Messages in this thread Patch in this message | | | From | Seth Forshee <> | Subject | [PATCH] userns: Fix k[ug]id_has_mapping() for CONFIG_USER_NS=n | Date | Wed, 3 Sep 2014 10:59:34 -0500 |
| |
Calling k[ug]id_has_mapping() with INVALID_[UG]ID returns false if user namespaces are enabled and true if they are disabled. INVALID_[UG]ID is never included in any mapping, so fix the CONFIG_USER_NS=n to return false for invalid ids.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> ---
Hi Eric,
I noticed this inconsistency when working on support for user namespaces in fuse. As far as I know this isn't breaking anything, but it seems like invalid should be invalid even with CONFIG_USER_NS=n. A compile/boot test with this patch didn't reveal any breakage.
Thanks, Seth
include/linux/uidgid.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/linux/uidgid.h b/include/linux/uidgid.h index 2d1f9b627f91..6c302267f7cc 100644 --- a/include/linux/uidgid.h +++ b/include/linux/uidgid.h @@ -165,12 +165,12 @@ static inline gid_t from_kgid_munged(struct user_namespace *to, kgid_t kgid) static inline bool kuid_has_mapping(struct user_namespace *ns, kuid_t uid) { - return true; + return uid_valid(uid); } static inline bool kgid_has_mapping(struct user_namespace *ns, kgid_t gid) { - return true; + return gid_valid(gid); } #endif /* CONFIG_USER_NS */ -- 1.9.1
| |