[PATCH v2] selftests/filesystems: Assume that TIOCGPTPEER is defined

Mark Brown posted 1 patch 1 month, 3 weeks ago
tools/testing/selftests/filesystems/devpts_pts.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
[PATCH v2] selftests/filesystems: Assume that TIOCGPTPEER is defined
Posted by Mark Brown 1 month, 3 weeks ago
The devpts_pts selftest has an ifdef in case an architecture does not
define TIOCGPTPEER, but the handling for this is broken since we need
errno to be set to EINVAL in order to skip the test as we should. Given
that this ioctl() has been defined since v4.15 we may as well just assume
it's there rather than write handling code which will probably never get
used.

Signed-off-by: Mark Brown <broonie@kernel.org>
---
Changes in v2:
- Rebase onto v6.19-rc1.
- Link to v1: https://patch.msgid.link/20251126-selftests-filesystems-devpts-tiocgptpeer-v1-1-92bd65d02981@kernel.org
---
 tools/testing/selftests/filesystems/devpts_pts.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/tools/testing/selftests/filesystems/devpts_pts.c b/tools/testing/selftests/filesystems/devpts_pts.c
index 54fea349204e..950e8b7f675b 100644
--- a/tools/testing/selftests/filesystems/devpts_pts.c
+++ b/tools/testing/selftests/filesystems/devpts_pts.c
@@ -100,7 +100,7 @@ static int resolve_procfd_symlink(int fd, char *buf, size_t buflen)
 static int do_tiocgptpeer(char *ptmx, char *expected_procfd_contents)
 {
 	int ret;
-	int master = -1, slave = -1, fret = -1;
+	int master = -1, slave, fret = -1;
 
 	master = open(ptmx, O_RDWR | O_NOCTTY | O_CLOEXEC);
 	if (master < 0) {
@@ -119,9 +119,7 @@ static int do_tiocgptpeer(char *ptmx, char *expected_procfd_contents)
 		goto do_cleanup;
 	}
 
-#ifdef TIOCGPTPEER
 	slave = ioctl(master, TIOCGPTPEER, O_RDWR | O_NOCTTY | O_CLOEXEC);
-#endif
 	if (slave < 0) {
 		if (errno == EINVAL) {
 			fprintf(stderr, "TIOCGPTPEER is not supported. "

---
base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8
change-id: 20251126-selftests-filesystems-devpts-tiocgptpeer-fbd30e579859

Best regards,
--  
Mark Brown <broonie@kernel.org>