* [PATCH] perf io: Make open and read calls more robust
@ 2022-06-15 6:03 Ian Rogers
2022-06-15 6:50 ` Adrian Hunter
2022-06-15 18:33 ` Andi Kleen
0 siblings, 2 replies; 7+ messages in thread
From: Ian Rogers @ 2022-06-15 6:03 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Alexey Bayduraev, Riccardo Mancini, Adrian Hunter,
Dave Marchevsky, Song Liu, Leo Yan, John Garry, Florian Fischer,
Andi Kleen, Kan Liang, Masami Hiramatsu, Zechuan Chen,
Yang Jihong, German Gomez, Thomas Richter, James Clark,
Denis Nikitin, Lexi Shao, Michael Petlan, Remi Bernon,
Fangrui Song, linux-kernel, linux-perf-users
Cc: Stephane Eranian, Ian Rogers
Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
causes the syscall to need to restart.
Signed-off-by: Ian Rogers <irogers@google.com>
---
tools/lib/api/io.h | 3 ++-
tools/lib/symbol/kallsyms.c | 4 ++-
tools/perf/builtin-daemon.c | 19 +++++++-------
tools/perf/builtin-ftrace.c | 18 ++++++-------
tools/perf/builtin-inject.c | 2 +-
tools/perf/builtin-kvm.c | 2 +-
tools/perf/builtin-record.c | 6 ++---
tools/perf/builtin-sched.c | 2 +-
tools/perf/builtin-script.c | 2 +-
tools/perf/util/copyfile.c | 2 +-
tools/perf/util/data.c | 11 ++++----
tools/perf/util/dso.c | 6 ++---
tools/perf/util/evlist.c | 6 ++---
tools/perf/util/lzma.c | 4 +--
tools/perf/util/namespaces.c | 5 ++--
tools/perf/util/parse-events.c | 6 ++---
tools/perf/util/pmu.c | 12 ++++-----
tools/perf/util/probe-event.c | 4 +--
tools/perf/util/probe-file.c | 10 ++++----
tools/perf/util/session.c | 2 +-
tools/perf/util/symbol-elf.c | 32 ++++++++++++------------
tools/perf/util/symbol-minimal.c | 6 ++---
tools/perf/util/symbol.c | 4 +--
tools/perf/util/synthetic-events.c | 7 +++---
tools/perf/util/trace-event-info.c | 6 ++---
tools/perf/util/trace-event-read.c | 5 ++--
tools/perf/util/unwind-libunwind-local.c | 4 +--
tools/perf/util/zlib.c | 5 ++--
28 files changed, 102 insertions(+), 93 deletions(-)
diff --git a/tools/lib/api/io.h b/tools/lib/api/io.h
index 777c20f6b604..b292bbd39fb0 100644
--- a/tools/lib/api/io.h
+++ b/tools/lib/api/io.h
@@ -7,6 +7,7 @@
#ifndef __API_IO__
#define __API_IO__
+#include <errno.h>
#include <stdlib.h>
#include <unistd.h>
@@ -45,7 +46,7 @@ static inline int io__get_char(struct io *io)
return -1;
if (ptr == io->end) {
- ssize_t n = read(io->fd, io->buf, io->buf_len);
+ ssize_t n = TEMP_FAILURE_RETRY(read(io->fd, io->buf, io->buf_len));
if (n <= 0) {
io->eof = true;
diff --git a/tools/lib/symbol/kallsyms.c b/tools/lib/symbol/kallsyms.c
index e335ac2b9e19..5436c038c379 100644
--- a/tools/lib/symbol/kallsyms.c
+++ b/tools/lib/symbol/kallsyms.c
@@ -3,7 +3,9 @@
#include "api/io.h"
#include <stdio.h>
#include <sys/stat.h>
+#include <errno.h>
#include <fcntl.h>
+#include <unistd.h>
u8 kallsyms2elf_type(char type)
{
@@ -36,7 +38,7 @@ int kallsyms__parse(const char *filename, void *arg,
char bf[BUFSIZ];
int err;
- io.fd = open(filename, O_RDONLY, 0);
+ io.fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY, 0));
if (io.fd < 0)
return -1;
diff --git a/tools/perf/builtin-daemon.c b/tools/perf/builtin-daemon.c
index 6cb3f6cc36d0..d4e1a5636056 100644
--- a/tools/perf/builtin-daemon.c
+++ b/tools/perf/builtin-daemon.c
@@ -351,7 +351,7 @@ static int daemon_session__run(struct daemon_session *session,
return -1;
}
- fd = open("/dev/null", O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open("/dev/null", O_RDONLY));
if (fd < 0) {
perror("failed: open /dev/null");
return -1;
@@ -360,7 +360,7 @@ static int daemon_session__run(struct daemon_session *session,
dup2(fd, 0);
close(fd);
- fd = open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644);
+ fd = TEMP_FAILURE_RETRY(open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644));
if (fd < 0) {
perror("failed: open session output");
return -1;
@@ -405,7 +405,8 @@ static pid_t handle_signalfd(struct daemon *daemon)
* coalesced in kernel and we can receive only single signal even
* if multiple SIGCHLD were generated.
*/
- err = read(daemon->signal_fd, &si, sizeof(struct signalfd_siginfo));
+ err = TEMP_FAILURE_RETRY(read(daemon->signal_fd, &si,
+ sizeof(struct signalfd_siginfo)));
if (err != sizeof(struct signalfd_siginfo)) {
pr_err("failed to read signal fd\n");
return -1;
@@ -522,7 +523,7 @@ static int daemon_session__control(struct daemon_session *session,
scnprintf(control_path, sizeof(control_path), "%s/%s",
session->base, SESSION_CONTROL);
- control = open(control_path, O_WRONLY|O_NONBLOCK);
+ control = TEMP_FAILURE_RETRY(open(control_path, O_WRONLY|O_NONBLOCK));
if (!control)
return -1;
@@ -531,7 +532,7 @@ static int daemon_session__control(struct daemon_session *session,
scnprintf(ack_path, sizeof(ack_path), "%s/%s",
session->base, SESSION_ACK);
- ack = open(ack_path, O_RDONLY, O_NONBLOCK);
+ ack = TEMP_FAILURE_RETRY(open(ack_path, O_RDONLY, O_NONBLOCK));
if (!ack) {
close(control);
return -1;
@@ -564,7 +565,7 @@ static int daemon_session__control(struct daemon_session *session,
goto out;
}
- err = read(ack, buf, sizeof(buf));
+ err = TEMP_FAILURE_RETRY(read(ack, buf, sizeof(buf)));
if (err > 0)
ret = strcmp(buf, "ack\n");
else
@@ -1078,7 +1079,7 @@ static int handle_config_changes(struct daemon *daemon, int conf_fd,
ssize_t len;
while (!(*config_changed)) {
- len = read(conf_fd, buf, sizeof(buf));
+ len = TEMP_FAILURE_RETRY(read(conf_fd, buf, sizeof(buf)));
if (len == -1) {
if (errno != EAGAIN) {
perror("failed: read");
@@ -1147,7 +1148,7 @@ static int check_lock(struct daemon *daemon)
scnprintf(path, sizeof(path), "%s/lock", daemon->base);
- fd = open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640);
+ fd = TEMP_FAILURE_RETRY(open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640));
if (fd < 0)
return -1;
@@ -1201,7 +1202,7 @@ static int go_background(struct daemon *daemon)
return -1;
}
- fd = open("output", O_RDWR|O_CREAT|O_TRUNC, 0644);
+ fd = TEMP_FAILURE_RETRY(open("output", O_RDWR|O_CREAT|O_TRUNC, 0644));
if (fd < 0) {
perror("failed: open");
return -1;
diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
index 7de07bb16d23..7553e20f38b2 100644
--- a/tools/perf/builtin-ftrace.c
+++ b/tools/perf/builtin-ftrace.c
@@ -79,7 +79,7 @@ static int __write_tracing_file(const char *name, const char *val, bool append)
else
flags |= O_TRUNC;
- fd = open(file, flags);
+ fd = TEMP_FAILURE_RETRY(open(file, flags));
if (fd < 0) {
pr_debug("cannot open tracing file: %s: %s\n",
name, str_error_r(errno, errbuf, sizeof(errbuf)));
@@ -132,7 +132,7 @@ static int read_tracing_file_to_stdout(const char *name)
return -1;
}
- fd = open(file, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
if (fd < 0) {
pr_debug("cannot open tracing file: %s: %s\n",
name, str_error_r(errno, buf, sizeof(buf)));
@@ -141,7 +141,7 @@ static int read_tracing_file_to_stdout(const char *name)
/* read contents to stdout */
while (true) {
- int n = read(fd, buf, sizeof(buf));
+ int n = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
if (n == 0)
break;
else if (n < 0)
@@ -608,7 +608,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
goto out_reset;
}
- trace_fd = open(trace_file, O_RDONLY);
+ trace_fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
put_tracing_file(trace_file);
@@ -645,7 +645,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
break;
if (pollfd.revents & POLLIN) {
- int n = read(trace_fd, buf, sizeof(buf));
+ int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
if (n < 0)
break;
if (fwrite(buf, n, 1, stdout) != 1)
@@ -665,7 +665,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
/* read remaining buffer contents */
while (true) {
- int n = read(trace_fd, buf, sizeof(buf));
+ int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
if (n <= 0)
break;
if (fwrite(buf, n, 1, stdout) != 1)
@@ -826,7 +826,7 @@ static int prepare_func_latency(struct perf_ftrace *ftrace)
return -1;
}
- fd = open(trace_file, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
if (fd < 0)
pr_err("failed to open trace_pipe\n");
@@ -913,7 +913,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
break;
if (pollfd.revents & POLLIN) {
- int n = read(trace_fd, buf, sizeof(buf) - 1);
+ int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
if (n < 0)
break;
@@ -931,7 +931,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
/* read remaining buffer contents */
while (!ftrace->target.use_bpf) {
- int n = read(trace_fd, buf, sizeof(buf) - 1);
+ int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
if (n <= 0)
break;
make_histogram(buckets, buf, n, line, ftrace->use_nsec);
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index a75bf11585b5..960b4a50583e 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -155,7 +155,7 @@ static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
int ret;
while (size > 0) {
- ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
+ ssz = TEMP_FAILURE_RETRY(read(fd, buf, min(size, (off_t)sizeof(buf))));
if (ssz < 0)
return -errno;
ret = output_bytes(inject, buf, ssz);
diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index 3696ae97f149..4dd8292d1941 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -880,7 +880,7 @@ static int perf_kvm__handle_timerfd(struct perf_kvm_stat *kvm)
uint64_t c;
int rc;
- rc = read(kvm->timerfd, &c, sizeof(uint64_t));
+ rc = TEMP_FAILURE_RETRY(read(kvm->timerfd, &c, sizeof(uint64_t)));
if (rc < 0) {
if (errno == EAGAIN)
return 0;
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 9a71f0330137..361610bc96b6 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -902,7 +902,7 @@ static bool record__kcore_readable(struct machine *machine)
scnprintf(kcore, sizeof(kcore), "%s/proc/kcore", machine->root_dir);
- fd = open(kcore, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(kcore, O_RDONLY));
if (fd < 0)
return false;
@@ -2049,7 +2049,7 @@ static int record__terminate_thread(struct record_thread *thread_data)
close(thread_data->pipes.msg[1]);
thread_data->pipes.msg[1] = -1;
- err = read(thread_data->pipes.ack[0], &ack, sizeof(ack));
+ err = TEMP_FAILURE_RETRY(read(thread_data->pipes.ack[0], &ack, sizeof(ack)));
if (err > 0)
pr_debug2("threads[%d]: sent %s\n", tid, thread_msg_tags[ack]);
else
@@ -2097,7 +2097,7 @@ static int record__start_threads(struct record *rec)
goto out_err;
}
- err = read(thread_data[t].pipes.ack[0], &msg, sizeof(msg));
+ err = TEMP_FAILURE_RETRY(read(thread_data[t].pipes.ack[0], &msg, sizeof(msg)));
if (err > 0)
pr_debug2("threads[%d]: sent %s\n", rec->thread_data[t].tid,
thread_msg_tags[msg]);
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 646bd938927a..149787268d02 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -604,7 +604,7 @@ static u64 get_cpu_usage_nsec_self(int fd)
u64 runtime;
int ret;
- ret = read(fd, &runtime, sizeof(runtime));
+ ret = TEMP_FAILURE_RETRY(read(fd, &runtime, sizeof(runtime)));
BUG_ON(ret != sizeof(runtime));
return runtime;
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index c689054002cc..844f5350bfdf 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -4207,7 +4207,7 @@ int cmd_script(int argc, const char **argv)
goto out_delete;
}
- input = open(data.path, O_RDONLY); /* input_name */
+ input = TEMP_FAILURE_RETRY(open(data.path, O_RDONLY)); /* input_name */
if (input < 0) {
err = -errno;
perror("failed to open file");
diff --git a/tools/perf/util/copyfile.c b/tools/perf/util/copyfile.c
index 47e03de7c235..c7bff4943936 100644
--- a/tools/perf/util/copyfile.c
+++ b/tools/perf/util/copyfile.c
@@ -112,7 +112,7 @@ static int copyfile_mode_ns(const char *from, const char *to, mode_t mode,
goto out_close_to;
nsinfo__mountns_enter(nsi, &nsc);
- fromfd = open(from, O_RDONLY);
+ fromfd = TEMP_FAILURE_RETRY(open(from, O_RDONLY));
nsinfo__mountns_exit(&nsc);
if (fromfd < 0)
goto out_close_to;
diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c
index caabeac24c69..a9066a3d914f 100644
--- a/tools/perf/util/data.c
+++ b/tools/perf/util/data.c
@@ -53,7 +53,8 @@ int perf_data__create_dir(struct perf_data *data, int nr)
goto out_err;
}
- ret = open(file->path, O_RDWR|O_CREAT|O_TRUNC, S_IRUSR|S_IWUSR);
+ ret = TEMP_FAILURE_RETRY(open(file->path, O_RDWR|O_CREAT|O_TRUNC,
+ S_IRUSR|S_IWUSR));
if (ret < 0) {
ret = -errno;
goto out_err;
@@ -123,7 +124,7 @@ int perf_data__open_dir(struct perf_data *data)
if (!file->path)
goto out_err;
- ret = open(file->path, O_RDONLY);
+ ret = TEMP_FAILURE_RETRY(open(file->path, O_RDONLY));
if (ret < 0)
goto out_err;
@@ -248,7 +249,7 @@ static int open_file_read(struct perf_data *data)
int fd;
char sbuf[STRERR_BUFSIZE];
- fd = open(data->file.path, flags);
+ fd = TEMP_FAILURE_RETRY(open(data->file.path, flags));
if (fd < 0) {
int err = errno;
@@ -288,8 +289,8 @@ static int open_file_write(struct perf_data *data)
int fd;
char sbuf[STRERR_BUFSIZE];
- fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
- S_IRUSR|S_IWUSR);
+ fd = TEMP_FAILURE_RETRY(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,
diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
index 5ac13958d1bd..40befa2429f9 100644
--- a/tools/perf/util/dso.c
+++ b/tools/perf/util/dso.c
@@ -300,7 +300,7 @@ int filename__decompress(const char *name, char *pathname,
* descriptor to the uncompressed file.
*/
if (!compressions[comp].is_compressed(name))
- return open(name, O_RDONLY);
+ return TEMP_FAILURE_RETRY(open(name, O_RDONLY));
fd = mkstemp(tmpbuf);
if (fd < 0) {
@@ -476,7 +476,7 @@ static int do_open(char *name)
char sbuf[STRERR_BUFSIZE];
do {
- fd = open(name, O_RDONLY|O_CLOEXEC);
+ fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY|O_CLOEXEC));
if (fd >= 0)
return fd;
@@ -898,7 +898,7 @@ static ssize_t file_read(struct dso *dso, struct machine *machine,
goto out;
}
- ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset);
+ ret = TEMP_FAILURE_RETRY(pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset));
out:
pthread_mutex_unlock(&dso__data_open_lock);
return ret;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 48af7d379d82..ba0dcdf91120 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -1440,7 +1440,7 @@ int evlist__prepare_workload(struct evlist *evlist, struct target *target, const
/*
* Wait until the parent tells us to go.
*/
- ret = read(go_pipe[0], &bf, 1);
+ ret = TEMP_FAILURE_RETRY(read(go_pipe[0], &bf, 1));
/*
* The parent will ask for the execvp() to be performed by
* writing exactly one byte, in workload.cork_fd, usually via
@@ -1837,7 +1837,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
* O_RDWR avoids POLLHUPs which is necessary to allow the other
* end of a FIFO to be repeatedly opened and closed.
*/
- fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC);
+ fd = TEMP_FAILURE_RETRY(open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC));
if (fd < 0) {
pr_err("Failed to open '%s'\n", s);
ret = -errno;
@@ -1848,7 +1848,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
if (p && *++p) {
/* O_RDWR | O_NONBLOCK means the other end need not be open */
- fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC);
+ fd = TEMP_FAILURE_RETRY(open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC));
if (fd < 0) {
pr_err("Failed to open '%s'\n", p);
ret = -errno;
diff --git a/tools/perf/util/lzma.c b/tools/perf/util/lzma.c
index 51424cdc3b68..93e0a6a53e28 100644
--- a/tools/perf/util/lzma.c
+++ b/tools/perf/util/lzma.c
@@ -109,7 +109,7 @@ int lzma_decompress_to_file(const char *input, int output_fd)
bool lzma_is_compressed(const char *input)
{
- int fd = open(input, O_RDONLY);
+ int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
const uint8_t magic[6] = { 0xFD, '7', 'z', 'X', 'Z', 0x00 };
char buf[6] = { 0 };
ssize_t rc;
@@ -117,7 +117,7 @@ bool lzma_is_compressed(const char *input)
if (fd < 0)
return -1;
- rc = read(fd, buf, sizeof(buf));
+ rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
close(fd);
return rc == sizeof(buf) ?
memcmp(buf, magic, sizeof(buf)) == 0 : false;
diff --git a/tools/perf/util/namespaces.c b/tools/perf/util/namespaces.c
index dd536220cdb9..89edfba6af06 100644
--- a/tools/perf/util/namespaces.c
+++ b/tools/perf/util/namespaces.c
@@ -9,6 +9,7 @@
#include "get_current_dir_name.h"
#include <sys/types.h>
#include <sys/stat.h>
+#include <errno.h>
#include <fcntl.h>
#include <limits.h>
#include <sched.h>
@@ -266,11 +267,11 @@ void nsinfo__mountns_enter(struct nsinfo *nsi,
if (!oldcwd)
return;
- oldns = open(curpath, O_RDONLY);
+ oldns = TEMP_FAILURE_RETRY(open(curpath, O_RDONLY));
if (oldns < 0)
goto errout;
- newns = open(nsi->mntns_path, O_RDONLY);
+ newns = TEMP_FAILURE_RETRY(open(nsi->mntns_path, O_RDONLY));
if (newns < 0)
goto errout;
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index 7ed235740431..75f8f2e44195 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -189,7 +189,7 @@ static int tp_event_has_id(const char *dir_path, struct dirent *evt_dir)
int fd;
snprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path, evt_dir->d_name);
- fd = open(evt_path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
if (fd < 0)
return -EINVAL;
close(fd);
@@ -233,10 +233,10 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config)
scnprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path,
evt_dirent->d_name);
- fd = open(evt_path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
if (fd < 0)
continue;
- if (read(fd, id_buf, sizeof(id_buf)) < 0) {
+ if (TEMP_FAILURE_RETRY(read(fd, id_buf, sizeof(id_buf))) < 0) {
close(fd);
continue;
}
diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
index 9a1c7e63e663..59252cfc2b5b 100644
--- a/tools/perf/util/pmu.c
+++ b/tools/perf/util/pmu.c
@@ -154,14 +154,14 @@ static int perf_pmu__parse_scale(struct perf_pmu_alias *alias, char *dir, char *
scnprintf(path, PATH_MAX, "%s/%s.scale", dir, name);
- fd = open(path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
if (fd == -1)
return -1;
if (fstat(fd, &st) < 0)
goto error;
- sret = read(fd, scale, sizeof(scale)-1);
+ sret = TEMP_FAILURE_RETRY(read(fd, scale, sizeof(scale)-1));
if (sret < 0)
goto error;
@@ -184,11 +184,11 @@ static int perf_pmu__parse_unit(struct perf_pmu_alias *alias, char *dir, char *n
scnprintf(path, PATH_MAX, "%s/%s.unit", dir, name);
- fd = open(path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
if (fd == -1)
return -1;
- sret = read(fd, alias->unit, UNIT_MAX_LEN);
+ sret = TEMP_FAILURE_RETRY(read(fd, alias->unit, UNIT_MAX_LEN));
if (sret < 0)
goto error;
@@ -214,7 +214,7 @@ perf_pmu__parse_per_pkg(struct perf_pmu_alias *alias, char *dir, char *name)
scnprintf(path, PATH_MAX, "%s/%s.per-pkg", dir, name);
- fd = open(path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
if (fd == -1)
return -1;
@@ -232,7 +232,7 @@ static int perf_pmu__parse_snapshot(struct perf_pmu_alias *alias,
scnprintf(path, PATH_MAX, "%s/%s.snapshot", dir, name);
- fd = open(path, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
if (fd == -1)
return -1;
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 062b5cbe67af..7838889f55df 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -276,7 +276,7 @@ static char *find_module_name(const char *module)
char *mod_name = NULL;
int name_offset;
- fd = open(module, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(module, O_RDONLY));
if (fd < 0)
return NULL;
@@ -598,7 +598,7 @@ static int get_text_start_address(const char *exec, u64 *address,
struct nscookie nsc;
nsinfo__mountns_enter(nsi, &nsc);
- fd = open(exec, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(exec, O_RDONLY));
nsinfo__mountns_exit(&nsc);
if (fd < 0)
return -errno;
diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
index 3d50de3217d5..b788932656e8 100644
--- a/tools/perf/util/probe-file.c
+++ b/tools/perf/util/probe-file.c
@@ -115,9 +115,9 @@ int open_trace_file(const char *trace_file, bool readwrite)
if (ret >= 0) {
pr_debug("Opening %s write=%d\n", buf, readwrite);
if (readwrite && !probe_event_dry_run)
- ret = open(buf, O_RDWR | O_APPEND, 0);
+ ret = TEMP_FAILURE_RETRY(open(buf, O_RDWR | O_APPEND, 0));
else
- ret = open(buf, O_RDONLY, 0);
+ ret = TEMP_FAILURE_RETRY(open(buf, O_RDONLY, 0));
if (ret < 0)
ret = -errno;
@@ -180,7 +180,7 @@ struct strlist *probe_file__get_rawlist(int fd)
if (sl == NULL)
return NULL;
- fddup = dup(fd);
+ fddup = TEMP_FAILURE_RETRY(dup(fd));
if (fddup < 0)
goto out_free_sl;
@@ -498,7 +498,7 @@ static int probe_cache__open(struct probe_cache *pcache, const char *target,
}
snprintf(cpath, PATH_MAX, "%s/probes", dir_name);
- fd = open(cpath, O_CREAT | O_RDWR, 0644);
+ fd = TEMP_FAILURE_RETRY(open(cpath, O_CREAT | O_RDWR, 0644));
if (fd < 0)
pr_debug("Failed to open cache(%d): %s\n", fd, cpath);
free(dir_name);
@@ -514,7 +514,7 @@ static int probe_cache__load(struct probe_cache *pcache)
int ret = 0, fddup;
FILE *fp;
- fddup = dup(pcache->fd);
+ fddup = TEMP_FAILURE_RETRY(dup(pcache->fd));
if (fddup < 0)
return -errno;
fp = fdopen(fddup, "r");
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 0acb9de54b06..b2fc389cc530 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -384,7 +384,7 @@ static int skipn(int fd, off_t n)
ssize_t ret;
while (n > 0) {
- ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
+ ret = TEMP_FAILURE_RETRY(read(fd, buf, min(n, (off_t)sizeof(buf))));
if (ret <= 0)
return ret;
n -= ret;
diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index ecd377938eea..02556070a2ce 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -577,7 +577,7 @@ static int read_build_id(const char *filename, struct build_id *bid)
if (size < BUILD_ID_SIZE)
goto out;
- fd = open(filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
if (fd < 0)
goto out;
@@ -638,7 +638,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
size_t size = sizeof(bid->data);
int fd, err = -1;
- fd = open(filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
if (fd < 0)
goto out;
@@ -647,24 +647,24 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
GElf_Nhdr nhdr;
size_t namesz, descsz;
- if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr))
+ if (TEMP_FAILURE_RETRY(read(fd, &nhdr, sizeof(nhdr))) != sizeof(nhdr))
break;
namesz = NOTE_ALIGN(nhdr.n_namesz);
descsz = NOTE_ALIGN(nhdr.n_descsz);
if (nhdr.n_type == NT_GNU_BUILD_ID &&
nhdr.n_namesz == sizeof("GNU")) {
- if (read(fd, bf, namesz) != (ssize_t)namesz)
+ if (TEMP_FAILURE_RETRY(read(fd, bf, namesz)) != (ssize_t)namesz)
break;
if (memcmp(bf, "GNU", sizeof("GNU")) == 0) {
size_t sz = min(descsz, size);
- if (read(fd, bid->data, sz) == (ssize_t)sz) {
+ if (TEMP_FAILURE_RETRY(read(fd, bid->data, sz)) == (ssize_t)sz) {
memset(bid->data + sz, 0, size - sz);
bid->size = sz;
err = 0;
break;
}
- } else if (read(fd, bf, descsz) != (ssize_t)descsz)
+ } else if (TEMP_FAILURE_RETRY(read(fd, bf, descsz)) != (ssize_t)descsz)
break;
} else {
int n = namesz + descsz;
@@ -674,7 +674,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
__func__, filename, nhdr.n_namesz, nhdr.n_descsz);
}
- if (read(fd, bf, n) != n)
+ if (TEMP_FAILURE_RETRY(read(fd, bf, n)) != n)
break;
}
}
@@ -732,7 +732,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,
Elf_Scn *sec;
Elf_Kind ek;
- fd = open(filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
if (fd < 0)
goto out;
@@ -844,7 +844,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
type = dso->symtab_type;
} else {
- fd = open(name, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
if (fd < 0) {
dso->load_errno = errno;
return -1;
@@ -1454,7 +1454,7 @@ static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len)
if (len < n)
n = len;
/* Use read because mmap won't work on proc files */
- r = read(from, buf, n);
+ r = TEMP_FAILURE_RETRY(read(from, buf, n));
if (r < 0)
goto out;
if (!r)
@@ -1485,7 +1485,7 @@ static int kcore__open(struct kcore *kcore, const char *filename)
{
GElf_Ehdr *ehdr;
- kcore->fd = open(filename, O_RDONLY);
+ kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
if (kcore->fd == -1)
return -1;
@@ -1518,7 +1518,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,
if (temp)
kcore->fd = mkstemp(filename);
else
- kcore->fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400);
+ kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400));
if (kcore->fd == -1)
return -1;
@@ -1966,7 +1966,7 @@ static int kcore_copy__compare_fds(int from, int to)
while (1) {
/* Use read because mmap won't work on proc files */
- ret = read(from, buf_from, page_size);
+ ret = TEMP_FAILURE_RETRY(read(from, buf_from, page_size));
if (ret < 0)
goto out;
@@ -1994,11 +1994,11 @@ static int kcore_copy__compare_files(const char *from_filename,
{
int from, to, err = -1;
- from = open(from_filename, O_RDONLY);
+ from = TEMP_FAILURE_RETRY(open(from_filename, O_RDONLY));
if (from < 0)
return -1;
- to = open(to_filename, O_RDONLY);
+ to = TEMP_FAILURE_RETRY(open(to_filename, O_RDONLY));
if (to < 0)
goto out_close_from;
@@ -2419,7 +2419,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)
Elf *elf;
int fd, ret;
- fd = open(target, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(target, O_RDONLY));
if (fd < 0)
return -EBADF;
diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
index f9eb0bee7f15..7e25d476b1b2 100644
--- a/tools/perf/util/symbol-minimal.c
+++ b/tools/perf/util/symbol-minimal.c
@@ -230,7 +230,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
size_t buf_size;
void *buf;
- fd = open(filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
if (fd < 0)
return -1;
@@ -242,7 +242,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
if (buf == NULL)
goto out;
- if (read(fd, buf, buf_size) != (ssize_t) buf_size)
+ if (TEMP_FAILURE_RETRY(read(fd, buf, buf_size)) != (ssize_t) buf_size)
goto out_free;
ret = read_build_id(buf, buf_size, bid, false);
@@ -256,7 +256,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
enum dso_binary_type type)
{
- int fd = open(name, O_RDONLY);
+ int fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
if (fd < 0)
goto out_errno;
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index f72baf636724..5212667ea67d 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1335,7 +1335,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
md.dso = dso;
INIT_LIST_HEAD(&md.maps);
- fd = open(kcore_filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(kcore_filename, O_RDONLY));
if (fd < 0) {
pr_debug("Failed to open %s. Note /proc/kcore requires CAP_SYS_RAWIO capability to access.\n",
kcore_filename);
@@ -2156,7 +2156,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
*/
static bool filename__readable(const char *file)
{
- int fd = open(file, O_RDONLY);
+ int fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
if (fd < 0)
return false;
close(fd);
diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
index 76beda3e1a10..9dcc268bd69c 100644
--- a/tools/perf/util/synthetic-events.c
+++ b/tools/perf/util/synthetic-events.c
@@ -40,6 +40,7 @@
#include <api/io.h>
#include <sys/types.h>
#include <sys/stat.h>
+#include <errno.h>
#include <fcntl.h>
#include <unistd.h>
@@ -86,13 +87,13 @@ static int perf_event__get_comm_ids(pid_t pid, pid_t tid, char *comm, size_t len
else
snprintf(bf, sizeof(bf), "/proc/%d/status", tid);
- fd = open(bf, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY));
if (fd < 0) {
pr_debug("couldn't open %s\n", bf);
return -1;
}
- n = read(fd, bf, sizeof(bf) - 1);
+ n = TEMP_FAILURE_RETRY(read(fd, bf, sizeof(bf) - 1));
close(fd);
if (n <= 0) {
pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
@@ -410,7 +411,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
snprintf(bf, sizeof(bf), "%s/proc/%d/task/%d/maps",
machine->root_dir, pid, pid);
- io.fd = open(bf, O_RDONLY, 0);
+ io.fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY, 0));
if (io.fd < 0) {
/*
* We raced with a task exiting - just return:
diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c
index a65f65d0857e..c917a4217c6f 100644
--- a/tools/perf/util/trace-event-info.c
+++ b/tools/perf/util/trace-event-info.c
@@ -48,7 +48,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
int r, fd;
int err = -EIO;
- fd = open(file, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
if (fd < 0) {
pr_debug("Can't read '%s'", file);
return -errno;
@@ -61,7 +61,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
}
do {
- r = read(fd, buf, BUFSIZ);
+ r = TEMP_FAILURE_RETRY(read(fd, buf, BUFSIZ));
if (r > 0) {
size += r;
if (write(output_fd, buf, r) != r)
@@ -519,7 +519,7 @@ struct tracing_data *tracing_data_get(struct list_head *pattrs,
return NULL;
}
- temp_fd = open(tdata->temp_file, O_RDWR);
+ temp_fd = TEMP_FAILURE_RETRY(open(tdata->temp_file, O_RDWR));
if (temp_fd < 0) {
pr_debug("Can't read '%s'", tdata->temp_file);
free(tdata);
diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c
index 8a01af783310..7086770ec225 100644
--- a/tools/perf/util/trace-event-read.c
+++ b/tools/perf/util/trace-event-read.c
@@ -11,6 +11,7 @@
#include <sys/stat.h>
#include <sys/wait.h>
#include <sys/mman.h>
+#include <errno.h>
#include <fcntl.h>
#include <unistd.h>
#include <errno.h>
@@ -28,7 +29,7 @@ static int __do_read(int fd, void *buf, int size)
int rsize = size;
while (size) {
- int ret = read(fd, buf, size);
+ int ret = TEMP_FAILURE_RETRY(read(fd, buf, size));
if (ret <= 0)
return -1;
@@ -105,7 +106,7 @@ static char *read_string(void)
char c;
for (;;) {
- r = read(input_fd, &c, 1);
+ r = TEMP_FAILURE_RETRY(read(input_fd, &c, 1));
if (r < 0) {
pr_debug("reading input file");
goto out;
diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c
index 37622699c91a..2ece387a73c6 100644
--- a/tools/perf/util/unwind-libunwind-local.c
+++ b/tools/perf/util/unwind-libunwind-local.c
@@ -362,7 +362,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
}
if (ofs <= 0) {
- fd = open(dso->symsrc_filename, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(dso->symsrc_filename, O_RDONLY));
if (fd >= 0) {
ofs = elf_section_offset(fd, ".debug_frame");
close(fd);
@@ -377,7 +377,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
dso, DSO_BINARY_TYPE__DEBUGLINK,
machine->root_dir, debuglink, PATH_MAX);
if (!ret) {
- fd = open(debuglink, O_RDONLY);
+ fd = TEMP_FAILURE_RETRY(open(debuglink, O_RDONLY));
if (fd >= 0) {
ofs = elf_section_offset(fd,
".debug_frame");
diff --git a/tools/perf/util/zlib.c b/tools/perf/util/zlib.c
index 78d2297c1b67..f4563ea094c8 100644
--- a/tools/perf/util/zlib.c
+++ b/tools/perf/util/zlib.c
@@ -1,4 +1,5 @@
// SPDX-License-Identifier: GPL-2.0
+#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <string.h>
@@ -82,7 +83,7 @@ int gzip_decompress_to_file(const char *input, int output_fd)
bool gzip_is_compressed(const char *input)
{
- int fd = open(input, O_RDONLY);
+ int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
const uint8_t magic[2] = { 0x1f, 0x8b };
char buf[2] = { 0 };
ssize_t rc;
@@ -90,7 +91,7 @@ bool gzip_is_compressed(const char *input)
if (fd < 0)
return -1;
- rc = read(fd, buf, sizeof(buf));
+ rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
close(fd);
return rc == sizeof(buf) ?
memcmp(buf, magic, sizeof(buf)) == 0 : false;
--
2.36.1.476.g0c4daa206d-goog
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 6:03 [PATCH] perf io: Make open and read calls more robust Ian Rogers
@ 2022-06-15 6:50 ` Adrian Hunter
2022-06-15 15:34 ` Ian Rogers
2022-06-15 18:33 ` Andi Kleen
1 sibling, 1 reply; 7+ messages in thread
From: Adrian Hunter @ 2022-06-15 6:50 UTC (permalink / raw)
To: Ian Rogers, Peter Zijlstra, Ingo Molnar,
Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin,
Jiri Olsa, Namhyung Kim, Alexey Bayduraev, Riccardo Mancini,
Dave Marchevsky, Song Liu, Leo Yan, John Garry, Florian Fischer,
Andi Kleen, Kan Liang, Masami Hiramatsu, Zechuan Chen,
Yang Jihong, German Gomez, Thomas Richter, James Clark,
Denis Nikitin, Lexi Shao, Michael Petlan, Remi Bernon,
Fangrui Song, linux-kernel, linux-perf-users
Cc: Stephane Eranian
On 15/06/22 09:03, Ian Rogers wrote:
> Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
> causes the syscall to need to restart.
Please expand on why you want this change. Can you give an example
of where it makes a difference?
It looks pretty ugly so maybe making replacement functions or macros
is better e.g.
#define open_retry(a, ...) \
({ long int __result; \
do __result = (long int) open((a), ##__VA_ARGS__); \
while (__result == -1L && errno == EINTR); \
__result; })
>
> Signed-off-by: Ian Rogers <irogers@google.com>
> ---
> tools/lib/api/io.h | 3 ++-
> tools/lib/symbol/kallsyms.c | 4 ++-
> tools/perf/builtin-daemon.c | 19 +++++++-------
> tools/perf/builtin-ftrace.c | 18 ++++++-------
> tools/perf/builtin-inject.c | 2 +-
> tools/perf/builtin-kvm.c | 2 +-
> tools/perf/builtin-record.c | 6 ++---
> tools/perf/builtin-sched.c | 2 +-
> tools/perf/builtin-script.c | 2 +-
> tools/perf/util/copyfile.c | 2 +-
> tools/perf/util/data.c | 11 ++++----
> tools/perf/util/dso.c | 6 ++---
> tools/perf/util/evlist.c | 6 ++---
> tools/perf/util/lzma.c | 4 +--
> tools/perf/util/namespaces.c | 5 ++--
> tools/perf/util/parse-events.c | 6 ++---
> tools/perf/util/pmu.c | 12 ++++-----
> tools/perf/util/probe-event.c | 4 +--
> tools/perf/util/probe-file.c | 10 ++++----
> tools/perf/util/session.c | 2 +-
> tools/perf/util/symbol-elf.c | 32 ++++++++++++------------
> tools/perf/util/symbol-minimal.c | 6 ++---
> tools/perf/util/symbol.c | 4 +--
> tools/perf/util/synthetic-events.c | 7 +++---
> tools/perf/util/trace-event-info.c | 6 ++---
> tools/perf/util/trace-event-read.c | 5 ++--
> tools/perf/util/unwind-libunwind-local.c | 4 +--
> tools/perf/util/zlib.c | 5 ++--
> 28 files changed, 102 insertions(+), 93 deletions(-)
>
> diff --git a/tools/lib/api/io.h b/tools/lib/api/io.h
> index 777c20f6b604..b292bbd39fb0 100644
> --- a/tools/lib/api/io.h
> +++ b/tools/lib/api/io.h
> @@ -7,6 +7,7 @@
> #ifndef __API_IO__
> #define __API_IO__
>
> +#include <errno.h>
> #include <stdlib.h>
> #include <unistd.h>
>
> @@ -45,7 +46,7 @@ static inline int io__get_char(struct io *io)
> return -1;
>
> if (ptr == io->end) {
> - ssize_t n = read(io->fd, io->buf, io->buf_len);
> + ssize_t n = TEMP_FAILURE_RETRY(read(io->fd, io->buf, io->buf_len));
>
> if (n <= 0) {
> io->eof = true;
> diff --git a/tools/lib/symbol/kallsyms.c b/tools/lib/symbol/kallsyms.c
> index e335ac2b9e19..5436c038c379 100644
> --- a/tools/lib/symbol/kallsyms.c
> +++ b/tools/lib/symbol/kallsyms.c
> @@ -3,7 +3,9 @@
> #include "api/io.h"
> #include <stdio.h>
> #include <sys/stat.h>
> +#include <errno.h>
> #include <fcntl.h>
> +#include <unistd.h>
>
> u8 kallsyms2elf_type(char type)
> {
> @@ -36,7 +38,7 @@ int kallsyms__parse(const char *filename, void *arg,
> char bf[BUFSIZ];
> int err;
>
> - io.fd = open(filename, O_RDONLY, 0);
> + io.fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY, 0));
>
> if (io.fd < 0)
> return -1;
> diff --git a/tools/perf/builtin-daemon.c b/tools/perf/builtin-daemon.c
> index 6cb3f6cc36d0..d4e1a5636056 100644
> --- a/tools/perf/builtin-daemon.c
> +++ b/tools/perf/builtin-daemon.c
> @@ -351,7 +351,7 @@ static int daemon_session__run(struct daemon_session *session,
> return -1;
> }
>
> - fd = open("/dev/null", O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open("/dev/null", O_RDONLY));
> if (fd < 0) {
> perror("failed: open /dev/null");
> return -1;
> @@ -360,7 +360,7 @@ static int daemon_session__run(struct daemon_session *session,
> dup2(fd, 0);
> close(fd);
>
> - fd = open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644);
> + fd = TEMP_FAILURE_RETRY(open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644));
> if (fd < 0) {
> perror("failed: open session output");
> return -1;
> @@ -405,7 +405,8 @@ static pid_t handle_signalfd(struct daemon *daemon)
> * coalesced in kernel and we can receive only single signal even
> * if multiple SIGCHLD were generated.
> */
> - err = read(daemon->signal_fd, &si, sizeof(struct signalfd_siginfo));
> + err = TEMP_FAILURE_RETRY(read(daemon->signal_fd, &si,
> + sizeof(struct signalfd_siginfo)));
> if (err != sizeof(struct signalfd_siginfo)) {
> pr_err("failed to read signal fd\n");
> return -1;
> @@ -522,7 +523,7 @@ static int daemon_session__control(struct daemon_session *session,
> scnprintf(control_path, sizeof(control_path), "%s/%s",
> session->base, SESSION_CONTROL);
>
> - control = open(control_path, O_WRONLY|O_NONBLOCK);
> + control = TEMP_FAILURE_RETRY(open(control_path, O_WRONLY|O_NONBLOCK));
> if (!control)
> return -1;
>
> @@ -531,7 +532,7 @@ static int daemon_session__control(struct daemon_session *session,
> scnprintf(ack_path, sizeof(ack_path), "%s/%s",
> session->base, SESSION_ACK);
>
> - ack = open(ack_path, O_RDONLY, O_NONBLOCK);
> + ack = TEMP_FAILURE_RETRY(open(ack_path, O_RDONLY, O_NONBLOCK));
> if (!ack) {
> close(control);
> return -1;
> @@ -564,7 +565,7 @@ static int daemon_session__control(struct daemon_session *session,
> goto out;
> }
>
> - err = read(ack, buf, sizeof(buf));
> + err = TEMP_FAILURE_RETRY(read(ack, buf, sizeof(buf)));
> if (err > 0)
> ret = strcmp(buf, "ack\n");
> else
> @@ -1078,7 +1079,7 @@ static int handle_config_changes(struct daemon *daemon, int conf_fd,
> ssize_t len;
>
> while (!(*config_changed)) {
> - len = read(conf_fd, buf, sizeof(buf));
> + len = TEMP_FAILURE_RETRY(read(conf_fd, buf, sizeof(buf)));
> if (len == -1) {
> if (errno != EAGAIN) {
> perror("failed: read");
> @@ -1147,7 +1148,7 @@ static int check_lock(struct daemon *daemon)
>
> scnprintf(path, sizeof(path), "%s/lock", daemon->base);
>
> - fd = open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640);
> + fd = TEMP_FAILURE_RETRY(open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640));
> if (fd < 0)
> return -1;
>
> @@ -1201,7 +1202,7 @@ static int go_background(struct daemon *daemon)
> return -1;
> }
>
> - fd = open("output", O_RDWR|O_CREAT|O_TRUNC, 0644);
> + fd = TEMP_FAILURE_RETRY(open("output", O_RDWR|O_CREAT|O_TRUNC, 0644));
> if (fd < 0) {
> perror("failed: open");
> return -1;
> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
> index 7de07bb16d23..7553e20f38b2 100644
> --- a/tools/perf/builtin-ftrace.c
> +++ b/tools/perf/builtin-ftrace.c
> @@ -79,7 +79,7 @@ static int __write_tracing_file(const char *name, const char *val, bool append)
> else
> flags |= O_TRUNC;
>
> - fd = open(file, flags);
> + fd = TEMP_FAILURE_RETRY(open(file, flags));
> if (fd < 0) {
> pr_debug("cannot open tracing file: %s: %s\n",
> name, str_error_r(errno, errbuf, sizeof(errbuf)));
> @@ -132,7 +132,7 @@ static int read_tracing_file_to_stdout(const char *name)
> return -1;
> }
>
> - fd = open(file, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> if (fd < 0) {
> pr_debug("cannot open tracing file: %s: %s\n",
> name, str_error_r(errno, buf, sizeof(buf)));
> @@ -141,7 +141,7 @@ static int read_tracing_file_to_stdout(const char *name)
>
> /* read contents to stdout */
> while (true) {
> - int n = read(fd, buf, sizeof(buf));
> + int n = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> if (n == 0)
> break;
> else if (n < 0)
> @@ -608,7 +608,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> goto out_reset;
> }
>
> - trace_fd = open(trace_file, O_RDONLY);
> + trace_fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
>
> put_tracing_file(trace_file);
>
> @@ -645,7 +645,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> break;
>
> if (pollfd.revents & POLLIN) {
> - int n = read(trace_fd, buf, sizeof(buf));
> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> if (n < 0)
> break;
> if (fwrite(buf, n, 1, stdout) != 1)
> @@ -665,7 +665,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
>
> /* read remaining buffer contents */
> while (true) {
> - int n = read(trace_fd, buf, sizeof(buf));
> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> if (n <= 0)
> break;
> if (fwrite(buf, n, 1, stdout) != 1)
> @@ -826,7 +826,7 @@ static int prepare_func_latency(struct perf_ftrace *ftrace)
> return -1;
> }
>
> - fd = open(trace_file, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
> if (fd < 0)
> pr_err("failed to open trace_pipe\n");
>
> @@ -913,7 +913,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
> break;
>
> if (pollfd.revents & POLLIN) {
> - int n = read(trace_fd, buf, sizeof(buf) - 1);
> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> if (n < 0)
> break;
>
> @@ -931,7 +931,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
>
> /* read remaining buffer contents */
> while (!ftrace->target.use_bpf) {
> - int n = read(trace_fd, buf, sizeof(buf) - 1);
> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> if (n <= 0)
> break;
> make_histogram(buckets, buf, n, line, ftrace->use_nsec);
> diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
> index a75bf11585b5..960b4a50583e 100644
> --- a/tools/perf/builtin-inject.c
> +++ b/tools/perf/builtin-inject.c
> @@ -155,7 +155,7 @@ static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
> int ret;
>
> while (size > 0) {
> - ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
> + ssz = TEMP_FAILURE_RETRY(read(fd, buf, min(size, (off_t)sizeof(buf))));
> if (ssz < 0)
> return -errno;
> ret = output_bytes(inject, buf, ssz);
> diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
> index 3696ae97f149..4dd8292d1941 100644
> --- a/tools/perf/builtin-kvm.c
> +++ b/tools/perf/builtin-kvm.c
> @@ -880,7 +880,7 @@ static int perf_kvm__handle_timerfd(struct perf_kvm_stat *kvm)
> uint64_t c;
> int rc;
>
> - rc = read(kvm->timerfd, &c, sizeof(uint64_t));
> + rc = TEMP_FAILURE_RETRY(read(kvm->timerfd, &c, sizeof(uint64_t)));
> if (rc < 0) {
> if (errno == EAGAIN)
> return 0;
> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> index 9a71f0330137..361610bc96b6 100644
> --- a/tools/perf/builtin-record.c
> +++ b/tools/perf/builtin-record.c
> @@ -902,7 +902,7 @@ static bool record__kcore_readable(struct machine *machine)
>
> scnprintf(kcore, sizeof(kcore), "%s/proc/kcore", machine->root_dir);
>
> - fd = open(kcore, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(kcore, O_RDONLY));
> if (fd < 0)
> return false;
>
> @@ -2049,7 +2049,7 @@ static int record__terminate_thread(struct record_thread *thread_data)
>
> close(thread_data->pipes.msg[1]);
> thread_data->pipes.msg[1] = -1;
> - err = read(thread_data->pipes.ack[0], &ack, sizeof(ack));
> + err = TEMP_FAILURE_RETRY(read(thread_data->pipes.ack[0], &ack, sizeof(ack)));
> if (err > 0)
> pr_debug2("threads[%d]: sent %s\n", tid, thread_msg_tags[ack]);
> else
> @@ -2097,7 +2097,7 @@ static int record__start_threads(struct record *rec)
> goto out_err;
> }
>
> - err = read(thread_data[t].pipes.ack[0], &msg, sizeof(msg));
> + err = TEMP_FAILURE_RETRY(read(thread_data[t].pipes.ack[0], &msg, sizeof(msg)));
> if (err > 0)
> pr_debug2("threads[%d]: sent %s\n", rec->thread_data[t].tid,
> thread_msg_tags[msg]);
> diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
> index 646bd938927a..149787268d02 100644
> --- a/tools/perf/builtin-sched.c
> +++ b/tools/perf/builtin-sched.c
> @@ -604,7 +604,7 @@ static u64 get_cpu_usage_nsec_self(int fd)
> u64 runtime;
> int ret;
>
> - ret = read(fd, &runtime, sizeof(runtime));
> + ret = TEMP_FAILURE_RETRY(read(fd, &runtime, sizeof(runtime)));
> BUG_ON(ret != sizeof(runtime));
>
> return runtime;
> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> index c689054002cc..844f5350bfdf 100644
> --- a/tools/perf/builtin-script.c
> +++ b/tools/perf/builtin-script.c
> @@ -4207,7 +4207,7 @@ int cmd_script(int argc, const char **argv)
> goto out_delete;
> }
>
> - input = open(data.path, O_RDONLY); /* input_name */
> + input = TEMP_FAILURE_RETRY(open(data.path, O_RDONLY)); /* input_name */
> if (input < 0) {
> err = -errno;
> perror("failed to open file");
> diff --git a/tools/perf/util/copyfile.c b/tools/perf/util/copyfile.c
> index 47e03de7c235..c7bff4943936 100644
> --- a/tools/perf/util/copyfile.c
> +++ b/tools/perf/util/copyfile.c
> @@ -112,7 +112,7 @@ static int copyfile_mode_ns(const char *from, const char *to, mode_t mode,
> goto out_close_to;
>
> nsinfo__mountns_enter(nsi, &nsc);
> - fromfd = open(from, O_RDONLY);
> + fromfd = TEMP_FAILURE_RETRY(open(from, O_RDONLY));
> nsinfo__mountns_exit(&nsc);
> if (fromfd < 0)
> goto out_close_to;
> diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c
> index caabeac24c69..a9066a3d914f 100644
> --- a/tools/perf/util/data.c
> +++ b/tools/perf/util/data.c
> @@ -53,7 +53,8 @@ int perf_data__create_dir(struct perf_data *data, int nr)
> goto out_err;
> }
>
> - ret = open(file->path, O_RDWR|O_CREAT|O_TRUNC, S_IRUSR|S_IWUSR);
> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDWR|O_CREAT|O_TRUNC,
> + S_IRUSR|S_IWUSR));
> if (ret < 0) {
> ret = -errno;
> goto out_err;
> @@ -123,7 +124,7 @@ int perf_data__open_dir(struct perf_data *data)
> if (!file->path)
> goto out_err;
>
> - ret = open(file->path, O_RDONLY);
> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDONLY));
> if (ret < 0)
> goto out_err;
>
> @@ -248,7 +249,7 @@ static int open_file_read(struct perf_data *data)
> int fd;
> char sbuf[STRERR_BUFSIZE];
>
> - fd = open(data->file.path, flags);
> + fd = TEMP_FAILURE_RETRY(open(data->file.path, flags));
> if (fd < 0) {
> int err = errno;
>
> @@ -288,8 +289,8 @@ static int open_file_write(struct perf_data *data)
> int fd;
> char sbuf[STRERR_BUFSIZE];
>
> - fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
> - S_IRUSR|S_IWUSR);
> + fd = TEMP_FAILURE_RETRY(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,
> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> index 5ac13958d1bd..40befa2429f9 100644
> --- a/tools/perf/util/dso.c
> +++ b/tools/perf/util/dso.c
> @@ -300,7 +300,7 @@ int filename__decompress(const char *name, char *pathname,
> * descriptor to the uncompressed file.
> */
> if (!compressions[comp].is_compressed(name))
> - return open(name, O_RDONLY);
> + return TEMP_FAILURE_RETRY(open(name, O_RDONLY));
>
> fd = mkstemp(tmpbuf);
> if (fd < 0) {
> @@ -476,7 +476,7 @@ static int do_open(char *name)
> char sbuf[STRERR_BUFSIZE];
>
> do {
> - fd = open(name, O_RDONLY|O_CLOEXEC);
> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY|O_CLOEXEC));
> if (fd >= 0)
> return fd;
>
> @@ -898,7 +898,7 @@ static ssize_t file_read(struct dso *dso, struct machine *machine,
> goto out;
> }
>
> - ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset);
> + ret = TEMP_FAILURE_RETRY(pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset));
> out:
> pthread_mutex_unlock(&dso__data_open_lock);
> return ret;
> diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> index 48af7d379d82..ba0dcdf91120 100644
> --- a/tools/perf/util/evlist.c
> +++ b/tools/perf/util/evlist.c
> @@ -1440,7 +1440,7 @@ int evlist__prepare_workload(struct evlist *evlist, struct target *target, const
> /*
> * Wait until the parent tells us to go.
> */
> - ret = read(go_pipe[0], &bf, 1);
> + ret = TEMP_FAILURE_RETRY(read(go_pipe[0], &bf, 1));
> /*
> * The parent will ask for the execvp() to be performed by
> * writing exactly one byte, in workload.cork_fd, usually via
> @@ -1837,7 +1837,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
> * O_RDWR avoids POLLHUPs which is necessary to allow the other
> * end of a FIFO to be repeatedly opened and closed.
> */
> - fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> + fd = TEMP_FAILURE_RETRY(open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> if (fd < 0) {
> pr_err("Failed to open '%s'\n", s);
> ret = -errno;
> @@ -1848,7 +1848,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
>
> if (p && *++p) {
> /* O_RDWR | O_NONBLOCK means the other end need not be open */
> - fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> + fd = TEMP_FAILURE_RETRY(open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> if (fd < 0) {
> pr_err("Failed to open '%s'\n", p);
> ret = -errno;
> diff --git a/tools/perf/util/lzma.c b/tools/perf/util/lzma.c
> index 51424cdc3b68..93e0a6a53e28 100644
> --- a/tools/perf/util/lzma.c
> +++ b/tools/perf/util/lzma.c
> @@ -109,7 +109,7 @@ int lzma_decompress_to_file(const char *input, int output_fd)
>
> bool lzma_is_compressed(const char *input)
> {
> - int fd = open(input, O_RDONLY);
> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> const uint8_t magic[6] = { 0xFD, '7', 'z', 'X', 'Z', 0x00 };
> char buf[6] = { 0 };
> ssize_t rc;
> @@ -117,7 +117,7 @@ bool lzma_is_compressed(const char *input)
> if (fd < 0)
> return -1;
>
> - rc = read(fd, buf, sizeof(buf));
> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> close(fd);
> return rc == sizeof(buf) ?
> memcmp(buf, magic, sizeof(buf)) == 0 : false;
> diff --git a/tools/perf/util/namespaces.c b/tools/perf/util/namespaces.c
> index dd536220cdb9..89edfba6af06 100644
> --- a/tools/perf/util/namespaces.c
> +++ b/tools/perf/util/namespaces.c
> @@ -9,6 +9,7 @@
> #include "get_current_dir_name.h"
> #include <sys/types.h>
> #include <sys/stat.h>
> +#include <errno.h>
> #include <fcntl.h>
> #include <limits.h>
> #include <sched.h>
> @@ -266,11 +267,11 @@ void nsinfo__mountns_enter(struct nsinfo *nsi,
> if (!oldcwd)
> return;
>
> - oldns = open(curpath, O_RDONLY);
> + oldns = TEMP_FAILURE_RETRY(open(curpath, O_RDONLY));
> if (oldns < 0)
> goto errout;
>
> - newns = open(nsi->mntns_path, O_RDONLY);
> + newns = TEMP_FAILURE_RETRY(open(nsi->mntns_path, O_RDONLY));
> if (newns < 0)
> goto errout;
>
> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> index 7ed235740431..75f8f2e44195 100644
> --- a/tools/perf/util/parse-events.c
> +++ b/tools/perf/util/parse-events.c
> @@ -189,7 +189,7 @@ static int tp_event_has_id(const char *dir_path, struct dirent *evt_dir)
> int fd;
>
> snprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path, evt_dir->d_name);
> - fd = open(evt_path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> if (fd < 0)
> return -EINVAL;
> close(fd);
> @@ -233,10 +233,10 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config)
>
> scnprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path,
> evt_dirent->d_name);
> - fd = open(evt_path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> if (fd < 0)
> continue;
> - if (read(fd, id_buf, sizeof(id_buf)) < 0) {
> + if (TEMP_FAILURE_RETRY(read(fd, id_buf, sizeof(id_buf))) < 0) {
> close(fd);
> continue;
> }
> diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> index 9a1c7e63e663..59252cfc2b5b 100644
> --- a/tools/perf/util/pmu.c
> +++ b/tools/perf/util/pmu.c
> @@ -154,14 +154,14 @@ static int perf_pmu__parse_scale(struct perf_pmu_alias *alias, char *dir, char *
>
> scnprintf(path, PATH_MAX, "%s/%s.scale", dir, name);
>
> - fd = open(path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> if (fd == -1)
> return -1;
>
> if (fstat(fd, &st) < 0)
> goto error;
>
> - sret = read(fd, scale, sizeof(scale)-1);
> + sret = TEMP_FAILURE_RETRY(read(fd, scale, sizeof(scale)-1));
> if (sret < 0)
> goto error;
>
> @@ -184,11 +184,11 @@ static int perf_pmu__parse_unit(struct perf_pmu_alias *alias, char *dir, char *n
>
> scnprintf(path, PATH_MAX, "%s/%s.unit", dir, name);
>
> - fd = open(path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> if (fd == -1)
> return -1;
>
> - sret = read(fd, alias->unit, UNIT_MAX_LEN);
> + sret = TEMP_FAILURE_RETRY(read(fd, alias->unit, UNIT_MAX_LEN));
> if (sret < 0)
> goto error;
>
> @@ -214,7 +214,7 @@ perf_pmu__parse_per_pkg(struct perf_pmu_alias *alias, char *dir, char *name)
>
> scnprintf(path, PATH_MAX, "%s/%s.per-pkg", dir, name);
>
> - fd = open(path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> if (fd == -1)
> return -1;
>
> @@ -232,7 +232,7 @@ static int perf_pmu__parse_snapshot(struct perf_pmu_alias *alias,
>
> scnprintf(path, PATH_MAX, "%s/%s.snapshot", dir, name);
>
> - fd = open(path, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> if (fd == -1)
> return -1;
>
> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> index 062b5cbe67af..7838889f55df 100644
> --- a/tools/perf/util/probe-event.c
> +++ b/tools/perf/util/probe-event.c
> @@ -276,7 +276,7 @@ static char *find_module_name(const char *module)
> char *mod_name = NULL;
> int name_offset;
>
> - fd = open(module, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(module, O_RDONLY));
> if (fd < 0)
> return NULL;
>
> @@ -598,7 +598,7 @@ static int get_text_start_address(const char *exec, u64 *address,
> struct nscookie nsc;
>
> nsinfo__mountns_enter(nsi, &nsc);
> - fd = open(exec, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(exec, O_RDONLY));
> nsinfo__mountns_exit(&nsc);
> if (fd < 0)
> return -errno;
> diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
> index 3d50de3217d5..b788932656e8 100644
> --- a/tools/perf/util/probe-file.c
> +++ b/tools/perf/util/probe-file.c
> @@ -115,9 +115,9 @@ int open_trace_file(const char *trace_file, bool readwrite)
> if (ret >= 0) {
> pr_debug("Opening %s write=%d\n", buf, readwrite);
> if (readwrite && !probe_event_dry_run)
> - ret = open(buf, O_RDWR | O_APPEND, 0);
> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDWR | O_APPEND, 0));
> else
> - ret = open(buf, O_RDONLY, 0);
> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDONLY, 0));
>
> if (ret < 0)
> ret = -errno;
> @@ -180,7 +180,7 @@ struct strlist *probe_file__get_rawlist(int fd)
> if (sl == NULL)
> return NULL;
>
> - fddup = dup(fd);
> + fddup = TEMP_FAILURE_RETRY(dup(fd));
> if (fddup < 0)
> goto out_free_sl;
>
> @@ -498,7 +498,7 @@ static int probe_cache__open(struct probe_cache *pcache, const char *target,
> }
>
> snprintf(cpath, PATH_MAX, "%s/probes", dir_name);
> - fd = open(cpath, O_CREAT | O_RDWR, 0644);
> + fd = TEMP_FAILURE_RETRY(open(cpath, O_CREAT | O_RDWR, 0644));
> if (fd < 0)
> pr_debug("Failed to open cache(%d): %s\n", fd, cpath);
> free(dir_name);
> @@ -514,7 +514,7 @@ static int probe_cache__load(struct probe_cache *pcache)
> int ret = 0, fddup;
> FILE *fp;
>
> - fddup = dup(pcache->fd);
> + fddup = TEMP_FAILURE_RETRY(dup(pcache->fd));
> if (fddup < 0)
> return -errno;
> fp = fdopen(fddup, "r");
> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> index 0acb9de54b06..b2fc389cc530 100644
> --- a/tools/perf/util/session.c
> +++ b/tools/perf/util/session.c
> @@ -384,7 +384,7 @@ static int skipn(int fd, off_t n)
> ssize_t ret;
>
> while (n > 0) {
> - ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
> + ret = TEMP_FAILURE_RETRY(read(fd, buf, min(n, (off_t)sizeof(buf))));
> if (ret <= 0)
> return ret;
> n -= ret;
> diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
> index ecd377938eea..02556070a2ce 100644
> --- a/tools/perf/util/symbol-elf.c
> +++ b/tools/perf/util/symbol-elf.c
> @@ -577,7 +577,7 @@ static int read_build_id(const char *filename, struct build_id *bid)
> if (size < BUILD_ID_SIZE)
> goto out;
>
> - fd = open(filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> if (fd < 0)
> goto out;
>
> @@ -638,7 +638,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> size_t size = sizeof(bid->data);
> int fd, err = -1;
>
> - fd = open(filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> if (fd < 0)
> goto out;
>
> @@ -647,24 +647,24 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> GElf_Nhdr nhdr;
> size_t namesz, descsz;
>
> - if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr))
> + if (TEMP_FAILURE_RETRY(read(fd, &nhdr, sizeof(nhdr))) != sizeof(nhdr))
> break;
>
> namesz = NOTE_ALIGN(nhdr.n_namesz);
> descsz = NOTE_ALIGN(nhdr.n_descsz);
> if (nhdr.n_type == NT_GNU_BUILD_ID &&
> nhdr.n_namesz == sizeof("GNU")) {
> - if (read(fd, bf, namesz) != (ssize_t)namesz)
> + if (TEMP_FAILURE_RETRY(read(fd, bf, namesz)) != (ssize_t)namesz)
> break;
> if (memcmp(bf, "GNU", sizeof("GNU")) == 0) {
> size_t sz = min(descsz, size);
> - if (read(fd, bid->data, sz) == (ssize_t)sz) {
> + if (TEMP_FAILURE_RETRY(read(fd, bid->data, sz)) == (ssize_t)sz) {
> memset(bid->data + sz, 0, size - sz);
> bid->size = sz;
> err = 0;
> break;
> }
> - } else if (read(fd, bf, descsz) != (ssize_t)descsz)
> + } else if (TEMP_FAILURE_RETRY(read(fd, bf, descsz)) != (ssize_t)descsz)
> break;
> } else {
> int n = namesz + descsz;
> @@ -674,7 +674,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
> __func__, filename, nhdr.n_namesz, nhdr.n_descsz);
> }
> - if (read(fd, bf, n) != n)
> + if (TEMP_FAILURE_RETRY(read(fd, bf, n)) != n)
> break;
> }
> }
> @@ -732,7 +732,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,
> Elf_Scn *sec;
> Elf_Kind ek;
>
> - fd = open(filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> if (fd < 0)
> goto out;
>
> @@ -844,7 +844,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
>
> type = dso->symtab_type;
> } else {
> - fd = open(name, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> if (fd < 0) {
> dso->load_errno = errno;
> return -1;
> @@ -1454,7 +1454,7 @@ static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len)
> if (len < n)
> n = len;
> /* Use read because mmap won't work on proc files */
> - r = read(from, buf, n);
> + r = TEMP_FAILURE_RETRY(read(from, buf, n));
> if (r < 0)
> goto out;
> if (!r)
> @@ -1485,7 +1485,7 @@ static int kcore__open(struct kcore *kcore, const char *filename)
> {
> GElf_Ehdr *ehdr;
>
> - kcore->fd = open(filename, O_RDONLY);
> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> if (kcore->fd == -1)
> return -1;
>
> @@ -1518,7 +1518,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,
> if (temp)
> kcore->fd = mkstemp(filename);
> else
> - kcore->fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400);
> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400));
> if (kcore->fd == -1)
> return -1;
>
> @@ -1966,7 +1966,7 @@ static int kcore_copy__compare_fds(int from, int to)
>
> while (1) {
> /* Use read because mmap won't work on proc files */
> - ret = read(from, buf_from, page_size);
> + ret = TEMP_FAILURE_RETRY(read(from, buf_from, page_size));
> if (ret < 0)
> goto out;
>
> @@ -1994,11 +1994,11 @@ static int kcore_copy__compare_files(const char *from_filename,
> {
> int from, to, err = -1;
>
> - from = open(from_filename, O_RDONLY);
> + from = TEMP_FAILURE_RETRY(open(from_filename, O_RDONLY));
> if (from < 0)
> return -1;
>
> - to = open(to_filename, O_RDONLY);
> + to = TEMP_FAILURE_RETRY(open(to_filename, O_RDONLY));
> if (to < 0)
> goto out_close_from;
>
> @@ -2419,7 +2419,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)
> Elf *elf;
> int fd, ret;
>
> - fd = open(target, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(target, O_RDONLY));
> if (fd < 0)
> return -EBADF;
>
> diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
> index f9eb0bee7f15..7e25d476b1b2 100644
> --- a/tools/perf/util/symbol-minimal.c
> +++ b/tools/perf/util/symbol-minimal.c
> @@ -230,7 +230,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> size_t buf_size;
> void *buf;
>
> - fd = open(filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> if (fd < 0)
> return -1;
>
> @@ -242,7 +242,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> if (buf == NULL)
> goto out;
>
> - if (read(fd, buf, buf_size) != (ssize_t) buf_size)
> + if (TEMP_FAILURE_RETRY(read(fd, buf, buf_size)) != (ssize_t) buf_size)
> goto out_free;
>
> ret = read_build_id(buf, buf_size, bid, false);
> @@ -256,7 +256,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> enum dso_binary_type type)
> {
> - int fd = open(name, O_RDONLY);
> + int fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> if (fd < 0)
> goto out_errno;
>
> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> index f72baf636724..5212667ea67d 100644
> --- a/tools/perf/util/symbol.c
> +++ b/tools/perf/util/symbol.c
> @@ -1335,7 +1335,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
> md.dso = dso;
> INIT_LIST_HEAD(&md.maps);
>
> - fd = open(kcore_filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(kcore_filename, O_RDONLY));
> if (fd < 0) {
> pr_debug("Failed to open %s. Note /proc/kcore requires CAP_SYS_RAWIO capability to access.\n",
> kcore_filename);
> @@ -2156,7 +2156,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
> */
> static bool filename__readable(const char *file)
> {
> - int fd = open(file, O_RDONLY);
> + int fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> if (fd < 0)
> return false;
> close(fd);
> diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
> index 76beda3e1a10..9dcc268bd69c 100644
> --- a/tools/perf/util/synthetic-events.c
> +++ b/tools/perf/util/synthetic-events.c
> @@ -40,6 +40,7 @@
> #include <api/io.h>
> #include <sys/types.h>
> #include <sys/stat.h>
> +#include <errno.h>
> #include <fcntl.h>
> #include <unistd.h>
>
> @@ -86,13 +87,13 @@ static int perf_event__get_comm_ids(pid_t pid, pid_t tid, char *comm, size_t len
> else
> snprintf(bf, sizeof(bf), "/proc/%d/status", tid);
>
> - fd = open(bf, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY));
> if (fd < 0) {
> pr_debug("couldn't open %s\n", bf);
> return -1;
> }
>
> - n = read(fd, bf, sizeof(bf) - 1);
> + n = TEMP_FAILURE_RETRY(read(fd, bf, sizeof(bf) - 1));
> close(fd);
> if (n <= 0) {
> pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
> @@ -410,7 +411,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
> snprintf(bf, sizeof(bf), "%s/proc/%d/task/%d/maps",
> machine->root_dir, pid, pid);
>
> - io.fd = open(bf, O_RDONLY, 0);
> + io.fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY, 0));
> if (io.fd < 0) {
> /*
> * We raced with a task exiting - just return:
> diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c
> index a65f65d0857e..c917a4217c6f 100644
> --- a/tools/perf/util/trace-event-info.c
> +++ b/tools/perf/util/trace-event-info.c
> @@ -48,7 +48,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> int r, fd;
> int err = -EIO;
>
> - fd = open(file, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> if (fd < 0) {
> pr_debug("Can't read '%s'", file);
> return -errno;
> @@ -61,7 +61,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> }
>
> do {
> - r = read(fd, buf, BUFSIZ);
> + r = TEMP_FAILURE_RETRY(read(fd, buf, BUFSIZ));
> if (r > 0) {
> size += r;
> if (write(output_fd, buf, r) != r)
> @@ -519,7 +519,7 @@ struct tracing_data *tracing_data_get(struct list_head *pattrs,
> return NULL;
> }
>
> - temp_fd = open(tdata->temp_file, O_RDWR);
> + temp_fd = TEMP_FAILURE_RETRY(open(tdata->temp_file, O_RDWR));
> if (temp_fd < 0) {
> pr_debug("Can't read '%s'", tdata->temp_file);
> free(tdata);
> diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c
> index 8a01af783310..7086770ec225 100644
> --- a/tools/perf/util/trace-event-read.c
> +++ b/tools/perf/util/trace-event-read.c
> @@ -11,6 +11,7 @@
> #include <sys/stat.h>
> #include <sys/wait.h>
> #include <sys/mman.h>
> +#include <errno.h>
> #include <fcntl.h>
> #include <unistd.h>
> #include <errno.h>
> @@ -28,7 +29,7 @@ static int __do_read(int fd, void *buf, int size)
> int rsize = size;
>
> while (size) {
> - int ret = read(fd, buf, size);
> + int ret = TEMP_FAILURE_RETRY(read(fd, buf, size));
>
> if (ret <= 0)
> return -1;
> @@ -105,7 +106,7 @@ static char *read_string(void)
> char c;
>
> for (;;) {
> - r = read(input_fd, &c, 1);
> + r = TEMP_FAILURE_RETRY(read(input_fd, &c, 1));
> if (r < 0) {
> pr_debug("reading input file");
> goto out;
> diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c
> index 37622699c91a..2ece387a73c6 100644
> --- a/tools/perf/util/unwind-libunwind-local.c
> +++ b/tools/perf/util/unwind-libunwind-local.c
> @@ -362,7 +362,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> }
>
> if (ofs <= 0) {
> - fd = open(dso->symsrc_filename, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(dso->symsrc_filename, O_RDONLY));
> if (fd >= 0) {
> ofs = elf_section_offset(fd, ".debug_frame");
> close(fd);
> @@ -377,7 +377,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> dso, DSO_BINARY_TYPE__DEBUGLINK,
> machine->root_dir, debuglink, PATH_MAX);
> if (!ret) {
> - fd = open(debuglink, O_RDONLY);
> + fd = TEMP_FAILURE_RETRY(open(debuglink, O_RDONLY));
> if (fd >= 0) {
> ofs = elf_section_offset(fd,
> ".debug_frame");
> diff --git a/tools/perf/util/zlib.c b/tools/perf/util/zlib.c
> index 78d2297c1b67..f4563ea094c8 100644
> --- a/tools/perf/util/zlib.c
> +++ b/tools/perf/util/zlib.c
> @@ -1,4 +1,5 @@
> // SPDX-License-Identifier: GPL-2.0
> +#include <errno.h>
> #include <fcntl.h>
> #include <stdio.h>
> #include <string.h>
> @@ -82,7 +83,7 @@ int gzip_decompress_to_file(const char *input, int output_fd)
>
> bool gzip_is_compressed(const char *input)
> {
> - int fd = open(input, O_RDONLY);
> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> const uint8_t magic[2] = { 0x1f, 0x8b };
> char buf[2] = { 0 };
> ssize_t rc;
> @@ -90,7 +91,7 @@ bool gzip_is_compressed(const char *input)
> if (fd < 0)
> return -1;
>
> - rc = read(fd, buf, sizeof(buf));
> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> close(fd);
> return rc == sizeof(buf) ?
> memcmp(buf, magic, sizeof(buf)) == 0 : false;
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 6:50 ` Adrian Hunter
@ 2022-06-15 15:34 ` Ian Rogers
2022-06-15 16:40 ` Adrian Hunter
0 siblings, 1 reply; 7+ messages in thread
From: Ian Rogers @ 2022-06-15 15:34 UTC (permalink / raw)
To: Adrian Hunter
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Alexey Bayduraev, Riccardo Mancini, Dave Marchevsky, Song Liu,
Leo Yan, John Garry, Florian Fischer, Andi Kleen, Kan Liang,
Masami Hiramatsu, Zechuan Chen, Yang Jihong, German Gomez,
Thomas Richter, James Clark, Denis Nikitin, Lexi Shao,
Michael Petlan, Remi Bernon, Fangrui Song, linux-kernel,
linux-perf-users, Stephane Eranian
On Tue, Jun 14, 2022 at 11:51 PM Adrian Hunter <adrian.hunter@intel.com> wrote:
>
> On 15/06/22 09:03, Ian Rogers wrote:
> > Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
> > causes the syscall to need to restart.
>
> Please expand on why you want this change. Can you give an example
> of where it makes a difference?
As we can get signals pretty much anywhere in the code many syscalls
are robust to this:
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n311
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n372
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n1585
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n2502
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-top.c?h=perf/core#n707
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/copyfile.c?h=perf/core#n59
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/evlist.c?h=perf/core#n1966
But as this change shows there are a large number that are not. They
will get EINTR and turn it into a failure which really isn't one.
Making an example of this is near impossible as you need to arrange
the signal delivery very carefully. I am however motivated to do this
as we're seeing missing events when running perf in fleetwide
profiling and I wish to discount EINTR as a cause.
> It looks pretty ugly so maybe making replacement functions or macros
> is better e.g.
>
> #define open_retry(a, ...) \
> ({ long int __result; \
> do __result = (long int) open((a), ##__VA_ARGS__); \
> while (__result == -1L && errno == EINTR); \
> __result; })
As open is a user land thing there is no Linux precedent to follow
here. TEMP_FAILURE_RETRY is the general norm, and I'd like to see it
used when a raw system call is made. ie, if you want open_retry:
#define open_retry(a, ...) TEMP_FAILURE_RETRY(open((a), ##__VA_ARGS__))
I don't think we need to be in the game of reinventing
TEMP_FAILURE_RETRY. I think we can follow on from this change with
changes not to use raw open and read calls, instead preferring some of
our existing EINTR robust reads like readn, etc. Doing that is beyond
the scope of what this change is trying to do.
Thanks,
Ian
> >
> > Signed-off-by: Ian Rogers <irogers@google.com>
> > ---
> > tools/lib/api/io.h | 3 ++-
> > tools/lib/symbol/kallsyms.c | 4 ++-
> > tools/perf/builtin-daemon.c | 19 +++++++-------
> > tools/perf/builtin-ftrace.c | 18 ++++++-------
> > tools/perf/builtin-inject.c | 2 +-
> > tools/perf/builtin-kvm.c | 2 +-
> > tools/perf/builtin-record.c | 6 ++---
> > tools/perf/builtin-sched.c | 2 +-
> > tools/perf/builtin-script.c | 2 +-
> > tools/perf/util/copyfile.c | 2 +-
> > tools/perf/util/data.c | 11 ++++----
> > tools/perf/util/dso.c | 6 ++---
> > tools/perf/util/evlist.c | 6 ++---
> > tools/perf/util/lzma.c | 4 +--
> > tools/perf/util/namespaces.c | 5 ++--
> > tools/perf/util/parse-events.c | 6 ++---
> > tools/perf/util/pmu.c | 12 ++++-----
> > tools/perf/util/probe-event.c | 4 +--
> > tools/perf/util/probe-file.c | 10 ++++----
> > tools/perf/util/session.c | 2 +-
> > tools/perf/util/symbol-elf.c | 32 ++++++++++++------------
> > tools/perf/util/symbol-minimal.c | 6 ++---
> > tools/perf/util/symbol.c | 4 +--
> > tools/perf/util/synthetic-events.c | 7 +++---
> > tools/perf/util/trace-event-info.c | 6 ++---
> > tools/perf/util/trace-event-read.c | 5 ++--
> > tools/perf/util/unwind-libunwind-local.c | 4 +--
> > tools/perf/util/zlib.c | 5 ++--
> > 28 files changed, 102 insertions(+), 93 deletions(-)
> >
> > diff --git a/tools/lib/api/io.h b/tools/lib/api/io.h
> > index 777c20f6b604..b292bbd39fb0 100644
> > --- a/tools/lib/api/io.h
> > +++ b/tools/lib/api/io.h
> > @@ -7,6 +7,7 @@
> > #ifndef __API_IO__
> > #define __API_IO__
> >
> > +#include <errno.h>
> > #include <stdlib.h>
> > #include <unistd.h>
> >
> > @@ -45,7 +46,7 @@ static inline int io__get_char(struct io *io)
> > return -1;
> >
> > if (ptr == io->end) {
> > - ssize_t n = read(io->fd, io->buf, io->buf_len);
> > + ssize_t n = TEMP_FAILURE_RETRY(read(io->fd, io->buf, io->buf_len));
> >
> > if (n <= 0) {
> > io->eof = true;
> > diff --git a/tools/lib/symbol/kallsyms.c b/tools/lib/symbol/kallsyms.c
> > index e335ac2b9e19..5436c038c379 100644
> > --- a/tools/lib/symbol/kallsyms.c
> > +++ b/tools/lib/symbol/kallsyms.c
> > @@ -3,7 +3,9 @@
> > #include "api/io.h"
> > #include <stdio.h>
> > #include <sys/stat.h>
> > +#include <errno.h>
> > #include <fcntl.h>
> > +#include <unistd.h>
> >
> > u8 kallsyms2elf_type(char type)
> > {
> > @@ -36,7 +38,7 @@ int kallsyms__parse(const char *filename, void *arg,
> > char bf[BUFSIZ];
> > int err;
> >
> > - io.fd = open(filename, O_RDONLY, 0);
> > + io.fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY, 0));
> >
> > if (io.fd < 0)
> > return -1;
> > diff --git a/tools/perf/builtin-daemon.c b/tools/perf/builtin-daemon.c
> > index 6cb3f6cc36d0..d4e1a5636056 100644
> > --- a/tools/perf/builtin-daemon.c
> > +++ b/tools/perf/builtin-daemon.c
> > @@ -351,7 +351,7 @@ static int daemon_session__run(struct daemon_session *session,
> > return -1;
> > }
> >
> > - fd = open("/dev/null", O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open("/dev/null", O_RDONLY));
> > if (fd < 0) {
> > perror("failed: open /dev/null");
> > return -1;
> > @@ -360,7 +360,7 @@ static int daemon_session__run(struct daemon_session *session,
> > dup2(fd, 0);
> > close(fd);
> >
> > - fd = open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644);
> > + fd = TEMP_FAILURE_RETRY(open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644));
> > if (fd < 0) {
> > perror("failed: open session output");
> > return -1;
> > @@ -405,7 +405,8 @@ static pid_t handle_signalfd(struct daemon *daemon)
> > * coalesced in kernel and we can receive only single signal even
> > * if multiple SIGCHLD were generated.
> > */
> > - err = read(daemon->signal_fd, &si, sizeof(struct signalfd_siginfo));
> > + err = TEMP_FAILURE_RETRY(read(daemon->signal_fd, &si,
> > + sizeof(struct signalfd_siginfo)));
> > if (err != sizeof(struct signalfd_siginfo)) {
> > pr_err("failed to read signal fd\n");
> > return -1;
> > @@ -522,7 +523,7 @@ static int daemon_session__control(struct daemon_session *session,
> > scnprintf(control_path, sizeof(control_path), "%s/%s",
> > session->base, SESSION_CONTROL);
> >
> > - control = open(control_path, O_WRONLY|O_NONBLOCK);
> > + control = TEMP_FAILURE_RETRY(open(control_path, O_WRONLY|O_NONBLOCK));
> > if (!control)
> > return -1;
> >
> > @@ -531,7 +532,7 @@ static int daemon_session__control(struct daemon_session *session,
> > scnprintf(ack_path, sizeof(ack_path), "%s/%s",
> > session->base, SESSION_ACK);
> >
> > - ack = open(ack_path, O_RDONLY, O_NONBLOCK);
> > + ack = TEMP_FAILURE_RETRY(open(ack_path, O_RDONLY, O_NONBLOCK));
> > if (!ack) {
> > close(control);
> > return -1;
> > @@ -564,7 +565,7 @@ static int daemon_session__control(struct daemon_session *session,
> > goto out;
> > }
> >
> > - err = read(ack, buf, sizeof(buf));
> > + err = TEMP_FAILURE_RETRY(read(ack, buf, sizeof(buf)));
> > if (err > 0)
> > ret = strcmp(buf, "ack\n");
> > else
> > @@ -1078,7 +1079,7 @@ static int handle_config_changes(struct daemon *daemon, int conf_fd,
> > ssize_t len;
> >
> > while (!(*config_changed)) {
> > - len = read(conf_fd, buf, sizeof(buf));
> > + len = TEMP_FAILURE_RETRY(read(conf_fd, buf, sizeof(buf)));
> > if (len == -1) {
> > if (errno != EAGAIN) {
> > perror("failed: read");
> > @@ -1147,7 +1148,7 @@ static int check_lock(struct daemon *daemon)
> >
> > scnprintf(path, sizeof(path), "%s/lock", daemon->base);
> >
> > - fd = open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640);
> > + fd = TEMP_FAILURE_RETRY(open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640));
> > if (fd < 0)
> > return -1;
> >
> > @@ -1201,7 +1202,7 @@ static int go_background(struct daemon *daemon)
> > return -1;
> > }
> >
> > - fd = open("output", O_RDWR|O_CREAT|O_TRUNC, 0644);
> > + fd = TEMP_FAILURE_RETRY(open("output", O_RDWR|O_CREAT|O_TRUNC, 0644));
> > if (fd < 0) {
> > perror("failed: open");
> > return -1;
> > diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
> > index 7de07bb16d23..7553e20f38b2 100644
> > --- a/tools/perf/builtin-ftrace.c
> > +++ b/tools/perf/builtin-ftrace.c
> > @@ -79,7 +79,7 @@ static int __write_tracing_file(const char *name, const char *val, bool append)
> > else
> > flags |= O_TRUNC;
> >
> > - fd = open(file, flags);
> > + fd = TEMP_FAILURE_RETRY(open(file, flags));
> > if (fd < 0) {
> > pr_debug("cannot open tracing file: %s: %s\n",
> > name, str_error_r(errno, errbuf, sizeof(errbuf)));
> > @@ -132,7 +132,7 @@ static int read_tracing_file_to_stdout(const char *name)
> > return -1;
> > }
> >
> > - fd = open(file, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> > if (fd < 0) {
> > pr_debug("cannot open tracing file: %s: %s\n",
> > name, str_error_r(errno, buf, sizeof(buf)));
> > @@ -141,7 +141,7 @@ static int read_tracing_file_to_stdout(const char *name)
> >
> > /* read contents to stdout */
> > while (true) {
> > - int n = read(fd, buf, sizeof(buf));
> > + int n = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> > if (n == 0)
> > break;
> > else if (n < 0)
> > @@ -608,7 +608,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> > goto out_reset;
> > }
> >
> > - trace_fd = open(trace_file, O_RDONLY);
> > + trace_fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
> >
> > put_tracing_file(trace_file);
> >
> > @@ -645,7 +645,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> > break;
> >
> > if (pollfd.revents & POLLIN) {
> > - int n = read(trace_fd, buf, sizeof(buf));
> > + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> > if (n < 0)
> > break;
> > if (fwrite(buf, n, 1, stdout) != 1)
> > @@ -665,7 +665,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> >
> > /* read remaining buffer contents */
> > while (true) {
> > - int n = read(trace_fd, buf, sizeof(buf));
> > + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> > if (n <= 0)
> > break;
> > if (fwrite(buf, n, 1, stdout) != 1)
> > @@ -826,7 +826,7 @@ static int prepare_func_latency(struct perf_ftrace *ftrace)
> > return -1;
> > }
> >
> > - fd = open(trace_file, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
> > if (fd < 0)
> > pr_err("failed to open trace_pipe\n");
> >
> > @@ -913,7 +913,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
> > break;
> >
> > if (pollfd.revents & POLLIN) {
> > - int n = read(trace_fd, buf, sizeof(buf) - 1);
> > + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> > if (n < 0)
> > break;
> >
> > @@ -931,7 +931,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
> >
> > /* read remaining buffer contents */
> > while (!ftrace->target.use_bpf) {
> > - int n = read(trace_fd, buf, sizeof(buf) - 1);
> > + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> > if (n <= 0)
> > break;
> > make_histogram(buckets, buf, n, line, ftrace->use_nsec);
> > diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
> > index a75bf11585b5..960b4a50583e 100644
> > --- a/tools/perf/builtin-inject.c
> > +++ b/tools/perf/builtin-inject.c
> > @@ -155,7 +155,7 @@ static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
> > int ret;
> >
> > while (size > 0) {
> > - ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
> > + ssz = TEMP_FAILURE_RETRY(read(fd, buf, min(size, (off_t)sizeof(buf))));
> > if (ssz < 0)
> > return -errno;
> > ret = output_bytes(inject, buf, ssz);
> > diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
> > index 3696ae97f149..4dd8292d1941 100644
> > --- a/tools/perf/builtin-kvm.c
> > +++ b/tools/perf/builtin-kvm.c
> > @@ -880,7 +880,7 @@ static int perf_kvm__handle_timerfd(struct perf_kvm_stat *kvm)
> > uint64_t c;
> > int rc;
> >
> > - rc = read(kvm->timerfd, &c, sizeof(uint64_t));
> > + rc = TEMP_FAILURE_RETRY(read(kvm->timerfd, &c, sizeof(uint64_t)));
> > if (rc < 0) {
> > if (errno == EAGAIN)
> > return 0;
> > diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> > index 9a71f0330137..361610bc96b6 100644
> > --- a/tools/perf/builtin-record.c
> > +++ b/tools/perf/builtin-record.c
> > @@ -902,7 +902,7 @@ static bool record__kcore_readable(struct machine *machine)
> >
> > scnprintf(kcore, sizeof(kcore), "%s/proc/kcore", machine->root_dir);
> >
> > - fd = open(kcore, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(kcore, O_RDONLY));
> > if (fd < 0)
> > return false;
> >
> > @@ -2049,7 +2049,7 @@ static int record__terminate_thread(struct record_thread *thread_data)
> >
> > close(thread_data->pipes.msg[1]);
> > thread_data->pipes.msg[1] = -1;
> > - err = read(thread_data->pipes.ack[0], &ack, sizeof(ack));
> > + err = TEMP_FAILURE_RETRY(read(thread_data->pipes.ack[0], &ack, sizeof(ack)));
> > if (err > 0)
> > pr_debug2("threads[%d]: sent %s\n", tid, thread_msg_tags[ack]);
> > else
> > @@ -2097,7 +2097,7 @@ static int record__start_threads(struct record *rec)
> > goto out_err;
> > }
> >
> > - err = read(thread_data[t].pipes.ack[0], &msg, sizeof(msg));
> > + err = TEMP_FAILURE_RETRY(read(thread_data[t].pipes.ack[0], &msg, sizeof(msg)));
> > if (err > 0)
> > pr_debug2("threads[%d]: sent %s\n", rec->thread_data[t].tid,
> > thread_msg_tags[msg]);
> > diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
> > index 646bd938927a..149787268d02 100644
> > --- a/tools/perf/builtin-sched.c
> > +++ b/tools/perf/builtin-sched.c
> > @@ -604,7 +604,7 @@ static u64 get_cpu_usage_nsec_self(int fd)
> > u64 runtime;
> > int ret;
> >
> > - ret = read(fd, &runtime, sizeof(runtime));
> > + ret = TEMP_FAILURE_RETRY(read(fd, &runtime, sizeof(runtime)));
> > BUG_ON(ret != sizeof(runtime));
> >
> > return runtime;
> > diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> > index c689054002cc..844f5350bfdf 100644
> > --- a/tools/perf/builtin-script.c
> > +++ b/tools/perf/builtin-script.c
> > @@ -4207,7 +4207,7 @@ int cmd_script(int argc, const char **argv)
> > goto out_delete;
> > }
> >
> > - input = open(data.path, O_RDONLY); /* input_name */
> > + input = TEMP_FAILURE_RETRY(open(data.path, O_RDONLY)); /* input_name */
> > if (input < 0) {
> > err = -errno;
> > perror("failed to open file");
> > diff --git a/tools/perf/util/copyfile.c b/tools/perf/util/copyfile.c
> > index 47e03de7c235..c7bff4943936 100644
> > --- a/tools/perf/util/copyfile.c
> > +++ b/tools/perf/util/copyfile.c
> > @@ -112,7 +112,7 @@ static int copyfile_mode_ns(const char *from, const char *to, mode_t mode,
> > goto out_close_to;
> >
> > nsinfo__mountns_enter(nsi, &nsc);
> > - fromfd = open(from, O_RDONLY);
> > + fromfd = TEMP_FAILURE_RETRY(open(from, O_RDONLY));
> > nsinfo__mountns_exit(&nsc);
> > if (fromfd < 0)
> > goto out_close_to;
> > diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c
> > index caabeac24c69..a9066a3d914f 100644
> > --- a/tools/perf/util/data.c
> > +++ b/tools/perf/util/data.c
> > @@ -53,7 +53,8 @@ int perf_data__create_dir(struct perf_data *data, int nr)
> > goto out_err;
> > }
> >
> > - ret = open(file->path, O_RDWR|O_CREAT|O_TRUNC, S_IRUSR|S_IWUSR);
> > + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDWR|O_CREAT|O_TRUNC,
> > + S_IRUSR|S_IWUSR));
> > if (ret < 0) {
> > ret = -errno;
> > goto out_err;
> > @@ -123,7 +124,7 @@ int perf_data__open_dir(struct perf_data *data)
> > if (!file->path)
> > goto out_err;
> >
> > - ret = open(file->path, O_RDONLY);
> > + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDONLY));
> > if (ret < 0)
> > goto out_err;
> >
> > @@ -248,7 +249,7 @@ static int open_file_read(struct perf_data *data)
> > int fd;
> > char sbuf[STRERR_BUFSIZE];
> >
> > - fd = open(data->file.path, flags);
> > + fd = TEMP_FAILURE_RETRY(open(data->file.path, flags));
> > if (fd < 0) {
> > int err = errno;
> >
> > @@ -288,8 +289,8 @@ static int open_file_write(struct perf_data *data)
> > int fd;
> > char sbuf[STRERR_BUFSIZE];
> >
> > - fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
> > - S_IRUSR|S_IWUSR);
> > + fd = TEMP_FAILURE_RETRY(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,
> > diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> > index 5ac13958d1bd..40befa2429f9 100644
> > --- a/tools/perf/util/dso.c
> > +++ b/tools/perf/util/dso.c
> > @@ -300,7 +300,7 @@ int filename__decompress(const char *name, char *pathname,
> > * descriptor to the uncompressed file.
> > */
> > if (!compressions[comp].is_compressed(name))
> > - return open(name, O_RDONLY);
> > + return TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> >
> > fd = mkstemp(tmpbuf);
> > if (fd < 0) {
> > @@ -476,7 +476,7 @@ static int do_open(char *name)
> > char sbuf[STRERR_BUFSIZE];
> >
> > do {
> > - fd = open(name, O_RDONLY|O_CLOEXEC);
> > + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY|O_CLOEXEC));
> > if (fd >= 0)
> > return fd;
> >
> > @@ -898,7 +898,7 @@ static ssize_t file_read(struct dso *dso, struct machine *machine,
> > goto out;
> > }
> >
> > - ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset);
> > + ret = TEMP_FAILURE_RETRY(pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset));
> > out:
> > pthread_mutex_unlock(&dso__data_open_lock);
> > return ret;
> > diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> > index 48af7d379d82..ba0dcdf91120 100644
> > --- a/tools/perf/util/evlist.c
> > +++ b/tools/perf/util/evlist.c
> > @@ -1440,7 +1440,7 @@ int evlist__prepare_workload(struct evlist *evlist, struct target *target, const
> > /*
> > * Wait until the parent tells us to go.
> > */
> > - ret = read(go_pipe[0], &bf, 1);
> > + ret = TEMP_FAILURE_RETRY(read(go_pipe[0], &bf, 1));
> > /*
> > * The parent will ask for the execvp() to be performed by
> > * writing exactly one byte, in workload.cork_fd, usually via
> > @@ -1837,7 +1837,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
> > * O_RDWR avoids POLLHUPs which is necessary to allow the other
> > * end of a FIFO to be repeatedly opened and closed.
> > */
> > - fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> > + fd = TEMP_FAILURE_RETRY(open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> > if (fd < 0) {
> > pr_err("Failed to open '%s'\n", s);
> > ret = -errno;
> > @@ -1848,7 +1848,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
> >
> > if (p && *++p) {
> > /* O_RDWR | O_NONBLOCK means the other end need not be open */
> > - fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> > + fd = TEMP_FAILURE_RETRY(open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> > if (fd < 0) {
> > pr_err("Failed to open '%s'\n", p);
> > ret = -errno;
> > diff --git a/tools/perf/util/lzma.c b/tools/perf/util/lzma.c
> > index 51424cdc3b68..93e0a6a53e28 100644
> > --- a/tools/perf/util/lzma.c
> > +++ b/tools/perf/util/lzma.c
> > @@ -109,7 +109,7 @@ int lzma_decompress_to_file(const char *input, int output_fd)
> >
> > bool lzma_is_compressed(const char *input)
> > {
> > - int fd = open(input, O_RDONLY);
> > + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> > const uint8_t magic[6] = { 0xFD, '7', 'z', 'X', 'Z', 0x00 };
> > char buf[6] = { 0 };
> > ssize_t rc;
> > @@ -117,7 +117,7 @@ bool lzma_is_compressed(const char *input)
> > if (fd < 0)
> > return -1;
> >
> > - rc = read(fd, buf, sizeof(buf));
> > + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> > close(fd);
> > return rc == sizeof(buf) ?
> > memcmp(buf, magic, sizeof(buf)) == 0 : false;
> > diff --git a/tools/perf/util/namespaces.c b/tools/perf/util/namespaces.c
> > index dd536220cdb9..89edfba6af06 100644
> > --- a/tools/perf/util/namespaces.c
> > +++ b/tools/perf/util/namespaces.c
> > @@ -9,6 +9,7 @@
> > #include "get_current_dir_name.h"
> > #include <sys/types.h>
> > #include <sys/stat.h>
> > +#include <errno.h>
> > #include <fcntl.h>
> > #include <limits.h>
> > #include <sched.h>
> > @@ -266,11 +267,11 @@ void nsinfo__mountns_enter(struct nsinfo *nsi,
> > if (!oldcwd)
> > return;
> >
> > - oldns = open(curpath, O_RDONLY);
> > + oldns = TEMP_FAILURE_RETRY(open(curpath, O_RDONLY));
> > if (oldns < 0)
> > goto errout;
> >
> > - newns = open(nsi->mntns_path, O_RDONLY);
> > + newns = TEMP_FAILURE_RETRY(open(nsi->mntns_path, O_RDONLY));
> > if (newns < 0)
> > goto errout;
> >
> > diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> > index 7ed235740431..75f8f2e44195 100644
> > --- a/tools/perf/util/parse-events.c
> > +++ b/tools/perf/util/parse-events.c
> > @@ -189,7 +189,7 @@ static int tp_event_has_id(const char *dir_path, struct dirent *evt_dir)
> > int fd;
> >
> > snprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path, evt_dir->d_name);
> > - fd = open(evt_path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> > if (fd < 0)
> > return -EINVAL;
> > close(fd);
> > @@ -233,10 +233,10 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config)
> >
> > scnprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path,
> > evt_dirent->d_name);
> > - fd = open(evt_path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> > if (fd < 0)
> > continue;
> > - if (read(fd, id_buf, sizeof(id_buf)) < 0) {
> > + if (TEMP_FAILURE_RETRY(read(fd, id_buf, sizeof(id_buf))) < 0) {
> > close(fd);
> > continue;
> > }
> > diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> > index 9a1c7e63e663..59252cfc2b5b 100644
> > --- a/tools/perf/util/pmu.c
> > +++ b/tools/perf/util/pmu.c
> > @@ -154,14 +154,14 @@ static int perf_pmu__parse_scale(struct perf_pmu_alias *alias, char *dir, char *
> >
> > scnprintf(path, PATH_MAX, "%s/%s.scale", dir, name);
> >
> > - fd = open(path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> > if (fd == -1)
> > return -1;
> >
> > if (fstat(fd, &st) < 0)
> > goto error;
> >
> > - sret = read(fd, scale, sizeof(scale)-1);
> > + sret = TEMP_FAILURE_RETRY(read(fd, scale, sizeof(scale)-1));
> > if (sret < 0)
> > goto error;
> >
> > @@ -184,11 +184,11 @@ static int perf_pmu__parse_unit(struct perf_pmu_alias *alias, char *dir, char *n
> >
> > scnprintf(path, PATH_MAX, "%s/%s.unit", dir, name);
> >
> > - fd = open(path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> > if (fd == -1)
> > return -1;
> >
> > - sret = read(fd, alias->unit, UNIT_MAX_LEN);
> > + sret = TEMP_FAILURE_RETRY(read(fd, alias->unit, UNIT_MAX_LEN));
> > if (sret < 0)
> > goto error;
> >
> > @@ -214,7 +214,7 @@ perf_pmu__parse_per_pkg(struct perf_pmu_alias *alias, char *dir, char *name)
> >
> > scnprintf(path, PATH_MAX, "%s/%s.per-pkg", dir, name);
> >
> > - fd = open(path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> > if (fd == -1)
> > return -1;
> >
> > @@ -232,7 +232,7 @@ static int perf_pmu__parse_snapshot(struct perf_pmu_alias *alias,
> >
> > scnprintf(path, PATH_MAX, "%s/%s.snapshot", dir, name);
> >
> > - fd = open(path, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> > if (fd == -1)
> > return -1;
> >
> > diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> > index 062b5cbe67af..7838889f55df 100644
> > --- a/tools/perf/util/probe-event.c
> > +++ b/tools/perf/util/probe-event.c
> > @@ -276,7 +276,7 @@ static char *find_module_name(const char *module)
> > char *mod_name = NULL;
> > int name_offset;
> >
> > - fd = open(module, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(module, O_RDONLY));
> > if (fd < 0)
> > return NULL;
> >
> > @@ -598,7 +598,7 @@ static int get_text_start_address(const char *exec, u64 *address,
> > struct nscookie nsc;
> >
> > nsinfo__mountns_enter(nsi, &nsc);
> > - fd = open(exec, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(exec, O_RDONLY));
> > nsinfo__mountns_exit(&nsc);
> > if (fd < 0)
> > return -errno;
> > diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
> > index 3d50de3217d5..b788932656e8 100644
> > --- a/tools/perf/util/probe-file.c
> > +++ b/tools/perf/util/probe-file.c
> > @@ -115,9 +115,9 @@ int open_trace_file(const char *trace_file, bool readwrite)
> > if (ret >= 0) {
> > pr_debug("Opening %s write=%d\n", buf, readwrite);
> > if (readwrite && !probe_event_dry_run)
> > - ret = open(buf, O_RDWR | O_APPEND, 0);
> > + ret = TEMP_FAILURE_RETRY(open(buf, O_RDWR | O_APPEND, 0));
> > else
> > - ret = open(buf, O_RDONLY, 0);
> > + ret = TEMP_FAILURE_RETRY(open(buf, O_RDONLY, 0));
> >
> > if (ret < 0)
> > ret = -errno;
> > @@ -180,7 +180,7 @@ struct strlist *probe_file__get_rawlist(int fd)
> > if (sl == NULL)
> > return NULL;
> >
> > - fddup = dup(fd);
> > + fddup = TEMP_FAILURE_RETRY(dup(fd));
> > if (fddup < 0)
> > goto out_free_sl;
> >
> > @@ -498,7 +498,7 @@ static int probe_cache__open(struct probe_cache *pcache, const char *target,
> > }
> >
> > snprintf(cpath, PATH_MAX, "%s/probes", dir_name);
> > - fd = open(cpath, O_CREAT | O_RDWR, 0644);
> > + fd = TEMP_FAILURE_RETRY(open(cpath, O_CREAT | O_RDWR, 0644));
> > if (fd < 0)
> > pr_debug("Failed to open cache(%d): %s\n", fd, cpath);
> > free(dir_name);
> > @@ -514,7 +514,7 @@ static int probe_cache__load(struct probe_cache *pcache)
> > int ret = 0, fddup;
> > FILE *fp;
> >
> > - fddup = dup(pcache->fd);
> > + fddup = TEMP_FAILURE_RETRY(dup(pcache->fd));
> > if (fddup < 0)
> > return -errno;
> > fp = fdopen(fddup, "r");
> > diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> > index 0acb9de54b06..b2fc389cc530 100644
> > --- a/tools/perf/util/session.c
> > +++ b/tools/perf/util/session.c
> > @@ -384,7 +384,7 @@ static int skipn(int fd, off_t n)
> > ssize_t ret;
> >
> > while (n > 0) {
> > - ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
> > + ret = TEMP_FAILURE_RETRY(read(fd, buf, min(n, (off_t)sizeof(buf))));
> > if (ret <= 0)
> > return ret;
> > n -= ret;
> > diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
> > index ecd377938eea..02556070a2ce 100644
> > --- a/tools/perf/util/symbol-elf.c
> > +++ b/tools/perf/util/symbol-elf.c
> > @@ -577,7 +577,7 @@ static int read_build_id(const char *filename, struct build_id *bid)
> > if (size < BUILD_ID_SIZE)
> > goto out;
> >
> > - fd = open(filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> > if (fd < 0)
> > goto out;
> >
> > @@ -638,7 +638,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > size_t size = sizeof(bid->data);
> > int fd, err = -1;
> >
> > - fd = open(filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> > if (fd < 0)
> > goto out;
> >
> > @@ -647,24 +647,24 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > GElf_Nhdr nhdr;
> > size_t namesz, descsz;
> >
> > - if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr))
> > + if (TEMP_FAILURE_RETRY(read(fd, &nhdr, sizeof(nhdr))) != sizeof(nhdr))
> > break;
> >
> > namesz = NOTE_ALIGN(nhdr.n_namesz);
> > descsz = NOTE_ALIGN(nhdr.n_descsz);
> > if (nhdr.n_type == NT_GNU_BUILD_ID &&
> > nhdr.n_namesz == sizeof("GNU")) {
> > - if (read(fd, bf, namesz) != (ssize_t)namesz)
> > + if (TEMP_FAILURE_RETRY(read(fd, bf, namesz)) != (ssize_t)namesz)
> > break;
> > if (memcmp(bf, "GNU", sizeof("GNU")) == 0) {
> > size_t sz = min(descsz, size);
> > - if (read(fd, bid->data, sz) == (ssize_t)sz) {
> > + if (TEMP_FAILURE_RETRY(read(fd, bid->data, sz)) == (ssize_t)sz) {
> > memset(bid->data + sz, 0, size - sz);
> > bid->size = sz;
> > err = 0;
> > break;
> > }
> > - } else if (read(fd, bf, descsz) != (ssize_t)descsz)
> > + } else if (TEMP_FAILURE_RETRY(read(fd, bf, descsz)) != (ssize_t)descsz)
> > break;
> > } else {
> > int n = namesz + descsz;
> > @@ -674,7 +674,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
> > __func__, filename, nhdr.n_namesz, nhdr.n_descsz);
> > }
> > - if (read(fd, bf, n) != n)
> > + if (TEMP_FAILURE_RETRY(read(fd, bf, n)) != n)
> > break;
> > }
> > }
> > @@ -732,7 +732,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,
> > Elf_Scn *sec;
> > Elf_Kind ek;
> >
> > - fd = open(filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> > if (fd < 0)
> > goto out;
> >
> > @@ -844,7 +844,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> >
> > type = dso->symtab_type;
> > } else {
> > - fd = open(name, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> > if (fd < 0) {
> > dso->load_errno = errno;
> > return -1;
> > @@ -1454,7 +1454,7 @@ static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len)
> > if (len < n)
> > n = len;
> > /* Use read because mmap won't work on proc files */
> > - r = read(from, buf, n);
> > + r = TEMP_FAILURE_RETRY(read(from, buf, n));
> > if (r < 0)
> > goto out;
> > if (!r)
> > @@ -1485,7 +1485,7 @@ static int kcore__open(struct kcore *kcore, const char *filename)
> > {
> > GElf_Ehdr *ehdr;
> >
> > - kcore->fd = open(filename, O_RDONLY);
> > + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> > if (kcore->fd == -1)
> > return -1;
> >
> > @@ -1518,7 +1518,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,
> > if (temp)
> > kcore->fd = mkstemp(filename);
> > else
> > - kcore->fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400);
> > + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400));
> > if (kcore->fd == -1)
> > return -1;
> >
> > @@ -1966,7 +1966,7 @@ static int kcore_copy__compare_fds(int from, int to)
> >
> > while (1) {
> > /* Use read because mmap won't work on proc files */
> > - ret = read(from, buf_from, page_size);
> > + ret = TEMP_FAILURE_RETRY(read(from, buf_from, page_size));
> > if (ret < 0)
> > goto out;
> >
> > @@ -1994,11 +1994,11 @@ static int kcore_copy__compare_files(const char *from_filename,
> > {
> > int from, to, err = -1;
> >
> > - from = open(from_filename, O_RDONLY);
> > + from = TEMP_FAILURE_RETRY(open(from_filename, O_RDONLY));
> > if (from < 0)
> > return -1;
> >
> > - to = open(to_filename, O_RDONLY);
> > + to = TEMP_FAILURE_RETRY(open(to_filename, O_RDONLY));
> > if (to < 0)
> > goto out_close_from;
> >
> > @@ -2419,7 +2419,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)
> > Elf *elf;
> > int fd, ret;
> >
> > - fd = open(target, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(target, O_RDONLY));
> > if (fd < 0)
> > return -EBADF;
> >
> > diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
> > index f9eb0bee7f15..7e25d476b1b2 100644
> > --- a/tools/perf/util/symbol-minimal.c
> > +++ b/tools/perf/util/symbol-minimal.c
> > @@ -230,7 +230,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > size_t buf_size;
> > void *buf;
> >
> > - fd = open(filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> > if (fd < 0)
> > return -1;
> >
> > @@ -242,7 +242,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > if (buf == NULL)
> > goto out;
> >
> > - if (read(fd, buf, buf_size) != (ssize_t) buf_size)
> > + if (TEMP_FAILURE_RETRY(read(fd, buf, buf_size)) != (ssize_t) buf_size)
> > goto out_free;
> >
> > ret = read_build_id(buf, buf_size, bid, false);
> > @@ -256,7 +256,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> > int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> > enum dso_binary_type type)
> > {
> > - int fd = open(name, O_RDONLY);
> > + int fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> > if (fd < 0)
> > goto out_errno;
> >
> > diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> > index f72baf636724..5212667ea67d 100644
> > --- a/tools/perf/util/symbol.c
> > +++ b/tools/perf/util/symbol.c
> > @@ -1335,7 +1335,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
> > md.dso = dso;
> > INIT_LIST_HEAD(&md.maps);
> >
> > - fd = open(kcore_filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(kcore_filename, O_RDONLY));
> > if (fd < 0) {
> > pr_debug("Failed to open %s. Note /proc/kcore requires CAP_SYS_RAWIO capability to access.\n",
> > kcore_filename);
> > @@ -2156,7 +2156,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
> > */
> > static bool filename__readable(const char *file)
> > {
> > - int fd = open(file, O_RDONLY);
> > + int fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> > if (fd < 0)
> > return false;
> > close(fd);
> > diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
> > index 76beda3e1a10..9dcc268bd69c 100644
> > --- a/tools/perf/util/synthetic-events.c
> > +++ b/tools/perf/util/synthetic-events.c
> > @@ -40,6 +40,7 @@
> > #include <api/io.h>
> > #include <sys/types.h>
> > #include <sys/stat.h>
> > +#include <errno.h>
> > #include <fcntl.h>
> > #include <unistd.h>
> >
> > @@ -86,13 +87,13 @@ static int perf_event__get_comm_ids(pid_t pid, pid_t tid, char *comm, size_t len
> > else
> > snprintf(bf, sizeof(bf), "/proc/%d/status", tid);
> >
> > - fd = open(bf, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY));
> > if (fd < 0) {
> > pr_debug("couldn't open %s\n", bf);
> > return -1;
> > }
> >
> > - n = read(fd, bf, sizeof(bf) - 1);
> > + n = TEMP_FAILURE_RETRY(read(fd, bf, sizeof(bf) - 1));
> > close(fd);
> > if (n <= 0) {
> > pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
> > @@ -410,7 +411,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
> > snprintf(bf, sizeof(bf), "%s/proc/%d/task/%d/maps",
> > machine->root_dir, pid, pid);
> >
> > - io.fd = open(bf, O_RDONLY, 0);
> > + io.fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY, 0));
> > if (io.fd < 0) {
> > /*
> > * We raced with a task exiting - just return:
> > diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c
> > index a65f65d0857e..c917a4217c6f 100644
> > --- a/tools/perf/util/trace-event-info.c
> > +++ b/tools/perf/util/trace-event-info.c
> > @@ -48,7 +48,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> > int r, fd;
> > int err = -EIO;
> >
> > - fd = open(file, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> > if (fd < 0) {
> > pr_debug("Can't read '%s'", file);
> > return -errno;
> > @@ -61,7 +61,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> > }
> >
> > do {
> > - r = read(fd, buf, BUFSIZ);
> > + r = TEMP_FAILURE_RETRY(read(fd, buf, BUFSIZ));
> > if (r > 0) {
> > size += r;
> > if (write(output_fd, buf, r) != r)
> > @@ -519,7 +519,7 @@ struct tracing_data *tracing_data_get(struct list_head *pattrs,
> > return NULL;
> > }
> >
> > - temp_fd = open(tdata->temp_file, O_RDWR);
> > + temp_fd = TEMP_FAILURE_RETRY(open(tdata->temp_file, O_RDWR));
> > if (temp_fd < 0) {
> > pr_debug("Can't read '%s'", tdata->temp_file);
> > free(tdata);
> > diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c
> > index 8a01af783310..7086770ec225 100644
> > --- a/tools/perf/util/trace-event-read.c
> > +++ b/tools/perf/util/trace-event-read.c
> > @@ -11,6 +11,7 @@
> > #include <sys/stat.h>
> > #include <sys/wait.h>
> > #include <sys/mman.h>
> > +#include <errno.h>
> > #include <fcntl.h>
> > #include <unistd.h>
> > #include <errno.h>
> > @@ -28,7 +29,7 @@ static int __do_read(int fd, void *buf, int size)
> > int rsize = size;
> >
> > while (size) {
> > - int ret = read(fd, buf, size);
> > + int ret = TEMP_FAILURE_RETRY(read(fd, buf, size));
> >
> > if (ret <= 0)
> > return -1;
> > @@ -105,7 +106,7 @@ static char *read_string(void)
> > char c;
> >
> > for (;;) {
> > - r = read(input_fd, &c, 1);
> > + r = TEMP_FAILURE_RETRY(read(input_fd, &c, 1));
> > if (r < 0) {
> > pr_debug("reading input file");
> > goto out;
> > diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c
> > index 37622699c91a..2ece387a73c6 100644
> > --- a/tools/perf/util/unwind-libunwind-local.c
> > +++ b/tools/perf/util/unwind-libunwind-local.c
> > @@ -362,7 +362,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> > }
> >
> > if (ofs <= 0) {
> > - fd = open(dso->symsrc_filename, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(dso->symsrc_filename, O_RDONLY));
> > if (fd >= 0) {
> > ofs = elf_section_offset(fd, ".debug_frame");
> > close(fd);
> > @@ -377,7 +377,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> > dso, DSO_BINARY_TYPE__DEBUGLINK,
> > machine->root_dir, debuglink, PATH_MAX);
> > if (!ret) {
> > - fd = open(debuglink, O_RDONLY);
> > + fd = TEMP_FAILURE_RETRY(open(debuglink, O_RDONLY));
> > if (fd >= 0) {
> > ofs = elf_section_offset(fd,
> > ".debug_frame");
> > diff --git a/tools/perf/util/zlib.c b/tools/perf/util/zlib.c
> > index 78d2297c1b67..f4563ea094c8 100644
> > --- a/tools/perf/util/zlib.c
> > +++ b/tools/perf/util/zlib.c
> > @@ -1,4 +1,5 @@
> > // SPDX-License-Identifier: GPL-2.0
> > +#include <errno.h>
> > #include <fcntl.h>
> > #include <stdio.h>
> > #include <string.h>
> > @@ -82,7 +83,7 @@ int gzip_decompress_to_file(const char *input, int output_fd)
> >
> > bool gzip_is_compressed(const char *input)
> > {
> > - int fd = open(input, O_RDONLY);
> > + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> > const uint8_t magic[2] = { 0x1f, 0x8b };
> > char buf[2] = { 0 };
> > ssize_t rc;
> > @@ -90,7 +91,7 @@ bool gzip_is_compressed(const char *input)
> > if (fd < 0)
> > return -1;
> >
> > - rc = read(fd, buf, sizeof(buf));
> > + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> > close(fd);
> > return rc == sizeof(buf) ?
> > memcmp(buf, magic, sizeof(buf)) == 0 : false;
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 15:34 ` Ian Rogers
@ 2022-06-15 16:40 ` Adrian Hunter
2022-06-15 17:38 ` Ian Rogers
0 siblings, 1 reply; 7+ messages in thread
From: Adrian Hunter @ 2022-06-15 16:40 UTC (permalink / raw)
To: Ian Rogers
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Alexey Bayduraev, Riccardo Mancini, Dave Marchevsky, Song Liu,
Leo Yan, John Garry, Florian Fischer, Andi Kleen, Kan Liang,
Masami Hiramatsu, Zechuan Chen, Yang Jihong, German Gomez,
Thomas Richter, James Clark, Denis Nikitin, Lexi Shao,
Michael Petlan, Remi Bernon, Fangrui Song, linux-kernel,
linux-perf-users, Stephane Eranian
On 15/06/22 18:34, Ian Rogers wrote:
> On Tue, Jun 14, 2022 at 11:51 PM Adrian Hunter <adrian.hunter@intel.com> wrote:
>>
>> On 15/06/22 09:03, Ian Rogers wrote:
>>> Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
>>> causes the syscall to need to restart.
>>
>> Please expand on why you want this change. Can you give an example
>> of where it makes a difference?
>
> As we can get signals pretty much anywhere in the code many syscalls
> are robust to this:
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n311
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n372
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n1585
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n2502
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-top.c?h=perf/core#n707
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/copyfile.c?h=perf/core#n59
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/evlist.c?h=perf/core#n1966
>
> But as this change shows there are a large number that are not. They
> will get EINTR and turn it into a failure which really isn't one.
> Making an example of this is near impossible as you need to arrange
> the signal delivery very carefully. I am however motivated to do this
> as we're seeing missing events when running perf in fleetwide
> profiling and I wish to discount EINTR as a cause.
Aren't most signals set to Ign or Term? What signals could you be getting?
Is signal() setting SA_RESTART by default? Does that make a difference?
>
>> It looks pretty ugly so maybe making replacement functions or macros
>> is better e.g.
>>
>> #define open_retry(a, ...) \
>> ({ long int __result; \
>> do __result = (long int) open((a), ##__VA_ARGS__); \
>> while (__result == -1L && errno == EINTR); \
>> __result; })
>
> As open is a user land thing there is no Linux precedent to follow
> here. TEMP_FAILURE_RETRY is the general norm, and I'd like to see it
> used when a raw system call is made. ie, if you want open_retry:
>
> #define open_retry(a, ...) TEMP_FAILURE_RETRY(open((a), ##__VA_ARGS__))
>
> I don't think we need to be in the game of reinventing
> TEMP_FAILURE_RETRY.
It was only an example. TEMP_FAILURE_RETRY seemed non-standard,
but I guess if we're defining _GNU_SOURCE anyway it doesn't matter.
> I think we can follow on from this change with
> changes not to use raw open and read calls, instead preferring some of
> our existing EINTR robust reads like readn, etc. Doing that is beyond
> the scope of what this change is trying to do.
>
> Thanks,
> Ian
>
>>>
>>> Signed-off-by: Ian Rogers <irogers@google.com>
>>> ---
>>> tools/lib/api/io.h | 3 ++-
>>> tools/lib/symbol/kallsyms.c | 4 ++-
>>> tools/perf/builtin-daemon.c | 19 +++++++-------
>>> tools/perf/builtin-ftrace.c | 18 ++++++-------
>>> tools/perf/builtin-inject.c | 2 +-
>>> tools/perf/builtin-kvm.c | 2 +-
>>> tools/perf/builtin-record.c | 6 ++---
>>> tools/perf/builtin-sched.c | 2 +-
>>> tools/perf/builtin-script.c | 2 +-
>>> tools/perf/util/copyfile.c | 2 +-
>>> tools/perf/util/data.c | 11 ++++----
>>> tools/perf/util/dso.c | 6 ++---
>>> tools/perf/util/evlist.c | 6 ++---
>>> tools/perf/util/lzma.c | 4 +--
>>> tools/perf/util/namespaces.c | 5 ++--
>>> tools/perf/util/parse-events.c | 6 ++---
>>> tools/perf/util/pmu.c | 12 ++++-----
>>> tools/perf/util/probe-event.c | 4 +--
>>> tools/perf/util/probe-file.c | 10 ++++----
>>> tools/perf/util/session.c | 2 +-
>>> tools/perf/util/symbol-elf.c | 32 ++++++++++++------------
>>> tools/perf/util/symbol-minimal.c | 6 ++---
>>> tools/perf/util/symbol.c | 4 +--
>>> tools/perf/util/synthetic-events.c | 7 +++---
>>> tools/perf/util/trace-event-info.c | 6 ++---
>>> tools/perf/util/trace-event-read.c | 5 ++--
>>> tools/perf/util/unwind-libunwind-local.c | 4 +--
>>> tools/perf/util/zlib.c | 5 ++--
>>> 28 files changed, 102 insertions(+), 93 deletions(-)
>>>
>>> diff --git a/tools/lib/api/io.h b/tools/lib/api/io.h
>>> index 777c20f6b604..b292bbd39fb0 100644
>>> --- a/tools/lib/api/io.h
>>> +++ b/tools/lib/api/io.h
>>> @@ -7,6 +7,7 @@
>>> #ifndef __API_IO__
>>> #define __API_IO__
>>>
>>> +#include <errno.h>
>>> #include <stdlib.h>
>>> #include <unistd.h>
>>>
>>> @@ -45,7 +46,7 @@ static inline int io__get_char(struct io *io)
>>> return -1;
>>>
>>> if (ptr == io->end) {
>>> - ssize_t n = read(io->fd, io->buf, io->buf_len);
>>> + ssize_t n = TEMP_FAILURE_RETRY(read(io->fd, io->buf, io->buf_len));
>>>
>>> if (n <= 0) {
>>> io->eof = true;
>>> diff --git a/tools/lib/symbol/kallsyms.c b/tools/lib/symbol/kallsyms.c
>>> index e335ac2b9e19..5436c038c379 100644
>>> --- a/tools/lib/symbol/kallsyms.c
>>> +++ b/tools/lib/symbol/kallsyms.c
>>> @@ -3,7 +3,9 @@
>>> #include "api/io.h"
>>> #include <stdio.h>
>>> #include <sys/stat.h>
>>> +#include <errno.h>
>>> #include <fcntl.h>
>>> +#include <unistd.h>
>>>
>>> u8 kallsyms2elf_type(char type)
>>> {
>>> @@ -36,7 +38,7 @@ int kallsyms__parse(const char *filename, void *arg,
>>> char bf[BUFSIZ];
>>> int err;
>>>
>>> - io.fd = open(filename, O_RDONLY, 0);
>>> + io.fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY, 0));
>>>
>>> if (io.fd < 0)
>>> return -1;
>>> diff --git a/tools/perf/builtin-daemon.c b/tools/perf/builtin-daemon.c
>>> index 6cb3f6cc36d0..d4e1a5636056 100644
>>> --- a/tools/perf/builtin-daemon.c
>>> +++ b/tools/perf/builtin-daemon.c
>>> @@ -351,7 +351,7 @@ static int daemon_session__run(struct daemon_session *session,
>>> return -1;
>>> }
>>>
>>> - fd = open("/dev/null", O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open("/dev/null", O_RDONLY));
>>> if (fd < 0) {
>>> perror("failed: open /dev/null");
>>> return -1;
>>> @@ -360,7 +360,7 @@ static int daemon_session__run(struct daemon_session *session,
>>> dup2(fd, 0);
>>> close(fd);
>>>
>>> - fd = open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644);
>>> + fd = TEMP_FAILURE_RETRY(open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644));
>>> if (fd < 0) {
>>> perror("failed: open session output");
>>> return -1;
>>> @@ -405,7 +405,8 @@ static pid_t handle_signalfd(struct daemon *daemon)
>>> * coalesced in kernel and we can receive only single signal even
>>> * if multiple SIGCHLD were generated.
>>> */
>>> - err = read(daemon->signal_fd, &si, sizeof(struct signalfd_siginfo));
>>> + err = TEMP_FAILURE_RETRY(read(daemon->signal_fd, &si,
>>> + sizeof(struct signalfd_siginfo)));
>>> if (err != sizeof(struct signalfd_siginfo)) {
>>> pr_err("failed to read signal fd\n");
>>> return -1;
>>> @@ -522,7 +523,7 @@ static int daemon_session__control(struct daemon_session *session,
>>> scnprintf(control_path, sizeof(control_path), "%s/%s",
>>> session->base, SESSION_CONTROL);
>>>
>>> - control = open(control_path, O_WRONLY|O_NONBLOCK);
>>> + control = TEMP_FAILURE_RETRY(open(control_path, O_WRONLY|O_NONBLOCK));
>>> if (!control)
>>> return -1;
>>>
>>> @@ -531,7 +532,7 @@ static int daemon_session__control(struct daemon_session *session,
>>> scnprintf(ack_path, sizeof(ack_path), "%s/%s",
>>> session->base, SESSION_ACK);
>>>
>>> - ack = open(ack_path, O_RDONLY, O_NONBLOCK);
>>> + ack = TEMP_FAILURE_RETRY(open(ack_path, O_RDONLY, O_NONBLOCK));
>>> if (!ack) {
>>> close(control);
>>> return -1;
>>> @@ -564,7 +565,7 @@ static int daemon_session__control(struct daemon_session *session,
>>> goto out;
>>> }
>>>
>>> - err = read(ack, buf, sizeof(buf));
>>> + err = TEMP_FAILURE_RETRY(read(ack, buf, sizeof(buf)));
>>> if (err > 0)
>>> ret = strcmp(buf, "ack\n");
>>> else
>>> @@ -1078,7 +1079,7 @@ static int handle_config_changes(struct daemon *daemon, int conf_fd,
>>> ssize_t len;
>>>
>>> while (!(*config_changed)) {
>>> - len = read(conf_fd, buf, sizeof(buf));
>>> + len = TEMP_FAILURE_RETRY(read(conf_fd, buf, sizeof(buf)));
>>> if (len == -1) {
>>> if (errno != EAGAIN) {
>>> perror("failed: read");
>>> @@ -1147,7 +1148,7 @@ static int check_lock(struct daemon *daemon)
>>>
>>> scnprintf(path, sizeof(path), "%s/lock", daemon->base);
>>>
>>> - fd = open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640);
>>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640));
>>> if (fd < 0)
>>> return -1;
>>>
>>> @@ -1201,7 +1202,7 @@ static int go_background(struct daemon *daemon)
>>> return -1;
>>> }
>>>
>>> - fd = open("output", O_RDWR|O_CREAT|O_TRUNC, 0644);
>>> + fd = TEMP_FAILURE_RETRY(open("output", O_RDWR|O_CREAT|O_TRUNC, 0644));
>>> if (fd < 0) {
>>> perror("failed: open");
>>> return -1;
>>> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
>>> index 7de07bb16d23..7553e20f38b2 100644
>>> --- a/tools/perf/builtin-ftrace.c
>>> +++ b/tools/perf/builtin-ftrace.c
>>> @@ -79,7 +79,7 @@ static int __write_tracing_file(const char *name, const char *val, bool append)
>>> else
>>> flags |= O_TRUNC;
>>>
>>> - fd = open(file, flags);
>>> + fd = TEMP_FAILURE_RETRY(open(file, flags));
>>> if (fd < 0) {
>>> pr_debug("cannot open tracing file: %s: %s\n",
>>> name, str_error_r(errno, errbuf, sizeof(errbuf)));
>>> @@ -132,7 +132,7 @@ static int read_tracing_file_to_stdout(const char *name)
>>> return -1;
>>> }
>>>
>>> - fd = open(file, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
>>> if (fd < 0) {
>>> pr_debug("cannot open tracing file: %s: %s\n",
>>> name, str_error_r(errno, buf, sizeof(buf)));
>>> @@ -141,7 +141,7 @@ static int read_tracing_file_to_stdout(const char *name)
>>>
>>> /* read contents to stdout */
>>> while (true) {
>>> - int n = read(fd, buf, sizeof(buf));
>>> + int n = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
>>> if (n == 0)
>>> break;
>>> else if (n < 0)
>>> @@ -608,7 +608,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
>>> goto out_reset;
>>> }
>>>
>>> - trace_fd = open(trace_file, O_RDONLY);
>>> + trace_fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
>>>
>>> put_tracing_file(trace_file);
>>>
>>> @@ -645,7 +645,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
>>> break;
>>>
>>> if (pollfd.revents & POLLIN) {
>>> - int n = read(trace_fd, buf, sizeof(buf));
>>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
>>> if (n < 0)
>>> break;
>>> if (fwrite(buf, n, 1, stdout) != 1)
>>> @@ -665,7 +665,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
>>>
>>> /* read remaining buffer contents */
>>> while (true) {
>>> - int n = read(trace_fd, buf, sizeof(buf));
>>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
>>> if (n <= 0)
>>> break;
>>> if (fwrite(buf, n, 1, stdout) != 1)
>>> @@ -826,7 +826,7 @@ static int prepare_func_latency(struct perf_ftrace *ftrace)
>>> return -1;
>>> }
>>>
>>> - fd = open(trace_file, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
>>> if (fd < 0)
>>> pr_err("failed to open trace_pipe\n");
>>>
>>> @@ -913,7 +913,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
>>> break;
>>>
>>> if (pollfd.revents & POLLIN) {
>>> - int n = read(trace_fd, buf, sizeof(buf) - 1);
>>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
>>> if (n < 0)
>>> break;
>>>
>>> @@ -931,7 +931,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
>>>
>>> /* read remaining buffer contents */
>>> while (!ftrace->target.use_bpf) {
>>> - int n = read(trace_fd, buf, sizeof(buf) - 1);
>>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
>>> if (n <= 0)
>>> break;
>>> make_histogram(buckets, buf, n, line, ftrace->use_nsec);
>>> diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
>>> index a75bf11585b5..960b4a50583e 100644
>>> --- a/tools/perf/builtin-inject.c
>>> +++ b/tools/perf/builtin-inject.c
>>> @@ -155,7 +155,7 @@ static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
>>> int ret;
>>>
>>> while (size > 0) {
>>> - ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
>>> + ssz = TEMP_FAILURE_RETRY(read(fd, buf, min(size, (off_t)sizeof(buf))));
>>> if (ssz < 0)
>>> return -errno;
>>> ret = output_bytes(inject, buf, ssz);
>>> diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
>>> index 3696ae97f149..4dd8292d1941 100644
>>> --- a/tools/perf/builtin-kvm.c
>>> +++ b/tools/perf/builtin-kvm.c
>>> @@ -880,7 +880,7 @@ static int perf_kvm__handle_timerfd(struct perf_kvm_stat *kvm)
>>> uint64_t c;
>>> int rc;
>>>
>>> - rc = read(kvm->timerfd, &c, sizeof(uint64_t));
>>> + rc = TEMP_FAILURE_RETRY(read(kvm->timerfd, &c, sizeof(uint64_t)));
>>> if (rc < 0) {
>>> if (errno == EAGAIN)
>>> return 0;
>>> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
>>> index 9a71f0330137..361610bc96b6 100644
>>> --- a/tools/perf/builtin-record.c
>>> +++ b/tools/perf/builtin-record.c
>>> @@ -902,7 +902,7 @@ static bool record__kcore_readable(struct machine *machine)
>>>
>>> scnprintf(kcore, sizeof(kcore), "%s/proc/kcore", machine->root_dir);
>>>
>>> - fd = open(kcore, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(kcore, O_RDONLY));
>>> if (fd < 0)
>>> return false;
>>>
>>> @@ -2049,7 +2049,7 @@ static int record__terminate_thread(struct record_thread *thread_data)
>>>
>>> close(thread_data->pipes.msg[1]);
>>> thread_data->pipes.msg[1] = -1;
>>> - err = read(thread_data->pipes.ack[0], &ack, sizeof(ack));
>>> + err = TEMP_FAILURE_RETRY(read(thread_data->pipes.ack[0], &ack, sizeof(ack)));
>>> if (err > 0)
>>> pr_debug2("threads[%d]: sent %s\n", tid, thread_msg_tags[ack]);
>>> else
>>> @@ -2097,7 +2097,7 @@ static int record__start_threads(struct record *rec)
>>> goto out_err;
>>> }
>>>
>>> - err = read(thread_data[t].pipes.ack[0], &msg, sizeof(msg));
>>> + err = TEMP_FAILURE_RETRY(read(thread_data[t].pipes.ack[0], &msg, sizeof(msg)));
>>> if (err > 0)
>>> pr_debug2("threads[%d]: sent %s\n", rec->thread_data[t].tid,
>>> thread_msg_tags[msg]);
>>> diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
>>> index 646bd938927a..149787268d02 100644
>>> --- a/tools/perf/builtin-sched.c
>>> +++ b/tools/perf/builtin-sched.c
>>> @@ -604,7 +604,7 @@ static u64 get_cpu_usage_nsec_self(int fd)
>>> u64 runtime;
>>> int ret;
>>>
>>> - ret = read(fd, &runtime, sizeof(runtime));
>>> + ret = TEMP_FAILURE_RETRY(read(fd, &runtime, sizeof(runtime)));
>>> BUG_ON(ret != sizeof(runtime));
>>>
>>> return runtime;
>>> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
>>> index c689054002cc..844f5350bfdf 100644
>>> --- a/tools/perf/builtin-script.c
>>> +++ b/tools/perf/builtin-script.c
>>> @@ -4207,7 +4207,7 @@ int cmd_script(int argc, const char **argv)
>>> goto out_delete;
>>> }
>>>
>>> - input = open(data.path, O_RDONLY); /* input_name */
>>> + input = TEMP_FAILURE_RETRY(open(data.path, O_RDONLY)); /* input_name */
>>> if (input < 0) {
>>> err = -errno;
>>> perror("failed to open file");
>>> diff --git a/tools/perf/util/copyfile.c b/tools/perf/util/copyfile.c
>>> index 47e03de7c235..c7bff4943936 100644
>>> --- a/tools/perf/util/copyfile.c
>>> +++ b/tools/perf/util/copyfile.c
>>> @@ -112,7 +112,7 @@ static int copyfile_mode_ns(const char *from, const char *to, mode_t mode,
>>> goto out_close_to;
>>>
>>> nsinfo__mountns_enter(nsi, &nsc);
>>> - fromfd = open(from, O_RDONLY);
>>> + fromfd = TEMP_FAILURE_RETRY(open(from, O_RDONLY));
>>> nsinfo__mountns_exit(&nsc);
>>> if (fromfd < 0)
>>> goto out_close_to;
>>> diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c
>>> index caabeac24c69..a9066a3d914f 100644
>>> --- a/tools/perf/util/data.c
>>> +++ b/tools/perf/util/data.c
>>> @@ -53,7 +53,8 @@ int perf_data__create_dir(struct perf_data *data, int nr)
>>> goto out_err;
>>> }
>>>
>>> - ret = open(file->path, O_RDWR|O_CREAT|O_TRUNC, S_IRUSR|S_IWUSR);
>>> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDWR|O_CREAT|O_TRUNC,
>>> + S_IRUSR|S_IWUSR));
>>> if (ret < 0) {
>>> ret = -errno;
>>> goto out_err;
>>> @@ -123,7 +124,7 @@ int perf_data__open_dir(struct perf_data *data)
>>> if (!file->path)
>>> goto out_err;
>>>
>>> - ret = open(file->path, O_RDONLY);
>>> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDONLY));
>>> if (ret < 0)
>>> goto out_err;
>>>
>>> @@ -248,7 +249,7 @@ static int open_file_read(struct perf_data *data)
>>> int fd;
>>> char sbuf[STRERR_BUFSIZE];
>>>
>>> - fd = open(data->file.path, flags);
>>> + fd = TEMP_FAILURE_RETRY(open(data->file.path, flags));
>>> if (fd < 0) {
>>> int err = errno;
>>>
>>> @@ -288,8 +289,8 @@ static int open_file_write(struct perf_data *data)
>>> int fd;
>>> char sbuf[STRERR_BUFSIZE];
>>>
>>> - fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
>>> - S_IRUSR|S_IWUSR);
>>> + fd = TEMP_FAILURE_RETRY(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,
>>> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
>>> index 5ac13958d1bd..40befa2429f9 100644
>>> --- a/tools/perf/util/dso.c
>>> +++ b/tools/perf/util/dso.c
>>> @@ -300,7 +300,7 @@ int filename__decompress(const char *name, char *pathname,
>>> * descriptor to the uncompressed file.
>>> */
>>> if (!compressions[comp].is_compressed(name))
>>> - return open(name, O_RDONLY);
>>> + return TEMP_FAILURE_RETRY(open(name, O_RDONLY));
>>>
>>> fd = mkstemp(tmpbuf);
>>> if (fd < 0) {
>>> @@ -476,7 +476,7 @@ static int do_open(char *name)
>>> char sbuf[STRERR_BUFSIZE];
>>>
>>> do {
>>> - fd = open(name, O_RDONLY|O_CLOEXEC);
>>> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY|O_CLOEXEC));
>>> if (fd >= 0)
>>> return fd;
>>>
>>> @@ -898,7 +898,7 @@ static ssize_t file_read(struct dso *dso, struct machine *machine,
>>> goto out;
>>> }
>>>
>>> - ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset);
>>> + ret = TEMP_FAILURE_RETRY(pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset));
>>> out:
>>> pthread_mutex_unlock(&dso__data_open_lock);
>>> return ret;
>>> diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
>>> index 48af7d379d82..ba0dcdf91120 100644
>>> --- a/tools/perf/util/evlist.c
>>> +++ b/tools/perf/util/evlist.c
>>> @@ -1440,7 +1440,7 @@ int evlist__prepare_workload(struct evlist *evlist, struct target *target, const
>>> /*
>>> * Wait until the parent tells us to go.
>>> */
>>> - ret = read(go_pipe[0], &bf, 1);
>>> + ret = TEMP_FAILURE_RETRY(read(go_pipe[0], &bf, 1));
>>> /*
>>> * The parent will ask for the execvp() to be performed by
>>> * writing exactly one byte, in workload.cork_fd, usually via
>>> @@ -1837,7 +1837,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
>>> * O_RDWR avoids POLLHUPs which is necessary to allow the other
>>> * end of a FIFO to be repeatedly opened and closed.
>>> */
>>> - fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC);
>>> + fd = TEMP_FAILURE_RETRY(open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC));
>>> if (fd < 0) {
>>> pr_err("Failed to open '%s'\n", s);
>>> ret = -errno;
>>> @@ -1848,7 +1848,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
>>>
>>> if (p && *++p) {
>>> /* O_RDWR | O_NONBLOCK means the other end need not be open */
>>> - fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC);
>>> + fd = TEMP_FAILURE_RETRY(open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC));
>>> if (fd < 0) {
>>> pr_err("Failed to open '%s'\n", p);
>>> ret = -errno;
>>> diff --git a/tools/perf/util/lzma.c b/tools/perf/util/lzma.c
>>> index 51424cdc3b68..93e0a6a53e28 100644
>>> --- a/tools/perf/util/lzma.c
>>> +++ b/tools/perf/util/lzma.c
>>> @@ -109,7 +109,7 @@ int lzma_decompress_to_file(const char *input, int output_fd)
>>>
>>> bool lzma_is_compressed(const char *input)
>>> {
>>> - int fd = open(input, O_RDONLY);
>>> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
>>> const uint8_t magic[6] = { 0xFD, '7', 'z', 'X', 'Z', 0x00 };
>>> char buf[6] = { 0 };
>>> ssize_t rc;
>>> @@ -117,7 +117,7 @@ bool lzma_is_compressed(const char *input)
>>> if (fd < 0)
>>> return -1;
>>>
>>> - rc = read(fd, buf, sizeof(buf));
>>> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
>>> close(fd);
>>> return rc == sizeof(buf) ?
>>> memcmp(buf, magic, sizeof(buf)) == 0 : false;
>>> diff --git a/tools/perf/util/namespaces.c b/tools/perf/util/namespaces.c
>>> index dd536220cdb9..89edfba6af06 100644
>>> --- a/tools/perf/util/namespaces.c
>>> +++ b/tools/perf/util/namespaces.c
>>> @@ -9,6 +9,7 @@
>>> #include "get_current_dir_name.h"
>>> #include <sys/types.h>
>>> #include <sys/stat.h>
>>> +#include <errno.h>
>>> #include <fcntl.h>
>>> #include <limits.h>
>>> #include <sched.h>
>>> @@ -266,11 +267,11 @@ void nsinfo__mountns_enter(struct nsinfo *nsi,
>>> if (!oldcwd)
>>> return;
>>>
>>> - oldns = open(curpath, O_RDONLY);
>>> + oldns = TEMP_FAILURE_RETRY(open(curpath, O_RDONLY));
>>> if (oldns < 0)
>>> goto errout;
>>>
>>> - newns = open(nsi->mntns_path, O_RDONLY);
>>> + newns = TEMP_FAILURE_RETRY(open(nsi->mntns_path, O_RDONLY));
>>> if (newns < 0)
>>> goto errout;
>>>
>>> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
>>> index 7ed235740431..75f8f2e44195 100644
>>> --- a/tools/perf/util/parse-events.c
>>> +++ b/tools/perf/util/parse-events.c
>>> @@ -189,7 +189,7 @@ static int tp_event_has_id(const char *dir_path, struct dirent *evt_dir)
>>> int fd;
>>>
>>> snprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path, evt_dir->d_name);
>>> - fd = open(evt_path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
>>> if (fd < 0)
>>> return -EINVAL;
>>> close(fd);
>>> @@ -233,10 +233,10 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config)
>>>
>>> scnprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path,
>>> evt_dirent->d_name);
>>> - fd = open(evt_path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
>>> if (fd < 0)
>>> continue;
>>> - if (read(fd, id_buf, sizeof(id_buf)) < 0) {
>>> + if (TEMP_FAILURE_RETRY(read(fd, id_buf, sizeof(id_buf))) < 0) {
>>> close(fd);
>>> continue;
>>> }
>>> diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
>>> index 9a1c7e63e663..59252cfc2b5b 100644
>>> --- a/tools/perf/util/pmu.c
>>> +++ b/tools/perf/util/pmu.c
>>> @@ -154,14 +154,14 @@ static int perf_pmu__parse_scale(struct perf_pmu_alias *alias, char *dir, char *
>>>
>>> scnprintf(path, PATH_MAX, "%s/%s.scale", dir, name);
>>>
>>> - fd = open(path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
>>> if (fd == -1)
>>> return -1;
>>>
>>> if (fstat(fd, &st) < 0)
>>> goto error;
>>>
>>> - sret = read(fd, scale, sizeof(scale)-1);
>>> + sret = TEMP_FAILURE_RETRY(read(fd, scale, sizeof(scale)-1));
>>> if (sret < 0)
>>> goto error;
>>>
>>> @@ -184,11 +184,11 @@ static int perf_pmu__parse_unit(struct perf_pmu_alias *alias, char *dir, char *n
>>>
>>> scnprintf(path, PATH_MAX, "%s/%s.unit", dir, name);
>>>
>>> - fd = open(path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
>>> if (fd == -1)
>>> return -1;
>>>
>>> - sret = read(fd, alias->unit, UNIT_MAX_LEN);
>>> + sret = TEMP_FAILURE_RETRY(read(fd, alias->unit, UNIT_MAX_LEN));
>>> if (sret < 0)
>>> goto error;
>>>
>>> @@ -214,7 +214,7 @@ perf_pmu__parse_per_pkg(struct perf_pmu_alias *alias, char *dir, char *name)
>>>
>>> scnprintf(path, PATH_MAX, "%s/%s.per-pkg", dir, name);
>>>
>>> - fd = open(path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
>>> if (fd == -1)
>>> return -1;
>>>
>>> @@ -232,7 +232,7 @@ static int perf_pmu__parse_snapshot(struct perf_pmu_alias *alias,
>>>
>>> scnprintf(path, PATH_MAX, "%s/%s.snapshot", dir, name);
>>>
>>> - fd = open(path, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
>>> if (fd == -1)
>>> return -1;
>>>
>>> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
>>> index 062b5cbe67af..7838889f55df 100644
>>> --- a/tools/perf/util/probe-event.c
>>> +++ b/tools/perf/util/probe-event.c
>>> @@ -276,7 +276,7 @@ static char *find_module_name(const char *module)
>>> char *mod_name = NULL;
>>> int name_offset;
>>>
>>> - fd = open(module, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(module, O_RDONLY));
>>> if (fd < 0)
>>> return NULL;
>>>
>>> @@ -598,7 +598,7 @@ static int get_text_start_address(const char *exec, u64 *address,
>>> struct nscookie nsc;
>>>
>>> nsinfo__mountns_enter(nsi, &nsc);
>>> - fd = open(exec, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(exec, O_RDONLY));
>>> nsinfo__mountns_exit(&nsc);
>>> if (fd < 0)
>>> return -errno;
>>> diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
>>> index 3d50de3217d5..b788932656e8 100644
>>> --- a/tools/perf/util/probe-file.c
>>> +++ b/tools/perf/util/probe-file.c
>>> @@ -115,9 +115,9 @@ int open_trace_file(const char *trace_file, bool readwrite)
>>> if (ret >= 0) {
>>> pr_debug("Opening %s write=%d\n", buf, readwrite);
>>> if (readwrite && !probe_event_dry_run)
>>> - ret = open(buf, O_RDWR | O_APPEND, 0);
>>> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDWR | O_APPEND, 0));
>>> else
>>> - ret = open(buf, O_RDONLY, 0);
>>> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDONLY, 0));
>>>
>>> if (ret < 0)
>>> ret = -errno;
>>> @@ -180,7 +180,7 @@ struct strlist *probe_file__get_rawlist(int fd)
>>> if (sl == NULL)
>>> return NULL;
>>>
>>> - fddup = dup(fd);
>>> + fddup = TEMP_FAILURE_RETRY(dup(fd));
>>> if (fddup < 0)
>>> goto out_free_sl;
>>>
>>> @@ -498,7 +498,7 @@ static int probe_cache__open(struct probe_cache *pcache, const char *target,
>>> }
>>>
>>> snprintf(cpath, PATH_MAX, "%s/probes", dir_name);
>>> - fd = open(cpath, O_CREAT | O_RDWR, 0644);
>>> + fd = TEMP_FAILURE_RETRY(open(cpath, O_CREAT | O_RDWR, 0644));
>>> if (fd < 0)
>>> pr_debug("Failed to open cache(%d): %s\n", fd, cpath);
>>> free(dir_name);
>>> @@ -514,7 +514,7 @@ static int probe_cache__load(struct probe_cache *pcache)
>>> int ret = 0, fddup;
>>> FILE *fp;
>>>
>>> - fddup = dup(pcache->fd);
>>> + fddup = TEMP_FAILURE_RETRY(dup(pcache->fd));
>>> if (fddup < 0)
>>> return -errno;
>>> fp = fdopen(fddup, "r");
>>> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
>>> index 0acb9de54b06..b2fc389cc530 100644
>>> --- a/tools/perf/util/session.c
>>> +++ b/tools/perf/util/session.c
>>> @@ -384,7 +384,7 @@ static int skipn(int fd, off_t n)
>>> ssize_t ret;
>>>
>>> while (n > 0) {
>>> - ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
>>> + ret = TEMP_FAILURE_RETRY(read(fd, buf, min(n, (off_t)sizeof(buf))));
>>> if (ret <= 0)
>>> return ret;
>>> n -= ret;
>>> diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
>>> index ecd377938eea..02556070a2ce 100644
>>> --- a/tools/perf/util/symbol-elf.c
>>> +++ b/tools/perf/util/symbol-elf.c
>>> @@ -577,7 +577,7 @@ static int read_build_id(const char *filename, struct build_id *bid)
>>> if (size < BUILD_ID_SIZE)
>>> goto out;
>>>
>>> - fd = open(filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
>>> if (fd < 0)
>>> goto out;
>>>
>>> @@ -638,7 +638,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> size_t size = sizeof(bid->data);
>>> int fd, err = -1;
>>>
>>> - fd = open(filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
>>> if (fd < 0)
>>> goto out;
>>>
>>> @@ -647,24 +647,24 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> GElf_Nhdr nhdr;
>>> size_t namesz, descsz;
>>>
>>> - if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr))
>>> + if (TEMP_FAILURE_RETRY(read(fd, &nhdr, sizeof(nhdr))) != sizeof(nhdr))
>>> break;
>>>
>>> namesz = NOTE_ALIGN(nhdr.n_namesz);
>>> descsz = NOTE_ALIGN(nhdr.n_descsz);
>>> if (nhdr.n_type == NT_GNU_BUILD_ID &&
>>> nhdr.n_namesz == sizeof("GNU")) {
>>> - if (read(fd, bf, namesz) != (ssize_t)namesz)
>>> + if (TEMP_FAILURE_RETRY(read(fd, bf, namesz)) != (ssize_t)namesz)
>>> break;
>>> if (memcmp(bf, "GNU", sizeof("GNU")) == 0) {
>>> size_t sz = min(descsz, size);
>>> - if (read(fd, bid->data, sz) == (ssize_t)sz) {
>>> + if (TEMP_FAILURE_RETRY(read(fd, bid->data, sz)) == (ssize_t)sz) {
>>> memset(bid->data + sz, 0, size - sz);
>>> bid->size = sz;
>>> err = 0;
>>> break;
>>> }
>>> - } else if (read(fd, bf, descsz) != (ssize_t)descsz)
>>> + } else if (TEMP_FAILURE_RETRY(read(fd, bf, descsz)) != (ssize_t)descsz)
>>> break;
>>> } else {
>>> int n = namesz + descsz;
>>> @@ -674,7 +674,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
>>> __func__, filename, nhdr.n_namesz, nhdr.n_descsz);
>>> }
>>> - if (read(fd, bf, n) != n)
>>> + if (TEMP_FAILURE_RETRY(read(fd, bf, n)) != n)
>>> break;
>>> }
>>> }
>>> @@ -732,7 +732,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,
>>> Elf_Scn *sec;
>>> Elf_Kind ek;
>>>
>>> - fd = open(filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
>>> if (fd < 0)
>>> goto out;
>>>
>>> @@ -844,7 +844,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
>>>
>>> type = dso->symtab_type;
>>> } else {
>>> - fd = open(name, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
>>> if (fd < 0) {
>>> dso->load_errno = errno;
>>> return -1;
>>> @@ -1454,7 +1454,7 @@ static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len)
>>> if (len < n)
>>> n = len;
>>> /* Use read because mmap won't work on proc files */
>>> - r = read(from, buf, n);
>>> + r = TEMP_FAILURE_RETRY(read(from, buf, n));
>>> if (r < 0)
>>> goto out;
>>> if (!r)
>>> @@ -1485,7 +1485,7 @@ static int kcore__open(struct kcore *kcore, const char *filename)
>>> {
>>> GElf_Ehdr *ehdr;
>>>
>>> - kcore->fd = open(filename, O_RDONLY);
>>> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
>>> if (kcore->fd == -1)
>>> return -1;
>>>
>>> @@ -1518,7 +1518,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,
>>> if (temp)
>>> kcore->fd = mkstemp(filename);
>>> else
>>> - kcore->fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400);
>>> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400));
>>> if (kcore->fd == -1)
>>> return -1;
>>>
>>> @@ -1966,7 +1966,7 @@ static int kcore_copy__compare_fds(int from, int to)
>>>
>>> while (1) {
>>> /* Use read because mmap won't work on proc files */
>>> - ret = read(from, buf_from, page_size);
>>> + ret = TEMP_FAILURE_RETRY(read(from, buf_from, page_size));
>>> if (ret < 0)
>>> goto out;
>>>
>>> @@ -1994,11 +1994,11 @@ static int kcore_copy__compare_files(const char *from_filename,
>>> {
>>> int from, to, err = -1;
>>>
>>> - from = open(from_filename, O_RDONLY);
>>> + from = TEMP_FAILURE_RETRY(open(from_filename, O_RDONLY));
>>> if (from < 0)
>>> return -1;
>>>
>>> - to = open(to_filename, O_RDONLY);
>>> + to = TEMP_FAILURE_RETRY(open(to_filename, O_RDONLY));
>>> if (to < 0)
>>> goto out_close_from;
>>>
>>> @@ -2419,7 +2419,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)
>>> Elf *elf;
>>> int fd, ret;
>>>
>>> - fd = open(target, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(target, O_RDONLY));
>>> if (fd < 0)
>>> return -EBADF;
>>>
>>> diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
>>> index f9eb0bee7f15..7e25d476b1b2 100644
>>> --- a/tools/perf/util/symbol-minimal.c
>>> +++ b/tools/perf/util/symbol-minimal.c
>>> @@ -230,7 +230,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> size_t buf_size;
>>> void *buf;
>>>
>>> - fd = open(filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
>>> if (fd < 0)
>>> return -1;
>>>
>>> @@ -242,7 +242,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> if (buf == NULL)
>>> goto out;
>>>
>>> - if (read(fd, buf, buf_size) != (ssize_t) buf_size)
>>> + if (TEMP_FAILURE_RETRY(read(fd, buf, buf_size)) != (ssize_t) buf_size)
>>> goto out_free;
>>>
>>> ret = read_build_id(buf, buf_size, bid, false);
>>> @@ -256,7 +256,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
>>> int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
>>> enum dso_binary_type type)
>>> {
>>> - int fd = open(name, O_RDONLY);
>>> + int fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
>>> if (fd < 0)
>>> goto out_errno;
>>>
>>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
>>> index f72baf636724..5212667ea67d 100644
>>> --- a/tools/perf/util/symbol.c
>>> +++ b/tools/perf/util/symbol.c
>>> @@ -1335,7 +1335,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
>>> md.dso = dso;
>>> INIT_LIST_HEAD(&md.maps);
>>>
>>> - fd = open(kcore_filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(kcore_filename, O_RDONLY));
>>> if (fd < 0) {
>>> pr_debug("Failed to open %s. Note /proc/kcore requires CAP_SYS_RAWIO capability to access.\n",
>>> kcore_filename);
>>> @@ -2156,7 +2156,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
>>> */
>>> static bool filename__readable(const char *file)
>>> {
>>> - int fd = open(file, O_RDONLY);
>>> + int fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
>>> if (fd < 0)
>>> return false;
>>> close(fd);
>>> diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
>>> index 76beda3e1a10..9dcc268bd69c 100644
>>> --- a/tools/perf/util/synthetic-events.c
>>> +++ b/tools/perf/util/synthetic-events.c
>>> @@ -40,6 +40,7 @@
>>> #include <api/io.h>
>>> #include <sys/types.h>
>>> #include <sys/stat.h>
>>> +#include <errno.h>
>>> #include <fcntl.h>
>>> #include <unistd.h>
>>>
>>> @@ -86,13 +87,13 @@ static int perf_event__get_comm_ids(pid_t pid, pid_t tid, char *comm, size_t len
>>> else
>>> snprintf(bf, sizeof(bf), "/proc/%d/status", tid);
>>>
>>> - fd = open(bf, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY));
>>> if (fd < 0) {
>>> pr_debug("couldn't open %s\n", bf);
>>> return -1;
>>> }
>>>
>>> - n = read(fd, bf, sizeof(bf) - 1);
>>> + n = TEMP_FAILURE_RETRY(read(fd, bf, sizeof(bf) - 1));
>>> close(fd);
>>> if (n <= 0) {
>>> pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
>>> @@ -410,7 +411,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
>>> snprintf(bf, sizeof(bf), "%s/proc/%d/task/%d/maps",
>>> machine->root_dir, pid, pid);
>>>
>>> - io.fd = open(bf, O_RDONLY, 0);
>>> + io.fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY, 0));
>>> if (io.fd < 0) {
>>> /*
>>> * We raced with a task exiting - just return:
>>> diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c
>>> index a65f65d0857e..c917a4217c6f 100644
>>> --- a/tools/perf/util/trace-event-info.c
>>> +++ b/tools/perf/util/trace-event-info.c
>>> @@ -48,7 +48,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
>>> int r, fd;
>>> int err = -EIO;
>>>
>>> - fd = open(file, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
>>> if (fd < 0) {
>>> pr_debug("Can't read '%s'", file);
>>> return -errno;
>>> @@ -61,7 +61,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
>>> }
>>>
>>> do {
>>> - r = read(fd, buf, BUFSIZ);
>>> + r = TEMP_FAILURE_RETRY(read(fd, buf, BUFSIZ));
>>> if (r > 0) {
>>> size += r;
>>> if (write(output_fd, buf, r) != r)
>>> @@ -519,7 +519,7 @@ struct tracing_data *tracing_data_get(struct list_head *pattrs,
>>> return NULL;
>>> }
>>>
>>> - temp_fd = open(tdata->temp_file, O_RDWR);
>>> + temp_fd = TEMP_FAILURE_RETRY(open(tdata->temp_file, O_RDWR));
>>> if (temp_fd < 0) {
>>> pr_debug("Can't read '%s'", tdata->temp_file);
>>> free(tdata);
>>> diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c
>>> index 8a01af783310..7086770ec225 100644
>>> --- a/tools/perf/util/trace-event-read.c
>>> +++ b/tools/perf/util/trace-event-read.c
>>> @@ -11,6 +11,7 @@
>>> #include <sys/stat.h>
>>> #include <sys/wait.h>
>>> #include <sys/mman.h>
>>> +#include <errno.h>
>>> #include <fcntl.h>
>>> #include <unistd.h>
>>> #include <errno.h>
>>> @@ -28,7 +29,7 @@ static int __do_read(int fd, void *buf, int size)
>>> int rsize = size;
>>>
>>> while (size) {
>>> - int ret = read(fd, buf, size);
>>> + int ret = TEMP_FAILURE_RETRY(read(fd, buf, size));
>>>
>>> if (ret <= 0)
>>> return -1;
>>> @@ -105,7 +106,7 @@ static char *read_string(void)
>>> char c;
>>>
>>> for (;;) {
>>> - r = read(input_fd, &c, 1);
>>> + r = TEMP_FAILURE_RETRY(read(input_fd, &c, 1));
>>> if (r < 0) {
>>> pr_debug("reading input file");
>>> goto out;
>>> diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c
>>> index 37622699c91a..2ece387a73c6 100644
>>> --- a/tools/perf/util/unwind-libunwind-local.c
>>> +++ b/tools/perf/util/unwind-libunwind-local.c
>>> @@ -362,7 +362,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
>>> }
>>>
>>> if (ofs <= 0) {
>>> - fd = open(dso->symsrc_filename, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(dso->symsrc_filename, O_RDONLY));
>>> if (fd >= 0) {
>>> ofs = elf_section_offset(fd, ".debug_frame");
>>> close(fd);
>>> @@ -377,7 +377,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
>>> dso, DSO_BINARY_TYPE__DEBUGLINK,
>>> machine->root_dir, debuglink, PATH_MAX);
>>> if (!ret) {
>>> - fd = open(debuglink, O_RDONLY);
>>> + fd = TEMP_FAILURE_RETRY(open(debuglink, O_RDONLY));
>>> if (fd >= 0) {
>>> ofs = elf_section_offset(fd,
>>> ".debug_frame");
>>> diff --git a/tools/perf/util/zlib.c b/tools/perf/util/zlib.c
>>> index 78d2297c1b67..f4563ea094c8 100644
>>> --- a/tools/perf/util/zlib.c
>>> +++ b/tools/perf/util/zlib.c
>>> @@ -1,4 +1,5 @@
>>> // SPDX-License-Identifier: GPL-2.0
>>> +#include <errno.h>
>>> #include <fcntl.h>
>>> #include <stdio.h>
>>> #include <string.h>
>>> @@ -82,7 +83,7 @@ int gzip_decompress_to_file(const char *input, int output_fd)
>>>
>>> bool gzip_is_compressed(const char *input)
>>> {
>>> - int fd = open(input, O_RDONLY);
>>> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
>>> const uint8_t magic[2] = { 0x1f, 0x8b };
>>> char buf[2] = { 0 };
>>> ssize_t rc;
>>> @@ -90,7 +91,7 @@ bool gzip_is_compressed(const char *input)
>>> if (fd < 0)
>>> return -1;
>>>
>>> - rc = read(fd, buf, sizeof(buf));
>>> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
>>> close(fd);
>>> return rc == sizeof(buf) ?
>>> memcmp(buf, magic, sizeof(buf)) == 0 : false;
>>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 16:40 ` Adrian Hunter
@ 2022-06-15 17:38 ` Ian Rogers
0 siblings, 0 replies; 7+ messages in thread
From: Ian Rogers @ 2022-06-15 17:38 UTC (permalink / raw)
To: Adrian Hunter
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Alexey Bayduraev, Riccardo Mancini, Dave Marchevsky, Song Liu,
Leo Yan, John Garry, Florian Fischer, Andi Kleen, Kan Liang,
Masami Hiramatsu, Zechuan Chen, Yang Jihong, German Gomez,
Thomas Richter, James Clark, Denis Nikitin, Lexi Shao,
Michael Petlan, Remi Bernon, Fangrui Song, linux-kernel,
linux-perf-users, Stephane Eranian
On Wed, Jun 15, 2022 at 9:40 AM Adrian Hunter <adrian.hunter@intel.com> wrote:
>
> On 15/06/22 18:34, Ian Rogers wrote:
> > On Tue, Jun 14, 2022 at 11:51 PM Adrian Hunter <adrian.hunter@intel.com> wrote:
> >>
> >> On 15/06/22 09:03, Ian Rogers wrote:
> >>> Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
> >>> causes the syscall to need to restart.
> >>
> >> Please expand on why you want this change. Can you give an example
> >> of where it makes a difference?
> >
> > As we can get signals pretty much anywhere in the code many syscalls
> > are robust to this:
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n311
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n372
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n1585
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-record.c?h=perf/core#n2502
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/builtin-top.c?h=perf/core#n707
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/copyfile.c?h=perf/core#n59
> > https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/evlist.c?h=perf/core#n1966
> >
> > But as this change shows there are a large number that are not. They
> > will get EINTR and turn it into a failure which really isn't one.
> > Making an example of this is near impossible as you need to arrange
> > the signal delivery very carefully. I am however motivated to do this
> > as we're seeing missing events when running perf in fleetwide
> > profiling and I wish to discount EINTR as a cause.
>
> Aren't most signals set to Ign or Term? What signals could you be getting?
>
> Is signal() setting SA_RESTART by default? Does that make a difference?
My understanding is that signal will specify SA_RESTART but sigaction
won't, so cases like:
https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/tree/tools/perf/util/evlist.c?h=perf/core#n1475
will cause EINTR. The existing code is partially EINTR resistant and
so I'm looking to see this be the norm. I think having a mix is
confusing, perhaps existing EINTR usage could be turned into asserts
to discover when this does happen. In general I think it is better to
be robust, which will matter more in contexts like libperf.
Thanks,
Ian
> >
> >> It looks pretty ugly so maybe making replacement functions or macros
> >> is better e.g.
> >>
> >> #define open_retry(a, ...) \
> >> ({ long int __result; \
> >> do __result = (long int) open((a), ##__VA_ARGS__); \
> >> while (__result == -1L && errno == EINTR); \
> >> __result; })
> >
> > As open is a user land thing there is no Linux precedent to follow
> > here. TEMP_FAILURE_RETRY is the general norm, and I'd like to see it
> > used when a raw system call is made. ie, if you want open_retry:
> >
> > #define open_retry(a, ...) TEMP_FAILURE_RETRY(open((a), ##__VA_ARGS__))
> >
> > I don't think we need to be in the game of reinventing
> > TEMP_FAILURE_RETRY.
>
> It was only an example. TEMP_FAILURE_RETRY seemed non-standard,
> but I guess if we're defining _GNU_SOURCE anyway it doesn't matter.
>
> > I think we can follow on from this change with
> > changes not to use raw open and read calls, instead preferring some of
> > our existing EINTR robust reads like readn, etc. Doing that is beyond
> > the scope of what this change is trying to do.
> >
> > Thanks,
> > Ian
> >
> >>>
> >>> Signed-off-by: Ian Rogers <irogers@google.com>
> >>> ---
> >>> tools/lib/api/io.h | 3 ++-
> >>> tools/lib/symbol/kallsyms.c | 4 ++-
> >>> tools/perf/builtin-daemon.c | 19 +++++++-------
> >>> tools/perf/builtin-ftrace.c | 18 ++++++-------
> >>> tools/perf/builtin-inject.c | 2 +-
> >>> tools/perf/builtin-kvm.c | 2 +-
> >>> tools/perf/builtin-record.c | 6 ++---
> >>> tools/perf/builtin-sched.c | 2 +-
> >>> tools/perf/builtin-script.c | 2 +-
> >>> tools/perf/util/copyfile.c | 2 +-
> >>> tools/perf/util/data.c | 11 ++++----
> >>> tools/perf/util/dso.c | 6 ++---
> >>> tools/perf/util/evlist.c | 6 ++---
> >>> tools/perf/util/lzma.c | 4 +--
> >>> tools/perf/util/namespaces.c | 5 ++--
> >>> tools/perf/util/parse-events.c | 6 ++---
> >>> tools/perf/util/pmu.c | 12 ++++-----
> >>> tools/perf/util/probe-event.c | 4 +--
> >>> tools/perf/util/probe-file.c | 10 ++++----
> >>> tools/perf/util/session.c | 2 +-
> >>> tools/perf/util/symbol-elf.c | 32 ++++++++++++------------
> >>> tools/perf/util/symbol-minimal.c | 6 ++---
> >>> tools/perf/util/symbol.c | 4 +--
> >>> tools/perf/util/synthetic-events.c | 7 +++---
> >>> tools/perf/util/trace-event-info.c | 6 ++---
> >>> tools/perf/util/trace-event-read.c | 5 ++--
> >>> tools/perf/util/unwind-libunwind-local.c | 4 +--
> >>> tools/perf/util/zlib.c | 5 ++--
> >>> 28 files changed, 102 insertions(+), 93 deletions(-)
> >>>
> >>> diff --git a/tools/lib/api/io.h b/tools/lib/api/io.h
> >>> index 777c20f6b604..b292bbd39fb0 100644
> >>> --- a/tools/lib/api/io.h
> >>> +++ b/tools/lib/api/io.h
> >>> @@ -7,6 +7,7 @@
> >>> #ifndef __API_IO__
> >>> #define __API_IO__
> >>>
> >>> +#include <errno.h>
> >>> #include <stdlib.h>
> >>> #include <unistd.h>
> >>>
> >>> @@ -45,7 +46,7 @@ static inline int io__get_char(struct io *io)
> >>> return -1;
> >>>
> >>> if (ptr == io->end) {
> >>> - ssize_t n = read(io->fd, io->buf, io->buf_len);
> >>> + ssize_t n = TEMP_FAILURE_RETRY(read(io->fd, io->buf, io->buf_len));
> >>>
> >>> if (n <= 0) {
> >>> io->eof = true;
> >>> diff --git a/tools/lib/symbol/kallsyms.c b/tools/lib/symbol/kallsyms.c
> >>> index e335ac2b9e19..5436c038c379 100644
> >>> --- a/tools/lib/symbol/kallsyms.c
> >>> +++ b/tools/lib/symbol/kallsyms.c
> >>> @@ -3,7 +3,9 @@
> >>> #include "api/io.h"
> >>> #include <stdio.h>
> >>> #include <sys/stat.h>
> >>> +#include <errno.h>
> >>> #include <fcntl.h>
> >>> +#include <unistd.h>
> >>>
> >>> u8 kallsyms2elf_type(char type)
> >>> {
> >>> @@ -36,7 +38,7 @@ int kallsyms__parse(const char *filename, void *arg,
> >>> char bf[BUFSIZ];
> >>> int err;
> >>>
> >>> - io.fd = open(filename, O_RDONLY, 0);
> >>> + io.fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY, 0));
> >>>
> >>> if (io.fd < 0)
> >>> return -1;
> >>> diff --git a/tools/perf/builtin-daemon.c b/tools/perf/builtin-daemon.c
> >>> index 6cb3f6cc36d0..d4e1a5636056 100644
> >>> --- a/tools/perf/builtin-daemon.c
> >>> +++ b/tools/perf/builtin-daemon.c
> >>> @@ -351,7 +351,7 @@ static int daemon_session__run(struct daemon_session *session,
> >>> return -1;
> >>> }
> >>>
> >>> - fd = open("/dev/null", O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open("/dev/null", O_RDONLY));
> >>> if (fd < 0) {
> >>> perror("failed: open /dev/null");
> >>> return -1;
> >>> @@ -360,7 +360,7 @@ static int daemon_session__run(struct daemon_session *session,
> >>> dup2(fd, 0);
> >>> close(fd);
> >>>
> >>> - fd = open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644);
> >>> + fd = TEMP_FAILURE_RETRY(open(SESSION_OUTPUT, O_RDWR|O_CREAT|O_TRUNC, 0644));
> >>> if (fd < 0) {
> >>> perror("failed: open session output");
> >>> return -1;
> >>> @@ -405,7 +405,8 @@ static pid_t handle_signalfd(struct daemon *daemon)
> >>> * coalesced in kernel and we can receive only single signal even
> >>> * if multiple SIGCHLD were generated.
> >>> */
> >>> - err = read(daemon->signal_fd, &si, sizeof(struct signalfd_siginfo));
> >>> + err = TEMP_FAILURE_RETRY(read(daemon->signal_fd, &si,
> >>> + sizeof(struct signalfd_siginfo)));
> >>> if (err != sizeof(struct signalfd_siginfo)) {
> >>> pr_err("failed to read signal fd\n");
> >>> return -1;
> >>> @@ -522,7 +523,7 @@ static int daemon_session__control(struct daemon_session *session,
> >>> scnprintf(control_path, sizeof(control_path), "%s/%s",
> >>> session->base, SESSION_CONTROL);
> >>>
> >>> - control = open(control_path, O_WRONLY|O_NONBLOCK);
> >>> + control = TEMP_FAILURE_RETRY(open(control_path, O_WRONLY|O_NONBLOCK));
> >>> if (!control)
> >>> return -1;
> >>>
> >>> @@ -531,7 +532,7 @@ static int daemon_session__control(struct daemon_session *session,
> >>> scnprintf(ack_path, sizeof(ack_path), "%s/%s",
> >>> session->base, SESSION_ACK);
> >>>
> >>> - ack = open(ack_path, O_RDONLY, O_NONBLOCK);
> >>> + ack = TEMP_FAILURE_RETRY(open(ack_path, O_RDONLY, O_NONBLOCK));
> >>> if (!ack) {
> >>> close(control);
> >>> return -1;
> >>> @@ -564,7 +565,7 @@ static int daemon_session__control(struct daemon_session *session,
> >>> goto out;
> >>> }
> >>>
> >>> - err = read(ack, buf, sizeof(buf));
> >>> + err = TEMP_FAILURE_RETRY(read(ack, buf, sizeof(buf)));
> >>> if (err > 0)
> >>> ret = strcmp(buf, "ack\n");
> >>> else
> >>> @@ -1078,7 +1079,7 @@ static int handle_config_changes(struct daemon *daemon, int conf_fd,
> >>> ssize_t len;
> >>>
> >>> while (!(*config_changed)) {
> >>> - len = read(conf_fd, buf, sizeof(buf));
> >>> + len = TEMP_FAILURE_RETRY(read(conf_fd, buf, sizeof(buf)));
> >>> if (len == -1) {
> >>> if (errno != EAGAIN) {
> >>> perror("failed: read");
> >>> @@ -1147,7 +1148,7 @@ static int check_lock(struct daemon *daemon)
> >>>
> >>> scnprintf(path, sizeof(path), "%s/lock", daemon->base);
> >>>
> >>> - fd = open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640);
> >>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDWR|O_CREAT|O_CLOEXEC, 0640));
> >>> if (fd < 0)
> >>> return -1;
> >>>
> >>> @@ -1201,7 +1202,7 @@ static int go_background(struct daemon *daemon)
> >>> return -1;
> >>> }
> >>>
> >>> - fd = open("output", O_RDWR|O_CREAT|O_TRUNC, 0644);
> >>> + fd = TEMP_FAILURE_RETRY(open("output", O_RDWR|O_CREAT|O_TRUNC, 0644));
> >>> if (fd < 0) {
> >>> perror("failed: open");
> >>> return -1;
> >>> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
> >>> index 7de07bb16d23..7553e20f38b2 100644
> >>> --- a/tools/perf/builtin-ftrace.c
> >>> +++ b/tools/perf/builtin-ftrace.c
> >>> @@ -79,7 +79,7 @@ static int __write_tracing_file(const char *name, const char *val, bool append)
> >>> else
> >>> flags |= O_TRUNC;
> >>>
> >>> - fd = open(file, flags);
> >>> + fd = TEMP_FAILURE_RETRY(open(file, flags));
> >>> if (fd < 0) {
> >>> pr_debug("cannot open tracing file: %s: %s\n",
> >>> name, str_error_r(errno, errbuf, sizeof(errbuf)));
> >>> @@ -132,7 +132,7 @@ static int read_tracing_file_to_stdout(const char *name)
> >>> return -1;
> >>> }
> >>>
> >>> - fd = open(file, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> >>> if (fd < 0) {
> >>> pr_debug("cannot open tracing file: %s: %s\n",
> >>> name, str_error_r(errno, buf, sizeof(buf)));
> >>> @@ -141,7 +141,7 @@ static int read_tracing_file_to_stdout(const char *name)
> >>>
> >>> /* read contents to stdout */
> >>> while (true) {
> >>> - int n = read(fd, buf, sizeof(buf));
> >>> + int n = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> >>> if (n == 0)
> >>> break;
> >>> else if (n < 0)
> >>> @@ -608,7 +608,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> >>> goto out_reset;
> >>> }
> >>>
> >>> - trace_fd = open(trace_file, O_RDONLY);
> >>> + trace_fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
> >>>
> >>> put_tracing_file(trace_file);
> >>>
> >>> @@ -645,7 +645,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> >>> break;
> >>>
> >>> if (pollfd.revents & POLLIN) {
> >>> - int n = read(trace_fd, buf, sizeof(buf));
> >>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> >>> if (n < 0)
> >>> break;
> >>> if (fwrite(buf, n, 1, stdout) != 1)
> >>> @@ -665,7 +665,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace)
> >>>
> >>> /* read remaining buffer contents */
> >>> while (true) {
> >>> - int n = read(trace_fd, buf, sizeof(buf));
> >>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf)));
> >>> if (n <= 0)
> >>> break;
> >>> if (fwrite(buf, n, 1, stdout) != 1)
> >>> @@ -826,7 +826,7 @@ static int prepare_func_latency(struct perf_ftrace *ftrace)
> >>> return -1;
> >>> }
> >>>
> >>> - fd = open(trace_file, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(trace_file, O_RDONLY));
> >>> if (fd < 0)
> >>> pr_err("failed to open trace_pipe\n");
> >>>
> >>> @@ -913,7 +913,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
> >>> break;
> >>>
> >>> if (pollfd.revents & POLLIN) {
> >>> - int n = read(trace_fd, buf, sizeof(buf) - 1);
> >>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> >>> if (n < 0)
> >>> break;
> >>>
> >>> @@ -931,7 +931,7 @@ static int __cmd_latency(struct perf_ftrace *ftrace)
> >>>
> >>> /* read remaining buffer contents */
> >>> while (!ftrace->target.use_bpf) {
> >>> - int n = read(trace_fd, buf, sizeof(buf) - 1);
> >>> + int n = TEMP_FAILURE_RETRY(read(trace_fd, buf, sizeof(buf) - 1));
> >>> if (n <= 0)
> >>> break;
> >>> make_histogram(buckets, buf, n, line, ftrace->use_nsec);
> >>> diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
> >>> index a75bf11585b5..960b4a50583e 100644
> >>> --- a/tools/perf/builtin-inject.c
> >>> +++ b/tools/perf/builtin-inject.c
> >>> @@ -155,7 +155,7 @@ static int copy_bytes(struct perf_inject *inject, int fd, off_t size)
> >>> int ret;
> >>>
> >>> while (size > 0) {
> >>> - ssz = read(fd, buf, min(size, (off_t)sizeof(buf)));
> >>> + ssz = TEMP_FAILURE_RETRY(read(fd, buf, min(size, (off_t)sizeof(buf))));
> >>> if (ssz < 0)
> >>> return -errno;
> >>> ret = output_bytes(inject, buf, ssz);
> >>> diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
> >>> index 3696ae97f149..4dd8292d1941 100644
> >>> --- a/tools/perf/builtin-kvm.c
> >>> +++ b/tools/perf/builtin-kvm.c
> >>> @@ -880,7 +880,7 @@ static int perf_kvm__handle_timerfd(struct perf_kvm_stat *kvm)
> >>> uint64_t c;
> >>> int rc;
> >>>
> >>> - rc = read(kvm->timerfd, &c, sizeof(uint64_t));
> >>> + rc = TEMP_FAILURE_RETRY(read(kvm->timerfd, &c, sizeof(uint64_t)));
> >>> if (rc < 0) {
> >>> if (errno == EAGAIN)
> >>> return 0;
> >>> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> >>> index 9a71f0330137..361610bc96b6 100644
> >>> --- a/tools/perf/builtin-record.c
> >>> +++ b/tools/perf/builtin-record.c
> >>> @@ -902,7 +902,7 @@ static bool record__kcore_readable(struct machine *machine)
> >>>
> >>> scnprintf(kcore, sizeof(kcore), "%s/proc/kcore", machine->root_dir);
> >>>
> >>> - fd = open(kcore, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(kcore, O_RDONLY));
> >>> if (fd < 0)
> >>> return false;
> >>>
> >>> @@ -2049,7 +2049,7 @@ static int record__terminate_thread(struct record_thread *thread_data)
> >>>
> >>> close(thread_data->pipes.msg[1]);
> >>> thread_data->pipes.msg[1] = -1;
> >>> - err = read(thread_data->pipes.ack[0], &ack, sizeof(ack));
> >>> + err = TEMP_FAILURE_RETRY(read(thread_data->pipes.ack[0], &ack, sizeof(ack)));
> >>> if (err > 0)
> >>> pr_debug2("threads[%d]: sent %s\n", tid, thread_msg_tags[ack]);
> >>> else
> >>> @@ -2097,7 +2097,7 @@ static int record__start_threads(struct record *rec)
> >>> goto out_err;
> >>> }
> >>>
> >>> - err = read(thread_data[t].pipes.ack[0], &msg, sizeof(msg));
> >>> + err = TEMP_FAILURE_RETRY(read(thread_data[t].pipes.ack[0], &msg, sizeof(msg)));
> >>> if (err > 0)
> >>> pr_debug2("threads[%d]: sent %s\n", rec->thread_data[t].tid,
> >>> thread_msg_tags[msg]);
> >>> diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
> >>> index 646bd938927a..149787268d02 100644
> >>> --- a/tools/perf/builtin-sched.c
> >>> +++ b/tools/perf/builtin-sched.c
> >>> @@ -604,7 +604,7 @@ static u64 get_cpu_usage_nsec_self(int fd)
> >>> u64 runtime;
> >>> int ret;
> >>>
> >>> - ret = read(fd, &runtime, sizeof(runtime));
> >>> + ret = TEMP_FAILURE_RETRY(read(fd, &runtime, sizeof(runtime)));
> >>> BUG_ON(ret != sizeof(runtime));
> >>>
> >>> return runtime;
> >>> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> >>> index c689054002cc..844f5350bfdf 100644
> >>> --- a/tools/perf/builtin-script.c
> >>> +++ b/tools/perf/builtin-script.c
> >>> @@ -4207,7 +4207,7 @@ int cmd_script(int argc, const char **argv)
> >>> goto out_delete;
> >>> }
> >>>
> >>> - input = open(data.path, O_RDONLY); /* input_name */
> >>> + input = TEMP_FAILURE_RETRY(open(data.path, O_RDONLY)); /* input_name */
> >>> if (input < 0) {
> >>> err = -errno;
> >>> perror("failed to open file");
> >>> diff --git a/tools/perf/util/copyfile.c b/tools/perf/util/copyfile.c
> >>> index 47e03de7c235..c7bff4943936 100644
> >>> --- a/tools/perf/util/copyfile.c
> >>> +++ b/tools/perf/util/copyfile.c
> >>> @@ -112,7 +112,7 @@ static int copyfile_mode_ns(const char *from, const char *to, mode_t mode,
> >>> goto out_close_to;
> >>>
> >>> nsinfo__mountns_enter(nsi, &nsc);
> >>> - fromfd = open(from, O_RDONLY);
> >>> + fromfd = TEMP_FAILURE_RETRY(open(from, O_RDONLY));
> >>> nsinfo__mountns_exit(&nsc);
> >>> if (fromfd < 0)
> >>> goto out_close_to;
> >>> diff --git a/tools/perf/util/data.c b/tools/perf/util/data.c
> >>> index caabeac24c69..a9066a3d914f 100644
> >>> --- a/tools/perf/util/data.c
> >>> +++ b/tools/perf/util/data.c
> >>> @@ -53,7 +53,8 @@ int perf_data__create_dir(struct perf_data *data, int nr)
> >>> goto out_err;
> >>> }
> >>>
> >>> - ret = open(file->path, O_RDWR|O_CREAT|O_TRUNC, S_IRUSR|S_IWUSR);
> >>> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDWR|O_CREAT|O_TRUNC,
> >>> + S_IRUSR|S_IWUSR));
> >>> if (ret < 0) {
> >>> ret = -errno;
> >>> goto out_err;
> >>> @@ -123,7 +124,7 @@ int perf_data__open_dir(struct perf_data *data)
> >>> if (!file->path)
> >>> goto out_err;
> >>>
> >>> - ret = open(file->path, O_RDONLY);
> >>> + ret = TEMP_FAILURE_RETRY(open(file->path, O_RDONLY));
> >>> if (ret < 0)
> >>> goto out_err;
> >>>
> >>> @@ -248,7 +249,7 @@ static int open_file_read(struct perf_data *data)
> >>> int fd;
> >>> char sbuf[STRERR_BUFSIZE];
> >>>
> >>> - fd = open(data->file.path, flags);
> >>> + fd = TEMP_FAILURE_RETRY(open(data->file.path, flags));
> >>> if (fd < 0) {
> >>> int err = errno;
> >>>
> >>> @@ -288,8 +289,8 @@ static int open_file_write(struct perf_data *data)
> >>> int fd;
> >>> char sbuf[STRERR_BUFSIZE];
> >>>
> >>> - fd = open(data->file.path, O_CREAT|O_RDWR|O_TRUNC|O_CLOEXEC,
> >>> - S_IRUSR|S_IWUSR);
> >>> + fd = TEMP_FAILURE_RETRY(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,
> >>> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
> >>> index 5ac13958d1bd..40befa2429f9 100644
> >>> --- a/tools/perf/util/dso.c
> >>> +++ b/tools/perf/util/dso.c
> >>> @@ -300,7 +300,7 @@ int filename__decompress(const char *name, char *pathname,
> >>> * descriptor to the uncompressed file.
> >>> */
> >>> if (!compressions[comp].is_compressed(name))
> >>> - return open(name, O_RDONLY);
> >>> + return TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> >>>
> >>> fd = mkstemp(tmpbuf);
> >>> if (fd < 0) {
> >>> @@ -476,7 +476,7 @@ static int do_open(char *name)
> >>> char sbuf[STRERR_BUFSIZE];
> >>>
> >>> do {
> >>> - fd = open(name, O_RDONLY|O_CLOEXEC);
> >>> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY|O_CLOEXEC));
> >>> if (fd >= 0)
> >>> return fd;
> >>>
> >>> @@ -898,7 +898,7 @@ static ssize_t file_read(struct dso *dso, struct machine *machine,
> >>> goto out;
> >>> }
> >>>
> >>> - ret = pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset);
> >>> + ret = TEMP_FAILURE_RETRY(pread(dso->data.fd, data, DSO__DATA_CACHE_SIZE, offset));
> >>> out:
> >>> pthread_mutex_unlock(&dso__data_open_lock);
> >>> return ret;
> >>> diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> >>> index 48af7d379d82..ba0dcdf91120 100644
> >>> --- a/tools/perf/util/evlist.c
> >>> +++ b/tools/perf/util/evlist.c
> >>> @@ -1440,7 +1440,7 @@ int evlist__prepare_workload(struct evlist *evlist, struct target *target, const
> >>> /*
> >>> * Wait until the parent tells us to go.
> >>> */
> >>> - ret = read(go_pipe[0], &bf, 1);
> >>> + ret = TEMP_FAILURE_RETRY(read(go_pipe[0], &bf, 1));
> >>> /*
> >>> * The parent will ask for the execvp() to be performed by
> >>> * writing exactly one byte, in workload.cork_fd, usually via
> >>> @@ -1837,7 +1837,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
> >>> * O_RDWR avoids POLLHUPs which is necessary to allow the other
> >>> * end of a FIFO to be repeatedly opened and closed.
> >>> */
> >>> - fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> >>> + fd = TEMP_FAILURE_RETRY(open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> >>> if (fd < 0) {
> >>> pr_err("Failed to open '%s'\n", s);
> >>> ret = -errno;
> >>> @@ -1848,7 +1848,7 @@ static int evlist__parse_control_fifo(const char *str, int *ctl_fd, int *ctl_fd_
> >>>
> >>> if (p && *++p) {
> >>> /* O_RDWR | O_NONBLOCK means the other end need not be open */
> >>> - fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC);
> >>> + fd = TEMP_FAILURE_RETRY(open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC));
> >>> if (fd < 0) {
> >>> pr_err("Failed to open '%s'\n", p);
> >>> ret = -errno;
> >>> diff --git a/tools/perf/util/lzma.c b/tools/perf/util/lzma.c
> >>> index 51424cdc3b68..93e0a6a53e28 100644
> >>> --- a/tools/perf/util/lzma.c
> >>> +++ b/tools/perf/util/lzma.c
> >>> @@ -109,7 +109,7 @@ int lzma_decompress_to_file(const char *input, int output_fd)
> >>>
> >>> bool lzma_is_compressed(const char *input)
> >>> {
> >>> - int fd = open(input, O_RDONLY);
> >>> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> >>> const uint8_t magic[6] = { 0xFD, '7', 'z', 'X', 'Z', 0x00 };
> >>> char buf[6] = { 0 };
> >>> ssize_t rc;
> >>> @@ -117,7 +117,7 @@ bool lzma_is_compressed(const char *input)
> >>> if (fd < 0)
> >>> return -1;
> >>>
> >>> - rc = read(fd, buf, sizeof(buf));
> >>> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> >>> close(fd);
> >>> return rc == sizeof(buf) ?
> >>> memcmp(buf, magic, sizeof(buf)) == 0 : false;
> >>> diff --git a/tools/perf/util/namespaces.c b/tools/perf/util/namespaces.c
> >>> index dd536220cdb9..89edfba6af06 100644
> >>> --- a/tools/perf/util/namespaces.c
> >>> +++ b/tools/perf/util/namespaces.c
> >>> @@ -9,6 +9,7 @@
> >>> #include "get_current_dir_name.h"
> >>> #include <sys/types.h>
> >>> #include <sys/stat.h>
> >>> +#include <errno.h>
> >>> #include <fcntl.h>
> >>> #include <limits.h>
> >>> #include <sched.h>
> >>> @@ -266,11 +267,11 @@ void nsinfo__mountns_enter(struct nsinfo *nsi,
> >>> if (!oldcwd)
> >>> return;
> >>>
> >>> - oldns = open(curpath, O_RDONLY);
> >>> + oldns = TEMP_FAILURE_RETRY(open(curpath, O_RDONLY));
> >>> if (oldns < 0)
> >>> goto errout;
> >>>
> >>> - newns = open(nsi->mntns_path, O_RDONLY);
> >>> + newns = TEMP_FAILURE_RETRY(open(nsi->mntns_path, O_RDONLY));
> >>> if (newns < 0)
> >>> goto errout;
> >>>
> >>> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> >>> index 7ed235740431..75f8f2e44195 100644
> >>> --- a/tools/perf/util/parse-events.c
> >>> +++ b/tools/perf/util/parse-events.c
> >>> @@ -189,7 +189,7 @@ static int tp_event_has_id(const char *dir_path, struct dirent *evt_dir)
> >>> int fd;
> >>>
> >>> snprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path, evt_dir->d_name);
> >>> - fd = open(evt_path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> >>> if (fd < 0)
> >>> return -EINVAL;
> >>> close(fd);
> >>> @@ -233,10 +233,10 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config)
> >>>
> >>> scnprintf(evt_path, MAXPATHLEN, "%s/%s/id", dir_path,
> >>> evt_dirent->d_name);
> >>> - fd = open(evt_path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(evt_path, O_RDONLY));
> >>> if (fd < 0)
> >>> continue;
> >>> - if (read(fd, id_buf, sizeof(id_buf)) < 0) {
> >>> + if (TEMP_FAILURE_RETRY(read(fd, id_buf, sizeof(id_buf))) < 0) {
> >>> close(fd);
> >>> continue;
> >>> }
> >>> diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> >>> index 9a1c7e63e663..59252cfc2b5b 100644
> >>> --- a/tools/perf/util/pmu.c
> >>> +++ b/tools/perf/util/pmu.c
> >>> @@ -154,14 +154,14 @@ static int perf_pmu__parse_scale(struct perf_pmu_alias *alias, char *dir, char *
> >>>
> >>> scnprintf(path, PATH_MAX, "%s/%s.scale", dir, name);
> >>>
> >>> - fd = open(path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> >>> if (fd == -1)
> >>> return -1;
> >>>
> >>> if (fstat(fd, &st) < 0)
> >>> goto error;
> >>>
> >>> - sret = read(fd, scale, sizeof(scale)-1);
> >>> + sret = TEMP_FAILURE_RETRY(read(fd, scale, sizeof(scale)-1));
> >>> if (sret < 0)
> >>> goto error;
> >>>
> >>> @@ -184,11 +184,11 @@ static int perf_pmu__parse_unit(struct perf_pmu_alias *alias, char *dir, char *n
> >>>
> >>> scnprintf(path, PATH_MAX, "%s/%s.unit", dir, name);
> >>>
> >>> - fd = open(path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> >>> if (fd == -1)
> >>> return -1;
> >>>
> >>> - sret = read(fd, alias->unit, UNIT_MAX_LEN);
> >>> + sret = TEMP_FAILURE_RETRY(read(fd, alias->unit, UNIT_MAX_LEN));
> >>> if (sret < 0)
> >>> goto error;
> >>>
> >>> @@ -214,7 +214,7 @@ perf_pmu__parse_per_pkg(struct perf_pmu_alias *alias, char *dir, char *name)
> >>>
> >>> scnprintf(path, PATH_MAX, "%s/%s.per-pkg", dir, name);
> >>>
> >>> - fd = open(path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> >>> if (fd == -1)
> >>> return -1;
> >>>
> >>> @@ -232,7 +232,7 @@ static int perf_pmu__parse_snapshot(struct perf_pmu_alias *alias,
> >>>
> >>> scnprintf(path, PATH_MAX, "%s/%s.snapshot", dir, name);
> >>>
> >>> - fd = open(path, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(path, O_RDONLY));
> >>> if (fd == -1)
> >>> return -1;
> >>>
> >>> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> >>> index 062b5cbe67af..7838889f55df 100644
> >>> --- a/tools/perf/util/probe-event.c
> >>> +++ b/tools/perf/util/probe-event.c
> >>> @@ -276,7 +276,7 @@ static char *find_module_name(const char *module)
> >>> char *mod_name = NULL;
> >>> int name_offset;
> >>>
> >>> - fd = open(module, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(module, O_RDONLY));
> >>> if (fd < 0)
> >>> return NULL;
> >>>
> >>> @@ -598,7 +598,7 @@ static int get_text_start_address(const char *exec, u64 *address,
> >>> struct nscookie nsc;
> >>>
> >>> nsinfo__mountns_enter(nsi, &nsc);
> >>> - fd = open(exec, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(exec, O_RDONLY));
> >>> nsinfo__mountns_exit(&nsc);
> >>> if (fd < 0)
> >>> return -errno;
> >>> diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
> >>> index 3d50de3217d5..b788932656e8 100644
> >>> --- a/tools/perf/util/probe-file.c
> >>> +++ b/tools/perf/util/probe-file.c
> >>> @@ -115,9 +115,9 @@ int open_trace_file(const char *trace_file, bool readwrite)
> >>> if (ret >= 0) {
> >>> pr_debug("Opening %s write=%d\n", buf, readwrite);
> >>> if (readwrite && !probe_event_dry_run)
> >>> - ret = open(buf, O_RDWR | O_APPEND, 0);
> >>> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDWR | O_APPEND, 0));
> >>> else
> >>> - ret = open(buf, O_RDONLY, 0);
> >>> + ret = TEMP_FAILURE_RETRY(open(buf, O_RDONLY, 0));
> >>>
> >>> if (ret < 0)
> >>> ret = -errno;
> >>> @@ -180,7 +180,7 @@ struct strlist *probe_file__get_rawlist(int fd)
> >>> if (sl == NULL)
> >>> return NULL;
> >>>
> >>> - fddup = dup(fd);
> >>> + fddup = TEMP_FAILURE_RETRY(dup(fd));
> >>> if (fddup < 0)
> >>> goto out_free_sl;
> >>>
> >>> @@ -498,7 +498,7 @@ static int probe_cache__open(struct probe_cache *pcache, const char *target,
> >>> }
> >>>
> >>> snprintf(cpath, PATH_MAX, "%s/probes", dir_name);
> >>> - fd = open(cpath, O_CREAT | O_RDWR, 0644);
> >>> + fd = TEMP_FAILURE_RETRY(open(cpath, O_CREAT | O_RDWR, 0644));
> >>> if (fd < 0)
> >>> pr_debug("Failed to open cache(%d): %s\n", fd, cpath);
> >>> free(dir_name);
> >>> @@ -514,7 +514,7 @@ static int probe_cache__load(struct probe_cache *pcache)
> >>> int ret = 0, fddup;
> >>> FILE *fp;
> >>>
> >>> - fddup = dup(pcache->fd);
> >>> + fddup = TEMP_FAILURE_RETRY(dup(pcache->fd));
> >>> if (fddup < 0)
> >>> return -errno;
> >>> fp = fdopen(fddup, "r");
> >>> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> >>> index 0acb9de54b06..b2fc389cc530 100644
> >>> --- a/tools/perf/util/session.c
> >>> +++ b/tools/perf/util/session.c
> >>> @@ -384,7 +384,7 @@ static int skipn(int fd, off_t n)
> >>> ssize_t ret;
> >>>
> >>> while (n > 0) {
> >>> - ret = read(fd, buf, min(n, (off_t)sizeof(buf)));
> >>> + ret = TEMP_FAILURE_RETRY(read(fd, buf, min(n, (off_t)sizeof(buf))));
> >>> if (ret <= 0)
> >>> return ret;
> >>> n -= ret;
> >>> diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
> >>> index ecd377938eea..02556070a2ce 100644
> >>> --- a/tools/perf/util/symbol-elf.c
> >>> +++ b/tools/perf/util/symbol-elf.c
> >>> @@ -577,7 +577,7 @@ static int read_build_id(const char *filename, struct build_id *bid)
> >>> if (size < BUILD_ID_SIZE)
> >>> goto out;
> >>>
> >>> - fd = open(filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> >>> if (fd < 0)
> >>> goto out;
> >>>
> >>> @@ -638,7 +638,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> size_t size = sizeof(bid->data);
> >>> int fd, err = -1;
> >>>
> >>> - fd = open(filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> >>> if (fd < 0)
> >>> goto out;
> >>>
> >>> @@ -647,24 +647,24 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> GElf_Nhdr nhdr;
> >>> size_t namesz, descsz;
> >>>
> >>> - if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr))
> >>> + if (TEMP_FAILURE_RETRY(read(fd, &nhdr, sizeof(nhdr))) != sizeof(nhdr))
> >>> break;
> >>>
> >>> namesz = NOTE_ALIGN(nhdr.n_namesz);
> >>> descsz = NOTE_ALIGN(nhdr.n_descsz);
> >>> if (nhdr.n_type == NT_GNU_BUILD_ID &&
> >>> nhdr.n_namesz == sizeof("GNU")) {
> >>> - if (read(fd, bf, namesz) != (ssize_t)namesz)
> >>> + if (TEMP_FAILURE_RETRY(read(fd, bf, namesz)) != (ssize_t)namesz)
> >>> break;
> >>> if (memcmp(bf, "GNU", sizeof("GNU")) == 0) {
> >>> size_t sz = min(descsz, size);
> >>> - if (read(fd, bid->data, sz) == (ssize_t)sz) {
> >>> + if (TEMP_FAILURE_RETRY(read(fd, bid->data, sz)) == (ssize_t)sz) {
> >>> memset(bid->data + sz, 0, size - sz);
> >>> bid->size = sz;
> >>> err = 0;
> >>> break;
> >>> }
> >>> - } else if (read(fd, bf, descsz) != (ssize_t)descsz)
> >>> + } else if (TEMP_FAILURE_RETRY(read(fd, bf, descsz)) != (ssize_t)descsz)
> >>> break;
> >>> } else {
> >>> int n = namesz + descsz;
> >>> @@ -674,7 +674,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
> >>> __func__, filename, nhdr.n_namesz, nhdr.n_descsz);
> >>> }
> >>> - if (read(fd, bf, n) != n)
> >>> + if (TEMP_FAILURE_RETRY(read(fd, bf, n)) != n)
> >>> break;
> >>> }
> >>> }
> >>> @@ -732,7 +732,7 @@ int filename__read_debuglink(const char *filename, char *debuglink,
> >>> Elf_Scn *sec;
> >>> Elf_Kind ek;
> >>>
> >>> - fd = open(filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> >>> if (fd < 0)
> >>> goto out;
> >>>
> >>> @@ -844,7 +844,7 @@ int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> >>>
> >>> type = dso->symtab_type;
> >>> } else {
> >>> - fd = open(name, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> >>> if (fd < 0) {
> >>> dso->load_errno = errno;
> >>> return -1;
> >>> @@ -1454,7 +1454,7 @@ static int copy_bytes(int from, off_t from_offs, int to, off_t to_offs, u64 len)
> >>> if (len < n)
> >>> n = len;
> >>> /* Use read because mmap won't work on proc files */
> >>> - r = read(from, buf, n);
> >>> + r = TEMP_FAILURE_RETRY(read(from, buf, n));
> >>> if (r < 0)
> >>> goto out;
> >>> if (!r)
> >>> @@ -1485,7 +1485,7 @@ static int kcore__open(struct kcore *kcore, const char *filename)
> >>> {
> >>> GElf_Ehdr *ehdr;
> >>>
> >>> - kcore->fd = open(filename, O_RDONLY);
> >>> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> >>> if (kcore->fd == -1)
> >>> return -1;
> >>>
> >>> @@ -1518,7 +1518,7 @@ static int kcore__init(struct kcore *kcore, char *filename, int elfclass,
> >>> if (temp)
> >>> kcore->fd = mkstemp(filename);
> >>> else
> >>> - kcore->fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400);
> >>> + kcore->fd = TEMP_FAILURE_RETRY(open(filename, O_WRONLY | O_CREAT | O_EXCL, 0400));
> >>> if (kcore->fd == -1)
> >>> return -1;
> >>>
> >>> @@ -1966,7 +1966,7 @@ static int kcore_copy__compare_fds(int from, int to)
> >>>
> >>> while (1) {
> >>> /* Use read because mmap won't work on proc files */
> >>> - ret = read(from, buf_from, page_size);
> >>> + ret = TEMP_FAILURE_RETRY(read(from, buf_from, page_size));
> >>> if (ret < 0)
> >>> goto out;
> >>>
> >>> @@ -1994,11 +1994,11 @@ static int kcore_copy__compare_files(const char *from_filename,
> >>> {
> >>> int from, to, err = -1;
> >>>
> >>> - from = open(from_filename, O_RDONLY);
> >>> + from = TEMP_FAILURE_RETRY(open(from_filename, O_RDONLY));
> >>> if (from < 0)
> >>> return -1;
> >>>
> >>> - to = open(to_filename, O_RDONLY);
> >>> + to = TEMP_FAILURE_RETRY(open(to_filename, O_RDONLY));
> >>> if (to < 0)
> >>> goto out_close_from;
> >>>
> >>> @@ -2419,7 +2419,7 @@ int get_sdt_note_list(struct list_head *head, const char *target)
> >>> Elf *elf;
> >>> int fd, ret;
> >>>
> >>> - fd = open(target, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(target, O_RDONLY));
> >>> if (fd < 0)
> >>> return -EBADF;
> >>>
> >>> diff --git a/tools/perf/util/symbol-minimal.c b/tools/perf/util/symbol-minimal.c
> >>> index f9eb0bee7f15..7e25d476b1b2 100644
> >>> --- a/tools/perf/util/symbol-minimal.c
> >>> +++ b/tools/perf/util/symbol-minimal.c
> >>> @@ -230,7 +230,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> size_t buf_size;
> >>> void *buf;
> >>>
> >>> - fd = open(filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(filename, O_RDONLY));
> >>> if (fd < 0)
> >>> return -1;
> >>>
> >>> @@ -242,7 +242,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> if (buf == NULL)
> >>> goto out;
> >>>
> >>> - if (read(fd, buf, buf_size) != (ssize_t) buf_size)
> >>> + if (TEMP_FAILURE_RETRY(read(fd, buf, buf_size)) != (ssize_t) buf_size)
> >>> goto out_free;
> >>>
> >>> ret = read_build_id(buf, buf_size, bid, false);
> >>> @@ -256,7 +256,7 @@ int sysfs__read_build_id(const char *filename, struct build_id *bid)
> >>> int symsrc__init(struct symsrc *ss, struct dso *dso, const char *name,
> >>> enum dso_binary_type type)
> >>> {
> >>> - int fd = open(name, O_RDONLY);
> >>> + int fd = TEMP_FAILURE_RETRY(open(name, O_RDONLY));
> >>> if (fd < 0)
> >>> goto out_errno;
> >>>
> >>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
> >>> index f72baf636724..5212667ea67d 100644
> >>> --- a/tools/perf/util/symbol.c
> >>> +++ b/tools/perf/util/symbol.c
> >>> @@ -1335,7 +1335,7 @@ static int dso__load_kcore(struct dso *dso, struct map *map,
> >>> md.dso = dso;
> >>> INIT_LIST_HEAD(&md.maps);
> >>>
> >>> - fd = open(kcore_filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(kcore_filename, O_RDONLY));
> >>> if (fd < 0) {
> >>> pr_debug("Failed to open %s. Note /proc/kcore requires CAP_SYS_RAWIO capability to access.\n",
> >>> kcore_filename);
> >>> @@ -2156,7 +2156,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz)
> >>> */
> >>> static bool filename__readable(const char *file)
> >>> {
> >>> - int fd = open(file, O_RDONLY);
> >>> + int fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> >>> if (fd < 0)
> >>> return false;
> >>> close(fd);
> >>> diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c
> >>> index 76beda3e1a10..9dcc268bd69c 100644
> >>> --- a/tools/perf/util/synthetic-events.c
> >>> +++ b/tools/perf/util/synthetic-events.c
> >>> @@ -40,6 +40,7 @@
> >>> #include <api/io.h>
> >>> #include <sys/types.h>
> >>> #include <sys/stat.h>
> >>> +#include <errno.h>
> >>> #include <fcntl.h>
> >>> #include <unistd.h>
> >>>
> >>> @@ -86,13 +87,13 @@ static int perf_event__get_comm_ids(pid_t pid, pid_t tid, char *comm, size_t len
> >>> else
> >>> snprintf(bf, sizeof(bf), "/proc/%d/status", tid);
> >>>
> >>> - fd = open(bf, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY));
> >>> if (fd < 0) {
> >>> pr_debug("couldn't open %s\n", bf);
> >>> return -1;
> >>> }
> >>>
> >>> - n = read(fd, bf, sizeof(bf) - 1);
> >>> + n = TEMP_FAILURE_RETRY(read(fd, bf, sizeof(bf) - 1));
> >>> close(fd);
> >>> if (n <= 0) {
> >>> pr_warning("Couldn't get COMM, tigd and ppid for pid %d\n",
> >>> @@ -410,7 +411,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool,
> >>> snprintf(bf, sizeof(bf), "%s/proc/%d/task/%d/maps",
> >>> machine->root_dir, pid, pid);
> >>>
> >>> - io.fd = open(bf, O_RDONLY, 0);
> >>> + io.fd = TEMP_FAILURE_RETRY(open(bf, O_RDONLY, 0));
> >>> if (io.fd < 0) {
> >>> /*
> >>> * We raced with a task exiting - just return:
> >>> diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c
> >>> index a65f65d0857e..c917a4217c6f 100644
> >>> --- a/tools/perf/util/trace-event-info.c
> >>> +++ b/tools/perf/util/trace-event-info.c
> >>> @@ -48,7 +48,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> >>> int r, fd;
> >>> int err = -EIO;
> >>>
> >>> - fd = open(file, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(file, O_RDONLY));
> >>> if (fd < 0) {
> >>> pr_debug("Can't read '%s'", file);
> >>> return -errno;
> >>> @@ -61,7 +61,7 @@ static int record_file(const char *file, ssize_t hdr_sz)
> >>> }
> >>>
> >>> do {
> >>> - r = read(fd, buf, BUFSIZ);
> >>> + r = TEMP_FAILURE_RETRY(read(fd, buf, BUFSIZ));
> >>> if (r > 0) {
> >>> size += r;
> >>> if (write(output_fd, buf, r) != r)
> >>> @@ -519,7 +519,7 @@ struct tracing_data *tracing_data_get(struct list_head *pattrs,
> >>> return NULL;
> >>> }
> >>>
> >>> - temp_fd = open(tdata->temp_file, O_RDWR);
> >>> + temp_fd = TEMP_FAILURE_RETRY(open(tdata->temp_file, O_RDWR));
> >>> if (temp_fd < 0) {
> >>> pr_debug("Can't read '%s'", tdata->temp_file);
> >>> free(tdata);
> >>> diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c
> >>> index 8a01af783310..7086770ec225 100644
> >>> --- a/tools/perf/util/trace-event-read.c
> >>> +++ b/tools/perf/util/trace-event-read.c
> >>> @@ -11,6 +11,7 @@
> >>> #include <sys/stat.h>
> >>> #include <sys/wait.h>
> >>> #include <sys/mman.h>
> >>> +#include <errno.h>
> >>> #include <fcntl.h>
> >>> #include <unistd.h>
> >>> #include <errno.h>
> >>> @@ -28,7 +29,7 @@ static int __do_read(int fd, void *buf, int size)
> >>> int rsize = size;
> >>>
> >>> while (size) {
> >>> - int ret = read(fd, buf, size);
> >>> + int ret = TEMP_FAILURE_RETRY(read(fd, buf, size));
> >>>
> >>> if (ret <= 0)
> >>> return -1;
> >>> @@ -105,7 +106,7 @@ static char *read_string(void)
> >>> char c;
> >>>
> >>> for (;;) {
> >>> - r = read(input_fd, &c, 1);
> >>> + r = TEMP_FAILURE_RETRY(read(input_fd, &c, 1));
> >>> if (r < 0) {
> >>> pr_debug("reading input file");
> >>> goto out;
> >>> diff --git a/tools/perf/util/unwind-libunwind-local.c b/tools/perf/util/unwind-libunwind-local.c
> >>> index 37622699c91a..2ece387a73c6 100644
> >>> --- a/tools/perf/util/unwind-libunwind-local.c
> >>> +++ b/tools/perf/util/unwind-libunwind-local.c
> >>> @@ -362,7 +362,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> >>> }
> >>>
> >>> if (ofs <= 0) {
> >>> - fd = open(dso->symsrc_filename, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(dso->symsrc_filename, O_RDONLY));
> >>> if (fd >= 0) {
> >>> ofs = elf_section_offset(fd, ".debug_frame");
> >>> close(fd);
> >>> @@ -377,7 +377,7 @@ static int read_unwind_spec_debug_frame(struct dso *dso,
> >>> dso, DSO_BINARY_TYPE__DEBUGLINK,
> >>> machine->root_dir, debuglink, PATH_MAX);
> >>> if (!ret) {
> >>> - fd = open(debuglink, O_RDONLY);
> >>> + fd = TEMP_FAILURE_RETRY(open(debuglink, O_RDONLY));
> >>> if (fd >= 0) {
> >>> ofs = elf_section_offset(fd,
> >>> ".debug_frame");
> >>> diff --git a/tools/perf/util/zlib.c b/tools/perf/util/zlib.c
> >>> index 78d2297c1b67..f4563ea094c8 100644
> >>> --- a/tools/perf/util/zlib.c
> >>> +++ b/tools/perf/util/zlib.c
> >>> @@ -1,4 +1,5 @@
> >>> // SPDX-License-Identifier: GPL-2.0
> >>> +#include <errno.h>
> >>> #include <fcntl.h>
> >>> #include <stdio.h>
> >>> #include <string.h>
> >>> @@ -82,7 +83,7 @@ int gzip_decompress_to_file(const char *input, int output_fd)
> >>>
> >>> bool gzip_is_compressed(const char *input)
> >>> {
> >>> - int fd = open(input, O_RDONLY);
> >>> + int fd = TEMP_FAILURE_RETRY(open(input, O_RDONLY));
> >>> const uint8_t magic[2] = { 0x1f, 0x8b };
> >>> char buf[2] = { 0 };
> >>> ssize_t rc;
> >>> @@ -90,7 +91,7 @@ bool gzip_is_compressed(const char *input)
> >>> if (fd < 0)
> >>> return -1;
> >>>
> >>> - rc = read(fd, buf, sizeof(buf));
> >>> + rc = TEMP_FAILURE_RETRY(read(fd, buf, sizeof(buf)));
> >>> close(fd);
> >>> return rc == sizeof(buf) ?
> >>> memcmp(buf, magic, sizeof(buf)) == 0 : false;
> >>
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 6:03 [PATCH] perf io: Make open and read calls more robust Ian Rogers
2022-06-15 6:50 ` Adrian Hunter
@ 2022-06-15 18:33 ` Andi Kleen
2022-06-15 19:17 ` Ian Rogers
1 sibling, 1 reply; 7+ messages in thread
From: Andi Kleen @ 2022-06-15 18:33 UTC (permalink / raw)
To: Ian Rogers, Peter Zijlstra, Ingo Molnar,
Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin,
Jiri Olsa, Namhyung Kim, Alexey Bayduraev, Riccardo Mancini,
Adrian Hunter, Dave Marchevsky, Song Liu, Leo Yan, John Garry,
Florian Fischer, Kan Liang, Masami Hiramatsu, Zechuan Chen,
Yang Jihong, German Gomez, Thomas Richter, James Clark,
Denis Nikitin, Lexi Shao, Michael Petlan, Remi Bernon,
Fangrui Song, linux-kernel, linux-perf-users
Cc: Stephane Eranian
On 6/14/2022 11:03 PM, Ian Rogers wrote:
> Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
> causes the syscall to need to restart.
Why not just use SA_RESTART on the signals?
-Andi
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] perf io: Make open and read calls more robust
2022-06-15 18:33 ` Andi Kleen
@ 2022-06-15 19:17 ` Ian Rogers
0 siblings, 0 replies; 7+ messages in thread
From: Ian Rogers @ 2022-06-15 19:17 UTC (permalink / raw)
To: Andi Kleen
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Alexey Bayduraev, Riccardo Mancini, Adrian Hunter,
Dave Marchevsky, Song Liu, Leo Yan, John Garry, Florian Fischer,
Kan Liang, Masami Hiramatsu, Zechuan Chen, Yang Jihong,
German Gomez, Thomas Richter, James Clark, Denis Nikitin,
Lexi Shao, Michael Petlan, Remi Bernon, Fangrui Song,
linux-kernel, linux-perf-users, Stephane Eranian
On Wed, Jun 15, 2022 at 11:47 AM Andi Kleen <ak@linux.intel.com> wrote:
>
>
> On 6/14/2022 11:03 PM, Ian Rogers wrote:
> > Wrap open and read calls with TEMP_FAILURE_RETRY in case a signal
> > causes the syscall to need to restart.
>
>
> Why not just use SA_RESTART on the signals?
As explained to Adrian, existing code is EINTR tolerant, some of the
code is in libraries and for portability shouldn't assume SA_RESTART.
Thanks,
Ian
>
> -Andi
>
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2022-06-15 19:17 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-15 6:03 [PATCH] perf io: Make open and read calls more robust Ian Rogers
2022-06-15 6:50 ` Adrian Hunter
2022-06-15 15:34 ` Ian Rogers
2022-06-15 16:40 ` Adrian Hunter
2022-06-15 17:38 ` Ian Rogers
2022-06-15 18:33 ` Andi Kleen
2022-06-15 19:17 ` Ian Rogers
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).