[LTP] [PATCH] syscalls/getxattr05: skip userns test cases when not supported

[LTP] [PATCH] syscalls/getxattr05: skip userns test cases when not supported

[Amir Goldstein]

These test cases are failing when CONFIG_USER_NS is not defined.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 testcases/kernel/syscalls/getxattr/getxattr05.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/testcases/kernel/syscalls/getxattr/getxattr05.c b/testcases/kernel/syscalls/getxattr/getxattr05.c
index 31b985ee0..c0d339f29 100644
--- a/testcases/kernel/syscalls/getxattr/getxattr05.c
+++ b/testcases/kernel/syscalls/getxattr/getxattr05.c
@@ -37,11 +37,13 @@
 #if defined(HAVE_SYS_XATTR_H) && defined(HAVE_LIBACL)
 
 #define TEST_FILE	"testfile"
+#define SELF_USERNS	"/proc/self/ns/user"
 #define MAX_USERNS	"/proc/sys/user/max_user_namespaces"
 #define UID_MAP	"/proc/self/uid_map"
 
 static acl_t acl;
 static int orig_max_userns = -1;
+static int user_ns_supported = 1;
 
 static struct tcase {
 	/* 0: without userns, 1: with userns */
@@ -110,6 +112,11 @@ static void do_getxattr(unsigned int n)
 	struct tcase *tc = &tcases[n];
 	pid_t pid;
 
+	if (tc->set_userns && !user_ns_supported) {
+		tst_res(TCONF, "user namespace not available");
+		return;
+	}
+
 	pid = SAFE_FORK();
 	if (!pid) {
 		if (tc->set_userns)
@@ -144,10 +151,13 @@ static void setup(void)
 	/* The default value of max_user_namespaces is set to 0 on some distros,
 	 * We need to change the default value to call unshare().
 	 */
-	if (!access(MAX_USERNS, F_OK)) {
+	if (access(SELF_USERNS, F_OK) != 0) {
+		user_ns_supported = 0;
+	} else if (!access(MAX_USERNS, F_OK)) {
 		SAFE_FILE_SCANF(MAX_USERNS, "%d", &orig_max_userns);
 		SAFE_FILE_PRINTF(MAX_USERNS, "%d", 10);
 	}
+
 }
 
 static void cleanup(void)
-- 
2.17.1

[LTP] [PATCH] syscalls/getxattr05: skip userns test cases when not supported

[Cyril Hrubis]

Hi!
Pushed, thanks.

-- 
Cyril Hrubis
chrubis@suse.cz

[LTP] [PATCH] syscalls/getxattr05: skip userns test cases when not supported

[Amir Goldstein]

On Wed, Jun 19, 2019 at 5:23 PM Cyril Hrubis <chrubis@suse.cz> wrote:
>
> Hi!
> Pushed, thanks.

FYI, some more false positive failures I ran into.
I've put them in my skip list, if somebody has time to fix them...

ioctl_ns04 and ioctl_ns06 also fail when CONFIG_USER_NS is not defined:

ioctl_ns04.c:24: CONF: namespace not available
safe_macros.c:225: BROK: ioctl_ns04.c:31:
open(/proc/self/ns/user,0,034641644410) failed: ENOENT

oom03 and oom05 fail when CONFIG_MEMCG is not defined:

mem.c:757: BROK: mount /dev/cgroup: ENOENT

And also does not cleanup /dev/cgroup afterwards, so error them becomes:

safe_macros.c:169: BROK: mem.c:747: mkdir(/dev/cgroup,0777) failed: EEXIST

Thanks,
Amir.