perf evlist: Add 'system_wide' option

Add an option to cause a selected event to be opened always without a
pid when configured by perf_evsel__config().

This is needed when using the sched_switch tracepoint to follow object
code execution.

sched_switch occurs before the task switch and so it cannot record it in
a context limited to that task.  Note that also means that sched_switch
is useless when capturing data per-thread, as is the 'context-switches'
software event for the same reason.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
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/1406786474-9306-9-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

3 files changed, 64 insertions, 13 deletions

diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 3b366c085021..c74d8ecc9c70 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -265,17 +265,27 @@ int perf_evlist__add_newtp(struct perf_evlist *evlist,
         return 0;
 }
 
+static int perf_evlist__nr_threads(struct perf_evlist *evlist,
+                                   struct perf_evsel *evsel)
+{
+        if (evsel->system_wide)
+                return 1;
+        else
+                return thread_map__nr(evlist->threads);
+}
+
 void perf_evlist__disable(struct perf_evlist *evlist)
 {
         int cpu, thread;
         struct perf_evsel *pos;
         int nr_cpus = cpu_map__nr(evlist->cpus);
-        int nr_threads = thread_map__nr(evlist->threads);
+        int nr_threads;
 
         for (cpu = 0; cpu < nr_cpus; cpu++) {
                 evlist__for_each(evlist, pos) {
                         if (!perf_evsel__is_group_leader(pos) || !pos->fd)
                                 continue;
+                        nr_threads = perf_evlist__nr_threads(evlist, pos);
                         for (thread = 0; thread < nr_threads; thread++)
                                 ioctl(FD(pos, cpu, thread),
                                       PERF_EVENT_IOC_DISABLE, 0);
@@ -288,12 +298,13 @@ void perf_evlist__enable(struct perf_evlist *evlist)
         int cpu, thread;
         struct perf_evsel *pos;
         int nr_cpus = cpu_map__nr(evlist->cpus);
-        int nr_threads = thread_map__nr(evlist->threads);
+        int nr_threads;
 
         for (cpu = 0; cpu < nr_cpus; cpu++) {
                 evlist__for_each(evlist, pos) {
                         if (!perf_evsel__is_group_leader(pos) || !pos->fd)
                                 continue;
+                        nr_threads = perf_evlist__nr_threads(evlist, pos);
                         for (thread = 0; thread < nr_threads; thread++)
                                 ioctl(FD(pos, cpu, thread),
                                       PERF_EVENT_IOC_ENABLE, 0);
@@ -305,12 +316,14 @@ int perf_evlist__disable_event(struct perf_evlist *evlist,
                                struct perf_evsel *evsel)
 {
         int cpu, thread, err;
+        int nr_cpus = cpu_map__nr(evlist->cpus);
+        int nr_threads = perf_evlist__nr_threads(evlist, evsel);
 
         if (!evsel->fd)
                 return 0;
 
-        for (cpu = 0; cpu < evlist->cpus->nr; cpu++) {
-                for (thread = 0; thread < evlist->threads->nr; thread++) {
+        for (cpu = 0; cpu < nr_cpus; cpu++) {
+                for (thread = 0; thread < nr_threads; thread++) {
                         err = ioctl(FD(evsel, cpu, thread),
                                     PERF_EVENT_IOC_DISABLE, 0);
                         if (err)
@@ -324,12 +337,14 @@ int perf_evlist__enable_event(struct perf_evlist *evlist,
                               struct perf_evsel *evsel)
 {
         int cpu, thread, err;
+        int nr_cpus = cpu_map__nr(evlist->cpus);
+        int nr_threads = perf_evlist__nr_threads(evlist, evsel);
 
         if (!evsel->fd)
                 return -EINVAL;
 
-        for (cpu = 0; cpu < evlist->cpus->nr; cpu++) {
-                for (thread = 0; thread < evlist->threads->nr; thread++) {
+        for (cpu = 0; cpu < nr_cpus; cpu++) {
+                for (thread = 0; thread < nr_threads; thread++) {
                         err = ioctl(FD(evsel, cpu, thread),
                                     PERF_EVENT_IOC_ENABLE, 0);
                         if (err)
@@ -343,7 +358,16 @@ static int perf_evlist__alloc_pollfd(struct perf_evlist *evlist)
 {
         int nr_cpus = cpu_map__nr(evlist->cpus);
         int nr_threads = thread_map__nr(evlist->threads);
-        int nfds = nr_cpus * nr_threads * evlist->nr_entries;
+        int nfds = 0;
+        struct perf_evsel *evsel;
+
+        list_for_each_entry(evsel, &evlist->entries, node) {
+                if (evsel->system_wide)
+                        nfds += nr_cpus;
+                else
+                        nfds += nr_cpus * nr_threads;
+        }
+
         evlist->pollfd = malloc(sizeof(struct pollfd) * nfds);
         return evlist->pollfd != NULL ? 0 : -ENOMEM;
 }
@@ -636,7 +660,12 @@ static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx,
         struct perf_evsel *evsel;
 
         evlist__for_each(evlist, evsel) {
-                int fd = FD(evsel, cpu, thread);
+                int fd;
+
+                if (evsel->system_wide && thread)
+                        continue;
+
+                fd = FD(evsel, cpu, thread);
 
                 if (*output == -1) {
                         *output = fd;
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 0c8919decac8..66de9a708163 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -695,6 +695,10 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
 int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
 {
         int cpu, thread;
+
+        if (evsel->system_wide)
+                nthreads = 1;
+
         evsel->fd = xyarray__new(ncpus, nthreads, sizeof(int));
 
         if (evsel->fd) {
@@ -713,6 +717,9 @@ static int perf_evsel__run_ioctl(struct perf_evsel *evsel, int ncpus, int nthrea
 {
         int cpu, thread;
 
+        if (evsel->system_wide)
+                nthreads = 1;
+
         for (cpu = 0; cpu < ncpus; cpu++) {
                 for (thread = 0; thread < nthreads; thread++) {
                         int fd = FD(evsel, cpu, thread),
@@ -743,6 +750,9 @@ int perf_evsel__enable(struct perf_evsel *evsel, int ncpus, int nthreads)
 
 int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
 {
+        if (evsel->system_wide)
+                nthreads = 1;
+
         evsel->sample_id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id));
         if (evsel->sample_id == NULL)
                 return -ENOMEM;
@@ -787,6 +797,9 @@ void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
 {
         int cpu, thread;
 
+        if (evsel->system_wide)
+                nthreads = 1;
+
         for (cpu = 0; cpu < ncpus; cpu++)
                 for (thread = 0; thread < nthreads; ++thread) {
                         close(FD(evsel, cpu, thread));
@@ -875,6 +888,9 @@ int __perf_evsel__read(struct perf_evsel *evsel,
         int cpu, thread;
         struct perf_counts_values *aggr = &evsel->counts->aggr, count;
 
+        if (evsel->system_wide)
+                nthreads = 1;
+
         aggr->val = aggr->ena = aggr->run = 0;
 
         for (cpu = 0; cpu < ncpus; cpu++) {
@@ -997,13 +1013,18 @@ static size_t perf_event_attr__fprintf(struct perf_event_attr *attr, FILE *fp)
 static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
                               struct thread_map *threads)
 {
-        int cpu, thread;
+        int cpu, thread, nthreads;
         unsigned long flags = PERF_FLAG_FD_CLOEXEC;
         int pid = -1, err;
         enum { NO_CHANGE, SET_TO_MAX, INCREASED_MAX } set_rlimit = NO_CHANGE;
 
+        if (evsel->system_wide)
+                nthreads = 1;
+        else
+                nthreads = threads->nr;
+
         if (evsel->fd == NULL &&
-            perf_evsel__alloc_fd(evsel, cpus->nr, threads->nr) < 0)
+            perf_evsel__alloc_fd(evsel, cpus->nr, nthreads) < 0)
                 return -ENOMEM;
 
         if (evsel->cgrp) {
@@ -1027,10 +1048,10 @@ retry_sample_id:
 
         for (cpu = 0; cpu < cpus->nr; cpu++) {
 
-                for (thread = 0; thread < threads->nr; thread++) {
+                for (thread = 0; thread < nthreads; thread++) {
                         int group_fd;
 
-                        if (!evsel->cgrp)
+                        if (!evsel->cgrp && !evsel->system_wide)
                                 pid = threads->map[thread];
 
                         group_fd = get_group_fd(evsel, cpu, thread);
@@ -1103,7 +1124,7 @@ out_close:
                         close(FD(evsel, cpu, thread));
                         FD(evsel, cpu, thread) = -1;
                 }
-                thread = threads->nr;
+                thread = nthreads;
         } while (--cpu >= 0);
         return err;
 }
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index d7f93ce0ebc1..dbb2a0d20907 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -85,6 +85,7 @@ struct perf_evsel {
         bool                    needs_swap;
         bool                    no_aux_samples;
         bool                    immediate;
+        bool                    system_wide;
         /* parse modifier helper */
         int                     exclude_GH;
         int                     nr_members;