[PATCH] selftests: binderfs: skip the stress test without binderfs

[PATCH] selftests: binderfs: skip the stress test without binderfs

[Eva Kurchatova]

binderfs_stress mounts binderfs and asserts on the first mount, so on
a kernel without CONFIG_ANDROID_BINDERFS the case fails rather than
reporting that the filesystem is not there.

Ask /proc/filesystems whether the kernel has it, the way
supports_filesystem() in the landlock tests and check_resctrlfs_support()
in the resctrl ones already do, and skip when it does not.

Probing with a mount instead would not work here: the test is meant to
run unprivileged, and path_mount() rejects an unprivileged caller in
may_mount() before do_new_mount() ever looks the filesystem type up, so
the probe would see EPERM whether binderfs is there or not. The child
would then still reach the real mount inside its userns and fail.

Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>
---
 .../filesystems/binderfs/binderfs_test.c      | 25 +++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/tools/testing/selftests/filesystems/binderfs/binderfs_test.c b/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
index a1a79a6fef17..70f0391136c4 100644
--- a/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
+++ b/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
@@ -36,6 +36,28 @@
 		fd = -EBADF;        \
 	}
 
+static bool binderfs_supported(void)
+{
+	char line[128];
+	bool ret = false;
+	FILE *f;
+
+	f = fopen("/proc/filesystems", "r");
+	if (!f)
+		return true;	/* Cannot tell, let the test run and report. */
+
+	while (fgets(line, sizeof(line), f)) {
+		/* binderfs has no backing device, hence the "nodev" prefix. */
+		if (!strcmp(line, "nodev\tbinder\n")) {
+			ret = true;
+			break;
+		}
+	}
+
+	fclose(f);
+	return ret;
+}
+
 static void change_mountns(struct __test_metadata *_metadata)
 {
 	int ret;
@@ -388,6 +410,9 @@ TEST(binderfs_stress)
 	char binderfs_mntpt[] = P_tmpdir "/binderfs_XXXXXX",
 		device_path[sizeof(P_tmpdir "/binderfs_XXXXXX/") + BINDERFS_MAX_NAME];
 
+	if (!binderfs_supported())
+		SKIP(return, "The Android binderfs filesystem is not available");
+
 	ret = socketpair(PF_LOCAL, SOCK_STREAM | SOCK_CLOEXEC, 0, syncfds);
 	ASSERT_EQ(ret, 0) {
 		TH_LOG("%s - Failed to create socket pair", strerror(errno));
-- 
2.55.0

Re: [PATCH] selftests: binderfs: skip the stress test without binderfs

[sashiko-bot]

> binderfs_stress mounts binderfs and asserts on the first mount, so on
> a kernel without CONFIG_ANDROID_BINDERFS the case fails rather than
> reporting that the filesystem is not there.
> 
> Ask /proc/filesystems whether the kernel has it, the way
> supports_filesystem() in the landlock tests and check_resctrlfs_support()
> in the resctrl ones already do, and skip when it does not.
> 
> Probing with a mount instead would not work here: the test is meant to
> run unprivileged, and path_mount() rejects an unprivileged caller in
> may_mount() before do_new_mount() ever looks the filesystem type up, so
> the probe would see EPERM whether binderfs is there or not. The child
> would then still reach the real mount inside its userns and fail.
> 
> Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260925133435.3635453-1-eva.kurchatova@virtuozzo.com?part=1