net: rxrpc: change call to sock_recv_ts_and_drops() on rxrpc recvmsg to sock_recv_tim...
authorEyal Birger <eyal.birger@gmail.com>
Sun, 1 Mar 2015 12:58:27 +0000 (14:58 +0200)
committerDavid S. Miller <davem@davemloft.net>
Mon, 2 Mar 2015 05:19:29 +0000 (00:19 -0500)
Commit 3b885787ea4112 ("net: Generalize socket rx gap / receive queue overflow cmsg")
allowed receiving packet dropcount information as a socket level option.
RXRPC sockets recvmsg function was changed to support this by calling
sock_recv_ts_and_drops() instead of sock_recv_timestamp().

However, protocol families wishing to receive dropcount should call
sock_queue_rcv_skb() or set the dropcount specifically (as done
in packet_rcv()). This was not done for rxrpc and thus this feature
never worked on these sockets.

Formalizing this by not calling sock_recv_ts_and_drops() in rxrpc as
part of an effort to move skb->dropcount into skb->cb[]

Signed-off-by: Eyal Birger <eyal.birger@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/rxrpc/ar-recvmsg.c

index 4575485ad1b4d02796717879e09ad9078e232617..d58ba702bd2c3200a3670832f02ca47957004cb0 100644 (file)
@@ -150,7 +150,7 @@ int rxrpc_recvmsg(struct kiocb *iocb, struct socket *sock,
                                       &call->conn->trans->peer->srx, len);
                                msg->msg_namelen = len;
                        }
-                       sock_recv_ts_and_drops(msg, &rx->sk, skb);
+                       sock_recv_timestamp(msg, &rx->sk, skb);
                }
 
                /* receive the message */