os/exec: in TestContextCancel, dump goroutines on failure

If this test fails, we want to know exactly what the os/exec
goroutines are doing. Panicking gives us a goroutine dump,
whereas t.Fatal does not.

While we're here, use exponential backoff instead of a hard-coded 1ms
sleep. We want to give the OS enough time to actually terminate the
subprocess.

For #42061

Change-Id: I3d50a71ac314853c68a935218e7f97ce18b08b5b
Reviewed-on: https://go-review.googlesource.com/c/go/+/368317
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
diff --git a/src/os/exec/exec_test.go b/src/os/exec/exec_test.go
index 6172c78..81de018 100644
--- a/src/os/exec/exec_test.go
+++ b/src/os/exec/exec_test.go
@@ -954,6 +954,10 @@
 }
 
 func TestContextCancel(t *testing.T) {
+	// To reduce noise in the final goroutine dump,
+	// let other parallel tests complete if possible.
+	t.Parallel()
+
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 	c := helperCommandContext(t, ctx, "cat")
@@ -978,14 +982,25 @@
 	// Calling cancel should have killed the process, so writes
 	// should now fail.  Give the process a little while to die.
 	start := time.Now()
+	delay := 1 * time.Millisecond
 	for {
 		if _, err := io.WriteString(stdin, "echo"); err != nil {
 			break
 		}
+
 		if time.Since(start) > time.Minute {
-			t.Fatal("canceling context did not stop program")
+			// Panic instead of calling t.Fatal so that we get a goroutine dump.
+			// We want to know exactly what the os/exec goroutines got stuck on.
+			panic("canceling context did not stop program")
 		}
-		time.Sleep(time.Millisecond)
+
+		// Back off exponentially (up to 1-second sleeps) to give the OS time to
+		// terminate the process.
+		delay *= 2
+		if delay > 1*time.Second {
+			delay = 1 * time.Second
+		}
+		time.Sleep(delay)
 	}
 
 	if err := c.Wait(); err == nil {