From: Andy Gospodarek <gospo@redhat.com>
Subject: [RHEL5.1 PATCH] ixgb: update to 1.0.109 to add pci error recovery
Date: Mon, 7 May 2007 17:36:52 -0400
Bugzilla: 211380
Message-Id: <20070507213651.GD2810@gospo.rdu.redhat.com>
Changelog: [net] ixgb: update to 1.0.109 to add pci error recovery



IBM specifically asked for PCI error handling to be included in the ixgb
driver for RHEL5.1.  Since we don't ship PCIE AER support for RHEL5
(though I do have some patches ready if/when the request is approved)
this update will go on all platforms but the error recovery stuff will
only be effective on ppc64 kernels.

This patch includes several upstream commits up to and including the
following upstream commit.  This updates us to 1.0.109-k4 -- the first
version to include error recovery support in the driver.

commit 01748fbb413d6f3b36c330544969d1d7254ee509
Author: Linas Vepstas <linas@austin.ibm.com>

    ixgb: Add PCI Error recovery callbacks


---

 ixgb.h         |    5 --
 ixgb_ethtool.c |    6 --
 ixgb_hw.c      |   11 ++++
 ixgb_ids.h     |    1 
 ixgb_main.c    |  141 ++++++++++++++++++++++++++++++++++++++++++++++++++++-----
 5 files changed, 143 insertions(+), 21 deletions(-)

--- linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_ethtool.c.gospo	2007-03-07 16:44:36.631512000 -0500
+++ linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_ethtool.c	2007-03-07 16:58:49.276613000 -0500
@@ -654,11 +654,7 @@ ixgb_phys_id(struct net_device *netdev, 
 
 	mod_timer(&adapter->blink_timer, jiffies);
 
-	if (data)
-		schedule_timeout_interruptible(data * HZ);
-	else
-		schedule_timeout_interruptible(MAX_SCHEDULE_TIMEOUT);
-
+	msleep_interruptible(data * 1000);
 	del_timer_sync(&adapter->blink_timer);
 	ixgb_led_off(&adapter->hw);
 	clear_bit(IXGB_LED_ON, &adapter->led_status);
--- linux-2.6.18.x86_64/drivers/net/ixgb/ixgb.h.gospo	2007-03-07 16:44:36.666481000 -0500
+++ linux-2.6.18.x86_64/drivers/net/ixgb/ixgb.h	2007-03-07 16:58:53.405480000 -0500
@@ -110,9 +110,6 @@ struct ixgb_adapter;
 #define IXGB_RXBUFFER_8192  8192
 #define IXGB_RXBUFFER_16384 16384
 
-/* How many Tx Descriptors do we need to call netif_wake_queue? */
-#define IXGB_TX_QUEUE_WAKE 16
-
 /* How many Rx Buffers do we bundle into one write to the hardware ? */
 #define IXGB_RX_BUFFER_WRITE	4	/* Must be power of 2 */
 
@@ -173,7 +170,7 @@ struct ixgb_adapter {
 	unsigned long led_status;
 
 	/* TX */
-	struct ixgb_desc_ring tx_ring;
+	struct ixgb_desc_ring tx_ring ____cacheline_aligned_in_smp;
 	unsigned long timeo_start;
 	uint32_t tx_cmd_type;
 	uint64_t hw_csum_tx_good;
--- linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_hw.c.gospo	2007-03-07 16:44:36.673476000 -0500
+++ linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_hw.c	2007-03-07 16:58:17.533393000 -0500
@@ -236,6 +236,17 @@ ixgb_identify_phy(struct ixgb_hw *hw)
 		DEBUGOUT("Identified G6104 optics\n");
 		phy_type = ixgb_phy_type_g6104;
 		break;
+	case IXGB_DEVICE_ID_82597EX_CX4:
+		DEBUGOUT("Identified CX4\n");
+		xpak_vendor = ixgb_identify_xpak_vendor(hw);
+		if (xpak_vendor == ixgb_xpak_vendor_intel) {
+			DEBUGOUT("Identified TXN17201 optics\n");
+			phy_type = ixgb_phy_type_txn17201;
+		} else {
+			DEBUGOUT("Identified G6005 optics\n");
+			phy_type = ixgb_phy_type_g6005;
+		}
+		break;
 	default:
 		DEBUGOUT("Unknown physical layer module\n");
 		phy_type = ixgb_phy_type_unknown;
--- linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_ids.h.gospo	2007-03-07 16:44:36.685471000 -0500
+++ linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_ids.h	2007-03-07 16:58:17.552376000 -0500
@@ -45,6 +45,7 @@
 
 #define IXGB_DEVICE_ID_82597EX_CX4   0x109E
 #define IXGB_SUBDEVICE_ID_A00C  0xA00C
+#define IXGB_SUBDEVICE_ID_A01C  0xA01C
 
 #endif /* #ifndef _IXGB_IDS_H_ */
 /* End of File */
--- linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_main.c.gospo	2007-03-07 16:44:36.695452000 -0500
+++ linux-2.6.18.x86_64/drivers/net/ixgb/ixgb_main.c	2007-03-07 16:58:57.802082000 -0500
@@ -36,7 +36,7 @@ static char ixgb_driver_string[] = "Inte
 #else
 #define DRIVERNAPI "-NAPI"
 #endif
-#define DRV_VERSION		"1.0.109-k2"DRIVERNAPI
+#define DRV_VERSION		"1.0.109-k4"DRIVERNAPI
 char ixgb_driver_version[] = DRV_VERSION;
 static char ixgb_copyright[] = "Copyright (c) 1999-2006 Intel Corporation.";
 
@@ -118,15 +118,26 @@ static void ixgb_restore_vlan(struct ixg
 static void ixgb_netpoll(struct net_device *dev);
 #endif
 
-/* Exported from other modules */
+static pci_ers_result_t ixgb_io_error_detected (struct pci_dev *pdev,
+	                     enum pci_channel_state state);
+static pci_ers_result_t ixgb_io_slot_reset (struct pci_dev *pdev);
+static void ixgb_io_resume (struct pci_dev *pdev);
 
+/* Exported from other modules */
 extern void ixgb_check_options(struct ixgb_adapter *adapter);
 
+static struct pci_error_handlers ixgb_err_handler = {
+	.error_detected = ixgb_io_error_detected,
+	.slot_reset = ixgb_io_slot_reset,
+	.resume = ixgb_io_resume,
+};
+
 static struct pci_driver ixgb_driver = {
 	.name     = ixgb_driver_name,
 	.id_table = ixgb_pci_tbl,
 	.probe    = ixgb_probe,
 	.remove   = __devexit_p(ixgb_remove),
+	.err_handler = &ixgb_err_handler
 };
 
 MODULE_AUTHOR("Intel Corporation, <linux.nics@intel.com>");
@@ -140,12 +151,12 @@ module_param(debug, int, 0);
 MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)");
 
 /* some defines for controlling descriptor fetches in h/w */
-#define RXDCTL_WTHRESH_DEFAULT 16	/* chip writes back at this many or RXT0 */
-#define RXDCTL_PTHRESH_DEFAULT 0		/* chip considers prefech below
-						 * this */
-#define RXDCTL_HTHRESH_DEFAULT 0		/* chip will only prefetch if tail
-						 * is pushed this many descriptors
-						 * from head */
+#define RXDCTL_WTHRESH_DEFAULT 15  /* chip writes back at this many or RXT0 */
+#define RXDCTL_PTHRESH_DEFAULT 0   /* chip considers prefech below
+                                    * this */
+#define RXDCTL_HTHRESH_DEFAULT 0   /* chip will only prefetch if tail
+                                    * is pushed this many descriptors
+                                    * from head */
 
 /**
  * ixgb_init_module - Driver Registration Routine
@@ -1174,6 +1185,7 @@ ixgb_tso(struct ixgb_adapter *adapter, s
 	int err;
 
 	if (likely(skb_is_gso(skb))) {
+		struct ixgb_buffer *buffer_info;
 		if (skb_header_cloned(skb)) {
 			err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
 			if (err)
@@ -1196,6 +1208,8 @@ ixgb_tso(struct ixgb_adapter *adapter, s
 
 		i = adapter->tx_ring.next_to_use;
 		context_desc = IXGB_CONTEXT_DESC(adapter->tx_ring, i);
+		buffer_info = &adapter->tx_ring.buffer_info[i];
+		WARN_ON(buffer_info->dma != 0);
 
 		context_desc->ipcss = ipcss;
 		context_desc->ipcso = ipcso;
@@ -1233,11 +1247,14 @@ ixgb_tx_csum(struct ixgb_adapter *adapte
 	uint8_t css, cso;
 
 	if(likely(skb->ip_summed == CHECKSUM_HW)) {
+		struct ixgb_buffer *buffer_info;
 		css = skb->h.raw - skb->data;
 		cso = (skb->h.raw + skb->csum) - skb->data;
 
 		i = adapter->tx_ring.next_to_use;
 		context_desc = IXGB_CONTEXT_DESC(adapter->tx_ring, i);
+		buffer_info = &adapter->tx_ring.buffer_info[i];
+		WARN_ON(buffer_info->dma != 0);
 
 		context_desc->tucss = css;
 		context_desc->tucso = cso;
@@ -1283,6 +1300,7 @@ ixgb_tx_map(struct ixgb_adapter *adapter
 		buffer_info = &tx_ring->buffer_info[i];
 		size = min(len, IXGB_MAX_DATA_PER_TXD);
 		buffer_info->length = size;
+		WARN_ON(buffer_info->dma != 0);
 		buffer_info->dma =
 			pci_map_single(adapter->pdev,
 				skb->data + offset,
@@ -1543,6 +1561,11 @@ void
 ixgb_update_stats(struct ixgb_adapter *adapter)
 {
 	struct net_device *netdev = adapter->netdev;
+	struct pci_dev *pdev = adapter->pdev;
+
+	/* Prevent stats update while adapter is being reset */
+	if (pdev->error_state && pdev->error_state != pci_channel_io_normal)
+		return;
 
 	if((netdev->flags & IFF_PROMISC) || (netdev->flags & IFF_ALLMULTI) ||
 	   (netdev->mc_count > IXGB_MAX_NUM_MULTICAST_ADDRESSES)) {
@@ -1787,7 +1810,7 @@ ixgb_clean_tx_irq(struct ixgb_adapter *a
 	if (unlikely(netif_queue_stopped(netdev))) {
 		spin_lock(&adapter->tx_lock);
 		if (netif_queue_stopped(netdev) && netif_carrier_ok(netdev) &&
-		    (IXGB_DESC_UNUSED(tx_ring) > IXGB_TX_QUEUE_WAKE))
+		    (IXGB_DESC_UNUSED(tx_ring) >= DESC_NEEDED))
 			netif_wake_queue(netdev);
 		spin_unlock(&adapter->tx_lock);
 	}
@@ -1948,7 +1971,7 @@ ixgb_clean_rx_irq(struct ixgb_adapter *a
 #define IXGB_CB_LENGTH 256
 		if (length < IXGB_CB_LENGTH) {
 			struct sk_buff *new_skb =
-			    dev_alloc_skb(length + NET_IP_ALIGN);
+			    netdev_alloc_skb(netdev, length + NET_IP_ALIGN);
 			if (new_skb) {
 				skb_reserve(new_skb, NET_IP_ALIGN);
 				new_skb->dev = netdev;
@@ -2032,7 +2055,7 @@ ixgb_alloc_rx_buffers(struct ixgb_adapte
 	while(--cleancount > 2) {
 		/* recycle! its good for you */
 		if (!(skb = buffer_info->skb))
-			skb = dev_alloc_skb(adapter->rx_buffer_len
+			skb = netdev_alloc_skb(netdev, adapter->rx_buffer_len
 			                    + NET_IP_ALIGN);
 		else {
 			skb_trim(skb, 0);
@@ -2190,7 +2213,7 @@ ixgb_restore_vlan(struct ixgb_adapter *a
 
 static void ixgb_netpoll(struct net_device *dev)
 {
-	struct ixgb_adapter *adapter = dev->priv;
+	struct ixgb_adapter *adapter = netdev_priv(dev);
 
 	disable_irq(adapter->pdev->irq);
 	ixgb_intr(adapter->pdev->irq, dev, NULL);
@@ -2198,4 +2221,98 @@ static void ixgb_netpoll(struct net_devi
 }
 #endif
 
+/**
+ * ixgb_io_error_detected() - called when PCI error is detected
+ * @pdev    pointer to pci device with error
+ * @state   pci channel state after error
+ *
+ * This callback is called by the PCI subsystem whenever
+ * a PCI bus error is detected.
+ */
+static pci_ers_result_t ixgb_io_error_detected (struct pci_dev *pdev,
+			             enum pci_channel_state state)
+{
+	struct net_device *netdev = pci_get_drvdata(pdev);
+	struct ixgb_adapter *adapter = netdev->priv;
+
+	if(netif_running(netdev))
+		ixgb_down(adapter, TRUE);
+
+	pci_disable_device(pdev);
+
+	/* Request a slot reset. */
+	return PCI_ERS_RESULT_NEED_RESET;
+}
+
+/**
+ * ixgb_io_slot_reset - called after the pci bus has been reset.
+ * @pdev    pointer to pci device with error
+ *
+ * This callback is called after the PCI buss has been reset.
+ * Basically, this tries to restart the card from scratch.
+ * This is a shortened version of the device probe/discovery code,
+ * it resembles the first-half of the ixgb_probe() routine.
+ */
+static pci_ers_result_t ixgb_io_slot_reset (struct pci_dev *pdev)
+{
+	struct net_device *netdev = pci_get_drvdata(pdev);
+	struct ixgb_adapter *adapter = netdev->priv;
+
+	if(pci_enable_device(pdev)) {
+		DPRINTK(PROBE, ERR, "Cannot re-enable PCI device after reset.\n");
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	/* Perform card reset only on one instance of the card */
+	if (0 != PCI_FUNC (pdev->devfn))
+		return PCI_ERS_RESULT_RECOVERED;
+
+	pci_set_master(pdev);
+
+	netif_carrier_off(netdev);
+	netif_stop_queue(netdev);
+	ixgb_reset(adapter);
+
+	/* Make sure the EEPROM is good */
+	if(!ixgb_validate_eeprom_checksum(&adapter->hw)) {
+		DPRINTK(PROBE, ERR, "After reset, the EEPROM checksum is not valid.\n");
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+	ixgb_get_ee_mac_addr(&adapter->hw, netdev->dev_addr);
+	memcpy(netdev->perm_addr, netdev->dev_addr, netdev->addr_len);
+
+	if(!is_valid_ether_addr(netdev->perm_addr)) {
+		DPRINTK(PROBE, ERR, "After reset, invalid MAC address.\n");
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	return PCI_ERS_RESULT_RECOVERED;
+}
+
+/**
+ * ixgb_io_resume - called when its OK to resume normal operations
+ * @pdev    pointer to pci device with error
+ *
+ * The error recovery driver tells us that its OK to resume
+ * normal operation. Implementation resembles the second-half
+ * of the ixgb_probe() routine.
+ */
+static void ixgb_io_resume (struct pci_dev *pdev)
+{
+	struct net_device *netdev = pci_get_drvdata(pdev);
+	struct ixgb_adapter *adapter = netdev->priv;
+
+	pci_set_master(pdev);
+
+	if(netif_running(netdev)) {
+		if(ixgb_up(adapter)) {
+			printk ("ixgb: can't bring device back up after reset\n");
+			return;
+		}
+	}
+
+	netif_device_attach(netdev);
+	mod_timer(&adapter->watchdog_timer, jiffies);
+}
+
 /* ixgb_main.c */