static int run_builtin(struct cmd_struct *p, int argc, const char **argv)
{
int status;
struct stat st;
const char *prefix;
char sbuf[STRERR_BUFSIZE];
prefix = NULL;
if (p->option & RUN_SETUP)
prefix = NULL; /* setup_perf_directory(); */
if (use_browser == -1)
use_browser = check_browser_config(p->cmd);
if (use_pager == -1 && p->option & RUN_SETUP)
use_pager = check_pager_config(p->cmd);
if (use_pager == -1 && p->option & USE_PAGER)
use_pager = 1;
commit_pager_choice();
perf_env__set_cmdline(&perf_env, argc, argv);
status = p->fn(argc, argv, prefix);
perf_config__exit();
exit_browser(status);
perf_env__exit(&perf_env);
bpf__clear();
if (status)
return status & 0xff;
/* Somebody closed stdout? */
if (fstat(fileno(stdout), &st))
return 0;
/* Ignore write errors for pipes and sockets.. */
if (S_ISFIFO(st.st_mode) || S_ISSOCK(st.st_mode))
return 0;
status = 1;
/* Check for ENOSPC and EIO errors.. */
if (fflush(stdout)) {
fprintf(stderr, "write failure on standard output: %s",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out;
}
if (ferror(stdout)) {
fprintf(stderr, "unknown write failure on standard output");
goto out;
}
if (fclose(stdout)) {
fprintf(stderr, "close failed on standard output: %s",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out;
}
status = 0;
out:
return status;
}
static int wait_or_whine(pid_t pid)
{
char sbuf[STRERR_BUFSIZE];
for (;;) {
int status, code;
pid_t waiting = waitpid(pid, &status, 0);
if (waiting < 0) {
if (errno == EINTR)
continue;
fprintf(stderr, " Error: waitpid failed (%s)",
str_error_r(errno, sbuf, sizeof(sbuf)));
return -ERR_RUN_COMMAND_WAITPID;
}
if (waiting != pid)
return -ERR_RUN_COMMAND_WAITPID_WRONG_PID;
if (WIFSIGNALED(status))
return -ERR_RUN_COMMAND_WAITPID_SIGNAL;
if (!WIFEXITED(status))
return -ERR_RUN_COMMAND_WAITPID_NOEXIT;
code = WEXITSTATUS(status);
switch (code) {
case 127:
return -ERR_RUN_COMMAND_EXEC;
case 0:
return 0;
default:
return -code;
}
}
}
int tracing_path__strerror_open_tp(int err, char *buf, size_t size,
const char *sys, const char *name)
{
char sbuf[128];
char filename[PATH_MAX];
snprintf(filename, PATH_MAX, "%s/%s", sys, name ?: "*");
switch (err) {
case ENOENT:
/*
* We will get here if we can't find the tracepoint, but one of
* debugfs or tracefs is configured, which means you probably
* want some tracepoint which wasn't compiled in your kernel.
* - jirka
*/
if (debugfs__configured() || tracefs__configured()) {
/* sdt markers */
if (!strncmp(filename, "sdt_", 4)) {
snprintf(buf, size,
"Error:\tFile %s/%s not found.\n"
"Hint:\tSDT event cannot be directly recorded on.\n"
"\tPlease first use 'perf probe %s:%s' before recording it.\n",
tracing_events_path, filename, sys, name);
} else {
snprintf(buf, size,
"Error:\tFile %s/%s not found.\n"
"Hint:\tPerhaps this kernel misses some CONFIG_ setting to enable this feature?.\n",
tracing_events_path, filename);
}
break;
}
snprintf(buf, size, "%s",
"Error:\tUnable to find debugfs/tracefs\n"
"Hint:\tWas your kernel compiled with debugfs/tracefs support?\n"
"Hint:\tIs the debugfs/tracefs filesystem mounted?\n"
"Hint:\tTry 'sudo mount -t debugfs nodev /sys/kernel/debug'");
break;
case EACCES: {
snprintf(buf, size,
"Error:\tNo permissions to read %s/%s\n"
"Hint:\tTry 'sudo mount -o remount,mode=755 %s'\n",
tracing_events_path, filename, tracing_path_mount());
}
break;
default:
snprintf(buf, size, "%s", str_error_r(err, sbuf, sizeof(sbuf)));
break;
}
return 0;
}
static int run_test(struct test *test, int subtest)
{
int status, err = -1, child = dont_fork ? 0 : fork();
char sbuf[STRERR_BUFSIZE];
if (child < 0) {
pr_err("failed to fork test: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
return -1;
}
if (!child) {
if (!dont_fork) {
pr_debug("test child forked, pid %d\n", getpid());
if (!verbose) {
int nullfd = open("/dev/null", O_WRONLY);
if (nullfd >= 0) {
close(STDERR_FILENO);
close(STDOUT_FILENO);
dup2(nullfd, STDOUT_FILENO);
dup2(STDOUT_FILENO, STDERR_FILENO);
close(nullfd);
}
} else {
signal(SIGSEGV, sighandler_dump_stack);
signal(SIGFPE, sighandler_dump_stack);
}
}
err = test->func(subtest);
if (!dont_fork)
exit(err);
}
if (!dont_fork) {
wait(&status);
if (WIFEXITED(status)) {
err = (signed char)WEXITSTATUS(status);
pr_debug("test child finished with %d\n", err);
} else if (WIFSIGNALED(status)) {
err = -1;
pr_debug("test child interrupted\n");
}
}
return err;
}
static int open_file_write(struct perf_data *data)
{
int fd;
char sbuf[STRERR_BUFSIZE];
if (check_backup(data))
return -1;
fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
S_IRUSR|S_IWUSR);
if (fd < 0)
pr_err("failed to open %s : %s\n", data->file.path,
str_error_r(errno, sbuf, sizeof(sbuf)));
return fd;
}
int llvm__get_nr_cpus(void)
{
static int nr_cpus_avail = 0;
char serr[STRERR_BUFSIZE];
if (nr_cpus_avail > 0)
return nr_cpus_avail;
nr_cpus_avail = sysconf(_SC_NPROCESSORS_CONF);
if (nr_cpus_avail <= 0) {
pr_err(
"WARNING:\tunable to get available CPUs in this system: %s\n"
" \tUse 128 instead.\n", str_error_r(errno, serr, sizeof(serr)));
nr_cpus_avail = 128;
}
return nr_cpus_avail;
}
static void exec_woman_emacs(const char *path, const char *page)
{
char sbuf[STRERR_BUFSIZE];
if (!check_emacsclient_version()) {
/* This works only with emacsclient version >= 22. */
char *man_page;
if (!path)
path = "emacsclient";
if (asprintf(&man_page, "(woman \"%s\")", page) > 0) {
execlp(path, "emacsclient", "-e", man_page, NULL);
free(man_page);
}
warning("failed to exec '%s': %s", path,
str_error_r(errno, sbuf, sizeof(sbuf)));
}
}
static int do_test(struct perf_evlist *evlist, int mmap_pages,
int *sample_count, int *comm_count)
{
int err;
char sbuf[STRERR_BUFSIZE];
err = perf_evlist__mmap(evlist, mmap_pages, true);
if (err < 0) {
pr_debug("perf_evlist__mmap: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
return TEST_FAIL;
}
perf_evlist__enable(evlist);
testcase();
perf_evlist__disable(evlist);
err = count_samples(evlist, sample_count, comm_count);
perf_evlist__munmap(evlist);
return err;
}
static int open_file_read(struct perf_data *data)
{
struct stat st;
int fd;
char sbuf[STRERR_BUFSIZE];
fd = open(data->file.path, O_RDONLY);
if (fd < 0) {
int err = errno;
pr_err("failed to open %s: %s", data->file.path,
str_error_r(err, sbuf, sizeof(sbuf)));
if (err == ENOENT && !strcmp(data->file.path, "perf.data"))
pr_err(" (try 'perf record' first)");
pr_err("\n");
return -err;
}
if (fstat(fd, &st) < 0)
goto out_close;
if (!data->force && st.st_uid && (st.st_uid != geteuid())) {
pr_err("File %s not owned by current user or root (use -f to override)\n",
data->file.path);
goto out_close;
}
if (!st.st_size) {
pr_info("zero-sized data (%s), nothing to do!\n",
data->file.path);
goto out_close;
}
data->size = st.st_size;
return fd;
out_close:
close(fd);
return -1;
}
int cmd_buildid_cache(int argc, const char **argv)
{
struct strlist *list;
struct str_node *pos;
int ret = 0;
int ns_id = -1;
bool force = false;
char const *add_name_list_str = NULL,
*remove_name_list_str = NULL,
*purge_name_list_str = NULL,
*missing_filename = NULL,
*update_name_list_str = NULL,
*kcore_filename = NULL;
char sbuf[STRERR_BUFSIZE];
struct perf_data data = {
.mode = PERF_DATA_MODE_READ,
};
struct perf_session *session = NULL;
struct nsinfo *nsi = NULL;
const struct option buildid_cache_options[] = {
OPT_STRING('a', "add", &add_name_list_str,
"file list", "file(s) to add"),
OPT_STRING('k', "kcore", &kcore_filename,
"file", "kcore file to add"),
OPT_STRING('r', "remove", &remove_name_list_str, "file list",
"file(s) to remove"),
OPT_STRING('p', "purge", &purge_name_list_str, "path list",
"path(s) to remove (remove old caches too)"),
OPT_STRING('M', "missing", &missing_filename, "file",
"to find missing build ids in the cache"),
OPT_BOOLEAN('f', "force", &force, "don't complain, do it"),
OPT_STRING('u', "update", &update_name_list_str, "file list",
"file(s) to update"),
OPT_INCR('v', "verbose", &verbose, "be more verbose"),
OPT_INTEGER(0, "target-ns", &ns_id, "target pid for namespace context"),
OPT_END()
};
const char * const buildid_cache_usage[] = {
"perf buildid-cache [<options>]",
NULL
};
argc = parse_options(argc, argv, buildid_cache_options,
buildid_cache_usage, 0);
if (argc || (!add_name_list_str && !kcore_filename &&
!remove_name_list_str && !purge_name_list_str &&
!missing_filename && !update_name_list_str))
usage_with_options(buildid_cache_usage, buildid_cache_options);
if (ns_id > 0)
nsi = nsinfo__new(ns_id);
if (missing_filename) {
data.file.path = missing_filename;
data.force = force;
session = perf_session__new(&data, false, NULL);
if (session == NULL)
return -1;
}
if (symbol__init(session ? &session->header.env : NULL) < 0)
goto out;
setup_pager();
if (add_name_list_str) {
list = strlist__new(add_name_list_str, NULL);
if (list) {
strlist__for_each_entry(pos, list)
if (build_id_cache__add_file(pos->s, nsi)) {
if (errno == EEXIST) {
pr_debug("%s already in the cache\n",
pos->s);
continue;
}
pr_warning("Couldn't add %s: %s\n",
pos->s, str_error_r(errno, sbuf, sizeof(sbuf)));
}
strlist__delete(list);
}
}
if (remove_name_list_str) {
list = strlist__new(remove_name_list_str, NULL);
if (list) {
strlist__for_each_entry(pos, list)
if (build_id_cache__remove_file(pos->s, nsi)) {
if (errno == ENOENT) {
pr_debug("%s wasn't in the cache\n",
pos->s);
continue;
}
pr_warning("Couldn't remove %s: %s\n",
pos->s, str_error_r(errno, sbuf, sizeof(sbuf)));
}
strlist__delete(list);
}
}
if (purge_name_list_str) {
list = strlist__new(purge_name_list_str, NULL);
if (list) {
strlist__for_each_entry(pos, list)
if (build_id_cache__purge_path(pos->s, nsi)) {
if (errno == ENOENT) {
pr_debug("%s wasn't in the cache\n",
pos->s);
continue;
}
pr_warning("Couldn't remove %s: %s\n",
pos->s, str_error_r(errno, sbuf, sizeof(sbuf)));
}
strlist__delete(list);
}
}
if (missing_filename)
ret = build_id_cache__fprintf_missing(session, stdout);
if (update_name_list_str) {
list = strlist__new(update_name_list_str, NULL);
if (list) {
strlist__for_each_entry(pos, list)
if (build_id_cache__update_file(pos->s, nsi)) {
if (errno == ENOENT) {
pr_debug("%s wasn't in the cache\n",
pos->s);
continue;
}
pr_warning("Couldn't update %s: %s\n",
pos->s, str_error_r(errno, sbuf, sizeof(sbuf)));
}
strlist__delete(list);
}
}
if (kcore_filename && build_id_cache__add_kcore(kcore_filename, force))
pr_warning("Couldn't add %s\n", kcore_filename);
out:
perf_session__delete(session);
nsinfo__zput(nsi);
return ret;
}
static int do_test(struct bpf_object *obj, int (*func)(void),
int expect)
{
struct record_opts opts = {
.target = {
.uid = UINT_MAX,
.uses_mmap = true,
},
.freq = 0,
.mmap_pages = 256,
.default_interval = 1,
};
char pid[16];
char sbuf[STRERR_BUFSIZE];
struct perf_evlist *evlist;
int i, ret = TEST_FAIL, err = 0, count = 0;
struct parse_events_evlist parse_evlist;
struct parse_events_error parse_error;
bzero(&parse_error, sizeof(parse_error));
bzero(&parse_evlist, sizeof(parse_evlist));
parse_evlist.error = &parse_error;
INIT_LIST_HEAD(&parse_evlist.list);
err = parse_events_load_bpf_obj(&parse_evlist, &parse_evlist.list, obj, NULL);
if (err || list_empty(&parse_evlist.list)) {
pr_debug("Failed to add events selected by BPF\n");
return TEST_FAIL;
}
snprintf(pid, sizeof(pid), "%d", getpid());
pid[sizeof(pid) - 1] = '\0';
opts.target.tid = opts.target.pid = pid;
/* Instead of perf_evlist__new_default, don't add default events */
evlist = perf_evlist__new();
if (!evlist) {
pr_debug("Not enough memory to create evlist\n");
return TEST_FAIL;
}
err = perf_evlist__create_maps(evlist, &opts.target);
if (err < 0) {
pr_debug("Not enough memory to create thread/cpu maps\n");
goto out_delete_evlist;
}
perf_evlist__splice_list_tail(evlist, &parse_evlist.list);
evlist->nr_groups = parse_evlist.nr_groups;
perf_evlist__config(evlist, &opts, NULL);
err = perf_evlist__open(evlist);
if (err < 0) {
pr_debug("perf_evlist__open: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
err = perf_evlist__mmap(evlist, opts.mmap_pages, false);
if (err < 0) {
pr_debug("perf_evlist__mmap: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
perf_evlist__enable(evlist);
(*func)();
perf_evlist__disable(evlist);
for (i = 0; i < evlist->nr_mmaps; i++) {
union perf_event *event;
while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
const u32 type = event->header.type;
if (type == PERF_RECORD_SAMPLE)
count ++;
}
}
if (count != expect) {
pr_debug("BPF filter result incorrect\n");
goto out_delete_evlist;
}
ret = TEST_OK;
out_delete_evlist:
perf_evlist__delete(evlist);
return ret;
}
int test__PERF_RECORD(int subtest __maybe_unused)
{
struct record_opts opts = {
.target = {
.uid = UINT_MAX,
.uses_mmap = true,
},
.no_buffering = true,
.mmap_pages = 256,
};
cpu_set_t cpu_mask;
size_t cpu_mask_size = sizeof(cpu_mask);
struct perf_evlist *evlist = perf_evlist__new_dummy();
struct perf_evsel *evsel;
struct perf_sample sample;
const char *cmd = "sleep";
const char *argv[] = { cmd, "1", NULL, };
char *bname, *mmap_filename;
u64 prev_time = 0;
bool found_cmd_mmap = false,
found_libc_mmap = false,
found_vdso_mmap = false,
found_ld_mmap = false;
int err = -1, errs = 0, i, wakeups = 0;
u32 cpu;
int total_events = 0, nr_events[PERF_RECORD_MAX] = { 0, };
char sbuf[STRERR_BUFSIZE];
if (evlist == NULL) /* Fallback for kernels lacking PERF_COUNT_SW_DUMMY */
evlist = perf_evlist__new_default();
if (evlist == NULL || argv == NULL) {
pr_debug("Not enough memory to create evlist\n");
goto out;
}
/*
* Create maps of threads and cpus to monitor. In this case
* we start with all threads and cpus (-1, -1) but then in
* perf_evlist__prepare_workload we'll fill in the only thread
* we're monitoring, the one forked there.
*/
err = perf_evlist__create_maps(evlist, &opts.target);
if (err < 0) {
pr_debug("Not enough memory to create thread/cpu maps\n");
goto out_delete_evlist;
}
/*
* Prepare the workload in argv[] to run, it'll fork it, and then wait
* for perf_evlist__start_workload() to exec it. This is done this way
* so that we have time to open the evlist (calling sys_perf_event_open
* on all the fds) and then mmap them.
*/
err = perf_evlist__prepare_workload(evlist, &opts.target, argv, false, NULL);
if (err < 0) {
pr_debug("Couldn't run the workload!\n");
goto out_delete_evlist;
}
/*
* Config the evsels, setting attr->comm on the first one, etc.
*/
evsel = perf_evlist__first(evlist);
perf_evsel__set_sample_bit(evsel, CPU);
perf_evsel__set_sample_bit(evsel, TID);
perf_evsel__set_sample_bit(evsel, TIME);
perf_evlist__config(evlist, &opts, NULL);
err = sched__get_first_possible_cpu(evlist->workload.pid, &cpu_mask);
if (err < 0) {
pr_debug("sched__get_first_possible_cpu: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
cpu = err;
/*
* So that we can check perf_sample.cpu on all the samples.
*/
if (sched_setaffinity(evlist->workload.pid, cpu_mask_size, &cpu_mask) < 0) {
pr_debug("sched_setaffinity: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
/*
* Call sys_perf_event_open on all the fds on all the evsels,
* grouping them if asked to.
*/
err = perf_evlist__open(evlist);
if (err < 0) {
pr_debug("perf_evlist__open: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
/*
* mmap the first fd on a given CPU and ask for events for the other
* fds in the same CPU to be injected in the same mmap ring buffer
* (using ioctl(PERF_EVENT_IOC_SET_OUTPUT)).
*/
err = perf_evlist__mmap(evlist, opts.mmap_pages, false);
if (err < 0) {
pr_debug("perf_evlist__mmap: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
/*
* Now that all is properly set up, enable the events, they will
* count just on workload.pid, which will start...
*/
perf_evlist__enable(evlist);
/*
* Now!
*/
perf_evlist__start_workload(evlist);
while (1) {
int before = total_events;
for (i = 0; i < evlist->nr_mmaps; i++) {
union perf_event *event;
while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
const u32 type = event->header.type;
const char *name = perf_event__name(type);
++total_events;
if (type < PERF_RECORD_MAX)
nr_events[type]++;
err = perf_evlist__parse_sample(evlist, event, &sample);
if (err < 0) {
if (verbose)
perf_event__fprintf(event, stderr);
pr_debug("Couldn't parse sample\n");
goto out_delete_evlist;
}
if (verbose) {
pr_info("%" PRIu64" %d ", sample.time, sample.cpu);
perf_event__fprintf(event, stderr);
}
if (prev_time > sample.time) {
pr_debug("%s going backwards in time, prev=%" PRIu64 ", curr=%" PRIu64 "\n",
name, prev_time, sample.time);
++errs;
}
prev_time = sample.time;
if (sample.cpu != cpu) {
pr_debug("%s with unexpected cpu, expected %d, got %d\n",
name, cpu, sample.cpu);
++errs;
}
if ((pid_t)sample.pid != evlist->workload.pid) {
pr_debug("%s with unexpected pid, expected %d, got %d\n",
name, evlist->workload.pid, sample.pid);
++errs;
}
if ((pid_t)sample.tid != evlist->workload.pid) {
pr_debug("%s with unexpected tid, expected %d, got %d\n",
name, evlist->workload.pid, sample.tid);
++errs;
}
if ((type == PERF_RECORD_COMM ||
type == PERF_RECORD_MMAP ||
type == PERF_RECORD_MMAP2 ||
type == PERF_RECORD_FORK ||
type == PERF_RECORD_EXIT) &&
(pid_t)event->comm.pid != evlist->workload.pid) {
pr_debug("%s with unexpected pid/tid\n", name);
++errs;
}
if ((type == PERF_RECORD_COMM ||
type == PERF_RECORD_MMAP ||
type == PERF_RECORD_MMAP2) &&
event->comm.pid != event->comm.tid) {
pr_debug("%s with different pid/tid!\n", name);
++errs;
}
switch (type) {
case PERF_RECORD_COMM:
if (strcmp(event->comm.comm, cmd)) {
pr_debug("%s with unexpected comm!\n", name);
++errs;
}
break;
case PERF_RECORD_EXIT:
goto found_exit;
case PERF_RECORD_MMAP:
mmap_filename = event->mmap.filename;
goto check_bname;
case PERF_RECORD_MMAP2:
mmap_filename = event->mmap2.filename;
check_bname:
bname = strrchr(mmap_filename, '/');
if (bname != NULL) {
if (!found_cmd_mmap)
found_cmd_mmap = !strcmp(bname + 1, cmd);
if (!found_libc_mmap)
found_libc_mmap = !strncmp(bname + 1, "libc", 4);
if (!found_ld_mmap)
found_ld_mmap = !strncmp(bname + 1, "ld", 2);
} else if (!found_vdso_mmap)
found_vdso_mmap = !strcmp(mmap_filename, "[vdso]");
break;
case PERF_RECORD_SAMPLE:
/* Just ignore samples for now */
break;
default:
pr_debug("Unexpected perf_event->header.type %d!\n",
type);
++errs;
}
perf_evlist__mmap_consume(evlist, i);
}
}
/*
* We don't use poll here because at least at 3.1 times the
* PERF_RECORD_{!SAMPLE} events don't honour
* perf_event_attr.wakeup_events, just PERF_EVENT_SAMPLE does.
*/
if (total_events == before && false)
perf_evlist__poll(evlist, -1);
sleep(1);
if (++wakeups > 5) {
pr_debug("No PERF_RECORD_EXIT event!\n");
break;
}
}
found_exit:
if (nr_events[PERF_RECORD_COMM] > 1) {
pr_debug("Excessive number of PERF_RECORD_COMM events!\n");
++errs;
}
if (nr_events[PERF_RECORD_COMM] == 0) {
pr_debug("Missing PERF_RECORD_COMM for %s!\n", cmd);
++errs;
}
if (!found_cmd_mmap) {
pr_debug("PERF_RECORD_MMAP for %s missing!\n", cmd);
++errs;
}
if (!found_libc_mmap) {
pr_debug("PERF_RECORD_MMAP for %s missing!\n", "libc");
++errs;
}
if (!found_ld_mmap) {
pr_debug("PERF_RECORD_MMAP for %s missing!\n", "ld");
++errs;
}
if (!found_vdso_mmap) {
pr_debug("PERF_RECORD_MMAP for %s missing!\n", "[vdso]");
++errs;
}
out_delete_evlist:
perf_evlist__delete(evlist);
out:
return (err < 0 || errs > 0) ? -1 : 0;
}
int main(int argc, const char **argv)
{
int err;
const char *cmd;
char sbuf[STRERR_BUFSIZE];
int value;
/* libsubcmd init */
exec_cmd_init("perf", PREFIX, PERF_EXEC_PATH, EXEC_PATH_ENVIRONMENT);
pager_init(PERF_PAGER_ENVIRONMENT);
/* The page_size is placed in util object. */
page_size = sysconf(_SC_PAGE_SIZE);
cache_line_size(&cacheline_size);
if (sysctl__read_int("kernel/perf_event_max_stack", &value) == 0)
sysctl_perf_event_max_stack = value;
if (sysctl__read_int("kernel/perf_event_max_contexts_per_stack", &value) == 0)
sysctl_perf_event_max_contexts_per_stack = value;
cmd = extract_argv0_path(argv[0]);
if (!cmd)
cmd = "perf-help";
srandom(time(NULL));
perf_config__init();
err = perf_config(perf_default_config, NULL);
if (err)
return err;
set_buildid_dir(NULL);
/* get debugfs/tracefs mount point from /proc/mounts */
tracing_path_mount();
/*
* "perf-xxxx" is the same as "perf xxxx", but we obviously:
*
* - cannot take flags in between the "perf" and the "xxxx".
* - cannot execute it externally (since it would just do
* the same thing over again)
*
* So we just directly call the internal command handler. If that one
* fails to handle this, then maybe we just run a renamed perf binary
* that contains a dash in its name. To handle this scenario, we just
* fall through and ignore the "xxxx" part of the command string.
*/
if (strstarts(cmd, "perf-")) {
cmd += 5;
argv[0] = cmd;
handle_internal_command(argc, argv);
/*
* If the command is handled, the above function does not
* return undo changes and fall through in such a case.
*/
cmd -= 5;
argv[0] = cmd;
}
if (strstarts(cmd, "trace")) {
#if defined(HAVE_LIBAUDIT_SUPPORT) || defined(HAVE_SYSCALL_TABLE_SUPPORT)
setup_path();
argv[0] = "trace";
return cmd_trace(argc, argv);
#else
fprintf(stderr,
"trace command not available: missing audit-libs devel package at build time.\n");
goto out;
#endif
}
/* Look for flags.. */
argv++;
argc--;
handle_options(&argv, &argc, NULL);
commit_pager_choice();
if (argc > 0) {
if (strstarts(argv[0], "--"))
argv[0] += 2;
} else {
/* The user didn't specify a command; give them help */
printf("\n usage: %s\n\n", perf_usage_string);
list_common_cmds_help();
printf("\n %s\n\n", perf_more_info_string);
goto out;
}
cmd = argv[0];
test_attr__init();
/*
* We use PATH to find perf commands, but we prepend some higher
* precedence paths: the "--exec-path" option, the PERF_EXEC_PATH
* environment, and the $(perfexecdir) from the Makefile at build
* time.
*/
setup_path();
/*
* Block SIGWINCH notifications so that the thread that wants it can
* unblock and get syscalls like select interrupted instead of waiting
* forever while the signal goes to some other non interested thread.
*/
pthread__block_sigwinch();
perf_debug_setup();
while (1) {
static int done_help;
run_argv(&argc, &argv);
if (errno != ENOENT)
break;
if (!done_help) {
cmd = argv[0] = help_unknown_cmd(cmd);
done_help = 1;
} else
break;
}
fprintf(stderr, "Failed to run command '%s': %s\n",
cmd, str_error_r(errno, sbuf, sizeof(sbuf)));
out:
return 1;
}
/*
* This test will open software clock events (cpu-clock, task-clock)
* then check their frequency -> period conversion has no artifact of
* setting period to 1 forcefully.
*/
static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
{
int i, err = -1;
volatile int tmp = 0;
u64 total_periods = 0;
int nr_samples = 0;
char sbuf[STRERR_BUFSIZE];
union perf_event *event;
struct perf_evsel *evsel;
struct perf_evlist *evlist;
struct perf_event_attr attr = {
.type = PERF_TYPE_SOFTWARE,
.config = clock_id,
.sample_type = PERF_SAMPLE_PERIOD,
.exclude_kernel = 1,
.disabled = 1,
.freq = 1,
};
struct cpu_map *cpus;
struct thread_map *threads;
attr.sample_freq = 500;
evlist = perf_evlist__new();
if (evlist == NULL) {
pr_debug("perf_evlist__new\n");
return -1;
}
evsel = perf_evsel__new(&attr);
if (evsel == NULL) {
pr_debug("perf_evsel__new\n");
goto out_delete_evlist;
}
perf_evlist__add(evlist, evsel);
cpus = cpu_map__dummy_new();
threads = thread_map__new_by_tid(getpid());
if (!cpus || !threads) {
err = -ENOMEM;
pr_debug("Not enough memory to create thread/cpu maps\n");
goto out_free_maps;
}
perf_evlist__set_maps(evlist, cpus, threads);
cpus = NULL;
threads = NULL;
if (perf_evlist__open(evlist)) {
const char *knob = "/proc/sys/kernel/perf_event_max_sample_rate";
err = -errno;
pr_debug("Couldn't open evlist: %s\nHint: check %s, using %" PRIu64 " in this test.\n",
str_error_r(errno, sbuf, sizeof(sbuf)),
knob, (u64)attr.sample_freq);
goto out_delete_evlist;
}
err = perf_evlist__mmap(evlist, 128);
if (err < 0) {
pr_debug("failed to mmap event: %d (%s)\n", errno,
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
perf_evlist__enable(evlist);
/* collect samples */
for (i = 0; i < NR_LOOPS; i++)
tmp++;
perf_evlist__disable(evlist);
while ((event = perf_evlist__mmap_read(evlist, 0)) != NULL) {
struct perf_sample sample;
if (event->header.type != PERF_RECORD_SAMPLE)
goto next_event;
err = perf_evlist__parse_sample(evlist, event, &sample);
if (err < 0) {
pr_debug("Error during parse sample\n");
goto out_delete_evlist;
}
total_periods += sample.period;
nr_samples++;
next_event:
perf_evlist__mmap_consume(evlist, 0);
}
if ((u64) nr_samples == total_periods) {
pr_debug("All (%d) samples have period value of 1!\n",
nr_samples);
err = -1;
}
out_free_maps:
cpu_map__put(cpus);
thread_map__put(threads);
out_delete_evlist:
perf_evlist__delete(evlist);
return err;
}
int test__sw_clock_freq(struct test *test __maybe_unused, int subtest __maybe_unused)
{
int ret;
ret = __test__sw_clock_freq(PERF_COUNT_SW_CPU_CLOCK);
if (!ret)
ret = __test__sw_clock_freq(PERF_COUNT_SW_TASK_CLOCK);
return ret;
}
int start_command(struct child_process *cmd)
{
int need_in, need_out, need_err;
int fdin[2], fdout[2], fderr[2];
char sbuf[STRERR_BUFSIZE];
/*
* In case of errors we must keep the promise to close FDs
* that have been passed in via ->in and ->out.
*/
need_in = !cmd->no_stdin && cmd->in < 0;
if (need_in) {
if (pipe(fdin) < 0) {
if (cmd->out > 0)
close(cmd->out);
return -ERR_RUN_COMMAND_PIPE;
}
cmd->in = fdin[1];
}
need_out = !cmd->no_stdout
&& !cmd->stdout_to_stderr
&& cmd->out < 0;
if (need_out) {
if (pipe(fdout) < 0) {
if (need_in)
close_pair(fdin);
else if (cmd->in)
close(cmd->in);
return -ERR_RUN_COMMAND_PIPE;
}
cmd->out = fdout[0];
}
need_err = !cmd->no_stderr && cmd->err < 0;
if (need_err) {
if (pipe(fderr) < 0) {
if (need_in)
close_pair(fdin);
else if (cmd->in)
close(cmd->in);
if (need_out)
close_pair(fdout);
else if (cmd->out)
close(cmd->out);
return -ERR_RUN_COMMAND_PIPE;
}
cmd->err = fderr[0];
}
fflush(NULL);
cmd->pid = fork();
if (!cmd->pid) {
if (cmd->no_stdin)
dup_devnull(0);
else if (need_in) {
dup2(fdin[0], 0);
close_pair(fdin);
} else if (cmd->in) {
dup2(cmd->in, 0);
close(cmd->in);
}
if (cmd->no_stderr)
dup_devnull(2);
else if (need_err) {
dup2(fderr[1], 2);
close_pair(fderr);
}
if (cmd->no_stdout)
dup_devnull(1);
else if (cmd->stdout_to_stderr)
dup2(2, 1);
else if (need_out) {
dup2(fdout[1], 1);
close_pair(fdout);
} else if (cmd->out > 1) {
dup2(cmd->out, 1);
close(cmd->out);
}
if (cmd->dir && chdir(cmd->dir))
die("exec %s: cd to %s failed (%s)", cmd->argv[0],
cmd->dir, str_error_r(errno, sbuf, sizeof(sbuf)));
if (cmd->env) {
for (; *cmd->env; cmd->env++) {
if (strchr(*cmd->env, '='))
putenv((char*)*cmd->env);
else
unsetenv(*cmd->env);
}
}
if (cmd->preexec_cb)
cmd->preexec_cb();
if (cmd->exec_cmd) {
execv_cmd(cmd->argv);
} else {
execvp(cmd->argv[0], (char *const*) cmd->argv);
}
exit(127);
}
if (cmd->pid < 0) {
int err = errno;
if (need_in)
close_pair(fdin);
else if (cmd->in)
close(cmd->in);
if (need_out)
close_pair(fdout);
else if (cmd->out)
close(cmd->out);
if (need_err)
close_pair(fderr);
return err == ENOENT ?
-ERR_RUN_COMMAND_EXEC :
-ERR_RUN_COMMAND_FORK;
}
if (need_in)
close(fdin[0]);
else if (cmd->in)
close(cmd->in);
if (need_out)
close(fdout[1]);
else if (cmd->out)
close(cmd->out);
if (need_err)
close(fderr[1]);
return 0;
}
int test__backward_ring_buffer(int subtest __maybe_unused)
{
int ret = TEST_SKIP, err, sample_count = 0, comm_count = 0;
char pid[16], sbuf[STRERR_BUFSIZE];
struct perf_evlist *evlist;
struct perf_evsel *evsel __maybe_unused;
struct parse_events_error parse_error;
struct record_opts opts = {
.target = {
.uid = UINT_MAX,
.uses_mmap = true,
},
.freq = 0,
.mmap_pages = 256,
.default_interval = 1,
};
snprintf(pid, sizeof(pid), "%d", getpid());
pid[sizeof(pid) - 1] = '\0';
opts.target.tid = opts.target.pid = pid;
evlist = perf_evlist__new();
if (!evlist) {
pr_debug("No ehough memory to create evlist\n");
return TEST_FAIL;
}
err = perf_evlist__create_maps(evlist, &opts.target);
if (err < 0) {
pr_debug("Not enough memory to create thread/cpu maps\n");
goto out_delete_evlist;
}
bzero(&parse_error, sizeof(parse_error));
/*
* Set backward bit, ring buffer should be writing from end. Record
* it in aux evlist
*/
err = parse_events(evlist, "syscalls:sys_enter_prctl/overwrite/", &parse_error);
if (err) {
pr_debug("Failed to parse tracepoint event, try use root\n");
ret = TEST_SKIP;
goto out_delete_evlist;
}
perf_evlist__config(evlist, &opts, NULL);
err = perf_evlist__open(evlist);
if (err < 0) {
pr_debug("perf_evlist__open: %s\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_delete_evlist;
}
ret = TEST_FAIL;
err = do_test(evlist, opts.mmap_pages, &sample_count,
&comm_count);
if (err != TEST_OK)
goto out_delete_evlist;
if ((sample_count != NR_ITERS) || (comm_count != NR_ITERS)) {
pr_err("Unexpected counter: sample_count=%d, comm_count=%d\n",
sample_count, comm_count);
goto out_delete_evlist;
}
err = do_test(evlist, 1, &sample_count, &comm_count);
if (err != TEST_OK)
goto out_delete_evlist;
ret = TEST_OK;
out_delete_evlist:
perf_evlist__delete(evlist);
return ret;
}
int test__openat_syscall_event_on_all_cpus(int subtest __maybe_unused)
{
int err = -1, fd, cpu;
struct cpu_map *cpus;
struct perf_evsel *evsel;
unsigned int nr_openat_calls = 111, i;
cpu_set_t cpu_set;
struct thread_map *threads = thread_map__new(-1, getpid(), UINT_MAX);
char sbuf[STRERR_BUFSIZE];
char errbuf[BUFSIZ];
if (threads == NULL) {
pr_debug("thread_map__new\n");
return -1;
}
cpus = cpu_map__new(NULL);
if (cpus == NULL) {
pr_debug("cpu_map__new\n");
goto out_thread_map_delete;
}
CPU_ZERO(&cpu_set);
evsel = perf_evsel__newtp("syscalls", "sys_enter_openat");
if (IS_ERR(evsel)) {
tracing_path__strerror_open_tp(errno, errbuf, sizeof(errbuf), "syscalls", "sys_enter_openat");
pr_debug("%s\n", errbuf);
goto out_thread_map_delete;
}
if (perf_evsel__open(evsel, cpus, threads) < 0) {
pr_debug("failed to open counter: %s, "
"tweak /proc/sys/kernel/perf_event_paranoid?\n",
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_evsel_delete;
}
for (cpu = 0; cpu < cpus->nr; ++cpu) {
unsigned int ncalls = nr_openat_calls + cpu;
/*
* XXX eventually lift this restriction in a way that
* keeps perf building on older glibc installations
* without CPU_ALLOC. 1024 cpus in 2010 still seems
* a reasonable upper limit tho :-)
*/
if (cpus->map[cpu] >= CPU_SETSIZE) {
pr_debug("Ignoring CPU %d\n", cpus->map[cpu]);
continue;
}
CPU_SET(cpus->map[cpu], &cpu_set);
if (sched_setaffinity(0, sizeof(cpu_set), &cpu_set) < 0) {
pr_debug("sched_setaffinity() failed on CPU %d: %s ",
cpus->map[cpu],
str_error_r(errno, sbuf, sizeof(sbuf)));
goto out_close_fd;
}
for (i = 0; i < ncalls; ++i) {
fd = openat(0, "/etc/passwd", O_RDONLY);
close(fd);
}
CPU_CLR(cpus->map[cpu], &cpu_set);
}
/*
* Here we need to explicitly preallocate the counts, as if
* we use the auto allocation it will allocate just for 1 cpu,
* as we start by cpu 0.
*/
if (perf_evsel__alloc_counts(evsel, cpus->nr, 1) < 0) {
pr_debug("perf_evsel__alloc_counts(ncpus=%d)\n", cpus->nr);
goto out_close_fd;
}
err = 0;
for (cpu = 0; cpu < cpus->nr; ++cpu) {
unsigned int expected;
if (cpus->map[cpu] >= CPU_SETSIZE)
continue;
if (perf_evsel__read_on_cpu(evsel, cpu, 0) < 0) {
pr_debug("perf_evsel__read_on_cpu\n");
err = -1;
break;
}
expected = nr_openat_calls + cpu;
if (perf_counts(evsel->counts, cpu, 0)->val != expected) {
pr_debug("perf_evsel__read_on_cpu: expected to intercept %d calls on cpu %d, got %" PRIu64 "\n",
expected, cpus->map[cpu], perf_counts(evsel->counts, cpu, 0)->val);
err = -1;
}
}
perf_evsel__free_counts(evsel);
out_close_fd:
perf_evsel__close_fd(evsel, 1, threads->nr);
out_evsel_delete:
perf_evsel__delete(evsel);
out_thread_map_delete:
thread_map__put(threads);
return err;
}
static int
read_from_pipe(const char *cmd, void **p_buf, size_t *p_read_sz)
{
int err = 0;
void *buf = NULL;
FILE *file = NULL;
size_t read_sz = 0, buf_sz = 0;
char serr[STRERR_BUFSIZE];
file = popen(cmd, "r");
if (!file) {
pr_err("ERROR: unable to popen cmd: %s\n",
str_error_r(errno, serr, sizeof(serr)));
return -EINVAL;
}
while (!feof(file) && !ferror(file)) {
/*
* Make buf_sz always have obe byte extra space so we
* can put '\0' there.
*/
if (buf_sz - read_sz < READ_SIZE + 1) {
void *new_buf;
buf_sz = read_sz + READ_SIZE + 1;
new_buf = realloc(buf, buf_sz);
if (!new_buf) {
pr_err("ERROR: failed to realloc memory\n");
err = -ENOMEM;
goto errout;
}
buf = new_buf;
}
read_sz += fread(buf + read_sz, 1, READ_SIZE, file);
}
if (buf_sz - read_sz < 1) {
pr_err("ERROR: internal error\n");
err = -EINVAL;
goto errout;
}
if (ferror(file)) {
pr_err("ERROR: error occurred when reading from pipe: %s\n",
str_error_r(errno, serr, sizeof(serr)));
err = -EIO;
goto errout;
}
err = WEXITSTATUS(pclose(file));
file = NULL;
if (err) {
err = -EINVAL;
goto errout;
}
/*
* If buf is string, give it terminal '\0' to make our life
* easier. If buf is not string, that '\0' is out of space
* indicated by read_sz so caller won't even notice it.
*/
((char *)buf)[read_sz] = '\0';
if (!p_buf)
free(buf);
else
*p_buf = buf;
if (p_read_sz)
*p_read_sz = read_sz;
return 0;
errout:
if (file)
pclose(file);
free(buf);
if (p_buf)
*p_buf = NULL;
if (p_read_sz)
*p_read_sz = 0;
return err;
}
