drbd: fix potential protocol error
authorLars Ellenberg <lars.ellenberg@linbit.com>
Thu, 1 Apr 2010 14:57:19 +0000 (16:57 +0200)
committerPhilipp Reisner <philipp.reisner@linbit.com>
Mon, 17 May 2010 23:07:38 +0000 (01:07 +0200)
Don't forget to drain the digest in case we cannot satisfy a
checksum based resync or online-verify request.

It would additionally cause a protocoll error,
dropping the connection.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
drivers/block/drbd/drbd_receiver.c

index 93106fb..c7285e1 100644 (file)
@@ -1331,6 +1331,9 @@ static int drbd_drain_block(struct drbd_conf *mdev, int data_size)
        int rr, rv = 1;
        void *data;
 
+       if (!data_size)
+               return TRUE;
+
        page = drbd_pp_alloc(mdev, 1);
 
        data = kmap(page);
@@ -1946,7 +1949,7 @@ static int receive_DataRequest(struct drbd_conf *mdev, struct p_header *h)
                            "no local data.\n");
                drbd_send_ack_rp(mdev, h->command == P_DATA_REQUEST ? P_NEG_DREPLY :
                                 P_NEG_RS_DREPLY , p);
-               return TRUE;
+               return drbd_drain_block(mdev, h->length - brps);
        }
 
        /* GFP_NOIO, because we must not cause arbitrary write-out: in a DRBD