ceph: ensure RNG is seeded before using
authorJason A. Donenfeld <Jason@zx2c4.com>
Thu, 8 Jun 2017 02:44:54 +0000 (22:44 -0400)
committerTheodore Ts'o <tytso@mit.edu>
Tue, 20 Jun 2017 02:06:28 +0000 (22:06 -0400)
Ceph uses the RNG for various nonce generations, and it shouldn't accept
using bad randomness. So, we wait for the RNG to be properly seeded. We
do this by calling wait_for_random_bytes() in a function that is
certainly called in process context, early on, so that all subsequent
calls to get_random_bytes are necessarily acceptable.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: "Yan, Zheng" <zyan@redhat.com>
Cc: Sage Weil <sage@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
net/ceph/ceph_common.c

index 4fd0283..26ab586 100644 (file)
@@ -611,7 +611,11 @@ struct ceph_client *ceph_create_client(struct ceph_options *opt, void *private)
 {
        struct ceph_client *client;
        struct ceph_entity_addr *myaddr = NULL;
-       int err = -ENOMEM;
+       int err;
+
+       err = wait_for_random_bytes();
+       if (err < 0)
+               return ERR_PTR(err);
 
        client = kzalloc(sizeof(*client), GFP_KERNEL);
        if (client == NULL)