Skip to contentSkip to conversation

ceph: support idmapped mounts

PATCH In review Christian Brauner 5 participants 21 messages Jan 4, 2022 – May 19, 2023 Ceph client

lore
Share & tools
lore
#0 lore raw

cover ceph: support idmapped mounts

From: Christian Brauner <[email protected]>

Hey everyone,

This patch series enables cephfs to support idmapped mounts, i.e. the ability to alter ownership information on a per-mount basis.

Container managers such as LXD support sharaing data via cephfs between the host and unprivileged containers and between unprivileged containers. They may all use different idmappings. Idmapped mounts can be used to create mounts with the idmapping used for the container (or a different one specific to the use-case).

There are in fact more use-cases such as remapping ownership for mountpoints on the host itself to grant or restrict access to different users or to make it possible to enforce that programs running as root will write with a non-zero {g,u}id to disk.

The patch series is simple overall and few changes are needed to cephfs. There is one cephfs specific issue that I would like to discuss and solve which I explain in detail in:

[PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

It has to do with how to handle mds serves which have id-based access restrictions configured. I would ask you to please take a look at the explanation in the aforementioned patch.

The patch series passes the vfs and idmapped mount testsuite as part of xfstests. To run it you will need a config like:

[ceph]
export FSTYP=ceph
export TEST_DIR=/mnt/test
export TEST_DEV=10.103.182.10:6789:/
export TEST_FS_MOUNT_OPTS="-o name=admin,secret=$password

and then simply call

sudo ./check -g idmapped

The patch series is on top of my patches scheduled for v5.17. The easiest way is to either fetch the branch (fs.idmapped.ceph.v1) or the tag (tag.fs.idmapped.ceph.v1):

git fetch git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux.git fs.idmapped.ceph.v1 git fetch git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux.git tag.fs.idmapped.ceph.v1

Thanks!
Christian

Christian Brauner (12):
ceph: stash idmapping in mdsc request
ceph: handle idmapped mounts in create_request_message()
ceph: allow idmapped mknod inode op
ceph: allow idmapped symlink inode op
ceph: allow idmapped mkdir inode op
ceph: allow idmapped rename inode op
ceph: allow idmapped getattr inode op
ceph: allow idmapped permission inode op
ceph: allow idmapped setattr inode op
ceph/acl: allow idmapped set_acl inode op ceph/file: allow idmapped atomic_open inode op
ceph: allow idmapped mounts

 fs/ceph/acl.c        |  2 +-
 fs/ceph/dir.c        |  4 ++++
 fs/ceph/file.c       | 16 ++++++++++++----
 fs/ceph/inode.c      | 15 +++++++++++----
 fs/ceph/mds_client.c | 29 +++++++++++++++++++++++++----
 fs/ceph/mds_client.h |  1 +
 fs/ceph/super.c      |  2 +-
 7 files changed, 55 insertions(+), 14 deletions(-)
Signature · 1 line
2.32.0
#1 lore raw

PATCH 1/12 ceph: stash idmapping in mdsc request

From: Christian Brauner <[email protected]>

When sending a mds request cephfs will send relevant data for the requested operation. For creation requests the caller's fs{g,u}id is used to set the ownership of the newly created filesystem object. For setattr requests the caller can pass in arbitrary {g,u}id values to which the relevant filesystem object is supposed to be changed.

If the caller is performing the relevant operation via an idmapped mount cephfs simply needs to take the idmapping into account when it sends the relevant mds request.

In order to support idmapped mounts for cephfs we stash the idmapping whenever they are relevant for the operation for the duration of the request. Since mds requests can be queued and performed asynchronously we make sure to keep the idmapping around and release it once the request has finished.

In follow-up patches we will use this to send correct ownership information over the wire. This patch just adds the basic infrastructure to keep the idmapping around. The actual conversion patches are all fairly minimal.

---
 fs/ceph/mds_client.c | 7 +++++++
 fs/ceph/mds_client.h | 1 +
 2 files changed, 8 insertions(+)
fs/ceph/mds_client.c +7 -0
index c30eefc0ac19..ae2cc4ce1d48 100644
@@ -12,6 +12,7 @@
 #include <linux/bits.h>
 #include <linux/ktime.h>
 #include <linux/bitmap.h>
+#include <linux/mnt_idmapping.h>
 
 #include "super.h"
 #include "mds_client.h"
@@ -862,6 +863,8 @@ void ceph_mdsc_release_request(struct kref *kref)
 	kfree(req->r_path1);
 	kfree(req->r_path2);
 	put_cred(req->r_cred);
+	if (!initial_idmapping(req->mnt_userns))
+		put_user_ns(req->mnt_userns);
 	if (req->r_pagelist)
 		ceph_pagelist_release(req->r_pagelist);
 	put_request_session(req);
@@ -918,6 +921,10 @@ static void __register_request(struct ceph_mds_client *mdsc,
 	insert_request(&mdsc->request_tree, req);
 
 	req->r_cred = get_current_cred();
+	if (!req->mnt_userns)
+		req->mnt_userns = &init_user_ns;
+	else
+		get_user_ns(req->mnt_userns);
 
 	if (mdsc->oldest_tid == 0 && req->r_op != CEPH_MDS_OP_SETFILELOCK)
 		mdsc->oldest_tid = req->r_tid;
fs/ceph/mds_client.h +1 -0
index 97c7f7bfa55f..1b648645e340 100644
@@ -275,6 +275,7 @@ struct ceph_mds_request {
 	union ceph_mds_request_args r_args;
 	int r_fmode;        /* file mode, if expecting cap */
 	const struct cred *r_cred;
+	struct user_namespace *mnt_userns;
 	int r_request_release_offset;
 	struct timespec64 r_stamp;
 
Signature · 1 line
2.32.0
#3 lore raw

PATCH 2/12 ceph: handle idmapped mounts in create_request_message()

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

From my perspective the root of the problem lies in the fact that creation requests implicitly infer the ownership from the {g,u}id that gets sent along with every mds request.

I have thought of multiple ways of addressing this problem but the one I prefer is to give all mds requests that create a filesystem object a proper, separate {g,u}id field entry in the argument struct. This is, for example how ->setattr mds requests work.

This way the caller's fs{g,u}id can be used consistenly for server access checks and is separated from the ownership for new filesystem objects.

Servers could then be updated to refuse creation requests whenever the {g,u}id used for access checking doesn't match the {g,u}id used for creating the filesystem object just as is done for setattr requests on a uid restricted server. But I am, of course, open to other suggestions.

---
 fs/ceph/mds_client.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
fs/ceph/mds_client.c +18 -4
index ae2cc4ce1d48..1fb43a8fd64c 100644
@@ -2459,6 +2459,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 	void *p, *end;
 	int ret;
 	bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME);
+	kuid_t caller_fsuid;
+	kgid_t caller_fsgid;
 
 	ret = set_request_path_attr(req->r_inode, req->r_dentry,
 			      req->r_parent, req->r_path1, req->r_ino1.ino,
@@ -2524,10 +2526,22 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 
 	head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
 	head->op = cpu_to_le32(req->r_op);
-	head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns,
-						 req->r_cred->fsuid));
-	head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns,
-						 req->r_cred->fsgid));
+	/*
+	 * Inode operations that create filesystem objects based on the
+	 * caller's fs{g,u}id like ->mknod(), ->create(), ->mkdir() etc. don't
+	 * have separate {g,u}id fields in their respective structs in the
+	 * ceph_mds_request_args union. Instead the caller_{g,u}id field is
+	 * used to set ownership of the newly created inode by the mds server.
+	 * For these inode operations we need to send the mapped fs{g,u}id over
+	 * the wire. For other cases we simple set req->mnt_userns to the
+	 * initial idmapping meaning the unmapped fs{g,u}id is sent.
+	 */
+	caller_fsuid = mapped_kuid_user(req->mnt_userns, &init_user_ns,
+					req->r_cred->fsuid);
+	caller_fsgid = mapped_kgid_user(req->mnt_userns, &init_user_ns,
+					req->r_cred->fsgid);
+	head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, caller_fsuid));
+	head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, caller_fsgid));
 	head->ino = cpu_to_le64(req->r_deleg_ino);
 	head->args = req->r_args;
 
Signature · 1 line
2.32.0
#11 lore raw

PATCH 3/12 ceph: allow idmapped mknod inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_mknod() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/dir.c | 1 +
 1 file changed, 1 insertion(+)
fs/ceph/dir.c +1 -0
index 133dbd9338e7..7278863fbd4a 100644
@@ -871,6 +871,7 @@ static int ceph_mknod(struct user_namespace *mnt_userns, struct inode *dir,
 	req->r_parent = dir;
 	ihold(dir);
 	set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags);
+	req->mnt_userns = mnt_userns;
 	req->r_args.mknod.mode = cpu_to_le32(mode);
 	req->r_args.mknod.rdev = cpu_to_le32(rdev);
 	req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL;
Signature · 1 line
2.32.0
#12 lore raw

PATCH 4/12 ceph: allow idmapped symlink inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_symlink() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/dir.c | 1 +
 1 file changed, 1 insertion(+)
fs/ceph/dir.c +1 -0
index 7278863fbd4a..67ce448a9ff5 100644
@@ -939,6 +939,7 @@ static int ceph_symlink(struct user_namespace *mnt_userns, struct inode *dir,
 	req->r_num_caps = 2;
 	req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL;
 	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
+	req->mnt_userns = mnt_userns;
 	if (as_ctx.pagelist) {
 		req->r_pagelist = as_ctx.pagelist;
 		as_ctx.pagelist = NULL;
Signature · 1 line
2.32.0
#13 lore raw

PATCH 5/12 ceph: allow idmapped mkdir inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_mkdir() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/dir.c | 1 +
 1 file changed, 1 insertion(+)
fs/ceph/dir.c +1 -0
index 67ce448a9ff5..210257afb346 100644
@@ -1002,6 +1002,7 @@ static int ceph_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
 	ihold(dir);
 	set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags);
 	req->r_args.mkdir.mode = cpu_to_le32(mode);
+	req->mnt_userns = mnt_userns;
 	req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL;
 	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
 	if (as_ctx.pagelist) {
Signature · 1 line
2.32.0
#14 lore raw

PATCH 6/12 ceph: allow idmapped rename inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_rename() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/dir.c | 1 +
 1 file changed, 1 insertion(+)
fs/ceph/dir.c +1 -0
index 210257afb346..9463c960f03b 100644
@@ -1248,6 +1248,7 @@ static int ceph_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
 	req->r_old_dentry_unless = CEPH_CAP_FILE_EXCL;
 	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
 	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
+	req->mnt_userns = mnt_userns;
 	/* release LINK_RDCACHE on source inode (mds will lock it) */
 	req->r_old_inode_drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL;
 	if (d_really_is_positive(new_dentry)) {
Signature · 1 line
2.32.0
#15 lore raw

PATCH 7/12 ceph: allow idmapped getattr inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_getattr() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ceph/inode.c +1 -1
index e3322fcb2e8d..f648aecc5827 100644
@@ -2364,7 +2364,7 @@ int ceph_getattr(struct user_namespace *mnt_userns, const struct path *path,
 			return err;
 	}
 
-	generic_fillattr(&init_user_ns, inode, stat);
+	generic_fillattr(mnt_userns, inode, stat);
 	stat->ino = ceph_present_inode(inode);
 
 	/*
Signature · 1 line
2.32.0
#16 lore raw

PATCH 8/12 ceph: allow idmapped permission inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_permission() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ceph/inode.c +1 -1
index f648aecc5827..d5cecc3519fa 100644
@@ -2307,7 +2307,7 @@ int ceph_permission(struct user_namespace *mnt_userns, struct inode *inode,
 	err = ceph_do_getattr(inode, CEPH_CAP_AUTH_SHARED, false);
 
 	if (!err)
-		err = generic_permission(&init_user_ns, inode, mask);
+		err = generic_permission(mnt_userns, inode, mask);
 	return err;
 }
 
Signature · 1 line
2.32.0
#17 lore raw

PATCH 9/12 ceph: allow idmapped setattr inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable __ceph_setattr() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/inode.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/inode.c +9 -2
index d5cecc3519fa..658b620efd50 100644
@@ -2032,6 +2032,13 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr)
 
 	dout("setattr %p issued %s\n", inode, ceph_cap_string(issued));
 
+	/*
+	 * The attr->ia_{g,u}id members contain the target {g,u}id we're
+	 * sending over the wire. The mount idmapping only matters when we
+	 * create new filesystem objects based on the caller's mapped
+	 * fs{g,u}id.
+	 */
+	req->mnt_userns = &init_user_ns;
 	if (ia_valid & ATTR_UID) {
 		dout("setattr %p uid %d -> %d\n", inode,
 		     from_kuid(&init_user_ns, inode->i_uid),
@@ -2222,7 +2229,7 @@ int ceph_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
 	if (ceph_inode_is_shutdown(inode))
 		return -ESTALE;
 
-	err = setattr_prepare(&init_user_ns, dentry, attr);
+	err = setattr_prepare(mnt_userns, dentry, attr);
 	if (err != 0)
 		return err;
 
@@ -2237,7 +2244,7 @@ int ceph_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
 	err = __ceph_setattr(inode, attr);
 
 	if (err >= 0 && (attr->ia_valid & ATTR_MODE))
-		err = posix_acl_chmod(&init_user_ns, inode, attr->ia_mode);
+		err = posix_acl_chmod(mnt_userns, inode, attr->ia_mode);
 
 	return err;
 }
Signature · 1 line
2.32.0
#18 lore raw

PATCH 10/12 ceph/acl: allow idmapped set_acl inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_set_acl() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/acl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ceph/acl.c +1 -1
index f4fc8e0b847c..7957d44bb27c 100644
@@ -104,7 +104,7 @@ int ceph_set_acl(struct user_namespace *mnt_userns, struct inode *inode,
 	case ACL_TYPE_ACCESS:
 		name = XATTR_NAME_POSIX_ACL_ACCESS;
 		if (acl) {
-			ret = posix_acl_update_mode(&init_user_ns, inode,
+			ret = posix_acl_update_mode(mnt_userns, inode,
 						    &new_mode, &acl);
 			if (ret)
 				goto out;
Signature · 1 line
2.32.0
#19 lore raw

PATCH 11/12 ceph/file: allow idmapped atomic_open inode op

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Enable ceph_atomic_open() to handle idmapped mounts. This is just a matter of passing down the mount's idmapping.

---
 fs/ceph/file.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)
fs/ceph/file.c +12 -4
index c138e8126286..7fecb41796c7 100644
@@ -608,7 +608,10 @@ static int ceph_finish_async_create(struct inode *dir, struct dentry *dentry,
 	in.truncate_seq = cpu_to_le32(1);
 	in.truncate_size = cpu_to_le64(-1ULL);
 	in.xattr_version = cpu_to_le64(1);
-	in.uid = cpu_to_le32(from_kuid(&init_user_ns, current_fsuid()));
+	in.uid = cpu_to_le32(from_kuid(&init_user_ns,
+				       mapped_kuid_user(req->mnt_userns,
+							&init_user_ns,
+							current_fsuid())));
 	if (dir->i_mode & S_ISGID) {
 		in.gid = cpu_to_le32(from_kgid(&init_user_ns, dir->i_gid));
 
@@ -616,11 +619,14 @@ static int ceph_finish_async_create(struct inode *dir, struct dentry *dentry,
 		if (S_ISDIR(mode))
 			mode |= S_ISGID;
 		else if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP) &&
-			 !in_group_p(dir->i_gid) &&
-			 !capable_wrt_inode_uidgid(&init_user_ns, dir, CAP_FSETID))
+			 !in_group_p(i_gid_into_mnt(req->mnt_userns, dir)) &&
+			 !capable_wrt_inode_uidgid(req->mnt_userns, dir, CAP_FSETID))
 			mode &= ~S_ISGID;
 	} else {
-		in.gid = cpu_to_le32(from_kgid(&init_user_ns, current_fsgid()));
+		in.gid = cpu_to_le32(from_kgid(&init_user_ns,
+				     mapped_kgid_user(req->mnt_userns,
+						      &init_user_ns,
+						      current_fsgid())));
 	}
 	in.mode = cpu_to_le32((u32)mode);
 
@@ -677,6 +683,7 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
 		     struct file *file, unsigned flags, umode_t mode)
 {
 	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
+	struct user_namespace *mnt_userns = file_mnt_user_ns(file);
 	struct ceph_mds_client *mdsc = fsc->mdsc;
 	struct ceph_mds_request *req;
 	struct dentry *dn;
@@ -719,6 +726,7 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
 		mask |= CEPH_CAP_XATTR_SHARED;
 	req->r_args.open.mask = cpu_to_le32(mask);
 	req->r_parent = dir;
+	req->mnt_userns = mnt_userns;
 	ihold(dir);
 
 	if (flags & O_CREAT) {
Signature · 1 line
2.32.0
#20 lore raw

PATCH 12/12 ceph: allow idmapped mounts

In reply to Christian Brauner

From: Christian Brauner <[email protected]>

Now that we converted cephfs internally to account for idmapped mounts allow the creation of idmapped mounts on by setting the FS_ALLOW_IDMAP flag.

---
 fs/ceph/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ceph/super.c +1 -1
index bab61232dc5a..eda4a26fcb0c 100644
@@ -1242,7 +1242,7 @@ static struct file_system_type ceph_fs_type = {
 	.name		= "ceph",
 	.init_fs_context = ceph_init_fs_context,
 	.kill_sb	= ceph_kill_sb,
-	.fs_flags	= FS_RENAME_DOES_D_MOVE,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE | FS_ALLOW_IDMAP,
 };
 MODULE_ALIAS_FS("ceph");
 
Signature · 1 line
2.32.0
#2 lore raw

Re: [PATCH 01/12] ceph: stash idmapping in mdsc request

In reply to Christian Brauner
Quoted 45 lines · Christian Brauner

From: Christian Brauner <[email protected]>

When sending a mds request cephfs will send relevant data for the requested operation. For creation requests the caller's fs{g,u}id is used to set the ownership of the newly created filesystem object. For setattr requests the caller can pass in arbitrary {g,u}id values to which the relevant filesystem object is supposed to be changed.

If the caller is performing the relevant operation via an idmapped mount cephfs simply needs to take the idmapping into account when it sends the relevant mds request.

In order to support idmapped mounts for cephfs we stash the idmapping whenever they are relevant for the operation for the duration of the request. Since mds requests can be queued and performed asynchronously we make sure to keep the idmapping around and release it once the request has finished.

In follow-up patches we will use this to send correct ownership information over the wire. This patch just adds the basic infrastructure to keep the idmapping around. The actual conversion patches are all fairly minimal.

---
 fs/ceph/mds_client.c | 7 +++++++
 fs/ceph/mds_client.h | 1 +
 2 files changed, 8 insertions(+)
fs/ceph/mds_client.c +1 -0
index c30eefc0ac19..ae2cc4ce1d48 100644
@@ -12,6 +12,7 @@
 #include <linux/bits.h>
 #include <linux/ktime.h>
 #include <linux/bitmap.h>
+#include <linux/mnt_idmapping.h>
 
 #include "super.h"
 #include "mds_client.h"

fs/ceph/mds_client.c void ceph_mdsc_release_request(struct kref *kref) · line 867

Quoted 5 lines
fs/ceph/mds_client.c +2 -0
@@ -862,6 +863,8 @@ void ceph_mdsc_release_request(struct kref *kref)
 	kfree(req->r_path1);
 	kfree(req->r_path2);
 	put_cred(req->r_cred);
+	if (!initial_idmapping(req->mnt_userns))
+		put_user_ns(req->mnt_userns);

I assume initial_idmapping returns true if it's init_user_ns ?

Quoted 19 lines
fs/ceph/mds_client.c +4 -0
 	if (req->r_pagelist)
 		ceph_pagelist_release(req->r_pagelist);
 	put_request_session(req);
@@ -918,6 +921,10 @@ static void __register_request(struct ceph_mds_client *mdsc,
 	insert_request(&mdsc->request_tree, req);
 
 	req->r_cred = get_current_cred();
+	if (!req->mnt_userns)
+		req->mnt_userns = &init_user_ns;
+	else
+		get_user_ns(req->mnt_userns);
 
 	if (mdsc->oldest_tid == 0 && req->r_op != CEPH_MDS_OP_SETFILELOCK)
 		mdsc->oldest_tid = req->r_tid;
fs/ceph/mds_client.h +0 -0
index 97c7f7bfa55f..1b648645e340 100644

fs/ceph/mds_client.h struct ceph_mds_request { · line 278

Quoted 4 lines
fs/ceph/mds_client.h +1 -0
@@ -275,6 +275,7 @@ struct ceph_mds_request {
 	union ceph_mds_request_args r_args;
 	int r_fmode;        /* file mode, if expecting cap */
 	const struct cred *r_cred;
+	struct user_namespace *mnt_userns;

Can we call this r_mnt_userns to match the other fields ?

Quoted 3 lines
fs/ceph/mds_client.h +0 -0
 	int r_request_release_offset;
 	struct timespec64 r_stamp;
 
Signature · 1 line
Jeff Layton <[email protected]>
#4 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

In reply to Christian Brauner
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Quoted 66 lines

From my perspective the root of the problem lies in the fact that creation requests implicitly infer the ownership from the {g,u}id that gets sent along with every mds request.

I have thought of multiple ways of addressing this problem but the one I prefer is to give all mds requests that create a filesystem object a proper, separate {g,u}id field entry in the argument struct. This is, for example how ->setattr mds requests work.

This way the caller's fs{g,u}id can be used consistenly for server access checks and is separated from the ownership for new filesystem objects.

Servers could then be updated to refuse creation requests whenever the {g,u}id used for access checking doesn't match the {g,u}id used for creating the filesystem object just as is done for setattr requests on a uid restricted server. But I am, of course, open to other suggestions.

---
 fs/ceph/mds_client.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
fs/ceph/mds_client.c +18 -4
index ae2cc4ce1d48..1fb43a8fd64c 100644
@@ -2459,6 +2459,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 	void *p, *end;
 	int ret;
 	bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME);
+	kuid_t caller_fsuid;
+	kgid_t caller_fsgid;
 
 	ret = set_request_path_attr(req->r_inode, req->r_dentry,
 			      req->r_parent, req->r_path1, req->r_ino1.ino,
@@ -2524,10 +2526,22 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 
 	head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
 	head->op = cpu_to_le32(req->r_op);
-	head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns,
-						 req->r_cred->fsuid));
-	head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns,
-						 req->r_cred->fsgid));
+	/*
+	 * Inode operations that create filesystem objects based on the
+	 * caller's fs{g,u}id like ->mknod(), ->create(), ->mkdir() etc. don't
+	 * have separate {g,u}id fields in their respective structs in the
+	 * ceph_mds_request_args union. Instead the caller_{g,u}id field is
+	 * used to set ownership of the newly created inode by the mds server.
+	 * For these inode operations we need to send the mapped fs{g,u}id over
+	 * the wire. For other cases we simple set req->mnt_userns to the
+	 * initial idmapping meaning the unmapped fs{g,u}id is sent.
+	 */
+	caller_fsuid = mapped_kuid_user(req->mnt_userns, &init_user_ns,
+					req->r_cred->fsuid);
+	caller_fsgid = mapped_kgid_user(req->mnt_userns, &init_user_ns,
+					req->r_cred->fsgid);
+	head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, caller_fsuid));
+	head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, caller_fsgid));
 	head->ino = cpu_to_le64(req->r_deleg_ino);
 	head->args = req->r_args;
 
Signature · 1 line
Jeff Layton <[email protected]>
#5 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

Quoted 82 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Yeah, I'm confused. We want the fs {g,u}id to be consistent throughout the request pipeline and to reflect the actual Ceph user all the way through the server-side pipeline. What if client.greg is only authorized to work as uid 12345 and access /users/greg/; why would you send in a bunch of requests as root just because I mounted that way inside my own space?

This might be more obvious in the userspace Client, which is already set up to be friendlier to mapped users for Ganesha etc:
mknod (https://github.com/ceph/ceph/blob/master/src/client/Client.cc#L7297)
and similar calls receive a "UserPerm" from the caller specifying who the call should be performed as, and they pass that in to the generic make_requst() function
(https://github.com/ceph/ceph/blob/master/src/client/Client.cc#L1778)
which uses it to set the uid and gid fields you found in the message. -Greg

Quoted 70 lines
Quoted 56 lines

From my perspective the root of the problem lies in the fact that creation requests implicitly infer the ownership from the {g,u}id that gets sent along with every mds request.

I have thought of multiple ways of addressing this problem but the one I prefer is to give all mds requests that create a filesystem object a proper, separate {g,u}id field entry in the argument struct. This is, for example how ->setattr mds requests work.

This way the caller's fs{g,u}id can be used consistenly for server access checks and is separated from the ownership for new filesystem objects.

Servers could then be updated to refuse creation requests whenever the {g,u}id used for access checking doesn't match the {g,u}id used for creating the filesystem object just as is done for setattr requests on a uid restricted server. But I am, of course, open to other suggestions.

---
 fs/ceph/mds_client.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
fs/ceph/mds_client.c +11 -4
index ae2cc4ce1d48..1fb43a8fd64c 100644
@@ -2459,6 +2459,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
      void *p, *end;
      int ret;
      bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME);
+     kuid_t caller_fsuid;
+     kgid_t caller_fsgid;
 
      ret = set_request_path_attr(req->r_inode, req->r_dentry,
                            req->r_parent, req->r_path1, req->r_ino1.ino,
@@ -2524,10 +2526,22 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 
      head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
      head->op = cpu_to_le32(req->r_op);
-     head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns,
-                                              req->r_cred->fsuid));
-     head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns,
-                                              req->r_cred->fsgid));
+     /*
+      * Inode operations that create filesystem objects based on the
+      * caller's fs{g,u}id like ->mknod(), ->create(), ->mkdir() etc. don't
+      * have separate {g,u}id fields in their respective structs in the
+      * ceph_mds_request_args union. Instead the caller_{g,u}id field is
+      * used to set ownership of the newly created inode by the mds server.
+      * For these inode operations we need to send the mapped fs{g,u}id over
+      * the wire. For other cases we simple set req->mnt_userns to the
+      * initial idmapping meaning the unmapped fs{g,u}id is sent.

fs/ceph/mds_client.c static struct ceph_msg *create_request_message(struct ceph_mds_session *session, · line 2547

Quoted 10 lines
fs/ceph/mds_client.c +7 -0
+      */
+     caller_fsuid = mapped_kuid_user(req->mnt_userns, &init_user_ns,
+                                     req->r_cred->fsuid);
+     caller_fsgid = mapped_kgid_user(req->mnt_userns, &init_user_ns,
+                                     req->r_cred->fsgid);
+     head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, caller_fsuid));
+     head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, caller_fsgid));
      head->ino = cpu_to_le64(req->r_deleg_ino);
      head->args = req->r_args;
 

--
Jeff Layton <[email protected]>

#7 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

In reply to Jeff Layton
Quoted 81 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Since the idmapping is a property of the mount and not a property of the caller the caller's fs{g,u}id aren't mapped. What is mapped are the inode's i{g,u}id when accessed from a particular mount.

The fs{g,u}id are only ever mapped when a new filesystem object is created. So if I have an idmapped mount that makes it so that files owned by 1000 on-disk appear to be owned by uid 0 then a user with uid 0 creating a new file will create files with uid 1000 on-disk when going through that mount. For cephfs that'd be the uid we would be sending with creation requests as I've currently written it.

So then when the user looks at the file it created it will see it as being owned by uid 0 from that idmapped mount (whereas on-disk it's 1000). But the user's fs{g,u}id isn't per se changed when going through that mount. So in my opinion I was thinking that the server with access permissions set would want to always check permissions on the users "raw" fs{g,u}id. That would mean I'd have to change the patch obviously. My suggestion was to send the {g,u}id the file will be created with separately. The alternative would be to not just pass the idmapping into the creation iop's but into all iops so that we can always map it for cephfs. But this would mean a lot of vfs changes for one filesystem. So if we could first explore alternatives approaches I'd be grateful.

(I'll be traveling for the latter half of this week starting today at CET afternoon so apologies but I'll probably take some time to respond.)

#6 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

In reply to Gregory Farnum
Quoted 95 lines · Gregory Farnum
Quoted 82 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Yeah, I'm confused. We want the fs {g,u}id to be consistent throughout the request pipeline and to reflect the actual Ceph user all the way through the server-side pipeline. What if client.greg is only authorized to work as uid 12345 and access /users/greg/; why would you send in a bunch of requests as root just because I mounted that way inside my own space?

Quoted 8 lines

This might be more obvious in the userspace Client, which is already set up to be friendlier to mapped users for Ganesha etc:
mknod (https://github.com/ceph/ceph/blob/master/src/client/Client.cc#L7297)
and similar calls receive a "UserPerm" from the caller specifying who the call should be performed as, and they pass that in to the generic make_requst() function
(https://github.com/ceph/ceph/blob/master/src/client/Client.cc#L1778)
which uses it to set the uid and gid fields you found in the message.

Thank you for those links. I think I read through this code before and I'll give it another read.

#8 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

In reply to Christian Brauner
Quoted 98 lines · Christian Brauner
Quoted 81 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Since the idmapping is a property of the mount and not a property of the caller the caller's fs{g,u}id aren't mapped. What is mapped are the inode's i{g,u}id when accessed from a particular mount.

The fs{g,u}id are only ever mapped when a new filesystem object is created. So if I have an idmapped mount that makes it so that files owned by 1000 on-disk appear to be owned by uid 0 then a user with uid 0 creating a new file will create files with uid 1000 on-disk when going through that mount. For cephfs that'd be the uid we would be sending with creation requests as I've currently written it.

Quoted 12 lines

So then when the user looks at the file it created it will see it as being owned by uid 0 from that idmapped mount (whereas on-disk it's 1000). But the user's fs{g,u}id isn't per se changed when going through that mount. So in my opinion I was thinking that the server with access permissions set would want to always check permissions on the users "raw" fs{g,u}id. That would mean I'd have to change the patch obviously. My suggestion was to send the {g,u}id the file will be created with separately. The alternative would be to not just pass the idmapping into the creation iop's but into all iops so that we can always map it for cephfs. But this would mean a lot of vfs changes for one filesystem. So if we could first explore alternatives approaches I'd be grateful.

You'll probably need to do this for NFS anyway, if you have plans in that direction. Extending the protocol there will be much more difficult. I think that approach sounds much cleaner overall.

Quoted 2 lines

(I'll be traveling for the latter half of this week starting today at CET afternoon so apologies but I'll probably take some time to respond.)

Ok. I guess you can get away with this on a local fs because the backend storage doesn't really care about uid/gids at all. The only permission checking is done in the kernel and you (presumably) can just map the inode's uid/gid prior to checking permissions.

I'm a little confused as to what you mean by "raw" id here. In your earlier example with a mapping of 1000:0:1, which one is the raw id for the actual user?

Signature · 1 line
Jeff Layton <[email protected]>
#9 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

Quoted 112 lines · Jeff Layton
Quoted 98 lines · Christian Brauner
Quoted 81 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Since the idmapping is a property of the mount and not a property of the caller the caller's fs{g,u}id aren't mapped. What is mapped are the inode's i{g,u}id when accessed from a particular mount.

The fs{g,u}id are only ever mapped when a new filesystem object is created. So if I have an idmapped mount that makes it so that files owned by 1000 on-disk appear to be owned by uid 0 then a user with uid 0 creating a new file will create files with uid 1000 on-disk when going through that mount. For cephfs that'd be the uid we would be sending with creation requests as I've currently written it.

Quoted 12 lines

So then when the user looks at the file it created it will see it as being owned by uid 0 from that idmapped mount (whereas on-disk it's 1000). But the user's fs{g,u}id isn't per se changed when going through that mount. So in my opinion I was thinking that the server with access permissions set would want to always check permissions on the users "raw" fs{g,u}id. That would mean I'd have to change the patch obviously. My suggestion was to send the {g,u}id the file will be created with separately. The alternative would be to not just pass the idmapping into the creation iop's but into all iops so that we can always map it for cephfs. But this would mean a lot of vfs changes for one filesystem. So if we could first explore alternatives approaches I'd be grateful.

Quoted 3 lines

You'll probably need to do this for NFS anyway, if you have plans in that direction. Extending the protocol there will be much more difficult. I think that approach sounds much cleaner overall.

Ok. Is it ok if I take a little while to work on this?
I have some other work I need to be looking at first and then I have Februrary "free".

Quoted 8 lines
Quoted 3 lines

(I'll be traveling for the latter half of this week starting today at CET afternoon so apologies but I'll probably take some time to respond.)

Ok. I guess you can get away with this on a local fs because the backend storage doesn't really care about uid/gids at all. The only permission checking is done in the kernel and you (presumably) can just map the inode's uid/gid prior to checking permissions.

Yes, we always map the inode as that's semantically cleaner and easier to reason about in my opinion.

Quoted 4 lines

I'm a little confused as to what you mean by "raw" id here. In your earlier example with a mapping of 1000:0:1, which one is the raw id for the actual user?

Oh, sorry. In this context I really just meant the values gotten from current_fs{g,u}id() as they are sent now.

#10 lore raw

Re: [PATCH 02/12] ceph: handle idmapped mounts in create_request_message()

In reply to Christian Brauner

Dear colleagues,

Quoted 98 lines · On Wed, 5 Jan 2022 15:10:23 +0100 Christian Brauner…
Quoted 81 lines · Jeff Layton
Quoted 73 lines · Christian Brauner

From: Christian Brauner <[email protected]>

Inode operations that create a new filesystem object such as ->mknod, ->create, ->mkdir() and others don't take a {g,u}id argument explicitly. Instead the caller's fs{g,u}id is used for the {g,u}id of the new filesystem object.

Cephfs mds creation request argument structures mirror this filesystem behavior. They don't encode a {g,u}id explicitly. Instead the caller's fs{g,u}id that is always sent as part of any mds request is used by the servers to set the {g,u}id of the new filesystem object.

In order to ensure that the correct {g,u}id is used map the caller's fs{g,u}id for creation requests. This doesn't require complex changes. It suffices to pass in the relevant idmapping recorded in the request message. If this request message was triggered from an inode operation that creates filesystem objects it will have passed down the relevant idmaping. If this is a request message that was triggered from an inode operation that doens't need to take idmappings into account the initial idmapping is passed down which is an identity mapping and thus is guaranteed to leave the caller's fs{g,u}id unchanged.,u}id is sent.

The last few weeks before Christmas 2021 I have spent time not just reading and poking the cephfs kernel code but also took a look at the ceph mds server userspace to ensure I didn't miss some subtlety.

This made me aware of one complication to solve. All requests send the caller's fs{g,u}id over the wire. The caller's fs{g,u}id matters for the server in exactly two cases:

1. to set the ownership for creation requests
2. to determine whether this client is allowed access on this server

Case 1. we already covered and explained. Case 2. is only relevant for servers where an explicit uid access restriction has been set. That is to say the mds server restricts access to requests coming from a specific uid. Servers without uid restrictions will grant access to requests from any uid by setting MDS_AUTH_UID_ANY.

Case 2. introduces the complication because the caller's fs{g,u}id is not just used to record ownership but also serves as the {g,u}id used when checking access to the server.

Consider a user mounting a cephfs client and creating an idmapped mount from it that maps files owned by uid 1000 to be owned uid 0:

mount -t cephfs -o [...] /unmapped
mount-idmapped --map-mount 1000:0:1 /idmapped

That is to say if the mounted cephfs filesystem contains a file "file1" which is owned by uid 1000:

- looking at it via /unmapped/file1 will report it as owned by uid 1000 (One can think of this as the on-disk value.)
- looking at it via /idmapped/file1 will report it as owned by uid 0

Now, consider creating new files via the idmapped mount at /idmapped. When a caller with fs{g,u}id 1000 creates a file "file2" by going through the idmapped mount mounted at /idmapped it will create a file that is owned by uid 1000 on-disk, i.e.:

- looking at it via /unmapped/file2 will report it as owned by uid 1000
- looking at it via /idmapped/file2 will report it as owned by uid 0

Now consider an mds server that has a uid access restriction set and only grants access to requests from uid 0.

If the client sends a creation request for a file e.g. /idmapped/file2 it will send the caller's fs{g,u}id idmapped according to the idmapped mount. So if the caller has fs{g,u}id 1000 it will be mapped to {g,u}id 0 in the idmapped mount and will be sent over the wire allowing the caller access to the mds server.

Quoted 6 lines

However, if the caller is not issuing a creation request the caller's fs{g,u}id will be send without the mount's idmapping applied. So if the caller that just successfully created a new file on the restricted mds server sends a request as fs{g,u}id 1000 access will be refused. This however is inconsistent.

Quoted 4 lines

IDGI, why would you send the fs{g,u}id without the mount's idmapping applied in this case? ISTM that idmapping is wholly a client-side feature, and that you should always map id's regardless of whether you're creating or not.

Since the idmapping is a property of the mount and not a property of the caller the caller's fs{g,u}id aren't mapped. What is mapped are the inode's i{g,u}id when accessed from a particular mount.

The fs{g,u}id are only ever mapped when a new filesystem object is created. So if I have an idmapped mount that makes it so that files owned by 1000 on-disk appear to be owned by uid 0 then a user with uid 0 creating a new file will create files with uid 1000 on-disk when going through that mount. For cephfs that'd be the uid we would be sending with creation requests as I've currently written it.

Quoted 11 lines

So then when the user looks at the file it created it will see it as being owned by uid 0 from that idmapped mount (whereas on-disk it's 1000). But the user's fs{g,u}id isn't per se changed when going through that mount. So in my opinion I was thinking that the server with access permissions set would want to always check permissions on the users "raw" fs{g,u}id. That would mean I'd have to change the patch obviously. My suggestion was to send the {g,u}id the file will be created with separately. The alternative would be to not just pass the idmapping into the creation iop's but into all iops so that we can always map it for cephfs. But this would mean a lot of vfs changes for one filesystem. So if we could first explore alternatives approaches I'd be grateful.

I can't understand which kind of operations we are talking about here.
Right now almost all inode_operations are taking struct mnt_idmap as a parameter (at the moment of this series was posted it was struct user_namespace, but that's not important).

The only iops those are not taking idmap is lookup, readlink, fiemap, update_time, atomic_open and a few more. So, we want to pass struct mnt_idmap to them to always map current_fs{g,u}id according to a mount idmapping? As Christian pointed above:

Quoted 3 lines

Since the idmapping is a property of the mount and not a property of the caller the caller's fs{g,u}id aren't mapped. What is mapped are the inode's i{g,u}id when accessed from a particular mount.

If we want to go this way then we don't need to pass mnt_idmap to any additional inode ops and the current approach works fine. Please, correct me if I'm wrong.

Quoted 4 lines

(I'll be traveling for the latter half of this week starting today at CET afternoon so apologies but I'll probably take some time to respond.)

P.S.

I'm trying to make a respin for this series, I've made a formal rebase on top of the current Linux kernel tree and fixed it according to the Jeff's review comment:
https://lore.kernel.org/all/[email protected]/

This thing is really important for LXD/LXC project so I'll be happy to help with pushing this forward.

Current tree:
https://github.com/mihalicyn/linux/commits/fs.idmapped.ceph.v2

Kind regards,
Alex