[PATCH v3 02/21] fs: Remove check of s_user_ns for existing mounts in fs_fully_visible()

[Seth Forshee]

fs_fully_visible() ignores MNT_LOCK_NODEV when FS_USERS_DEV_MOUNT
is not set for the filesystem, but there is a bug in the logic
that may cause mounting to fail. It is doing this only when the
existing mount is not in init_user_ns but should check the new
mount instead. But the new mount is always in a non-init
namespace when fs_fully_visible() is called, so that condition
can simply be removed.

Signed-off-by: Seth Forshee <seth.forshee at canonical.com>
---
 fs/namespace.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/fs/namespace.c b/fs/namespace.c
index f20c82f91ecb..c133318bec35 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -3234,8 +3234,7 @@ static bool fs_fully_visible(struct file_system_type *type, int *new_mnt_flags)
 		mnt_flags = mnt->mnt.mnt_flags;
 		if (mnt->mnt.mnt_sb->s_iflags & SB_I_NOEXEC)
 			mnt_flags &= ~(MNT_LOCK_NOSUID | MNT_LOCK_NOEXEC);
-		if (mnt->mnt.mnt_sb->s_user_ns != &init_user_ns &&
-		    !(mnt->mnt.mnt_sb->s_type->fs_flags & FS_USERNS_DEV_MOUNT))
+		if (!(mnt->mnt.mnt_sb->s_type->fs_flags & FS_USERNS_DEV_MOUNT))
 			mnt_flags &= ~(MNT_LOCK_NODEV);
 
 		/* Verify the mount flags are equal to or more permissive
-- 
1.9.1