xfs: port to vfs{g,u}id_t and associated helpers
PATCH 1/1 xfs: port to vfs{g,u}id_t and associated helpers
A while ago we introduced a dedicated vfs{g,u}id_t type in commit 1e5267cd0895 ("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_t in favor of the new type safe helpers that operate on vfs{g,u}id_t.
--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)
fs/xfs/xfs_inode.c +2 -3
@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;
fs/xfs/xfs_iops.c +4 -2
@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);
@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;
fs/xfs/xfs_itable.c +6 -2
@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;
@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid)); buf->bs_size = ip->i_disk_size; buf->bs_nlink = inode->i_nlink;
base-commit: 7e18e42e4b280c85b76967a9106a13ca61c16179
Signature · 1 line
2.34.1
Quoted 77 lines · Christian Brauner
A while ago we introduced a dedicated vfs{g,u}id_t type in commit
1e5267cd0895("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_t in favor of the new type safe helpers that operate on vfs{g,u}id_t.Cc Dave Chinner <[email protected]>Cc Seth Forshee (Digital Ocean) <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Christian Brauner (Microsoft) <[email protected]>--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)fs/xfs/xfs_inode.c +2
-3@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;fs/xfs/xfs_iops.c +4
-2@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;fs/xfs/xfs_itable.c +4
-0@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);
fs/xfs/xfs_itable.c xfs_bulkstat_one_int( · line 95
Quoted 10 lines
fs/xfs/xfs_itable.c +2
-2/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid));
So we convert i_uid into vfsguid into kuid into the user space visible one here. I wonder if there is any good short cut, as reading this chaing feels odd. But I guess that's what we get for a non-VFS ioctl interface exposint uids/gids.
Quoted 79 lines · Christoph Hellwig
Quoted 78 lines · Christian Brauner
A while ago we introduced a dedicated vfs{g,u}id_t type in commit
1e5267cd0895("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_t in favor of the new type safe helpers that operate on vfs{g,u}id_t.Cc Dave Chinner <[email protected]>Cc Seth Forshee (Digital Ocean) <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Christian Brauner (Microsoft) <[email protected]>--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)fs/xfs/xfs_inode.c +2
-3@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;fs/xfs/xfs_iops.c +4
-2@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;fs/xfs/xfs_itable.c +4
-0@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);
Quoted 14 lines
fs/xfs/xfs_itable.c xfs_bulkstat_one_int( · line 95
Quoted 9 lines
fs/xfs/xfs_itable.c +2
-2/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid));So we convert
i_uidinto vfsguid into kuid into the user space visible one here. I wonder if there is any good short cut, as reading this chaing feels odd. But I guess that's what we get for a non-VFS ioctl interface exposint uids/gids.
We could add a tiny helper for this for sure but I think there's so little need for it that it wouldn't make much sense. But if more cases pop up during conversion of all direct users we should definitely add one.
Quoted 4 lines · Christian Brauner
A while ago we introduced a dedicated vfs{g,u}id_t type in commit
1e5267cd0895("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_t
Does this mean i_uid_into_mnt is going away?
Quoted 74 lines
in favor of the
new type safe helpers that operate on vfs{g,u}id_t.Cc Dave Chinner <[email protected]>Cc Seth Forshee (Digital Ocean) <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Christian Brauner (Microsoft) <[email protected]>--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)fs/xfs/xfs_inode.c +2
-3@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;fs/xfs/xfs_iops.c +4
-2@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;fs/xfs/xfs_itable.c +4
-0@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);
fs/xfs/xfs_itable.c xfs_bulkstat_one_int( · line 95
Quoted 10 lines
fs/xfs/xfs_itable.c +2
-2/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid));
Hmm. How does this work, exactly?
First we map inode.i_uid into a uid in the (file) mount userns, which I guess is what a vfsuid_t is? And then we map that vfsuid_t into a kernel uid (hence kuid_t)? After which we map that kuid_t into the... filesystem(?)'s userns, which is what we return to userspace?
/me confused, need picture. :/
(Admittedly my unfamiliarity with idmapping isn't helping here -- this looks consistent with what the VFS does, but I don't quite grok the layers...)
--D
Quoted 8 lines
fs/xfs/xfs_itable.c +0
-0buf->bs_size = ip->i_disk_size; buf->bs_nlink = inode->i_nlink;base-commit:
7e18e42e4b280c85b76967a9106a13ca61c16179Signature · 1 line
2.34.1
Quoted 7 lines · Darrick J. Wong
Quoted 4 lines · Christian Brauner
A while ago we introduced a dedicated vfs{g,u}id_t type in commit
1e5267cd0895("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_tDoes this mean
i_uid_into_mntis going away?
Yes, all of the non-vfs{g,u}id_t helper will be removed once we finished converting. That hopefully won't take long. I was just a bit busy with posix acls (still am tbh) so didn't get around to removing it completely yet. But the fs/attr.c and fs/posix_acl.c codepaths are already switched.
Quoted 88 lines
Quoted 75 lines
in favor of the
new type safe helpers that operate on vfs{g,u}id_t.Cc Dave Chinner <[email protected]>Cc Seth Forshee (Digital Ocean) <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Christian Brauner (Microsoft) <[email protected]>--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)fs/xfs/xfs_inode.c +2
-3@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;fs/xfs/xfs_iops.c +4
-2@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;fs/xfs/xfs_itable.c +4
-0@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);fs/xfs/xfs_itable.c xfs_bulkstat_one_int( · line 95
Quoted 10 lines
fs/xfs/xfs_itable.c +2
-2/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid));Hmm. How does this work, exactly?
Quoted 2 lines
First we map inode.i_uid into a uid in the (file) mount userns, which I guess is what a
vfsuid_tis? And then we map thatvfsuid_tinto a
Yes.
Quoted 2 lines
kernel uid (hence
kuid_t)? After which we map thatkuid_tinto the... filesystem(?)'s userns, which is what we return to userspace?
Yes.
Quoted 2 lines
/me confused, need picture. :/
Without going into all of [1] the gist is that we undo the filesystem idmapping and apply the mount idmapping. For this we use a new type to clearly indicate that this is a {g,u}id as seen from the mount.
Since this ioctl reports ownership information we now need to translate that vfs{g,u}id_t into a userspace {g,u}id.
A filesystem would usually not see this unless it has something like xfs does where it allows to report ownership outside of the vfs.
Note that the bulkstat code also doesn't currently support idmapped mounts - and won't because it can operate on the whole filesystem not just within a mount. So i_uid_into_vfsuid() is always a nop and just changes the type.
Quoted 4 lines
(Admittedly my unfamiliarity with idmapping isn't helping here -- this looks consistent with what the VFS does, but I don't quite grok the layers...)
[1]: So it is fully documented in
Documentations/filesystems/idmappings.rst.
I'll update that with the new helpers and types in a bit.
Darrick J. Wong
Reviewed-byQuoted 120 lines · Christian Brauner
Quoted 7 lines · Darrick J. Wong
Quoted 4 lines · Christian Brauner
A while ago we introduced a dedicated vfs{g,u}id_t type in commit
1e5267cd0895("mnt_idmapping: add vfs{g,u}id_t"). We already switched over a good part of the VFS. Ultimately we will remove all legacy idmapped mount helpers that operate only on k{g,u}id_tDoes this mean
i_uid_into_mntis going away?Yes, all of the non-vfs{g,u}id_t helper will be removed once we finished converting. That hopefully won't take long. I was just a bit busy with posix acls (still am tbh) so didn't get around to removing it completely yet. But the
fs/attr.candfs/posix_acl.ccodepaths are already switched.Quoted 88 lines
Quoted 75 lines
in favor of the
new type safe helpers that operate on vfs{g,u}id_t.Cc Dave Chinner <[email protected]>Cc Seth Forshee (Digital Ocean) <[email protected]>Cc Christoph Hellwig <[email protected]>Cc Darrick J. Wong <[email protected]>Signed-off-by Christian Brauner (Microsoft) <[email protected]>--- fs/xfs/xfs_inode.c | 5 ++--- fs/xfs/xfs_iops.c | 6 ++++-- fs/xfs/xfs_itable.c | 8 ++++++-- 3 files changed, 12 insertions(+), 7 deletions(-)fs/xfs/xfs_inode.c +2
-3@@ -835,9 +835,8 @@ xfs_init_new_inode(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if (irix_sgid_inherit && - (inode->i_mode &S_ISGID ) && - !in_group_p (i_gid_into_mnt (mnt_userns, inode))) +if (irix_sgid_inherit && (inode->i_mode &S_ISGID ) && + !vfsgid_in_group_p (i_gid_into_vfsgid (mnt_userns, inode))) inode->i_mode &= ~S_ISGID ; ip->i_disk_size =0 ;fs/xfs/xfs_iops.c +4
-2@@ -558,6 +558,8 @@ xfs_vn_getattr(struct inode *inode =d_inode (path->dentry);struct xfs_inode *ip =XFS_I (inode);struct xfs_mount *mp = ip->i_mount; +vfsuid_t vfsuid =i_uid_into_vfsuid (mnt_userns, inode); +vfsgid_t vfsgid =i_gid_into_vfsgid (mnt_userns, inode);trace_xfs_getattr (ip);@@ -568,8 +570,8 @@ xfs_vn_getattr( stat->dev = inode->i_sb->s_dev; stat->mode = inode->i_mode; stat->nlink = inode->i_nlink; - stat->uid =i_uid_into_mnt (mnt_userns, inode); - stat->gid =i_gid_into_mnt (mnt_userns, inode); + stat->uid =vfsuid_into_kuid (vfsuid); + stat->gid =vfsgid_into_kgid (vfsgid); stat->ino = ip->i_ino; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime;fs/xfs/xfs_itable.c +4
-0@@ -66,6 +66,8 @@ xfs_bulkstat_one_int(struct xfs_bulkstat *buf = bc->buf;xfs_extnum_t nextents;int error = -EINVAL ; +vfsuid_t vfsuid; +vfsgid_t vfsgid;if (xfs_internal_inum (mp, ino)) goto out_advance;@@ -81,14 +83,16 @@ xfs_bulkstat_one_int(ASSERT (ip !=NULL );ASSERT (ip->i_imap.im_blkno !=0 ); inode =VFS_I (ip); + vfsuid =i_uid_into_vfsuid (mnt_userns, inode); + vfsgid =i_gid_into_vfsgid (mnt_userns, inode);fs/xfs/xfs_itable.c xfs_bulkstat_one_int( · line 95
Quoted 10 lines
fs/xfs/xfs_itable.c +2
-2/* xfs_iget returns the following without needing * further change. */ buf->bs_projectid = ip->i_projid; buf->bs_ino = ino; - buf->bs_uid =from_kuid (sb_userns,i_uid_into_mnt (mnt_userns, inode)); - buf->bs_gid =from_kgid (sb_userns,i_gid_into_mnt (mnt_userns, inode)); + buf->bs_uid =from_kuid (sb_userns,vfsuid_into_kuid (vfsuid)); + buf->bs_gid =from_kgid (sb_userns,vfsgid_into_kgid (vfsgid));Hmm. How does this work, exactly?
Quoted 2 lines
First we map inode.i_uid into a uid in the (file) mount userns, which I guess is what a
vfsuid_tis? And then we map thatvfsuid_tinto aYes.
Quoted 2 lines
kernel uid (hence
kuid_t)? After which we map thatkuid_tinto the... filesystem(?)'s userns, which is what we return to userspace?Yes.
Quoted 2 lines
/me confused, need picture. :/
Without going into all of [1] the gist is that we undo the filesystem idmapping and apply the mount idmapping. For this we use a new type to clearly indicate that this is a {g,u}id as seen from the mount.
Quoted 5 lines
Since this ioctl reports ownership information we now need to translate that vfs{g,u}id_t into a userspace {g,u}id.
A filesystem would usually not see this unless it has something like xfs does where it allows to report ownership outside of the vfs.
Heh, I'm surprised I understood all that. :)
Quoted 6 lines
Note that the bulkstat code also doesn't currently support idmapped mounts - and won't because it can operate on the whole filesystem not just within a mount. So
i_uid_into_vfsuid()is always a nop and just changes the type.Quoted 1 line
Quoted 7 lines
Quoted 3 lines
(Admittedly my unfamiliarity with idmapping isn't helping here -- this looks consistent with what the VFS does, but I don't quite grok the layers...)
[1]: So it is fully documented in
Documentations/filesystems/idmappings.rst.
I'll update that with the new helpers and types in a bit.
Ok. I did take a look at
https://www.kernel.org/doc/html/latest/filesystems/idmappings.html
but it was the lack of any mention of the new helpers that I got stuck on. :(
With documentation updates :D,
--D
--D