pfcp_endpoint: fix final PFCP retrans resp_cb

After the final retransmission of a sent request, still keep the message
in the queue for its expiry period, so that a later response is matched
to the request.

The osmo_pfcp_msg.resp_cb() depends on the sent message to remain in the
queue until it times out. That was not the case in an earlier stage of
libosmo-pfcp development.

I noticed this during ttcn3 testing, where osmo-hnbgw continuously
resends PFCP Association Setup Requests, and fails to associate if ttcn3
happens to respond to the final retransmission of a request.

Related: SYS#5599
Change-Id: Iaca396891921f7057015ce6e1e4528b955757809

1 files changed, 8 insertions, 7 deletions

diff --git a/src/libosmo-pfcp/pfcp_endpoint.c b/src/libosmo-pfcp/pfcp_endpoint.c
index 0ee3a9a..68d0a21 100644
--- a/src/libosmo-pfcp/pfcp_endpoint.c
+++ b/src/libosmo-pfcp/pfcp_endpoint.c
@@ -181,18 +181,19 @@ static bool pfcp_queue_retrans(struct osmo_pfcp_queue_entry *qe)
 	struct osmo_pfcp_msg *m = qe->m;
 	int rc;
 
+	/* if no more attempts remaining, drop from queue */
+	if (!qe->n1_remaining)
+		return false;
+
 	/* re-transmit */
-	if (qe->n1_remaining)
-		qe->n1_remaining--;
-	OSMO_LOG_PFCP_MSG(m, LOGL_INFO, "re-sending (%u attempts remaining)\n", qe->n1_remaining);
+	qe->n1_remaining--;
+	OSMO_LOG_PFCP_MSG(m, LOGL_INFO, "re-sending (%u attempts remaining after this)\n", qe->n1_remaining);
 
 	rc = osmo_pfcp_endpoint_tx_data_no_logging(endpoint, m);
-	/* If encoding failed, it cannot ever succeed. Drop the queue entry. */
+	/* If encoding failed, it cannot ever succeed. Drop the queue entry. (Error logging already taken care of in
+	 * osmo_pfcp_endpoint_tx_data_no_logging().) */
 	if (rc)
 		return false;
-	/* if no more attempts remaining, drop from queue */
-	if (!qe->n1_remaining)
-		return false;
 	/* re-schedule timer, keep in queue */
 	osmo_timer_schedule(&qe->t1, t1_ms/1000, (t1_ms % 1000) * 1000);
 	return true;