Fix a few issues with transmit handling in em(4) and igb(4):

- Do not define the foo_start() methods or set if_start in the ifnet if
  multiq transmit is enabled.  Also, set if_transmit and if_qflush before
  ether_ifattach rather than after when multiq transmit is enabled.  This
  helps to ensure that the drivers never try to mix different transmit
  methods.
- Properly restart transmit during resume.  igb(4) was not restarting it
  at all, and em(4) was restarting even if the link was down and was
  calling the wrong method if multiq transmit was enabled.
- Remove all the 'more' handling for transmit completions.  Transmit
  completion processing does not have a processing limit, so it always
  runs to completion and never has more work to do when it returns.
  Instead, the previous code was returning 'true' anytime there were
  packets in the queue that weren't still in the process of being
  transmitted.  The effect was that the driver would continuously
  reschedule a task to process TX completions in effect running at 100%
  CPU polling the hardware until it finished transmitting all of the
  packets in the ring.  Now it will just wait for the next TX completion
  interrupt.
- Restart packet transmission when the link becomes active.
- Fix the MSI-X queue interrupt handlers to restart packet transmission if
  there are pending packets in the relevant software queue (IFQ or buf_ring)
  after processing TX completions.  This is the root cause for the OACTIVE
  hangs as if the MSI-X queue handler drained all the pending packets from
  the TX ring, nothing would ever restart it.  As such, remove some
  previously-added workarounds to reschedule a task to poll the TX ring
  anytime OACTIVE was set.

Tested by:	sbruno
Reviewed by:	jfv
MFC after:	1 week

1 files changed, 63 insertions, 33 deletions

diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
index 8ab511d..9c31dad 100644
--- a/sys/dev/e1000/if_em.c
+++ b/sys/dev/e1000/if_em.c
@@ -193,13 +193,14 @@ static int	em_detach(device_t);
 static int	em_shutdown(device_t);
 static int	em_suspend(device_t);
 static int	em_resume(device_t);
-static void	em_start(struct ifnet *);
-static void	em_start_locked(struct ifnet *, struct tx_ring *);
 #ifdef EM_MULTIQUEUE
 static int	em_mq_start(struct ifnet *, struct mbuf *);
 static int	em_mq_start_locked(struct ifnet *,
 		    struct tx_ring *, struct mbuf *);
 static void	em_qflush(struct ifnet *);
+#else
+static void	em_start(struct ifnet *);
+static void	em_start_locked(struct ifnet *, struct tx_ring *);
 #endif
 static int	em_ioctl(struct ifnet *, u_long, caddr_t);
 static void	em_init(void *);
@@ -234,7 +235,7 @@ static void	em_enable_intr(struct adapter *);
 static void	em_disable_intr(struct adapter *);
 static void	em_update_stats_counters(struct adapter *);
 static void	em_add_hw_stats(struct adapter *adapter);
-static bool	em_txeof(struct tx_ring *);
+static void	em_txeof(struct tx_ring *);
 static bool	em_rxeof(struct rx_ring *, int, int *);
 #ifndef __NO_STRICT_ALIGNMENT
 static int	em_fixup_rx(struct rx_ring *);
@@ -847,6 +848,7 @@ static int
 em_resume(device_t dev)
 {
 	struct adapter *adapter = device_get_softc(dev);
+	struct tx_ring	*txr = adapter->tx_rings;
 	struct ifnet *ifp = adapter->ifp;
 
 	EM_CORE_LOCK(adapter);
@@ -854,8 +856,22 @@ em_resume(device_t dev)
 		e1000_resume_workarounds_pchlan(&adapter->hw);
 	em_init_locked(adapter);
 	em_init_manageability(adapter);
+
+	if ((ifp->if_flags & IFF_UP) &&
+	    (ifp->if_drv_flags & IFF_DRV_RUNNING) && adapter->link_active) {
+		for (int i = 0; i < adapter->num_queues; i++, txr++) {
+			EM_TX_LOCK(txr);
+#ifdef EM_MULTIQUEUE
+			if (!drbr_empty(ifp, txr->br))
+				em_mq_start_locked(ifp, txr, NULL);
+#else
+			if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+				em_start_locked(ifp, txr);
+#endif
+			EM_TX_UNLOCK(txr);
+		}
+	}
 	EM_CORE_UNLOCK(adapter);
-	em_start(ifp);
 
 	return bus_generic_resume(dev);
 }
@@ -959,7 +975,7 @@ em_qflush(struct ifnet *ifp)
 	}
 	if_qflush(ifp);
 }
-#endif /* EM_MULTIQUEUE */
+#else  /* !EM_MULTIQUEUE */
 
 static void
 em_start_locked(struct ifnet *ifp, struct tx_ring *txr)
@@ -1020,14 +1036,9 @@ em_start(struct ifnet *ifp)
 		em_start_locked(ifp, txr);
 		EM_TX_UNLOCK(txr);
 	}
-	/*
-	** If we went inactive schedule
-	** a task to clean up.
-	*/
-	if (ifp->if_drv_flags & IFF_DRV_OACTIVE)
-		taskqueue_enqueue(txr->tq, &txr->tx_task);
 	return;
 }
+#endif /* EM_MULTIQUEUE */
 
 /*********************************************************************
  *  Ioctl entry point
@@ -1424,7 +1435,8 @@ em_poll(struct ifnet *ifp, enum poll_cmd cmd, int count)
 	if (!drbr_empty(ifp, txr->br))
 		em_mq_start_locked(ifp, txr, NULL);
 #else
-	em_start_locked(ifp, txr);
+	if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+		em_start_locked(ifp, txr);
 #endif
 	EM_TX_UNLOCK(txr);
 
@@ -1497,10 +1509,11 @@ em_handle_que(void *context, int pending)
 		if (!drbr_empty(ifp, txr->br))
 			em_mq_start_locked(ifp, txr, NULL);
 #else
-		em_start_locked(ifp, txr);
+		if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+			em_start_locked(ifp, txr);
 #endif
 		EM_TX_UNLOCK(txr);
-		if (more || (ifp->if_drv_flags & IFF_DRV_OACTIVE)) {
+		if (more) {
 			taskqueue_enqueue(adapter->tq, &adapter->que_task);
 			return;
 		}
@@ -1521,17 +1534,21 @@ em_msix_tx(void *arg)
 {
 	struct tx_ring *txr = arg;
 	struct adapter *adapter = txr->adapter;
-	bool		more;
+	struct ifnet	*ifp = adapter->ifp;
 
 	++txr->tx_irq;
 	EM_TX_LOCK(txr);
-	more = em_txeof(txr);
+	em_txeof(txr);
+#ifdef EM_MULTIQUEUE
+	if (!drbr_empty(ifp, txr->br))
+		em_mq_start_locked(ifp, txr, NULL);
+#else
+	if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+		em_start_locked(ifp, txr);
+#endif
+	/* Reenable this interrupt */
+	E1000_WRITE_REG(&adapter->hw, E1000_IMS, txr->ims);
 	EM_TX_UNLOCK(txr);
-	if (more)
-		taskqueue_enqueue(txr->tq, &txr->tx_task);
-	else
-		/* Reenable this interrupt */
-		E1000_WRITE_REG(&adapter->hw, E1000_IMS, txr->ims);
 	return;
 }
 
@@ -1609,7 +1626,8 @@ em_handle_tx(void *context, int pending)
 	if (!drbr_empty(ifp, txr->br))
 		em_mq_start_locked(ifp, txr, NULL);
 #else
-	em_start_locked(ifp, txr);
+	if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+		em_start_locked(ifp, txr);
 #endif
 	E1000_WRITE_REG(&adapter->hw, E1000_IMS, txr->ims);
 	EM_TX_UNLOCK(txr);
@@ -1619,6 +1637,7 @@ static void
 em_handle_link(void *context, int pending)
 {
 	struct adapter	*adapter = context;
+	struct tx_ring	*txr = adapter->tx_rings;
 	struct ifnet *ifp = adapter->ifp;
 
 	if (!(ifp->if_drv_flags & IFF_DRV_RUNNING))
@@ -1630,6 +1649,19 @@ em_handle_link(void *context, int pending)
 	callout_reset(&adapter->timer, hz, em_local_timer, adapter);
 	E1000_WRITE_REG(&adapter->hw, E1000_IMS,
 	    EM_MSIX_LINK | E1000_IMS_LSC);
+	if (adapter->link_active) {
+		for (int i = 0; i < adapter->num_queues; i++, txr++) {
+			EM_TX_LOCK(txr);
+#ifdef EM_MULTIQUEUE
+			if (!drbr_empty(ifp, txr->br))
+				em_mq_start_locked(ifp, txr, NULL);
+#else
+			if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+				em_start_locked(ifp, txr);
+#endif
+			EM_TX_UNLOCK(txr);
+		}
+	}
 	EM_CORE_UNLOCK(adapter);
 }
 
@@ -2902,20 +2934,21 @@ em_setup_interface(device_t dev, struct adapter *adapter)
 	ifp->if_softc = adapter;
 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
 	ifp->if_ioctl = em_ioctl;
+#ifdef EM_MULTIQUEUE
+	/* Multiqueue stack interface */
+	ifp->if_transmit = em_mq_start;
+	ifp->if_qflush = em_qflush;
+#else
 	ifp->if_start = em_start;
 	IFQ_SET_MAXLEN(&ifp->if_snd, adapter->num_tx_desc - 1);
 	ifp->if_snd.ifq_drv_maxlen = adapter->num_tx_desc - 1;
 	IFQ_SET_READY(&ifp->if_snd);
+#endif	
 
 	ether_ifattach(ifp, adapter->hw.mac.addr);
 
 	ifp->if_capabilities = ifp->if_capenable = 0;
 
-#ifdef EM_MULTIQUEUE
-	/* Multiqueue stack interface */
-	ifp->if_transmit = em_mq_start;
-	ifp->if_qflush = em_qflush;
-#endif	
 
 	ifp->if_capabilities |= IFCAP_HWCSUM | IFCAP_VLAN_HWCSUM;
 	ifp->if_capabilities |= IFCAP_TSO4;
@@ -3742,7 +3775,7 @@ em_tso_setup(struct tx_ring *txr, struct mbuf *mp, int ip_off,
  *  tx_buffer is put back on the free queue.
  *
  **********************************************************************/
-static bool
+static void
 em_txeof(struct tx_ring *txr)
 {
 	struct adapter	*adapter = txr->adapter;
@@ -3762,14 +3795,14 @@ em_txeof(struct tx_ring *txr)
 		selwakeuppri(&na->tx_si, PI_NET);
 		EM_CORE_UNLOCK(adapter);
 		EM_TX_LOCK(txr);
-		return (FALSE);
+		return;
 	}
 #endif /* DEV_NETMAP */
 
 	/* No work, make sure watchdog is off */
         if (txr->tx_avail == adapter->num_tx_desc) {
 		txr->queue_status = EM_QUEUE_IDLE;
-                return (FALSE);
+                return;
 	}
 
 	processed = 0;
@@ -3858,10 +3891,7 @@ em_txeof(struct tx_ring *txr)
 	/* Disable watchdog if all clean */
 	if (txr->tx_avail == adapter->num_tx_desc) {
 		txr->queue_status = EM_QUEUE_IDLE;
-		return (FALSE);
 	} 
-
-	return (TRUE);
 }