linux/tools/perf/util/record.c
Adrian Hunter 46ec69add5 perf tools: Fix probing the kernel API with cpu-wide events
Fall back to probing with the current pid if cpu-wide probing fails.
This primarily affects the setting of comm_exec flag when the user is
un-privileged and /proc/sys/kernel/perf_event_paranoid > 0.

The change to comm_exec can be observed by using -vv with perf record
and a kernel that supports comm_exec.

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/1407855871-15024-4-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2014-08-13 16:22:06 -03:00

253 lines
5.2 KiB
C

#include "evlist.h"
#include "evsel.h"
#include "cpumap.h"
#include "parse-events.h"
#include <api/fs/fs.h>
#include "util.h"
#include "cloexec.h"
typedef void (*setup_probe_fn_t)(struct perf_evsel *evsel);
static int perf_do_probe_api(setup_probe_fn_t fn, int cpu, const char *str)
{
struct perf_evlist *evlist;
struct perf_evsel *evsel;
unsigned long flags = perf_event_open_cloexec_flag();
int err = -EAGAIN, fd;
static pid_t pid = -1;
evlist = perf_evlist__new();
if (!evlist)
return -ENOMEM;
if (parse_events(evlist, str))
goto out_delete;
evsel = perf_evlist__first(evlist);
while (1) {
fd = sys_perf_event_open(&evsel->attr, pid, cpu, -1, flags);
if (fd < 0) {
if (pid == -1 && errno == EACCES) {
pid = 0;
continue;
}
goto out_delete;
}
break;
}
close(fd);
fn(evsel);
fd = sys_perf_event_open(&evsel->attr, pid, cpu, -1, flags);
if (fd < 0) {
if (errno == EINVAL)
err = -EINVAL;
goto out_delete;
}
close(fd);
err = 0;
out_delete:
perf_evlist__delete(evlist);
return err;
}
static bool perf_probe_api(setup_probe_fn_t fn)
{
const char *try[] = {"cycles:u", "instructions:u", "cpu-clock:u", NULL};
struct cpu_map *cpus;
int cpu, ret, i = 0;
cpus = cpu_map__new(NULL);
if (!cpus)
return false;
cpu = cpus->map[0];
cpu_map__delete(cpus);
do {
ret = perf_do_probe_api(fn, cpu, try[i++]);
if (!ret)
return true;
} while (ret == -EAGAIN && try[i]);
return false;
}
static void perf_probe_sample_identifier(struct perf_evsel *evsel)
{
evsel->attr.sample_type |= PERF_SAMPLE_IDENTIFIER;
}
static void perf_probe_comm_exec(struct perf_evsel *evsel)
{
evsel->attr.comm_exec = 1;
}
bool perf_can_sample_identifier(void)
{
return perf_probe_api(perf_probe_sample_identifier);
}
static bool perf_can_comm_exec(void)
{
return perf_probe_api(perf_probe_comm_exec);
}
void perf_evlist__config(struct perf_evlist *evlist, struct record_opts *opts)
{
struct perf_evsel *evsel;
bool use_sample_identifier = false;
bool use_comm_exec;
/*
* Set the evsel leader links before we configure attributes,
* since some might depend on this info.
*/
if (opts->group)
perf_evlist__set_leader(evlist);
if (evlist->cpus->map[0] < 0)
opts->no_inherit = true;
use_comm_exec = perf_can_comm_exec();
evlist__for_each(evlist, evsel) {
perf_evsel__config(evsel, opts);
if (!evsel->idx && use_comm_exec)
evsel->attr.comm_exec = 1;
}
if (evlist->nr_entries > 1) {
struct perf_evsel *first = perf_evlist__first(evlist);
evlist__for_each(evlist, evsel) {
if (evsel->attr.sample_type == first->attr.sample_type)
continue;
use_sample_identifier = perf_can_sample_identifier();
break;
}
evlist__for_each(evlist, evsel)
perf_evsel__set_sample_id(evsel, use_sample_identifier);
}
perf_evlist__set_id_pos(evlist);
}
static int get_max_rate(unsigned int *rate)
{
char path[PATH_MAX];
const char *procfs = procfs__mountpoint();
if (!procfs)
return -1;
snprintf(path, PATH_MAX,
"%s/sys/kernel/perf_event_max_sample_rate", procfs);
return filename__read_int(path, (int *) rate);
}
static int record_opts__config_freq(struct record_opts *opts)
{
bool user_freq = opts->user_freq != UINT_MAX;
unsigned int max_rate;
if (opts->user_interval != ULLONG_MAX)
opts->default_interval = opts->user_interval;
if (user_freq)
opts->freq = opts->user_freq;
/*
* User specified count overrides default frequency.
*/
if (opts->default_interval)
opts->freq = 0;
else if (opts->freq) {
opts->default_interval = opts->freq;
} else {
pr_err("frequency and count are zero, aborting\n");
return -1;
}
if (get_max_rate(&max_rate))
return 0;
/*
* User specified frequency is over current maximum.
*/
if (user_freq && (max_rate < opts->freq)) {
pr_err("Maximum frequency rate (%u) reached.\n"
"Please use -F freq option with lower value or consider\n"
"tweaking /proc/sys/kernel/perf_event_max_sample_rate.\n",
max_rate);
return -1;
}
/*
* Default frequency is over current maximum.
*/
if (max_rate < opts->freq) {
pr_warning("Lowering default frequency rate to %u.\n"
"Please consider tweaking "
"/proc/sys/kernel/perf_event_max_sample_rate.\n",
max_rate);
opts->freq = max_rate;
}
return 0;
}
int record_opts__config(struct record_opts *opts)
{
return record_opts__config_freq(opts);
}
bool perf_evlist__can_select_event(struct perf_evlist *evlist, const char *str)
{
struct perf_evlist *temp_evlist;
struct perf_evsel *evsel;
int err, fd, cpu;
bool ret = false;
pid_t pid = -1;
temp_evlist = perf_evlist__new();
if (!temp_evlist)
return false;
err = parse_events(temp_evlist, str);
if (err)
goto out_delete;
evsel = perf_evlist__last(temp_evlist);
if (!evlist || cpu_map__empty(evlist->cpus)) {
struct cpu_map *cpus = cpu_map__new(NULL);
cpu = cpus ? cpus->map[0] : 0;
cpu_map__delete(cpus);
} else {
cpu = evlist->cpus->map[0];
}
while (1) {
fd = sys_perf_event_open(&evsel->attr, pid, cpu, -1,
perf_event_open_cloexec_flag());
if (fd < 0) {
if (pid == -1 && errno == EACCES) {
pid = 0;
continue;
}
goto out_delete;
}
break;
}
close(fd);
ret = true;
out_delete:
perf_evlist__delete(temp_evlist);
return ret;
}