WSL2-Linux-Kernel/tools/perf/util/header.c

1195 строки
28 KiB
C
Исходник Обычный вид История

#define _FILE_OFFSET_BITS 64
#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 17:23:10 +03:00
#include <byteswap.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#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 17:23:10 +03:00
#include <linux/kernel.h>
#include "evlist.h"
#include "evsel.h"
#include "util.h"
#include "header.h"
#include "../perf.h"
#include "trace-event.h"
#include "session.h"
#include "symbol.h"
#include "debug.h"
static bool no_buildid_cache = false;
static int event_count;
static struct perf_trace_event_type *events;
int perf_header__push_event(u64 id, const char *name)
{
if (strlen(name) > MAX_EVENT_NAME)
pr_warning("Event %s will be truncated\n", name);
if (!events) {
events = malloc(sizeof(struct perf_trace_event_type));
if (events == NULL)
return -ENOMEM;
} else {
struct perf_trace_event_type *nevents;
nevents = realloc(events, (event_count + 1) * sizeof(*events));
if (nevents == NULL)
return -ENOMEM;
events = nevents;
}
memset(&events[event_count], 0, sizeof(struct perf_trace_event_type));
events[event_count].event_id = id;
strncpy(events[event_count].name, name, MAX_EVENT_NAME - 1);
event_count++;
return 0;
}
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;
}
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 14:02:48 +04:00
struct perf_event_attr attr;
struct perf_file_section ids;
};
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 23:50:43 +03:00
{
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 23:50:43 +03:00
}
void perf_header__clear_feat(struct perf_header *header, int feat)
{
clear_bit(feat, header->adds_features);
}
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 23:50:43 +03:00
{
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 23:50:43 +03:00
}
static int do_write(int fd, const void *buf, size_t size)
{
while (size) {
int ret = write(fd, buf, size);
if (ret < 0)
return -errno;
size -= ret;
buf += ret;
}
return 0;
}
#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 record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
#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;
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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)
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
{
const size_t size = PATH_MAX;
char *realname, *filename = zalloc(size),
*linkname = zalloc(size), *targetname;
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
int len, err = -1;
perf symbols: Handle /proc/sys/kernel/kptr_restrict Perf uses /proc/modules to figure out where kernel modules are loaded. With the advent of kptr_restrict, non root users get zeroes for all module start addresses. So check if kptr_restrict is non zero and don't generate the syntethic PERF_RECORD_MMAP events for them. Warn the user about it in perf record and in perf report. In perf report the reference relocation symbol being zero means that kptr_restrict was set, thus /proc/kallsyms has only zeroed addresses, so don't use it to fixup symbol addresses when using a valid kallsyms (in the buildid cache) or vmlinux (in the vmlinux path) build-id located automatically or specified by the user. Provide an explanation about it in 'perf report' if kernel samples were taken, checking if a suitable vmlinux or kallsyms was found/specified. Restricted /proc/kallsyms don't go to the buildid cache anymore. Example: [acme@emilia ~]$ perf record -F 100000 sleep 1 WARNING: Kernel address maps (/proc/{kallsyms,modules}) are restricted, check /proc/sys/kernel/kptr_restrict. Samples in kernel functions may not be resolved if a suitable vmlinux file is not found in the buildid cache or in the vmlinux path. Samples in kernel modules won't be resolved at all. If some relocation was applied (e.g. kexec) symbols may be misresolved even with a suitable vmlinux or kallsyms file. [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.005 MB perf.data (~231 samples) ] [acme@emilia ~]$ [acme@emilia ~]$ perf report --stdio Kernel address maps (/proc/{kallsyms,modules}) were restricted, check /proc/sys/kernel/kptr_restrict before running 'perf record'. If some relocation was applied (e.g. kexec) symbols may be misresolved. Samples in kernel modules can't be resolved as well. # Events: 13 cycles # # Overhead Command Shared Object Symbol # ........ ....... ................. ..................... # 20.24% sleep [kernel.kallsyms] [k] page_fault 20.04% sleep [kernel.kallsyms] [k] filemap_fault 19.78% sleep [kernel.kallsyms] [k] __lru_cache_add 19.69% sleep ld-2.12.so [.] memcpy 14.71% sleep [kernel.kallsyms] [k] dput 4.70% sleep [kernel.kallsyms] [k] flush_signal_handlers 0.73% sleep [kernel.kallsyms] [k] perf_event_comm 0.11% sleep [kernel.kallsyms] [k] native_write_msr_safe # # (For a higher level overview, try: perf report --sort comm,dso) # [acme@emilia ~]$ This is because it found a suitable vmlinux (build-id checked) in /lib/modules/2.6.39-rc7+/build/vmlinux (use -v in perf report to see the long file name). If we remove that file from the vmlinux path: [root@emilia ~]# mv /lib/modules/2.6.39-rc7+/build/vmlinux \ /lib/modules/2.6.39-rc7+/build/vmlinux.OFF [acme@emilia ~]$ perf report --stdio [kernel.kallsyms] with build id 57298cdbe0131f6871667ec0eaab4804dcf6f562 not found, continuing without symbols Kernel address maps (/proc/{kallsyms,modules}) were restricted, check /proc/sys/kernel/kptr_restrict before running 'perf record'. As no suitable kallsyms nor vmlinux was found, kernel samples can't be resolved. Samples in kernel modules can't be resolved as well. # Events: 13 cycles # # Overhead Command Shared Object Symbol # ........ ....... ................. ...... # 80.31% sleep [kernel.kallsyms] [k] 0xffffffff8103425a 19.69% sleep ld-2.12.so [.] memcpy # # (For a higher level overview, try: perf report --sort comm,dso) # [acme@emilia ~]$ Reported-by: Stephane Eranian <eranian@google.com> Suggested-by: David Miller <davem@davemloft.net> Cc: Dave Jones <davej@redhat.com> Cc: David Miller <davem@davemloft.net> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Kees Cook <kees.cook@canonical.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Tom Zanussi <tzanussi@gmail.com> Link: http://lkml.kernel.org/n/tip-mt512joaxxbhhp1odop04yit@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-05-26 16:53:51 +04:00
if (is_kallsyms) {
if (symbol_conf.kptr_restrict) {
pr_debug("Not caching a kptr_restrict'ed /proc/kallsyms\n");
return 0;
}
realname = (char *)name;
perf symbols: Handle /proc/sys/kernel/kptr_restrict Perf uses /proc/modules to figure out where kernel modules are loaded. With the advent of kptr_restrict, non root users get zeroes for all module start addresses. So check if kptr_restrict is non zero and don't generate the syntethic PERF_RECORD_MMAP events for them. Warn the user about it in perf record and in perf report. In perf report the reference relocation symbol being zero means that kptr_restrict was set, thus /proc/kallsyms has only zeroed addresses, so don't use it to fixup symbol addresses when using a valid kallsyms (in the buildid cache) or vmlinux (in the vmlinux path) build-id located automatically or specified by the user. Provide an explanation about it in 'perf report' if kernel samples were taken, checking if a suitable vmlinux or kallsyms was found/specified. Restricted /proc/kallsyms don't go to the buildid cache anymore. Example: [acme@emilia ~]$ perf record -F 100000 sleep 1 WARNING: Kernel address maps (/proc/{kallsyms,modules}) are restricted, check /proc/sys/kernel/kptr_restrict. Samples in kernel functions may not be resolved if a suitable vmlinux file is not found in the buildid cache or in the vmlinux path. Samples in kernel modules won't be resolved at all. If some relocation was applied (e.g. kexec) symbols may be misresolved even with a suitable vmlinux or kallsyms file. [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.005 MB perf.data (~231 samples) ] [acme@emilia ~]$ [acme@emilia ~]$ perf report --stdio Kernel address maps (/proc/{kallsyms,modules}) were restricted, check /proc/sys/kernel/kptr_restrict before running 'perf record'. If some relocation was applied (e.g. kexec) symbols may be misresolved. Samples in kernel modules can't be resolved as well. # Events: 13 cycles # # Overhead Command Shared Object Symbol # ........ ....... ................. ..................... # 20.24% sleep [kernel.kallsyms] [k] page_fault 20.04% sleep [kernel.kallsyms] [k] filemap_fault 19.78% sleep [kernel.kallsyms] [k] __lru_cache_add 19.69% sleep ld-2.12.so [.] memcpy 14.71% sleep [kernel.kallsyms] [k] dput 4.70% sleep [kernel.kallsyms] [k] flush_signal_handlers 0.73% sleep [kernel.kallsyms] [k] perf_event_comm 0.11% sleep [kernel.kallsyms] [k] native_write_msr_safe # # (For a higher level overview, try: perf report --sort comm,dso) # [acme@emilia ~]$ This is because it found a suitable vmlinux (build-id checked) in /lib/modules/2.6.39-rc7+/build/vmlinux (use -v in perf report to see the long file name). If we remove that file from the vmlinux path: [root@emilia ~]# mv /lib/modules/2.6.39-rc7+/build/vmlinux \ /lib/modules/2.6.39-rc7+/build/vmlinux.OFF [acme@emilia ~]$ perf report --stdio [kernel.kallsyms] with build id 57298cdbe0131f6871667ec0eaab4804dcf6f562 not found, continuing without symbols Kernel address maps (/proc/{kallsyms,modules}) were restricted, check /proc/sys/kernel/kptr_restrict before running 'perf record'. As no suitable kallsyms nor vmlinux was found, kernel samples can't be resolved. Samples in kernel modules can't be resolved as well. # Events: 13 cycles # # Overhead Command Shared Object Symbol # ........ ....... ................. ...... # 80.31% sleep [kernel.kallsyms] [k] 0xffffffff8103425a 19.69% sleep ld-2.12.so [.] memcpy # # (For a higher level overview, try: perf report --sort comm,dso) # [acme@emilia ~]$ Reported-by: Stephane Eranian <eranian@google.com> Suggested-by: David Miller <davem@davemloft.net> Cc: Dave Jones <davej@redhat.com> Cc: David Miller <davem@davemloft.net> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Kees Cook <kees.cook@canonical.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Tom Zanussi <tzanussi@gmail.com> Link: http://lkml.kernel.org/n/tip-mt512joaxxbhhp1odop04yit@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-05-26 16:53:51 +04:00
} else
realname = realpath(name, NULL);
if (realname == NULL || filename == NULL || linkname == NULL)
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
goto out_free;
len = snprintf(filename, size, "%s%s%s",
debugdir, is_kallsyms ? "/" : "", realname);
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
if (mkdir_p(filename, 0755))
goto out_free;
snprintf(filename + len, sizeof(filename) - len, "/%s", sbuild_id);
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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;
}
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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);
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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) < 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);
}
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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)
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
{
struct rb_node *nd;
int ret;
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
char debugdir[PATH_MAX];
perf buildid: add perfconfig option to specify buildid cache dir This patch adds the ability to specify an alternate directory to store the buildid cache (buildids, copy of binaries). By default, it is hardcoded to $HOME/.debug. This directory contains immutable data. The layout of the directory is such that no conflicts in filenames are possible. A modification in a file, yields a different buildid and thus a different location in the subdir hierarchy. You may want to put the buildid cache elsewhere because of disk space limitation or simply to share the cache between users. It is also useful for remote collect vs. local analysis of profiles. This patch adds a new config option to the perfconfig file. Under the tag 'buildid', there is a dir option. For instance, if you have: $ cat /etc/perfconfig [buildid] dir = /var/cache/perf-buildid All buildids and binaries are be saved in the directory specified. The perf record, buildid-list, buildid-cache, report, annotate, and archive commands will it to pull information out. The option can be set in the system-wide perfconfig file or in the $HOME/.perfconfig file. Cc: David S. Miller <davem@davemloft.net> Cc: Frédéric 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: Tom Zanussi <tzanussi@gmail.com> LKML-Reference: <4c055fb7.df0ce30a.5f0d.ffffae52@mx.google.com> Signed-off-by: Stephane Eranian <eranian@google.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-06-01 23:25:01 +04:00
snprintf(debugdir, sizeof(debugdir), "%s", buildid_dir);
perf record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
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 record: Introduce a symtab cache Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. 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: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-12-28 02:37:06 +03:00
}
static int perf_header__adds_write(struct perf_header *header,
struct perf_evlist *evlist, int fd)
{
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
int nr_sections;
struct perf_session *session;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
struct perf_file_section *feat_sec;
int sec_size;
u64 sec_start;
int idx = 0, err;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
session = container_of(header, struct perf_session, header);
if (perf_header__has_feat(header, HEADER_BUILD_ID &&
!perf_session__read_build_ids(session, true)))
perf_header__clear_feat(header, HEADER_BUILD_ID);
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
if (!nr_sections)
return 0;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
feat_sec = calloc(sizeof(*feat_sec), nr_sections);
if (feat_sec == NULL)
return -ENOMEM;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
sec_size = sizeof(*feat_sec) * nr_sections;
sec_start = header->data_offset + header->data_size;
lseek(fd, sec_start + sec_size, SEEK_SET);
if (perf_header__has_feat(header, HEADER_TRACE_INFO)) {
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
struct perf_file_section *trace_sec;
trace_sec = &feat_sec[idx++];
/* Write trace info */
trace_sec->offset = lseek(fd, 0, SEEK_CUR);
read_tracing_data(fd, &evlist->entries);
trace_sec->size = lseek(fd, 0, SEEK_CUR) - trace_sec->offset;
}
if (perf_header__has_feat(header, HEADER_BUILD_ID)) {
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
struct perf_file_section *buildid_sec;
buildid_sec = &feat_sec[idx++];
/* Write build-ids */
buildid_sec->offset = lseek(fd, 0, SEEK_CUR);
err = dsos__write_buildid_table(header, fd);
if (err < 0) {
pr_debug("failed to write buildid table\n");
goto out_free;
}
buildid_sec->size = lseek(fd, 0, SEEK_CUR) -
buildid_sec->offset;
if (!no_buildid_cache)
perf_session__cache_build_ids(session);
}
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
lseek(fd, sec_start, SEEK_SET);
err = do_write(fd, feat_sec, sec_size);
if (err < 0)
pr_debug("failed to write feature section\n");
out_free:
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
free(feat_sec);
return err;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
}
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;
}
int perf_session__write_header(struct perf_session *session,
struct perf_evlist *evlist,
int fd, bool at_exit)
{
struct perf_file_header f_header;
struct perf_file_attr f_attr;
struct perf_header *header = &session->header;
struct perf_evsel *attr, *pair = NULL;
int err;
lseek(fd, sizeof(f_header), SEEK_SET);
if (session->evlist != evlist)
pair = list_entry(session->evlist->entries.next, struct perf_evsel, node);
list_for_each_entry(attr, &evlist->entries, node) {
attr->id_offset = lseek(fd, 0, SEEK_CUR);
err = do_write(fd, attr->id, attr->ids * sizeof(u64));
if (err < 0) {
out_err_write:
pr_debug("failed to write perf header\n");
return err;
}
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);
}
}
header->attr_offset = lseek(fd, 0, SEEK_CUR);
list_for_each_entry(attr, &evlist->entries, node) {
f_attr = (struct perf_file_attr){
.attr = attr->attr,
.ids = {
.offset = attr->id_offset,
.size = attr->ids * sizeof(u64),
}
};
err = do_write(fd, &f_attr, sizeof(f_attr));
if (err < 0) {
pr_debug("failed to write perf header attribute\n");
return err;
}
}
header->event_offset = lseek(fd, 0, SEEK_CUR);
header->event_size = event_count * sizeof(struct perf_trace_event_type);
if (events) {
err = do_write(fd, events, header->event_size);
if (err < 0) {
pr_debug("failed to write perf header events\n");
return err;
}
}
header->data_offset = lseek(fd, 0, SEEK_CUR);
if (at_exit) {
err = perf_header__adds_write(header, evlist, fd);
if (err < 0)
return err;
}
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
f_header = (struct perf_file_header){
.magic = PERF_MAGIC,
.size = sizeof(f_header),
.attr_size = sizeof(f_attr),
.attrs = {
.offset = header->attr_offset,
.size = evlist->nr_entries * sizeof(f_attr),
},
.data = {
.offset = header->data_offset,
.size = header->data_size,
},
.event_types = {
.offset = header->event_offset,
.size = header->event_size,
},
};
memcpy(&f_header.adds_features, &header->adds_features, sizeof(header->adds_features));
lseek(fd, 0, SEEK_SET);
err = do_write(fd, &f_header, sizeof(f_header));
if (err < 0) {
pr_debug("failed to write perf header\n");
return err;
}
lseek(fd, header->data_offset + header->data_size, SEEK_SET);
header->frozen = 1;
return 0;
}
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 17:23:10 +03:00
int fd, void *buf, size_t size)
{
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 17:23:10 +03:00
return -1;
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 17:23:10 +03:00
mem_bswap_64(buf, size);
return 0;
}
int perf_header__process_sections(struct perf_header *header, int fd,
int (*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 17:23:10 +03:00
struct perf_header *ph,
int feat, int fd))
{
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
struct perf_file_section *feat_sec;
int nr_sections;
int sec_size;
int idx = 0;
int err = -1, feat = 1;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS);
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
if (!nr_sections)
return 0;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
feat_sec = calloc(sizeof(*feat_sec), nr_sections);
if (!feat_sec)
return -1;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
sec_size = sizeof(*feat_sec) * nr_sections;
lseek(fd, header->data_offset + header->data_size, SEEK_SET);
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
if (perf_header__getbuffer64(header, fd, feat_sec, sec_size))
goto out_free;
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
err = 0;
while (idx < nr_sections && feat < HEADER_LAST_FEATURE) {
if (perf_header__has_feat(header, feat)) {
struct perf_file_section *sec = &feat_sec[idx++];
err = process(sec, header, feat, fd);
if (err < 0)
break;
}
++feat;
}
out_free:
free(feat_sec);
return err;
}
int perf_file_header__read(struct perf_file_header *header,
struct perf_header *ph, int fd)
{
lseek(fd, 0, SEEK_SET);
if (readn(fd, header, sizeof(*header)) <= 0 ||
memcmp(&header->magic, __perf_magic, sizeof(header->magic)))
return -1;
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 17:23:10 +03:00
if (attr_size != sizeof(struct perf_file_attr))
return -1;
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 17:23:10 +03:00
adds_features));
ph->needs_swap = true;
}
if (header->size != sizeof(*header)) {
/* Support the previous format */
if (header->size == offsetof(typeof(*header), adds_features))
bitmap_zero(header->adds_features, HEADER_FEAT_BITS);
else
return -1;
}
perf tools: Bring linear set of section headers for features Build a set of section headers for features right after the datas. Each implemented feature will have one of such section header that provides the offset and the size of the data manipulated by the feature. The trace informations have moved after the data and are recorded on exit time. The new layout is as follows: ----------------------- ___ [ magic ] | [ header size ] | [ attr size ] | [ attr content offset ] | [ attr content size ] | [ data offset ] File Headers [ data size ] | [ event_types offset ] | [ event_types size ] | [ feature bitmap ] v [ attr section ] [ events section ] ___ [ X ] | [ X ] | [ X ] Datas [ X ] | [ X ] v ___ [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 2 offset ] | [ Feature 2 size ] v [ Feature 1 content ] [ Feature 2 content ] ----------------------- We have as many feature's section headers as we have features in use for the current file. Say Feat 1 and Feat 3 are used by the file, but not Feat 2. Then the feature headers will be like follows: [ Feature 1 offset ] | [ Feature 1 size ] Features headers [ Feature 3 offset ] | [ Feature 3 size ] v There is no hole to cover Feature 2 that is not in use here. We only need to cover the needed headers in order, from the lowest feature bit to the highest. Currently we have two features: HEADER_TRACE_INFO and HEADER_BUILD_ID. Both have their contents that follow the feature headers. Putting the contents right after the feature headers is not mandatory though. While we keep the feature headers right after the data and in order, their offsets can point everywhere. We have just put the two above feature contents in the end of the file for convenience. The purpose of this layout change is to have a file format that scales while keeping it simple: having such linear feature headers is less error prone wrt forward/backward compatibility as the content of a feature can be put anywhere, its location can even change by the time, it's fine because its headers will tell where it is. And we know how to find these headers, following the above rules. Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Paul Mackerras <paulus@samba.org> Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> LKML-Reference: <1257911467-28276-6-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-11 06:51:07 +03:00
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 17:23:10 +03:00
sizeof(ph->adds_features));
/*
* FIXME: hack that assumes that if we need swap the perf.data file
* may be coming from an arch with a different word-size, ergo different
* DEFINE_BITMAP format, investigate more later, but for now its mostly
* safe to assume that we have a build-id section. Trace files probably
* have several other issues in this realm anyway...
*/
if (ph->needs_swap) {
memset(&ph->adds_features, 0, sizeof(ph->adds_features));
perf_header__set_feat(ph, HEADER_BUILD_ID);
}
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;
return 0;
}
static int __event_process_build_id(struct build_id_event *bev,
char *filename,
struct perf_session *session)
{
int err = -1;
struct list_head *head;
struct machine *machine;
u16 misc;
struct dso *dso;
enum dso_kernel_type dso_type;
machine = perf_session__findnew_machine(session, bev->pid);
if (!machine)
goto out;
misc = bev->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
switch (misc) {
case PERF_RECORD_MISC_KERNEL:
dso_type = DSO_TYPE_KERNEL;
head = &machine->kernel_dsos;
break;
case PERF_RECORD_MISC_GUEST_KERNEL:
dso_type = DSO_TYPE_GUEST_KERNEL;
head = &machine->kernel_dsos;
break;
case PERF_RECORD_MISC_USER:
case PERF_RECORD_MISC_GUEST_USER:
dso_type = DSO_TYPE_USER;
head = &machine->user_dsos;
break;
default:
goto out;
}
dso = __dsos__findnew(head, filename);
if (dso != NULL) {
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
char sbuild_id[BUILD_ID_SIZE * 2 + 1];
dso__set_build_id(dso, &bev->build_id);
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
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);
}
err = 0;
out:
return err;
}
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;
perf report: Use properly build_id kernel binaries If we bring the recorded perf data together with kernel binary from another machine using: on server A: perf archive on server B: tar xjvf perf.data.tar.bz2 -C ~/.debug the build_id kernel dso is not properly recognized during the "perf report" command on server B. The reason is, that build_id dsos are added during the session initialization, while the kernel maps are created during the sample event processing. The machine__create_kernel_maps functions ends up creating new dso object for kernel, but it does not check if we already have one added by build_id processing. Also the build_id reading ABI quirk added in commit: - commit b25114817a73bbd2b84ce9dba02ee1ef8989a947 perf build-id: Add quirk to deal with perf.data file format breakage populates the "struct build_id_event::pid" with 0, which is later interpreted as DEFAULT_GUEST_KERNEL_ID. This is not always correct, so it's better to guess the pid value based on the "struct build_id_event::header::misc" value. - Tested with data generated on x86 kernel version v2.6.34 and reported back on x86_64 current kernel. - Not tested for guest kernel case. Note the problem stays for PERF_RECORD_MMAP events recorded by perf that does not use proper pid (HOST_KERNEL_ID/DEFAULT_GUEST_KERNEL_ID). They are misinterpreted within the current perf code. Probably there's not much we can do about that. Cc: Avi Kivity <avi@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Yanmin Zhang <yanmin_zhang@linux.intel.com> Link: http://lkml.kernel.org/r/20110601194346.GB1934@jolsa.brq.redhat.com Signed-off-by: Jiri Olsa <jolsa@redhat.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2011-06-01 23:43:46 +04:00
/*
* 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;
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;
}
static int perf_header__read_build_ids(struct perf_header *header,
int input, u64 offset, u64 size)
{
struct perf_session *session = container_of(header, struct perf_session, header);
struct build_id_event bev;
char filename[PATH_MAX];
u64 limit = offset + size, orig_offset = offset;
int err = -1;
while (offset < limit) {
ssize_t len;
if (read(input, &bev, sizeof(bev)) != sizeof(bev))
goto out;
if (header->needs_swap)
perf_event_header__bswap(&bev.header);
len = bev.header.size - sizeof(bev);
if (read(input, filename, len) != len)
goto out;
/*
* 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);
}
__event_process_build_id(&bev, filename, session);
offset += bev.header.size;
}
err = 0;
out:
return err;
}
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 17:23:10 +03:00
struct perf_header *ph,
int feat, int fd)
{
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;
}
switch (feat) {
case HEADER_TRACE_INFO:
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
trace_report(fd, false);
break;
case HEADER_BUILD_ID:
if (perf_header__read_build_ids(ph, fd, section->offset, section->size))
pr_debug("Failed to read buildids, continuing...\n");
break;
default:
pr_debug("unknown feature %d, continuing...\n", feat);
}
return 0;
}
static int perf_file_header__read_pipe(struct perf_pipe_file_header *header,
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
struct perf_header *ph, int fd,
bool repipe)
{
if (readn(fd, header, sizeof(*header)) <= 0 ||
memcmp(&header->magic, __perf_magic, sizeof(header->magic)))
return -1;
if (repipe && do_write(STDOUT_FILENO, header, sizeof(*header)) < 0)
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
return -1;
if (header->size != sizeof(*header)) {
u64 size = bswap_64(header->size);
if (size != sizeof(*header))
return -1;
ph->needs_swap = true;
}
return 0;
}
static int perf_header__read_pipe(struct perf_session *session, int fd)
{
struct perf_header *header = &session->header;
struct perf_pipe_file_header f_header;
if (perf_file_header__read_pipe(&f_header, header, fd,
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
session->repipe) < 0) {
pr_debug("incompatible file format\n");
return -EINVAL;
}
session->fd = fd;
return 0;
}
int perf_session__read_header(struct perf_session *session, int fd)
{
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 17:23:10 +03:00
struct perf_file_header f_header;
struct perf_file_attr f_attr;
u64 f_id;
int nr_attrs, nr_ids, i, j;
session->evlist = perf_evlist__new(NULL, NULL);
if (session->evlist == NULL)
return -ENOMEM;
if (session->fd_pipe)
return perf_header__read_pipe(session, fd);
if (perf_file_header__read(&f_header, header, fd) < 0) {
pr_debug("incompatible file format\n");
return -EINVAL;
}
nr_attrs = f_header.attrs.size / sizeof(f_attr);
lseek(fd, f_header.attrs.offset, SEEK_SET);
for (i = 0; i < nr_attrs; i++) {
struct perf_evsel *evsel;
off_t tmp;
if (readn(fd, &f_attr, sizeof(f_attr)) <= 0)
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 17:23:10 +03:00
if (header->needs_swap)
perf_event__attr_swap(&f_attr.attr);
tmp = lseek(fd, 0, SEEK_CUR);
evsel = perf_evsel__new(&f_attr.attr, i);
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);
nr_ids = f_attr.ids.size / sizeof(u64);
/*
* 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;
lseek(fd, f_attr.ids.offset, SEEK_SET);
for (j = 0; j < nr_ids; j++) {
if (perf_header__getbuffer64(header, fd, &f_id, sizeof(f_id)))
goto out_errno;
perf_evlist__id_add(session->evlist, evsel, 0, j, f_id);
}
lseek(fd, tmp, SEEK_SET);
}
if (f_header.event_types.size) {
lseek(fd, f_header.event_types.offset, SEEK_SET);
events = malloc(f_header.event_types.size);
if (events == NULL)
return -ENOMEM;
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 17:23:10 +03:00
f_header.event_types.size))
goto out_errno;
event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type);
}
perf_header__process_sections(header, fd, perf_file_section__process);
lseek(fd, header->data_offset, SEEK_SET);
header->frozen = 1;
return 0;
out_errno:
return -errno;
out_delete_evlist:
perf_evlist__delete(session->evlist);
session->evlist = NULL;
return -ENOMEM;
}
int perf_event__synthesize_attr(struct perf_event_attr *attr, u16 ids, u64 *id,
perf_event__handler_t process,
struct perf_session *session)
{
union perf_event *ev;
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);
if (ev == NULL)
return -ENOMEM;
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;
err = process(ev, NULL, session);
free(ev);
return err;
}
int perf_session__synthesize_attrs(struct perf_session *session,
perf_event__handler_t process)
{
struct perf_evsel *attr;
int err = 0;
list_for_each_entry(attr, &session->evlist->entries, node) {
err = perf_event__synthesize_attr(&attr->attr, attr->ids,
attr->id, process, session);
if (err) {
pr_debug("failed to create perf header attribute\n");
return err;
}
}
return err;
}
int perf_event__process_attr(union perf_event *event,
struct perf_session *session)
{
unsigned int i, ids, n_ids;
struct perf_evsel *evsel;
if (session->evlist == NULL) {
session->evlist = perf_evlist__new(NULL, NULL);
if (session->evlist == NULL)
return -ENOMEM;
}
evsel = perf_evsel__new(&event->attr.attr,
session->evlist->nr_entries);
if (evsel == NULL)
return -ENOMEM;
perf_evlist__add(session->evlist, evsel);
ids = event->header.size;
ids -= (void *)&event->attr.id - (void *)event;
n_ids = ids / sizeof(u64);
/*
* 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;
for (i = 0; i < n_ids; i++) {
perf_evlist__id_add(session->evlist, evsel, 0, i,
event->attr.id[i]);
}
perf_session__update_sample_type(session);
return 0;
}
int perf_event__synthesize_event_type(u64 event_id, char *name,
perf_event__handler_t process,
struct perf_session *session)
{
union perf_event ev;
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);
err = process(&ev, NULL, session);
return err;
}
int perf_event__synthesize_event_types(perf_event__handler_t process,
struct perf_session *session)
{
struct perf_trace_event_type *type;
int i, err = 0;
for (i = 0; i < event_count; i++) {
type = &events[i];
err = perf_event__synthesize_event_type(type->event_id,
type->name, process,
session);
if (err) {
pr_debug("failed to create perf header event type\n");
return err;
}
}
return err;
}
int perf_event__process_event_type(union perf_event *event,
struct perf_session *session __unused)
{
if (perf_header__push_event(event->event_type.event_type.event_id,
event->event_type.event_type.name) < 0)
return -ENOMEM;
return 0;
}
int perf_event__synthesize_tracing_data(int fd, struct perf_evlist *evlist,
perf_event__handler_t process,
struct perf_session *session __unused)
{
union perf_event ev;
ssize_t size = 0, aligned_size = 0, padding;
int err __used = 0;
memset(&ev, 0, sizeof(ev));
ev.tracing_data.header.type = PERF_RECORD_HEADER_TRACING_DATA;
size = read_tracing_data_size(fd, &evlist->entries);
if (size <= 0)
return size;
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;
process(&ev, NULL, session);
err = read_tracing_data(fd, &evlist->entries);
write_padded(fd, NULL, 0, padding);
return aligned_size;
}
int perf_event__process_tracing_data(union perf_event *event,
struct perf_session *session)
{
ssize_t size_read, padding, size = event->tracing_data.size;
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);
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
size_read = trace_report(session->fd, session->repipe);
padding = ALIGN(size_read, sizeof(u64)) - size_read;
if (read(session->fd, buf, padding) < 0)
die("reading input file");
perf: add perf-inject builtin Currently, perf 'live mode' writes build-ids at the end of the session, which isn't actually useful for processing live mode events. What would be better would be to have the build-ids sent before any of the samples that reference them, which can be done by processing the event stream and retrieving the build-ids on the first hit. Doing that in perf-record itself, however, is off-limits. This patch introduces perf-inject, which does the same job while leaving perf-record untouched. Normal mode perf still records the build-ids at the end of the session as it should, but for live mode, perf-inject can be injected in between the record and report steps e.g.: perf record -o - ./hackbench 10 | perf inject -v -b | perf report -v -i - perf-inject reads a perf-record event stream and repipes it to stdout. At any point the processing code can inject other events into the event stream - in this case build-ids (-b option) are read and injected as needed into the event stream. Build-ids are just the first user of perf-inject - potentially anything that needs userspace processing to augment the trace stream with additional information could make use of this facility. Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frédéric Weisbecker <fweisbec@gmail.com> LKML-Reference: <1272696080-16435-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2010-05-01 10:41:20 +04:00
if (session->repipe) {
int retw = write(STDOUT_FILENO, buf, padding);
if (retw <= 0 || retw != padding)
die("repiping tracing data padding");
}
if (size_read + padding != size)
die("tracing data size mismatch");
return size_read + padding;
}
int perf_event__synthesize_build_id(struct dso *pos, u16 misc,
perf_event__handler_t process,
struct machine *machine,
struct perf_session *session)
{
union perf_event ev;
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;
ev.build_id.pid = machine->pid;
ev.build_id.header.size = sizeof(ev.build_id) + len;
memcpy(&ev.build_id.filename, pos->long_name, pos->long_name_len);
err = process(&ev, NULL, session);
return err;
}
int perf_event__process_build_id(union perf_event *event,
struct perf_session *session)
{
__event_process_build_id(&event->build_id,
event->build_id.filename,
session);
return 0;
}
void disable_buildid_cache(void)
{
no_buildid_cache = true;
}