WSL2-Linux-Kernel/tools/perf/builtin-inject.c

641 строка
16 KiB
C
Исходник Обычный вид История

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
/*
* builtin-inject.c
*
* Builtin inject command: Examine the live mode (stdin) event stream
* and repipe it to stdout while optionally injecting additional
* events into it.
*/
#include "builtin.h"
#include "perf.h"
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
#include "util/color.h"
#include "util/evlist.h"
#include "util/evsel.h"
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
#include "util/session.h"
#include "util/tool.h"
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
#include "util/debug.h"
#include "util/build-id.h"
#include "util/data.h"
#include "util/auxtrace.h"
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
#include "util/parse-options.h"
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
#include <linux/list.h>
struct perf_inject {
struct perf_tool tool;
struct perf_session *session;
bool build_ids;
bool sched_stat;
bool have_auxtrace;
const char *input_name;
struct perf_data_file output;
u64 bytes_written;
struct list_head samples;
struct itrace_synth_opts itrace_synth_opts;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
};
struct event_entry {
struct list_head node;
u32 tid;
union perf_event event[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
static int output_bytes(struct perf_inject *inject, void *buf, size_t sz)
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
{
ssize_t size;
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 = perf_data_file__write(&inject->output, buf, sz);
if (size < 0)
return -errno;
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
inject->bytes_written += size;
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 0;
}
static int perf_event__repipe_synth(struct perf_tool *tool,
union perf_event *event)
{
struct perf_inject *inject = container_of(tool, struct perf_inject,
tool);
return output_bytes(inject, event, event->header.size);
}
static int perf_event__repipe_oe_synth(struct perf_tool *tool,
union perf_event *event,
struct ordered_events *oe __maybe_unused)
{
return perf_event__repipe_synth(tool, event);
}
static int perf_event__repipe_op2_synth(struct perf_tool *tool,
union perf_event *event,
perf tools: Use __maybe_used for unused variables perf defines both __used and __unused variables to use for marking unused variables. The variable __used is defined to __attribute__((__unused__)), which contradicts the kernel definition to __attribute__((__used__)) for new gcc versions. On Android, __used is also defined in system headers and this leads to warnings like: warning: '__used__' attribute ignored __unused is not defined in the kernel and is not a standard definition. If __unused is included everywhere instead of __used, this leads to conflicts with glibc headers, since glibc has a variables with this name in its headers. The best approach is to use __maybe_unused, the definition used in the kernel for __attribute__((unused)). In this way there is only one definition in perf sources (instead of 2 definitions that point to the same thing: __used and __unused) and it works on both Linux and Android. This patch simply replaces all instances of __used and __unused with __maybe_unused. Signed-off-by: Irina Tirdea <irina.tirdea@intel.com> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com [ committer note: fixed up conflict with a116e05 in builtin-sched.c ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-11 02:15:03 +04:00
struct perf_session *session
__maybe_unused)
{
return perf_event__repipe_synth(tool, event);
}
static int perf_event__repipe_attr(struct perf_tool *tool,
union perf_event *event,
struct perf_evlist **pevlist)
{
struct perf_inject *inject = container_of(tool, struct perf_inject,
tool);
int ret;
ret = perf_event__process_attr(tool, event, pevlist);
if (ret)
return ret;
if (!inject->output.is_pipe)
return 0;
return perf_event__repipe_synth(tool, event);
}
#ifdef HAVE_AUXTRACE_SUPPORT
static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
{
char buf[4096];
ssize_t ssz;
int ret;
while (size > 0) {
ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
if (ssz < 0)
return -errno;
ret = output_bytes(inject, buf, ssz);
if (ret)
return ret;
size -= ssz;
}
return 0;
}
static s64 perf_event__repipe_auxtrace(struct perf_tool *tool,
union perf_event *event,
struct perf_session *session
__maybe_unused)
{
struct perf_inject *inject = container_of(tool, struct perf_inject,
tool);
int ret;
inject->have_auxtrace = true;
if (!inject->output.is_pipe) {
off_t offset;
offset = lseek(inject->output.fd, 0, SEEK_CUR);
if (offset == -1)
return -errno;
ret = auxtrace_index__auxtrace_event(&session->auxtrace_index,
event, offset);
if (ret < 0)
return ret;
}
if (perf_data_file__is_pipe(session->file) || !session->one_mmap) {
ret = output_bytes(inject, event, event->header.size);
if (ret < 0)
return ret;
ret = copy_bytes(inject, perf_data_file__fd(session->file),
event->auxtrace.size);
} else {
ret = output_bytes(inject, event,
event->header.size + event->auxtrace.size);
}
if (ret < 0)
return ret;
return event->auxtrace.size;
}
#else
static s64
perf_event__repipe_auxtrace(struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_session *session __maybe_unused)
{
pr_err("AUX area tracing not supported\n");
return -EINVAL;
}
#endif
static int perf_event__repipe(struct perf_tool *tool,
union perf_event *event,
perf tools: Use __maybe_used for unused variables perf defines both __used and __unused variables to use for marking unused variables. The variable __used is defined to __attribute__((__unused__)), which contradicts the kernel definition to __attribute__((__used__)) for new gcc versions. On Android, __used is also defined in system headers and this leads to warnings like: warning: '__used__' attribute ignored __unused is not defined in the kernel and is not a standard definition. If __unused is included everywhere instead of __used, this leads to conflicts with glibc headers, since glibc has a variables with this name in its headers. The best approach is to use __maybe_unused, the definition used in the kernel for __attribute__((unused)). In this way there is only one definition in perf sources (instead of 2 definitions that point to the same thing: __used and __unused) and it works on both Linux and Android. This patch simply replaces all instances of __used and __unused with __maybe_unused. Signed-off-by: Irina Tirdea <irina.tirdea@intel.com> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com [ committer note: fixed up conflict with a116e05 in builtin-sched.c ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-11 02:15:03 +04:00
struct perf_sample *sample __maybe_unused,
struct machine *machine __maybe_unused)
{
return perf_event__repipe_synth(tool, event);
}
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
typedef int (*inject_handler)(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct perf_evsel *evsel,
struct machine *machine);
static int perf_event__repipe_sample(struct perf_tool *tool,
union perf_event *event,
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
struct perf_sample *sample,
struct perf_evsel *evsel,
struct machine *machine)
{
if (evsel->handler) {
inject_handler f = evsel->handler;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
return f(tool, event, sample, evsel, machine);
}
build_id__mark_dso_hit(tool, event, sample, evsel, machine);
return perf_event__repipe_synth(tool, event);
}
static int perf_event__repipe_mmap(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct machine *machine)
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
{
int err;
err = perf_event__process_mmap(tool, event, sample, machine);
perf_event__repipe(tool, event, sample, machine);
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 err;
}
static int perf_event__repipe_mmap2(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct machine *machine)
{
int err;
err = perf_event__process_mmap2(tool, event, sample, machine);
perf_event__repipe(tool, event, sample, machine);
return err;
}
static int perf_event__repipe_fork(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct machine *machine)
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
{
int err;
err = perf_event__process_fork(tool, event, sample, machine);
perf_event__repipe(tool, event, sample, machine);
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 err;
}
static int perf_event__repipe_comm(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct machine *machine)
{
int err;
err = perf_event__process_comm(tool, event, sample, machine);
perf_event__repipe(tool, event, sample, machine);
return err;
}
static int perf_event__repipe_exit(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct machine *machine)
{
int err;
err = perf_event__process_exit(tool, event, sample, machine);
perf_event__repipe(tool, event, sample, machine);
return err;
}
static int perf_event__repipe_tracing_data(struct perf_tool *tool,
union perf_event *event,
struct perf_session *session)
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
{
int err;
perf_event__repipe_synth(tool, event);
err = perf_event__process_tracing_data(tool, event, session);
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 err;
}
static int perf_event__repipe_id_index(struct perf_tool *tool,
union perf_event *event,
struct perf_session *session)
{
int err;
perf_event__repipe_synth(tool, event);
err = perf_event__process_id_index(tool, event, session);
return err;
}
static int dso__read_build_id(struct dso *dso)
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 (dso->has_build_id)
return 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
if (filename__read_build_id(dso->long_name, dso->build_id,
sizeof(dso->build_id)) > 0) {
dso->has_build_id = true;
return 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;
}
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
static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool,
struct machine *machine)
{
u16 misc = PERF_RECORD_MISC_USER;
int err;
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 (dso__read_build_id(dso) < 0) {
pr_debug("no build_id found for %s\n", dso->long_name);
return -1;
}
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 (dso->kernel)
misc = PERF_RECORD_MISC_KERNEL;
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
err = perf_event__synthesize_build_id(tool, dso, misc, perf_event__repipe,
machine);
if (err) {
pr_err("Can't synthesize build_id event for %s\n", dso->long_name);
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;
}
return 0;
}
static int perf_event__inject_buildid(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
perf tools: Use __maybe_used for unused variables perf defines both __used and __unused variables to use for marking unused variables. The variable __used is defined to __attribute__((__unused__)), which contradicts the kernel definition to __attribute__((__used__)) for new gcc versions. On Android, __used is also defined in system headers and this leads to warnings like: warning: '__used__' attribute ignored __unused is not defined in the kernel and is not a standard definition. If __unused is included everywhere instead of __used, this leads to conflicts with glibc headers, since glibc has a variables with this name in its headers. The best approach is to use __maybe_unused, the definition used in the kernel for __attribute__((unused)). In this way there is only one definition in perf sources (instead of 2 definitions that point to the same thing: __used and __unused) and it works on both Linux and Android. This patch simply replaces all instances of __used and __unused with __maybe_unused. Signed-off-by: Irina Tirdea <irina.tirdea@intel.com> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com [ committer note: fixed up conflict with a116e05 in builtin-sched.c ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-11 02:15:03 +04:00
struct perf_evsel *evsel __maybe_unused,
struct machine *machine)
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 addr_location al;
struct thread *thread;
u8 cpumode;
cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
thread = machine__findnew_thread(machine, sample->pid, sample->tid);
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 (thread == NULL) {
pr_err("problem processing %d event, skipping it.\n",
event->header.type);
goto repipe;
}
thread__find_addr_map(thread, cpumode, MAP__FUNCTION, sample->ip, &al);
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 (al.map != NULL) {
if (!al.map->dso->hit) {
al.map->dso->hit = 1;
if (map__load(al.map, NULL) >= 0) {
dso__inject_build_id(al.map->dso, tool, machine);
/*
* If this fails, too bad, let the other side
* account this as unresolved.
*/
} else {
#ifdef HAVE_LIBELF_SUPPORT
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
pr_warning("no symbols found in %s, maybe "
"install a debug package?\n",
al.map->dso->long_name);
#endif
}
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
}
}
repipe:
perf_event__repipe(tool, event, sample, machine);
return 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
}
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
static int perf_inject__sched_process_exit(struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
struct perf_evsel *evsel __maybe_unused,
struct machine *machine __maybe_unused)
{
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
struct event_entry *ent;
list_for_each_entry(ent, &inject->samples, node) {
if (sample->tid == ent->tid) {
list_del_init(&ent->node);
free(ent);
break;
}
}
return 0;
}
static int perf_inject__sched_switch(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample,
struct perf_evsel *evsel,
struct machine *machine)
{
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
struct event_entry *ent;
perf_inject__sched_process_exit(tool, event, sample, evsel, machine);
ent = malloc(event->header.size + sizeof(struct event_entry));
if (ent == NULL) {
color_fprintf(stderr, PERF_COLOR_RED,
"Not enough memory to process sched switch event!");
return -1;
}
ent->tid = sample->tid;
memcpy(&ent->event, event, event->header.size);
list_add(&ent->node, &inject->samples);
return 0;
}
static int perf_inject__sched_stat(struct perf_tool *tool,
union perf_event *event __maybe_unused,
struct perf_sample *sample,
struct perf_evsel *evsel,
struct machine *machine)
{
struct event_entry *ent;
union perf_event *event_sw;
struct perf_sample sample_sw;
struct perf_inject *inject = container_of(tool, struct perf_inject, tool);
u32 pid = perf_evsel__intval(evsel, sample, "pid");
list_for_each_entry(ent, &inject->samples, node) {
if (pid == ent->tid)
goto found;
}
return 0;
found:
event_sw = &ent->event[0];
perf_evsel__parse_sample(evsel, event_sw, &sample_sw);
sample_sw.period = sample->period;
sample_sw.time = sample->time;
perf_event__synthesize_sample(event_sw, evsel->attr.sample_type,
evsel->attr.read_format, &sample_sw,
false);
build_id__mark_dso_hit(tool, event_sw, &sample_sw, evsel, machine);
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
return perf_event__repipe(tool, event_sw, &sample_sw, machine);
}
perf tools: Use __maybe_used for unused variables perf defines both __used and __unused variables to use for marking unused variables. The variable __used is defined to __attribute__((__unused__)), which contradicts the kernel definition to __attribute__((__used__)) for new gcc versions. On Android, __used is also defined in system headers and this leads to warnings like: warning: '__used__' attribute ignored __unused is not defined in the kernel and is not a standard definition. If __unused is included everywhere instead of __used, this leads to conflicts with glibc headers, since glibc has a variables with this name in its headers. The best approach is to use __maybe_unused, the definition used in the kernel for __attribute__((unused)). In this way there is only one definition in perf sources (instead of 2 definitions that point to the same thing: __used and __unused) and it works on both Linux and Android. This patch simply replaces all instances of __used and __unused with __maybe_unused. Signed-off-by: Irina Tirdea <irina.tirdea@intel.com> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com [ committer note: fixed up conflict with a116e05 in builtin-sched.c ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-11 02:15:03 +04:00
static void sig_handler(int sig __maybe_unused)
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_done = 1;
}
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
static int perf_evsel__check_stype(struct perf_evsel *evsel,
u64 sample_type, const char *sample_msg)
{
struct perf_event_attr *attr = &evsel->attr;
const char *name = perf_evsel__name(evsel);
if (!(attr->sample_type & sample_type)) {
pr_err("Samples for %s event do not have %s attribute set.",
name, sample_msg);
return -EINVAL;
}
return 0;
}
static int __cmd_inject(struct perf_inject *inject)
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
{
int ret = -EINVAL;
struct perf_session *session = inject->session;
struct perf_data_file *file_out = &inject->output;
int fd = perf_data_file__fd(file_out);
u64 output_data_offset;
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
signal(SIGINT, sig_handler);
if (inject->build_ids || inject->sched_stat ||
inject->itrace_synth_opts.set) {
inject->tool.mmap = perf_event__repipe_mmap;
inject->tool.mmap2 = perf_event__repipe_mmap2;
inject->tool.fork = perf_event__repipe_fork;
inject->tool.tracing_data = perf_event__repipe_tracing_data;
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
}
output_data_offset = session->header.data_offset;
if (inject->build_ids) {
inject->tool.sample = perf_event__inject_buildid;
} else if (inject->sched_stat) {
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
struct perf_evsel *evsel;
evlist__for_each(session->evlist, evsel) {
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
const char *name = perf_evsel__name(evsel);
if (!strcmp(name, "sched:sched_switch")) {
if (perf_evsel__check_stype(evsel, PERF_SAMPLE_TID, "TID"))
return -EINVAL;
evsel->handler = perf_inject__sched_switch;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
} else if (!strcmp(name, "sched:sched_process_exit"))
evsel->handler = perf_inject__sched_process_exit;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
else if (!strncmp(name, "sched:sched_stat_", 17))
evsel->handler = perf_inject__sched_stat;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
}
} else if (inject->itrace_synth_opts.set) {
session->itrace_synth_opts = &inject->itrace_synth_opts;
inject->itrace_synth_opts.inject = true;
inject->tool.comm = perf_event__repipe_comm;
inject->tool.exit = perf_event__repipe_exit;
inject->tool.id_index = perf_event__repipe_id_index;
inject->tool.auxtrace_info = perf_event__process_auxtrace_info;
inject->tool.auxtrace = perf_event__process_auxtrace;
inject->tool.ordered_events = true;
inject->tool.ordering_requires_timestamps = true;
/* Allow space in the header for new attributes */
output_data_offset = 4096;
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
}
if (!inject->itrace_synth_opts.set)
auxtrace_index__free(&session->auxtrace_index);
if (!file_out->is_pipe)
lseek(fd, output_data_offset, SEEK_SET);
ret = perf_session__process_events(session);
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 (!file_out->is_pipe) {
if (inject->build_ids) {
perf_header__set_feat(&session->header,
HEADER_BUILD_ID);
if (inject->have_auxtrace)
dsos__hit_all(session);
}
/*
* The AUX areas have been removed and replaced with
* synthesized hardware events, so clear the feature flag.
*/
if (inject->itrace_synth_opts.set)
perf_header__clear_feat(&session->header,
HEADER_AUXTRACE);
session->header.data_offset = output_data_offset;
session->header.data_size = inject->bytes_written;
perf_session__write_header(session, session->evlist, fd, true);
}
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 ret;
}
perf tools: Use __maybe_used for unused variables perf defines both __used and __unused variables to use for marking unused variables. The variable __used is defined to __attribute__((__unused__)), which contradicts the kernel definition to __attribute__((__used__)) for new gcc versions. On Android, __used is also defined in system headers and this leads to warnings like: warning: '__used__' attribute ignored __unused is not defined in the kernel and is not a standard definition. If __unused is included everywhere instead of __used, this leads to conflicts with glibc headers, since glibc has a variables with this name in its headers. The best approach is to use __maybe_unused, the definition used in the kernel for __attribute__((unused)). In this way there is only one definition in perf sources (instead of 2 definitions that point to the same thing: __used and __unused) and it works on both Linux and Android. This patch simply replaces all instances of __used and __unused with __maybe_unused. Signed-off-by: Irina Tirdea <irina.tirdea@intel.com> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com [ committer note: fixed up conflict with a116e05 in builtin-sched.c ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-11 02:15:03 +04:00
int cmd_inject(int argc, const char **argv, const char *prefix __maybe_unused)
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_inject inject = {
.tool = {
.sample = perf_event__repipe_sample,
.mmap = perf_event__repipe,
.mmap2 = perf_event__repipe,
.comm = perf_event__repipe,
.fork = perf_event__repipe,
.exit = perf_event__repipe,
.lost = perf_event__repipe,
.aux = perf_event__repipe,
.itrace_start = perf_event__repipe,
.read = perf_event__repipe_sample,
.throttle = perf_event__repipe,
.unthrottle = perf_event__repipe,
.attr = perf_event__repipe_attr,
.tracing_data = perf_event__repipe_op2_synth,
.auxtrace_info = perf_event__repipe_op2_synth,
.auxtrace = perf_event__repipe_auxtrace,
.auxtrace_error = perf_event__repipe_op2_synth,
.finished_round = perf_event__repipe_oe_synth,
.build_id = perf_event__repipe_op2_synth,
perf tools: Add id index Add an index of the event identifiers, in preparation for Intel PT. The event id (also called the sample id) is a unique number allocated by the kernel to the event created by perf_event_open(). Events can include the event id by having a sample type including PERF_SAMPLE_ID or PERF_SAMPLE_IDENTIFIER. Currently the main use of the event id is to match an event back to the evsel to which it belongs i.e. perf_evlist__id2evsel() The purpose of this patch is to make it possible to match an event back to the mmap from which it was read. The reason that is useful is because the mmap represents a time-ordered context (either for a cpu or for a thread). Intel PT decodes trace information on that basis. In full-trace mode, that information can be recorded when the Intel PT trace is read, but in sample-mode the Intel PT trace data is embedded in a sample and it is in that case that the "id index" is needed. So the mmaps are numbered (idx) and the cpu and tid recorded against the id by perf_evlist__set_sid_idx() which is called by perf_evlist__mmap_per_evsel(). That information is recorded on the perf.data file in the new "id index". idx, cpu and tid are added to struct perf_sample_id (which is the node of evlist's hash table to match ids to evsels). The information can be retrieved using perf_evlist__id2sid(). Note however this all depends on having a sample type including PERF_SAMPLE_ID or PERF_SAMPLE_IDENTIFIER, otherwise ids are not recorded. The "id index" is a synthesized event record which will be created when Intel PT sampling is used by calling perf_event__synthesize_id_index(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Jiri Olsa <jolsa@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/1414417770-18602-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-10-27 16:49:22 +03:00
.id_index = perf_event__repipe_op2_synth,
},
.input_name = "-",
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
.samples = LIST_HEAD_INIT(inject.samples),
.output = {
.path = "-",
.mode = PERF_DATA_MODE_WRITE,
},
};
struct perf_data_file file = {
.mode = PERF_DATA_MODE_READ,
};
int ret;
const struct option options[] = {
OPT_BOOLEAN('b', "build-ids", &inject.build_ids,
"Inject build-ids into the output stream"),
OPT_STRING('i', "input", &inject.input_name, "file",
"input file name"),
OPT_STRING('o', "output", &inject.output.path, "file",
"output file name"),
perf inject: Merge sched_stat_* and sched_switch events You may want to know where and how long a task is sleeping. A callchain may be found in sched_switch and a time slice in stat_iowait, so I add handler in perf inject for merging this events. My code saves sched_switch event for each process and when it meets stat_iowait, it reports the sched_switch event, because this event contains a correct callchain. By another words it replaces all stat_iowait events on proper sched_switch events. I use the next sequence of commands for testing: perf record -e sched:sched_stat_sleep -e sched:sched_switch \ -e sched:sched_process_exit -g -o ~/perf.data.raw \ ~/test-program perf inject -v -s -i ~/perf.data.raw -o ~/perf.data perf report --stdio -i ~/perf.data 100.00% foo [kernel.kallsyms] [k] __schedule | --- __schedule schedule | |--79.75%-- schedule_hrtimeout_range_clock | schedule_hrtimeout_range | poll_schedule_timeout | do_select | core_sys_select | sys_select | system_call_fastpath | __select | __libc_start_main | --20.25%-- do_nanosleep hrtimer_nanosleep sys_nanosleep system_call_fastpath __GI___libc_nanosleep __libc_start_main And here is test-program.c: #include<unistd.h> #include<time.h> #include<sys/select.h> int main() { struct timespec ts1; struct timeval tv1; int i; long s; for (i = 0; i < 10; i++) { ts1.tv_sec = 0; ts1.tv_nsec = 10000000; nanosleep(&ts1, NULL); tv1.tv_sec = 0; tv1.tv_usec = 40000; select(0, NULL, NULL, NULL,&tv1); } return 1; } Signed-off-by: Andrew Vagin <avagin@openvz.org> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1344344165-369636-4-git-send-email-avagin@openvz.org [ committer note: Made it use evsel->handler ] Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-08-07 16:56:04 +04:00
OPT_BOOLEAN('s', "sched-stat", &inject.sched_stat,
"Merge sched-stat and sched-switch for getting events "
"where and how long tasks slept"),
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show build ids, etc)"),
OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name, "file",
"kallsyms pathname"),
perf inject: Support using -f to override perf.data file ownership Enable perf inject to use perf.data when it is not owned by current user or root. Example: # perf record ls # chown Yunlong.Song:Yunlong.Song perf.data # ls -al perf.data -rw------- 1 Yunlong.Song Yunlong.Song 28260 Apr 2 10:37 perf.data # id uid=0(root) gid=0(root) groups=0(root),64(pkcs11) Before this patch: # perf inject -v -b -i perf.data -o perf.data.new File perf.data not owned by current user or root (use -f to override) # perf inject -v -b -i perf.data -o perf.data.new -f Error: unknown switch `f' usage: perf inject [<options>] -b, --build-ids Inject build-ids into the output stream -i, --input <file> input file name -o, --output <file> output file name -s, --sched-stat Merge sched-stat and sched-switch for getting events where and how long tasks slept -v, --verbose be more verbose (show build ids, etc) --kallsyms <file> kallsyms pathname As shown above, the -f option does not work at all. After this patch: # perf inject -v -b -i perf.data -o perf.data.new File perf.data not owned by current user or root (use -f to override) # perf inject -v -b -i perf.data -o perf.data.new -f build id event received for [kernel.kallsyms]: f6dcb66d8b98f1c0d9eb87bf043444b69f91d30c symsrc__init: cannot get elf header. Looking at the vmlinux_path (7 entries long) Using /proc/kcore for kernel object code Using /proc/kallsyms for symbols As shown above, the -f option really works now. Signed-off-by: Yunlong Song <yunlong.song@huawei.com> Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Wang Nan <wangnan0@huawei.com> Link: http://lkml.kernel.org/r/1427982439-27388-3-git-send-email-yunlong.song@huawei.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2015-04-02 16:47:11 +03:00
OPT_BOOLEAN('f', "force", &file.force, "don't complain, do it"),
OPT_CALLBACK_OPTARG(0, "itrace", &inject.itrace_synth_opts,
NULL, "opts", "Instruction Tracing options",
itrace_parse_synth_opts),
OPT_END()
};
const char * const inject_usage[] = {
"perf inject [<options>]",
NULL
};
argc = parse_options(argc, argv, options, inject_usage, 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
/*
* Any (unrecognized) arguments left?
*/
if (argc)
usage_with_options(inject_usage, options);
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 (perf_data_file__open(&inject.output)) {
perror("failed to create output file");
return -1;
}
inject.tool.ordered_events = inject.sched_stat;
file.path = inject.input_name;
inject.session = perf_session__new(&file, true, &inject.tool);
if (inject.session == NULL)
return -1;
perf tools: Check recorded kernel version when finding vmlinux Currently vmlinux_path__init() only tries to find vmlinux file from current directory, /boot and some canonical directories with version number of the running kernel. This can be a problem when reporting old data recorded on a kernel version not running currently. We can use --symfs option for this but it's annoying for user to do it always. As we already have the info in the perf.data file, it can be changed to use it for the search automatically. Before: $ perf report ... # Samples: 4K of event 'cpu-clock' # Event count (approx.): 1067250000 # # Overhead Command Shared Object Symbol # ........ .......... ................. .............................. 71.87% swapper [kernel.kallsyms] [k] recover_probed_instruction After: # Overhead Command Shared Object Symbol # ........ .......... ................. .................... 71.87% swapper [kernel.kallsyms] [k] native_safe_halt This requires to change signature of symbol__init() to receive struct perf_session_env *. Reported-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Namhyung Kim <namhyung.kim@lge.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/1407825645-24586-14-git-send-email-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-08-12 10:40:45 +04:00
if (symbol__init(&inject.session->header.env) < 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;
ret = __cmd_inject(&inject);
perf_session__delete(inject.session);
return ret;
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
}