[PATCH v4 1/3] perf sched stats: Fix SIGCHLD vs pause() race in schedstat_record()

[Swapnil Sapkal]

If the profiled workload exits very quickly, SIGCHLD can be delivered
and consumed by the empty signal handler before the process enters
pause(), causing an indefinite hang.

Fix this with a simpler approach:

 - The signal handler now sets a 'volatile sig_atomic_t done' flag.
   Reset 'done' before registering signal handlers so that an early
   signal during setup is not discarded by a later reset.

 - Replace pause() with a loop that checks 'done' and uses
   waitpid(WNOHANG) to detect child exit without blocking.  This
   handles both workload mode (child exits) and system-wide mode
   (user sends SIGINT/SIGTERM).  Using WNOHANG avoids the SA_RESTART
   problem where a blocking waitpid() would auto-restart and ignore
   the done flag if the child doesn't exit on signal.

Suggested-by: Namhyung Kim <namhyung@xxxxxxxxxx>
Assisted-by: Claude:claude-opus-4.6
Signed-off-by: Swapnil Sapkal <swapnil.sapkal@xxxxxxx>
---
 tools/perf/builtin-sched.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index 53a93aa18853..7da71c372e25 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -36,6 +36,7 @@
 #include <linux/zalloc.h>
 #include <sys/prctl.h>
 #include <sys/resource.h>
+#include <sys/wait.h>
 #include <inttypes.h>
 
 #include <errno.h>
@@ -3758,8 +3759,11 @@ static int process_synthesized_schedstat_event(const struct perf_tool *tool,
 	return 0;
 }
 
+static volatile sig_atomic_t done;
+
 static void sighandler(int sig __maybe_unused)
 {
+	done = 1;
 }
 
 static int enable_sched_schedstats(int *reset)
@@ -3819,6 +3823,7 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 		.mode  = PERF_DATA_MODE_WRITE,
 	};
 
+	done = 0;
 	signal(SIGINT, sighandler);
 	signal(SIGCHLD, sighandler);
 	signal(SIGTERM, sighandler);
@@ -3903,8 +3908,11 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 	if (argc)
 		evlist__start_workload(evlist);
 
-	/* wait for signal */
-	pause();
+	while (!done) {
+		if (argc && waitpid(evlist->workload.pid, NULL, WNOHANG) > 0)
+			break;
+		sleep(1);
+	}
 
 	if (reset) {
 		err = disable_sched_schedstat();
-- 
2.43.0