Reapply r223198 which was reverted in the previous vendor import. Some

portions were already reapplied in r233708:
- Use a dedicated task to handle deferred transmits from the if_transmit
  method instead of reusing the existing per-queue interrupt task.
  Reusing the per-queue interrupt task could result in both an interrupt
  thread and the taskqueue thread trying to handle received packets on a
  single queue resulting in out-of-order packet processing.
- Call ether_ifdetach() earlier in igb_detach().
- Drain tasks and free taskqueues during igb_detach().

MFC after:	1 week

2 files changed, 42 insertions, 4 deletions

diff --git a/sys/dev/e1000/if_igb.c b/sys/dev/e1000/if_igb.c
index 5a5ec81..b15c38e 100644
--- a/sys/dev/e1000/if_igb.c
+++ b/sys/dev/e1000/if_igb.c
@@ -176,6 +176,7 @@ static int	igb_mq_start(struct ifnet *, struct mbuf *);
 static int	igb_mq_start_locked(struct ifnet *,
 		    struct tx_ring *, struct mbuf *);
 static void	igb_qflush(struct ifnet *);
+static void	igb_deferred_mq_start(void *, int);
 #else
 static void	igb_start(struct ifnet *);
 static void	igb_start_locked(struct tx_ring *, struct ifnet *ifp);
@@ -715,6 +716,8 @@ igb_detach(device_t dev)
 		return (EBUSY);
 	}
 
+	ether_ifdetach(adapter->ifp);
+
 	if (adapter->led_dev != NULL)
 		led_destroy(adapter->led_dev);
 
@@ -746,8 +749,6 @@ igb_detach(device_t dev)
 	if (adapter->vlan_detach != NULL)
 		EVENTHANDLER_DEREGISTER(vlan_unconfig, adapter->vlan_detach);
 
-	ether_ifdetach(adapter->ifp);
-
 	callout_drain(&adapter->timer);
 
 #ifdef DEV_NETMAP
@@ -943,7 +944,7 @@ igb_mq_start(struct ifnet *ifp, struct mbuf *m)
 		IGB_TX_UNLOCK(txr);
 	} else {
 		err = drbr_enqueue(ifp, txr->br, m);
-		taskqueue_enqueue(que->tq, &que->que_task);
+		taskqueue_enqueue(que->tq, &txr->txq_task);
 	}
 
 	return (err);
@@ -1003,6 +1004,22 @@ igb_mq_start_locked(struct ifnet *ifp, struct tx_ring *txr, struct mbuf *m)
 }
 
 /*
+ * Called from a taskqueue to drain queued transmit packets.
+ */
+static void
+igb_deferred_mq_start(void *arg, int pending)
+{
+	struct tx_ring *txr = arg;
+	struct adapter *adapter = txr->adapter;
+	struct ifnet *ifp = adapter->ifp;
+
+	IGB_TX_LOCK(txr);
+	if (!drbr_empty(ifp, txr->br))
+		igb_mq_start_locked(ifp, txr, NULL);
+	IGB_TX_UNLOCK(txr);
+}
+
+/*
 ** Flush all ring buffers
 */
 static void
@@ -2382,6 +2399,7 @@ igb_allocate_legacy(struct adapter *adapter)
 {
 	device_t		dev = adapter->dev;
 	struct igb_queue	*que = adapter->queues;
+	struct tx_ring		*txr = adapter->tx_rings;
 	int			error, rid = 0;
 
 	/* Turn off all interrupts */
@@ -2400,6 +2418,10 @@ igb_allocate_legacy(struct adapter *adapter)
 		return (ENXIO);
 	}
 
+#if __FreeBSD_version >= 800000
+	TASK_INIT(&txr->txq_task, 0, igb_deferred_mq_start, txr);
+#endif
+
 	/*
 	 * Try allocating a fast interrupt and the associated deferred
 	 * processing contexts.
@@ -2473,9 +2495,13 @@ igb_allocate_msix(struct adapter *adapter)
 		*/
 		if (adapter->num_queues > 1)
 			bus_bind_intr(dev, que->res, i);
+#if __FreeBSD_version >= 800000
+		TASK_INIT(&que->txr->txq_task, 0, igb_deferred_mq_start,
+		    que->txr);
+#endif
 		/* Make tasklet for deferred handling */
 		TASK_INIT(&que->que_task, 0, igb_handle_que, que);
-		que->tq = taskqueue_create_fast("igb_que", M_NOWAIT,
+		que->tq = taskqueue_create("igb_que", M_NOWAIT,
 		    taskqueue_thread_enqueue, &que->tq);
 		taskqueue_start_threads(&que->tq, 1, PI_NET, "%s que",
 		    device_get_nameunit(adapter->dev));
@@ -2682,13 +2708,24 @@ igb_free_pci_resources(struct adapter *adapter)
 	else
 		(adapter->msix != 0) ? (rid = 1):(rid = 0);
 
+	que = adapter->queues;
 	if (adapter->tag != NULL) {
+		taskqueue_drain(que->tq, &adapter->link_task);
 		bus_teardown_intr(dev, adapter->res, adapter->tag);
 		adapter->tag = NULL;
 	}
 	if (adapter->res != NULL)
 		bus_release_resource(dev, SYS_RES_IRQ, rid, adapter->res);
 
+	for (int i = 0; i < adapter->num_queues; i++, que++) {
+		if (que->tq != NULL) {
+#if __FreeBSD_version >= 800000
+			taskqueue_drain(que->tq, &que->txr->txq_task);
+#endif
+			taskqueue_drain(que->tq, &que->que_task);
+			taskqueue_free(que->tq);
+		}
+	}
 mem:
 	if (adapter->msix)
 		pci_release_msi(dev);
diff --git a/sys/dev/e1000/if_igb.h b/sys/dev/e1000/if_igb.h
index 9989206..8ff1b44 100644
--- a/sys/dev/e1000/if_igb.h
+++ b/sys/dev/e1000/if_igb.h
@@ -301,6 +301,7 @@ struct tx_ring {
 	struct buf_ring		*br;
 #endif
 	bus_dma_tag_t		txtag;
+	struct task		txq_task;
 
 	u32			bytes;
 	u32			packets;