[PATCH 3.2 35/77] RDS: fix race condition when sending a message on unbound socket

[Ben Hutchings]

3.2.75-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Quentin Casasnovas <quentin.casasnovas@xxxxxxxxxx>

commit 8c7188b23474cca017b3ef354c4a58456f68303a upstream.

Sasha's found a NULL pointer dereference in the RDS connection code when
sending a message to an apparently unbound socket.  The problem is caused
by the code checking if the socket is bound in rds_sendmsg(), which checks
the rs_bound_addr field without taking a lock on the socket.  This opens a
race where rs_bound_addr is temporarily set but where the transport is not
in rds_bind(), leading to a NULL pointer dereference when trying to
dereference 'trans' in __rds_conn_create().

Vegard wrote a reproducer for this issue, so kindly ask him to share if
you're interested.

I cannot reproduce the NULL pointer dereference using Vegard's reproducer
with this patch, whereas I could without.

Complete earlier incomplete fix to CVE-2015-6937:

  74e98eb08588 ("RDS: verify the underlying transport exists before creating a connection")

Cc: David S. Miller <davem@xxxxxxxxxxxxx>

Reviewed-by: Vegard Nossum <vegard.nossum@xxxxxxxxxx>
Reviewed-by: Sasha Levin <sasha.levin@xxxxxxxxxx>
Acked-by: Santosh Shilimkar <santosh.shilimkar@xxxxxxxxxx>
Signed-off-by: Quentin Casasnovas <quentin.casasnovas@xxxxxxxxxx>
Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
 net/rds/connection.c | 6 ------
 net/rds/send.c       | 4 +++-
 2 files changed, 3 insertions(+), 7 deletions(-)

--- a/net/rds/connection.c
+++ b/net/rds/connection.c
@@ -178,12 +178,6 @@ static struct rds_connection *__rds_conn
 		}
 	}
 
-	if (trans == NULL) {
-		kmem_cache_free(rds_conn_slab, conn);
-		conn = ERR_PTR(-ENODEV);
-		goto out;
-	}
-
 	conn->c_trans = trans;
 
 	ret = trans->conn_alloc(conn, gfp);
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -955,11 +955,13 @@ int rds_sendmsg(struct kiocb *iocb, stru
 		release_sock(sk);
 	}
 
-	/* racing with another thread binding seems ok here */
+	lock_sock(sk);
 	if (daddr == 0 || rs->rs_bound_addr == 0) {
+		release_sock(sk);
 		ret = -ENOTCONN; /* XXX not a great errno */
 		goto out;
 	}
+	release_sock(sk);
 
 	/* size of rm including all sgs */
 	ret = rds_rm_size(msg, payload_len);

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/