diff options
| author |   Christian Brauner <christian.brauner@ubuntu.com> | 2021-03-20 13:26:22 +0100 |
|---|---|---|
| committer | Christian Brauner <christian.brauner@ubuntu.com> | 2021-03-23 11:13:32 +0100 |
| commit | a65e58e791a1690da8de731c8391816a22f5555c (patch) | |
| tree | 8837e13e3bfc20a2771b108c0ef2c6fc00dcff7d /fs/namei.c | |
| parent | fs: document mapping helpers (diff) | |
| download | linux-dev-a65e58e791a1690da8de731c8391816a22f5555c.tar.xz linux-dev-a65e58e791a1690da8de731c8391816a22f5555c.zip | |
fs: document and rename fsid helpers
Vivek pointed out that the fs{g,u}id_into_mnt() naming scheme can be
misleading as it could be understood as implying they do the exact same
thing as i_{g,u}id_into_mnt(). The original motivation for this naming
scheme was to signal to callers that the helpers will always take care
to map the k{g,u}id such that the ownership is expressed in terms of the
mnt_users.
Get rid of the confusion by renaming those helpers to something more
sensible. Al suggested mapped_fs{g,u}id() which seems a really good fit.
Usually filesystems don't need to bother with these helpers directly
only in some cases where they allocate objects that carry {g,u}ids which
are either filesystem specific (e.g. xfs quota objects) or don't have a
clean set of helpers as inodes have.
Link: https://lore.kernel.org/r/20210320122623.599086-3-christian.brauner@ubuntu.com
Inspired-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Darrick J. Wong <djwong@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Diffstat (limited to '')
| -rw-r--r-- | fs/namei.c | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/fs/namei.c b/fs/namei.c index 216f16e74351..6b5424d34962 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2830,8 +2830,8 @@ static inline int may_create(struct user_namespace *mnt_userns, if (IS_DEADDIR(dir)) return -ENOENT; s_user_ns = dir->i_sb->s_user_ns; - if (!kuid_has_mapping(s_user_ns, fsuid_into_mnt(mnt_userns)) || - !kgid_has_mapping(s_user_ns, fsgid_into_mnt(mnt_userns))) + if (!kuid_has_mapping(s_user_ns, mapped_fsuid(mnt_userns)) || + !kgid_has_mapping(s_user_ns, mapped_fsgid(mnt_userns))) return -EOVERFLOW; return inode_permission(mnt_userns, dir, MAY_WRITE | MAY_EXEC); } @@ -3040,8 +3040,8 @@ static int may_o_create(struct user_namespace *mnt_userns, return error; s_user_ns = dir->dentry->d_sb->s_user_ns; - if (!kuid_has_mapping(s_user_ns, fsuid_into_mnt(mnt_userns)) || - !kgid_has_mapping(s_user_ns, fsgid_into_mnt(mnt_userns))) + if (!kuid_has_mapping(s_user_ns, mapped_fsuid(mnt_userns)) || + !kgid_has_mapping(s_user_ns, mapped_fsgid(mnt_userns))) return -EOVERFLOW; error = inode_permission(mnt_userns, dir->dentry->d_inode, |