The user_notification_wait_killable_after_reply test fails due to an
unhandled error when a traced syscall is interrupted by a signal.
When a signal arrives after the tracer has received a seccomp
notification but before it has replied, the notification can become
stale. Any subsequent reply (like with SECCOMP_IOCTL_NOTIF_ADDFD)
will fail with -ENOENT.
This patch fixes the test by handling the -ENOENT return value from
SECCOMP_IOCTL_NOTIF_ADDFD, preventing the test from failing
incorrectly. The loop counter is decremented to re-run the iteration
for the restarted syscall.
Signed-off-by: Wake Liu <wakel@google.com>
---
tools/testing/selftests/seccomp/seccomp_bpf.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c
index 574fdd102eb5..c3e598c9c4ee 100644
--- a/tools/testing/selftests/seccomp/seccomp_bpf.c
+++ b/tools/testing/selftests/seccomp/seccomp_bpf.c
@@ -5048,8 +5048,12 @@ TEST(user_notification_wait_killable_after_reply)
addfd.id = req.id;
addfd.flags = SECCOMP_ADDFD_FLAG_SEND;
addfd.srcfd = 0;
- ASSERT_GE(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), 0)
- kill(pid, SIGKILL);
+ ret = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd);
+ if (ret < 0 && errno == ENOENT) {
+ i--;
+ continue;
+ }
+ ASSERT_GE(ret, 0);
}
/*
--
2.52.0.rc1.455.g30608eb744-goog