ext4: check the acl's validity before setting

Before setting the acl, call posix_acl_valid() to check if it is valid or not.

---
 fs/ext4/acl.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

index d40c8db..97f7650 100644
@@ -197,6 +197,12 @@ __ext4_set_acl(handle_t *handle, struct inode *inode, int type,
 	size_t size = 0;
 	int error;
 
+	if (acl) {
+		error = posix_acl_valid(acl);
+		if (error < 0)
+			return error;
+	}
+
 	switch (type) {
 	case ACL_TYPE_ACCESS:
 		name_index = EXT4_XATTR_INDEX_POSIX_ACL_ACCESS;

1.6.0.2

Why? posix_acl_xattr_set already does that for you, and it's the only way to feed in an ACL from userspace.

Hi Hellwig,

You are right, this patch is useless. I'm sorry for this spam mail.

But btrfs_set_acl checks the validity of the ACL too. Should we delete it? Another choice is we can check the validity in every fs tree, and delete the check from posix_acl_xattr_set.

How should we do?

Thanks!

Doing it in common code is the right thing. I suspect btrfs having it is an oversight, but please verify it with the btrfs maintainers.