ip6: set local flag on outbound echo reply

Type: fix

When VPP generates an ICMP echo reply in response to an incoming
echo request to a local address, set VNET_BUFFER_F_LOCALLY_ORIGINATED
on the buffer. It will prevent ip6-rewrite from decrementing the hop
limit.

Outbound IPv4 echo replies also get this flag set.

Change-Id: Iaa229294eb158edb58cf1bf1b7a90da281321406
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
diff --git a/src/vnet/ip/icmp6.c b/src/vnet/ip/icmp6.c
index b6ed3ea..b0fdadb 100644
--- a/src/vnet/ip/icmp6.c
+++ b/src/vnet/ip/icmp6.c
@@ -351,6 +351,9 @@
 				vnet_buffer (p1)->sw_if_index[VLIB_RX]);
 	  vnet_buffer (p1)->sw_if_index[VLIB_TX] = fib_index1;
 
+	  p0->flags |= VNET_BUFFER_F_LOCALLY_ORIGINATED;
+	  p1->flags |= VNET_BUFFER_F_LOCALLY_ORIGINATED;
+
 	  /* verify speculative enqueues, maybe switch current next frame */
 	  /* if next0==next1==next_index then nothing special needs to be done */
 	  vlib_validate_buffer_enqueue_x2 (vm, node, next_index,
@@ -404,6 +407,7 @@
 				vnet_buffer (p0)->sw_if_index[VLIB_RX]);
 	  vnet_buffer (p0)->sw_if_index[VLIB_TX] = fib_index0;
 
+	  p0->flags |= VNET_BUFFER_F_LOCALLY_ORIGINATED;
 	  /* Verify speculative enqueue, maybe switch current next frame */
 	  vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
 					   to_next, n_left_to_next,