8, continuing...
unknown feature 9, continuing...
unknown feature 10, continuing...
unknown feature 11, continuing...
unknown feature 12, continuing...
unknown feature 13, continuing...
unknown feature 14, continuing...
These are all new features added by fbe96f2. Update
perf_file_section__process to know they are valid feature ids.
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1318984464-20650-1-git-send-email-dsahern@gmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-10-19 00:34:24 +00:00
|
|
|
|
|
|
|
case HEADER_HOSTNAME:
|
|
|
|
case HEADER_OSRELEASE:
|
|
|
|
case HEADER_VERSION:
|
|
|
|
case HEADER_ARCH:
|
|
|
|
case HEADER_NRCPUS:
|
|
|
|
case HEADER_CPUDESC:
|
|
|
|
case HEADER_CPUID:
|
|
|
|
case HEADER_TOTAL_MEM:
|
|
|
|
case HEADER_CMDLINE:
|
|
|
|
case HEADER_EVENT_DESC:
|
|
|
|
case HEADER_CPU_TOPOLOGY:
|
|
|
|
case HEADER_NUMA_TOPOLOGY:
|
|
|
|
break;
|
|
|
|
|
2009-11-16 18:32:43 +00:00
|
|
|
default:
|
|
|
|
pr_debug("unknown feature %d, continuing...\n", feat);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2009-10-17 15:12:34 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
static int perf_file_header__read_pipe(struct perf_pipe_file_header *header,
|
2010-05-01 06:41:20 +00:00
|
|
|
struct perf_header *ph, int fd,
|
|
|
|
bool repipe)
|
2009-06-25 15:05:54 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
if (readn(fd, header, sizeof(*header)) <= 0 ||
|
|
|
|
memcmp(&header->magic, __perf_magic, sizeof(header->magic)))
|
2010-04-02 04:59:15 +00:00
|
|
|
return -1;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (repipe && do_write(STDOUT_FILENO, header, sizeof(*header)) < 0)
|
2010-05-01 06:41:20 +00:00
|
|
|
return -1;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (header->size != sizeof(*header)) {
|
|
|
|
u64 size = bswap_64(header->size);
|
2010-04-02 04:59:15 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (size != sizeof(*header))
|
2010-04-02 04:59:15 +00:00
|
|
|
return -1;
|
|
|
|
|
|
|
|
ph->needs_swap = true;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int perf_header__read_pipe(struct perf_session *session, int fd)
|
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
struct perf_header *header = &session->header;
|
2010-04-02 04:59:15 +00:00
|
|
|
struct perf_pipe_file_header f_header;
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_file_header__read_pipe(&f_header, header, fd,
|
2010-05-01 06:41:20 +00:00
|
|
|
session->repipe) < 0) {
|
2010-04-02 04:59:15 +00:00
|
|
|
pr_debug("incompatible file format\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
session->fd = fd;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
int perf_session__read_header(struct perf_session *session, int fd)
|
2010-04-02 04:59:15 +00:00
|
|
|
{
|
2011-03-09 11:13:19 +00:00
|
|
|
struct perf_header *header = &session->header;
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
struct perf_file_header f_header;
|
2009-06-25 15:05:54 +00:00
|
|
|
struct perf_file_attr f_attr;
|
|
|
|
u64 f_id;
|
|
|
|
int nr_attrs, nr_ids, i, j;
|
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
session->evlist = perf_evlist__new(NULL, NULL);
|
|
|
|
if (session->evlist == NULL)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2010-04-02 04:59:15 +00:00
|
|
|
if (session->fd_pipe)
|
|
|
|
return perf_header__read_pipe(session, fd);
|
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_file_header__read(&f_header, header, fd) < 0) {
|
2009-11-19 16:55:55 +00:00
|
|
|
pr_debug("incompatible file format\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2009-06-25 15:05:54 +00:00
|
|
|
|
|
|
|
nr_attrs = f_header.attrs.size / sizeof(f_attr);
|
|
|
|
lseek(fd, f_header.attrs.offset, SEEK_SET);
|
|
|
|
|
|
|
|
for (i = 0; i < nr_attrs; i++) {
|
2011-03-10 14:15:54 +00:00
|
|
|
struct perf_evsel *evsel;
|
2009-08-06 18:57:41 +00:00
|
|
|
off_t tmp;
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-07-15 18:34:09 +00:00
|
|
|
if (readn(fd, &f_attr, sizeof(f_attr)) <= 0)
|
2009-12-29 00:48:32 +00:00
|
|
|
goto out_errno;
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
|
2011-07-15 18:34:09 +00:00
|
|
|
if (header->needs_swap)
|
|
|
|
perf_event__attr_swap(&f_attr.attr);
|
|
|
|
|
2009-08-06 18:57:41 +00:00
|
|
|
tmp = lseek(fd, 0, SEEK_CUR);
|
2011-03-10 14:15:54 +00:00
|
|
|
evsel = perf_evsel__new(&f_attr.attr, i);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
if (evsel == NULL)
|
|
|
|
goto out_delete_evlist;
|
|
|
|
/*
|
|
|
|
* Do it before so that if perf_evsel__alloc_id fails, this
|
|
|
|
* entry gets purged too at perf_evlist__delete().
|
|
|
|
*/
|
|
|
|
perf_evlist__add(session->evlist, evsel);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
|
|
|
nr_ids = f_attr.ids.size / sizeof(u64);
|
2011-03-10 14:15:54 +00:00
|
|
|
/*
|
|
|
|
* We don't have the cpu and thread maps on the header, so
|
|
|
|
* for allocating the perf_sample_id table we fake 1 cpu and
|
|
|
|
* hattr->ids threads.
|
|
|
|
*/
|
|
|
|
if (perf_evsel__alloc_id(evsel, 1, nr_ids))
|
|
|
|
goto out_delete_evlist;
|
|
|
|
|
2009-06-25 15:05:54 +00:00
|
|
|
lseek(fd, f_attr.ids.offset, SEEK_SET);
|
|
|
|
|
|
|
|
for (j = 0; j < nr_ids; j++) {
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_header__getbuffer64(header, fd, &f_id, sizeof(f_id)))
|
2009-12-29 00:48:32 +00:00
|
|
|
goto out_errno;
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
perf_evlist__id_add(session->evlist, evsel, 0, j, f_id);
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
2009-11-17 03:18:09 +00:00
|
|
|
|
2009-06-25 15:05:54 +00:00
|
|
|
lseek(fd, tmp, SEEK_SET);
|
|
|
|
}
|
|
|
|
|
2011-11-12 00:17:32 +00:00
|
|
|
symbol_conf.nr_events = nr_attrs;
|
|
|
|
|
2009-09-12 05:52:51 +00:00
|
|
|
if (f_header.event_types.size) {
|
|
|
|
lseek(fd, f_header.event_types.offset, SEEK_SET);
|
|
|
|
events = malloc(f_header.event_types.size);
|
2009-11-19 16:55:55 +00:00
|
|
|
if (events == NULL)
|
|
|
|
return -ENOMEM;
|
2011-03-09 11:13:19 +00:00
|
|
|
if (perf_header__getbuffer64(header, fd, events,
|
perf tools: Cross platform perf.data analysis support
There are still some problems related to loading vmlinux files,
but those are unrelated to the feature implemented in this
patch, so will get fixed in the next patches, but here are some
results:
1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit
userland
2. transfer it to a Debian Testing machine, PARISC64, 32-bit
userland
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5
74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]
55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko
41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko
90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko
984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko
acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5
22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0
353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0
d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0
a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf
d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0
acme@parisc:~/git/linux-2.6-tip$
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm
The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms...
^^^^ The problem related to vmlinux handling, it shouldn't be trying this
^^^^ rather alien /proc/kallsyms at all...
/lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols
/lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols
/home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols
/usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols
Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols
<SNIP more complaints about not finding the right build-ids,
those will have to wait for 'perf archive' or plain
copying what was collected by 'perf record' on the x86_64,
source machine, see further below for an example of this >
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get when running the same command for the
same perf.data file on the F12, x86_64, source machine:
[root@doppio linux-2.6-tip]# perf report --sort comm
# Samples: 293085637
#
# Overhead Command
# ........ ...............
#
61.70% find
23.50% perf
5.86% swapper
3.12% sshd
2.39% init
0.87% bash
0.86% sleep
0.59% dbus-daemon
0.25% hald
0.24% NetworkManager
0.19% hald-addon-rfki
0.15% openvpn
0.07% phy0
0.07% events/0
0.05% iwl3945
0.05% events/1
0.03% kondemand/0
[root@doppio linux-2.6-tip]#
The other modes work as well, modulo the problem with vmlinux:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object
# ........ ............... .................................
#
35.11% find ffffffff81002b5a
18.25% perf ffffffff8102235f
16.17% find libc-2.10.2.so
9.07% find find
5.80% swapper ffffffff8102235f
3.95% perf libc-2.10.2.so
2.33% init ffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k
1.35% find [e1000e]
0.68% sleep libc-2.10.2.so
acme@parisc:~/git/linux-2.6-tip$
And the lack of the right buildids:
acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15
# Samples: 293085637
#
# Overhead Command Shared Object Symbol
# ........ ............... ................................. ......
#
35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a
18.25% perf ffffffff8102235f [k] 0xffffffff8102235f
16.17% find libc-2.10.2.so [.] 0x00000000045782
9.07% find find [.] 0x0000000000fb0e
5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f
3.95% perf libc-2.10.2.so [.] 0x0000000007f398
2.33% init ffffffff810091b9 [k] 0xffffffff810091b9
1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440
1.35% find [e1000e] [k] 0x00000000010948
0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b
acme@parisc:~/git/linux-2.6-tip$
But if we:
acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug
ls: cannot access /home/acme/.debug: No such file or directory
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/
acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/
acme@doppio's password:
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02
acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb
acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
acme@parisc:~/git/linux-2.6-tip$
Which matches what we get on the source, F12, x86_64 machine:
[root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so
# dso: libc-2.10.2.so
# Samples: 64281170
#
# Overhead Command Symbol
# ........ ............... ......
#
14.98% perf [.] __GI_strcmp
12.30% find [.] __GI_memmove
9.25% find [.] _int_malloc
7.60% find [.] _IO_vfprintf_internal
6.10% find [.] _IO_new_file_xsputn
6.02% find [.] __GI_close
3.08% find [.] _IO_file_overflow_internal
3.08% find [.] malloc_consolidate
3.08% find [.] _int_free
3.08% find [.] __strchrnul
3.08% find [.] __getdents64
3.08% find [.] __write_nocancel
3.08% sleep [.] __GI__dl_addr
3.08% sshd [.] __libc_select
3.08% find [.] _IO_new_file_write
3.07% find [.] _IO_new_do_write
3.06% find [.] __GI___errno_location
3.05% find [.] __GI___libc_malloc
3.04% perf [.] __GI_memcpy
1.71% find [.] __fprintf_chk
1.29% bash [.] __gconv_transform_utf8_internal
0.79% dbus-daemon [.] __GI_strlen
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
[root@doppio linux-2.6-tip]#
So I think this is really, really nice in that it demonstrates
the portability of perf.data files and the use of build-ids
accross such aliens worlds :-)
There are some things to fix tho, like the bitmap on the header,
but things are looking good.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-01-14 14:23:10 +00:00
|
|
|
f_header.event_types.size))
|
2009-12-29 00:48:32 +00:00
|
|
|
goto out_errno;
|
2009-09-12 05:52:51 +00:00
|
|
|
event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type);
|
|
|
|
}
|
2009-10-06 21:36:47 +00:00
|
|
|
|
perf tools: Make perf.data more self-descriptive (v8)
The goal of this patch is to include more information about the host
environment into the perf.data so it is more self-descriptive. Overtime,
profiles are captured on various machines and it becomes hard to track
what was recorded, on what machine and when.
This patch provides a way to solve this by extending the perf.data file
with basic information about the host machine. To add those extensions,
we leverage the feature bits capabilities of the perf.data format. The
change is backward compatible with existing perf.data files.
We define the following useful new extensions:
- HEADER_HOSTNAME: the hostname
- HEADER_OSRELEASE: the kernel release number
- HEADER_ARCH: the hw architecture
- HEADER_CPUDESC: generic CPU description
- HEADER_NRCPUS: number of online/avail cpus
- HEADER_CMDLINE: perf command line
- HEADER_VERSION: perf version
- HEADER_TOPOLOGY: cpu topology
- HEADER_EVENT_DESC: full event description (attrs)
- HEADER_CPUID: easy-to-parse low level CPU identication
The small granularity for the entries is to make it easier to extend
without breaking backward compatiblity. Many entries are provided as
ASCII strings.
Perf report/script have been modified to print the basic information as
easy-to-parse ASCII strings. Extended information about CPU and NUMA
topology may be requested with the -I option.
Thanks to David Ahern for reviewing and testing the many versions of
this patch.
$ perf report --stdio
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# HEADER_CPU_TOPOLOGY info available, use -I to display
# HEADER_NUMA_TOPOLOGY info available, use -I to display
# ========
#
...
$ perf report --stdio -I
# ========
# captured on : Mon Sep 26 15:22:14 2011
# hostname : quad
# os release : 3.1.0-rc4-tip
# perf version : 3.1.0-rc4
# arch : x86_64
# nrcpus online : 4
# nrcpus avail : 4
# cpudesc : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
# cpuid : GenuineIntel,6,15,11
# total memory : 8105360 kB
# cmdline : /home/eranian/perfmon/official/tip/build/tools/perf/perf record date
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, id = { 29, 30, 31,
# sibling cores : 0-3
# sibling threads : 0
# sibling threads : 1
# sibling threads : 2
# sibling threads : 3
# node0 meminfo : total = 8320608 kB, free = 7571024 kB
# node0 cpu list : 0-3
# ========
#
...
Reviewed-by: David Ahern <dsahern@gmail.com>
Tested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20110930134040.GA5575@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
[ committer notes: Use --show-info in the tools as was in the docs, rename
perf_header_fprintf_info to perf_file_section__fprintf_info, fixup
conflict with f69b64f7 "perf: Support setting the disassembler style" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-09-30 13:40:40 +00:00
|
|
|
perf_header__process_sections(header, fd, NULL,
|
|
|
|
perf_file_section__process);
|
2009-11-11 03:51:05 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
lseek(fd, header->data_offset, SEEK_SET);
|
2009-06-25 15:05:54 +00:00
|
|
|
|
2011-03-09 11:13:19 +00:00
|
|
|
header->frozen = 1;
|
2009-11-19 16:55:55 +00:00
|
|
|
return 0;
|
2009-12-29 00:48:32 +00:00
|
|
|
out_errno:
|
|
|
|
return -errno;
|
2011-03-10 14:15:54 +00:00
|
|
|
|
|
|
|
out_delete_evlist:
|
|
|
|
perf_evlist__delete(session->evlist);
|
|
|
|
session->evlist = NULL;
|
|
|
|
return -ENOMEM;
|
2009-06-25 15:05:54 +00:00
|
|
|
}
|
2009-08-16 18:56:37 +00:00
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_attr(struct perf_tool *tool,
|
2011-11-25 10:19:45 +00:00
|
|
|
struct perf_event_attr *attr, u16 ids, u64 *id,
|
2011-11-28 09:56:39 +00:00
|
|
|
perf_event__handler_t process)
|
2010-04-02 04:59:19 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
union perf_event *ev;
|
2010-04-02 04:59:19 +00:00
|
|
|
size_t size;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
size = sizeof(struct perf_event_attr);
|
|
|
|
size = ALIGN(size, sizeof(u64));
|
|
|
|
size += sizeof(struct perf_event_header);
|
|
|
|
size += ids * sizeof(u64);
|
|
|
|
|
|
|
|
ev = malloc(size);
|
|
|
|
|
2010-11-13 02:35:06 +00:00
|
|
|
if (ev == NULL)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2010-04-02 04:59:19 +00:00
|
|
|
ev->attr.attr = *attr;
|
|
|
|
memcpy(ev->attr.id, id, ids * sizeof(u64));
|
|
|
|
|
|
|
|
ev->attr.header.type = PERF_RECORD_HEADER_ATTR;
|
|
|
|
ev->attr.header.size = size;
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
err = process(tool, ev, NULL, NULL);
|
2010-04-02 04:59:19 +00:00
|
|
|
|
|
|
|
free(ev);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_attrs(struct perf_tool *tool,
|
2011-11-25 10:19:45 +00:00
|
|
|
struct perf_session *session,
|
2011-03-10 14:15:54 +00:00
|
|
|
perf_event__handler_t process)
|
2010-04-02 04:59:19 +00:00
|
|
|
{
|
2011-03-10 14:15:54 +00:00
|
|
|
struct perf_evsel *attr;
|
|
|
|
int err = 0;
|
2010-04-02 04:59:19 +00:00
|
|
|
|
2011-03-10 14:15:54 +00:00
|
|
|
list_for_each_entry(attr, &session->evlist->entries, node) {
|
2011-11-28 10:30:20 +00:00
|
|
|
err = perf_event__synthesize_attr(tool, &attr->attr, attr->ids,
|
2011-11-28 09:56:39 +00:00
|
|
|
attr->id, process);
|
2010-04-02 04:59:19 +00:00
|
|
|
if (err) {
|
|
|
|
pr_debug("failed to create perf header attribute\n");
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-01-29 16:01:45 +00:00
|
|
|
int perf_event__process_attr(union perf_event *event,
|
2011-11-12 00:45:41 +00:00
|
|
|
struct perf_evlist **pevlist)
|
2010-04-02 04:59:19 +00:00
|
|
|
{
|
|
|
|
unsigned int i, ids, n_ids;
|
2011-03-10 14:15:54 +00:00
|
|
|
struct perf_evsel *evsel;
|
2011-11-12 00:45:41 +00:00
|
|
|
struct perf_evlist *evlist = *pevlist;
|
2010-04-02 04:59:19 +00:00
|
|
|
|
2011-11-12 00:45:41 +00:00
|
|
|
if (evlist == NULL) {
|
|
|
|
*pevlist = evlist = perf_evlist__new(NULL, NULL);
|
|
|
|
if (evlist == NULL)
|
2011-03-10 14:15:54 +00:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2011-11-12 00:45:41 +00:00
|
|
|
evsel = perf_evsel__new(&event->attr.attr, evlist->nr_entries);
|
2011-03-10 14:15:54 +00:00
|
|
|
if (evsel == NULL)
|
2010-04-02 04:59:19 +00:00
|
|
|
return -ENOMEM;
|
|
|
|
|
2011-11-12 00:45:41 +00:00
|
|
|
perf_evlist__add(evlist, evsel);
|
2011-03-10 14:15:54 +00:00
|
|
|
|
2011-01-29 16:01:45 +00:00
|
|
|
ids = event->header.size;
|
|
|
|
ids -= (void *)&event->attr.id - (void *)event;
|
2010-04-02 04:59:19 +00:00
|
|
|
n_ids = ids / sizeof(u64);
|
2011-03-10 14:15:54 +00:00
|
|
|
/*
|
|
|
|
* We don't have the cpu and thread maps on the header, so
|
|
|
|
* for allocating the perf_sample_id table we fake 1 cpu and
|
|
|
|
* hattr->ids threads.
|
|
|
|
*/
|
|
|
|
if (perf_evsel__alloc_id(evsel, 1, n_ids))
|
|
|
|
return -ENOMEM;
|
2010-04-02 04:59:19 +00:00
|
|
|
|
|
|
|
for (i = 0; i < n_ids; i++) {
|
2011-11-12 00:45:41 +00:00
|
|
|
perf_evlist__id_add(evlist, evsel, 0, i, event->attr.id[i]);
|
2010-04-02 04:59:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2010-04-02 04:59:20 +00:00
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_event_type(struct perf_tool *tool,
|
2011-11-25 10:19:45 +00:00
|
|
|
u64 event_id, char *name,
|
2011-01-29 16:01:45 +00:00
|
|
|
perf_event__handler_t process,
|
2011-11-28 09:56:39 +00:00
|
|
|
struct machine *machine)
|
2010-04-02 04:59:20 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
union perf_event ev;
|
2010-04-02 04:59:20 +00:00
|
|
|
size_t size = 0;
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
memset(&ev, 0, sizeof(ev));
|
|
|
|
|
|
|
|
ev.event_type.event_type.event_id = event_id;
|
|
|
|
memset(ev.event_type.event_type.name, 0, MAX_EVENT_NAME);
|
|
|
|
strncpy(ev.event_type.event_type.name, name, MAX_EVENT_NAME - 1);
|
|
|
|
|
|
|
|
ev.event_type.header.type = PERF_RECORD_HEADER_EVENT_TYPE;
|
|
|
|
size = strlen(name);
|
|
|
|
size = ALIGN(size, sizeof(u64));
|
|
|
|
ev.event_type.header.size = sizeof(ev.event_type) -
|
|
|
|
(sizeof(ev.event_type.event_type.name) - size);
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
err = process(tool, &ev, NULL, machine);
|
2010-04-02 04:59:20 +00:00
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_event_types(struct perf_tool *tool,
|
2011-11-25 10:19:45 +00:00
|
|
|
perf_event__handler_t process,
|
2011-11-28 09:56:39 +00:00
|
|
|
struct machine *machine)
|
2010-04-02 04:59:20 +00:00
|
|
|
{
|
|
|
|
struct perf_trace_event_type *type;
|
|
|
|
int i, err = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < event_count; i++) {
|
|
|
|
type = &events[i];
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
err = perf_event__synthesize_event_type(tool, type->event_id,
|
2011-01-29 16:01:45 +00:00
|
|
|
type->name, process,
|
2011-11-28 09:56:39 +00:00
|
|
|
machine);
|
2010-04-02 04:59:20 +00:00
|
|
|
if (err) {
|
|
|
|
pr_debug("failed to create perf header event type\n");
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__process_event_type(struct perf_tool *tool __unused,
|
2011-11-28 09:56:39 +00:00
|
|
|
union perf_event *event)
|
2010-04-02 04:59:20 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
if (perf_header__push_event(event->event_type.event_type.event_id,
|
|
|
|
event->event_type.event_type.name) < 0)
|
2010-04-02 04:59:20 +00:00
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2010-04-02 04:59:21 +00:00
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_tracing_data(struct perf_tool *tool, int fd,
|
2011-11-25 10:19:45 +00:00
|
|
|
struct perf_evlist *evlist,
|
2011-11-28 09:56:39 +00:00
|
|
|
perf_event__handler_t process)
|
2010-04-02 04:59:21 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
union perf_event ev;
|
2011-10-20 13:59:43 +00:00
|
|
|
struct tracing_data *tdata;
|
2010-04-02 04:59:21 +00:00
|
|
|
ssize_t size = 0, aligned_size = 0, padding;
|
2011-01-24 16:13:04 +00:00
|
|
|
int err __used = 0;
|
2010-04-02 04:59:21 +00:00
|
|
|
|
2011-10-20 13:59:43 +00:00
|
|
|
/*
|
|
|
|
* We are going to store the size of the data followed
|
|
|
|
* by the data contents. Since the fd descriptor is a pipe,
|
|
|
|
* we cannot seek back to store the size of the data once
|
|
|
|
* we know it. Instead we:
|
|
|
|
*
|
|
|
|
* - write the tracing data to the temp file
|
|
|
|
* - get/write the data size to pipe
|
|
|
|
* - write the tracing data from the temp file
|
|
|
|
* to the pipe
|
|
|
|
*/
|
|
|
|
tdata = tracing_data_get(&evlist->entries, fd, true);
|
|
|
|
if (!tdata)
|
|
|
|
return -1;
|
|
|
|
|
2010-04-02 04:59:21 +00:00
|
|
|
memset(&ev, 0, sizeof(ev));
|
|
|
|
|
|
|
|
ev.tracing_data.header.type = PERF_RECORD_HEADER_TRACING_DATA;
|
2011-10-20 13:59:43 +00:00
|
|
|
size = tdata->size;
|
2010-04-02 04:59:21 +00:00
|
|
|
aligned_size = ALIGN(size, sizeof(u64));
|
|
|
|
padding = aligned_size - size;
|
|
|
|
ev.tracing_data.header.size = sizeof(ev.tracing_data);
|
|
|
|
ev.tracing_data.size = aligned_size;
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
process(tool, &ev, NULL, NULL);
|
2010-04-02 04:59:21 +00:00
|
|
|
|
2011-10-20 13:59:43 +00:00
|
|
|
/*
|
|
|
|
* The put function will copy all the tracing data
|
|
|
|
* stored in temp file to the pipe.
|
|
|
|
*/
|
|
|
|
tracing_data_put(tdata);
|
|
|
|
|
2010-04-02 04:59:21 +00:00
|
|
|
write_padded(fd, NULL, 0, padding);
|
|
|
|
|
|
|
|
return aligned_size;
|
|
|
|
}
|
|
|
|
|
2011-01-29 16:01:45 +00:00
|
|
|
int perf_event__process_tracing_data(union perf_event *event,
|
|
|
|
struct perf_session *session)
|
2010-04-02 04:59:21 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
ssize_t size_read, padding, size = event->tracing_data.size;
|
2010-04-02 04:59:21 +00:00
|
|
|
off_t offset = lseek(session->fd, 0, SEEK_CUR);
|
|
|
|
char buf[BUFSIZ];
|
|
|
|
|
|
|
|
/* setup for reading amidst mmap */
|
|
|
|
lseek(session->fd, offset + sizeof(struct tracing_data_event),
|
|
|
|
SEEK_SET);
|
|
|
|
|
2010-05-01 06:41:20 +00:00
|
|
|
size_read = trace_report(session->fd, session->repipe);
|
2010-04-02 04:59:21 +00:00
|
|
|
|
|
|
|
padding = ALIGN(size_read, sizeof(u64)) - size_read;
|
|
|
|
|
|
|
|
if (read(session->fd, buf, padding) < 0)
|
|
|
|
die("reading input file");
|
2010-05-01 06:41:20 +00:00
|
|
|
if (session->repipe) {
|
|
|
|
int retw = write(STDOUT_FILENO, buf, padding);
|
|
|
|
if (retw <= 0 || retw != padding)
|
|
|
|
die("repiping tracing data padding");
|
|
|
|
}
|
2010-04-02 04:59:21 +00:00
|
|
|
|
|
|
|
if (size_read + padding != size)
|
|
|
|
die("tracing data size mismatch");
|
|
|
|
|
|
|
|
return size_read + padding;
|
|
|
|
}
|
2010-04-02 04:59:22 +00:00
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__synthesize_build_id(struct perf_tool *tool,
|
2011-11-25 10:19:45 +00:00
|
|
|
struct dso *pos, u16 misc,
|
2011-01-29 16:01:45 +00:00
|
|
|
perf_event__handler_t process,
|
2011-11-28 09:56:39 +00:00
|
|
|
struct machine *machine)
|
2010-04-02 04:59:22 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
union perf_event ev;
|
2010-04-02 04:59:22 +00:00
|
|
|
size_t len;
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
if (!pos->hit)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
memset(&ev, 0, sizeof(ev));
|
|
|
|
|
|
|
|
len = pos->long_name_len + 1;
|
|
|
|
len = ALIGN(len, NAME_ALIGN);
|
|
|
|
memcpy(&ev.build_id.build_id, pos->build_id, sizeof(pos->build_id));
|
|
|
|
ev.build_id.header.type = PERF_RECORD_HEADER_BUILD_ID;
|
|
|
|
ev.build_id.header.misc = misc;
|
2010-04-28 00:17:50 +00:00
|
|
|
ev.build_id.pid = machine->pid;
|
2010-04-02 04:59:22 +00:00
|
|
|
ev.build_id.header.size = sizeof(ev.build_id) + len;
|
|
|
|
memcpy(&ev.build_id.filename, pos->long_name, pos->long_name_len);
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
err = process(tool, &ev, NULL, machine);
|
2010-04-02 04:59:22 +00:00
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-28 10:30:20 +00:00
|
|
|
int perf_event__process_build_id(struct perf_tool *tool __used,
|
2011-11-25 10:19:45 +00:00
|
|
|
union perf_event *event,
|
2011-01-29 16:01:45 +00:00
|
|
|
struct perf_session *session)
|
2010-04-02 04:59:22 +00:00
|
|
|
{
|
2011-01-29 16:01:45 +00:00
|
|
|
__event_process_build_id(&event->build_id,
|
|
|
|
event->build_id.filename,
|
2010-04-19 05:32:50 +00:00
|
|
|
session);
|
2010-04-02 04:59:22 +00:00
|
|
|
return 0;
|
|
|
|
}
|
2010-06-17 09:39:01 +00:00
|
|
|
|
|
|
|
void disable_buildid_cache(void)
|
|
|
|
{
|
|
|
|
no_buildid_cache = true;
|
|
|
|
}
|