add bnx2x driver for BCM57710

Signed-off-by: Eliezer Tamir <eliezert@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 files changed, 23846 insertions, 0 deletions

diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 7ae9024b583..b034410b7ab 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -2597,6 +2597,15 @@ config TEHUTI
 	help
 	  Tehuti Networks 10G Ethernet NIC
 
+config BNX2X
+	tristate "Broadcom NetXtremeII 10Gb support"
+	depends on PCI
+	help
+	  This driver supports Broadcom NetXtremeII 10 gigabit Ethernet cards.
+	  To compile this driver as a module, choose M here: the module
+	  will be called bnx2x.  This is recommended.
+
+
 endif # NETDEV_10000
 
 source "drivers/net/tokenring/Kconfig"
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 5dd2d5eb191..5e36f203ce5 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -65,6 +65,7 @@ obj-$(CONFIG_STNIC) += stnic.o 8390.o
 obj-$(CONFIG_FEALNX) += fealnx.o
 obj-$(CONFIG_TIGON3) += tg3.o
 obj-$(CONFIG_BNX2) += bnx2.o
+obj-$(CONFIG_BNX2X) += bnx2x.o
 spidernet-y += spider_net.o spider_net_ethtool.o
 obj-$(CONFIG_SPIDER_NET) += spidernet.o sungem_phy.o
 obj-$(CONFIG_GELIC_NET) += ps3_gelic.o
diff --git a/drivers/net/bnx2x.c b/drivers/net/bnx2x.c
new file mode 100644
index 00000000000..e8c5754798e
--- /dev/null
+++ b/drivers/net/bnx2x.c
@@ -0,0 +1,9065 @@
+/* bnx2x.c: Broadcom Everest network driver.
+ *
+ * Copyright (c) 2007 Broadcom Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation.
+ *
+ * Written by: Eliezer Tamir <eliezert@broadcom.com>
+ * Based on code from Michael Chan's bnx2 driver
+ * UDP CSUM errata workaround by Arik Gendelman
+ * Slowpath rework by Vladislav Zolotarov
+ * Statistics and Link managment by Yitchak Gertner
+ *
+ */
+
+/* define this to make the driver freeze on error
+ * to allow getting debug info
+ * (you will need to reboot afterwords)
+ */
+/*#define BNX2X_STOP_ON_ERROR*/
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/kernel.h>
+#include <linux/device.h>  /* for dev_info() */
+#include <linux/timer.h>
+#include <linux/errno.h>
+#include <linux/ioport.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/interrupt.h>
+#include <linux/pci.h>
+#include <linux/init.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/dma-mapping.h>
+#include <linux/bitops.h>
+#include <linux/irq.h>
+#include <linux/delay.h>
+#include <asm/byteorder.h>
+#include <linux/time.h>
+#include <linux/ethtool.h>
+#include <linux/mii.h>
+#ifdef NETIF_F_HW_VLAN_TX
+	#include <linux/if_vlan.h>
+	#define BCM_VLAN 1
+#endif
+#include <net/ip.h>
+#include <net/tcp.h>
+#include <net/checksum.h>
+#include <linux/workqueue.h>
+#include <linux/crc32.h>
+#include <linux/prefetch.h>
+#include <linux/zlib.h>
+#include <linux/version.h>
+#include <linux/io.h>
+
+#include "bnx2x_reg.h"
+#include "bnx2x_fw_defs.h"
+#include "bnx2x_hsi.h"
+#include "bnx2x.h"
+#include "bnx2x_init.h"
+
+#define DRV_MODULE_VERSION      "0.40.15"
+#define DRV_MODULE_RELDATE      "$DateTime: 2007/11/15 07:28:37 $"
+#define BNX2X_BC_VER    	0x040009
+
+/* Time in jiffies before concluding the transmitter is hung. */
+#define TX_TIMEOUT      	(5*HZ)
+
+static const char version[] __devinitdata =
+	"Broadcom NetXtreme II 577xx 10Gigabit Ethernet Driver "
+	DRV_MODULE_NAME " " DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
+
+MODULE_AUTHOR("Eliezer Tamir <eliezert@broadcom.com>");
+MODULE_DESCRIPTION("Broadcom NetXtreme II BCM57710 Driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(DRV_MODULE_VERSION);
+MODULE_INFO(cvs_version, "$Revision: #356 $");
+
+static int use_inta;
+static int poll;
+static int onefunc;
+static int nomcp;
+static int debug;
+static int use_multi;
+
+module_param(use_inta, int, 0);
+module_param(poll, int, 0);
+module_param(onefunc, int, 0);
+module_param(debug, int, 0);
+MODULE_PARM_DESC(use_inta, "use INT#A instead of MSI-X");
+MODULE_PARM_DESC(poll, "use polling (for debug)");
+MODULE_PARM_DESC(onefunc, "enable only first function");
+MODULE_PARM_DESC(nomcp, "ignore managment CPU (Implies onefunc)");
+MODULE_PARM_DESC(debug, "defualt debug msglevel");
+
+#ifdef BNX2X_MULTI
+module_param(use_multi, int, 0);
+MODULE_PARM_DESC(use_multi, "use per-CPU queues");
+#endif
+
+enum bnx2x_board_type {
+	BCM57710 = 0,
+};
+
+/* indexed by board_t, above */
+static const struct {
+	char *name;
+} board_info[] __devinitdata = {
+	{ "Broadcom NetXtreme II BCM57710 XGb" }
+};
+
+static const struct pci_device_id bnx2x_pci_tbl[] = {
+	{ PCI_VENDOR_ID_BROADCOM, PCI_DEVICE_ID_NX2_57710,
+		PCI_ANY_ID, PCI_ANY_ID, 0, 0, BCM57710 },
+	{ 0 }
+};
+
+MODULE_DEVICE_TABLE(pci, bnx2x_pci_tbl);
+
+/****************************************************************************
+* General service functions
+****************************************************************************/
+
+/* used only at init
+ * locking is done by mcp
+ */
+static void bnx2x_reg_wr_ind(struct bnx2x *bp, u32 addr, u32 val)
+{
+	pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS, addr);
+	pci_write_config_dword(bp->pdev, PCICFG_GRC_DATA, val);
+	pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
+			       PCICFG_VENDOR_ID_OFFSET);
+}
+
+#ifdef BNX2X_IND_RD
+static u32 bnx2x_reg_rd_ind(struct bnx2x *bp, u32 addr)
+{
+	u32 val;
+
+	pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS, addr);
+	pci_read_config_dword(bp->pdev, PCICFG_GRC_DATA, &val);
+	pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
+			       PCICFG_VENDOR_ID_OFFSET);
+
+	return val;
+}
+#endif
+
+static const u32 dmae_reg_go_c[] = {
+	DMAE_REG_GO_C0, DMAE_REG_GO_C1, DMAE_REG_GO_C2, DMAE_REG_GO_C3,
+	DMAE_REG_GO_C4, DMAE_REG_GO_C5, DMAE_REG_GO_C6, DMAE_REG_GO_C7,
+	DMAE_REG_GO_C8, DMAE_REG_GO_C9, DMAE_REG_GO_C10, DMAE_REG_GO_C11,
+	DMAE_REG_GO_C12, DMAE_REG_GO_C13, DMAE_REG_GO_C14, DMAE_REG_GO_C15
+};
+
+/* copy command into DMAE command memory and set DMAE command go */
+static void bnx2x_post_dmae(struct bnx2x *bp, struct dmae_command *dmae,
+			    int idx)
+{
+	u32 cmd_offset;
+	int i;
+
+	cmd_offset = (DMAE_REG_CMD_MEM + sizeof(struct dmae_command) * idx);
+	for (i = 0; i < (sizeof(struct dmae_command)/4); i++) {
+		REG_WR(bp, cmd_offset + i*4, *(((u32 *)dmae) + i));
+
+/*      	DP(NETIF_MSG_DMAE, "DMAE cmd[%d].%d (0x%08x) : 0x%08x\n",
+		   idx, i, cmd_offset + i*4, *(((u32 *)dmae) + i)); */
+	}
+	REG_WR(bp, dmae_reg_go_c[idx], 1);
+}
+
+static void bnx2x_write_dmae(struct bnx2x *bp, dma_addr_t dma_addr,
+			     u32 dst_addr, u32 len32)
+{
+	struct dmae_command *dmae = &bp->dmae;
+	int port = bp->port;
+	u32 *wb_comp = bnx2x_sp(bp, wb_comp);
+	int timeout = 200;
+
+	memset(dmae, 0, sizeof(struct dmae_command));
+
+	dmae->opcode = (DMAE_CMD_SRC_PCI | DMAE_CMD_DST_GRC |
+			DMAE_CMD_C_DST_PCI | DMAE_CMD_C_ENABLE |
+			DMAE_CMD_SRC_RESET | DMAE_CMD_DST_RESET |
+#ifdef __BIG_ENDIAN
+			DMAE_CMD_ENDIANITY_B_DW_SWAP |
+#else
+			DMAE_CMD_ENDIANITY_DW_SWAP |
+#endif
+			(port ? DMAE_CMD_PORT_1 : DMAE_CMD_PORT_0));
+	dmae->src_addr_lo = U64_LO(dma_addr);
+	dmae->src_addr_hi = U64_HI(dma_addr);
+	dmae->dst_addr_lo = dst_addr >> 2;
+	dmae->dst_addr_hi = 0;
+	dmae->len = len32;
+	dmae->comp_addr_lo = U64_LO(bnx2x_sp_mapping(bp, wb_comp));
+	dmae->comp_addr_hi = U64_HI(bnx2x_sp_mapping(bp, wb_comp));
+	dmae->comp_val = BNX2X_WB_COMP_VAL;
+
+/*
+	DP(NETIF_MSG_DMAE, "dmae: opcode 0x%08x\n"
+	   DP_LEVEL "src_addr  [%x:%08x]  len [%d *4]  "
+		    "dst_addr [%x:%08x (%08x)]\n"
+	   DP_LEVEL "comp_addr [%x:%08x]  comp_val 0x%08x\n",
+	   dmae->opcode, dmae->src_addr_hi, dmae->src_addr_lo,
+	   dmae->len, dmae->dst_addr_hi, dmae->dst_addr_lo, dst_addr,
+	   dmae->comp_addr_hi, dmae->comp_addr_lo, dmae->comp_val);
+*/
+/*
+	DP(NETIF_MSG_DMAE, "data [0x%08x 0x%08x 0x%08x 0x%08x]\n",
+	   bp->slowpath->wb_data[0], bp->slowpath->wb_data[1],
+	   bp->slowpath->wb_data[2], bp->slowpath->wb_data[3]);
+*/
+
+	*wb_comp = 0;
+
+	bnx2x_post_dmae(bp, dmae, port * 8);
+
+	udelay(5);
+	/* adjust timeout for emulation/FPGA */
+	if (CHIP_REV_IS_SLOW(bp))
+		timeout *= 100;
+	while (*wb_comp != BNX2X_WB_COMP_VAL) {
+/*      	DP(NETIF_MSG_DMAE, "wb_comp 0x%08x\n", *wb_comp); */
+		udelay(5);
+		if (!timeout) {
+			BNX2X_ERR("dmae timeout!\n");
+			break;
+		}
+		timeout--;
+	}
+}
+
+#ifdef BNX2X_DMAE_RD
+static void bnx2x_read_dmae(struct bnx2x *bp, u32 src_addr, u32 len32)
+{
+	struct dmae_command *dmae = &bp->dmae;
+	int port = bp->port;
+	u32 *wb_comp = bnx2x_sp(bp, wb_comp);
+	int timeout = 200;
+
+	memset(bnx2x_sp(bp, wb_data[0]), 0, sizeof(u32) * 4);
+	memset(dmae, 0, sizeof(struct dmae_command));
+
+	dmae->opcode = (DMAE_CMD_SRC_GRC | DMAE_CMD_DST_PCI |
+			DMAE_CMD_C_DST_PCI | DMAE_CMD_C_ENABLE |
+			DMAE_CMD_SRC_RESET | DMAE_CMD_DST_RESET |
+#ifdef __BIG_ENDIAN
+			DMAE_CMD_ENDIANITY_B_DW_SWAP |
+#else
+			DMAE_CMD_ENDIANITY_DW_SWAP |
+#endif
+			(port ? DMAE_CMD_PORT_1 : DMAE_CMD_PORT_0));
+	dmae->src_addr_lo = src_addr >> 2;
+	dmae->src_addr_hi = 0;
+	dmae->dst_addr_lo = U64_LO(bnx2x_sp_mapping(bp, wb_data));
+	dmae->dst_addr_hi = U64_HI(bnx2x_sp_mapping(bp, wb_data));
+	dmae->len = len32;
+	dmae->comp_addr_lo = U64_LO(bnx2x_sp_mapping(bp, wb_comp));
+	dmae->comp_addr_hi = U64_HI(bnx2x_sp_mapping(bp, wb_comp));
+	dmae->comp_val = BNX2X_WB_COMP_VAL;
+
+/*
+	DP(NETIF_MSG_DMAE, "dmae: opcode 0x%08x\n"
+	   DP_LEVEL "src_addr  [%x:%08x]  len [%d *4]  "
+		    "dst_addr [%x:%08x (%08x)]\n"
+	   DP_LEVEL "comp_addr [%x:%08x]  comp_val 0x%08x\n",
+	   dmae->opcode, dmae->src_addr_hi, dmae->src_addr_lo,
+	   dmae->len, dmae->dst_addr_hi, dmae->dst_addr_lo, src_addr,
+	   dmae->comp_addr_hi, dmae->comp_addr_lo, dmae->comp_val);
+*/
+
+	*wb_comp = 0;
+
+	bnx2x_post_dmae(bp, dmae, port * 8);
+
+	udelay(5);
+	while (*wb_comp != BNX2X_WB_COMP_VAL) {
+		udelay(5);
+		if (!timeout) {
+			BNX2X_ERR("dmae timeout!\n");
+			break;
+		}
+		timeout--;
+	}
+/*
+	DP(NETIF_MSG_DMAE, "data [0x%08x 0x%08x 0x%08x 0x%08x]\n",
+	   bp->slowpath->wb_data[0], bp->slowpath->wb_data[1],
+	   bp->slowpath->wb_data[2], bp->slowpath->wb_data[3]);
+*/
+}
+#endif
+
+static int bnx2x_mc_assert(struct bnx2x *bp)
+{
+	int i, j;
+	int rc = 0;
+	char last_idx;
+	const char storm[] = {"XTCU"};
+	const u32 intmem_base[] = {
+		BAR_XSTRORM_INTMEM,
+		BAR_TSTRORM_INTMEM,
+		BAR_CSTRORM_INTMEM,
+		BAR_USTRORM_INTMEM
+	};
+
+	/* Go through all instances of all SEMIs */
+	for (i = 0; i < 4; i++) {
+		last_idx = REG_RD8(bp, XSTORM_ASSERT_LIST_INDEX_OFFSET +
+				   intmem_base[i]);
+		BNX2X_ERR("DATA %cSTORM_ASSERT_LIST_INDEX 0x%x\n",
+			  storm[i], last_idx);
+
+		/* print the asserts */
+		for (j = 0; j < STROM_ASSERT_ARRAY_SIZE; j++) {
+			u32 row0, row1, row2, row3;
+
+			row0 = REG_RD(bp, XSTORM_ASSERT_LIST_OFFSET(j) +
+				      intmem_base[i]);
+			row1 = REG_RD(bp, XSTORM_ASSERT_LIST_OFFSET(j) + 4 +
+				      intmem_base[i]);
+			row2 = REG_RD(bp, XSTORM_ASSERT_LIST_OFFSET(j) + 8 +
+				      intmem_base[i]);
+			row3 = REG_RD(bp, XSTORM_ASSERT_LIST_OFFSET(j) + 12 +
+				      intmem_base[i]);
+
+			if (row0 != COMMON_ASM_INVALID_ASSERT_OPCODE) {
+				BNX2X_ERR("DATA %cSTORM_ASSERT_INDEX 0x%x ="
+					  " 0x%08x 0x%08x 0x%08x 0x%08x\n",
+					  storm[i], j, row3, row2, row1, row0);
+				rc++;
+			} else {
+				break;
+			}
+		}
+	}
+	return rc;
+}
+static void bnx2x_fw_dump(struct bnx2x *bp)
+{
+	u32 mark, offset;
+	u32 data[9];
+	int word;
+
+	mark = REG_RD(bp, MCP_REG_MCPR_SCRATCH + 0xf104);
+	printk(KERN_ERR PFX "begin fw dump (mark 0x%x)\n", mark);
+
+	for (offset = mark - 0x08000000; offset <= 0xF900; offset += 0x8*4) {
+		for (word = 0; word < 8; word++)
+			data[word] = htonl(REG_RD(bp, MCP_REG_MCPR_SCRATCH +
+						  offset + 4*word));
+		data[8] = 0x0;
+		printk(KERN_ERR PFX "%s", (char *)data);
+	}
+	for (offset = 0xF108; offset <= mark - 0x08000000; offset += 0x8*4) {
+		for (word = 0; word < 8; word++)
+			data[word] = htonl(REG_RD(bp, MCP_REG_MCPR_SCRATCH +
+						  offset + 4*word));
+		data[8] = 0x0;
+		printk(KERN_ERR PFX "%s", (char *)data);
+	}
+	printk("\n" KERN_ERR PFX "end of fw dump\n");
+}
+
+static void bnx2x_panic_dump(struct bnx2x *bp)
+{
+	int i;
+	u16 j, start, end;
+
+	BNX2X_ERR("begin crash dump -----------------\n");
+
+	for_each_queue(bp, i) {
+		struct bnx2x_fastpath *fp = &bp->fp[i];
+		struct eth_tx_db_data *hw_prods = fp->hw_tx_prods;
+
+		BNX2X_ERR("queue[%d]: tx_pkt_prod(%x)  tx_pkt_cons(%x)"
+			  "  tx_bd_prod(%x)  tx_bd_cons(%x)  *tx_cons_sb(%x)"
+			  "  *rx_cons_sb(%x)  rx_comp_prod(%x)"
+			  "  rx_comp_cons(%x)  fp_c_idx(%x)  fp_u_idx(%x)"
+			  "  bd data(%x,%x)\n",
+			  i, fp->tx_pkt_prod, fp->tx_pkt_cons, fp->tx_bd_prod,
+			  fp->tx_bd_cons, *fp->tx_cons_sb, *fp->rx_cons_sb,
+			  fp->rx_comp_prod, fp->rx_comp_cons, fp->fp_c_idx,
+			  fp->fp_u_idx, hw_prods->packets_prod,
+			  hw_prods->bds_prod);
+
+		start = TX_BD(le16_to_cpu(*fp->tx_cons_sb) - 10);
+		end = TX_BD(le16_to_cpu(*fp->tx_cons_sb) + 245);
+		for (j = start; j < end; j++) {
+			struct sw_tx_bd *sw_bd = &fp->tx_buf_ring[j];
+
+			BNX2X_ERR("packet[%x]=[%p,%x]\n", j,
+				  sw_bd->skb, sw_bd->first_bd);
+		}
+
+		start = TX_BD(fp->tx_bd_cons - 10);
+		end = TX_BD(fp->tx_bd_cons + 254);
+		for (j = start; j < end; j++) {
+			u32 *tx_bd = (u32 *)&fp->tx_desc_ring[j];
+
+			BNX2X_ERR("tx_bd[%x]=[%x:%x:%x:%x]\n",
+				  j, tx_bd[0], tx_bd[1], tx_bd[2], tx_bd[3]);
+		}
+
+		start = RX_BD(le16_to_cpu(*fp->rx_cons_sb) - 10);
+		end = RX_BD(le16_to_cpu(*fp->rx_cons_sb) + 503);
+		for (j = start; j < end; j++) {
+			u32 *rx_bd = (u32 *)&fp->rx_desc_ring[j];
+			struct sw_rx_bd *sw_bd = &fp->rx_buf_ring[j];
+
+			BNX2X_ERR("rx_bd[%x]=[%x:%x]  sw_bd=[%p]\n",
+				  j, rx_bd[0], rx_bd[1], sw_bd->skb);
+		}
+
+		start = RCQ_BD(fp->rx_comp_cons - 10);
+		end = RCQ_BD(fp->rx_comp_cons + 503);
+		for (j = start; j < end; j++) {
+			u32 *cqe = (u32 *)&fp->rx_comp_ring[j];
+
+			BNX2X_ERR("cqe[%x]=[%x:%x:%x:%x]\n",
+				  j, cqe[0], cqe[1], cqe[2], cqe[3]);
+		}
+	}
+
+	BNX2X_ERR("def_c_idx(%u)  def_u_idx(%u)  def_t_idx(%u)"
+		  "  def_x_idx(%u)  def_att_idx(%u)  attn_state(%u)"
+		  "  spq_prod_idx(%u)\n",
+		  bp->def_c_idx, bp->def_u_idx, bp->def_t_idx, bp->def_x_idx,
+		  bp->def_att_idx, bp->attn_state, bp->spq_prod_idx);
+
+
+	bnx2x_mc_assert(bp);
+	BNX2X_ERR("end crash dump -----------------\n");
+
+	bp->stats_state = STATS_STATE_DISABLE;
+	DP(BNX2X_MSG_STATS, "stats_state - DISABLE\n");
+}
+
+static void bnx2x_enable_int(struct bnx2x *bp)
+{
+	int port = bp->port;
+	u32 addr = port ? HC_REG_CONFIG_1 : HC_REG_CONFIG_0;
+	u32 val = REG_RD(bp, addr);
+	int msix = (bp->flags & USING_MSIX_FLAG) ? 1 : 0;
+
+	if (msix) {
+		val &= ~HC_CONFIG_0_REG_SINGLE_ISR_EN_0;
+		val |= (HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
+			HC_CONFIG_0_REG_ATTN_BIT_EN_0);
+	} else {
+		val |= (HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
+			HC_CONFIG_0_REG_INT_LINE_EN_0 |
+			HC_CONFIG_0_REG_ATTN_BIT_EN_0);
+		val &= ~HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0;
+	}
+
+	DP(NETIF_MSG_INTR, "write %x to HC %d (addr 0x%x)  msi %d\n",
+	   val, port, addr, msix);
+
+	REG_WR(bp, addr, val);
+}
+
+static void bnx2x_disable_int(struct bnx2x *bp)
+{
+	int port = bp->port;
+	u32 addr = port ? HC_REG_CONFIG_1 : HC_REG_CONFIG_0;
+	u32 val = REG_RD(bp, addr);
+
+	val &= ~(HC_CONFIG_0_REG_SINGLE_ISR_EN_0 |
+		 HC_CONFIG_0_REG_MSI_MSIX_INT_EN_0 |
+		 HC_CONFIG_0_REG_INT_LINE_EN_0 |
+		 HC_CONFIG_0_REG_ATTN_BIT_EN_0);
+
+	DP(NETIF_MSG_INTR, "write %x to HC %d (addr 0x%x)\n",
+	   val, port, addr);
+
+	REG_WR(bp, addr, val);
+	if (REG_RD(bp, addr) != val)
+		BNX2X_ERR("BUG! proper val not read from IGU!\n");
+}
+
+static void bnx2x_disable_int_sync(struct bnx2x *bp)
+{
+
+	int msix = (bp->flags & USING_MSIX_FLAG) ? 1 : 0;
+	int i;
+
+	atomic_inc(&bp->intr_sem);
+	/* prevent the HW from sending interrupts*/
+	bnx2x_disable_int(bp);
+
+	/* make sure all ISRs are done */
+	if (msix) {
+		for_each_queue(bp, i)
+			synchronize_irq(bp->msix_table[i].vector);
+
+		/* one more for the Slow Path IRQ */
+		synchronize_irq(bp->msix_table[i].vector);
+	} else
+		synchronize_irq(bp->pdev->irq);
+
+	/* make sure sp_task is not running */
+	cancel_work_sync(&bp->sp_task);
+
+}
+
+/* fast path code */
+
+/*
+ * general service functions
+ */
+
+static inline void bnx2x_ack_sb(struct bnx2x *bp, u8 id,
+				u8 storm, u16 index, u8 op, u8 update)
+{
+	u32 igu_addr = (IGU_ADDR_INT_ACK + IGU_PORT_BASE * bp->port) * 8;
+	struct igu_ack_register igu_ack;
+
+	igu_ack.status_block_index = index;
+	igu_ack.sb_id_and_flags =
+			((id << IGU_ACK_REGISTER_STATUS_BLOCK_ID_SHIFT) |
+			 (storm << IGU_ACK_REGISTER_STORM_ID_SHIFT) |
+			 (update << IGU_ACK_REGISTER_UPDATE_INDEX_SHIFT) |
+			 (op << IGU_ACK_REGISTER_INTERRUPT_MODE_SHIFT));
+
+/*      DP(NETIF_MSG_INTR, "write 0x%08x to IGU addr 0x%x\n",
+	   (*(u32 *)&igu_ack), BAR_IGU_INTMEM + igu_addr); */
+	REG_WR(bp, BAR_IGU_INTMEM + igu_addr, (*(u32 *)&igu_ack));
+}
+
+static inline u16 bnx2x_update_fpsb_idx(struct bnx2x_fastpath *fp)
+{
+	struct host_status_block *fpsb = fp->status_blk;
+	u16 rc = 0;
+
+	barrier(); /* status block is written to by the chip */
+	if (fp->fp_c_idx != fpsb->c_status_block.status_block_index) {
+		fp->fp_c_idx = fpsb->c_status_block.status_block_index;
+		rc |= 1;
+	}
+	if (fp->fp_u_idx != fpsb->u_status_block.status_block_index) {
+		fp->fp_u_idx = fpsb->u_status_block.status_block_index;
+		rc |= 2;
+	}
+	return rc;
+}
+
+static inline int bnx2x_has_work(struct bnx2x_fastpath *fp)
+{
+	u16 rx_cons_sb = le16_to_cpu(*fp->rx_cons_sb);
+
+	if ((rx_cons_sb & MAX_RCQ_DESC_CNT) == MAX_RCQ_DESC_CNT)
+		rx_cons_sb++;
+
+	if ((rx_cons_sb != fp->rx_comp_cons) ||
+	    (le16_to_cpu(*fp->tx_cons_sb) != fp->tx_pkt_cons))
+		return 1;
+
+	return 0;
+}
+
+static u16 bnx2x_ack_int(struct bnx2x *bp)
+{
+	u32 igu_addr = (IGU_ADDR_SIMD_MASK + IGU_PORT_BASE * bp->port) * 8;
+	u32 result = REG_RD(bp, BAR_IGU_INTMEM + igu_addr);
+
+/*      DP(NETIF_MSG_INTR, "read 0x%08x from IGU addr 0x%x\n",
+	   result, BAR_IGU_INTMEM + igu_addr); */
+
+#ifdef IGU_DEBUG
+#warning IGU_DEBUG active
+	if (result == 0) {
+		BNX2X_ERR("read %x from IGU\n", result);
+		REG_WR(bp, TM_REG_TIMER_SOFT_RST, 0);
+	}
+#endif
+	return result;
+}
+
+
+/*
+ * fast path service functions
+ */
+
+/* free skb in the packet ring at pos idx
+ * return idx of last bd freed
+ */
+static u16 bnx2x_free_tx_pkt(struct bnx2x *bp, struct bnx2x_fastpath *fp,
+			     u16 idx)
+{
+	struct sw_tx_bd *tx_buf = &fp->tx_buf_ring[idx];
+	struct eth_tx_bd *tx_bd;
+	struct sk_buff *skb = tx_buf->skb;
+	u16 bd_idx = tx_buf->first_bd;
+	int nbd;
+
+	DP(BNX2X_MSG_OFF, "pkt_idx %d  buff @(%p)->skb %p\n",
+	   idx, tx_buf, skb);
+
+	/* unmap first bd */
+	DP(BNX2X_MSG_OFF, "free bd_idx %d\n", bd_idx);
+	tx_bd = &fp->tx_desc_ring[bd_idx];
+	pci_unmap_single(bp->pdev, BD_UNMAP_ADDR(tx_bd),
+			 BD_UNMAP_LEN(tx_bd), PCI_DMA_TODEVICE);
+
+	nbd = le16_to_cpu(tx_bd->nbd) - 1;
+#ifdef BNX2X_STOP_ON_ERROR
+	if (nbd > (MAX_SKB_FRAGS + 2)) {
+		BNX2X_ERR("bad nbd!\n");
+		bnx2x_panic();
+	}
+#endif
+
+	/* Skip a parse bd and the TSO split header bd
+	   since they have no mapping */
+	if (nbd)
+		bd_idx = TX_BD(NEXT_TX_IDX(bd_idx));
+
+	if (tx_bd->bd_flags.as_bitfield & (ETH_TX_BD_FLAGS_IP_CSUM |
+					   ETH_TX_BD_FLAGS_TCP_CSUM |
+					   ETH_TX_BD_FLAGS_SW_LSO)) {
+		if (--nbd)
+			bd_idx = TX_BD(NEXT_TX_IDX(bd_idx));
+		tx_bd = &fp->tx_desc_ring[bd_idx];
+		/* is this a TSO split header bd? */
+		if (tx_bd->bd_flags.as_bitfield & ETH_TX_BD_FLAGS_SW_LSO) {
+			if (--nbd)
+				bd_idx = TX_BD(NEXT_TX_IDX(bd_idx));
+		}
+	}
+
+	/* now free frags */
+	while (nbd > 0) {
+
+		DP(BNX2X_MSG_OFF, "free frag bd_idx %d\n", bd_idx);
+		tx_bd = &fp->tx_desc_ring[bd_idx];
+		pci_unmap_page(bp->pdev, BD_UNMAP_ADDR(tx_bd),
+			       BD_UNMAP_LEN(tx_bd), PCI_DMA_TODEVICE);
+		if (--nbd)
+			bd_idx = TX_BD(NEXT_TX_IDX(bd_idx));
+	}
+
+	/* release skb */
+	BUG_TRAP(skb);
+	dev_kfree_skb(skb);
+	tx_buf->first_bd = 0;
+	tx_buf->skb = NULL;
+
+	return bd_idx;
+}
+
+static inline u32 bnx2x_tx_avail(struct bnx2x_fastpath *fp)
+{
+	u16 used;
+	u32 prod;
+	u32 cons;
+
+	/* Tell compiler that prod and cons can change */
+	barrier();
+	prod = fp->tx_bd_prod;
+	cons = fp->tx_bd_cons;
+
+	used = (NUM_TX_BD - NUM_TX_RINGS + prod - cons +
+		(cons / TX_DESC_CNT) - (prod / TX_DESC_CNT));
+
+	if (prod >= cons) {
+		/* used = prod - cons - prod/size + cons/size */
+		used -= NUM_TX_BD - NUM_TX_RINGS;
+	}
+
+	BUG_TRAP(used <= fp->bp->tx_ring_size);
+	BUG_TRAP((fp->bp->tx_ring_size - used) <= MAX_TX_AVAIL);
+
+	return (fp->bp->tx_ring_size - used);
+}
+
+static void bnx2x_tx_int(struct bnx2x_fastpath *fp, int work)
+{
+	struct bnx2x *bp = fp->bp;
+	u16 hw_cons, sw_cons, bd_cons = fp->tx_bd_cons;
+	int done = 0;
+
+#ifdef BNX2X_STOP_ON_ERROR
+	if (unlikely(bp->panic))
+		return;
+#endif
+
+	hw_cons = le16_to_cpu(*fp->tx_cons_sb);
+	sw_cons = fp->tx_pkt_cons;
+
+	while (sw_cons != hw_cons) {
+		u16 pkt_cons;
+
+		pkt_cons = TX_BD(sw_cons);
+
+		/* prefetch(bp->tx_buf_ring[pkt_cons].skb); */
+
+		DP(NETIF_MSG_TX_DONE, "hw_cons %u  sw_cons %u  pkt_cons %d\n",
+		   hw_cons, sw_cons, pkt_cons);
+
+/*      	if (NEXT_TX_IDX(sw_cons) != hw_cons) {
+			rmb();
+			prefetch(fp->tx_buf_ring[NEXT_TX_IDX(sw_cons)].skb);
+		}
+*/
+		bd_cons = bnx2x_free_tx_pkt(bp, fp, pkt_cons);
+		sw_cons++;
+		done++;
+
+		if (done == work)
+			break;
+	}
+
+	fp->tx_pkt_cons = sw_cons;
+	fp->tx_bd_cons = bd_cons;
+
+	/* Need to make the tx_cons update visible to start_xmit()
+	 * before checking for netif_queue_stopped().  Without the
+	 * memory barrier, there is a small possibility that start_xmit()
+	 * will miss it and cause the queue to be stopped forever.
+	 */
+	smp_mb();
+
+	/* TBD need a thresh? */
+	if (unlikely(netif_queue_stopped(bp->dev))) {
+
+		netif_tx_lock(bp->dev);
+
+		if (netif_queue_stopped(bp->dev) &&
+		    (bnx2x_tx_avail(fp) >= MAX_SKB_FRAGS + 3))
+			netif_wake_queue(bp->dev);
+
+		netif_tx_unlock(bp->dev);
+
+	}
+}
+
+static void bnx2x_sp_event(struct bnx2x_fastpath *fp,
+			   union eth_rx_cqe *rr_cqe)
+{
+	struct bnx2x *bp = fp->bp;
+	int cid = SW_CID(rr_cqe->ramrod_cqe.conn_and_cmd_data);
+	int command = CQE_CMD(rr_cqe->ramrod_cqe.conn_and_cmd_data);
+
+	DP(NETIF_MSG_RX_STATUS,
+	   "fp %d  cid %d  got ramrod #%d  state is %x  type is %d\n",
+	   fp->index, cid, command, bp->state, rr_cqe->ramrod_cqe.type);
+
+	bp->spq_left++;
+
+	if (fp->index) {
+		switch (command | fp->state) {
+		case (RAMROD_CMD_ID_ETH_CLIENT_SETUP |
+						BNX2X_FP_STATE_OPENING):
+			DP(NETIF_MSG_IFUP, "got MULTI[%d] setup ramrod\n",
+			   cid);
+			fp->state = BNX2X_FP_STATE_OPEN;
+			break;
+
+		case (RAMROD_CMD_ID_ETH_HALT | BNX2X_FP_STATE_HALTING):
+			DP(NETIF_MSG_IFDOWN, "got MULTI[%d] halt ramrod\n",
+			   cid);
+			fp->state = BNX2X_FP_STATE_HALTED;
+			break;
+
+		default:
+			BNX2X_ERR("unexpected MC reply(%d)  state is %x\n",
+				  command, fp->state);
+		}
+		mb(); /* force bnx2x_wait_ramrod to see the change */
+		return;
+	}
+	switch (command | bp->state) {
+	case (RAMROD_CMD_ID_ETH_PORT_SETUP | BNX2X_STATE_OPENING_WAIT4_PORT):
+		DP(NETIF_MSG_IFUP, "got setup ramrod\n");
+		bp->state = BNX2X_STATE_OPEN;
+		break;
+
+	case (RAMROD_CMD_ID_ETH_HALT | BNX2X_STATE_CLOSING_WAIT4_HALT):
+		DP(NETIF_MSG_IFDOWN, "got halt ramrod\n");
+		bp->state = BNX2X_STATE_CLOSING_WAIT4_DELETE;
+		fp->state = BNX2X_FP_STATE_HALTED;
+		break;
+
+	case (RAMROD_CMD_ID_ETH_PORT_DEL | BNX2X_STATE_CLOSING_WAIT4_DELETE):
+		DP(NETIF_MSG_IFDOWN, "got delete ramrod\n");
+		bp->state = BNX2X_STATE_CLOSING_WAIT4_UNLOAD;
+		break;
+
+	case (RAMROD_CMD_ID_ETH_CFC_DEL | BNX2X_STATE_CLOSING_WAIT4_HALT):
+		DP(NETIF_MSG_IFDOWN, "got delete ramrod for MULTI[%d]\n", cid);
+		bnx2x_fp(bp, cid, state) = BNX2X_FP_STATE_DELETED;
+		break;
+
+	case (RAMROD_CMD_ID_ETH_SET_MAC | BNX2X_STATE_OPEN):
+		DP(NETIF_MSG_IFUP, "got set mac ramrod\n");
+		break;
+
+	default:
+		BNX2X_ERR("unexpected ramrod (%d)  state is %x\n",
+			  command, bp->state);
+	}
+
+	mb(); /* force bnx2x_wait_ramrod to see the change */
+}
+
+static inline int bnx2x_alloc_rx_skb(struct bnx2x *bp,
+				     struct bnx2x_fastpath *fp, u16 index)
+{
+	struct sk_buff *skb;
+	struct sw_rx_bd *rx_buf = &fp->rx_buf_ring[index];
+	struct eth_rx_bd *rx_bd = &fp->rx_desc_ring[index];
+	dma_addr_t mapping;
+
+	skb = netdev_alloc_skb(bp->dev, bp->rx_buf_size);
+	if (unlikely(skb == NULL))
+		return -ENOMEM;
+
+	mapping = pci_map_single(bp->pdev, skb->data, bp->rx_buf_use_size,
+				 PCI_DMA_FROMDEVICE);
+	if (unlikely(dma_mapping_error(mapping))) {
+
+		dev_kfree_skb(skb);
+		return -ENOMEM;
+	}
+
+	rx_buf->skb = skb;
+	pci_unmap_addr_set(rx_buf, mapping, mapping);
+
+	rx_bd->addr_hi = cpu_to_le32(U64_HI(mapping));
+	rx_bd->addr_lo = cpu_to_le32(U64_LO(mapping));
+
+	return 0;
+}
+
+/* note that we are not allocating a new skb,
+ * we are just moving one from cons to prod
+ * we are not creating a new mapping,
+ * so there is no need to check for dma_mapping_error().
+ */
+static void bnx2x_reuse_rx_skb(struct bnx2x_fastpath *fp,
+			       struct sk_buff *skb, u16 cons, u16 prod)
+{
+	struct bnx2x *bp = fp->bp;
+	struct sw_rx_bd *cons_rx_buf = &fp->rx_buf_ring[cons];
+	struct sw_rx_bd *prod_rx_buf = &fp->rx_buf_ring[prod];
+	struct eth_rx_bd *cons_bd = &fp->rx_desc_ring[cons];
+	struct eth_rx_bd *prod_bd = &fp->rx_desc_ring[prod];
+
+	pci_dma_sync_single_for_device(bp->pdev,
+				       pci_unmap_addr(cons_rx_buf, mapping),
+				       bp->rx_offset + RX_COPY_THRESH,
+				       PCI_DMA_FROMDEVICE);
+
+	prod_rx_buf->skb = cons_rx_buf->skb;
+	pci_unmap_addr_set(prod_rx_buf, mapping,
+			   pci_unmap_addr(cons_rx_buf, mapping));
+	*prod_bd = *cons_bd;
+}
+
+static int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget)
+{
+	struct bnx2x *bp = fp->bp;
+	u16 bd_cons, bd_prod, comp_ring_cons;
+	u16 hw_comp_cons, sw_comp_cons, sw_comp_prod;
+	int rx_pkt = 0;
+
+#ifdef BNX2X_STOP_ON_ERROR
+	if (unlikely(bp->panic))
+		return 0;
+#endif
+
+	hw_comp_cons = le16_to_cpu(*fp->rx_cons_sb);
+	if ((hw_comp_cons & MAX_RCQ_DESC_CNT) == MAX_RCQ_DESC_CNT)
+		hw_comp_cons++;
+
+	bd_cons = fp->rx_bd_cons;
+	bd_prod = fp->rx_bd_prod;
+	sw_comp_cons = fp->rx_comp_cons;
+	sw_comp_prod = fp->rx_comp_prod;
+
+	/* Memory barrier necessary as speculative reads of the rx
+	 * buffer can be ahead of the index in the status block
+	 */
+	rmb();
+
+	DP(NETIF_MSG_RX_STATUS,
+	   "queue[%d]:  hw_comp_cons %u  sw_comp_cons %u\n",
+	   fp->index, hw_comp_cons, sw_comp_cons);
+
+	while (sw_comp_cons != hw_comp_cons) {
+		unsigned int len, pad;
+		struct sw_rx_bd *rx_buf;
+		struct sk_buff *skb;
+		union eth_rx_cqe *cqe;
+
+		comp_ring_cons = RCQ_BD(sw_comp_cons);
+		bd_prod = RX_BD(bd_prod);
+		bd_cons = RX_BD(bd_cons);
+
+		cqe = &fp->rx_comp_ring[comp_ring_cons];
+
+		DP(NETIF_MSG_RX_STATUS, "hw_comp_cons %u  sw_comp_cons %u"
+		   "  comp_ring (%u)  bd_ring (%u,%u)\n",
+		   hw_comp_cons, sw_comp_cons,
+		   comp_ring_cons, bd_prod, bd_cons);
+		DP(NETIF_MSG_RX_STATUS, "CQE type %x  err %x  status %x"
+		   "  queue %x  vlan %x  len %x\n",
+		   cqe->fast_path_cqe.type,
+		   cqe->fast_path_cqe.error_type_flags,
+		   cqe->fast_path_cqe.status_flags,
+		   cqe->fast_path_cqe.rss_hash_result,
+		   cqe->fast_path_cqe.vlan_tag, cqe->fast_path_cqe.pkt_len);
+
+		/* is this a slowpath msg? */
+		if (unlikely(cqe->fast_path_cqe.type)) {
+			bnx2x_sp_event(fp, cqe);
+			goto next_cqe;
+
+		/* this is an rx packet */
+		} else {
+			rx_buf = &fp->rx_buf_ring[bd_cons];
+			skb = rx_buf->skb;
+
+			len = le16_to_cpu(cqe->fast_path_cqe.pkt_len);
+			pad = cqe->fast_path_cqe.placement_offset;
+
+			pci_dma_sync_single_for_device(bp->pdev,
+					pci_unmap_addr(rx_buf, mapping),
+						       pad + RX_COPY_THRESH,
+						       PCI_DMA_FROMDEVICE);
+			prefetch(skb);
+			prefetch(((char *)(skb)) + 128);
+
+			/* is this an error packet? */
+			if (unlikely(cqe->fast_path_cqe.error_type_flags &
+							ETH_RX_ERROR_FALGS)) {
+			/* do we sometimes forward error packets anyway? */
+				DP(NETIF_MSG_RX_ERR,
+				   "ERROR flags(%u) Rx packet(%u)\n",
+				   cqe->fast_path_cqe.error_type_flags,
+				   sw_comp_cons);
+				/* TBD make sure MC counts this as a drop */
+				goto reuse_rx;
+			}
+
+			/* Since we don't have a jumbo ring
+			 * copy small packets if mtu > 1500
+			 */
+			if ((bp->dev->mtu > ETH_MAX_PACKET_SIZE) &&
+			    (len <= RX_COPY_THRESH)) {
+				struct sk_buff *new_skb;
+
+				new_skb = netdev_alloc_skb(bp->dev,
+							   len + pad);
+				if (new_skb == NULL) {
+					DP(NETIF_MSG_RX_ERR,
+					   "ERROR packet dropped "
+					   "because of alloc failure\n");
+					/* TBD count this as a drop? */
+					goto reuse_rx;
+				}
+
+				/* aligned copy */
+				skb_copy_from_linear_data_offset(skb, pad,
+						    new_skb->data + pad, len);
+				skb_reserve(new_skb, pad);
+				skb_put(new_skb, len);
+
+				bnx2x_reuse_rx_skb(fp, skb, bd_cons, bd_prod);
+
+				skb = new_skb;
+
+			} else if (bnx2x_alloc_rx_skb(bp, fp, bd_prod) == 0) {
+				pci_unmap_single(bp->pdev,
+					pci_unmap_addr(rx_buf, mapping),
+						 bp->rx_buf_use_size,
+						 PCI_DMA_FROMDEVICE);
+				skb_reserve(skb, pad);
+				skb_put(skb, len);
+
+			} else {
+				DP(NETIF_MSG_RX_ERR,
+				   "ERROR packet dropped because "
+				   "of alloc failure\n");
+reuse_rx:
+				bnx2x_reuse_rx_skb(fp, skb, bd_cons, bd_prod);
+				goto next_rx;
+			}
+
+			skb->protocol = eth_type_trans(skb, bp->dev);
+
+			skb->ip_summed = CHECKSUM_NONE;
+			if (bp->rx_csum && BNX2X_RX_SUM_OK(cqe))
+				skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+			/* TBD do we pass bad csum packets in promisc */
+		}
+
+#ifdef BCM_VLAN
+		if ((le16_to_cpu(cqe->fast_path_cqe.pars_flags.flags)
+				& PARSING_FLAGS_NUMBER_OF_NESTED_VLANS)
+		    && (bp->vlgrp != NULL))
+			vlan_hwaccel_receive_skb(skb, bp->vlgrp,
+				le16_to_cpu(cqe->fast_path_cqe.vlan_tag));
+		else
+#endif
+		netif_receive_skb(skb);
+
+		bp->dev->last_rx = jiffies;
+
+next_rx:
+		rx_buf->skb = NULL;
+
+		bd_cons = NEXT_RX_IDX(bd_cons);
+		bd_prod = NEXT_RX_IDX(bd_prod);
+next_cqe:
+		sw_comp_prod = NEXT_RCQ_IDX(sw_comp_prod);
+		sw_comp_cons = NEXT_RCQ_IDX(sw_comp_cons);
+		rx_pkt++;
+
+		if ((rx_pkt == budget))
+			break;
+	} /* while */
+
+	fp->rx_bd_cons = bd_cons;
+	fp->rx_bd_prod = bd_prod;
+	fp->rx_comp_cons = sw_comp_cons;
+	fp->rx_comp_prod = sw_comp_prod;
+
+	REG_WR(bp, BAR_TSTRORM_INTMEM +
+	       TSTORM_RCQ_PROD_OFFSET(bp->port, fp->index), sw_comp_prod);
+
+	mmiowb(); /* keep prod updates ordered