port: remove tap link status workaround

Remove the workaround that assumed tap interfaces are always running.
This was needed when tap interfaces were moved to network namespaces,
which caused the kernel to signal them as down even though they could
still transmit and receive packets.

Since the DPDK tap drive now supports netns change, we no longer need
this workaround.

Signed-off-by: Robin Jarry <[email protected]>
Reviewed-by: Christophe Fontaine <[email protected]>

Author: rjarry

modules/infra/control/port.c

@@ -751,19 +751,6 @@ static void link_event_cb(evutil_socket_t, short /*what*/, void * /*priv*/) {
 
 			port = iface_info_port(iface);
 
-			// XXX: net_tap devices are signaled down by the kernel when they
-			// are moved to another netns although they still can receive and
-			// transmit packets. Ignore link status updates for this driver and
-			// always assume they are running.
-			if (strncmp(port->devargs, "net_tap", strlen("net_tap")) == 0) {
-				if (!(iface->state & GR_IFACE_S_RUNNING)) {
-					LOG(INFO, "%s: link status up", iface->name);
-					iface->state |= GR_IFACE_S_RUNNING;
-					gr_event_push(GR_EVENT_IFACE_STATUS_UP, iface);
-				}
-				continue;
-			}
-
 			if (rte_eth_link_get_nowait(qmap->port_id, &link) < 0) {
 				LOG(WARNING, "rte_eth_link_get_nowait: %s", strerror(rte_errno));
 				continue;