A number of important fixes:

  - mbuf reused after an RX_COPY optimized operation can sometimes have
    a bogus cached address, resulting in TCP hangs. Add critical save points
    to the cached address. Thanks to Michael and the team at Verisign for
    finding this problem.
  - A couple more spots where the rxbuf->flags member should be cleared just
    to be sure no incorrect RX_COPY state is left around. Thanks to Adrian
    for tracking these down.
  - Remove the rearm_queues function from the driver, this was found to be
    responsible for some out-of-order packets by Verisign, and was always a
    bandaid, with the other fixes in this delta the bandaid can finally be
    removed.
  - In the other/link interrupt handler the entire state of the EICS register
    was being writen back into EICR (which clears causes and thus re-enables
    those interrupts), this was wrong, so now mask off the queue portion of
    the register value, so we only clear the other/link interrupt we intend.
    Marc from Verisign found this.
  - Make the SFP+ unsupported option tuneable now, by customer request.
  - Finally, just a couple of minor DEBUG string fixes.

I want to call out and thank all the participants in the 10G community/Intel
calls for helping track down these problems and make the driver better for
everyone!

MFC after:	3 days, these are critical fixes for 9.2!

1 files changed, 20 insertions, 24 deletions

diff --git a/sys/dev/ixgbe/ixgbe.c b/sys/dev/ixgbe/ixgbe.c
index 3a444e2..42fadd4 100644
--- a/sys/dev/ixgbe/ixgbe.c
+++ b/sys/dev/ixgbe/ixgbe.c
@@ -45,7 +45,7 @@ int             ixgbe_display_debug_stats = 0;
 /*********************************************************************
  *  Driver version
  *********************************************************************/
-char ixgbe_driver_version[] = "2.5.13";
+char ixgbe_driver_version[] = "2.5.15";
 
 /*********************************************************************
  *  PCI Device ID Table
@@ -297,6 +297,7 @@ TUNABLE_INT("hw.ixgbe.rxd", &ixgbe_rxd);
 ** doing so you are on your own :)
 */
 static int allow_unsupported_sfp = FALSE;
+TUNABLE_INT("hw.ixgbe.unsupported_sfp", &allow_unsupported_sfp);
 
 /*
 ** HW RSC control: 
@@ -1071,7 +1072,7 @@ ixgbe_init_locked(struct adapter *adapter)
 	u32		rxdctl, rxctrl;
 
 	mtx_assert(&adapter->core_mtx, MA_OWNED);
-	INIT_DEBUGOUT("ixgbe_init: begin");
+	INIT_DEBUGOUT("ixgbe_init_locked: begin");
 	hw->adapter_stopped = FALSE;
 	ixgbe_stop_adapter(hw);
         callout_stop(&adapter->timer);
@@ -1382,23 +1383,6 @@ ixgbe_disable_queue(struct adapter *adapter, u32 vector)
 	}
 }
 
-static inline void
-ixgbe_rearm_queues(struct adapter *adapter, u64 queues)
-{
-	u32 mask;
-
-	if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
-		mask = (IXGBE_EIMS_RTX_QUEUE & queues);
-		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS, mask);
-	} else {
-		mask = (queues & 0xFFFFFFFF);
-		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS_EX(0), mask);
-		mask = (queues >> 32);
-		IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS_EX(1), mask);
-	}
-}
-
-
 static void
 ixgbe_handle_que(void *context, int pending)
 {
@@ -1506,6 +1490,10 @@ ixgbe_msix_que(void *arg)
 	bool		more;
 	u32		newitr = 0;
 
+	/* Protect against spurious interrupts */
+	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
+		return;
+
 	ixgbe_disable_queue(adapter, que->msix);
 	++que->irqs;
 
@@ -1592,6 +1580,8 @@ ixgbe_msix_link(void *arg)
 
 	/* First get the cause */
 	reg_eicr = IXGBE_READ_REG(hw, IXGBE_EICS);
+	/* Be sure the queue bits are not cleared */
+	reg_eicr = ~IXGBE_EICR_RTX_QUEUE;
 	/* Clear interrupt with write */
 	IXGBE_WRITE_REG(hw, IXGBE_EICR, reg_eicr);
 
@@ -2067,7 +2057,6 @@ ixgbe_local_timer(void *arg)
                 goto watchdog;
 
 out:
-	ixgbe_rearm_queues(adapter, adapter->que_mask);
 	callout_reset(&adapter->timer, hz, ixgbe_local_timer, adapter);
 	return;
 
@@ -3201,7 +3190,7 @@ ixgbe_free_transmit_buffers(struct tx_ring *txr)
 	struct ixgbe_tx_buf *tx_buffer;
 	int             i;
 
-	INIT_DEBUGOUT("free_transmit_ring: begin");
+	INIT_DEBUGOUT("ixgbe_free_transmit_ring: begin");
 
 	if (txr->tx_buffers == NULL)
 		return;
@@ -4005,11 +3994,13 @@ ixgbe_setup_receive_ring(struct rx_ring *rxr)
 
 			addr = PNMB(slot + sj, &paddr);
 			netmap_load_map(rxr->ptag, rxbuf->pmap, addr);
-			/* Update descriptor */
+			/* Update descriptor and the cached value */
 			rxr->rx_base[j].read.pkt_addr = htole64(paddr);
+			rxbuf->addr = htole64(paddr);
 			continue;
 		}
 #endif /* DEV_NETMAP */
+		rxbuf->flags = 0; 
 		rxbuf->buf = m_getjcl(M_NOWAIT, MT_DATA,
 		    M_PKTHDR, adapter->rx_mbuf_sz);
 		if (rxbuf->buf == NULL) {
@@ -4026,8 +4017,9 @@ ixgbe_setup_receive_ring(struct rx_ring *rxr)
                         goto fail;
 		bus_dmamap_sync(rxr->ptag,
 		    rxbuf->pmap, BUS_DMASYNC_PREREAD);
-		/* Update descriptor */
+		/* Update the descriptor and the cached value */
 		rxr->rx_base[j].read.pkt_addr = htole64(seg[0].ds_addr);
+		rxbuf->addr = htole64(seg[0].ds_addr);
 	}
 
 
@@ -4244,6 +4236,8 @@ ixgbe_free_receive_structures(struct adapter *adapter)
 {
 	struct rx_ring *rxr = adapter->rx_rings;
 
+	INIT_DEBUGOUT("ixgbe_free_receive_structures: begin");
+
 	for (int i = 0; i < adapter->num_queues; i++, rxr++) {
 		struct lro_ctrl		*lro = &rxr->lro;
 		ixgbe_free_receive_buffers(rxr);
@@ -4268,7 +4262,7 @@ ixgbe_free_receive_buffers(struct rx_ring *rxr)
 	struct adapter		*adapter = rxr->adapter;
 	struct ixgbe_rx_buf	*rxbuf;
 
-	INIT_DEBUGOUT("free_receive_structures: begin");
+	INIT_DEBUGOUT("ixgbe_free_receive_buffers: begin");
 
 	/* Cleanup any existing buffers */
 	if (rxr->rx_buffers != NULL) {
@@ -4358,6 +4352,8 @@ ixgbe_rx_discard(struct rx_ring *rxr, int i)
 		m_free(rbuf->buf);
 		rbuf->buf = NULL;
 	}
+
+	rbuf->flags = 0;
  
 	return;
 }