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:   =?UTF-8?q?Lu=C3=ADs=20Henriques?= <lhenriques@suse.de>
To:     Xiubo Li <xiubli@redhat.com>, Ilya Dryomov <idryomov@gmail.com>,
        Jeff Layton <jlayton@kernel.org>
Cc:     ceph-devel@vger.kernel.org, linux-kernel@vger.kernel.org,
        =?UTF-8?q?Lu=C3=ADs=20Henriques?= <lhenriques@suse.de>
Subject: [RFC PATCH v2] ceph: allow encrypting a directory while not having Ax caps
Date:   Thu,  3 Nov 2022 14:06:40 +0000
Message-Id: <20221103140640.30850-1-lhenriques@suse.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

If a client doesn't have Fx caps on a directory, it will get errors while
trying encrypt it:

ceph: handle_cap_grant: cap grant attempt to change fscrypt_auth on non-I_NEW inode (old len 0 new len 48)
fscrypt (ceph, inode 1099511627812): Error -105 getting encryption context

A simple way to reproduce this is to use two clients:

    client1 # mkdir /mnt/mydir

    client2 # ls /mnt/mydir

    client1 # fscrypt encrypt /mnt/mydir
    client1 # echo hello > /mnt/mydir/world

This happens because, in __ceph_setattr(), we only initialize
ci->fscrypt_auth if we have Ax and ceph_fill_inode() won't use the
fscrypt_auth received if the inode state isn't I_NEW.  Fix it by allowing
ceph_fill_inode() to also set ci->fscrypt_auth if the inode doesn't have
it set already.

Signed-off-by: Luís Henriques <lhenriques@suse.de>
---
 fs/ceph/inode.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index fb507d57cb26..c7831f801911 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -1016,7 +1016,8 @@ int ceph_fill_inode(struct inode *inode, struct page *locked_page,
 	__ceph_update_quota(ci, iinfo->max_bytes, iinfo->max_files);
 
 #ifdef CONFIG_FS_ENCRYPTION
-	if (iinfo->fscrypt_auth_len && (inode->i_state & I_NEW)) {
+	if (iinfo->fscrypt_auth_len &&
+	    ((inode->i_state & I_NEW) || (ci->fscrypt_auth_len == 0))) {
 		kfree(ci->fscrypt_auth);
 		ci->fscrypt_auth_len = iinfo->fscrypt_auth_len;
 		ci->fscrypt_auth = iinfo->fscrypt_auth;