Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Alexander Mikhalitsyn <aleksandr.mikhalitsyn@canonical.com>
To:     xiubli@redhat.com
Cc:     brauner@kernel.org, stgraber@ubuntu.com,
        linux-fsdevel@vger.kernel.org, Jeff Layton <jlayton@kernel.org>,
        Ilya Dryomov <idryomov@gmail.com>, ceph-devel@vger.kernel.org,
        Alexander Mikhalitsyn <aleksandr.mikhalitsyn@canonical.com>,
        linux-kernel@vger.kernel.org
Subject: [PATCH v10 04/12] ceph: add enable_unsafe_idmap module parameter
Date:   Mon,  7 Aug 2023 15:26:18 +0200
Message-Id: <20230807132626.182101-5-aleksandr.mikhalitsyn@canonical.com>
In-Reply-To: <20230807132626.182101-1-aleksandr.mikhalitsyn@canonical.com>
References: <20230807132626.182101-1-aleksandr.mikhalitsyn@canonical.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

This parameter is used to decide if we allow
to perform IO on idmapped mount in case when MDS lacks
support of CEPHFS_FEATURE_HAS_OWNER_UIDGID feature.

In this case we can't properly handle MDS permission
checks and if UID/GID-based restrictions are enabled
on the MDS side then IO requests which go through an
idmapped mount may fail with -EACCESS/-EPERM.
Fortunately, for most of users it's not a case and
everything should work fine. But we put work "unsafe"
in the module parameter name to warn users about
possible problems with this feature and encourage
update of cephfs MDS.

Cc: Xiubo Li <xiubli@redhat.com>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: ceph-devel@vger.kernel.org
Suggested-by: Stéphane Graber <stgraber@ubuntu.com>
Signed-off-by: Alexander Mikhalitsyn <aleksandr.mikhalitsyn@canonical.com>
Acked-by: Christian Brauner <brauner@kernel.org>
---
 fs/ceph/mds_client.c | 28 +++++++++++++++++++++-------
 fs/ceph/mds_client.h |  2 ++
 fs/ceph/super.c      |  5 +++++
 3 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 5114de5ea65e..90c4b0689cd6 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -2949,6 +2949,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 	int ret;
 	bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME);
 	u16 request_head_version = mds_supported_head_version(session);
+	kuid_t caller_fsuid = req->r_cred->fsuid;
+	kgid_t caller_fsgid = req->r_cred->fsgid;
 
 	ret = set_request_path_attr(mdsc, req->r_inode, req->r_dentry,
 			      req->r_parent, req->r_path1, req->r_ino1.ino,
@@ -3046,12 +3048,24 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 	    !test_bit(CEPHFS_FEATURE_HAS_OWNER_UIDGID, &session->s_features)) {
 		WARN_ON_ONCE(!IS_CEPH_MDS_OP_NEWINODE(req->r_op));
 
-		pr_err_ratelimited_client(cl,
-			"idmapped mount is used and CEPHFS_FEATURE_HAS_OWNER_UIDGID"
-			" is not supported by MDS. Fail request with -EIO.\n");
+		if (enable_unsafe_idmap) {
+			pr_warn_once_client(cl,
+				"idmapped mount is used and CEPHFS_FEATURE_HAS_OWNER_UIDGID"
+				" is not supported by MDS. UID/GID-based restrictions may"
+				" not work properly.\n");
 
-		ret = -EIO;
-		goto out_err;
+			caller_fsuid = from_vfsuid(req->r_mnt_idmap, &init_user_ns,
+						   VFSUIDT_INIT(req->r_cred->fsuid));
+			caller_fsgid = from_vfsgid(req->r_mnt_idmap, &init_user_ns,
+						   VFSGIDT_INIT(req->r_cred->fsgid));
+		} else {
+			pr_err_ratelimited_client(cl,
+				"idmapped mount is used and CEPHFS_FEATURE_HAS_OWNER_UIDGID"
+				" is not supported by MDS. Fail request with -EIO.\n");
+
+			ret = -EIO;
+			goto out_err;
+		}
 	}
 
 	/*
@@ -3103,9 +3117,9 @@ static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
 	lhead->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
 	lhead->op = cpu_to_le32(req->r_op);
 	lhead->caller_uid = cpu_to_le32(from_kuid(&init_user_ns,
-						  req->r_cred->fsuid));
+						  caller_fsuid));
 	lhead->caller_gid = cpu_to_le32(from_kgid(&init_user_ns,
-						  req->r_cred->fsgid));
+						  caller_fsgid));
 	lhead->ino = cpu_to_le64(req->r_deleg_ino);
 	lhead->args = req->r_args;
 
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index 8f683e8203bd..0945ae4cf3c5 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -619,4 +619,6 @@ static inline int ceph_wait_on_async_create(struct inode *inode)
 extern int ceph_wait_on_conflict_unlink(struct dentry *dentry);
 extern u64 ceph_get_deleg_ino(struct ceph_mds_session *session);
 extern int ceph_restore_deleg_ino(struct ceph_mds_session *session, u64 ino);
+
+extern bool enable_unsafe_idmap;
 #endif
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 49fd17fbba9f..18bfdfd48cef 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -1680,6 +1680,11 @@ static const struct kernel_param_ops param_ops_mount_syntax = {
 module_param_cb(mount_syntax_v1, &param_ops_mount_syntax, &mount_support, 0444);
 module_param_cb(mount_syntax_v2, &param_ops_mount_syntax, &mount_support, 0444);
 
+bool enable_unsafe_idmap = false;
+module_param(enable_unsafe_idmap, bool, 0644);
+MODULE_PARM_DESC(enable_unsafe_idmap,
+		 "Allow to use idmapped mounts with MDS without CEPHFS_FEATURE_HAS_OWNER_UIDGID");
+
 module_init(init_ceph);
 module_exit(exit_ceph);
 
-- 
2.34.1