blob: f0e06b879504ce375282d6e3d57ab276d0605b31 [file] [log] [blame]
// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// +build aix darwin dragonfly freebsd linux netbsd openbsd solaris
package signal
import (
"bytes"
"flag"
"fmt"
"internal/testenv"
"io/ioutil"
"os"
"os/exec"
"runtime"
"strconv"
"sync"
"syscall"
"testing"
"time"
)
// settleTime is an upper bound on how long we expect signals to take to be
// delivered. Lower values make the test faster, but also flakier — especially
// on heavily loaded systems.
//
// The current value is set based on flakes observed in the Go builders.
var settleTime = 100 * time.Millisecond
func init() {
if testenv.Builder() == "solaris-amd64-oraclerel" {
// The solaris-amd64-oraclerel builder has been observed to time out in
// TestNohup even with a 250ms settle time.
//
// Use a much longer settle time on that builder to try to suss out whether
// the test is flaky due to builder slowness (which may mean we need a
// longer GO_TEST_TIMEOUT_SCALE) or due to a dropped signal (which may
// instead need a test-skip and upstream bug filed against the Solaris
// kernel).
//
// This constant is chosen so as to make the test as generous as possible
// while still reliably completing within 3 minutes in non-short mode.
//
// See https://golang.org/issue/33174.
settleTime = 11 * time.Second
} else if s := os.Getenv("GO_TEST_TIMEOUT_SCALE"); s != "" {
if scale, err := strconv.Atoi(s); err == nil {
settleTime *= time.Duration(scale)
}
}
}
func waitSig(t *testing.T, c <-chan os.Signal, sig os.Signal) {
t.Helper()
waitSig1(t, c, sig, false)
}
func waitSigAll(t *testing.T, c <-chan os.Signal, sig os.Signal) {
t.Helper()
waitSig1(t, c, sig, true)
}
func waitSig1(t *testing.T, c <-chan os.Signal, sig os.Signal, all bool) {
t.Helper()
// Sleep multiple times to give the kernel more tries to
// deliver the signal.
start := time.Now()
timer := time.NewTimer(settleTime / 10)
defer timer.Stop()
// If the caller notified for all signals on c, filter out SIGURG,
// which is used for runtime preemption and can come at unpredictable times.
// General user code should filter out all unexpected signals instead of just
// SIGURG, but since os/signal is tightly coupled to the runtime it seems
// appropriate to be stricter here.
for time.Since(start) < settleTime {
select {
case s := <-c:
if s == sig {
return
}
if !all || s != syscall.SIGURG {
t.Fatalf("signal was %v, want %v", s, sig)
}
case <-timer.C:
timer.Reset(settleTime / 10)
}
}
t.Fatalf("timeout after %v waiting for %v", settleTime, sig)
}
// quiesce waits until we can be reasonably confident that all pending signals
// have been delivered by the OS.
func quiesce() {
// The kernel will deliver a signal as a thread returns
// from a syscall. If the only active thread is sleeping,
// and the system is busy, the kernel may not get around
// to waking up a thread to catch the signal.
// We try splitting up the sleep to give the kernel
// many chances to deliver the signal.
start := time.Now()
for time.Since(start) < settleTime {
time.Sleep(settleTime / 10)
}
}
// Test that basic signal handling works.
func TestSignal(t *testing.T) {
// Ask for SIGHUP
c := make(chan os.Signal, 1)
Notify(c, syscall.SIGHUP)
defer Stop(c)
// Send this process a SIGHUP
t.Logf("sighup...")
syscall.Kill(syscall.Getpid(), syscall.SIGHUP)
waitSig(t, c, syscall.SIGHUP)
// Ask for everything we can get. The buffer size has to be
// more than 1, since the runtime might send SIGURG signals.
// Using 10 is arbitrary.
c1 := make(chan os.Signal, 10)
Notify(c1)
// Send this process a SIGWINCH
t.Logf("sigwinch...")
syscall.Kill(syscall.Getpid(), syscall.SIGWINCH)
waitSigAll(t, c1, syscall.SIGWINCH)
// Send two more SIGHUPs, to make sure that
// they get delivered on c1 and that not reading
// from c does not block everything.
t.Logf("sighup...")
syscall.Kill(syscall.Getpid(), syscall.SIGHUP)
waitSigAll(t, c1, syscall.SIGHUP)
t.Logf("sighup...")
syscall.Kill(syscall.Getpid(), syscall.SIGHUP)
waitSigAll(t, c1, syscall.SIGHUP)
// The first SIGHUP should be waiting for us on c.
waitSig(t, c, syscall.SIGHUP)
}
func TestStress(t *testing.T) {
dur := 3 * time.Second
if testing.Short() {
dur = 100 * time.Millisecond
}
defer runtime.GOMAXPROCS(runtime.GOMAXPROCS(4))
sig := make(chan os.Signal, 1)
Notify(sig, syscall.SIGUSR1)
go func() {
stop := time.After(dur)
for {
select {
case <-stop:
// Allow enough time for all signals to be delivered before we stop
// listening for them.
quiesce()
Stop(sig)
// According to its documentation, “[w]hen Stop returns, it in
// guaranteed that c will receive no more signals.” So we can safely
// close sig here: if there is a send-after-close race here, that is a
// bug in Stop and we would like to detect it.
close(sig)
return
default:
syscall.Kill(syscall.Getpid(), syscall.SIGUSR1)
runtime.Gosched()
}
}
}()
for range sig {
// Receive signals until the sender closes sig.
}
}
func testCancel(t *testing.T, ignore bool) {
// Ask to be notified on c1 when a SIGWINCH is received.
c1 := make(chan os.Signal, 1)
Notify(c1, syscall.SIGWINCH)
defer Stop(c1)
// Ask to be notified on c2 when a SIGHUP is received.
c2 := make(chan os.Signal, 1)
Notify(c2, syscall.SIGHUP)
defer Stop(c2)
// Send this process a SIGWINCH and wait for notification on c1.
syscall.Kill(syscall.Getpid(), syscall.SIGWINCH)
waitSig(t, c1, syscall.SIGWINCH)
// Send this process a SIGHUP and wait for notification on c2.
syscall.Kill(syscall.Getpid(), syscall.SIGHUP)
waitSig(t, c2, syscall.SIGHUP)
// Ignore, or reset the signal handlers for, SIGWINCH and SIGHUP.
// Either way, this should undo both calls to Notify above.
if ignore {
Ignore(syscall.SIGWINCH, syscall.SIGHUP)
// Don't bother deferring a call to Reset: it is documented to undo Notify,
// but its documentation says nothing about Ignore, and (as of the time of
// writing) it empirically does not undo an Ignore.
} else {
Reset(syscall.SIGWINCH, syscall.SIGHUP)
}
// Send this process a SIGWINCH. It should be ignored.
syscall.Kill(syscall.Getpid(), syscall.SIGWINCH)
// If ignoring, Send this process a SIGHUP. It should be ignored.
if ignore {
syscall.Kill(syscall.Getpid(), syscall.SIGHUP)
}
quiesce()
select {
case s := <-c1:
t.Errorf("unexpected signal %v", s)
default:
// nothing to read - good
}
select {
case s := <-c2:
t.Errorf("unexpected signal %v", s)
default:
// nothing to read - good
}
// One or both of the signals may have been blocked for this process
// by the calling process.
// Discard any queued signals now to avoid interfering with other tests.
Notify(c1, syscall.SIGWINCH)
Notify(c2, syscall.SIGHUP)
quiesce()
}
// Test that Reset cancels registration for listed signals on all channels.
func TestReset(t *testing.T) {
testCancel(t, false)
}
// Test that Ignore cancels registration for listed signals on all channels.
func TestIgnore(t *testing.T) {
testCancel(t, true)
}
// Test that Ignored correctly detects changes to the ignored status of a signal.
func TestIgnored(t *testing.T) {
// Ask to be notified on SIGWINCH.
c := make(chan os.Signal, 1)
Notify(c, syscall.SIGWINCH)
// If we're being notified, then the signal should not be ignored.
if Ignored(syscall.SIGWINCH) {
t.Errorf("expected SIGWINCH to not be ignored.")
}
Stop(c)
Ignore(syscall.SIGWINCH)
// We're no longer paying attention to this signal.
if !Ignored(syscall.SIGWINCH) {
t.Errorf("expected SIGWINCH to be ignored when explicitly ignoring it.")
}
Reset()
}
var checkSighupIgnored = flag.Bool("check_sighup_ignored", false, "if true, TestDetectNohup will fail if SIGHUP is not ignored.")
// Test that Ignored(SIGHUP) correctly detects whether it is being run under nohup.
func TestDetectNohup(t *testing.T) {
if *checkSighupIgnored {
if !Ignored(syscall.SIGHUP) {
t.Fatal("SIGHUP is not ignored.")
} else {
t.Log("SIGHUP is ignored.")
}
} else {
defer Reset()
// Ugly: ask for SIGHUP so that child will not have no-hup set
// even if test is running under nohup environment.
// We have no intention of reading from c.
c := make(chan os.Signal, 1)
Notify(c, syscall.SIGHUP)
if out, err := exec.Command(os.Args[0], "-test.run=TestDetectNohup", "-check_sighup_ignored").CombinedOutput(); err == nil {
t.Errorf("ran test with -check_sighup_ignored and it succeeded: expected failure.\nOutput:\n%s", out)
}
Stop(c)
// Again, this time with nohup, assuming we can find it.
_, err := os.Stat("/usr/bin/nohup")
if err != nil {
t.Skip("cannot find nohup; skipping second half of test")
}
Ignore(syscall.SIGHUP)
os.Remove("nohup.out")
out, err := exec.Command("/usr/bin/nohup", os.Args[0], "-test.run=TestDetectNohup", "-check_sighup_ignored").CombinedOutput()
data, _ := ioutil.ReadFile("nohup.out")
os.Remove("nohup.out")
if err != nil {
t.Errorf("ran test with -check_sighup_ignored under nohup and it failed: expected success.\nError: %v\nOutput:\n%s%s", err, out, data)
}
}
}
var (
sendUncaughtSighup = flag.Int("send_uncaught_sighup", 0, "send uncaught SIGHUP during TestStop")
dieFromSighup = flag.Bool("die_from_sighup", false, "wait to die from uncaught SIGHUP")
)
// Test that Stop cancels the channel's registrations.
func TestStop(t *testing.T) {
sigs := []syscall.Signal{
syscall.SIGWINCH,
syscall.SIGHUP,
syscall.SIGUSR1,
}
for _, sig := range sigs {
sig := sig
t.Run(fmt.Sprint(sig), func(t *testing.T) {
// When calling Notify with a specific signal,
// independent signals should not interfere with each other,
// and we end up needing to wait for signals to quiesce a lot.
// Test the three different signals concurrently.
t.Parallel()
// If the signal is not ignored, send the signal before registering a
// channel to verify the behavior of the default Go handler.
// If it's SIGWINCH or SIGUSR1 we should not see it.
// If it's SIGHUP, maybe we'll die. Let the flag tell us what to do.
mayHaveBlockedSignal := false
if !Ignored(sig) && (sig != syscall.SIGHUP || *sendUncaughtSighup == 1) {
syscall.Kill(syscall.Getpid(), sig)
quiesce()
// We don't know whether sig is blocked for this process; see
// https://golang.org/issue/38165. Assume that it could be.
mayHaveBlockedSignal = true
}
// Ask for signal
c := make(chan os.Signal, 1)
Notify(c, sig)
// Send this process the signal again.
syscall.Kill(syscall.Getpid(), sig)
waitSig(t, c, sig)
if mayHaveBlockedSignal {
// We may have received a queued initial signal in addition to the one
// that we sent after Notify. If so, waitSig may have observed that
// initial signal instead of the second one, and we may need to wait for
// the second signal to clear. Do that now.
quiesce()
select {
case <-c:
default:
}
}
// Stop watching for the signal and send it again.
// If it's SIGHUP, maybe we'll die. Let the flag tell us what to do.
Stop(c)
if sig != syscall.SIGHUP || *sendUncaughtSighup == 2 {
syscall.Kill(syscall.Getpid(), sig)
quiesce()
select {
case s := <-c:
t.Errorf("unexpected signal %v", s)
default:
// nothing to read - good
}
// If we're going to receive a signal, it has almost certainly been
// received by now. However, it may have been blocked for this process —
// we don't know. Explicitly unblock it and wait for it to clear now.
Notify(c, sig)
quiesce()
Stop(c)
}
})
}
}
// Test that when run under nohup, an uncaught SIGHUP does not kill the program.
func TestNohup(t *testing.T) {
// Ugly: ask for SIGHUP so that child will not have no-hup set
// even if test is running under nohup environment.
// We have no intention of reading from c.
c := make(chan os.Signal, 1)
Notify(c, syscall.SIGHUP)
// When run without nohup, the test should crash on an uncaught SIGHUP.
// When run under nohup, the test should ignore uncaught SIGHUPs,
// because the runtime is not supposed to be listening for them.
// Either way, TestStop should still be able to catch them when it wants them
// and then when it stops wanting them, the original behavior should resume.
//
// send_uncaught_sighup=1 sends the SIGHUP before starting to listen for SIGHUPs.
// send_uncaught_sighup=2 sends the SIGHUP after no longer listening for SIGHUPs.
//
// Both should fail without nohup and succeed with nohup.
var subTimeout time.Duration
var wg sync.WaitGroup
wg.Add(2)
if deadline, ok := t.Deadline(); ok {
subTimeout = time.Until(deadline)
subTimeout -= subTimeout / 10 // Leave 10% headroom for propagating output.
}
for i := 1; i <= 2; i++ {
i := i
go t.Run(fmt.Sprintf("uncaught-%d", i), func(t *testing.T) {
defer wg.Done()
args := []string{
"-test.v",
"-test.run=TestStop",
"-send_uncaught_sighup=" + strconv.Itoa(i),
"-die_from_sighup",
}
if subTimeout != 0 {
args = append(args, fmt.Sprintf("-test.timeout=%v", subTimeout))
}
out, err := exec.Command(os.Args[0], args...).CombinedOutput()
if err == nil {
t.Errorf("ran test with -send_uncaught_sighup=%d and it succeeded: expected failure.\nOutput:\n%s", i, out)
} else {
t.Logf("test with -send_uncaught_sighup=%d failed as expected.\nError: %v\nOutput:\n%s", i, err, out)
}
})
}
wg.Wait()
Stop(c)
// Skip the nohup test below when running in tmux on darwin, since nohup
// doesn't work correctly there. See issue #5135.
if runtime.GOOS == "darwin" && os.Getenv("TMUX") != "" {
t.Skip("Skipping nohup test due to running in tmux on darwin")
}
// Again, this time with nohup, assuming we can find it.
_, err := exec.LookPath("nohup")
if err != nil {
t.Skip("cannot find nohup; skipping second half of test")
}
wg.Add(2)
if deadline, ok := t.Deadline(); ok {
subTimeout = time.Until(deadline)
subTimeout -= subTimeout / 10 // Leave 10% headroom for propagating output.
}
for i := 1; i <= 2; i++ {
i := i
go t.Run(fmt.Sprintf("nohup-%d", i), func(t *testing.T) {
defer wg.Done()
// POSIX specifies that nohup writes to a file named nohup.out if standard
// output is a terminal. However, for an exec.Command, standard output is
// not a terminal — so we don't need to read or remove that file (and,
// indeed, cannot even create it if the current user is unable to write to
// GOROOT/src, such as when GOROOT is installed and owned by root).
args := []string{
os.Args[0],
"-test.v",
"-test.run=TestStop",
"-send_uncaught_sighup=" + strconv.Itoa(i),
}
if subTimeout != 0 {
args = append(args, fmt.Sprintf("-test.timeout=%v", subTimeout))
}
out, err := exec.Command("nohup", args...).CombinedOutput()
if err != nil {
t.Errorf("ran test with -send_uncaught_sighup=%d under nohup and it failed: expected success.\nError: %v\nOutput:\n%s", i, err, out)
} else {
t.Logf("ran test with -send_uncaught_sighup=%d under nohup.\nOutput:\n%s", i, out)
}
})
}
wg.Wait()
}
// Test that SIGCONT works (issue 8953).
func TestSIGCONT(t *testing.T) {
c := make(chan os.Signal, 1)
Notify(c, syscall.SIGCONT)
defer Stop(c)
syscall.Kill(syscall.Getpid(), syscall.SIGCONT)
waitSig(t, c, syscall.SIGCONT)
}
// Test race between stopping and receiving a signal (issue 14571).
func TestAtomicStop(t *testing.T) {
if os.Getenv("GO_TEST_ATOMIC_STOP") != "" {
atomicStopTestProgram(t)
t.Fatal("atomicStopTestProgram returned")
}
testenv.MustHaveExec(t)
// Call Notify for SIGINT before starting the child process.
// That ensures that SIGINT is not ignored for the child.
// This is necessary because if SIGINT is ignored when a
// Go program starts, then it remains ignored, and closing
// the last notification channel for SIGINT will switch it
// back to being ignored. In that case the assumption of
// atomicStopTestProgram, that it will either die from SIGINT
// or have it be reported, breaks down, as there is a third
// option: SIGINT might be ignored.
cs := make(chan os.Signal, 1)
Notify(cs, syscall.SIGINT)
defer Stop(cs)
const execs = 10
for i := 0; i < execs; i++ {
timeout := "0"
if deadline, ok := t.Deadline(); ok {
timeout = time.Until(deadline).String()
}
cmd := exec.Command(os.Args[0], "-test.run=TestAtomicStop", "-test.timeout="+timeout)
cmd.Env = append(os.Environ(), "GO_TEST_ATOMIC_STOP=1")
out, err := cmd.CombinedOutput()
if err == nil {
if len(out) > 0 {
t.Logf("iteration %d: output %s", i, out)
}
} else {
t.Logf("iteration %d: exit status %q: output: %s", i, err, out)
}
lost := bytes.Contains(out, []byte("lost signal"))
if lost {
t.Errorf("iteration %d: lost signal", i)
}
// The program should either die due to SIGINT,
// or exit with success without printing "lost signal".
if err == nil {
if len(out) > 0 && !lost {
t.Errorf("iteration %d: unexpected output", i)
}
} else {
if ee, ok := err.(*exec.ExitError); !ok {
t.Errorf("iteration %d: error (%v) has type %T; expected exec.ExitError", i, err, err)
} else if ws, ok := ee.Sys().(syscall.WaitStatus); !ok {
t.Errorf("iteration %d: error.Sys (%v) has type %T; expected syscall.WaitStatus", i, ee.Sys(), ee.Sys())
} else if !ws.Signaled() || ws.Signal() != syscall.SIGINT {
t.Errorf("iteration %d: got exit status %v; expected SIGINT", i, ee)
}
}
}
}
// atomicStopTestProgram is run in a subprocess by TestAtomicStop.
// It tries to trigger a signal delivery race. This function should
// either catch a signal or die from it.
func atomicStopTestProgram(t *testing.T) {
// This test won't work if SIGINT is ignored here.
if Ignored(syscall.SIGINT) {
fmt.Println("SIGINT is ignored")
os.Exit(1)
}
const tries = 10
timeout := 2 * time.Second
if deadline, ok := t.Deadline(); ok {
// Give each try an equal slice of the deadline, with one slice to spare for
// cleanup.
timeout = time.Until(deadline) / (tries + 1)
}
pid := syscall.Getpid()
printed := false
for i := 0; i < tries; i++ {
cs := make(chan os.Signal, 1)
Notify(cs, syscall.SIGINT)
var wg sync.WaitGroup
wg.Add(1)
go func() {
defer wg.Done()
Stop(cs)
}()
syscall.Kill(pid, syscall.SIGINT)
// At this point we should either die from SIGINT or
// get a notification on cs. If neither happens, we
// dropped the signal. It is given 2 seconds to
// deliver, as needed for gccgo on some loaded test systems.
select {
case <-cs:
case <-time.After(timeout):
if !printed {
fmt.Print("lost signal on tries:")
printed = true
}
fmt.Printf(" %d", i)
}
wg.Wait()
}
if printed {
fmt.Print("\n")
}
os.Exit(0)
}
func TestTime(t *testing.T) {
// Test that signal works fine when we are in a call to get time,
// which on some platforms is using VDSO. See issue #34391.
dur := 3 * time.Second
if testing.Short() {
dur = 100 * time.Millisecond
}
defer runtime.GOMAXPROCS(runtime.GOMAXPROCS(4))
sig := make(chan os.Signal, 1)
Notify(sig, syscall.SIGUSR1)
stop := make(chan struct{})
go func() {
for {
select {
case <-stop:
// Allow enough time for all signals to be delivered before we stop
// listening for them.
quiesce()
Stop(sig)
// According to its documentation, “[w]hen Stop returns, it in
// guaranteed that c will receive no more signals.” So we can safely
// close sig here: if there is a send-after-close race, that is a bug in
// Stop and we would like to detect it.
close(sig)
return
default:
syscall.Kill(syscall.Getpid(), syscall.SIGUSR1)
runtime.Gosched()
}
}
}()
done := make(chan struct{})
go func() {
for range sig {
// Receive signals until the sender closes sig.
}
close(done)
}()
t0 := time.Now()
for t1 := t0; t1.Sub(t0) < dur; t1 = time.Now() {
} // hammering on getting time
close(stop)
<-done
}