Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (37 commits) smc91c92_cs: fix the problem of "Unable to find hardware address" r8169: clean up my printk uglyness net: Hook up cxgb4 to Kconfig and Makefile cxgb4: Add main driver file and driver Makefile cxgb4: Add remaining driver headers and L2T management cxgb4: Add packet queues and packet DMA code cxgb4: Add HW and FW support code cxgb4: Add register, message, and FW definitions netlabel: Fix several rcu_dereference() calls used without RCU read locks bonding: fix potential deadlock in bond_uninit() net: check the length of the socket address passed to connect(2) stmmac: add documentation for the driver. stmmac: fix kconfig for crc32 build error be2net: fix bug in vlan rx path for big endian architecture be2net: fix flashing on big endian architectures be2net: fix a bug in flashing the redboot section bonding: bond_xmit_roundrobin() fix drivers/net: Add missing unlock net: gianfar - align BD ring size console messages net: gianfar - initialize per-queue statistics ...
author: Linus Torvalds <torvalds@linux-foundation.org> 2010-04-06 19:34:06 +0400
committer: Linus Torvalds <torvalds@linux-foundation.org> 2010-04-06 19:34:06 +0400
commit: cb4361c1dc29cd870f664c004b1817106fbce0fa (patch)
tree: c31533210d738c9074b21b8f31fe0399b39545f0
parent: 309361e09ca9e9670dc8664e5d14125bf82078af (diff)
parent: fb9e2d887243499b8d28efcf80821c4f6a092395 (diff)
download: linux-cb4361c1dc29cd870f664c004b1817106fbce0fa.tar.xz
47 files changed, 14221 insertions, 131 deletions
diff --git a/Documentation/networking/stmmac.txt b/Documentation/networking/stmmac.txt
new file mode 100644
index 000000000000..7ee770b5ef5f
--- /dev/null
+++ b/Documentation/networking/stmmac.txt
@@ -0,0 +1,143 @@
+       STMicroelectronics 10/100/1000 Synopsys Ethernet driver
+
+Copyright (C) 2007-2010  STMicroelectronics Ltd
+Author: Giuseppe Cavallaro <peppe.cavallaro@st.com>
+
+This is the driver for the MAC 10/100/1000 on-chip Ethernet controllers
+(Synopsys IP blocks); it has been fully tested on STLinux platforms.
+
+Currently this network device driver is for all STM embedded MAC/GMAC
+(7xxx SoCs).
+
+DWC Ether MAC 10/100/1000 Universal version 3.41a and DWC Ether MAC 10/100
+Universal version 4.0 have been used for developing the first code
+implementation.
+
+Please, for more information also visit: www.stlinux.com
+
+1) Kernel Configuration
+The kernel configuration option is STMMAC_ETH:
+ Device Drivers ---> Network device support ---> Ethernet (1000 Mbit) --->
+ STMicroelectronics 10/100/1000 Ethernet driver (STMMAC_ETH)
+
+2) Driver parameters list:
+	debug: message level (0: no output, 16: all);
+	phyaddr: to manually provide the physical address to the PHY device;
+	dma_rxsize: DMA rx ring size;
+	dma_txsize: DMA tx ring size;
+	buf_sz: DMA buffer size;
+	tc: control the HW FIFO threshold;
+	tx_coe: Enable/Disable Tx Checksum Offload engine;
+	watchdog: transmit timeout (in milliseconds);
+	flow_ctrl: Flow control ability [on/off];
+	pause: Flow Control Pause Time;
+	tmrate: timer period (only if timer optimisation is configured).
+
+3) Command line options
+Driver parameters can be also passed in command line by using:
+	stmmaceth=dma_rxsize:128,dma_txsize:512
+
+4) Driver information and notes
+
+4.1) Transmit process
+The xmit method is invoked when the kernel needs to transmit a packet; it sets
+the descriptors in the ring and informs the DMA engine that there is a packet
+ready to be transmitted.
+Once the controller has finished transmitting the packet, an interrupt is
+triggered; So the driver will be able to release the socket buffers.
+By default, the driver sets the NETIF_F_SG bit in the features field of the
+net_device structure enabling the scatter/gather feature.
+
+4.2) Receive process
+When one or more packets are received, an interrupt happens. The interrupts
+are not queued so the driver has to scan all the descriptors in the ring during
+the receive process.
+This is based on NAPI so the interrupt handler signals only if there is work to be
+done, and it exits.
+Then the poll method will be scheduled at some future point.
+The incoming packets are stored, by the DMA, in a list of pre-allocated socket
+buffers in order to avoid the memcpy (Zero-copy).
+
+4.3) Timer-Driver Interrupt
+Instead of having the device that asynchronously notifies the frame receptions, the
+driver configures a timer to generate an interrupt at regular intervals.
+Based on the granularity of the timer, the frames that are received by the device
+will experience different levels of latency. Some NICs have dedicated timer
+device to perform this task. STMMAC can use either the RTC device or the TMU
+channel 2  on STLinux platforms.
+The timers frequency can be passed to the driver as parameter; when change it,
+take care of both hardware capability and network stability/performance impact.
+Several performance tests on STM platforms showed this optimisation allows to spare
+the CPU while having the maximum throughput.
+
+4.4) WOL
+Wake up on Lan feature through Magic Frame is only supported for the GMAC
+core.
+
+4.5) DMA descriptors
+Driver handles both normal and enhanced descriptors. The latter has been only
+tested on DWC Ether MAC 10/100/1000 Universal version 3.41a.
+
+4.6) Ethtool support
+Ethtool is supported. Driver statistics and internal errors can be taken using:
+ethtool -S ethX command. It is possible to dump registers etc.
+
+4.7) Jumbo and Segmentation Offloading
+Jumbo frames are supported and tested for the GMAC.
+The GSO has been also added but it's performed in software.
+LRO is not supported.
+
+4.8) Physical
+The driver is compatible with PAL to work with PHY and GPHY devices.
+
+4.9) Platform information
+Several information came from the platform; please refer to the
+driver's Header file in include/linux directory.
+
+struct plat_stmmacenet_data {
+        int bus_id;
+        int pbl;
+        int has_gmac;
+        void (*fix_mac_speed)(void *priv, unsigned int speed);
+        void (*bus_setup)(unsigned long ioaddr);
+#ifdef CONFIG_STM_DRIVERS
+        struct stm_pad_config *pad_config;
+#endif
+        void *bsp_priv;
+};
+
+Where:
+- pbl (Programmable Burst Length) is maximum number of
+  beats to be transferred in one DMA transaction.
+  GMAC also enables the 4xPBL by default.
+- fix_mac_speed and bus_setup are used to configure internal target
+  registers (on STM platforms);
+- has_gmac: GMAC core is on board (get it at run-time in the next step);
+- bus_id: bus identifier.
+
+struct plat_stmmacphy_data {
+        int bus_id;
+        int phy_addr;
+        unsigned int phy_mask;
+        int interface;
+        int (*phy_reset)(void *priv);
+        void *priv;
+};
+
+Where:
+- bus_id: bus identifier;
+- phy_addr: physical address used for the attached phy device;
+            set it to -1 to get it at run-time;
+- interface: physical MII interface mode;
+- phy_reset: hook to reset HW function.
+
+TODO:
+- Continue to make the driver more generic and suitable for other Synopsys
+  Ethernet controllers used on other architectures (i.e. ARM).
+- 10G controllers are not supported.
+- MAC uses Normal descriptors and GMAC uses enhanced ones.
+  This is a limit that should be reviewed. MAC could want to
+  use the enhanced structure.
+- Checksumming: Rx/Tx csum is done in HW in case of GMAC only.
+- Review the timer optimisation code to use an embedded device that seems to be
+  available in new chip generations.
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 0ba5b8e50a7c..7b832c727f87 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -2582,6 +2582,31 @@ config CHELSIO_T3
 	  To compile this driver as a module, choose M here: the module
 	  will be called cxgb3.
 
+config CHELSIO_T4_DEPENDS
+	tristate
+	depends on PCI && INET
+	default y
+
+config CHELSIO_T4
+	tristate "Chelsio Communications T4 Ethernet support"
+	depends on CHELSIO_T4_DEPENDS
+	select FW_LOADER
+	select MDIO
+	help
+	  This driver supports Chelsio T4-based gigabit and 10Gb Ethernet
+	  adapters.
+
+	  For general information about Chelsio and our products, visit
+	  our website at <http://www.chelsio.com>.
+
+	  For customer support, please visit our customer support page at
+	  <http://www.chelsio.com/support.htm>.
+
+	  Please send feedback to <linux-bugs@chelsio.com>.
+
+	  To compile this driver as a module choose M here; the module
+	  will be called cxgb4.
+
 config EHEA
 	tristate "eHEA Ethernet support"
 	depends on IBMEBUS && INET && SPARSEMEM
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 478886234c28..a583b50d9de8 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -19,6 +19,7 @@ obj-$(CONFIG_IXGB) += ixgb/
 obj-$(CONFIG_IP1000) += ipg.o
 obj-$(CONFIG_CHELSIO_T1) += chelsio/
 obj-$(CONFIG_CHELSIO_T3) += cxgb3/
+obj-$(CONFIG_CHELSIO_T4) += cxgb4/
 obj-$(CONFIG_EHEA) += ehea/
 obj-$(CONFIG_CAN) += can/
 obj-$(CONFIG_BONDING) += bonding/
diff --git a/drivers/net/benet/be_cmds.c b/drivers/net/benet/be_cmds.c
index 50e6259b50e4..d0ef4ac987cd 100644
--- a/drivers/net/benet/be_cmds.c
+++ b/drivers/net/benet/be_cmds.c
@@ -1464,8 +1464,8 @@ int be_cmd_get_flash_crc(struct be_adapter *adapter, u8 *flashed_crc,
 
 	req->params.op_type = cpu_to_le32(IMG_TYPE_REDBOOT);
 	req->params.op_code = cpu_to_le32(FLASHROM_OPER_REPORT);
-	req->params.offset = offset;
-	req->params.data_buf_size = 0x4;
+	req->params.offset = cpu_to_le32(offset);
+	req->params.data_buf_size = cpu_to_le32(0x4);
 
 	status = be_mcc_notify_wait(adapter);
 	if (!status)
diff --git a/drivers/net/benet/be_main.c b/drivers/net/benet/be_main.c
index 43e8032f9236..ec6ace802256 100644
--- a/drivers/net/benet/be_main.c
+++ b/drivers/net/benet/be_main.c
@@ -807,7 +807,7 @@ static void be_rx_compl_process(struct be_adapter *adapter,
 			return;
 		}
 		vid = AMAP_GET_BITS(struct amap_eth_rx_compl, vlan_tag, rxcp);
-		vid = be16_to_cpu(vid);
+		vid = swab16(vid);
 		vlan_hwaccel_receive_skb(skb, adapter->vlan_grp, vid);
 	} else {
 		netif_receive_skb(skb);
@@ -884,7 +884,7 @@ static void be_rx_compl_process_gro(struct be_adapter *adapter,
 		napi_gro_frags(&eq_obj->napi);
 	} else {
 		vid = AMAP_GET_BITS(struct amap_eth_rx_compl, vlan_tag, rxcp);
-		vid = be16_to_cpu(vid);
+		vid = swab16(vid);
 
 		if (!adapter->vlan_grp || adapter->vlans_added == 0)
 			return;
@@ -1855,7 +1855,7 @@ static bool be_flash_redboot(struct be_adapter *adapter,
 	p += crc_offset;
 
 	status = be_cmd_get_flash_crc(adapter, flashed_crc,
-			(img_start + image_size - 4));
+			(image_size - 4));
 	if (status) {
 		dev_err(&adapter->pdev->dev,
 		"could not get crc from flash, not flashing redboot\n");
@@ -1991,7 +1991,7 @@ int be_load_fw(struct be_adapter *adapter, u8 *func)
 	struct flash_file_hdr_g3 *fhdr3;
 	struct image_hdr *img_hdr_ptr = NULL;
 	struct be_dma_mem flash_cmd;
-	int status, i = 0;
+	int status, i = 0, num_imgs = 0;
 	const u8 *p;
 
 	strcpy(fw_file, func);
@@ -2017,15 +2017,14 @@ int be_load_fw(struct be_adapter *adapter, u8 *func)
 	if ((adapter->generation == BE_GEN3) &&
 			(get_ufigen_type(fhdr) == BE_GEN3)) {
 		fhdr3 = (struct flash_file_hdr_g3 *) fw->data;
-		for (i = 0; i < fhdr3->num_imgs; i++) {
+		num_imgs = le32_to_cpu(fhdr3->num_imgs);
+		for (i = 0; i < num_imgs; i++) {
 			img_hdr_ptr = (struct image_hdr *) (fw->data +
 					(sizeof(struct flash_file_hdr_g3) +
-					i * sizeof(struct image_hdr)));
-			if (img_hdr_ptr->imageid == 1) {
-				status = be_flash_data(adapter, fw,
-						&flash_cmd, fhdr3->num_imgs);
-			}
-
+					 i * sizeof(struct image_hdr)));
+			if (le32_to_cpu(img_hdr_ptr->imageid) == 1)
+				status = be_flash_data(adapter, fw, &flash_cmd,
+							num_imgs);
 		}
 	} else if ((adapter->generation == BE_GEN2) &&
 			(get_ufigen_type(fhdr) == BE_GEN2)) {
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 5b92fbff431d..0075514bf32f 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -4156,7 +4156,7 @@ static int bond_xmit_roundrobin(struct sk_buff *skb, struct net_device *bond_dev
 	 * send the join/membership reports.  The curr_active_slave found
 	 * will send all of this type of traffic.
 	 */
-	if ((iph->protocol == htons(IPPROTO_IGMP)) &&
+	if ((iph->protocol == IPPROTO_IGMP) &&
 	    (skb->protocol == htons(ETH_P_IP))) {
 
 		read_lock(&bond->curr_slave_lock);
@@ -4450,6 +4450,14 @@ static const struct net_device_ops bond_netdev_ops = {
 	.ndo_vlan_rx_kill_vid	= bond_vlan_rx_kill_vid,
 };
 
+static void bond_destructor(struct net_device *bond_dev)
+{
+	struct bonding *bond = netdev_priv(bond_dev);
+	if (bond->wq)
+		destroy_workqueue(bond->wq);
+	free_netdev(bond_dev);
+}
+
 static void bond_setup(struct net_device *bond_dev)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
@@ -4470,7 +4478,7 @@ static void bond_setup(struct net_device *bond_dev)
 	bond_dev->ethtool_ops = &bond_ethtool_ops;
 	bond_set_mode_ops(bond, bond->params.mode);
 
-	bond_dev->destructor = free_netdev;
+	bond_dev->destructor = bond_destructor;
 
 	/* Initialize the device options */
 	bond_dev->tx_queue_len = 0;
@@ -4542,9 +4550,6 @@ static void bond_uninit(struct net_device *bond_dev)
 
 	bond_remove_proc_entry(bond);
 
-	if (bond->wq)
-		destroy_workqueue(bond->wq);
-
 	netif_addr_lock_bh(bond_dev);
 	bond_mc_list_destroy(bond);
 	netif_addr_unlock_bh(bond_dev);
@@ -4956,8 +4961,8 @@ int bond_create(struct net *net, const char *name)
 				bond_setup);
 	if (!bond_dev) {
 		pr_err("%s: eek! can't alloc netdev!\n", name);
-		res = -ENOMEM;
-		goto out;
+		rtnl_unlock();
+		return -ENOMEM;
 	}
 
 	dev_net_set(bond_dev, net);
@@ -4966,19 +4971,16 @@ int bond_create(struct net *net, const char *name)
 	if (!name) {
 		res = dev_alloc_name(bond_dev, "bond%d");
 		if (res < 0)
-			goto out_netdev;
+			goto out;
 	}
 
 	res = register_netdevice(bond_dev);
-	if (res < 0)
-		goto out_netdev;
 
 out:
 	rtnl_unlock();
+	if (res < 0)
+		bond_destructor(bond_dev);
 	return res;
-out_netdev:
-	free_netdev(bond_dev);
-	goto out;
 }
 
 static int __net_init bond_net_init(struct net *net)
diff --git a/drivers/net/cxgb4/Makefile b/drivers/net/cxgb4/Makefile
new file mode 100644
index 000000000000..498667487f52
--- /dev/null
+++ b/drivers/net/cxgb4/Makefile
@@ -0,0 +1,7 @@
+#
+# Chelsio T4 driver
+#
+
+obj-$(CONFIG_CHELSIO_T4) += cxgb4.o
+
+cxgb4-objs := cxgb4_main.o l2t.o t4_hw.o sge.o
diff --git a/drivers/net/cxgb4/cxgb4.h b/drivers/net/cxgb4/cxgb4.h
new file mode 100644
index 000000000000..3d8ff4889b56
--- /dev/null
+++ b/drivers/net/cxgb4/cxgb4.h
@@ -0,0 +1,741 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_H__
+#define __CXGB4_H__
+
+#include <linux/bitops.h>
+#include <linux/cache.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/netdevice.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
+#include <asm/io.h>
+#include "cxgb4_uld.h"
+#include "t4_hw.h"
+
+#define FW_VERSION_MAJOR 1
+#define FW_VERSION_MINOR 1
+#define FW_VERSION_MICRO 0
+
+enum {
+	MAX_NPORTS = 4,     /* max # of ports */
+	SERNUM_LEN = 16,    /* Serial # length */
+	EC_LEN     = 16,    /* E/C length */
+	ID_LEN     = 16,    /* ID length */
+};
+
+enum {
+	MEM_EDC0,
+	MEM_EDC1,
+	MEM_MC
+};
+
+enum dev_master {
+	MASTER_CANT,
+	MASTER_MAY,
+	MASTER_MUST
+};
+
+enum dev_state {
+	DEV_STATE_UNINIT,
+	DEV_STATE_INIT,
+	DEV_STATE_ERR
+};
+
+enum {
+	PAUSE_RX      = 1 << 0,
+	PAUSE_TX      = 1 << 1,
+	PAUSE_AUTONEG = 1 << 2
+};
+
+struct port_stats {
+	u64 tx_octets;            /* total # of octets in good frames */
+	u64 tx_frames;            /* all good frames */
+	u64 tx_bcast_frames;      /* all broadcast frames */
+	u64 tx_mcast_frames;      /* all multicast frames */
+	u64 tx_ucast_frames;      /* all unicast frames */
+	u64 tx_error_frames;      /* all error frames */
+
+	u64 tx_frames_64;         /* # of Tx frames in a particular range */
+	u64 tx_frames_65_127;
+	u64 tx_frames_128_255;
+	u64 tx_frames_256_511;
+	u64 tx_frames_512_1023;
+	u64 tx_frames_1024_1518;
+	u64 tx_frames_1519_max;
+
+	u64 tx_drop;              /* # of dropped Tx frames */
+	u64 tx_pause;             /* # of transmitted pause frames */
+	u64 tx_ppp0;              /* # of transmitted PPP prio 0 frames */
+	u64 tx_ppp1;              /* # of transmitted PPP prio 1 frames */
+	u64 tx_ppp2;              /* # of transmitted PPP prio 2 frames */
+	u64 tx_ppp3;              /* # of transmitted PPP prio 3 frames */
+	u64 tx_ppp4;              /* # of transmitted PPP prio 4 frames */
+	u64 tx_ppp5;              /* # of transmitted PPP prio 5 frames */
+	u64 tx_ppp6;              /* # of transmitted PPP prio 6 frames */
+	u64 tx_ppp7;              /* # of transmitted PPP prio 7 frames */
+
+	u64 rx_octets;            /* total # of octets in good frames */
+	u64 rx_frames;            /* all good frames */
+	u64 rx_bcast_frames;      /* all broadcast frames */
+	u64 rx_mcast_frames;      /* all multicast frames */
+	u64 rx_ucast_frames;      /* all unicast frames */
+	u64 rx_too_long;          /* # of frames exceeding MTU */
+	u64 rx_jabber;            /* # of jabber frames */
+	u64 rx_fcs_err;           /* # of received frames with bad FCS */
+	u64 rx_len_err;           /* # of received frames with length error */
+	u64 rx_symbol_err;        /* symbol errors */
+	u64 rx_runt;              /* # of short frames */
+
+	u64 rx_frames_64;         /* # of Rx frames in a particular range */
+	u64 rx_frames_65_127;
+	u64 rx_frames_128_255;
+	u64 rx_frames_256_511;
+	u64 rx_frames_512_1023;
+	u64 rx_frames_1024_1518;
+	u64 rx_frames_1519_max;
+
+	u64 rx_pause;             /* # of received pause frames */
+	u64 rx_ppp0;              /* # of received PPP prio 0 frames */
+	u64 rx_ppp1;              /* # of received PPP prio 1 frames */
+	u64 rx_ppp2;              /* # of received PPP prio 2 frames */
+	u64 rx_ppp3;              /* # of received PPP prio 3 frames */
+	u64 rx_ppp4;              /* # of received PPP prio 4 frames */
+	u64 rx_ppp5;              /* # of received PPP prio 5 frames */
+	u64 rx_ppp6;              /* # of received PPP prio 6 frames */
+	u64 rx_ppp7;              /* # of received PPP prio 7 frames */
+
+	u64 rx_ovflow0;           /* drops due to buffer-group 0 overflows */
+	u64 rx_ovflow1;           /* drops due to buffer-group 1 overflows */
+	u64 rx_ovflow2;           /* drops due to buffer-group 2 overflows */
+	u64 rx_ovflow3;           /* drops due to buffer-group 3 overflows */
+	u64 rx_trunc0;            /* buffer-group 0 truncated packets */
+	u64 rx_trunc1;            /* buffer-group 1 truncated packets */
+	u64 rx_trunc2;            /* buffer-group 2 truncated packets */
+	u64 rx_trunc3;            /* buffer-group 3 truncated packets */
+};
+
+struct lb_port_stats {
+	u64 octets;
+	u64 frames;
+	u64 bcast_frames;
+	u64 mcast_frames;
+	u64 ucast_frames;
+	u64 error_frames;
+
+	u64 frames_64;
+	u64 frames_65_127;
+	u64 frames_128_255;
+	u64 frames_256_511;
+	u64 frames_512_1023;
+	u64 frames_1024_1518;
+	u64 frames_1519_max;
+
+	u64 drop;
+
+	u64 ovflow0;
+	u64 ovflow1;
+	u64 ovflow2;
+	u64 ovflow3;
+	u64 trunc0;
+	u64 trunc1;
+	u64 trunc2;
+	u64 trunc3;
+};
+
+struct tp_tcp_stats {
+	u32 tcpOutRsts;
+	u64 tcpInSegs;
+	u64 tcpOutSegs;
+	u64 tcpRetransSegs;
+};
+
+struct tp_err_stats {
+	u32 macInErrs[4];
+	u32 hdrInErrs[4];
+	u32 tcpInErrs[4];
+	u32 tnlCongDrops[4];
+	u32 ofldChanDrops[4];
+	u32 tnlTxDrops[4];
+	u32 ofldVlanDrops[4];
+	u32 tcp6InErrs[4];
+	u32 ofldNoNeigh;
+	u32 ofldCongDefer;
+};
+
+struct tp_params {
+	unsigned int ntxchan;        /* # of Tx channels */
+	unsigned int tre;            /* log2 of core clocks per TP tick */
+};
+
+struct vpd_params {
+	unsigned int cclk;
+	u8 ec[EC_LEN + 1];
+	u8 sn[SERNUM_LEN + 1];
+	u8 id[ID_LEN + 1];
+};
+
+struct pci_params {
+	unsigned char speed;
+	unsigned char width;
+};
+
+struct adapter_params {
+	struct tp_params  tp;
+	struct vpd_params vpd;
+	struct pci_params pci;
+
+	unsigned int fw_vers;
+	unsigned int tp_vers;
+	u8 api_vers[7];
+
+	unsigned short mtus[NMTUS];
+	unsigned short a_wnd[NCCTRL_WIN];
+	unsigned short b_wnd[NCCTRL_WIN];
+
+	unsigned char nports;             /* # of ethernet ports */
+	unsigned char portvec;
+	unsigned char rev;                /* chip revision */
+	unsigned char offload;
+
+	unsigned int ofldq_wr_cred;
+};
+
+struct trace_params {
+	u32 data[TRACE_LEN / 4];
+	u32 mask[TRACE_LEN / 4];
+	unsigned short snap_len;
+	unsigned short min_len;
+	unsigned char skip_ofst;
+	unsigned char skip_len;
+	unsigned char invert;
+	unsigned char port;
+};
+
+struct link_config {
+	unsigned short supported;        /* link capabilities */
+	unsigned short advertising;      /* advertised capabilities */
+	unsigned short requested_speed;  /* speed user has requested */
+	unsigned short speed;            /* actual link speed */
+	unsigned char  requested_fc;     /* flow control user has requested */
+	unsigned char  fc;               /* actual link flow control */
+	unsigned char  autoneg;          /* autonegotiating? */
+	unsigned char  link_ok;          /* link up? */
+};
+
+#define FW_LEN16(fw_struct) FW_CMD_LEN16(sizeof(fw_struct) / 16)
+
+enum {
+	MAX_ETH_QSETS = 32,           /* # of Ethernet Tx/Rx queue sets */
+	MAX_OFLD_QSETS = 16,          /* # of offload Tx/Rx queue sets */
+	MAX_CTRL_QUEUES = NCHAN,      /* # of control Tx queues */
+	MAX_RDMA_QUEUES = NCHAN,      /* # of streaming RDMA Rx queues */
+};
+
+enum {
+	MAX_EGRQ = 128,         /* max # of egress queues, including FLs */
+	MAX_INGQ = 64           /* max # of interrupt-capable ingress queues */
+};
+
+struct adapter;
+struct vlan_group;
+struct sge_rspq;
+
+struct port_info {
+	struct adapter *adapter;
+	struct vlan_group *vlan_grp;
+	u16    viid;
+	s16    xact_addr_filt;        /* index of exact MAC address filter */
+	u16    rss_size;              /* size of VI's RSS table slice */
+	s8     mdio_addr;
+	u8     port_type;
+	u8     mod_type;
+	u8     port_id;
+	u8     tx_chan;
+	u8     lport;                 /* associated offload logical port */
+	u8     rx_offload;            /* CSO, etc */
+	u8     nqsets;                /* # of qsets */
+	u8     first_qset;            /* index of first qset */
+	struct link_config link_cfg;
+};
+
+/* port_info.rx_offload flags */
+enum {
+	RX_CSO = 1 << 0,
+};
+
+struct dentry;
+struct work_struct;
+
+enum {                                 /* adapter flags */
+	FULL_INIT_DONE     = (1 << 0),
+	USING_MSI          = (1 << 1),
+	USING_MSIX         = (1 << 2),
+	QUEUES_BOUND       = (1 << 3),
+	FW_OK              = (1 << 4),
+};
+
+struct rx_sw_desc;
+
+struct sge_fl {                     /* SGE free-buffer queue state */
+	unsigned int avail;         /* # of available Rx buffers */
+	unsigned int pend_cred;     /* new buffers since last FL DB ring */
+	unsigned int cidx;          /* consumer index */
+	unsigned int pidx;          /* producer index */
+	unsigned long alloc_failed; /* # of times buffer allocation failed */
+	unsigned long large_alloc_failed;
+	unsigned long starving;
+	/* RO fields */
+	unsigned int cntxt_id;      /* SGE context id for the free list */
+	unsigned int size;          /* capacity of free list */
+	struct rx_sw_desc *sdesc;   /* address of SW Rx descriptor ring */
+	__be64 *desc;               /* address of HW Rx descriptor ring */
+	dma_addr_t addr;            /* bus address of HW ring start */
+};
+
+/* A packet gather list */
+struct pkt_gl {
+	skb_frag_t frags[MAX_SKB_FRAGS];
+	void *va;                         /* virtual address of first byte */
+	unsigned int nfrags;              /* # of fragments */
+	unsigned int tot_len;             /* total length of fragments */
+};
+
+typedef int (*rspq_handler_t)(struct sge_rspq *q, const __be64 *rsp,
+			      const struct pkt_gl *gl);
+
+struct sge_rspq {                   /* state for an SGE response queue */
+	struct napi_struct napi;
+	const __be64 *cur_desc;     /* current descriptor in queue */
+	unsigned int cidx;          /* consumer index */
+	u8 gen;                     /* current generation bit */
+	u8 intr_params;             /* interrupt holdoff parameters */
+	u8 next_intr_params;        /* holdoff params for next interrupt */
+	u8 pktcnt_idx;              /* interrupt packet threshold */
+	u8 uld;                     /* ULD handling this queue */
+	u8 idx;                     /* queue index within its group */
+	int offset;                 /* offset into current Rx buffer */
+	u16 cntxt_id;               /* SGE context id for the response q */
+	u16 abs_id;                 /* absolute SGE id for the response q */
+	__be64 *desc;               /* address of HW response ring */
+	dma_addr_t phys_addr;       /* physical address of the ring */
+	unsigned int iqe_len;       /* entry size */
+	unsigned int size;          /* capacity of response queue */
+	struct adapter *adap;
+	struct net_device *netdev;  /* associated net device */
+	rspq_handler_t handler;
+};
+
+struct sge_eth_stats {              /* Ethernet queue statistics */
+	unsigned long pkts;         /* # of ethernet packets */
+	unsigned long lro_pkts;     /* # of LRO super packets */
+	unsigned long lro_merged;   /* # of wire packets merged by LRO */
+	unsigned long rx_cso;       /* # of Rx checksum offloads */
+	unsigned long vlan_ex;      /* # of Rx VLAN extractions */
+	unsigned long rx_drops;     /* # of packets dropped due to no mem */
+};
+
+struct sge_eth_rxq {                /* SW Ethernet Rx queue */
+	struct sge_rspq rspq;
+	struct sge_fl fl;
+	struct sge_eth_stats stats;
+} ____cacheline_aligned_in_smp;
+
+struct sge_ofld_stats {             /* offload queue statistics */
+	unsigned long pkts;         /* # of packets */
+	unsigned long imm;          /* # of immediate-data packets */
+	unsigned long an;           /* # of asynchronous notifications */
+	unsigned long nomem;        /* # of responses deferred due to no mem */
+};
+
+struct sge_ofld_rxq {               /* SW offload Rx queue */
+	struct sge_rspq rspq;
+	struct sge_fl fl;
+	struct sge_ofld_stats stats;
+} ____cacheline_aligned_in_smp;
+
+struct tx_desc {
+	__be64 flit[8];
+};
+
+struct tx_sw_desc;
+
+struct sge_txq {
+	unsigned int  in_use;       /* # of in-use Tx descriptors */
+	unsigned int  size;         /* # of descriptors */
+	unsigned int  cidx;         /* SW consumer index */
+	unsigned int  pidx;         /* producer index */
+	unsigned long stops;        /* # of times q has been stopped */
+	unsigned long restarts;     /* # of queue restarts */
+	unsigned int  cntxt_id;     /* SGE context id for the Tx q */
+	struct tx_desc *desc;       /* address of HW Tx descriptor ring */
+	struct tx_sw_desc *sdesc;   /* address of SW Tx descriptor ring */
+	struct sge_qstat *stat;     /* queue status entry */
+	dma_addr_t    phys_addr;    /* physical address of the ring */
+};
+
+struct sge_eth_txq {                /* state for an SGE Ethernet Tx queue */
+	struct sge_txq q;
+	struct netdev_queue *txq;   /* associated netdev TX queue */
+	unsigned long tso;          /* # of TSO requests */
+	unsigned long tx_cso;       /* # of Tx checksum offloads */
+	unsigned long vlan_ins;     /* # of Tx VLAN insertions */
+	unsigned long mapping_err;  /* # of I/O MMU packet mapping errors */
+} ____cacheline_aligned_in_smp;
+
+struct sge_ofld_txq {               /* state for an SGE offload Tx queue */
+	struct sge_txq q;
+	struct adapter *adap;
+	struct sk_buff_head sendq;  /* list of backpressured packets */
+	struct tasklet_struct qresume_tsk; /* restarts the queue */
+	u8 full;                    /* the Tx ring is full */
+	unsigned long mapping_err;  /* # of I/O MMU packet mapping errors */
+} ____cacheline_aligned_in_smp;
+
+struct sge_ctrl_txq {               /* state for an SGE control Tx queue */
+	struct sge_txq q;
+	struct adapter *adap;
+	struct sk_buff_head sendq;  /* list of backpressured packets */
+	struct tasklet_struct qresume_tsk; /* restarts the queue */
+	u8 full;                    /* the Tx ring is full */
+} ____cacheline_aligned_in_smp;
+
+struct sge {
+	struct sge_eth_txq ethtxq[MAX_ETH_QSETS];
+	struct sge_ofld_txq ofldtxq[MAX_OFLD_QSETS];
+	struct sge_ctrl_txq ctrlq[MAX_CTRL_QUEUES];
+
+	struct sge_eth_rxq ethrxq[MAX_ETH_QSETS];
+	struct sge_ofld_rxq ofldrxq[MAX_OFLD_QSETS];
+	struct sge_ofld_rxq rdmarxq[MAX_RDMA_QUEUES];
+	struct sge_rspq fw_evtq ____cacheline_aligned_in_smp;
+
+	struct sge_rspq intrq ____cacheline_aligned_in_smp;
+	spinlock_t intrq_lock;
+
+	u16 max_ethqsets;           /* # of available Ethernet queue sets */
+	u16 ethqsets;               /* # of active Ethernet queue sets */
+	u16 ethtxq_rover;           /* Tx queue to clean up next */
+	u16 ofldqsets;              /* # of active offload queue sets */
+	u16 rdmaqs;                 /* # of available RDMA Rx queues */
+	u16 ofld_rxq[MAX_OFLD_QSETS];
+	u16 rdma_rxq[NCHAN];
+	u16 timer_val[SGE_NTIMERS];
+	u8 counter_val[SGE_NCOUNTERS];
+	unsigned int starve_thres;
+	u8 idma_state[2];
+	void *egr_map[MAX_EGRQ];    /* qid->queue egress queue map */
+	struct sge_rspq *ingr_map[MAX_INGQ]; /* qid->queue ingress queue map */
+	DECLARE_BITMAP(starving_fl, MAX_EGRQ);
+	DECLARE_BITMAP(txq_maperr, MAX_EGRQ);
+	struct timer_list rx_timer; /* refills starving FLs */
+	struct timer_list tx_timer; /* checks Tx queues */
+};
+
+#define for_each_ethrxq(sge, i) for (i = 0; i < (sge)->ethqsets; i++)
+#define for_each_ofldrxq(sge, i) for (i = 0; i < (sge)->ofldqsets; i++)
+#define for_each_rdmarxq(sge, i) for (i = 0; i < (sge)->rdmaqs; i++)
+
+struct l2t_data;
+
+struct adapter {
+	void __iomem *regs;
+	struct pci_dev *pdev;
+	struct device *pdev_dev;
+	unsigned long registered_device_map;
+	unsigned long open_device_map;
+	unsigned long flags;
+
+	const char *name;
+	int msg_enable;
+
+	struct adapter_params params;
+	struct cxgb4_virt_res vres;
+	unsigned int swintr;
+
+	unsigned int wol;
+
+	struct {
+		unsigned short vec;
+		char desc[14];
+	} msix_info[MAX_INGQ + 1];
+
+	struct sge sge;
+
+	struct net_device *port[MAX_NPORTS];
+	u8 chan_map[NCHAN];                   /* channel -> port map */
+
+	struct l2t_data *l2t;
+	void *uld_handle[CXGB4_ULD_MAX];
+	struct list_head list_node;
+
+	struct tid_info tids;
+	void **tid_release_head;
+	spinlock_t tid_release_lock;
+	struct work_struct tid_release_task;
+	bool tid_release_task_busy;
+
+	struct dentry *debugfs_root;
+
+	spinlock_t stats_lock;
+};
+
+static inline u32 t4_read_reg(struct adapter *adap, u32 reg_addr)
+{
+	return readl(adap->regs + reg_addr);
+}
+
+static inline void t4_write_reg(struct adapter *adap, u32 reg_addr, u32 val)
+{
+	writel(val, adap->regs + reg_addr);
+}
+
+#ifndef readq
+static inline u64 readq(const volatile void __iomem *addr)
+{
+	return readl(addr) + ((u64)readl(addr + 4) << 32);
+}
+
+static inline void writeq(u64 val, volatile void __iomem *addr)
+{
+	writel(val, addr);
+	writel(val >> 32, addr + 4);
+}
+#endif
+
+static inline u64 t4_read_reg64(struct adapter *adap, u32 reg_addr)
+{
+	return readq(adap->regs + reg_addr);
+}
+
+static inline void t4_write_reg64(struct adapter *adap, u32 reg_addr, u64 val)
+{
+	writeq(val, adap->regs + reg_addr);
+}
+
+/**
+ * netdev2pinfo - return the port_info structure associated with a net_device
+ * @dev: the netdev
+ *
+ * Return the struct port_info associated with a net_device
+ */
+static inline struct port_info *netdev2pinfo(const struct net_device *dev)
+{
+	return netdev_priv(dev);
+}
+
+/**
+ * adap2pinfo - return the port_info of a port
+ * @adap: the adapter
+ * @idx: the port index
+ *
+ * Return the port_info structure for the port of the given index.
+ */
+static inline struct port_info *adap2pinfo(struct adapter *adap, int idx)
+{
+	return netdev_priv(adap->port[idx]);
+}
+
+/**
+ * netdev2adap - return the adapter structure associated with a net_device
+ * @dev: the netdev
+ *
+ * Return the struct adapter associated with a net_device
+ */
+static inline struct adapter *netdev2adap(const struct net_device *dev)
+{
+	return netdev2pinfo(dev)->adapter;
+}
+
+void t4_os_portmod_changed(const struct adapter *adap, int port_id);
+void t4_os_link_changed(struct adapter *adap, int port_id, int link_stat);
+
+void *t4_alloc_mem(size_t size);
+void t4_free_mem(void *addr);
+
+void t4_free_sge_resources(struct adapter *adap);
+irq_handler_t t4_intr_handler(struct adapter *adap);
+netdev_tx_t t4_eth_xmit(struct sk_buff *skb, struct net_device *dev);
+int t4_ethrx_handler(struct sge_rspq *q, const __be64 *rsp,
+		     const struct pkt_gl *gl);
+int t4_mgmt_tx(struct adapter *adap, struct sk_buff *skb);
+int t4_ofld_send(struct adapter *adap, struct sk_buff *skb);
+int t4_sge_alloc_rxq(struct adapter *adap, struct sge_rspq *iq, bool fwevtq,
+		     struct net_device *dev, int intr_idx,
+		     struct sge_fl *fl, rspq_handler_t hnd);
+int t4_sge_alloc_eth_txq(struct adapter *adap, struct sge_eth_txq *txq,
+			 struct net_device *dev, struct netdev_queue *netdevq,
+			 unsigned int iqid);
+int t4_sge_alloc_ctrl_txq(struct adapter *adap, struct sge_ctrl_txq *txq,
+			  struct net_device *dev, unsigned int iqid,
+			  unsigned int cmplqid);
+int t4_sge_alloc_ofld_txq(struct adapter *adap, struct sge_ofld_txq *txq,
+			  struct net_device *dev, unsigned int iqid);
+irqreturn_t t4_sge_intr_msix(int irq, void *cookie);
+void t4_sge_init(struct adapter *adap);
+void t4_sge_start(struct adapter *adap);
+void t4_sge_stop(struct adapter *adap);
+
+#define for_each_port(adapter, iter) \
+	for (iter = 0; iter < (adapter)->params.nports; ++iter)
+
+static inline unsigned int core_ticks_per_usec(const struct adapter *adap)
+{
+	return adap->params.vpd.cclk / 1000;
+}
+
+static inline unsigned int us_to_core_ticks(const struct adapter *adap,
+					    unsigned int us)
+{
+	return (us * adap->params.vpd.cclk) / 1000;
+}
+
+void t4_set_reg_field(struct adapter *adap, unsigned int addr, u32 mask,
+		      u32 val);
+
+int t4_wr_mbox_meat(struct adapter *adap, int mbox, const void *cmd, int size,
+		    void *rpl, bool sleep_ok);
+
+static inline int t4_wr_mbox(struct adapter *adap, int mbox, const void *cmd,
+			     int size, void *rpl)
+{
+	return t4_wr_mbox_meat(adap, mbox, cmd, size, rpl, true);
+}
+
+static inline int t4_wr_mbox_ns(struct adapter *adap, int mbox, const void *cmd,
+				int size, void *rpl)
+{
+	return t4_wr_mbox_meat(adap, mbox, cmd, size, rpl, false);
+}
+
+void t4_intr_enable(struct adapter *adapter);
+void t4_intr_disable(struct adapter *adapter);
+void t4_intr_clear(struct adapter *adapter);
+int t4_slow_intr_handler(struct adapter *adapter);
+
+int t4_link_start(struct adapter *adap, unsigned int mbox, unsigned int port,
+		  struct link_config *lc);
+int t4_restart_aneg(struct adapter *adap, unsigned int mbox, unsigned int port);
+int t4_seeprom_wp(struct adapter *adapter, bool enable);
+int t4_read_flash(struct adapter *adapter, unsigned int addr,
+		  unsigned int nwords, u32 *data, int byte_oriented);
+int t4_load_fw(struct adapter *adapter, const u8 *fw_data, unsigned int size);
+int t4_check_fw_version(struct adapter *adapter);
+int t4_prep_adapter(struct adapter *adapter);
+int t4_port_init(struct adapter *adap, int mbox, int pf, int vf);
+void t4_fatal_err(struct adapter *adapter);
+void t4_set_vlan_accel(struct adapter *adapter, unsigned int ports, int on);
+int t4_set_trace_filter(struct adapter *adapter, const struct trace_params *tp,
+			int filter_index, int enable);
+void t4_get_trace_filter(struct adapter *adapter, struct trace_params *tp,
+			 int filter_index, int *enabled);
+int t4_config_rss_range(struct adapter *adapter, int mbox, unsigned int viid,
+			int start, int n, const u16 *rspq, unsigned int nrspq);
+int t4_config_glbl_rss(struct adapter *adapter, int mbox, unsigned int mode,
+		       unsigned int flags);
+int t4_read_rss(struct adapter *adapter, u16 *entries);
+int t4_mc_read(struct adapter *adap, u32 addr, __be32 *data, u64 *parity);
+int t4_edc_read(struct adapter *adap, int idx, u32 addr, __be32 *data,
+		u64 *parity);
+
+void t4_get_port_stats(struct adapter *adap, int idx, struct port_stats *p);
+void t4_get_lb_stats(struct adapter *adap, int idx, struct lb_port_stats *p);
+
+void t4_read_mtu_tbl(struct adapter *adap, u16 *mtus, u8 *mtu_log);
+void t4_tp_get_err_stats(struct adapter *adap, struct tp_err_stats *st);
+void t4_tp_get_tcp_stats(struct adapter *adap, struct tp_tcp_stats *v4,
+			 struct tp_tcp_stats *v6);
+void t4_load_mtus(struct adapter *adap, const unsigned short *mtus,
+		  const unsigned short *alpha, const unsigned short *beta);
+
+void t4_wol_magic_enable(struct adapter *adap, unsigned int port,
+			 const u8 *addr);
+int t4_wol_pat_enable(struct adapter *adap, unsigned int port, unsigned int map,
+		      u64 mask0, u64 mask1, unsigned int crc, bool enable);
+
+int t4_fw_hello(struct adapter *adap, unsigned int mbox, unsigned int evt_mbox,
+		enum dev_master master, enum dev_state *state);
+int t4_fw_bye(struct adapter *adap, unsigned int mbox);
+int t4_early_init(struct adapter *adap, unsigned int mbox);
+int t4_fw_reset(struct adapter *adap, unsigned int mbox, int reset);
+int t4_query_params(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int nparams, const u32 *params,
+		    u32 *val);
+int t4_set_params(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		  unsigned int vf, unsigned int nparams, const u32 *params,
+		  const u32 *val);
+int t4_cfg_pfvf(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		unsigned int vf, unsigned int txq, unsigned int txq_eth_ctrl,
+		unsigned int rxqi, unsigned int rxq, unsigned int tc,
+		unsigned int vi, unsigned int cmask, unsigned int pmask,
+		unsigned int nexact, unsigned int rcaps, unsigned int wxcaps);
+int t4_alloc_vi(struct adapter *adap, unsigned int mbox, unsigned int port,
+		unsigned int pf, unsigned int vf, unsigned int nmac, u8 *mac,
+		unsigned int *rss_size);
+int t4_free_vi(struct adapter *adap, unsigned int mbox, unsigned int pf,
+	       unsigned int vf, unsigned int viid);
+int t4_set_rxmode(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		int mtu, int promisc, int all_multi, int bcast, bool sleep_ok);
+int t4_alloc_mac_filt(struct adapter *adap, unsigned int mbox,
+		      unsigned int viid, bool free, unsigned int naddr,
+		      const u8 **addr, u16 *idx, u64 *hash, bool sleep_ok);
+int t4_change_mac(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		  int idx, const u8 *addr, bool persist, bool add_smt);
+int t4_set_addr_hash(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		     bool ucast, u64 vec, bool sleep_ok);
+int t4_enable_vi(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		 bool rx_en, bool tx_en);
+int t4_identify_port(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		     unsigned int nblinks);
+int t4_mdio_rd(struct adapter *adap, unsigned int mbox, unsigned int phy_addr,
+	       unsigned int mmd, unsigned int reg, u16 *valp);
+int t4_mdio_wr(struct adapter *adap, unsigned int mbox, unsigned int phy_addr,
+	       unsigned int mmd, unsigned int reg, u16 val);
+int t4_iq_start_stop(struct adapter *adap, unsigned int mbox, bool start,
+		     unsigned int pf, unsigned int vf, unsigned int iqid,
+		     unsigned int fl0id, unsigned int fl1id);
+int t4_iq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+	       unsigned int vf, unsigned int iqtype, unsigned int iqid,
+	       unsigned int fl0id, unsigned int fl1id);
+int t4_eth_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		   unsigned int vf, unsigned int eqid);
+int t4_ctrl_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int eqid);
+int t4_ofld_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int eqid);
+int t4_handle_fw_rpl(struct adapter *adap, const __be64 *rpl);
+#endif /* __CXGB4_H__ */
diff --git a/drivers/net/cxgb4/cxgb4_main.c b/drivers/net/cxgb4/cxgb4_main.c
new file mode 100644
index 000000000000..a7e30a23d322
--- /dev/null
+++ b/drivers/net/cxgb4/cxgb4_main.c
@@ -0,0 +1,3388 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/bitmap.h>
+#include <linux/crc32.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <linux/err.h>
+#include <linux/etherdevice.h>
+#include <linux/firmware.h>
+#include <linux/if_vlan.h>
+#include <linux/init.h>
+#include <linux/log2.h>
+#include <linux/mdio.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/mutex.h>
+#include <linux/netdevice.h>
+#include <linux/pci.h>
+#include <linux/aer.h>
+#include <linux/rtnetlink.h>
+#include <linux/sched.h>
+#include <linux/seq_file.h>
+#include <linux/sockios.h>
+#include <linux/vmalloc.h>
+#include <linux/workqueue.h>
+#include <net/neighbour.h>
+#include <net/netevent.h>
+#include <asm/uaccess.h>
+
+#include "cxgb4.h"
+#include "t4_regs.h"
+#include "t4_msg.h"
+#include "t4fw_api.h"
+#include "l2t.h"
+
+#define DRV_VERSION "1.0.0-ko"
+#define DRV_DESC "Chelsio T4 Network Driver"
+
+/*
+ * Max interrupt hold-off timer value in us.  Queues fall back to this value
+ * under extreme memory pressure so it's largish to give the system time to
+ * recover.
+ */
+#define MAX_SGE_TIMERVAL 200U
+
+enum {
+	MEMWIN0_APERTURE = 65536,
+	MEMWIN0_BASE     = 0x30000,
+	MEMWIN1_APERTURE = 32768,
+	MEMWIN1_BASE     = 0x28000,
+	MEMWIN2_APERTURE = 2048,
+	MEMWIN2_BASE     = 0x1b800,
+};
+
+enum {
+	MAX_TXQ_ENTRIES      = 16384,
+	MAX_CTRL_TXQ_ENTRIES = 1024,
+	MAX_RSPQ_ENTRIES     = 16384,
+	MAX_RX_BUFFERS       = 16384,
+	MIN_TXQ_ENTRIES      = 32,
+	MIN_CTRL_TXQ_ENTRIES = 32,
+	MIN_RSPQ_ENTRIES     = 128,
+	MIN_FL_ENTRIES       = 16
+};
+
+#define DFLT_MSG_ENABLE (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK | \
+			 NETIF_MSG_TIMER | NETIF_MSG_IFDOWN | NETIF_MSG_IFUP |\
+			 NETIF_MSG_RX_ERR | NETIF_MSG_TX_ERR)
+
+#define CH_DEVICE(devid) { PCI_VDEVICE(CHELSIO, devid), 0 }
+
+static DEFINE_PCI_DEVICE_TABLE(cxgb4_pci_tbl) = {
+	CH_DEVICE(0xa000),  /* PE10K */
+	{ 0, }
+};
+
+#define FW_FNAME "cxgb4/t4fw.bin"
+
+MODULE_DESCRIPTION(DRV_DESC);
+MODULE_AUTHOR("Chelsio Communications");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRV_VERSION);
+MODULE_DEVICE_TABLE(pci, cxgb4_pci_tbl);
+MODULE_FIRMWARE(FW_FNAME);
+
+static int dflt_msg_enable = DFLT_MSG_ENABLE;
+
+module_param(dflt_msg_enable, int, 0644);
+MODULE_PARM_DESC(dflt_msg_enable, "Chelsio T4 default message enable bitmap");
+
+/*
+ * The driver uses the best interrupt scheme available on a platform in the
+ * order MSI-X, MSI, legacy INTx interrupts.  This parameter determines which
+ * of these schemes the driver may consider as follows:
+ *
+ * msi = 2: choose from among all three options
+ * msi = 1: only consider MSI and INTx interrupts
+ * msi = 0: force INTx interrupts
+ */
+static int msi = 2;
+
+module_param(msi, int, 0644);
+MODULE_PARM_DESC(msi, "whether to use INTx (0), MSI (1) or MSI-X (2)");
+
+/*
+ * Queue interrupt hold-off timer values.  Queues default to the first of these
+ * upon creation.
+ */
+static unsigned int intr_holdoff[SGE_NTIMERS - 1] = { 5, 10, 20, 50, 100 };
+
+module_param_array(intr_holdoff, uint, NULL, 0644);
+MODULE_PARM_DESC(intr_holdoff, "values for queue interrupt hold-off timers "
+		 "0..4 in microseconds");
+
+static unsigned int intr_cnt[SGE_NCOUNTERS - 1] = { 4, 8, 16 };
+
+module_param_array(intr_cnt, uint, NULL, 0644);
+MODULE_PARM_DESC(intr_cnt,
+		 "thresholds 1..3 for queue interrupt packet counters");
+
+static int vf_acls;
+
+#ifdef CONFIG_PCI_IOV
+module_param(vf_acls, bool, 0644);
+MODULE_PARM_DESC(vf_acls, "if set enable virtualization L2 ACL enforcement");
+
+static unsigned int num_vf[4];
+
+module_param_array(num_vf, uint, NULL, 0644);
+MODULE_PARM_DESC(num_vf, "number of VFs for each of PFs 0-3");
+#endif
+
+static struct dentry *cxgb4_debugfs_root;
+
+static LIST_HEAD(adapter_list);
+static DEFINE_MUTEX(uld_mutex);
+static struct cxgb4_uld_info ulds[CXGB4_ULD_MAX];
+static const char *uld_str[] = { "RDMA", "iSCSI" };
+
+static void link_report(struct net_device *dev)
+{
+	if (!netif_carrier_ok(dev))
+		netdev_info(dev, "link down\n");
+	else {
+		static const char *fc[] = { "no", "Rx", "Tx", "Tx/Rx" };
+
+		const char *s = "10Mbps";
+		const struct port_info *p = netdev_priv(dev);
+
+		switch (p->link_cfg.speed) {
+		case SPEED_10000:
+			s = "10Gbps";
+			break;
+		case SPEED_1000:
+			s = "1000Mbps";
+			break;
+		case SPEED_100:
+			s = "100Mbps";
+			break;
+		}
+
+		netdev_info(dev, "link up, %s, full-duplex, %s PAUSE\n", s,
+			    fc[p->link_cfg.fc]);
+	}
+}
+
+void t4_os_link_changed(struct adapter *adapter, int port_id, int link_stat)
+{
+	struct net_device *dev = adapter->port[port_id];
+
+	/* Skip changes from disabled ports. */
+	if (netif_running(dev) && link_stat != netif_carrier_ok(dev)) {
+		if (link_stat)
+			netif_carrier_on(dev);
+		else
+			netif_carrier_off(dev);
+
+		link_report(dev);
+	}
+}
+
+void t4_os_portmod_changed(const struct adapter *adap, int port_id)
+{
+	static const char *mod_str[] = {
+		NULL, "LR", "SR", "ER", "passive DA", "active DA"
+	};
+
+	const struct net_device *dev = adap->port[port_id];
+	const struct port_info *pi = netdev_priv(dev);
+
+	if (pi->mod_type == FW_PORT_MOD_TYPE_NONE)
+		netdev_info(dev, "port module unplugged\n");
+	else
+		netdev_info(dev, "%s module inserted\n", mod_str[pi->mod_type]);
+}
+
+/*
+ * Configure the exact and hash address filters to handle a port's multicast
+ * and secondary unicast MAC addresses.
+ */
+static int set_addr_filters(const struct net_device *dev, bool sleep)
+{
+	u64 mhash = 0;
+	u64 uhash = 0;
+	bool free = true;
+	u16 filt_idx[7];
+	const u8 *addr[7];
+	int ret, naddr = 0;
+	const struct dev_addr_list *d;
+	const struct netdev_hw_addr *ha;
+	int uc_cnt = netdev_uc_count(dev);
+	const struct port_info *pi = netdev_priv(dev);
+
+	/* first do the secondary unicast addresses */
+	netdev_for_each_uc_addr(ha, dev) {
+		addr[naddr++] = ha->addr;
+		if (--uc_cnt == 0 || naddr >= ARRAY_SIZE(addr)) {
+			ret = t4_alloc_mac_filt(pi->adapter, 0, pi->viid, free,
+					naddr, addr, filt_idx, &uhash, sleep);
+			if (ret < 0)
+				return ret;
+
+			free = false;
+			naddr = 0;
+		}
+	}
+
+	/* next set up the multicast addresses */
+	netdev_for_each_mc_addr(d, dev) {
+		addr[naddr++] = d->dmi_addr;
+		if (naddr >= ARRAY_SIZE(addr) || d->next == NULL) {
+			ret = t4_alloc_mac_filt(pi->adapter, 0, pi->viid, free,
+					naddr, addr, filt_idx, &mhash, sleep);
+			if (ret < 0)
+				return ret;
+
+			free = false;
+			naddr = 0;
+		}
+	}
+
+	return t4_set_addr_hash(pi->adapter, 0, pi->viid, uhash != 0,
+				uhash | mhash, sleep);
+}
+
+/*
+ * Set Rx properties of a port, such as promiscruity, address filters, and MTU.
+ * If @mtu is -1 it is left unchanged.
+ */
+static int set_rxmode(struct net_device *dev, int mtu, bool sleep_ok)
+{
+	int ret;
+	struct port_info *pi = netdev_priv(dev);
+
+	ret = set_addr_filters(dev, sleep_ok);
+	if (ret == 0)
+		ret = t4_set_rxmode(pi->adapter, 0, pi->viid, mtu,
+				    (dev->flags & IFF_PROMISC) ? 1 : 0,
+				    (dev->flags & IFF_ALLMULTI) ? 1 : 0, 1,
+				    sleep_ok);
+	return ret;
+}
+
+/**
+ *	link_start - enable a port
+ *	@dev: the port to enable
+ *
+ *	Performs the MAC and PHY actions needed to enable a port.
+ */
+static int link_start(struct net_device *dev)
+{
+	int ret;
+	struct port_info *pi = netdev_priv(dev);
+
+	/*
+	 * We do not set address filters and promiscuity here, the stack does
+	 * that step explicitly.
+	 */
+	ret = t4_set_rxmode(pi->adapter, 0, pi->viid, dev->mtu, -1, -1, -1,
+			    true);
+	if (ret == 0) {
+		ret = t4_change_mac(pi->adapter, 0, pi->viid,
+				    pi->xact_addr_filt, dev->dev_addr, true,
+				    false);
+		if (ret >= 0) {
+			pi->xact_addr_filt = ret;
+			ret = 0;
+		}
+	}
+	if (ret == 0)
+		ret = t4_link_start(pi->adapter, 0, pi->tx_chan, &pi->link_cfg);
+	if (ret == 0)
+		ret = t4_enable_vi(pi->adapter, 0, pi->viid, true, true);
+	return ret;
+}
+
+/*
+ * Response queue handler for the FW event queue.
+ */
+static int fwevtq_handler(struct sge_rspq *q, const __be64 *rsp,
+			  const struct pkt_gl *gl)
+{
+	u8 opcode = ((const struct rss_header *)rsp)->opcode;
+
+	rsp++;                                          /* skip RSS header */
+	if (likely(opcode == CPL_SGE_EGR_UPDATE)) {
+		const struct cpl_sge_egr_update *p = (void *)rsp;
+		unsigned int qid = EGR_QID(ntohl(p->opcode_qid));
+		struct sge_txq *txq = q->adap->sge.egr_map[qid];
+
+		txq->restarts++;
+		if ((u8 *)txq < (u8 *)q->adap->sge.ethrxq) {
+			struct sge_eth_txq *eq;
+
+			eq = container_of(txq, struct sge_eth_txq, q);
+			netif_tx_wake_queue(eq->txq);
+		} else {
+			struct sge_ofld_txq *oq;
+
+			oq = container_of(txq, struct sge_ofld_txq, q);
+			tasklet_schedule(&oq->qresume_tsk);
+		}
+	} else if (opcode == CPL_FW6_MSG || opcode == CPL_FW4_MSG) {
+		const struct cpl_fw6_msg *p = (void *)rsp;
+
+		if (p->type == 0)
+			t4_handle_fw_rpl(q->adap, p->data);
+	} else if (opcode == CPL_L2T_WRITE_RPL) {
+		const struct cpl_l2t_write_rpl *p = (void *)rsp;
+
+		do_l2t_write_rpl(q->adap, p);
+	} else
+		dev_err(q->adap->pdev_dev,
+			"unexpected CPL %#x on FW event queue\n", opcode);
+	return 0;
+}
+
+/**
+ *	uldrx_handler - response queue handler for ULD queues
+ *	@q: the response queue that received the packet
+ *	@rsp: the response queue descriptor holding the offload message
+ *	@gl: the gather list of packet fragments
+ *
+ *	Deliver an ingress offload packet to a ULD.  All processing is done by
+ *	the ULD, we just maintain statistics.
+ */
+static int uldrx_handler(struct sge_rspq *q, const __be64 *rsp,
+			 const struct pkt_gl *gl)
+{
+	struct sge_ofld_rxq *rxq = container_of(q, struct sge_ofld_rxq, rspq);
+
+	if (ulds[q->uld].rx_handler(q->adap->uld_handle[q->uld], rsp, gl)) {
+		rxq->stats.nomem++;
+		return -1;
+	}
+	if (gl == NULL)
+		rxq->stats.imm++;
+	else if (gl == CXGB4_MSG_AN)
+		rxq->stats.an++;
+	else
+		rxq->stats.pkts++;
+	return 0;
+}
+
+static void disable_msi(struct adapter *adapter)
+{
+	if (adapter->flags & USING_MSIX) {
+		pci_disable_msix(adapter->pdev);
+		adapter->flags &= ~USING_MSIX;
+	} else if (adapter->flags & USING_MSI) {
+		pci_disable_msi(adapter->pdev);
+		adapter->flags &= ~USING_MSI;
+	}
+}
+
+/*
+ * Interrupt handler for non-data events used with MSI-X.
+ */
+static irqreturn_t t4_nondata_intr(int irq, void *cookie)
+{
+	struct adapter *adap = cookie;
+
+	u32 v = t4_read_reg(adap, MYPF_REG(PL_PF_INT_CAUSE));
+	if (v & PFSW) {
+		adap->swintr = 1;
+		t4_write_reg(adap, MYPF_REG(PL_PF_INT_CAUSE), v);
+	}
+	t4_slow_intr_handler(adap);
+	return IRQ_HANDLED;
+}
+
+/*
+ * Name the MSI-X interrupts.
+ */
+static void name_msix_vecs(struct adapter *adap)
+{
+	int i, j, msi_idx = 2, n = sizeof(adap->msix_info[0].desc) - 1;
+
+	/* non-data interrupts */
+	snprintf(adap->msix_info[0].desc, n, "%s", adap->name);
+	adap->msix_info[0].desc[n] = 0;
+
+	/* FW events */
+	snprintf(adap->msix_info[1].desc, n, "%s-FWeventq", adap->name);
+	adap->msix_info[1].desc[n] = 0;
+
+	/* Ethernet queues */
+	for_each_port(adap, j) {
+		struct net_device *d = adap->port[j];
+		const struct port_info *pi = netdev_priv(d);
+
+		for (i = 0; i < pi->nqsets; i++, msi_idx++) {
+			snprintf(adap->msix_info[msi_idx].desc, n, "%s-Rx%d",
+				 d->name, i);
+			adap->msix_info[msi_idx].desc[n] = 0;
+		}
+	}
+
+	/* offload queues */
+	for_each_ofldrxq(&adap->sge, i) {
+		snprintf(adap->msix_info[msi_idx].desc, n, "%s-ofld%d",
+			 adap->name, i);
+		adap->msix_info[msi_idx++].desc[n] = 0;
+	}
+	for_each_rdmarxq(&adap->sge, i) {
+		snprintf(adap->msix_info[msi_idx].desc, n, "%s-rdma%d",
+			 adap->name, i);
+		adap->msix_info[msi_idx++].desc[n] = 0;
+	}
+}
+
+static int request_msix_queue_irqs(struct adapter *adap)
+{
+	struct sge *s = &adap->sge;
+	int err, ethqidx, ofldqidx = 0, rdmaqidx = 0, msi = 2;
+
+	err = request_irq(adap->msix_info[1].vec, t4_sge_intr_msix, 0,
+			  adap->msix_info[1].desc, &s->fw_evtq);
+	if (err)
+		return err;
+
+	for_each_ethrxq(s, ethqidx) {
+		err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+				  adap->msix_info[msi].desc,
+				  &s->ethrxq[ethqidx].rspq);
+		if (err)
+			goto unwind;
+		msi++;
+	}
+	for_each_ofldrxq(s, ofldqidx) {
+		err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+				  adap->msix_info[msi].desc,
+				  &s->ofldrxq[ofldqidx].rspq);
+		if (err)
+			goto unwind;
+		msi++;
+	}
+	for_each_rdmarxq(s, rdmaqidx) {
+		err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+				  adap->msix_info[msi].desc,
+				  &s->rdmarxq[rdmaqidx].rspq);
+		if (err)
+			goto unwind;
+		msi++;
+	}
+	return 0;
+
+unwind:
+	while (--rdmaqidx >= 0)
+		free_irq(adap->msix_info[--msi].vec,
+			 &s->rdmarxq[rdmaqidx].rspq);
+	while (--ofldqidx >= 0)
+		free_irq(adap->msix_info[--msi].vec,
+			 &s->ofldrxq[ofldqidx].rspq);
+	while (--ethqidx >= 0)
+		free_irq(adap->msix_info[--msi].vec, &s->ethrxq[ethqidx].rspq);
+	free_irq(adap->msix_info[1].vec, &s->fw_evtq);
+	return err;
+}
+
+static void free_msix_queue_irqs(struct adapter *adap)
+{
+	int i, msi = 2;
+	struct sge *s = &adap->sge;
+
+	free_irq(adap->msix_info[1].vec, &s->fw_evtq);
+	for_each_ethrxq(s, i)
+		free_irq(adap->msix_info[msi++].vec, &s->ethrxq[i].rspq);
+	for_each_ofldrxq(s, i)
+		free_irq(adap->msix_info[msi++].vec, &s->ofldrxq[i].rspq);
+	for_each_rdmarxq(s, i)
+		free_irq(adap->msix_info[msi++].vec, &s->rdmarxq[i].rspq);
+}
+
+/**
+ *	setup_rss - configure RSS
+ *	@adap: the adapter
+ *
+ *	Sets up RSS to distribute packets to multiple receive queues.  We
+ *	configure the RSS CPU lookup table to distribute to the number of HW
+ *	receive queues, and the response queue lookup table to narrow that
+ *	down to the response queues actually configured for each port.
+ *	We always configure the RSS mapping for all ports since the mapping
+ *	table has plenty of entries.
+ */
+static int setup_rss(struct adapter *adap)
+{
+	int i, j, err;
+	u16 rss[MAX_ETH_QSETS];
+
+	for_each_port(adap, i) {
+		const struct port_info *pi = adap2pinfo(adap, i);
+		const struct sge_eth_rxq *q = &adap->sge.ethrxq[pi->first_qset];
+
+		for (j = 0; j < pi->nqsets; j++)
+			rss[j] = q[j].rspq.abs_id;
+
+		err = t4_config_rss_range(adap, 0, pi->viid, 0, pi->rss_size,
+					  rss, pi->nqsets);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+/*
+ * Wait until all NAPI handlers are descheduled.
+ */
+static void quiesce_rx(struct adapter *adap)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(adap->sge.ingr_map); i++) {
+		struct sge_rspq *q = adap->sge.ingr_map[i];
+
+		if (q && q->handler)
+			napi_disable(&q->napi);
+	}
+}
+
+/*
+ * Enable NAPI scheduling and interrupt generation for all Rx queues.
+ */
+static void enable_rx(struct adapter *adap)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(adap->sge.ingr_map); i++) {
+		struct sge_rspq *q = adap->sge.ingr_map[i];
+
+		if (!q)
+			continue;
+		if (q->handler)
+			napi_enable(&q->napi);
+		/* 0-increment GTS to start the timer and enable interrupts */
+		t4_write_reg(adap, MYPF_REG(SGE_PF_GTS),
+			     SEINTARM(q->intr_params) |
+			     INGRESSQID(q->cntxt_id));
+	}
+}
+
+/**
+ *	setup_sge_queues - configure SGE Tx/Rx/response queues
+ *	@adap: the adapter
+ *
+ *	Determines how many sets of SGE queues to use and initializes them.
+ *	We support multiple queue sets per port if we have MSI-X, otherwise
+ *	just one queue set per port.
+ */
+static int setup_sge_queues(struct adapter *adap)
+{
+	int err, msi_idx, i, j;
+	struct sge *s = &adap->sge;
+
+	bitmap_zero(s->starving_fl, MAX_EGRQ);
+	bitmap_zero(s->txq_maperr, MAX_EGRQ);
+
+	if (adap->flags & USING_MSIX)
+		msi_idx = 1;         /* vector 0 is for non-queue interrupts */
+	else {
+		err = t4_sge_alloc_rxq(adap, &s->intrq, false, adap->port[0], 0,
+				       NULL, NULL);
+		if (err)
+			return err;
+		msi_idx = -((int)s->intrq.abs_id + 1);
+	}
+
+	err = t4_sge_alloc_rxq(adap, &s->fw_evtq, true, adap->port[0],
+			       msi_idx, NULL, fwevtq_handler);
+	if (err) {
+freeout:	t4_free_sge_resources(adap);
+		return err;
+	}
+
+	for_each_port(adap, i) {
+		struct net_device *dev = adap->port[i];
+		struct port_info *pi = netdev_priv(dev);
+		struct sge_eth_rxq *q = &s->ethrxq[pi->first_qset];
+		struct sge_eth_txq *t = &s->ethtxq[pi->first_qset];
+
+		for (j = 0; j < pi->nqsets; j++, q++) {
+			if (msi_idx > 0)
+				msi_idx++;
+			err = t4_sge_alloc_rxq(adap, &q->rspq, false, dev,
+					       msi_idx, &q->fl,
+					       t4_ethrx_handler);
+			if (err)
+				goto freeout;
+			q->rspq.idx = j;
+			memset(&q->stats, 0, sizeof(q->stats));
+		}
+		for (j = 0; j < pi->nqsets; j++, t++) {
+			err = t4_sge_alloc_eth_txq(adap, t, dev,
+					netdev_get_tx_queue(dev, j),
+					s->fw_evtq.cntxt_id);
+			if (err)
+				goto freeout;
+		}
+	}
+
+	j = s->ofldqsets / adap->params.nports; /* ofld queues per channel */
+	for_each_ofldrxq(s, i) {
+		struct sge_ofld_rxq *q = &s->ofldrxq[i];
+		struct net_device *dev = adap->port[i / j];
+
+		if (msi_idx > 0)
+			msi_idx++;
+		err = t4_sge_alloc_rxq(adap, &q->rspq, false, dev, msi_idx,
+				       &q->fl, uldrx_handler);
+		if (err)
+			goto freeout;
+		memset(&q->stats, 0, sizeof(q->stats));
+		s->ofld_rxq[i] = q->rspq.abs_id;
+		err = t4_sge_alloc_ofld_txq(adap, &s->ofldtxq[i], dev,
+					    s->fw_evtq.cntxt_id);
+		if (err)
+			goto freeout;
+	}
+
+	for_each_rdmarxq(s, i) {
+		struct sge_ofld_rxq *q = &s->rdmarxq[i];
+
+		if (msi_idx > 0)
+			msi_idx++;
+		err = t4_sge_alloc_rxq(adap, &q->rspq, false, adap->port[i],
+				       msi_idx, &q->fl, uldrx_handler);
+		if (err)
+			goto freeout;
+		memset(&q->stats, 0, sizeof(q->stats));
+		s->rdma_rxq[i] = q->rspq.abs_id;
+	}
+
+	for_each_port(adap, i) {
+		/*
+		 * Note that ->rdmarxq[i].rspq.cntxt_id below is 0 if we don't
+		 * have RDMA queues, and that's the right value.
+		 */
+		err = t4_sge_alloc_ctrl_txq(adap, &s->ctrlq[i], adap->port[i],
+					    s->fw_evtq.cntxt_id,
+					    s->rdmarxq[i].rspq.cntxt_id);
+		if (err)
+			goto freeout;
+	}
+
+	t4_write_reg(adap, MPS_TRC_RSS_CONTROL,
+		     RSSCONTROL(netdev2pinfo(adap->port[0])->tx_chan) |
+		     QUEUENUMBER(s->ethrxq[0].rspq.abs_id));
+	return 0;
+}
+
+/*
+ * Returns 0 if new FW was successfully loaded, a positive errno if a load was
+ * started but failed, and a negative errno if flash load couldn't start.
+ */
+static int upgrade_fw(struct adapter *adap)
+{
+	int ret;
+	u32 vers;
+	const struct fw_hdr *hdr;
+	const struct firmware *fw;
+	struct device *dev = adap->pdev_dev;
+
+	ret = request_firmware(&fw, FW_FNAME, dev);
+	if (ret < 0) {
+		dev_err(dev, "unable to load firmware image " FW_FNAME
+			", error %d\n", ret);
+		return ret;
+	}
+
+	hdr = (const struct fw_hdr *)fw->data;
+	vers = ntohl(hdr->fw_ver);
+	if (FW_HDR_FW_VER_MAJOR_GET(vers) != FW_VERSION_MAJOR) {
+		ret = -EINVAL;              /* wrong major version, won't do */
+		goto out;
+	}
+
+	/*
+	 * If the flash FW is unusable or we found something newer, load it.
+	 */
+	if (FW_HDR_FW_VER_MAJOR_GET(adap->params.fw_vers) != FW_VERSION_MAJOR ||
+	    vers > adap->params.fw_vers) {
+		ret = -t4_load_fw(adap, fw->data, fw->size);
+		if (!ret)
+			dev_info(dev, "firmware upgraded to version %pI4 from "
+				 FW_FNAME "\n", &hdr->fw_ver);
+	}
+out:	release_firmware(fw);
+	return ret;
+}
+
+/*
+ * Allocate a chunk of memory using kmalloc or, if that fails, vmalloc.
+ * The allocated memory is cleared.
+ */
+void *t4_alloc_mem(size_t size)
+{
+	void *p = kmalloc(size, GFP_KERNEL);
+
+	if (!p)
+		p = vmalloc(size);
+	if (p)
+		memset(p, 0, size);
+	return p;
+}
+
+/*
+ * Free memory allocated through alloc_mem().
+ */
+void t4_free_mem(void *addr)
+{
+	if (is_vmalloc_addr(addr))
+		vfree(addr);
+	else
+		kfree(addr);
+}
+
+static inline int is_offload(const struct adapter *adap)
+{
+	return adap->params.offload;
+}
+
+/*
+ * Implementation of ethtool operations.
+ */
+
+static u32 get_msglevel(struct net_device *dev)
+{
+	return netdev2adap(dev)->msg_enable;
+}
+
+static void set_msglevel(struct net_device *dev, u32 val)
+{
+	netdev2adap(dev)->msg_enable = val;
+}
+
+static char stats_strings[][ETH_GSTRING_LEN] = {
+	"TxOctetsOK         ",
+	"TxFramesOK         ",
+	"TxBroadcastFrames  ",
+	"TxMulticastFrames  ",
+	"TxUnicastFrames    ",
+	"TxErrorFrames      ",
+
+	"TxFrames64         ",
+	"TxFrames65To127    ",
+	"TxFrames128To255   ",
+	"TxFrames256To511   ",
+	"TxFrames512To1023  ",
+	"TxFrames1024To1518 ",
+	"TxFrames1519ToMax  ",
+
+	"TxFramesDropped    ",
+	"TxPauseFrames      ",
+	"TxPPP0Frames       ",
+	"TxPPP1Frames       ",
+	"TxPPP2Frames       ",
+	"TxPPP3Frames       ",
+	"TxPPP4Frames       ",
+	"TxPPP5Frames       ",
+	"TxPPP6Frames       ",
+	"TxPPP7Frames       ",
+
+	"RxOctetsOK         ",
+	"RxFramesOK         ",
+	"RxBroadcastFrames  ",
+	"RxMulticastFrames  ",
+	"RxUnicastFrames    ",
+
+	"RxFramesTooLong    ",
+	"RxJabberErrors     ",
+	"RxFCSErrors        ",
+	"RxLengthErrors     ",
+	"RxSymbolErrors     ",
+	"RxRuntFrames       ",
+
+	"RxFrames64         ",
+	"RxFrames65To127    ",
+	"RxFrames128To255   ",
+	"RxFrames256To511   ",
+	"RxFrames512To1023  ",
+	"RxFrames1024To1518 ",
+	"RxFrames1519ToMax  ",
+
+	"RxPauseFrames      ",
+	"RxPPP0Frames       ",
+	"RxPPP1Frames       ",
+	"RxPPP2Frames       ",
+	"RxPPP3Frames       ",
+	"RxPPP4Frames       ",
+	"RxPPP5Frames       ",
+	"RxPPP6Frames       ",
+	"RxPPP7Frames       ",
+
+	"RxBG0FramesDropped ",
+	"RxBG1FramesDropped ",
+	"RxBG2FramesDropped ",
+	"RxBG3FramesDropped ",
+	"RxBG0FramesTrunc   ",
+	"RxBG1FramesTrunc   ",
+	"RxBG2FramesTrunc   ",
+	"RxBG3FramesTrunc   ",
+
+	"TSO                ",
+	"TxCsumOffload      ",
+	"RxCsumGood         ",
+	"VLANextractions    ",
+	"VLANinsertions     ",
+};
+
+static int get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return ARRAY_SIZE(stats_strings);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+#define T4_REGMAP_SIZE (160 * 1024)
+
+static int get_regs_len(struct net_device *dev)
+{
+	return T4_REGMAP_SIZE;
+}
+
+static int get_eeprom_len(struct net_device *dev)
+{
+	return EEPROMSIZE;
+}
+
+static void get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
+{
+	struct adapter *adapter = netdev2adap(dev);
+
+	strcpy(info->driver, KBUILD_MODNAME);
+	strcpy(info->version, DRV_VERSION);
+	strcpy(info->bus_info, pci_name(adapter->pdev));
+
+	if (!adapter->params.fw_vers)
+		strcpy(info->fw_version, "N/A");
+	else
+		snprintf(info->fw_version, sizeof(info->fw_version),
+			"%u.%u.%u.%u, TP %u.%u.%u.%u",
+			FW_HDR_FW_VER_MAJOR_GET(adapter->params.fw_vers),
+			FW_HDR_FW_VER_MINOR_GET(adapter->params.fw_vers),
+			FW_HDR_FW_VER_MICRO_GET(adapter->params.fw_vers),
+			FW_HDR_FW_VER_BUILD_GET(adapter->params.fw_vers),
+			FW_HDR_FW_VER_MAJOR_GET(adapter->params.tp_vers),
+			FW_HDR_FW_VER_MINOR_GET(adapter->params.tp_vers),
+			FW_HDR_FW_VER_MICRO_GET(adapter->params.tp_vers),
+			FW_HDR_FW_VER_BUILD_GET(adapter->params.tp_vers));
+}
+
+static void get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+	if (stringset == ETH_SS_STATS)
+		memcpy(data, stats_strings, sizeof(stats_strings));
+}
+
+/*
+ * port stats maintained per queue of the port.  They should be in the same
+ * order as in stats_strings above.
+ */
+struct queue_port_stats {
+	u64 tso;
+	u64 tx_csum;
+	u64 rx_csum;
+	u64 vlan_ex;
+	u64 vlan_ins;
+};
+
+static void collect_sge_port_stats(const struct adapter *adap,
+		const struct port_info *p, struct queue_port_stats *s)
+{
+	int i;
+	const struct sge_eth_txq *tx = &adap->sge.ethtxq[p->first_qset];
+	const struct sge_eth_rxq *rx = &adap->sge.ethrxq[p->first_qset];
+
+	memset(s, 0, sizeof(*s));
+	for (i = 0; i < p->nqsets; i++, rx++, tx++) {
+		s->tso += tx->tso;
+		s->tx_csum += tx->tx_cso;
+		s->rx_csum += rx->stats.rx_cso;
+		s->vlan_ex += rx->stats.vlan_ex;
+		s->vlan_ins += tx->vlan_ins;
+	}
+}
+
+static void get_stats(struct net_device *dev, struct ethtool_stats *stats,
+		      u64 *data)
+{
+	struct port_info *pi = netdev_priv(dev);
+	struct adapter *adapter = pi->adapter;
+
+	t4_get_port_stats(adapter, pi->tx_chan, (struct port_stats *)data);
+
+	data += sizeof(struct port_stats) / sizeof(u64);
+	collect_sge_port_stats(adapter, pi, (struct queue_port_stats *)data);
+}
+
+/*
+ * Return a version number to identify the type of adapter.  The scheme is:
+ * - bits 0..9: chip version
+ * - bits 10..15: chip revision
+ */
+static inline unsigned int mk_adap_vers(const struct adapter *ap)
+{
+	return 4 | (ap->params.rev << 10);
+}
+
+static void reg_block_dump(struct adapter *ap, void *buf, unsigned int start,
+			   unsigned int end)
+{
+	u32 *p = buf + start;
+
+	for ( ; start <= end; start += sizeof(u32))
+		*p++ = t4_read_reg(ap, start);
+}
+
+static void get_regs(struct net_device *dev, struct ethtool_regs *regs,
+		     void *buf)
+{
+	static const unsigned int reg_ranges[] = {
+		0x1008, 0x1108,
+		0x1180, 0x11b4,
+		0x11fc, 0x123c,
+		0x1300, 0x173c,
+		0x1800, 0x18fc,
+		0x3000, 0x30d8,
+		0x30e0, 0x5924,
+		0x5960, 0x59d4,
+		0x5a00, 0x5af8,
+		0x6000, 0x6098,
+		0x6100, 0x6150,
+		0x6200, 0x6208,
+		0x6240, 0x6248,
+		0x6280, 0x6338,
+		0x6370, 0x638c,
+		0x6400, 0x643c,
+		0x6500, 0x6524,
+		0x6a00, 0x6a38,
+		0x6a60, 0x6a78,
+		0x6b00, 0x6b84,
+		0x6bf0, 0x6c84,
+		0x6cf0, 0x6d84,
+		0x6df0, 0x6e84,
+		0x6ef0, 0x6f84,
+		0x6ff0, 0x7084,
+		0x70f0, 0x7184,
+		0x71f0, 0x7284,
+		0x72f0, 0x7384,
+		0x73f0, 0x7450,
+		0x7500, 0x7530,
+		0x7600, 0x761c,
+		0x7680, 0x76cc,
+		0x7700, 0x7798,
+		0x77c0, 0x77fc,
+		0x7900, 0x79fc,
+		0x7b00, 0x7c38,
+		0x7d00, 0x7efc,
+		0x8dc0, 0x8e1c,
+		0x8e30, 0x8e78,
+		0x8ea0, 0x8f6c,
+		0x8fc0, 0x9074,
+		0x90fc, 0x90fc,
+		0x9400, 0x9458,
+		0x9600, 0x96bc,
+		0x9800, 0x9808,
+		0x9820, 0x983c,
+		0x9850, 0x9864,
+		0x9c00, 0x9c6c,
+		0x9c80, 0x9cec,
+		0x9d00, 0x9d6c,
+		0x9d80, 0x9dec,
+		0x9e00, 0x9e6c,
+		0x9e80, 0x9eec,
+		0x9f00, 0x9f6c,
+		0x9f80, 0x9fec,
+		0xd004, 0xd03c,
+		0xdfc0, 0xdfe0,
+		0xe000, 0xea7c,
+		0xf000, 0x11190,
+		0x19040, 0x19124,
+		0x19150, 0x191b0,
+		0x191d0, 0x191e8,
+		0x19238, 0x1924c,
+		0x193f8, 0x19474,
+		0x19490, 0x194f8,
+		0x19800, 0x19f30,
+		0x1a000, 0x1a06c,
+		0x1a0b0, 0x1a120,
+		0x1a128, 0x1a138,
+		0x1a190, 0x1a1c4,
+		0x1a1fc, 0x1a1fc,
+		0x1e040, 0x1e04c,
+		0x1e240, 0x1e28c,
+		0x1e2c0, 0x1e2c0,
+		0x1e2e0, 0x1e2e0,
+		0x1e300, 0x1e384,
+		0x1e3c0, 0x1e3c8,
+		0x1e440, 0x1e44c,
+		0x1e640, 0x1e68c,
+		0x1e6c0, 0x1e6c0,
+		0x1e6e0, 0x1e6e0,
+		0x1e700, 0x1e784,
+		0x1e7c0, 0x1e7c8,
+		0x1e840, 0x1e84c,
+		0x1ea40, 0x1ea8c,
+		0x1eac0, 0x1eac0,
+		0x1eae0, 0x1eae0,
+		0x1eb00, 0x1eb84,
+		0x1ebc0, 0x1ebc8,
+		0x1ec40, 0x1ec4c,
+		0x1ee40, 0x1ee8c,
+		0x1eec0, 0x1eec0,
+		0x1eee0, 0x1eee0,
+		0x1ef00, 0x1ef84,
+		0x1efc0, 0x1efc8,
+		0x1f040, 0x1f04c,
+		0x1f240, 0x1f28c,
+		0x1f2c0, 0x1f2c0,
+		0x1f2e0, 0x1f2e0,
+		0x1f300, 0x1f384,
+		0x1f3c0, 0x1f3c8,
+		0x1f440, 0x1f44c,
+		0x1f640, 0x1f68c,
+		0x1f6c0, 0x1f6c0,
+		0x1f6e0, 0x1f6e0,
+		0x1f700, 0x1f784,
+		0x1f7c0, 0x1f7c8,
+		0x1f840, 0x1f84c,
+		0x1fa40, 0x1fa8c,
+		0x1fac0, 0x1fac0,
+		0x1fae0, 0x1fae0,
+		0x1fb00, 0x1fb84,
+		0x1fbc0, 0x1fbc8,
+		0x1fc40, 0x1fc4c,
+		0x1fe40, 0x1fe8c,
+		0x1fec0, 0x1fec0,
+		0x1fee0, 0x1fee0,
+		0x1ff00, 0x1ff84,
+		0x1ffc0, 0x1ffc8,
+		0x20000, 0x2002c,
+		0x20100, 0x2013c,
+		0x20190, 0x201c8,
+		0x20200, 0x20318,
+		0x20400, 0x20528,
+		0x20540, 0x20614,
+		0x21000, 0x21040,
+		0x2104c, 0x21060,
+		0x210c0, 0x210ec,
+		0x21200, 0x21268,
+		0x21270, 0x21284,
+		0x212fc, 0x21388,
+		0x21400, 0x21404,
+		0x21500, 0x21518,
+		0x2152c, 0x2153c,
+		0x21550, 0x21554,
+		0x21600, 0x21600,
+		0x21608, 0x21628,
+		0x21630, 0x2163c,
+		0x21700, 0x2171c,
+		0x21780, 0x2178c,
+		0x21800, 0x21c38,
+		0x21c80, 0x21d7c,
+		0x21e00, 0x21e04,
+		0x22000, 0x2202c,
+		0x22100, 0x2213c,
+		0x22190, 0x221c8,
+		0x22200, 0x22318,
+		0x22400, 0x22528,
+		0x22540, 0x22614,
+		0x23000, 0x23040,
+		0x2304c, 0x23060,
+		0x230c0, 0x230ec,
+		0x23200, 0x23268,
+		0x23270, 0x23284,
+		0x232fc, 0x23388,
+		0x23400, 0x23404,
+		0x23500, 0x23518,
+		0x2352c, 0x2353c,
+		0x23550, 0x23554,
+		0x23600, 0x23600,
+		0x23608, 0x23628,
+		0x23630, 0x2363c,
+		0x23700, 0x2371c,
+		0x23780, 0x2378c,
+		0x23800, 0x23c38,
+		0x23c80, 0x23d7c,
+		0x23e00, 0x23e04,
+		0x24000, 0x2402c,
+		0x24100, 0x2413c,
+		0x24190, 0x241c8,
+		0x24200, 0x24318,
+		0x24400, 0x24528,
+		0x24540, 0x24614,
+		0x25000, 0x25040,
+		0x2504c, 0x25060,
+		0x250c0, 0x250ec,
+		0x25200, 0x25268,
+		0x25270, 0x25284,
+		0x252fc, 0x25388,
+		0x25400, 0x25404,
+		0x25500, 0x25518,
+		0x2552c, 0x2553c,
+		0x25550, 0x25554,
+		0x25600, 0x25600,
+		0x25608, 0x25628,
+		0x25630, 0x2563c,
+		0x25700, 0x2571c,
+		0x25780, 0x2578c,
+		0x25800, 0x25c38,
+		0x25c80, 0x25d7c,
+		0x25e00, 0x25e04,
+		0x26000, 0x2602c,
+		0x26100, 0x2613c,
+		0x26190, 0x261c8,
+		0x26200, 0x26318,
+		0x26400, 0x26528,
+		0x26540, 0x26614,
+		0x27000, 0x27040,
+		0x2704c, 0x27060,
+		0x270c0, 0x270ec,
+		0x27200, 0x27268,
+		0x27270, 0x27284,
+		0x272fc, 0x27388,
+		0x27400, 0x27404,
+		0x27500, 0x27518,
+		0x2752c, 0x2753c,
+		0x27550, 0x27554,
+		0x27600, 0x27600,
+		0x27608, 0x27628,
+		0x27630, 0x2763c,
+		0x27700, 0x2771c,
+		0x27780, 0x2778c,
+		0x27800, 0x27c38,
+		0x27c80, 0x27d7c,
+		0x27e00, 0x27e04
+	};
+
+	int i;
+	struct adapter *ap = netdev2adap(dev);
+
+	regs->version = mk_adap_vers(ap);
+
+	memset(buf, 0, T4_REGMAP_SIZE);
+	for (i = 0; i < ARRAY_SIZE(reg_ranges); i += 2)
+		reg_block_dump(ap, buf, reg_ranges[i], reg_ranges[i + 1]);
+}
+
+static int restart_autoneg(struct net_device *dev)
+{
+	struct port_info *p = netdev_priv(dev);
+
+	if (!netif_running(dev))
+		return -EAGAIN;
+	if (p->link_cfg.autoneg != AUTONEG_ENABLE)
+		return -EINVAL;
+	t4_restart_aneg(p->adapter, 0, p->tx_chan);
+	return 0;
+}
+
+static int identify_port(struct net_device *dev, u32 data)
+{
+	if (data == 0)
+		data = 2;     /* default to 2 seconds */
+
+	return t4_identify_port(netdev2adap(dev), 0, netdev2pinfo(dev)->viid,
+				data * 5);
+}
+
+static unsigned int from_fw_linkcaps(unsigned int type, unsigned int caps)
+{
+	unsigned int v = 0;
+
+	if (type == FW_PORT_TYPE_BT_SGMII || type == FW_PORT_TYPE_BT_XAUI) {
+		v |= SUPPORTED_TP;
+		if (caps & FW_PORT_CAP_SPEED_100M)
+			v |= SUPPORTED_100baseT_Full;
+		if (caps & FW_PORT_CAP_SPEED_1G)
+			v |= SUPPORTED_1000baseT_Full;
+		if (caps & FW_PORT_CAP_SPEED_10G)
+			v |= SUPPORTED_10000baseT_Full;
+	} else if (type == FW_PORT_TYPE_KX4 || type == FW_PORT_TYPE_KX) {
+		v |= SUPPORTED_Backplane;
+		if (caps & FW_PORT_CAP_SPEED_1G)
+			v |= SUPPORTED_1000baseKX_Full;
+		if (caps & FW_PORT_CAP_SPEED_10G)
+			v |= SUPPORTED_10000baseKX4_Full;
+	} else if (type == FW_PORT_TYPE_KR)
+		v |= SUPPORTED_Backplane | SUPPORTED_10000baseKR_Full;
+	else if (type == FW_PORT_TYPE_FIBER)
+		v |= SUPPORTED_FIBRE;
+
+	if (caps & FW_PORT_CAP_ANEG)
+		v |= SUPPORTED_Autoneg;
+	return v;
+}
+
+static unsigned int to_fw_linkcaps(unsigned int caps)
+{
+	unsigned int v = 0;
+
+	if (caps & ADVERTISED_100baseT_Full)
+		v |= FW_PORT_CAP_SPEED_100M;
+	if (caps & ADVERTISED_1000baseT_Full)
+		v |= FW_PORT_CAP_SPEED_1G;
+	if (caps & ADVERTISED_10000baseT_Full)
+		v |= FW_PORT_CAP_SPEED_10G;
+	return v;
+}
+
+static int get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	const struct port_info *p = netdev_priv(dev);
+
+	if (p->port_type == FW_PORT_TYPE_BT_SGMII ||
+	    p->port_type == FW_PORT_TYPE_BT_XAUI)
+		cmd->port = PORT_TP;
+	else if (p->port_type == FW_PORT_TYPE_FIBER)
+		cmd->port = PORT_FIBRE;
+	else if (p->port_type == FW_PORT_TYPE_TWINAX)
+		cmd->port = PORT_DA;
+	else
+		cmd->port = PORT_OTHER;
+
+	if (p->mdio_addr >= 0) {
+		cmd->phy_address = p->mdio_addr;
+		cmd->transceiver = XCVR_EXTERNAL;
+		cmd->mdio_support = p->port_type == FW_PORT_TYPE_BT_SGMII ?
+			MDIO_SUPPORTS_C22 : MDIO_SUPPORTS_C45;
+	} else {
+		cmd->phy_address = 0;  /* not really, but no better option */
+		cmd->transceiver = XCVR_INTERNAL;
+		cmd->mdio_support = 0;
+	}
+
+	cmd->supported = from_fw_linkcaps(p->port_type, p->link_cfg.supported);
+	cmd->advertising = from_fw_linkcaps(p->port_type,
+					    p->link_cfg.advertising);
+	cmd->speed = netif_carrier_ok(dev) ? p->link_cfg.speed : 0;
+	cmd->duplex = DUPLEX_FULL;
+	cmd->autoneg = p->link_cfg.autoneg;
+	cmd->maxtxpkt = 0;
+	cmd->maxrxpkt = 0;
+	return 0;
+}
+
+static unsigned int speed_to_caps(int speed)
+{
+	if (speed == SPEED_100)
+		return FW_PORT_CAP_SPEED_100M;
+	if (speed == SPEED_1000)
+		return FW_PORT_CAP_SPEED_1G;
+	if (speed == SPEED_10000)
+		return FW_PORT_CAP_SPEED_10G;
+	return 0;
+}
+
+static int set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	unsigned int cap;
+	struct port_info *p = netdev_priv(dev);
+	struct link_config *lc = &p->link_cfg;
+
+	if (cmd->duplex != DUPLEX_FULL)     /* only full-duplex supported */
+		return -EINVAL;
+
+	if (!(lc->supported & FW_PORT_CAP_ANEG)) {
+		/*
+		 * PHY offers a single speed.  See if that's what's
+		 * being requested.
+		 */
+		if (cmd->autoneg == AUTONEG_DISABLE &&
+		    (lc->supported & speed_to_caps(cmd->speed)))
+				return 0;
+		return -EINVAL;
+	}
+
+	if (cmd->autoneg == AUTONEG_DISABLE) {
+		cap = speed_to_caps(cmd->speed);
+
+		if (!(lc->supported & cap) || cmd->speed == SPEED_1000 ||
+		    cmd->speed == SPEED_10000)
+			return -EINVAL;
+		lc->requested_speed = cap;
+		lc->advertising = 0;
+	} else {
+		cap = to_fw_linkcaps(cmd->advertising);
+		if (!(lc->supported & cap))
+			return -EINVAL;
+		lc->requested_speed = 0;
+		lc->advertising = cap | FW_PORT_CAP_ANEG;
+	}
+	lc->autoneg = cmd->autoneg;
+
+	if (netif_running(dev))
+		return t4_link_start(p->adapter, 0, p->tx_chan, lc);
+	return 0;
+}
+
+static void get_pauseparam(struct net_device *dev,
+			   struct ethtool_pauseparam *epause)
+{
+	struct port_info *p = netdev_priv(dev);
+
+	epause->autoneg = (p->link_cfg.requested_fc & PAUSE_AUTONEG) != 0;
+	epause->rx_pause = (p->link_cfg.fc & PAUSE_RX) != 0;
+	epause->tx_pause = (p->link_cfg.fc & PAUSE_TX) != 0;
+}
+
+static int set_pauseparam(struct net_device *dev,
+			  struct ethtool_pauseparam *epause)
+{
+	struct port_info *p = netdev_priv(dev);
+	struct link_config *lc = &p->link_cfg;
+
+	if (epause->autoneg == AUTONEG_DISABLE)
+		lc->requested_fc = 0;
+	else if (lc->supported & FW_PORT_CAP_ANEG)
+		lc->requested_fc = PAUSE_AUTONEG;
+	else
+		return -EINVAL;
+
+	if (epause->rx_pause)
+		lc->requested_fc |= PAUSE_RX;
+	if (epause->tx_pause)
+		lc->requested_fc |= PAUSE_TX;
+	if (netif_running(dev))
+		return t4_link_start(p->adapter, 0, p->tx_chan, lc);
+	return 0;
+}
+
+static u32 get_rx_csum(struct net_device *dev)
+{
+	struct port_info *p = netdev_priv(dev);
+
+	return p->rx_offload & RX_CSO;
+}
+
+static int set_rx_csum(struct net_device *dev, u32 data)
+{
+	struct port_info *p = netdev_priv(dev);
+
+	if (data)
+		p->rx_offload |= RX_CSO;
+	else
+		p->rx_offload &= ~RX_CSO;
+	return 0;
+}
+
+static void get_sge_param(struct net_device *dev, struct ethtool_ringparam *e)
+{
+	const struct port_info *pi = netdev_priv(dev);
+	const struct sge *s = &pi->adapter->sge;
+
+	e->rx_max_pending = MAX_RX_BUFFERS;
+	e->rx_mini_max_pending = MAX_RSPQ_ENTRIES;
+	e->rx_jumbo_max_pending = 0;
+	e->tx_max_pending = MAX_TXQ_ENTRIES;
+
+	e->rx_pending = s->ethrxq[pi->first_qset].fl.size - 8;
+	e->rx_mini_pending = s->ethrxq[pi->first_qset].rspq.size;
+	e->rx_jumbo_pending = 0;
+	e->tx_pending = s->ethtxq[pi->first_qset].q.size;
+}
+
+static int set_sge_param(struct net_device *dev, struct ethtool_ringparam *e)
+{
+	int i;
+	const struct port_info *pi = netdev_priv(dev);
+	struct adapter *adapter = pi->adapter;
+	struct sge *s = &adapter->sge;
+
+	if (e->rx_pending > MAX_RX_BUFFERS || e->rx_jumbo_pending ||
+	    e->tx_pending > MAX_TXQ_ENTRIES ||
+	    e->rx_mini_pending > MAX_RSPQ_ENTRIES ||
+	    e->rx_mini_pending < MIN_RSPQ_ENTRIES ||
+	    e->rx_pending < MIN_FL_ENTRIES || e->tx_pending < MIN_TXQ_ENTRIES)
+		return -EINVAL;
+
+	if (adapter->flags & FULL_INIT_DONE)
+		return -EBUSY;
+
+	for (i = 0; i < pi->nqsets; ++i) {
+		s->ethtxq[pi->first_qset + i].q.size = e->tx_pending;
+		s->ethrxq[pi->first_qset + i].fl.size = e->rx_pending + 8;
+		s->ethrxq[pi->first_qset + i].rspq.size = e->rx_mini_pending;
+	}
+	return 0;
+}
+
+static int closest_timer(const struct sge *s, int time)
+{
+	int i, delta, match = 0, min_delta = INT_MAX;
+
+	for (i = 0; i < ARRAY_SIZE(s->timer_val); i++) {
+		delta = time - s->timer_val[i];
+		if (delta < 0)
+			delta = -delta;
+		if (delta < min_delta) {
+			min_delta = delta;
+			match = i;
+		}
+	}
+	return match;
+}
+
+static int closest_thres(const struct sge *s, int thres)
+{
+	int i, delta, match = 0, min_delta = INT_MAX;
+
+	for (i = 0; i < ARRAY_SIZE(s->counter_val); i++) {
+		delta = thres - s->counter_val[i];
+		if (delta < 0)
+			delta = -delta;
+		if (delta < min_delta) {
+			min_delta = delta;
+			match = i;
+		}
+	}
+	return match;
+}
+
+/*
+ * Return a queue's interrupt hold-off time in us.  0 means no timer.
+ */
+static unsigned int qtimer_val(const struct adapter *adap,
+			       const struct sge_rspq *q)
+{
+	unsigned int idx = q->intr_params >> 1;
+
+	return idx < SGE_NTIMERS ? adap->sge.timer_val[idx] : 0;
+}
+
+/**
+ *	set_rxq_intr_params - set a queue's interrupt holdoff parameters
+ *	@adap: the adapter
+ *	@q: the Rx queue
+ *	@us: the hold-off time in us, or 0 to disable timer
+ *	@cnt: the hold-off packet count, or 0 to disable counter
+ *
+ *	Sets an Rx queue's interrupt hold-off time and packet count.  At least
+ *	one of the two needs to be enabled for the queue to generate interrupts.
+ */
+static int set_rxq_intr_params(struct adapter *adap, struct sge_rspq *q,
+			       unsigned int us, unsigned int cnt)
+{
+	if ((us | cnt) == 0)
+		cnt = 1;
+
+	if (cnt) {
+		int err;
+		u32 v, new_idx;
+
+		new_idx = closest_thres(&adap->sge, cnt);
+		if (q->desc && q->pktcnt_idx != new_idx) {
+			/* the queue has already been created, update it */
+			v = FW_PARAMS_MNEM(FW_PARAMS_MNEM_DMAQ) |
+			    FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DMAQ_IQ_INTCNTTHRESH) |
+			    FW_PARAMS_PARAM_YZ(q->cntxt_id);
+			err = t4_set_params(adap, 0, 0, 0, 1, &v, &new_idx);
+			if (err)
+				return err;
+		}
+		q->pktcnt_idx = new_idx;
+	}
+
+	us = us == 0 ? 6 : closest_timer(&adap->sge, us);
+	q->intr_params = QINTR_TIMER_IDX(us) | (cnt > 0 ? QINTR_CNT_EN : 0);
+	return 0;
+}
+
+static int set_coalesce(struct net_device *dev, struct ethtool_coalesce *c)
+{
+	const struct port_info *pi = netdev_priv(dev);
+	struct adapter *adap = pi->adapter;
+
+	return set_rxq_intr_params(adap, &adap->sge.ethrxq[pi->first_qset].rspq,
+			c->rx_coalesce_usecs, c->rx_max_coalesced_frames);
+}
+
+static int get_coalesce(struct net_device *dev, struct ethtool_coalesce *c)
+{
+	const struct port_info *pi = netdev_priv(dev);
+	const struct adapter *adap = pi->adapter;
+	const struct sge_rspq *rq = &adap->sge.ethrxq[pi->first_qset].rspq;
+
+	c->rx_coalesce_usecs = qtimer_val(adap, rq);
+	c->rx_max_coalesced_frames = (rq->intr_params & QINTR_CNT_EN) ?
+		adap->sge.counter_val[rq->pktcnt_idx] : 0;
+	return 0;
+}
+
+/*
+ * Translate a physical EEPROM address to virtual.  The first 1K is accessed
+ * through virtual addresses starting at 31K, the rest is accessed through
+ * virtual addresses starting at 0.  This mapping is correct only for PF0.
+ */
+static int eeprom_ptov(unsigned int phys_addr)
+{
+	if (phys_addr < 1024)
+		return phys_addr + (31 << 10);
+	if (phys_addr < EEPROMSIZE)
+		return phys_addr - 1024;
+	return -EINVAL;
+}
+
+/*
+ * The next two routines implement eeprom read/write from physical addresses.
+ * The physical->virtual translation is correct only for PF0.
+ */
+static int eeprom_rd_phys(struct adapter *adap, unsigned int phys_addr, u32 *v)
+{
+	int vaddr = eeprom_ptov(phys_addr);
+
+	if (vaddr >= 0)
+		vaddr = pci_read_vpd(adap->pdev, vaddr, sizeof(u32), v);
+	return vaddr < 0 ? vaddr : 0;
+}
+
+static int eeprom_wr_phys(struct adapter *adap, unsigned int phys_addr, u32 v)
+{
+	int vaddr = eeprom_ptov(phys_addr);
+
+	if (vaddr >= 0)
+		vaddr = pci_write_vpd(adap->pdev, vaddr, sizeof(u32), &v);
+	return vaddr < 0 ? vaddr : 0;
+}
+
+#define EEPROM_MAGIC 0x38E2F10C
+
+static int get_eeprom(struct net_device *dev, struct ethtool_eeprom *e,
+		      u8 *data)
+{
+	int i, err = 0;
+	struct adapter *adapter = netdev2adap(dev);
+
+	u8 *buf = kmalloc(EEPROMSIZE, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	e->magic = EEPROM_MAGIC;
+	for (i = e->offset & ~3; !err && i < e->offset + e->len; i += 4)
+		err = eeprom_rd_phys(adapter, i, (u32 *)&buf[i]);
+
+	if (!err)
+		memcpy(data, buf + e->offset, e->len);
+	kfree(buf);
+	return err;
+}
+
+static int set_eeprom(struct net_device *dev, struct ethtool_eeprom *eeprom,
+		      u8 *data)
+{
+	u8 *buf;
+	int err = 0;
+	u32 aligned_offset, aligned_len, *p;
+	struct adapter *adapter = netdev2adap(dev);
+
+	if (eeprom->magic != EEPROM_MAGIC)
+		return -EINVAL;
+
+	aligned_offset = eeprom->offset & ~3;
+	aligned_len = (eeprom->len + (eeprom->offset & 3) + 3) & ~3;
+
+	if (aligned_offset != eeprom->offset || aligned_len != eeprom->len) {
+		/*
+		 * RMW possibly needed for first or last words.
+		 */
+		buf = kmalloc(aligned_len, GFP_KERNEL);
+		if (!buf)
+			return -ENOMEM;
+		err = eeprom_rd_phys(adapter, aligned_offset, (u32 *)buf);
+		if (!err && aligned_len > 4)
+			err = eeprom_rd_phys(adapter,
+					     aligned_offset + aligned_len - 4,
+					     (u32 *)&buf[aligned_len - 4]);
+		if (err)
+			goto out;
+		memcpy(buf + (eeprom->offset & 3), data, eeprom->len);
+	} else
+		buf = data;
+
+	err = t4_seeprom_wp(adapter, false);
+	if (err)
+		goto out;
+
+	for (p = (u32 *)buf; !err && aligned_len; aligned_len -= 4, p++) {
+		err = eeprom_wr_phys(adapter, aligned_offset, *p);
+		aligned_offset += 4;
+	}
+
+	if (!err)
+		err = t4_seeprom_wp(adapter, true);
+out:
+	if (buf != data)
+		kfree(buf);
+	return err;
+}
+
+static int set_flash(struct net_device *netdev, struct ethtool_flash *ef)
+{
+	int ret;
+	const struct firmware *fw;
+	struct adapter *adap = netdev2adap(netdev);
+
+	ef->data[sizeof(ef->data) - 1] = '\0';
+	ret = request_firmware(&fw, ef->data, adap->pdev_dev);
+	if (ret < 0)
+		return ret;
+
+	ret = t4_load_fw(adap, fw->data, fw->size);
+	release_firmware(fw);
+	if (!ret)
+		dev_info(adap->pdev_dev, "loaded firmware %s\n", ef->data);
+	return ret;
+}
+
+#define WOL_SUPPORTED (WAKE_BCAST | WAKE_MAGIC)
+#define BCAST_CRC 0xa0ccc1a6
+
+static void get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
+{
+	wol->supported = WAKE_BCAST | WAKE_MAGIC;
+	wol->wolopts = netdev2adap(dev)->wol;
+	memset(&wol->sopass, 0, sizeof(wol->sopass));
+}
+
+static int set_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
+{
+	int err = 0;
+	struct port_info *pi = netdev_priv(dev);
+
+	if (wol->wolopts & ~WOL_SUPPORTED)
+		return -EINVAL;
+	t4_wol_magic_enable(pi->adapter, pi->tx_chan,
+			    (wol->wolopts & WAKE_MAGIC) ? dev->dev_addr : NULL);
+	if (wol->wolopts & WAKE_BCAST) {
+		err = t4_wol_pat_enable(pi->adapter, pi->tx_chan, 0xfe, ~0ULL,
+					~0ULL, 0, false);
+		if (!err)
+			err = t4_wol_pat_enable(pi->adapter, pi->tx_chan, 1,
+						~6ULL, ~0ULL, BCAST_CRC, true);
+	} else
+		t4_wol_pat_enable(pi->adapter, pi->tx_chan, 0, 0, 0, 0, false);
+	return err;
+}
+
+static int set_tso(struct net_device *dev, u32 value)
+{
+	if (value)
+		dev->features |= NETIF_F_TSO | NETIF_F_TSO6;
+	else
+		dev->features &= ~(NETIF_F_TSO | NETIF_F_TSO6);
+	return 0;
+}
+
+static struct ethtool_ops cxgb_ethtool_ops = {
+	.get_settings      = get_settings,
+	.set_settings      = set_settings,
+	.get_drvinfo       = get_drvinfo,
+	.get_msglevel      = get_msglevel,
+	.set_msglevel      = set_msglevel,
+	.get_ringparam     = get_sge_param,
+	.set_ringparam     = set_sge_param,
+	.get_coalesce      = get_coalesce,
+	.set_coalesce      = set_coalesce,
+	.get_eeprom_len    = get_eeprom_len,
+	.get_eeprom        = get_eeprom,
+	.set_eeprom        = set_eeprom,
+	.get_pauseparam    = get_pauseparam,
+	.set_pauseparam    = set_pauseparam,
+	.get_rx_csum       = get_rx_csum,
+	.set_rx_csum       = set_rx_csum,
+	.set_tx_csum       = ethtool_op_set_tx_ipv6_csum,
+	.set_sg            = ethtool_op_set_sg,
+	.get_link          = ethtool_op_get_link,
+	.get_strings       = get_strings,
+	.phys_id           = identify_port,
+	.nway_reset        = restart_autoneg,
+	.get_sset_count    = get_sset_count,
+	.get_ethtool_stats = get_stats,
+	.get_regs_len      = get_regs_len,
+	.get_regs          = get_regs,
+	.get_wol           = get_wol,
+	.set_wol           = set_wol,
+	.set_tso           = set_tso,
+	.flash_device      = set_flash,
+};
+
+/*
+ * debugfs support
+ */
+
+static int mem_open(struct inode *inode, struct file *file)
+{
+	file->private_data = inode->i_private;
+	return 0;
+}
+
+static ssize_t mem_read(struct file *file, char __user *buf, size_t count,
+			loff_t *ppos)
+{
+	loff_t pos = *ppos;
+	loff_t avail = file->f_path.dentry->d_inode->i_size;
+	unsigned int mem = (uintptr_t)file->private_data & 3;
+	struct adapter *adap = file->private_data - mem;
+
+	if (pos < 0)
+		return -EINVAL;
+	if (pos >= avail)
+		return 0;
+	if (count > avail - pos)
+		count = avail - pos;
+
+	while (count) {
+		size_t len;
+		int ret, ofst;
+		__be32 data[16];
+
+		if (mem == MEM_MC)
+			ret = t4_mc_read(adap, pos, data, NULL);
+		else
+			ret = t4_edc_read(adap, mem, pos, data, NULL);
+		if (ret)
+			return ret;
+
+		ofst = pos % sizeof(data);
+		len = min(count, sizeof(data) - ofst);
+		if (copy_to_user(buf, (u8 *)data + ofst, len))
+			return -EFAULT;
+
+		buf += len;
+		pos += len;
+		count -= len;
+	}
+	count = pos - *ppos;
+	*ppos = pos;
+	return count;
+}
+
+static const struct file_operations mem_debugfs_fops = {
+	.owner   = THIS_MODULE,
+	.open    = mem_open,
+	.read    = mem_read,
+};
+
+static void __devinit add_debugfs_mem(struct adapter *adap, const char *name,
+				      unsigned int idx, unsigned int size_mb)
+{
+	struct dentry *de;
+
+	de = debugfs_create_file(name, S_IRUSR, adap->debugfs_root,
+				 (void *)adap + idx, &mem_debugfs_fops);
+	if (de && de->d_inode)
+		de->d_inode->i_size = size_mb << 20;
+}
+
+static int __devinit setup_debugfs(struct adapter *adap)
+{
+	int i;
+
+	if (IS_ERR_OR_NULL(adap->debugfs_root))
+		return -1;
+
+	i = t4_read_reg(adap, MA_TARGET_MEM_ENABLE);
+	if (i & EDRAM0_ENABLE)
+		add_debugfs_mem(adap, "edc0", MEM_EDC0, 5);
+	if (i & EDRAM1_ENABLE)
+		add_debugfs_mem(adap, "edc1", MEM_EDC1, 5);
+	if (i & EXT_MEM_ENABLE)
+		add_debugfs_mem(adap, "mc", MEM_MC,
+			EXT_MEM_SIZE_GET(t4_read_reg(adap, MA_EXT_MEMORY_BAR)));
+	if (adap->l2t)
+		debugfs_create_file("l2t", S_IRUSR, adap->debugfs_root, adap,
+				    &t4_l2t_fops);
+	return 0;
+}
+
+/*
+ * upper-layer driver support
+ */
+
+/*
+ * Allocate an active-open TID and set it to the supplied value.
+ */
+int cxgb4_alloc_atid(struct tid_info *t, void *data)
+{
+	int atid = -1;
+
+	spin_lock_bh(&t->atid_lock);
+	if (t->afree) {
+		union aopen_entry *p = t->afree;
+
+		atid = p - t->atid_tab;
+		t->afree = p->next;
+		p->data = data;
+		t->atids_in_use++;
+	}
+	spin_unlock_bh(&t->atid_lock);
+	return atid;
+}
+EXPORT_SYMBOL(cxgb4_alloc_atid);
+
+/*
+ * Release an active-open TID.
+ */
+void cxgb4_free_atid(struct tid_info *t, unsigned int atid)
+{
+	union aopen_entry *p = &t->atid_tab[atid];
+
+	spin_lock_bh(&t->atid_lock);
+	p->next = t->afree;
+	t->afree = p;
+	t->atids_in_use--;
+	spin_unlock_bh(&t->atid_lock);
+}
+EXPORT_SYMBOL(cxgb4_free_atid);
+
+/*
+ * Allocate a server TID and set it to the supplied value.
+ */
+int cxgb4_alloc_stid(struct tid_info *t, int family, void *data)
+{
+	int stid;
+
+	spin_lock_bh(&t->stid_lock);
+	if (family == PF_INET) {
+		stid = find_first_zero_bit(t->stid_bmap, t->nstids);
+		if (stid < t->nstids)
+			__set_bit(stid, t->stid_bmap);
+		else
+			stid = -1;
+	} else {
+		stid = bitmap_find_free_region(t->stid_bmap, t->nstids, 2);
+		if (stid < 0)
+			stid = -1;
+	}
+	if (stid >= 0) {
+		t->stid_tab[stid].data = data;
+		stid += t->stid_base;
+		t->stids_in_use++;
+	}
+	spin_unlock_bh(&t->stid_lock);
+	return stid;
+}
+EXPORT_SYMBOL(cxgb4_alloc_stid);
+
+/*
+ * Release a server TID.
+ */
+void cxgb4_free_stid(struct tid_info *t, unsigned int stid, int family)
+{
+	stid -= t->stid_base;
+	spin_lock_bh(&t->stid_lock);
+	if (family == PF_INET)
+		__clear_bit(stid, t->stid_bmap);
+	else
+		bitmap_release_region(t->stid_bmap, stid, 2);
+	t->stid_tab[stid].data = NULL;
+	t->stids_in_use--;
+	spin_unlock_bh(&t->stid_lock);
+}
+EXPORT_SYMBOL(cxgb4_free_stid);
+
+/*
+ * Populate a TID_RELEASE WR.  Caller must properly size the skb.
+ */
+static void mk_tid_release(struct sk_buff *skb, unsigned int chan,
+			   unsigned int tid)
+{
+	struct cpl_tid_release *req;
+
+	set_wr_txq(skb, CPL_PRIORITY_SETUP, chan);
+	req = (struct cpl_tid_release *)__skb_put(skb, sizeof(*req));
+	INIT_TP_WR(req, tid);
+	OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_TID_RELEASE, tid));
+}
+
+/*
+ * Queue a TID release request and if necessary schedule a work queue to
+ * process it.
+ */
+void cxgb4_queue_tid_release(struct tid_info *t, unsigned int chan,
+			     unsigned int tid)
+{
+	void **p = &t->tid_tab[tid];
+	struct adapter *adap = container_of(t, struct adapter, tids);
+
+	spin_lock_bh(&adap->tid_release_lock);
+	*p = adap->tid_release_head;
+	/* Low 2 bits encode the Tx channel number */
+	adap->tid_release_head = (void **)((uintptr_t)p | chan);
+	if (!adap->tid_release_task_busy) {
+		adap->tid_release_task_busy = true;
+		schedule_work(&adap->tid_release_task);
+	}
+	spin_unlock_bh(&adap->tid_release_lock);
+}
+EXPORT_SYMBOL(cxgb4_queue_tid_release);
+
+/*
+ * Process the list of pending TID release requests.
+ */
+static void process_tid_release_list(struct work_struct *work)
+{
+	struct sk_buff *skb;
+	struct adapter *adap;
+
+	adap = container_of(work, struct adapter, tid_release_task);
+
+	spin_lock_bh(&adap->tid_release_lock);
+	while (adap->tid_release_head) {
+		void **p = adap->tid_release_head;
+		unsigned int chan = (uintptr_t)p & 3;
+		p = (void *)p - chan;
+
+		adap->tid_release_head = *p;
+		*p = NULL;
+		spin_unlock_bh(&adap->tid_release_lock);
+
+		while (!(skb = alloc_skb(sizeof(struct cpl_tid_release),
+					 GFP_KERNEL)))
+			schedule_timeout_uninterruptible(1);
+
+		mk_tid_release(skb, chan, p - adap->tids.tid_tab);
+		t4_ofld_send(adap, skb);
+		spin_lock_bh(&adap->tid_release_lock);
+	}
+	adap->tid_release_task_busy = false;
+	spin_unlock_bh(&adap->tid_release_lock);
+}
+
+/*
+ * Release a TID and inform HW.  If we are unable to allocate the release
+ * message we defer to a work queue.
+ */
+void cxgb4_remove_tid(struct tid_info *t, unsigned int chan, unsigned int tid)
+{
+	void *old;
+	struct sk_buff *skb;
+	struct adapter *adap = container_of(t, struct adapter, tids);
+
+	old = t->tid_tab[tid];
+	skb = alloc_skb(sizeof(struct cpl_tid_release), GFP_ATOMIC);
+	if (likely(skb)) {
+		t->tid_tab[tid] = NULL;
+		mk_tid_release(skb, chan, tid);
+		t4_ofld_send(adap, skb);
+	} else
+		cxgb4_queue_tid_release(t, chan, tid);
+	if (old)
+		atomic_dec(&t->tids_in_use);
+}
+EXPORT_SYMBOL(cxgb4_remove_tid);
+
+/*
+ * Allocate and initialize the TID tables.  Returns 0 on success.
+ */
+static int tid_init(struct tid_info *t)
+{
+	size_t size;
+	unsigned int natids = t->natids;
+
+	size = t->ntids * sizeof(*t->tid_tab) + natids * sizeof(*t->atid_tab) +
+	       t->nstids * sizeof(*t->stid_tab) +
+	       BITS_TO_LONGS(t->nstids) * sizeof(long);
+	t->tid_tab = t4_alloc_mem(size);
+	if (!t->tid_tab)
+		return -ENOMEM;
+
+	t->atid_tab = (union aopen_entry *)&t->tid_tab[t->ntids];
+	t->stid_tab = (struct serv_entry *)&t->atid_tab[natids];
+	t->stid_bmap = (unsigned long *)&t->stid_tab[t->nstids];
+	spin_lock_init(&t->stid_lock);
+	spin_lock_init(&t->atid_lock);
+
+	t->stids_in_use = 0;
+	t->afree = NULL;
+	t->atids_in_use = 0;
+	atomic_set(&t->tids_in_use, 0);
+
+	/* Setup the free list for atid_tab and clear the stid bitmap. */
+	if (natids) {
+		while (--natids)
+			t->atid_tab[natids - 1].next = &t->atid_tab[natids];
+		t->afree = t->atid_tab;
+	}
+	bitmap_zero(t->stid_bmap, t->nstids);
+	return 0;
+}
+
+/**
+ *	cxgb4_create_server - create an IP server
+ *	@dev: the device
+ *	@stid: the server TID
+ *	@sip: local IP address to bind server to
+ *	@sport: the server's TCP port
+ *	@queue: queue to direct messages from this server to
+ *
+ *	Create an IP server for the given port and address.
+ *	Returns <0 on error and one of the %NET_XMIT_* values on success.
+ */
+int cxgb4_create_server(const struct net_device *dev, unsigned int stid,
+			__be32 sip, __be16 sport, unsigned int queue)
+{
+	unsigned int chan;
+	struct sk_buff *skb;
+	struct adapter *adap;
+	struct cpl_pass_open_req *req;
+
+	skb = alloc_skb(sizeof(*req), GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	adap = netdev2adap(dev);
+	req = (struct cpl_pass_open_req *)__skb_put(skb, sizeof(*req));
+	INIT_TP_WR(req, 0);
+	OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ, stid));
+	req->local_port = sport;
+	req->peer_port = htons(0);
+	req->local_ip = sip;
+	req->peer_ip = htonl(0);
+	chan = netdev2pinfo(adap->sge.ingr_map[queue]->netdev)->tx_chan;
+	req->opt0 = cpu_to_be64(TX_CHAN(chan));
+	req->opt1 = cpu_to_be64(CONN_POLICY_ASK |
+				SYN_RSS_ENABLE | SYN_RSS_QUEUE(queue));
+	return t4_mgmt_tx(adap, skb);
+}
+EXPORT_SYMBOL(cxgb4_create_server);
+
+/**
+ *	cxgb4_create_server6 - create an IPv6 server
+ *	@dev: the device
+ *	@stid: the server TID
+ *	@sip: local IPv6 address to bind server to
+ *	@sport: the server's TCP port
+ *	@queue: queue to direct messages from this server to
+ *
+ *	Create an IPv6 server for the given port and address.
+ *	Returns <0 on error and one of the %NET_XMIT_* values on success.
+ */
+int cxgb4_create_server6(const struct net_device *dev, unsigned int stid,
+			 const struct in6_addr *sip, __be16 sport,
+			 unsigned int queue)
+{
+	unsigned int chan;
+	struct sk_buff *skb;
+	struct adapter *adap;
+	struct cpl_pass_open_req6 *req;
+
+	skb = alloc_skb(sizeof(*req), GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
+	adap = netdev2adap(dev);
+	req = (struct cpl_pass_open_req6 *)__skb_put(skb, sizeof(*req));
+	INIT_TP_WR(req, 0);
+	OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ6, stid));
+	req->local_port = sport;
+	req->peer_port = htons(0);
+	req->local_ip_hi = *(__be64 *)(sip->s6_addr);
+	req->local_ip_lo = *(__be64 *)(sip->s6_addr + 8);
+	req->peer_ip_hi = cpu_to_be64(0);
+	req->peer_ip_lo = cpu_to_be64(0);
+	chan = netdev2pinfo(adap->sge.ingr_map[queue]->netdev)->tx_chan;
+	req->opt0 = cpu_to_be64(TX_CHAN(chan));
+	req->opt1 = cpu_to_be64(CONN_POLICY_ASK |
+				SYN_RSS_ENABLE | SYN_RSS_QUEUE(queue));
+	return t4_mgmt_tx(adap, skb);
+}
+EXPORT_SYMBOL(cxgb4_create_server6);
+
+/**
+ *	cxgb4_best_mtu - find the entry in the MTU table closest to an MTU
+ *	@mtus: the HW MTU table
+ *	@mtu: the target MTU
+ *	@idx: index of selected entry in the MTU table
+ *
+ *	Returns the index and the value in the HW MTU table that is closest to
+ *	but does not exceed @mtu, unless @mtu is smaller than any value in the
+ *	table, in which case that smallest available value is selected.
+ */
+unsigned int cxgb4_best_mtu(const unsigned short *mtus, unsigned short mtu,
+			    unsigned int *idx)
+{
+	unsigned int i = 0;
+
+	while (i < NMTUS - 1 && mtus[i + 1] <= mtu)
+		++i;
+	if (idx)
+		*idx = i;
+	return mtus[i];
+}
+EXPORT_SYMBOL(cxgb4_best_mtu);
+
+/**
+ *	cxgb4_port_chan - get the HW channel of a port
+ *	@dev: the net device for the port
+ *
+ *	Return the HW Tx channel of the given port.
+ */
+unsigned int cxgb4_port_chan(const struct net_device *dev)
+{
+	return netdev2pinfo(dev)->tx_chan;
+}
+EXPORT_SYMBOL(cxgb4_port_chan);
+
+/**
+ *	cxgb4_port_viid - get the VI id of a port
+ *	@dev: the net device for the port
+ *
+ *	Return the VI id of the given port.
+ */
+unsigned int cxgb4_port_viid(const struct net_device *dev)
+{
+	return netdev2pinfo(dev)->viid;
+}
+EXPORT_SYMBOL(cxgb4_port_viid);
+
+/**
+ *	cxgb4_port_idx - get the index of a port
+ *	@dev: the net device for the port
+ *
+ *	Return the index of the given port.
+ */
+unsigned int cxgb4_port_idx(const struct net_device *dev)
+{
+	return netdev2pinfo(dev)->port_id;
+}
+EXPORT_SYMBOL(cxgb4_port_idx);
+
+/**
+ *	cxgb4_netdev_by_hwid - return the net device of a HW port
+ *	@pdev: identifies the adapter
+ *	@id: the HW port id
+ *
+ *	Return the net device associated with the interface with the given HW
+ *	id.
+ */
+struct net_device *cxgb4_netdev_by_hwid(struct pci_dev *pdev, unsigned int id)
+{
+	const struct adapter *adap = pci_get_drvdata(pdev);
+
+	if (!adap || id >= NCHAN)
+		return NULL;
+	id = adap->chan_map[id];
+	return id < MAX_NPORTS ? adap->port[id] : NULL;
+}
+EXPORT_SYMBOL(cxgb4_netdev_by_hwid);
+
+void cxgb4_get_tcp_stats(struct pci_dev *pdev, struct tp_tcp_stats *v4,
+			 struct tp_tcp_stats *v6)
+{
+	struct adapter *adap = pci_get_drvdata(pdev);
+
+	spin_lock(&adap->stats_lock);
+	t4_tp_get_tcp_stats(adap, v4, v6);
+	spin_unlock(&adap->stats_lock);
+}
+EXPORT_SYMBOL(cxgb4_get_tcp_stats);
+
+void cxgb4_iscsi_init(struct net_device *dev, unsigned int tag_mask,
+		      const unsigned int *pgsz_order)
+{
+	struct adapter *adap = netdev2adap(dev);
+
+	t4_write_reg(adap, ULP_RX_ISCSI_TAGMASK, tag_mask);
+	t4_write_reg(adap, ULP_RX_ISCSI_PSZ, HPZ0(pgsz_order[0]) |
+		     HPZ1(pgsz_order[1]) | HPZ2(pgsz_order[2]) |
+		     HPZ3(pgsz_order[3]));
+}
+EXPORT_SYMBOL(cxgb4_iscsi_init);
+
+static struct pci_driver cxgb4_driver;
+
+static void check_neigh_update(struct neighbour *neigh)
+{
+	const struct device *parent;
+	const struct net_device *netdev = neigh->dev;
+
+	if (netdev->priv_flags & IFF_802_1Q_VLAN)
+		netdev = vlan_dev_real_dev(netdev);
+	parent = netdev->dev.parent;
+	if (parent && parent->driver == &cxgb4_driver.driver)
+		t4_l2t_update(dev_get_drvdata(parent), neigh);
+}
+
+static int netevent_cb(struct notifier_block *nb, unsigned long event,
+		       void *data)
+{
+	switch (event) {
+	case NETEVENT_NEIGH_UPDATE:
+		check_neigh_update(data);
+		break;
+	case NETEVENT_PMTU_UPDATE:
+	case NETEVENT_REDIRECT:
+	default:
+		break;
+	}
+	return 0;
+}
+
+static bool netevent_registered;
+static struct notifier_block cxgb4_netevent_nb = {
+	.notifier_call = netevent_cb
+};
+
+static void uld_attach(struct adapter *adap, unsigned int uld)
+{
+	void *handle;
+	struct cxgb4_lld_info lli;
+
+	lli.pdev = adap->pdev;
+	lli.l2t = adap->l2t;
+	lli.tids = &adap->tids;
+	lli.ports = adap->port;
+	lli.vr = &adap->vres;
+	lli.mtus = adap->params.mtus;
+	if (uld == CXGB4_ULD_RDMA) {
+		lli.rxq_ids = adap->sge.rdma_rxq;
+		lli.nrxq = adap->sge.rdmaqs;
+	} else if (uld == CXGB4_ULD_ISCSI) {
+		lli.rxq_ids = adap->sge.ofld_rxq;
+		lli.nrxq = adap->sge.ofldqsets;
+	}
+	lli.ntxq = adap->sge.ofldqsets;
+	lli.nchan = adap->params.nports;
+	lli.nports = adap->params.nports;
+	lli.wr_cred = adap->params.ofldq_wr_cred;
+	lli.adapter_type = adap->params.rev;
+	lli.iscsi_iolen = MAXRXDATA_GET(t4_read_reg(adap, TP_PARA_REG2));
+	lli.udb_density = 1 << QUEUESPERPAGEPF0_GET(
+			t4_read_reg(adap, SGE_EGRESS_QUEUES_PER_PAGE_PF));
+	lli.ucq_density = 1 << QUEUESPERPAGEPF0_GET(
+			t4_read_reg(adap, SGE_INGRESS_QUEUES_PER_PAGE_PF));
+	lli.gts_reg = adap->regs + MYPF_REG(SGE_PF_GTS);
+	lli.db_reg = adap->regs + MYPF_REG(SGE_PF_KDOORBELL);
+	lli.fw_vers = adap->params.fw_vers;
+
+	handle = ulds[uld].add(&lli);
+	if (IS_ERR(handle)) {
+		dev_warn(adap->pdev_dev,
+			 "could not attach to the %s driver, error %ld\n",
+			 uld_str[uld], PTR_ERR(handle));
+		return;
+	}
+
+	adap->uld_handle[uld] = handle;
+
+	if (!netevent_registered) {
+		register_netevent_notifier(&cxgb4_netevent_nb);
+		netevent_registered = true;
+	}
+}
+
+static void attach_ulds(struct adapter *adap)
+{
+	unsigned int i;
+
+	mutex_lock(&uld_mutex);
+	list_add_tail(&adap->list_node, &adapter_list);
+	for (i = 0; i < CXGB4_ULD_MAX; i++)
+		if (ulds[i].add)
+			uld_attach(adap, i);
+	mutex_unlock(&uld_mutex);
+}
+
+static void detach_ulds(struct adapter *adap)
+{
+	unsigned int i;
+
+	mutex_lock(&uld_mutex);
+	list_del(&adap->list_node);
+	for (i = 0; i < CXGB4_ULD_MAX; i++)
+		if (adap->uld_handle[i]) {
+			ulds[i].state_change(adap->uld_handle[i],
+					     CXGB4_STATE_DETACH);
+			adap->uld_handle[i] = NULL;
+		}
+	if (netevent_registered && list_empty(&adapter_list)) {
+		unregister_netevent_notifier(&cxgb4_netevent_nb);
+		netevent_registered = false;
+	}
+	mutex_unlock(&uld_mutex);
+}
+
+static void notify_ulds(struct adapter *adap, enum cxgb4_state new_state)
+{
+	unsigned int i;
+
+	mutex_lock(&uld_mutex);
+	for (i = 0; i < CXGB4_ULD_MAX; i++)
+		if (adap->uld_handle[i])
+			ulds[i].state_change(adap->uld_handle[i], new_state);
+	mutex_unlock(&uld_mutex);
+}
+
+/**
+ *	cxgb4_register_uld - register an upper-layer driver
+ *	@type: the ULD type
+ *	@p: the ULD methods
+ *
+ *	Registers an upper-layer driver with this driver and notifies the ULD
+ *	about any presently available devices that support its type.  Returns
+ *	%-EBUSY if a ULD of the same type is already registered.
+ */
+int cxgb4_register_uld(enum cxgb4_uld type, const struct cxgb4_uld_info *p)
+{
+	int ret = 0;
+	struct adapter *adap;
+
+	if (type >= CXGB4_ULD_MAX)
+		return -EINVAL;
+	mutex_lock(&uld_mutex);
+	if (ulds[type].add) {
+		ret = -EBUSY;
+		goto out;
+	}
+	ulds[type] = *p;
+	list_for_each_entry(adap, &adapter_list, list_node)
+		uld_attach(adap, type);
+out:	mutex_unlock(&uld_mutex);
+	return ret;
+}
+EXPORT_SYMBOL(cxgb4_register_uld);
+
+/**
+ *	cxgb4_unregister_uld - unregister an upper-layer driver
+ *	@type: the ULD type
+ *
+ *	Unregisters an existing upper-layer driver.
+ */
+int cxgb4_unregister_uld(enum cxgb4_uld type)
+{
+	struct adapter *adap;
+
+	if (type >= CXGB4_ULD_MAX)
+		return -EINVAL;
+	mutex_lock(&uld_mutex);
+	list_for_each_entry(adap, &adapter_list, list_node)
+		adap->uld_handle[type] = NULL;
+	ulds[type].add = NULL;
+	mutex_unlock(&uld_mutex);
+	return 0;
+}
+EXPORT_SYMBOL(cxgb4_unregister_uld);
+
+/**
+ *	cxgb_up - enable the adapter
+ *	@adap: adapter being enabled
+ *
+ *	Called when the first port is enabled, this function performs the
+ *	actions necessary to make an adapter operational, such as completing
+ *	the initialization of HW modules, and enabling interrupts.
+ *
+ *	Must be called with the rtnl lock held.
+ */
+static int cxgb_up(struct adapter *adap)
+{
+	int err = 0;
+
+	if (!(adap->flags & FULL_INIT_DONE)) {
+		err = setup_sge_queues(adap);
+		if (err)
+			goto out;
+		err = setup_rss(adap);
+		if (err) {
+			t4_free_sge_resources(adap);
+			goto out;
+		}
+		if (adap->flags & USING_MSIX)
+			name_msix_vecs(adap);
+		adap->flags |= FULL_INIT_DONE;
+	}
+
+	if (adap->flags & USING_MSIX) {
+		err = request_irq(adap->msix_info[0].vec, t4_nondata_intr, 0,
+				  adap->msix_info[0].desc, adap);
+		if (err)
+			goto irq_err;
+
+		err = request_msix_queue_irqs(adap);
+		if (err) {
+			free_irq(adap->msix_info[0].vec, adap);
+			goto irq_err;
+		}
+	} else {
+		err = request_irq(adap->pdev->irq, t4_intr_handler(adap),
+				  (adap->flags & USING_MSI) ? 0 : IRQF_SHARED,
+				  adap->name, adap);
+		if (err)
+			goto irq_err;
+	}
+	enable_rx(adap);
+	t4_sge_start(adap);
+	t4_intr_enable(adap);
+	notify_ulds(adap, CXGB4_STATE_UP);
+ out:
+	return err;
+ irq_err:
+	dev_err(adap->pdev_dev, "request_irq failed, err %d\n", err);
+	goto out;
+}
+
+static void cxgb_down(struct adapter *adapter)
+{
+	t4_intr_disable(adapter);
+	cancel_work_sync(&adapter->tid_release_task);
+	adapter->tid_release_task_busy = false;
+
+	if (adapter->flags & USING_MSIX) {
+		free_msix_queue_irqs(adapter);
+		free_irq(adapter->msix_info[0].vec, adapter);
+	} else
+		free_irq(adapter->pdev->irq, adapter);
+	quiesce_rx(adapter);
+}
+
+/*
+ * net_device operations
+ */
+static int cxgb_open(struct net_device *dev)
+{
+	int err;
+	struct port_info *pi = netdev_priv(dev);
+	struct adapter *adapter = pi->adapter;
+
+	if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0)
+		return err;
+
+	dev->real_num_tx_queues = pi->nqsets;
+	set_bit(pi->tx_chan, &adapter->open_device_map);
+	link_start(dev);
+	netif_tx_start_all_queues(dev);
+	return 0;
+}
+
+static int cxgb_close(struct net_device *dev)
+{
+	int ret;
+	struct port_info *pi = netdev_priv(dev);
+	struct adapter *adapter = pi->adapter;
+
+	netif_tx_stop_all_queues(dev);
+	netif_carrier_off(dev);
+	ret = t4_enable_vi(adapter, 0, pi->viid, false, false);
+
+	clear_bit(pi->tx_chan, &adapter->open_device_map);
+
+	if (!adapter->open_device_map)
+		cxgb_down(adapter);
+	return 0;
+}
+
+static struct net_device_stats *cxgb_get_stats(struct net_device *dev)
+{
+	struct port_stats stats;
+	struct port_info *p = netdev_priv(dev);
+	struct adapter *adapter = p->adapter;
+	struct net_device_stats *ns = &dev->stats;
+
+	spin_lock(&adapter->stats_lock);
+	t4_get_port_stats(adapter, p->tx_chan, &stats);
+	spin_unlock(&adapter->stats_lock);
+
+	ns->tx_bytes   = stats.tx_octets;
+	ns->tx_packets = stats.tx_frames;
+	ns->rx_bytes   = stats.rx_octets;
+	ns->rx_packets = stats.rx_frames;
+	ns->multicast  = stats.rx_mcast_frames;
+
+	/* detailed rx_errors */
+	ns->rx_length_errors = stats.rx_jabber + stats.rx_too_long +
+			       stats.rx_runt;
+	ns->rx_over_errors   = 0;
+	ns->rx_crc_errors    = stats.rx_fcs_err;
+	ns->rx_frame_errors  = stats.rx_symbol_err;
+	ns->rx_fifo_errors   = stats.rx_ovflow0 + stats.rx_ovflow1 +
+			       stats.rx_ovflow2 + stats.rx_ovflow3 +
+			       stats.rx_trunc0 + stats.rx_trunc1 +
+			       stats.rx_trunc2 + stats.rx_trunc3;
+	ns->rx_missed_errors = 0;
+
+	/* detailed tx_errors */
+	ns->tx_aborted_errors   = 0;
+	ns->tx_carrier_errors   = 0;
+	ns->tx_fifo_errors      = 0;
+	ns->tx_heartbeat_errors = 0;
+	ns->tx_window_errors    = 0;
+
+	ns->tx_errors = stats.tx_error_frames;
+	ns->rx_errors = stats.rx_symbol_err + stats.rx_fcs_err +
+		ns->rx_length_errors + stats.rx_len_err + ns->rx_fifo_errors;
+	return ns;
+}
+
+static int cxgb_ioctl(struct net_device *dev, struct ifreq *req, int cmd)
+{
+	int ret = 0, prtad, devad;
+	struct port_info *pi = netdev_priv(dev);
+	struct mii_ioctl_data *data = (struct mii_ioctl_data *)&req->ifr_data;
+
+	switch (cmd) {
+	case SIOCGMIIPHY:
+		if (pi->mdio_addr < 0)
+			return -EOPNOTSUPP;
+		data->phy_id = pi->mdio_addr;
+		break;
+	case SIOCGMIIREG:
+	case SIOCSMIIREG:
+		if (mdio_phy_id_is_c45(data->phy_id)) {
+			prtad = mdio_phy_id_prtad(data->phy_id);
+			devad = mdio_phy_id_devad(data->phy_id);
+		} else if (data->phy_id < 32) {
+			prtad = data->phy_id;
+			devad = 0;
+			data->reg_num &= 0x1f;
+		} else
+			return -EINVAL;
+
+		if (cmd == SIOCGMIIREG)
+			ret = t4_mdio_rd(pi->adapter, 0, prtad, devad,
+					 data->reg_num, &data->val_out);
+		else
+			ret = t4_mdio_wr(pi->adapter, 0, prtad, devad,
+					 data->reg_num, data->val_in);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+	return ret;
+}
+
+static void cxgb_set_rxmode(struct net_device *dev)
+{
+	/* unfortunately we can't return errors to the stack */
+	set_rxmode(dev, -1, false);
+}
+
+static int cxgb_change_mtu(struct net_device *dev, int new_mtu)
+{
+	int ret;
+	struct port_info *pi = netdev_priv(dev);
+
+	if (new_mtu < 81 || new_mtu > MAX_MTU)         /* accommodate SACK */
+		return -EINVAL;
+	ret = t4_set_rxmode(pi->adapter, 0, pi->viid, new_mtu, -1, -1, -1,
+			    true);
+	if (!ret)
+		dev->mtu = new_mtu;
+	return ret;
+}
+
+static int cxgb_set_mac_addr(struct net_device *dev, void *p)
+{
+	int ret;
+	struct sockaddr *addr = p;
+	struct port_info *pi = netdev_priv(dev);
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EINVAL;
+
+	ret = t4_change_mac(pi->adapter, 0, pi->viid, pi->xact_addr_filt,
+			    addr->sa_data, true, true);
+	if (ret < 0)
+		return ret;
+
+	memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
+	pi->xact_addr_filt = ret;
+	return 0;
+}
+
+static void vlan_rx_register(struct net_device *dev, struct vlan_group *grp)
+{
+	struct port_info *pi = netdev_priv(dev);
+
+	pi->vlan_grp = grp;
+	t4_set_vlan_accel(pi->adapter, 1 << pi->tx_chan, grp != NULL);
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void cxgb_netpoll(struct net_device *dev)
+{
+	struct port_info *pi = netdev_priv(dev);
+	struct adapter *adap = pi->adapter;
+
+	if (adap->flags & USING_MSIX) {
+		int i;
+		struct sge_eth_rxq *rx = &adap->sge.ethrxq[pi->first_qset];
+
+		for (i = pi->nqsets; i; i--, rx++)
+			t4_sge_intr_msix(0, &rx->rspq);
+	} else
+		t4_intr_handler(adap)(0, adap);
+}
+#endif
+
+static const struct net_device_ops cxgb4_netdev_ops = {
+	.ndo_open             = cxgb_open,
+	.ndo_stop             = cxgb_close,
+	.ndo_start_xmit       = t4_eth_xmit,
+	.ndo_get_stats        = cxgb_get_stats,
+	.ndo_set_rx_mode      = cxgb_set_rxmode,
+	.ndo_set_mac_address  = cxgb_set_mac_addr,
+	.ndo_validate_addr    = eth_validate_addr,
+	.ndo_do_ioctl         = cxgb_ioctl,
+	.ndo_change_mtu       = cxgb_change_mtu,
+	.ndo_vlan_rx_register = vlan_rx_register,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+	.ndo_poll_controller  = cxgb_netpoll,
+#endif
+};
+
+void t4_fatal_err(struct adapter *adap)
+{
+	t4_set_reg_field(adap, SGE_CONTROL, GLOBALENABLE, 0);
+	t4_intr_disable(adap);
+	dev_alert(adap->pdev_dev, "encountered fatal error, adapter stopped\n");
+}
+
+static void setup_memwin(struct adapter *adap)
+{
+	u32 bar0;
+
+	bar0 = pci_resource_start(adap->pdev, 0);  /* truncation intentional */
+	t4_write_reg(adap, PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 0),
+		     (bar0 + MEMWIN0_BASE) | BIR(0) |
+		     WINDOW(ilog2(MEMWIN0_APERTURE) - 10));
+	t4_write_reg(adap, PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 1),
+		     (bar0 + MEMWIN1_BASE) | BIR(0) |
+		     WINDOW(ilog2(MEMWIN1_APERTURE) - 10));
+	t4_write_reg(adap, PCIE_MEM_ACCESS_REG(PCIE_MEM_ACCESS_BASE_WIN, 2),
+		     (bar0 + MEMWIN2_BASE) | BIR(0) |
+		     WINDOW(ilog2(MEMWIN2_APERTURE) - 10));
+}
+
+/*
+ * Max # of ATIDs.  The absolute HW max is 16K but we keep it lower.
+ */
+#define MAX_ATIDS 8192U
+
+/*
+ * Phase 0 of initialization: contact FW, obtain config, perform basic init.
+ */
+static int adap_init0(struct adapter *adap)
+{
+	int ret;
+	u32 v, port_vec;
+	enum dev_state state;
+	u32 params[7], val[7];
+	struct fw_caps_config_cmd c;
+
+	ret = t4_check_fw_version(adap);
+	if (ret == -EINVAL || ret > 0) {
+		if (upgrade_fw(adap) >= 0)             /* recache FW version */
+			ret = t4_check_fw_version(adap);
+	}
+	if (ret < 0)
+		return ret;
+
+	/* contact FW, request master */
+	ret = t4_fw_hello(adap, 0, 0, MASTER_MUST, &state);
+	if (ret < 0) {
+		dev_err(adap->pdev_dev, "could not connect to FW, error %d\n",
+			ret);
+		return ret;
+	}
+
+	/* reset device */
+	ret = t4_fw_reset(adap, 0, PIORSTMODE | PIORST);
+	if (ret < 0)
+		goto bye;
+
+	/* get device capabilities */
+	memset(&c, 0, sizeof(c));
+	c.op_to_write = htonl(FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+			      FW_CMD_REQUEST | FW_CMD_READ);
+	c.retval_len16 = htonl(FW_LEN16(c));
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+	if (ret < 0)
+		goto bye;
+
+	/* select capabilities we'll be using */
+	if (c.niccaps & htons(FW_CAPS_CONFIG_NIC_VM)) {
+		if (!vf_acls)
+			c.niccaps ^= htons(FW_CAPS_CONFIG_NIC_VM);
+		else
+			c.niccaps = htons(FW_CAPS_CONFIG_NIC_VM);
+	} else if (vf_acls) {
+		dev_err(adap->pdev_dev, "virtualization ACLs not supported");
+		goto bye;
+	}
+	c.op_to_write = htonl(FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+			      FW_CMD_REQUEST | FW_CMD_WRITE);
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), NULL);
+	if (ret < 0)
+		goto bye;
+
+	ret = t4_config_glbl_rss(adap, 0,
+				 FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL,
+				 FW_RSS_GLB_CONFIG_CMD_TNLMAPEN |
+				 FW_RSS_GLB_CONFIG_CMD_TNLALLLKP);
+	if (ret < 0)
+		goto bye;
+
+	ret = t4_cfg_pfvf(adap, 0, 0, 0, 64, 64, 64, 0, 0, 4, 0xf, 0xf, 16,
+			  FW_CMD_CAP_PF, FW_CMD_CAP_PF);
+	if (ret < 0)
+		goto bye;
+
+	for (v = 0; v < SGE_NTIMERS - 1; v++)
+		adap->sge.timer_val[v] = min(intr_holdoff[v], MAX_SGE_TIMERVAL);
+	adap->sge.timer_val[SGE_NTIMERS - 1] = MAX_SGE_TIMERVAL;
+	adap->sge.counter_val[0] = 1;
+	for (v = 1; v < SGE_NCOUNTERS; v++)
+		adap->sge.counter_val[v] = min(intr_cnt[v - 1],
+					       THRESHOLD_3_MASK);
+	t4_sge_init(adap);
+
+	/* get basic stuff going */
+	ret = t4_early_init(adap, 0);
+	if (ret < 0)
+		goto bye;
+
+#define FW_PARAM_DEV(param) \
+	(FW_PARAMS_MNEM(FW_PARAMS_MNEM_DEV) | \
+	 FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DEV_##param))
+
+#define FW_PARAM_PFVF(param) \
+	(FW_PARAMS_MNEM(FW_PARAMS_MNEM_PFVF) | \
+	 FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_PFVF_##param))
+
+	params[0] = FW_PARAM_DEV(PORTVEC);
+	params[1] = FW_PARAM_PFVF(L2T_START);
+	params[2] = FW_PARAM_PFVF(L2T_END);
+	params[3] = FW_PARAM_PFVF(FILTER_START);
+	params[4] = FW_PARAM_PFVF(FILTER_END);
+	ret = t4_query_params(adap, 0, 0, 0, 5, params, val);
+	if (ret < 0)
+		goto bye;
+	port_vec = val[0];
+	adap->tids.ftid_base = val[3];
+	adap->tids.nftids = val[4] - val[3] + 1;
+
+	if (c.ofldcaps) {
+		/* query offload-related parameters */
+		params[0] = FW_PARAM_DEV(NTID);
+		params[1] = FW_PARAM_PFVF(SERVER_START);
+		params[2] = FW_PARAM_PFVF(SERVER_END);
+		params[3] = FW_PARAM_PFVF(TDDP_START);
+		params[4] = FW_PARAM_PFVF(TDDP_END);
+		params[5] = FW_PARAM_DEV(FLOWC_BUFFIFO_SZ);
+		ret = t4_query_params(adap, 0, 0, 0, 6, params, val);
+		if (ret < 0)
+			goto bye;
+		adap->tids.ntids = val[0];
+		adap->tids.natids = min(adap->tids.ntids / 2, MAX_ATIDS);
+		adap->tids.stid_base = val[1];
+		adap->tids.nstids = val[2] - val[1] + 1;
+		adap->vres.ddp.start = val[3];
+		adap->vres.ddp.size = val[4] - val[3] + 1;
+		adap->params.ofldq_wr_cred = val[5];
+		adap->params.offload = 1;
+	}
+	if (c.rdmacaps) {
+		params[0] = FW_PARAM_PFVF(STAG_START);
+		params[1] = FW_PARAM_PFVF(STAG_END);
+		params[2] = FW_PARAM_PFVF(RQ_START);
+		params[3] = FW_PARAM_PFVF(RQ_END);
+		params[4] = FW_PARAM_PFVF(PBL_START);
+		params[5] = FW_PARAM_PFVF(PBL_END);
+		ret = t4_query_params(adap, 0, 0, 0, 6, params, val);
+		if (ret < 0)
+			goto bye;
+		adap->vres.stag.start = val[0];
+		adap->vres.stag.size = val[1] - val[0] + 1;
+		adap->vres.rq.start = val[2];
+		adap->vres.rq.size = val[3] - val[2] + 1;
+		adap->vres.pbl.start = val[4];
+		adap->vres.pbl.size = val[5] - val[4] + 1;
+	}
+	if (c.iscsicaps) {
+		params[0] = FW_PARAM_PFVF(ISCSI_START);
+		params[1] = FW_PARAM_PFVF(ISCSI_END);
+		ret = t4_query_params(adap, 0, 0, 0, 2, params, val);
+		if (ret < 0)
+			goto bye;
+		adap->vres.iscsi.start = val[0];
+		adap->vres.iscsi.size = val[1] - val[0] + 1;
+	}
+#undef FW_PARAM_PFVF
+#undef FW_PARAM_DEV
+
+	adap->params.nports = hweight32(port_vec);
+	adap->params.portvec = port_vec;
+	adap->flags |= FW_OK;
+
+	/* These are finalized by FW initialization, load their values now */
+	v = t4_read_reg(adap, TP_TIMER_RESOLUTION);
+	adap->params.tp.tre = TIMERRESOLUTION_GET(v);
+	t4_read_mtu_tbl(adap, adap->params.mtus, NULL);
+	t4_load_mtus(adap, adap->params.mtus, adap->params.a_wnd,
+		     adap->params.b_wnd);
+
+	/* tweak some settings */
+	t4_write_reg(adap, TP_SHIFT_CNT, 0x64f8849);
+	t4_write_reg(adap, ULP_RX_TDDP_PSZ, HPZ0(PAGE_SHIFT - 12));
+	t4_write_reg(adap, TP_PIO_ADDR, TP_INGRESS_CONFIG);
+	v = t4_read_reg(adap, TP_PIO_DATA);
+	t4_write_reg(adap, TP_PIO_DATA, v & ~CSUM_HAS_PSEUDO_HDR);
+	setup_memwin(adap);
+	return 0;
+
+	/*
+	 * If a command timed out or failed with EIO FW does not operate within
+	 * its spec or something catastrophic happened to HW/FW, stop issuing
+	 * commands.
+	 */
+bye:	if (ret != -ETIMEDOUT && ret != -EIO)
+		t4_fw_bye(adap, 0);
+	return ret;
+}
+
+static inline bool is_10g_port(const struct link_config *lc)
+{
+	return (lc->supported & FW_PORT_CAP_SPEED_10G) != 0;
+}
+
+static inline void init_rspq(struct sge_rspq *q, u8 timer_idx, u8 pkt_cnt_idx,
+			     unsigned int size, unsigned int iqe_size)
+{
+	q->intr_params = QINTR_TIMER_IDX(timer_idx) |
+			 (pkt_cnt_idx < SGE_NCOUNTERS ? QINTR_CNT_EN : 0);
+	q->pktcnt_idx = pkt_cnt_idx < SGE_NCOUNTERS ? pkt_cnt_idx : 0;
+	q->iqe_len = iqe_size;
+	q->size = size;
+}
+
+/*
+ * Perform default configuration of DMA queues depending on the number and type
+ * of ports we found and the number of available CPUs.  Most settings can be
+ * modified by the admin prior to actual use.
+ */
+static void __devinit cfg_queues(struct adapter *adap)
+{
+	struct sge *s = &adap->sge;
+	int i, q10g = 0, n10g = 0, qidx = 0;
+
+	for_each_port(adap, i)
+		n10g += is_10g_port(&adap2pinfo(adap, i)->link_cfg);
+
+	/*
+	 * We default to 1 queue per non-10G port and up to # of cores queues
+	 * per 10G port.
+	 */
+	if (n10g)
+		q10g = (MAX_ETH_QSETS - (adap->params.nports - n10g)) / n10g;
+	if (q10g > num_online_cpus())
+		q10g = num_online_cpus();
+
+	for_each_port(adap, i) {
+		struct port_info *pi = adap2pinfo(adap, i);
+
+		pi->first_qset = qidx;
+		pi->nqsets = is_10g_port(&pi->link_cfg) ? q10g : 1;
+		qidx += pi->nqsets;
+	}
+
+	s->ethqsets = qidx;
+	s->max_ethqsets = qidx;   /* MSI-X may lower it later */
+
+	if (is_offload(adap)) {
+		/*
+		 * For offload we use 1 queue/channel if all ports are up to 1G,
+		 * otherwise we divide all available queues amongst the channels
+		 * capped by the number of available cores.
+		 */
+		if (n10g) {
+			i = min_t(int, ARRAY_SIZE(s->ofldrxq),
+				  num_online_cpus());
+			s->ofldqsets = roundup(i, adap->params.nports);
+		} else
+			s->ofldqsets = adap->params.nports;
+		/* For RDMA one Rx queue per channel suffices */
+		s->rdmaqs = adap->params.nports;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(s->ethrxq); i++) {
+		struct sge_eth_rxq *r = &s->ethrxq[i];
+
+		init_rspq(&r->rspq, 0, 0, 1024, 64);
+		r->fl.size = 72;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(s->ethtxq); i++)
+		s->ethtxq[i].q.size = 1024;
+
+	for (i = 0; i < ARRAY_SIZE(s->ctrlq); i++)
+		s->ctrlq[i].q.size = 512;
+
+	for (i = 0; i < ARRAY_SIZE(s->ofldtxq); i++)
+		s->ofldtxq[i].q.size = 1024;
+
+	for (i = 0; i < ARRAY_SIZE(s->ofldrxq); i++) {
+		struct sge_ofld_rxq *r = &s->ofldrxq[i];
+
+		init_rspq(&r->rspq, 0, 0, 1024, 64);
+		r->rspq.uld = CXGB4_ULD_ISCSI;
+		r->fl.size = 72;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(s->rdmarxq); i++) {
+		struct sge_ofld_rxq *r = &s->rdmarxq[i];
+
+		init_rspq(&r->rspq, 0, 0, 511, 64);
+		r->rspq.uld = CXGB4_ULD_RDMA;
+		r->fl.size = 72;
+	}
+
+	init_rspq(&s->fw_evtq, 6, 0, 512, 64);
+	init_rspq(&s->intrq, 6, 0, 2 * MAX_INGQ, 64);
+}
+
+/*
+ * Reduce the number of Ethernet queues across all ports to at most n.
+ * n provides at least one queue per port.
+ */
+static void __devinit reduce_ethqs(struct adapter *adap, int n)
+{
+	int i;
+	struct port_info *pi;
+
+	while (n < adap->sge.ethqsets)
+		for_each_port(adap, i) {
+			pi = adap2pinfo(adap, i);
+			if (pi->nqsets > 1) {
+				pi->nqsets--;
+				adap->sge.ethqsets--;
+				if (adap->sge.ethqsets <= n)
+					break;
+			}
+		}
+
+	n = 0;
+	for_each_port(adap, i) {
+		pi = adap2pinfo(adap, i);
+		pi->first_qset = n;
+		n += pi->nqsets;
+	}
+}
+
+/* 2 MSI-X vectors needed for the FW queue and non-data interrupts */
+#define EXTRA_VECS 2
+
+static int __devinit enable_msix(struct adapter *adap)
+{
+	int ofld_need = 0;
+	int i, err, want, need;
+	struct sge *s = &adap->sge;
+	unsigned int nchan = adap->params.nports;
+	struct msix_entry entries[MAX_INGQ + 1];
+
+	for (i = 0; i < ARRAY_SIZE(entries); ++i)
+		entries[i].entry = i;
+
+	want = s->max_ethqsets + EXTRA_VECS;
+	if (is_offload(adap)) {
+		want += s->rdmaqs + s->ofldqsets;
+		/* need nchan for each possible ULD */
+		ofld_need = 2 * nchan;
+	}
+	need = adap->params.nports + EXTRA_VECS + ofld_need;
+
+	while ((err = pci_enable_msix(adap->pdev, entries, want)) >= need)
+		want = err;
+
+	if (!err) {
+		/*
+		 * Distribute available vectors to the various queue groups.
+		 * Every group gets its minimum requirement and NIC gets top
+		 * priority for leftovers.
+		 */
+		i = want - EXTRA_VECS - ofld_need;
+		if (i < s->max_ethqsets) {
+			s->max_ethqsets = i;
+			if (i < s->ethqsets)
+				reduce_ethqs(adap, i);
+		}
+		if (is_offload(adap)) {
+			i = want - EXTRA_VECS - s->max_ethqsets;
+			i -= ofld_need - nchan;
+			s->ofldqsets = (i / nchan) * nchan;  /* round down */
+		}
+		for (i = 0; i < want; ++i)
+			adap->msix_info[i].vec = entries[i].vector;
+	} else if (err > 0)
+		dev_info(adap->pdev_dev,
+			 "only %d MSI-X vectors left, not using MSI-X\n", err);
+	return err;
+}
+
+#undef EXTRA_VECS
+
+static void __devinit print_port_info(struct adapter *adap)
+{
+	static const char *base[] = {
+		"R", "KX4", "T", "KX", "T", "KR", "CX4"
+	};
+
+	int i;
+	char buf[80];
+
+	for_each_port(adap, i) {
+		struct net_device *dev = adap->port[i];
+		const struct port_info *pi = netdev_priv(dev);
+		char *bufp = buf;
+
+		if (!test_bit(i, &adap->registered_device_map))
+			continue;
+
+		if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_100M)
+			bufp += sprintf(bufp, "100/");
+		if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_1G)
+			bufp += sprintf(bufp, "1000/");
+		if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_10G)
+			bufp += sprintf(bufp, "10G/");
+		if (bufp != buf)
+			--bufp;
+		sprintf(bufp, "BASE-%s", base[pi->port_type]);
+
+		netdev_info(dev, "Chelsio %s rev %d %s %sNIC PCIe x%d%s\n",
+			    adap->params.vpd.id, adap->params.rev,
+			    buf, is_offload(adap) ? "R" : "",
+			    adap->params.pci.width,
+			    (adap->flags & USING_MSIX) ? " MSI-X" :
+			    (adap->flags & USING_MSI) ? " MSI" : "");
+		if (adap->name == dev->name)
+			netdev_info(dev, "S/N: %s, E/C: %s\n",
+				    adap->params.vpd.sn, adap->params.vpd.ec);
+	}
+}
+
+#define VLAN_FEAT (NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO | NETIF_F_TSO6 |\
+		   NETIF_F_IPV6_CSUM | NETIF_F_HIGHDMA)
+
+static int __devinit init_one(struct pci_dev *pdev,
+			      const struct pci_device_id *ent)
+{
+	int func, i, err;
+	struct port_info *pi;
+	unsigned int highdma = 0;
+	struct adapter *adapter = NULL;
+
+	printk_once(KERN_INFO "%s - version %s\n", DRV_DESC, DRV_VERSION);
+
+	err = pci_request_regions(pdev, KBUILD_MODNAME);
+	if (err) {
+		/* Just info, some other driver may have claimed the device. */
+		dev_info(&pdev->dev, "cannot obtain PCI resources\n");
+		return err;
+	}
+
+	/* We control everything through PF 0 */
+	func = PCI_FUNC(pdev->devfn);
+	if (func > 0)
+		goto sriov;
+
+	err = pci_enable_device(pdev);
+	if (err) {
+		dev_err(&pdev->dev, "cannot enable PCI device\n");
+		goto out_release_regions;
+	}
+
+	if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) {
+		highdma = NETIF_F_HIGHDMA;
+		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+		if (err) {
+			dev_err(&pdev->dev, "unable to obtain 64-bit DMA for "
+				"coherent allocations\n");
+			goto out_disable_device;
+		}
+	} else {
+		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+		if (err) {
+			dev_err(&pdev->dev, "no usable DMA configuration\n");
+			goto out_disable_device;
+		}
+	}
+
+	pci_enable_pcie_error_reporting(pdev);
+	pci_set_master(pdev);
+	pci_save_state(pdev);
+
+	adapter = kzalloc(sizeof(*adapter), GFP_KERNEL);
+	if (!adapter) {
+		err = -ENOMEM;
+		goto out_disable_device;
+	}
+
+	adapter->regs = pci_ioremap_bar(pdev, 0);
+	if (!adapter->regs) {
+		dev_err(&pdev->dev, "cannot map device registers\n");
+		err = -ENOMEM;
+		goto out_free_adapter;
+	}
+
+	adapter->pdev = pdev;
+	adapter->pdev_dev = &pdev->dev;
+	adapter->name = pci_name(pdev);
+	adapter->msg_enable = dflt_msg_enable;
+	memset(adapter->chan_map, 0xff, sizeof(adapter->chan_map));
+
+	spin_lock_init(&adapter->stats_lock);
+	spin_lock_init(&adapter->tid_release_lock);
+
+	INIT_WORK(&adapter->tid_release_task, process_tid_release_list);
+
+	err = t4_prep_adapter(adapter);
+	if (err)
+		goto out_unmap_bar;
+	err = adap_init0(adapter);
+	if (err)
+		goto out_unmap_bar;
+
+	for_each_port(adapter, i) {
+		struct net_device *netdev;
+
+		netdev = alloc_etherdev_mq(sizeof(struct port_info),
+					   MAX_ETH_QSETS);
+		if (!netdev) {
+			err = -ENOMEM;
+			goto out_free_dev;
+		}
+
+		SET_NETDEV_DEV(netdev, &pdev->dev);
+
+		adapter->port[i] = netdev;
+		pi = netdev_priv(netdev);
+		pi->adapter = adapter;
+		pi->xact_addr_filt = -1;
+		pi->rx_offload = RX_CSO;
+		pi->port_id = i;
+		netif_carrier_off(netdev);
+		netif_tx_stop_all_queues(netdev);
+		netdev->irq = pdev->irq;
+
+		netdev->features |= NETIF_F_SG | NETIF_F_TSO | NETIF_F_TSO6;
+		netdev->features |= NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM;
+		netdev->features |= NETIF_F_GRO | highdma;
+		netdev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX;
+		netdev->vlan_features = netdev->features & VLAN_FEAT;
+
+		netdev->netdev_ops = &cxgb4_netdev_ops;
+		SET_ETHTOOL_OPS(netdev, &cxgb_ethtool_ops);
+	}
+
+	pci_set_drvdata(pdev, adapter);
+
+	if (adapter->flags & FW_OK) {
+		err = t4_port_init(adapter, 0, 0, 0);
+		if (err)
+			goto out_free_dev;
+	}
+
+	/*
+	 * Configure queues and allocate tables now, they can be needed as
+	 * soon as the first register_netdev completes.
+	 */
+	cfg_queues(adapter);
+
+	adapter->l2t = t4_init_l2t();
+	if (!adapter->l2t) {
+		/* We tolerate a lack of L2T, giving up some functionality */
+		dev_warn(&pdev->dev, "could not allocate L2T, continuing\n");
+		adapter->params.offload = 0;
+	}
+
+	if (is_offload(adapter) && tid_init(&adapter->tids) < 0) {
+		dev_warn(&pdev->dev, "could not allocate TID table, "
+			 "continuing\n");
+		adapter->params.offload = 0;
+	}
+
+	/*
+	 * The card is now ready to go.  If any errors occur during device
+	 * registration we do not fail the whole card but rather proceed only
+	 * with the ports we manage to register successfully.  However we must
+	 * register at least one net device.
+	 */
+	for_each_port(adapter, i) {
+		err = register_netdev(adapter->port[i]);
+		if (err)
+			dev_warn(&pdev->dev,
+				 "cannot register net device %s, skipping\n",
+				 adapter->port[i]->name);
+		else {
+			/*
+			 * Change the name we use for messages to the name of
+			 * the first successfully registered interface.
+			 */
+			if (!adapter->registered_device_map)
+				adapter->name = adapter->port[i]->name;
+
+			__set_bit(i, &adapter->registered_device_map);
+			adapter->chan_map[adap2pinfo(adapter, i)->tx_chan] = i;
+		}
+	}
+	if (!adapter->registered_device_map) {
+		dev_err(&pdev->dev, "could not register any net devices\n");
+		goto out_free_dev;
+	}
+
+	if (cxgb4_debugfs_root) {
+		adapter->debugfs_root = debugfs_create_dir(pci_name(pdev),
+							   cxgb4_debugfs_root);
+		setup_debugfs(adapter);
+	}
+
+	/* See what interrupts we'll be using */
+	if (msi > 1 && enable_msix(adapter) == 0)
+		adapter->flags |= USING_MSIX;
+	else if (msi > 0 && pci_enable_msi(pdev) == 0)
+		adapter->flags |= USING_MSI;
+
+	if (is_offload(adapter))
+		attach_ulds(adapter);
+
+	print_port_info(adapter);
+
+sriov:
+#ifdef CONFIG_PCI_IOV
+	if (func < ARRAY_SIZE(num_vf) && num_vf[func] > 0)
+		if (pci_enable_sriov(pdev, num_vf[func]) == 0)
+			dev_info(&pdev->dev,
+				 "instantiated %u virtual functions\n",
+				 num_vf[func]);
+#endif
+	return 0;
+
+ out_free_dev:
+	t4_free_mem(adapter->tids.tid_tab);
+	t4_free_mem(adapter->l2t);
+	for_each_port(adapter, i)
+		if (adapter->port[i])
+			free_netdev(adapter->port[i]);
+	if (adapter->flags & FW_OK)
+		t4_fw_bye(adapter, 0);
+ out_unmap_bar:
+	iounmap(adapter->regs);
+ out_free_adapter:
+	kfree(adapter);
+ out_disable_device:
+	pci_disable_pcie_error_reporting(pdev);
+	pci_disable_device(pdev);
+ out_release_regions:
+	pci_release_regions(pdev);
+	pci_set_drvdata(pdev, NULL);
+	return err;
+}
+
+static void __devexit remove_one(struct pci_dev *pdev)
+{
+	struct adapter *adapter = pci_get_drvdata(pdev);
+
+	pci_disable_sriov(pdev);
+
+	if (adapter) {
+		int i;
+
+		if (is_offload(adapter))
+			detach_ulds(adapter);
+
+		for_each_port(adapter, i)
+			if (test_bit(i, &adapter->registered_device_map))
+				unregister_netdev(adapter->port[i]);
+
+		if (adapter->debugfs_root)
+			debugfs_remove_recursive(adapter->debugfs_root);
+
+		t4_sge_stop(adapter);
+		t4_free_sge_resources(adapter);
+		t4_free_mem(adapter->l2t);
+		t4_free_mem(adapter->tids.tid_tab);
+		disable_msi(adapter);
+
+		for_each_port(adapter, i)
+			if (adapter->port[i])
+				free_netdev(adapter->port[i]);
+
+		if (adapter->flags & FW_OK)
+			t4_fw_bye(adapter, 0);
+		iounmap(adapter->regs);
+		kfree(adapter);
+		pci_disable_pcie_error_reporting(pdev);
+		pci_disable_device(pdev);
+		pci_release_regions(pdev);
+		pci_set_drvdata(pdev, NULL);
+	} else if (PCI_FUNC(pdev->devfn) > 0)
+		pci_release_regions(pdev);
+}
+
+static struct pci_driver cxgb4_driver = {
+	.name     = KBUILD_MODNAME,
+	.id_table = cxgb4_pci_tbl,
+	.probe    = init_one,
+	.remove   = __devexit_p(remove_one),
+};
+
+static int __init cxgb4_init_module(void)
+{
+	int ret;
+
+	/* Debugfs support is optional, just warn if this fails */
+	cxgb4_debugfs_root = debugfs_create_dir(KBUILD_MODNAME, NULL);
+	if (!cxgb4_debugfs_root)
+		pr_warning("could not create debugfs entry, continuing\n");
+
+	ret = pci_register_driver(&cxgb4_driver);
+	if (ret < 0)
+		debugfs_remove(cxgb4_debugfs_root);
+	return ret;
+}
+
+static void __exit cxgb4_cleanup_module(void)
+{
+	pci_unregister_driver(&cxgb4_driver);
+	debugfs_remove(cxgb4_debugfs_root);  /* NULL ok */
+}
+
+module_init(cxgb4_init_module);
+module_exit(cxgb4_cleanup_module);
diff --git a/drivers/net/cxgb4/cxgb4_uld.h b/drivers/net/cxgb4/cxgb4_uld.h
new file mode 100644
index 000000000000..5b98546ac92d
--- /dev/null
+++ b/drivers/net/cxgb4/cxgb4_uld.h
@@ -0,0 +1,239 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_OFLD_H
+#define __CXGB4_OFLD_H
+
+#include <linux/cache.h>
+#include <linux/spinlock.h>
+#include <linux/skbuff.h>
+#include <asm/atomic.h>
+
+/* CPL message priority levels */
+enum {
+	CPL_PRIORITY_DATA     = 0,  /* data messages */
+	CPL_PRIORITY_SETUP    = 1,  /* connection setup messages */
+	CPL_PRIORITY_TEARDOWN = 0,  /* connection teardown messages */
+	CPL_PRIORITY_LISTEN   = 1,  /* listen start/stop messages */
+	CPL_PRIORITY_ACK      = 1,  /* RX ACK messages */
+	CPL_PRIORITY_CONTROL  = 1   /* control messages */
+};
+
+#define INIT_TP_WR(w, tid) do { \
+	(w)->wr.wr_hi = htonl(FW_WR_OP(FW_TP_WR) | \
+			      FW_WR_IMMDLEN(sizeof(*w) - sizeof(w->wr))); \
+	(w)->wr.wr_mid = htonl(FW_WR_LEN16(DIV_ROUND_UP(sizeof(*w), 16)) | \
+			       FW_WR_FLOWID(tid)); \
+	(w)->wr.wr_lo = cpu_to_be64(0); \
+} while (0)
+
+#define INIT_TP_WR_CPL(w, cpl, tid) do { \
+	INIT_TP_WR(w, tid); \
+	OPCODE_TID(w) = htonl(MK_OPCODE_TID(cpl, tid)); \
+} while (0)
+
+#define INIT_ULPTX_WR(w, wrlen, atomic, tid) do { \
+	(w)->wr.wr_hi = htonl(FW_WR_OP(FW_ULPTX_WR) | FW_WR_ATOMIC(atomic)); \
+	(w)->wr.wr_mid = htonl(FW_WR_LEN16(DIV_ROUND_UP(wrlen, 16)) | \
+			       FW_WR_FLOWID(tid)); \
+	(w)->wr.wr_lo = cpu_to_be64(0); \
+} while (0)
+
+/* Special asynchronous notification message */
+#define CXGB4_MSG_AN ((void *)1)
+
+struct serv_entry {
+	void *data;
+};
+
+union aopen_entry {
+	void *data;
+	union aopen_entry *next;
+};
+
+/*
+ * Holds the size, base address, free list start, etc of the TID, server TID,
+ * and active-open TID tables.  The tables themselves are allocated dynamically.
+ */
+struct tid_info {
+	void **tid_tab;
+	unsigned int ntids;
+
+	struct serv_entry *stid_tab;
+	unsigned long *stid_bmap;
+	unsigned int nstids;
+	unsigned int stid_base;
+
+	union aopen_entry *atid_tab;
+	unsigned int natids;
+
+	unsigned int nftids;
+	unsigned int ftid_base;
+
+	spinlock_t atid_lock ____cacheline_aligned_in_smp;
+	union aopen_entry *afree;
+	unsigned int atids_in_use;
+
+	spinlock_t stid_lock;
+	unsigned int stids_in_use;
+
+	atomic_t tids_in_use;
+};
+
+static inline void *lookup_tid(const struct tid_info *t, unsigned int tid)
+{
+	return tid < t->ntids ? t->tid_tab[tid] : NULL;
+}
+
+static inline void *lookup_atid(const struct tid_info *t, unsigned int atid)
+{
+	return atid < t->natids ? t->atid_tab[atid].data : NULL;
+}
+
+static inline void *lookup_stid(const struct tid_info *t, unsigned int stid)
+{
+	stid -= t->stid_base;
+	return stid < t->nstids ? t->stid_tab[stid].data : NULL;
+}
+
+static inline void cxgb4_insert_tid(struct tid_info *t, void *data,
+				    unsigned int tid)
+{
+	t->tid_tab[tid] = data;
+	atomic_inc(&t->tids_in_use);
+}
+
+int cxgb4_alloc_atid(struct tid_info *t, void *data);
+int cxgb4_alloc_stid(struct tid_info *t, int family, void *data);
+void cxgb4_free_atid(struct tid_info *t, unsigned int atid);
+void cxgb4_free_stid(struct tid_info *t, unsigned int stid, int family);
+void cxgb4_remove_tid(struct tid_info *t, unsigned int qid, unsigned int tid);
+void cxgb4_queue_tid_release(struct tid_info *t, unsigned int chan,
+			     unsigned int tid);
+
+struct in6_addr;
+
+int cxgb4_create_server(const struct net_device *dev, unsigned int stid,
+			__be32 sip, __be16 sport, unsigned int queue);
+int cxgb4_create_server6(const struct net_device *dev, unsigned int stid,
+			 const struct in6_addr *sip, __be16 sport,
+			 unsigned int queue);
+
+static inline void set_wr_txq(struct sk_buff *skb, int prio, int queue)
+{
+	skb_set_queue_mapping(skb, (queue << 1) | prio);
+}
+
+enum cxgb4_uld {
+	CXGB4_ULD_RDMA,
+	CXGB4_ULD_ISCSI,
+	CXGB4_ULD_MAX
+};
+
+enum cxgb4_state {
+	CXGB4_STATE_UP,
+	CXGB4_STATE_START_RECOVERY,
+	CXGB4_STATE_DOWN,
+	CXGB4_STATE_DETACH
+};
+
+struct pci_dev;
+struct l2t_data;
+struct net_device;
+struct pkt_gl;
+struct tp_tcp_stats;
+
+struct cxgb4_range {
+	unsigned int start;
+	unsigned int size;
+};
+
+struct cxgb4_virt_res {                      /* virtualized HW resources */
+	struct cxgb4_range ddp;
+	struct cxgb4_range iscsi;
+	struct cxgb4_range stag;
+	struct cxgb4_range rq;
+	struct cxgb4_range pbl;
+};
+
+/*
+ * Block of information the LLD provides to ULDs attaching to a device.
+ */
+struct cxgb4_lld_info {
+	struct pci_dev *pdev;                /* associated PCI device */
+	struct l2t_data *l2t;                /* L2 table */
+	struct tid_info *tids;               /* TID table */
+	struct net_device **ports;           /* device ports */
+	const struct cxgb4_virt_res *vr;     /* assorted HW resources */
+	const unsigned short *mtus;          /* MTU table */
+	const unsigned short *rxq_ids;       /* the ULD's Rx queue ids */
+	unsigned short nrxq;                 /* # of Rx queues */
+	unsigned short ntxq;                 /* # of Tx queues */
+	unsigned char nchan:4;               /* # of channels */
+	unsigned char nports:4;              /* # of ports */
+	unsigned char wr_cred;               /* WR 16-byte credits */
+	unsigned char adapter_type;          /* type of adapter */
+	unsigned char fw_api_ver;            /* FW API version */
+	unsigned int fw_vers;                /* FW version */
+	unsigned int iscsi_iolen;            /* iSCSI max I/O length */
+	unsigned short udb_density;          /* # of user DB/page */
+	unsigned short ucq_density;          /* # of user CQs/page */
+	void __iomem *gts_reg;               /* address of GTS register */
+	void __iomem *db_reg;                /* address of kernel doorbell */
+};
+
+struct cxgb4_uld_info {
+	const char *name;
+	void *(*add)(const struct cxgb4_lld_info *p);
+	int (*rx_handler)(void *handle, const __be64 *rsp,
+			  const struct pkt_gl *gl);
+	int (*state_change)(void *handle, enum cxgb4_state new_state);
+};
+
+int cxgb4_register_uld(enum cxgb4_uld type, const struct cxgb4_uld_info *p);
+int cxgb4_unregister_uld(enum cxgb4_uld type);
+int cxgb4_ofld_send(struct net_device *dev, struct sk_buff *skb);
+unsigned int cxgb4_port_chan(const struct net_device *dev);
+unsigned int cxgb4_port_viid(const struct net_device *dev);
+unsigned int cxgb4_port_idx(const struct net_device *dev);
+struct net_device *cxgb4_netdev_by_hwid(struct pci_dev *pdev, unsigned int id);
+unsigned int cxgb4_best_mtu(const unsigned short *mtus, unsigned short mtu,
+			    unsigned int *idx);
+void cxgb4_get_tcp_stats(struct pci_dev *pdev, struct tp_tcp_stats *v4,
+			 struct tp_tcp_stats *v6);
+void cxgb4_iscsi_init(struct net_device *dev, unsigned int tag_mask,
+		      const unsigned int *pgsz_order);
+struct sk_buff *cxgb4_pktgl_to_skb(const struct pkt_gl *gl,
+				   unsigned int skb_len, unsigned int pull_len);
+#endif  /* !__CXGB4_OFLD_H */
diff --git a/drivers/net/cxgb4/l2t.c b/drivers/net/cxgb4/l2t.c
new file mode 100644
index 000000000000..9f96724a133a
--- /dev/null
+++ b/drivers/net/cxgb4/l2t.c
@@ -0,0 +1,624 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <linux/if.h>
+#include <linux/if_vlan.h>
+#include <linux/jhash.h>
+#include <net/neighbour.h>
+#include "cxgb4.h"
+#include "l2t.h"
+#include "t4_msg.h"
+#include "t4fw_api.h"
+
+#define VLAN_NONE 0xfff
+
+/* identifies sync vs async L2T_WRITE_REQs */
+#define F_SYNC_WR    (1 << 12)
+
+enum {
+	L2T_STATE_VALID,      /* entry is up to date */
+	L2T_STATE_STALE,      /* entry may be used but needs revalidation */
+	L2T_STATE_RESOLVING,  /* entry needs address resolution */
+	L2T_STATE_SYNC_WRITE, /* synchronous write of entry underway */
+
+	/* when state is one of the below the entry is not hashed */
+	L2T_STATE_SWITCHING,  /* entry is being used by a switching filter */
+	L2T_STATE_UNUSED      /* entry not in use */
+};
+
+struct l2t_data {
+	rwlock_t lock;
+	atomic_t nfree;             /* number of free entries */
+	struct l2t_entry *rover;    /* starting point for next allocation */
+	struct l2t_entry l2tab[L2T_SIZE];
+};
+
+static inline unsigned int vlan_prio(const struct l2t_entry *e)
+{
+	return e->vlan >> 13;
+}
+
+static inline void l2t_hold(struct l2t_data *d, struct l2t_entry *e)
+{
+	if (atomic_add_return(1, &e->refcnt) == 1)  /* 0 -> 1 transition */
+		atomic_dec(&d->nfree);
+}
+
+/*
+ * To avoid having to check address families we do not allow v4 and v6
+ * neighbors to be on the same hash chain.  We keep v4 entries in the first
+ * half of available hash buckets and v6 in the second.
+ */
+enum {
+	L2T_SZ_HALF = L2T_SIZE / 2,
+	L2T_HASH_MASK = L2T_SZ_HALF - 1
+};
+
+static inline unsigned int arp_hash(const u32 *key, int ifindex)
+{
+	return jhash_2words(*key, ifindex, 0) & L2T_HASH_MASK;
+}
+
+static inline unsigned int ipv6_hash(const u32 *key, int ifindex)
+{
+	u32 xor = key[0] ^ key[1] ^ key[2] ^ key[3];
+
+	return L2T_SZ_HALF + (jhash_2words(xor, ifindex, 0) & L2T_HASH_MASK);
+}
+
+static unsigned int addr_hash(const u32 *addr, int addr_len, int ifindex)
+{
+	return addr_len == 4 ? arp_hash(addr, ifindex) :
+			       ipv6_hash(addr, ifindex);
+}
+
+/*
+ * Checks if an L2T entry is for the given IP/IPv6 address.  It does not check
+ * whether the L2T entry and the address are of the same address family.
+ * Callers ensure an address is only checked against L2T entries of the same
+ * family, something made trivial by the separation of IP and IPv6 hash chains
+ * mentioned above.  Returns 0 if there's a match,
+ */
+static int addreq(const struct l2t_entry *e, const u32 *addr)
+{
+	if (e->v6)
+		return (e->addr[0] ^ addr[0]) | (e->addr[1] ^ addr[1]) |
+		       (e->addr[2] ^ addr[2]) | (e->addr[3] ^ addr[3]);
+	return e->addr[0] ^ addr[0];
+}
+
+static void neigh_replace(struct l2t_entry *e, struct neighbour *n)
+{
+	neigh_hold(n);
+	if (e->neigh)
+		neigh_release(e->neigh);
+	e->neigh = n;
+}
+
+/*
+ * Write an L2T entry.  Must be called with the entry locked.
+ * The write may be synchronous or asynchronous.
+ */
+static int write_l2e(struct adapter *adap, struct l2t_entry *e, int sync)
+{
+	struct sk_buff *skb;
+	struct cpl_l2t_write_req *req;
+
+	skb = alloc_skb(sizeof(*req), GFP_ATOMIC);
+	if (!skb)
+		return -ENOMEM;
+
+	req = (struct cpl_l2t_write_req *)__skb_put(skb, sizeof(*req));
+	INIT_TP_WR(req, 0);
+
+	OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_L2T_WRITE_REQ,
+					e->idx | (sync ? F_SYNC_WR : 0) |
+					TID_QID(adap->sge.fw_evtq.abs_id)));
+	req->params = htons(L2T_W_PORT(e->lport) | L2T_W_NOREPLY(!sync));
+	req->l2t_idx = htons(e->idx);
+	req->vlan = htons(e->vlan);
+	if (e->neigh)
+		memcpy(e->dmac, e->neigh->ha, sizeof(e->dmac));
+	memcpy(req->dst_mac, e->dmac, sizeof(req->dst_mac));
+
+	set_wr_txq(skb, CPL_PRIORITY_CONTROL, 0);
+	t4_ofld_send(adap, skb);
+
+	if (sync && e->state != L2T_STATE_SWITCHING)
+		e->state = L2T_STATE_SYNC_WRITE;
+	return 0;
+}
+
+/*
+ * Send packets waiting in an L2T entry's ARP queue.  Must be called with the
+ * entry locked.
+ */
+static void send_pending(struct adapter *adap, struct l2t_entry *e)
+{
+	while (e->arpq_head) {
+		struct sk_buff *skb = e->arpq_head;
+
+		e->arpq_head = skb->next;
+		skb->next = NULL;
+		t4_ofld_send(adap, skb);
+	}
+	e->arpq_tail = NULL;
+}
+
+/*
+ * Process a CPL_L2T_WRITE_RPL.  Wake up the ARP queue if it completes a
+ * synchronous L2T_WRITE.  Note that the TID in the reply is really the L2T
+ * index it refers to.
+ */
+void do_l2t_write_rpl(struct adapter *adap, const struct cpl_l2t_write_rpl *rpl)
+{
+	unsigned int tid = GET_TID(rpl);
+	unsigned int idx = tid & (L2T_SIZE - 1);
+
+	if (unlikely(rpl->status != CPL_ERR_NONE)) {
+		dev_err(adap->pdev_dev,
+			"Unexpected L2T_WRITE_RPL status %u for entry %u\n",
+			rpl->status, idx);
+		return;
+	}
+
+	if (tid & F_SYNC_WR) {
+		struct l2t_entry *e = &adap->l2t->l2tab[idx];
+
+		spin_lock(&e->lock);
+		if (e->state != L2T_STATE_SWITCHING) {
+			send_pending(adap, e);
+			e->state = (e->neigh->nud_state & NUD_STALE) ?
+					L2T_STATE_STALE : L2T_STATE_VALID;
+		}
+		spin_unlock(&e->lock);
+	}
+}
+
+/*
+ * Add a packet to an L2T entry's queue of packets awaiting resolution.
+ * Must be called with the entry's lock held.
+ */
+static inline void arpq_enqueue(struct l2t_entry *e, struct sk_buff *skb)
+{
+	skb->next = NULL;
+	if (e->arpq_head)
+		e->arpq_tail->next = skb;
+	else
+		e->arpq_head = skb;
+	e->arpq_tail = skb;
+}
+
+int cxgb4_l2t_send(struct net_device *dev, struct sk_buff *skb,
+		   struct l2t_entry *e)
+{
+	struct adapter *adap = netdev2adap(dev);
+
+again:
+	switch (e->state) {
+	case L2T_STATE_STALE:     /* entry is stale, kick off revalidation */
+		neigh_event_send(e->neigh, NULL);
+		spin_lock_bh(&e->lock);
+		if (e->state == L2T_STATE_STALE)
+			e->state = L2T_STATE_VALID;
+		spin_unlock_bh(&e->lock);
+	case L2T_STATE_VALID:     /* fast-path, send the packet on */
+		return t4_ofld_send(adap, skb);
+	case L2T_STATE_RESOLVING:
+	case L2T_STATE_SYNC_WRITE:
+		spin_lock_bh(&e->lock);
+		if (e->state != L2T_STATE_SYNC_WRITE &&
+		    e->state != L2T_STATE_RESOLVING) {
+			spin_unlock_bh(&e->lock);
+			goto again;
+		}
+		arpq_enqueue(e, skb);
+		spin_unlock_bh(&e->lock);
+
+		if (e->state == L2T_STATE_RESOLVING &&
+		    !neigh_event_send(e->neigh, NULL)) {
+			spin_lock_bh(&e->lock);
+			if (e->state == L2T_STATE_RESOLVING && e->arpq_head)
+				write_l2e(adap, e, 1);
+			spin_unlock_bh(&e->lock);
+		}
+	}
+	return 0;
+}
+EXPORT_SYMBOL(cxgb4_l2t_send);
+
+/*
+ * Allocate a free L2T entry.  Must be called with l2t_data.lock held.
+ */
+static struct l2t_entry *alloc_l2e(struct l2t_data *d)
+{
+	struct l2t_entry *end, *e, **p;
+
+	if (!atomic_read(&d->nfree))
+		return NULL;
+
+	/* there's definitely a free entry */
+	for (e = d->rover, end = &d->l2tab[L2T_SIZE]; e != end; ++e)
+		if (atomic_read(&e->refcnt) == 0)
+			goto found;
+
+	for (e = d->l2tab; atomic_read(&e->refcnt); ++e)
+		;
+found:
+	d->rover = e + 1;
+	atomic_dec(&d->nfree);
+
+	/*
+	 * The entry we found may be an inactive entry that is
+	 * presently in the hash table.  We need to remove it.
+	 */
+	if (e->state < L2T_STATE_SWITCHING)
+		for (p = &d->l2tab[e->hash].first; *p; p = &(*p)->next)
+			if (*p == e) {
+				*p = e->next;
+				e->next = NULL;
+				break;
+			}
+
+	e->state = L2T_STATE_UNUSED;
+	return e;
+}
+
+/*
+ * Called when an L2T entry has no more users.
+ */
+static void t4_l2e_free(struct l2t_entry *e)
+{
+	struct l2t_data *d;
+
+	spin_lock_bh(&e->lock);
+	if (atomic_read(&e->refcnt) == 0) {  /* hasn't been recycled */
+		if (e->neigh) {
+			neigh_release(e->neigh);
+			e->neigh = NULL;
+		}
+	}
+	spin_unlock_bh(&e->lock);
+
+	d = container_of(e, struct l2t_data, l2tab[e->idx]);
+	atomic_inc(&d->nfree);
+}
+
+void cxgb4_l2t_release(struct l2t_entry *e)
+{
+	if (atomic_dec_and_test(&e->refcnt))
+		t4_l2e_free(e);
+}
+EXPORT_SYMBOL(cxgb4_l2t_release);
+
+/*
+ * Update an L2T entry that was previously used for the same next hop as neigh.
+ * Must be called with softirqs disabled.
+ */
+static void reuse_entry(struct l2t_entry *e, struct neighbour *neigh)
+{
+	unsigned int nud_state;
+
+	spin_lock(&e->lock);                /* avoid race with t4_l2t_free */
+	if (neigh != e->neigh)
+		neigh_replace(e, neigh);
+	nud_state = neigh->nud_state;
+	if (memcmp(e->dmac, neigh->ha, sizeof(e->dmac)) ||
+	    !(nud_state & NUD_VALID))
+		e->state = L2T_STATE_RESOLVING;
+	else if (nud_state & NUD_CONNECTED)
+		e->state = L2T_STATE_VALID;
+	else
+		e->state = L2T_STATE_STALE;
+	spin_unlock(&e->lock);
+}
+
+struct l2t_entry *cxgb4_l2t_get(struct l2t_data *d, struct neighbour *neigh,
+				const struct net_device *physdev,
+				unsigned int priority)
+{
+	u8 lport;
+	u16 vlan;
+	struct l2t_entry *e;
+	int addr_len = neigh->tbl->key_len;
+	u32 *addr = (u32 *)neigh->primary_key;
+	int ifidx = neigh->dev->ifindex;
+	int hash = addr_hash(addr, addr_len, ifidx);
+
+	if (neigh->dev->flags & IFF_LOOPBACK)
+		lport = netdev2pinfo(physdev)->tx_chan + 4;
+	else
+		lport = netdev2pinfo(physdev)->lport;
+
+	if (neigh->dev->priv_flags & IFF_802_1Q_VLAN)
+		vlan = vlan_dev_vlan_id(neigh->dev);
+	else
+		vlan = VLAN_NONE;
+
+	write_lock_bh(&d->lock);
+	for (e = d->l2tab[hash].first; e; e = e->next)
+		if (!addreq(e, addr) && e->ifindex == ifidx &&
+		    e->vlan == vlan && e->lport == lport) {
+			l2t_hold(d, e);
+			if (atomic_read(&e->refcnt) == 1)
+				reuse_entry(e, neigh);
+			goto done;
+		}
+
+	/* Need to allocate a new entry */
+	e = alloc_l2e(d);
+	if (e) {
+		spin_lock(&e->lock);          /* avoid race with t4_l2t_free */
+		e->state = L2T_STATE_RESOLVING;
+		memcpy(e->addr, addr, addr_len);
+		e->ifindex = ifidx;
+		e->hash = hash;
+		e->lport = lport;
+		e->v6 = addr_len == 16;
+		atomic_set(&e->refcnt, 1);
+		neigh_replace(e, neigh);
+		e->vlan = vlan;
+		e->next = d->l2tab[hash].first;
+		d->l2tab[hash].first = e;
+		spin_unlock(&e->lock);
+	}
+done:
+	write_unlock_bh(&d->lock);
+	return e;
+}
+EXPORT_SYMBOL(cxgb4_l2t_get);
+
+/*
+ * Called when address resolution fails for an L2T entry to handle packets
+ * on the arpq head.  If a packet specifies a failure handler it is invoked,
+ * otherwise the packet is sent to the device.
+ */
+static void handle_failed_resolution(struct adapter *adap, struct sk_buff *arpq)
+{
+	while (arpq) {
+		struct sk_buff *skb = arpq;
+		const struct l2t_skb_cb *cb = L2T_SKB_CB(skb);
+
+		arpq = skb->next;
+		skb->next = NULL;
+		if (cb->arp_err_handler)
+			cb->arp_err_handler(cb->handle, skb);
+		else
+			t4_ofld_send(adap, skb);
+	}
+}
+
+/*
+ * Called when the host's neighbor layer makes a change to some entry that is
+ * loaded into the HW L2 table.
+ */
+void t4_l2t_update(struct adapter *adap, struct neighbour *neigh)
+{
+	struct l2t_entry *e;
+	struct sk_buff *arpq = NULL;
+	struct l2t_data *d = adap->l2t;
+	int addr_len = neigh->tbl->key_len;
+	u32 *addr = (u32 *) neigh->primary_key;
+	int ifidx = neigh->dev->ifindex;
+	int hash = addr_hash(addr, addr_len, ifidx);
+
+	read_lock_bh(&d->lock);
+	for (e = d->l2tab[hash].first; e; e = e->next)
+		if (!addreq(e, addr) && e->ifindex == ifidx) {
+			spin_lock(&e->lock);
+			if (atomic_read(&e->refcnt))
+				goto found;
+			spin_unlock(&e->lock);
+			break;
+		}
+	read_unlock_bh(&d->lock);
+	return;
+
+ found:
+	read_unlock(&d->lock);
+
+	if (neigh != e->neigh)
+		neigh_replace(e, neigh);
+
+	if (e->state == L2T_STATE_RESOLVING) {
+		if (neigh->nud_state & NUD_FAILED) {
+			arpq = e->arpq_head;
+			e->arpq_head = e->arpq_tail = NULL;
+		} else if ((neigh->nud_state & (NUD_CONNECTED | NUD_STALE)) &&
+			   e->arpq_head) {
+			write_l2e(adap, e, 1);
+		}
+	} else {
+		e->state = neigh->nud_state & NUD_CONNECTED ?
+			L2T_STATE_VALID : L2T_STATE_STALE;
+		if (memcmp(e->dmac, neigh->ha, sizeof(e->dmac)))
+			write_l2e(adap, e, 0);
+	}
+
+	spin_unlock_bh(&e->lock);
+
+	if (arpq)
+		handle_failed_resolution(adap, arpq);
+}
+
+/*
+ * Allocate an L2T entry for use by a switching rule.  Such entries need to be
+ * explicitly freed and while busy they are not on any hash chain, so normal
+ * address resolution updates do not see them.
+ */
+struct l2t_entry *t4_l2t_alloc_switching(struct l2t_data *d)
+{
+	struct l2t_entry *e;
+
+	write_lock_bh(&d->lock);
+	e = alloc_l2e(d);
+	if (e) {
+		spin_lock(&e->lock);          /* avoid race with t4_l2t_free */
+		e->state = L2T_STATE_SWITCHING;
+		atomic_set(&e->refcnt, 1);
+		spin_unlock(&e->lock);
+	}
+	write_unlock_bh(&d->lock);
+	return e;
+}
+
+/*
+ * Sets/updates the contents of a switching L2T entry that has been allocated
+ * with an earlier call to @t4_l2t_alloc_switching.
+ */
+int t4_l2t_set_switching(struct adapter *adap, struct l2t_entry *e, u16 vlan,
+			 u8 port, u8 *eth_addr)
+{
+	e->vlan = vlan;
+	e->lport = port;
+	memcpy(e->dmac, eth_addr, ETH_ALEN);
+	return write_l2e(adap, e, 0);
+}
+
+struct l2t_data *t4_init_l2t(void)
+{
+	int i;
+	struct l2t_data *d;
+
+	d = t4_alloc_mem(sizeof(*d));
+	if (!d)
+		return NULL;
+
+	d->rover = d->l2tab;
+	atomic_set(&d->nfree, L2T_SIZE);
+	rwlock_init(&d->lock);
+
+	for (i = 0; i < L2T_SIZE; ++i) {
+		d->l2tab[i].idx = i;
+		d->l2tab[i].state = L2T_STATE_UNUSED;
+		spin_lock_init(&d->l2tab[i].lock);
+		atomic_set(&d->l2tab[i].refcnt, 0);
+	}
+	return d;
+}
+
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+
+static inline void *l2t_get_idx(struct seq_file *seq, loff_t pos)
+{
+	struct l2t_entry *l2tab = seq->private;
+
+	return pos >= L2T_SIZE ? NULL : &l2tab[pos];
+}
+
+static void *l2t_seq_start(struct seq_file *seq, loff_t *pos)
+{
+	return *pos ? l2t_get_idx(seq, *pos - 1) : SEQ_START_TOKEN;
+}
+
+static void *l2t_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+	v = l2t_get_idx(seq, *pos);
+	if (v)
+		++*pos;
+	return v;
+}
+
+static void l2t_seq_stop(struct seq_file *seq, void *v)
+{
+}
+
+static char l2e_state(const struct l2t_entry *e)
+{
+	switch (e->state) {
+	case L2T_STATE_VALID: return 'V';
+	case L2T_STATE_STALE: return 'S';
+	case L2T_STATE_SYNC_WRITE: return 'W';
+	case L2T_STATE_RESOLVING: return e->arpq_head ? 'A' : 'R';
+	case L2T_STATE_SWITCHING: return 'X';
+	default:
+		return 'U';
+	}
+}
+
+static int l2t_seq_show(struct seq_file *seq, void *v)
+{
+	if (v == SEQ_START_TOKEN)
+		seq_puts(seq, " Idx IP address                "
+			 "Ethernet address  VLAN/P LP State Users Port\n");
+	else {
+		char ip[60];
+		struct l2t_entry *e = v;
+
+		spin_lock_bh(&e->lock);
+		if (e->state == L2T_STATE_SWITCHING)
+			ip[0] = '\0';
+		else
+			sprintf(ip, e->v6 ? "%pI6c" : "%pI4", e->addr);
+		seq_printf(seq, "%4u %-25s %17pM %4d %u %2u   %c   %5u %s\n",
+			   e->idx, ip, e->dmac,
+			   e->vlan & VLAN_VID_MASK, vlan_prio(e), e->lport,
+			   l2e_state(e), atomic_read(&e->refcnt),
+			   e->neigh ? e->neigh->dev->name : "");
+		spin_unlock_bh(&e->lock);
+	}
+	return 0;
+}
+
+static const struct seq_operations l2t_seq_ops = {
+	.start = l2t_seq_start,
+	.next = l2t_seq_next,
+	.stop = l2t_seq_stop,
+	.show = l2t_seq_show
+};
+
+static int l2t_seq_open(struct inode *inode, struct file *file)
+{
+	int rc = seq_open(file, &l2t_seq_ops);
+
+	if (!rc) {
+		struct adapter *adap = inode->i_private;
+		struct seq_file *seq = file->private_data;
+
+		seq->private = adap->l2t->l2tab;
+	}
+	return rc;
+}
+
+const struct file_operations t4_l2t_fops = {
+	.owner = THIS_MODULE,
+	.open = l2t_seq_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = seq_release,
+};
diff --git a/drivers/net/cxgb4/l2t.h b/drivers/net/cxgb4/l2t.h
new file mode 100644
index 000000000000..643f27ed3cf4
--- /dev/null
+++ b/drivers/net/cxgb4/l2t.h
@@ -0,0 +1,110 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __CXGB4_L2T_H
+#define __CXGB4_L2T_H
+
+#include <linux/spinlock.h>
+#include <linux/if_ether.h>
+#include <asm/atomic.h>
+
+struct adapter;
+struct l2t_data;
+struct neighbour;
+struct net_device;
+struct file_operations;
+struct cpl_l2t_write_rpl;
+
+/*
+ * Each L2T entry plays multiple roles.  First of all, it keeps state for the
+ * corresponding entry of the HW L2 table and maintains a queue of offload
+ * packets awaiting address resolution.  Second, it is a node of a hash table
+ * chain, where the nodes of the chain are linked together through their next
+ * pointer.  Finally, each node is a bucket of a hash table, pointing to the
+ * first element in its chain through its first pointer.
+ */
+struct l2t_entry {
+	u16 state;                  /* entry state */
+	u16 idx;                    /* entry index */
+	u32 addr[4];                /* next hop IP or IPv6 address */
+	int ifindex;                /* neighbor's net_device's ifindex */
+	struct neighbour *neigh;    /* associated neighbour */
+	struct l2t_entry *first;    /* start of hash chain */
+	struct l2t_entry *next;     /* next l2t_entry on chain */
+	struct sk_buff *arpq_head;  /* queue of packets awaiting resolution */
+	struct sk_buff *arpq_tail;
+	spinlock_t lock;
+	atomic_t refcnt;            /* entry reference count */
+	u16 hash;                   /* hash bucket the entry is on */
+	u16 vlan;                   /* VLAN TCI (id: bits 0-11, prio: 13-15 */
+	u8 v6;                      /* whether entry is for IPv6 */
+	u8 lport;                   /* associated offload logical interface */
+	u8 dmac[ETH_ALEN];          /* neighbour's MAC address */
+};
+
+typedef void (*arp_err_handler_t)(void *handle, struct sk_buff *skb);
+
+/*
+ * Callback stored in an skb to handle address resolution failure.
+ */
+struct l2t_skb_cb {
+	void *handle;
+	arp_err_handler_t arp_err_handler;
+};
+
+#define L2T_SKB_CB(skb) ((struct l2t_skb_cb *)(skb)->cb)
+
+static inline void t4_set_arp_err_handler(struct sk_buff *skb, void *handle,
+					  arp_err_handler_t handler)
+{
+	L2T_SKB_CB(skb)->handle = handle;
+	L2T_SKB_CB(skb)->arp_err_handler = handler;
+}
+
+void cxgb4_l2t_release(struct l2t_entry *e);
+int cxgb4_l2t_send(struct net_device *dev, struct sk_buff *skb,
+		   struct l2t_entry *e);
+struct l2t_entry *cxgb4_l2t_get(struct l2t_data *d, struct neighbour *neigh,
+				const struct net_device *physdev,
+				unsigned int priority);
+
+void t4_l2t_update(struct adapter *adap, struct neighbour *neigh);
+struct l2t_entry *t4_l2t_alloc_switching(struct l2t_data *d);
+int t4_l2t_set_switching(struct adapter *adap, struct l2t_entry *e, u16 vlan,
+			 u8 port, u8 *eth_addr);
+struct l2t_data *t4_init_l2t(void);
+void do_l2t_write_rpl(struct adapter *p, const struct cpl_l2t_write_rpl *rpl);
+
+extern const struct file_operations t4_l2t_fops;
+#endif  /* __CXGB4_L2T_H */
diff --git a/drivers/net/cxgb4/sge.c b/drivers/net/cxgb4/sge.c
new file mode 100644
index 000000000000..14adc58e71c3
--- /dev/null
+++ b/drivers/net/cxgb4/sge.c
@@ -0,0 +1,2431 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/if_vlan.h>
+#include <linux/ip.h>
+#include <linux/dma-mapping.h>
+#include <linux/jiffies.h>
+#include <net/ipv6.h>
+#include <net/tcp.h>
+#include "cxgb4.h"
+#include "t4_regs.h"
+#include "t4_msg.h"
+#include "t4fw_api.h"
+
+/*
+ * Rx buffer size.  We use largish buffers if possible but settle for single
+ * pages under memory shortage.
+ */
+#if PAGE_SHIFT >= 16
+# define FL_PG_ORDER 0
+#else
+# define FL_PG_ORDER (16 - PAGE_SHIFT)
+#endif
+
+/* RX_PULL_LEN should be <= RX_COPY_THRES */
+#define RX_COPY_THRES    256
+#define RX_PULL_LEN      128
+
+/*
+ * Main body length for sk_buffs used for Rx Ethernet packets with fragments.
+ * Should be >= RX_PULL_LEN but possibly bigger to give pskb_may_pull some room.
+ */
+#define RX_PKT_SKB_LEN   512
+
+/* Ethernet header padding prepended to RX_PKTs */
+#define RX_PKT_PAD 2
+
+/*
+ * Max number of Tx descriptors we clean up at a time.  Should be modest as
+ * freeing skbs isn't cheap and it happens while holding locks.  We just need
+ * to free packets faster than they arrive, we eventually catch up and keep
+ * the amortized cost reasonable.  Must be >= 2 * TXQ_STOP_THRES.
+ */
+#define MAX_TX_RECLAIM 16
+
+/*
+ * Max number of Rx buffers we replenish at a time.  Again keep this modest,
+ * allocating buffers isn't cheap either.
+ */
+#define MAX_RX_REFILL 16U
+
+/*
+ * Period of the Rx queue check timer.  This timer is infrequent as it has
+ * something to do only when the system experiences severe memory shortage.
+ */
+#define RX_QCHECK_PERIOD (HZ / 2)
+
+/*
+ * Period of the Tx queue check timer.
+ */
+#define TX_QCHECK_PERIOD (HZ / 2)
+
+/*
+ * Max number of Tx descriptors to be reclaimed by the Tx timer.
+ */
+#define MAX_TIMER_TX_RECLAIM 100
+
+/*
+ * Timer index used when backing off due to memory shortage.
+ */
+#define NOMEM_TMR_IDX (SGE_NTIMERS - 1)
+
+/*
+ * An FL with <= FL_STARVE_THRES buffers is starving and a periodic timer will
+ * attempt to refill it.
+ */
+#define FL_STARVE_THRES 4
+
+/*
+ * Suspend an Ethernet Tx queue with fewer available descriptors than this.
+ * This is the same as calc_tx_descs() for a TSO packet with
+ * nr_frags == MAX_SKB_FRAGS.
+ */
+#define ETHTXQ_STOP_THRES \
+	(1 + DIV_ROUND_UP((3 * MAX_SKB_FRAGS) / 2 + (MAX_SKB_FRAGS & 1), 8))
+
+/*
+ * Suspension threshold for non-Ethernet Tx queues.  We require enough room
+ * for a full sized WR.
+ */
+#define TXQ_STOP_THRES (SGE_MAX_WR_LEN / sizeof(struct tx_desc))
+
+/*
+ * Max Tx descriptor space we allow for an Ethernet packet to be inlined
+ * into a WR.
+ */
+#define MAX_IMM_TX_PKT_LEN 128
+
+/*
+ * Max size of a WR sent through a control Tx queue.
+ */
+#define MAX_CTRL_WR_LEN SGE_MAX_WR_LEN
+
+enum {
+	/* packet alignment in FL buffers */
+	FL_ALIGN = L1_CACHE_BYTES < 32 ? 32 : L1_CACHE_BYTES,
+	/* egress status entry size */
+	STAT_LEN = L1_CACHE_BYTES > 64 ? 128 : 64
+};
+
+struct tx_sw_desc {                /* SW state per Tx descriptor */
+	struct sk_buff *skb;
+	struct ulptx_sgl *sgl;
+};
+
+struct rx_sw_desc {                /* SW state per Rx descriptor */
+	struct page *page;
+	dma_addr_t dma_addr;
+};
+
+/*
+ * The low bits of rx_sw_desc.dma_addr have special meaning.
+ */
+enum {
+	RX_LARGE_BUF    = 1 << 0, /* buffer is larger than PAGE_SIZE */
+	RX_UNMAPPED_BUF = 1 << 1, /* buffer is not mapped */
+};
+
+static inline dma_addr_t get_buf_addr(const struct rx_sw_desc *d)
+{
+	return d->dma_addr & ~(dma_addr_t)(RX_LARGE_BUF | RX_UNMAPPED_BUF);
+}
+
+static inline bool is_buf_mapped(const struct rx_sw_desc *d)
+{
+	return !(d->dma_addr & RX_UNMAPPED_BUF);
+}
+
+/**
+ *	txq_avail - return the number of available slots in a Tx queue
+ *	@q: the Tx queue
+ *
+ *	Returns the number of descriptors in a Tx queue available to write new
+ *	packets.
+ */
+static inline unsigned int txq_avail(const struct sge_txq *q)
+{
+	return q->size - 1 - q->in_use;
+}
+
+/**
+ *	fl_cap - return the capacity of a free-buffer list
+ *	@fl: the FL
+ *
+ *	Returns the capacity of a free-buffer list.  The capacity is less than
+ *	the size because one descriptor needs to be left unpopulated, otherwise
+ *	HW will think the FL is empty.
+ */
+static inline unsigned int fl_cap(const struct sge_fl *fl)
+{
+	return fl->size - 8;   /* 1 descriptor = 8 buffers */
+}
+
+static inline bool fl_starving(const struct sge_fl *fl)
+{
+	return fl->avail - fl->pend_cred <= FL_STARVE_THRES;
+}
+
+static int map_skb(struct device *dev, const struct sk_buff *skb,
+		   dma_addr_t *addr)
+{
+	const skb_frag_t *fp, *end;
+	const struct skb_shared_info *si;
+
+	*addr = dma_map_single(dev, skb->data, skb_headlen(skb), DMA_TO_DEVICE);
+	if (dma_mapping_error(dev, *addr))
+		goto out_err;
+
+	si = skb_shinfo(skb);
+	end = &si->frags[si->nr_frags];
+
+	for (fp = si->frags; fp < end; fp++) {
+		*++addr = dma_map_page(dev, fp->page, fp->page_offset, fp->size,
+				       DMA_TO_DEVICE);
+		if (dma_mapping_error(dev, *addr))
+			goto unwind;
+	}
+	return 0;
+
+unwind:
+	while (fp-- > si->frags)
+		dma_unmap_page(dev, *--addr, fp->size, DMA_TO_DEVICE);
+
+	dma_unmap_single(dev, addr[-1], skb_headlen(skb), DMA_TO_DEVICE);
+out_err:
+	return -ENOMEM;
+}
+
+#ifdef CONFIG_NEED_DMA_MAP_STATE
+static void unmap_skb(struct device *dev, const struct sk_buff *skb,
+		      const dma_addr_t *addr)
+{
+	const skb_frag_t *fp, *end;
+	const struct skb_shared_info *si;
+
+	dma_unmap_single(dev, *addr++, skb_headlen(skb), DMA_TO_DEVICE);
+
+	si = skb_shinfo(skb);
+	end = &si->frags[si->nr_frags];
+	for (fp = si->frags; fp < end; fp++)
+		dma_unmap_page(dev, *addr++, fp->size, DMA_TO_DEVICE);
+}
+
+/**
+ *	deferred_unmap_destructor - unmap a packet when it is freed
+ *	@skb: the packet
+ *
+ *	This is the packet destructor used for Tx packets that need to remain
+ *	mapped until they are freed rather than until their Tx descriptors are
+ *	freed.
+ */
+static void deferred_unmap_destructor(struct sk_buff *skb)
+{
+	unmap_skb(skb->dev->dev.parent, skb, (dma_addr_t *)skb->head);
+}
+#endif
+
+static void unmap_sgl(struct device *dev, const struct sk_buff *skb,
+		      const struct ulptx_sgl *sgl, const struct sge_txq *q)
+{
+	const struct ulptx_sge_pair *p;
+	unsigned int nfrags = skb_shinfo(skb)->nr_frags;
+
+	if (likely(skb_headlen(skb)))
+		dma_unmap_single(dev, be64_to_cpu(sgl->addr0), ntohl(sgl->len0),
+				 DMA_TO_DEVICE);
+	else {
+		dma_unmap_page(dev, be64_to_cpu(sgl->addr0), ntohl(sgl->len0),
+			       DMA_TO_DEVICE);
+		nfrags--;
+	}
+
+	/*
+	 * the complexity below is because of the possibility of a wrap-around
+	 * in the middle of an SGL
+	 */
+	for (p = sgl->sge; nfrags >= 2; nfrags -= 2) {
+		if (likely((u8 *)(p + 1) <= (u8 *)q->stat)) {
+unmap:			dma_unmap_page(dev, be64_to_cpu(p->addr[0]),
+				       ntohl(p->len[0]), DMA_TO_DEVICE);
+			dma_unmap_page(dev, be64_to_cpu(p->addr[1]),
+				       ntohl(p->len[1]), DMA_TO_DEVICE);
+			p++;
+		} else if ((u8 *)p == (u8 *)q->stat) {
+			p = (const struct ulptx_sge_pair *)q->desc;
+			goto unmap;
+		} else if ((u8 *)p + 8 == (u8 *)q->stat) {
+			const __be64 *addr = (const __be64 *)q->desc;
+
+			dma_unmap_page(dev, be64_to_cpu(addr[0]),
+				       ntohl(p->len[0]), DMA_TO_DEVICE);
+			dma_unmap_page(dev, be64_to_cpu(addr[1]),
+				       ntohl(p->len[1]), DMA_TO_DEVICE);
+			p = (const struct ulptx_sge_pair *)&addr[2];
+		} else {
+			const __be64 *addr = (const __be64 *)q->desc;
+
+			dma_unmap_page(dev, be64_to_cpu(p->addr[0]),
+				       ntohl(p->len[0]), DMA_TO_DEVICE);
+			dma_unmap_page(dev, be64_to_cpu(addr[0]),
+				       ntohl(p->len[1]), DMA_TO_DEVICE);
+			p = (const struct ulptx_sge_pair *)&addr[1];
+		}
+	}
+	if (nfrags) {
+		__be64 addr;
+
+		if ((u8 *)p == (u8 *)q->stat)
+			p = (const struct ulptx_sge_pair *)q->desc;
+		addr = (u8 *)p + 16 <= (u8 *)q->stat ? p->addr[0] :
+						       *(const __be64 *)q->desc;
+		dma_unmap_page(dev, be64_to_cpu(addr), ntohl(p->len[0]),
+			       DMA_TO_DEVICE);
+	}
+}
+
+/**
+ *	free_tx_desc - reclaims Tx descriptors and their buffers
+ *	@adapter: the adapter
+ *	@q: the Tx queue to reclaim descriptors from
+ *	@n: the number of descriptors to reclaim
+ *	@unmap: whether the buffers should be unmapped for DMA
+ *
+ *	Reclaims Tx descriptors from an SGE Tx queue and frees the associated
+ *	Tx buffers.  Called with the Tx queue lock held.
+ */
+static void free_tx_desc(struct adapter *adap, struct sge_txq *q,
+			 unsigned int n, bool unmap)
+{
+	struct tx_sw_desc *d;
+	unsigned int cidx = q->cidx;
+	struct device *dev = adap->pdev_dev;
+
+	d = &q->sdesc[cidx];
+	while (n--) {
+		if (d->skb) {                       /* an SGL is present */
+			if (unmap)
+				unmap_sgl(dev, d->skb, d->sgl, q);
+			kfree_skb(d->skb);
+			d->skb = NULL;
+		}
+		++d;
+		if (++cidx == q->size) {
+			cidx = 0;
+			d = q->sdesc;
+		}
+	}
+	q->cidx = cidx;
+}
+
+/*
+ * Return the number of reclaimable descriptors in a Tx queue.
+ */
+static inline int reclaimable(const struct sge_txq *q)
+{
+	int hw_cidx = ntohs(q->stat->cidx);
+	hw_cidx -= q->cidx;
+	return hw_cidx < 0 ? hw_cidx + q->size : hw_cidx;
+}
+
+/**
+ *	reclaim_completed_tx - reclaims completed Tx descriptors
+ *	@adap: the adapter
+ *	@q: the Tx queue to reclaim completed descriptors from
+ *	@unmap: whether the buffers should be unmapped for DMA
+ *
+ *	Reclaims Tx descriptors that the SGE has indicated it has processed,
+ *	and frees the associated buffers if possible.  Called with the Tx
+ *	queue locked.
+ */
+static inline void reclaim_completed_tx(struct adapter *adap, struct sge_txq *q,
+					bool unmap)
+{
+	int avail = reclaimable(q);
+
+	if (avail) {
+		/*
+		 * Limit the amount of clean up work we do at a time to keep
+		 * the Tx lock hold time O(1).
+		 */
+		if (avail > MAX_TX_RECLAIM)
+			avail = MAX_TX_RECLAIM;
+
+		free_tx_desc(adap, q, avail, unmap);
+		q->in_use -= avail;
+	}
+}
+
+static inline int get_buf_size(const struct rx_sw_desc *d)
+{
+#if FL_PG_ORDER > 0
+	return (d->dma_addr & RX_LARGE_BUF) ? (PAGE_SIZE << FL_PG_ORDER) :
+					      PAGE_SIZE;
+#else
+	return PAGE_SIZE;
+#endif
+}
+
+/**
+ *	free_rx_bufs - free the Rx buffers on an SGE free list
+ *	@adap: the adapter
+ *	@q: the SGE free list to free buffers from
+ *	@n: how many buffers to free
+ *
+ *	Release the next @n buffers on an SGE free-buffer Rx queue.   The
+ *	buffers must be made inaccessible to HW before calling this function.
+ */
+static void free_rx_bufs(struct adapter *adap, struct sge_fl *q, int n)
+{
+	while (n--) {
+		struct rx_sw_desc *d = &q->sdesc[q->cidx];
+
+		if (is_buf_mapped(d))
+			dma_unmap_page(adap->pdev_dev, get_buf_addr(d),
+				       get_buf_size(d), PCI_DMA_FROMDEVICE);
+		put_page(d->page);
+		d->page = NULL;
+		if (++q->cidx == q->size)
+			q->cidx = 0;
+		q->avail--;
+	}
+}
+
+/**
+ *	unmap_rx_buf - unmap the current Rx buffer on an SGE free list
+ *	@adap: the adapter
+ *	@q: the SGE free list
+ *
+ *	Unmap the current buffer on an SGE free-buffer Rx queue.   The
+ *	buffer must be made inaccessible to HW before calling this function.
+ *
+ *	This is similar to @free_rx_bufs above but does not free the buffer.
+ *	Do note that the FL still loses any further access to the buffer.
+ */
+static void unmap_rx_buf(struct adapter *adap, struct sge_fl *q)
+{
+	struct rx_sw_desc *d = &q->sdesc[q->cidx];
+
+	if (is_buf_mapped(d))
+		dma_unmap_page(adap->pdev_dev, get_buf_addr(d),
+			       get_buf_size(d), PCI_DMA_FROMDEVICE);
+	d->page = NULL;
+	if (++q->cidx == q->size)
+		q->cidx = 0;
+	q->avail--;
+}
+
+static inline void ring_fl_db(struct adapter *adap, struct sge_fl *q)
+{
+	if (q->pend_cred >= 8) {
+		wmb();
+		t4_write_reg(adap, MYPF_REG(SGE_PF_KDOORBELL), DBPRIO |
+			     QID(q->cntxt_id) | PIDX(q->pend_cred / 8));
+		q->pend_cred &= 7;
+	}
+}
+
+static inline void set_rx_sw_desc(struct rx_sw_desc *sd, struct page *pg,
+				  dma_addr_t mapping)
+{
+	sd->page = pg;
+	sd->dma_addr = mapping;      /* includes size low bits */
+}
+
+/**
+ *	refill_fl - refill an SGE Rx buffer ring
+ *	@adap: the adapter
+ *	@q: the ring to refill
+ *	@n: the number of new buffers to allocate
+ *	@gfp: the gfp flags for the allocations
+ *
+ *	(Re)populate an SGE free-buffer queue with up to @n new packet buffers,
+ *	allocated with the supplied gfp flags.  The caller must assure that
+ *	@n does not exceed the queue's capacity.  If afterwards the queue is
+ *	found critically low mark it as starving in the bitmap of starving FLs.
+ *
+ *	Returns the number of buffers allocated.
+ */
+static unsigned int refill_fl(struct adapter *adap, struct sge_fl *q, int n,
+			      gfp_t gfp)
+{
+	struct page *pg;
+	dma_addr_t mapping;
+	unsigned int cred = q->avail;
+	__be64 *d = &q->desc[q->pidx];
+	struct rx_sw_desc *sd = &q->sdesc[q->pidx];
+
+	gfp |= __GFP_NOWARN;         /* failures are expected */
+
+#if FL_PG_ORDER > 0
+	/*
+	 * Prefer large buffers
+	 */
+	while (n) {
+		pg = alloc_pages(gfp | __GFP_COMP, FL_PG_ORDER);
+		if (unlikely(!pg)) {
+			q->large_alloc_failed++;
+			break;       /* fall back to single pages */
+		}
+
+		mapping = dma_map_page(adap->pdev_dev, pg, 0,
+				       PAGE_SIZE << FL_PG_ORDER,
+				       PCI_DMA_FROMDEVICE);
+		if (unlikely(dma_mapping_error(adap->pdev_dev, mapping))) {
+			__free_pages(pg, FL_PG_ORDER);
+			goto out;   /* do not try small pages for this error */
+		}
+		mapping |= RX_LARGE_BUF;
+		*d++ = cpu_to_be64(mapping);
+
+		set_rx_sw_desc(sd, pg, mapping);
+		sd++;
+
+		q->avail++;
+		if (++q->pidx == q->size) {
+			q->pidx = 0;
+			sd = q->sdesc;
+			d = q->desc;
+		}
+		n--;
+	}
+#endif
+
+	while (n--) {
+		pg = __netdev_alloc_page(adap->port[0], gfp);
+		if (unlikely(!pg)) {
+			q->alloc_failed++;
+			break;
+		}
+
+		mapping = dma_map_page(adap->pdev_dev, pg, 0, PAGE_SIZE,
+				       PCI_DMA_FROMDEVICE);
+		if (unlikely(dma_mapping_error(adap->pdev_dev, mapping))) {
+			netdev_free_page(adap->port[0], pg);
+			goto out;
+		}
+		*d++ = cpu_to_be64(mapping);
+
+		set_rx_sw_desc(sd, pg, mapping);
+		sd++;
+
+		q->avail++;
+		if (++q->pidx == q->size) {
+			q->pidx = 0;
+			sd = q->sdesc;
+			d = q->desc;
+		}
+	}
+
+out:	cred = q->avail - cred;
+	q->pend_cred += cred;
+	ring_fl_db(adap, q);
+
+	if (unlikely(fl_starving(q))) {
+		smp_wmb();
+		set_bit(q->cntxt_id, adap->sge.starving_fl);
+	}
+
+	return cred;
+}
+
+static inline void __refill_fl(struct adapter *adap, struct sge_fl *fl)
+{
+	refill_fl(adap, fl, min(MAX_RX_REFILL, fl_cap(fl) - fl->avail),
+		  GFP_ATOMIC);
+}
+
+/**
+ *	alloc_ring - allocate resources for an SGE descriptor ring
+ *	@dev: the PCI device's core device
+ *	@nelem: the number of descriptors
+ *	@elem_size: the size of each descriptor
+ *	@sw_size: the size of the SW state associated with each ring element
+ *	@phys: the physical address of the allocated ring
+ *	@metadata: address of the array holding the SW state for the ring
+ *	@stat_size: extra space in HW ring for status information
+ *
+ *	Allocates resources for an SGE descriptor ring, such as Tx queues,
+ *	free buffer lists, or response queues.  Each SGE ring requires
+ *	space for its HW descriptors plus, optionally, space for the SW state
+ *	associated with each HW entry (the metadata).  The function returns
+ *	three values: the virtual address for the HW ring (the return value
+ *	of the function), the bus address of the HW ring, and the address
+ *	of the SW ring.
+ */
+static void *alloc_ring(struct device *dev, size_t nelem, size_t elem_size,
+			size_t sw_size, dma_addr_t *phys, void *metadata,
+			size_t stat_size)
+{
+	size_t len = nelem * elem_size + stat_size;
+	void *s = NULL;
+	void *p = dma_alloc_coherent(dev, len, phys, GFP_KERNEL);
+
+	if (!p)
+		return NULL;
+	if (sw_size) {
+		s = kcalloc(nelem, sw_size, GFP_KERNEL);
+
+		if (!s) {
+			dma_free_coherent(dev, len, p, *phys);
+			return NULL;
+		}
+	}
+	if (metadata)
+		*(void **)metadata = s;
+	memset(p, 0, len);
+	return p;
+}
+
+/**
+ *	sgl_len - calculates the size of an SGL of the given capacity
+ *	@n: the number of SGL entries
+ *
+ *	Calculates the number of flits needed for a scatter/gather list that
+ *	can hold the given number of entries.
+ */
+static inline unsigned int sgl_len(unsigned int n)
+{
+	n--;
+	return (3 * n) / 2 + (n & 1) + 2;
+}
+
+/**
+ *	flits_to_desc - returns the num of Tx descriptors for the given flits
+ *	@n: the number of flits
+ *
+ *	Returns the number of Tx descriptors needed for the supplied number
+ *	of flits.
+ */
+static inline unsigned int flits_to_desc(unsigned int n)
+{
+	BUG_ON(n > SGE_MAX_WR_LEN / 8);
+	return DIV_ROUND_UP(n, 8);
+}
+
+/**
+ *	is_eth_imm - can an Ethernet packet be sent as immediate data?
+ *	@skb: the packet
+ *
+ *	Returns whether an Ethernet packet is small enough to fit as
+ *	immediate data.
+ */
+static inline int is_eth_imm(const struct sk_buff *skb)
+{
+	return skb->len <= MAX_IMM_TX_PKT_LEN - sizeof(struct cpl_tx_pkt);
+}
+
+/**
+ *	calc_tx_flits - calculate the number of flits for a packet Tx WR
+ *	@skb: the packet
+ *
+ *	Returns the number of flits needed for a Tx WR for the given Ethernet
+ *	packet, including the needed WR and CPL headers.
+ */
+static inline unsigned int calc_tx_flits(const struct sk_buff *skb)
+{
+	unsigned int flits;
+
+	if (is_eth_imm(skb))
+		return DIV_ROUND_UP(skb->len + sizeof(struct cpl_tx_pkt), 8);
+
+	flits = sgl_len(skb_shinfo(skb)->nr_frags + 1) + 4;
+	if (skb_shinfo(skb)->gso_size)
+		flits += 2;
+	return flits;
+}
+
+/**
+ *	calc_tx_descs - calculate the number of Tx descriptors for a packet
+ *	@skb: the packet
+ *
+ *	Returns the number of Tx descriptors needed for the given Ethernet
+ *	packet, including the needed WR and CPL headers.
+ */
+static inline unsigned int calc_tx_descs(const struct sk_buff *skb)
+{
+	return flits_to_desc(calc_tx_flits(skb));
+}
+
+/**
+ *	write_sgl - populate a scatter/gather list for a packet
+ *	@skb: the packet
+ *	@q: the Tx queue we are writing into
+ *	@sgl: starting location for writing the SGL
+ *	@end: points right after the end of the SGL
+ *	@start: start offset into skb main-body data to include in the SGL
+ *	@addr: the list of bus addresses for the SGL elements
+ *
+ *	Generates a gather list for the buffers that make up a packet.
+ *	The caller must provide adequate space for the SGL that will be written.
+ *	The SGL includes all of the packet's page fragments and the data in its
+ *	main body except for the first @start bytes.  @sgl must be 16-byte
+ *	aligned and within a Tx descriptor with available space.  @end points
+ *	right after the end of the SGL but does not account for any potential
+ *	wrap around, i.e., @end > @sgl.
+ */
+static void write_sgl(const struct sk_buff *skb, struct sge_txq *q,
+		      struct ulptx_sgl *sgl, u64 *end, unsigned int start,
+		      const dma_addr_t *addr)
+{
+	unsigned int i, len;
+	struct ulptx_sge_pair *to;
+	const struct skb_shared_info *si = skb_shinfo(skb);
+	unsigned int nfrags = si->nr_frags;
+	struct ulptx_sge_pair buf[MAX_SKB_FRAGS / 2 + 1];
+
+	len = skb_headlen(skb) - start;
+	if (likely(len)) {
+		sgl->len0 = htonl(len);
+		sgl->addr0 = cpu_to_be64(addr[0] + start);
+		nfrags++;
+	} else {
+		sgl->len0 = htonl(si->frags[0].size);
+		sgl->addr0 = cpu_to_be64(addr[1]);
+	}
+
+	sgl->cmd_nsge = htonl(ULPTX_CMD(ULP_TX_SC_DSGL) | ULPTX_NSGE(nfrags));
+	if (likely(--nfrags == 0))
+		return;
+	/*
+	 * Most of the complexity below deals with the possibility we hit the
+	 * end of the queue in the middle of writing the SGL.  For this case
+	 * only we create the SGL in a temporary buffer and then copy it.
+	 */
+	to = (u8 *)end > (u8 *)q->stat ? buf : sgl->sge;
+
+	for (i = (nfrags != si->nr_frags); nfrags >= 2; nfrags -= 2, to++) {
+		to->len[0] = cpu_to_be32(si->frags[i].size);
+		to->len[1] = cpu_to_be32(si->frags[++i].size);
+		to->addr[0] = cpu_to_be64(addr[i]);
+		to->addr[1] = cpu_to_be64(addr[++i]);
+	}
+	if (nfrags) {
+		to->len[0] = cpu_to_be32(si->frags[i].size);
+		to->len[1] = cpu_to_be32(0);
+		to->addr[0] = cpu_to_be64(addr[i + 1]);
+	}
+	if (unlikely((u8 *)end > (u8 *)q->stat)) {
+		unsigned int part0 = (u8 *)q->stat - (u8 *)sgl->sge, part1;
+
+		if (likely(part0))
+			memcpy(sgl->sge, buf, part0);
+		part1 = (u8 *)end - (u8 *)q->stat;
+		memcpy(q->desc, (u8 *)buf + part0, part1);
+		end = (void *)q->desc + part1;
+	}
+	if ((uintptr_t)end & 8)           /* 0-pad to multiple of 16 */
+		*(u64 *)end = 0;
+}
+
+/**
+ *	ring_tx_db - check and potentially ring a Tx queue's doorbell
+ *	@adap: the adapter
+ *	@q: the Tx queue
+ *	@n: number of new descriptors to give to HW
+ *
+ *	Ring the doorbel for a Tx queue.
+ */
+static inline void ring_tx_db(struct adapter *adap, struct sge_txq *q, int n)
+{
+	wmb();            /* write descriptors before telling HW */
+	t4_write_reg(adap, MYPF_REG(SGE_PF_KDOORBELL),
+		     QID(q->cntxt_id) | PIDX(n));
+}
+
+/**
+ *	inline_tx_skb - inline a packet's data into Tx descriptors
+ *	@skb: the packet
+ *	@q: the Tx queue where the packet will be inlined
+ *	@pos: starting position in the Tx queue where to inline the packet
+ *
+ *	Inline a packet's contents directly into Tx descriptors, starting at
+ *	the given position within the Tx DMA ring.
+ *	Most of the complexity of this operation is dealing with wrap arounds
+ *	in the middle of the packet we want to inline.
+ */
+static void inline_tx_skb(const struct sk_buff *skb, const struct sge_txq *q,
+			  void *pos)
+{
+	u64 *p;
+	int left = (void *)q->stat - pos;
+
+	if (likely(skb->len <= left)) {
+		if (likely(!skb->data_len))
+			skb_copy_from_linear_data(skb, pos, skb->len);
+		else
+			skb_copy_bits(skb, 0, pos, skb->len);
+		pos += skb->len;
+	} else {
+		skb_copy_bits(skb, 0, pos, left);
+		skb_copy_bits(skb, left, q->desc, skb->len - left);
+		pos = (void *)q->desc + (skb->len - left);
+	}
+
+	/* 0-pad to multiple of 16 */
+	p = PTR_ALIGN(pos, 8);
+	if ((uintptr_t)p & 8)
+		*p = 0;
+}
+
+/*
+ * Figure out what HW csum a packet wants and return the appropriate control
+ * bits.
+ */
+static u64 hwcsum(const struct sk_buff *skb)
+{
+	int csum_type;
+	const struct iphdr *iph = ip_hdr(skb);
+
+	if (iph->version == 4) {
+		if (iph->protocol == IPPROTO_TCP)
+			csum_type = TX_CSUM_TCPIP;
+		else if (iph->protocol == IPPROTO_UDP)
+			csum_type = TX_CSUM_UDPIP;
+		else {
+nocsum:			/*
+			 * unknown protocol, disable HW csum
+			 * and hope a bad packet is detected
+			 */
+			return TXPKT_L4CSUM_DIS;
+		}
+	} else {
+		/*
+		 * this doesn't work with extension headers
+		 */
+		const struct ipv6hdr *ip6h = (const struct ipv6hdr *)iph;
+
+		if (ip6h->nexthdr == IPPROTO_TCP)
+			csum_type = TX_CSUM_TCPIP6;
+		else if (ip6h->nexthdr == IPPROTO_UDP)
+			csum_type = TX_CSUM_UDPIP6;
+		else
+			goto nocsum;
+	}
+
+	if (likely(csum_type >= TX_CSUM_TCPIP))
+		return TXPKT_CSUM_TYPE(csum_type) |
+			TXPKT_IPHDR_LEN(skb_network_header_len(skb)) |
+			TXPKT_ETHHDR_LEN(skb_network_offset(skb) - ETH_HLEN);
+	else {
+		int start = skb_transport_offset(skb);
+
+		return TXPKT_CSUM_TYPE(csum_type) | TXPKT_CSUM_START(start) |
+			TXPKT_CSUM_LOC(start + skb->csum_offset);
+	}
+}
+
+static void eth_txq_stop(struct sge_eth_txq *q)
+{
+	netif_tx_stop_queue(q->txq);
+	q->q.stops++;
+}
+
+static inline void txq_advance(struct sge_txq *q, unsigned int n)
+{
+	q->in_use += n;
+	q->pidx += n;
+	if (q->pidx >= q->size)
+		q->pidx -= q->size;
+}
+
+/**
+ *	t4_eth_xmit - add a packet to an Ethernet Tx queue
+ *	@skb: the packet
+ *	@dev: the egress net device
+ *
+ *	Add a packet to an SGE Ethernet Tx queue.  Runs with softirqs disabled.
+ */
+netdev_tx_t t4_eth_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	u32 wr_mid;
+	u64 cntrl, *end;
+	int qidx, credits;
+	unsigned int flits, ndesc;
+	struct adapter *adap;
+	struct sge_eth_txq *q;
+	const struct port_info *pi;
+	struct fw_eth_tx_pkt_wr *wr;
+	struct cpl_tx_pkt_core *cpl;
+	const struct skb_shared_info *ssi;
+	dma_addr_t addr[MAX_SKB_FRAGS + 1];
+
+	/*
+	 * The chip min packet length is 10 octets but play safe and reject
+	 * anything shorter than an Ethernet header.
+	 */
+	if (unlikely(skb->len < ETH_HLEN)) {
+out_free:	dev_kfree_skb(skb);
+		return NETDEV_TX_OK;
+	}
+
+	pi = netdev_priv(dev);
+	adap = pi->adapter;
+	qidx = skb_get_queue_mapping(skb);
+	q = &adap->sge.ethtxq[qidx + pi->first_qset];
+
+	reclaim_completed_tx(adap, &q->q, true);
+
+	flits = calc_tx_flits(skb);
+	ndesc = flits_to_desc(flits);
+	credits = txq_avail(&q->q) - ndesc;
+
+	if (unlikely(credits < 0)) {
+		eth_txq_stop(q);
+		dev_err(adap->pdev_dev,
+			"%s: Tx ring %u full while queue awake!\n",
+			dev->name, qidx);
+		return NETDEV_TX_BUSY;
+	}
+
+	if (!is_eth_imm(skb) &&
+	    unlikely(map_skb(adap->pdev_dev, skb, addr) < 0)) {
+		q->mapping_err++;
+		goto out_free;
+	}
+
+	wr_mid = FW_WR_LEN16(DIV_ROUND_UP(flits, 2));
+	if (unlikely(credits < ETHTXQ_STOP_THRES)) {
+		eth_txq_stop(q);
+		wr_mid |= FW_WR_EQUEQ | FW_WR_EQUIQ;
+	}
+
+	wr = (void *)&q->q.desc[q->q.pidx];
+	wr->equiq_to_len16 = htonl(wr_mid);
+	wr->r3 = cpu_to_be64(0);
+	end = (u64 *)wr + flits;
+
+	ssi = skb_shinfo(skb);
+	if (ssi->gso_size) {
+		struct cpl_tx_pkt_lso *lso = (void *)wr;
+		bool v6 = (ssi->gso_type & SKB_GSO_TCPV6) != 0;
+		int l3hdr_len = skb_network_header_len(skb);
+		int eth_xtra_len = skb_network_offset(skb) - ETH_HLEN;
+
+		wr->op_immdlen = htonl(FW_WR_OP(FW_ETH_TX_PKT_WR) |
+				       FW_WR_IMMDLEN(sizeof(*lso)));
+		lso->lso_ctrl = htonl(LSO_OPCODE(CPL_TX_PKT_LSO) |
+				      LSO_FIRST_SLICE | LSO_LAST_SLICE |
+				      LSO_IPV6(v6) |
+				      LSO_ETHHDR_LEN(eth_xtra_len / 4) |
+				      LSO_IPHDR_LEN(l3hdr_len / 4) |
+				      LSO_TCPHDR_LEN(tcp_hdr(skb)->doff));
+		lso->ipid_ofst = htons(0);
+		lso->mss = htons(ssi->gso_size);
+		lso->seqno_offset = htonl(0);
+		lso->len = htonl(skb->len);
+		cpl = (void *)(lso + 1);
+		cntrl = TXPKT_CSUM_TYPE(v6 ? TX_CSUM_TCPIP6 : TX_CSUM_TCPIP) |
+			TXPKT_IPHDR_LEN(l3hdr_len) |
+			TXPKT_ETHHDR_LEN(eth_xtra_len);
+		q->tso++;
+		q->tx_cso += ssi->gso_segs;
+	} else {
+		int len;
+
+		len = is_eth_imm(skb) ? skb->len + sizeof(*cpl) : sizeof(*cpl);
+		wr->op_immdlen = htonl(FW_WR_OP(FW_ETH_TX_PKT_WR) |
+				       FW_WR_IMMDLEN(len));
+		cpl = (void *)(wr + 1);
+		if (skb->ip_summed == CHECKSUM_PARTIAL) {
+			cntrl = hwcsum(skb) | TXPKT_IPCSUM_DIS;
+			q->tx_cso++;
+		} else
+			cntrl = TXPKT_L4CSUM_DIS | TXPKT_IPCSUM_DIS;
+	}
+
+	if (vlan_tx_tag_present(skb)) {
+		q->vlan_ins++;
+		cntrl |= TXPKT_VLAN_VLD | TXPKT_VLAN(vlan_tx_tag_get(skb));
+	}
+
+	cpl->ctrl0 = htonl(TXPKT_OPCODE(CPL_TX_PKT_XT) |
+			   TXPKT_INTF(pi->tx_chan) | TXPKT_PF(0));
+	cpl->pack = htons(0);
+	cpl->len = htons(skb->len);
+	cpl->ctrl1 = cpu_to_be64(cntrl);
+
+	if (is_eth_imm(skb)) {
+		inline_tx_skb(skb, &q->q, cpl + 1);
+		dev_kfree_skb(skb);
+	} else {
+		int last_desc;
+
+		write_sgl(skb, &q->q, (struct ulptx_sgl *)(cpl + 1), end, 0,
+			  addr);
+		skb_orphan(skb);
+
+		last_desc = q->q.pidx + ndesc - 1;
+		if (last_desc >= q->q.size)
+			last_desc -= q->q.size;
+		q->q.sdesc[last_desc].skb = skb;
+		q->q.sdesc[last_desc].sgl = (struct ulptx_sgl *)(cpl + 1);
+	}
+
+	txq_advance(&q->q, ndesc);
+
+	ring_tx_db(adap, &q->q, ndesc);
+	return NETDEV_TX_OK;
+}
+
+/**
+ *	reclaim_completed_tx_imm - reclaim completed control-queue Tx descs
+ *	@q: the SGE control Tx queue
+ *
+ *	This is a variant of reclaim_completed_tx() that is used for Tx queues
+ *	that send only immediate data (presently just the control queues) and
+ *	thus do not have any sk_buffs to release.
+ */
+static inline void reclaim_completed_tx_imm(struct sge_txq *q)
+{
+	int hw_cidx = ntohs(q->stat->cidx);
+	int reclaim = hw_cidx - q->cidx;
+
+	if (reclaim < 0)
+		reclaim += q->size;
+
+	q->in_use -= reclaim;
+	q->cidx = hw_cidx;
+}
+
+/**
+ *	is_imm - check whether a packet can be sent as immediate data
+ *	@skb: the packet
+ *
+ *	Returns true if a packet can be sent as a WR with immediate data.
+ */
+static inline int is_imm(const struct sk_buff *skb)
+{
+	return skb->len <= MAX_CTRL_WR_LEN;
+}
+
+/**
+ *	ctrlq_check_stop - check if a control queue is full and should stop
+ *	@q: the queue
+ *	@wr: most recent WR written to the queue
+ *
+ *	Check if a control queue has become full and should be stopped.
+ *	We clean up control queue descriptors very lazily, only when we are out.
+ *	If the queue is still full after reclaiming any completed descriptors
+ *	we suspend it and have the last WR wake it up.
+ */
+static void ctrlq_check_stop(struct sge_ctrl_txq *q, struct fw_wr_hdr *wr)
+{
+	reclaim_completed_tx_imm(&q->q);
+	if (unlikely(txq_avail(&q->q) < TXQ_STOP_THRES)) {
+		wr->lo |= htonl(FW_WR_EQUEQ | FW_WR_EQUIQ);
+		q->q.stops++;
+		q->full = 1;
+	}
+}
+
+/**
+ *	ctrl_xmit - send a packet through an SGE control Tx queue
+ *	@q: the control queue
+ *	@skb: the packet
+ *
+ *	Send a packet through an SGE control Tx queue.  Packets sent through
+ *	a control queue must fit entirely as immediate data.
+ */
+static int ctrl_xmit(struct sge_ctrl_txq *q, struct sk_buff *skb)
+{
+	unsigned int ndesc;
+	struct fw_wr_hdr *wr;
+
+	if (unlikely(!is_imm(skb))) {
+		WARN_ON(1);
+		dev_kfree_skb(skb);
+		return NET_XMIT_DROP;
+	}
+
+	ndesc = DIV_ROUND_UP(skb->len, sizeof(struct tx_desc));
+	spin_lock(&q->sendq.lock);
+
+	if (unlikely(q->full)) {
+		skb->priority = ndesc;                  /* save for restart */
+		__skb_queue_tail(&q->sendq, skb);
+		spin_unlock(&q->sendq.lock);
+		return NET_XMIT_CN;
+	}
+
+	wr = (struct fw_wr_hdr *)&q->q.desc[q->q.pidx];
+	inline_tx_skb(skb, &q->q, wr);
+
+	txq_advance(&q->q, ndesc);
+	if (unlikely(txq_avail(&q->q) < TXQ_STOP_THRES))
+		ctrlq_check_stop(q, wr);
+
+	ring_tx_db(q->adap, &q->q, ndesc);
+	spin_unlock(&q->sendq.lock);
+
+	kfree_skb(skb);
+	return NET_XMIT_SUCCESS;
+}
+
+/**
+ *	restart_ctrlq - restart a suspended control queue
+ *	@data: the control queue to restart
+ *
+ *	Resumes transmission on a suspended Tx control queue.
+ */
+static void restart_ctrlq(unsigned long data)
+{
+	struct sk_buff *skb;
+	unsigned int written = 0;
+	struct sge_ctrl_txq *q = (struct sge_ctrl_txq *)data;
+
+	spin_lock(&q->sendq.lock);
+	reclaim_completed_tx_imm(&q->q);
+	BUG_ON(txq_avail(&q->q) < TXQ_STOP_THRES);  /* q should be empty */
+
+	while ((skb = __skb_dequeue(&q->sendq)) != NULL) {
+		struct fw_wr_hdr *wr;
+		unsigned int ndesc = skb->priority;     /* previously saved */
+
+		/*
+		 * Write descriptors and free skbs outside the lock to limit
+		 * wait times.  q->full is still set so new skbs will be queued.
+		 */
+		spin_unlock(&q->sendq.lock);
+
+		wr = (struct fw_wr_hdr *)&q->q.desc[q->q.pidx];
+		inline_tx_skb(skb, &q->q, wr);
+		kfree_skb(skb);
+
+		written += ndesc;
+		txq_advance(&q->q, ndesc);
+		if (unlikely(txq_avail(&q->q) < TXQ_STOP_THRES)) {
+			unsigned long old = q->q.stops;
+
+			ctrlq_check_stop(q, wr);
+			if (q->q.stops != old) {          /* suspended anew */
+				spin_lock(&q->sendq.lock);
+				goto ringdb;
+			}
+		}
+		if (written > 16) {
+			ring_tx_db(q->adap, &q->q, written);
+			written = 0;
+		}
+		spin_lock(&q->sendq.lock);
+	}
+	q->full = 0;
+ringdb: if (written)
+		ring_tx_db(q->adap, &q->q, written);
+	spin_unlock(&q->sendq.lock);
+}
+
+/**
+ *	t4_mgmt_tx - send a management message
+ *	@adap: the adapter
+ *	@skb: the packet containing the management message
+ *
+ *	Send a management message through control queue 0.
+ */
+int t4_mgmt_tx(struct adapter *adap, struct sk_buff *skb)
+{
+	int ret;
+
+	local_bh_disable();
+	ret = ctrl_xmit(&adap->sge.ctrlq[0], skb);
+	local_bh_enable();
+	return ret;
+}
+
+/**
+ *	is_ofld_imm - check whether a packet can be sent as immediate data
+ *	@skb: the packet
+ *
+ *	Returns true if a packet can be sent as an offload WR with immediate
+ *	data.  We currently use the same limit as for Ethernet packets.
+ */
+static inline int is_ofld_imm(const struct sk_buff *skb)
+{
+	return skb->len <= MAX_IMM_TX_PKT_LEN;
+}
+
+/**
+ *	calc_tx_flits_ofld - calculate # of flits for an offload packet
+ *	@skb: the packet
+ *
+ *	Returns the number of flits needed for the given offload packet.
+ *	These packets are already fully constructed and no additional headers
+ *	will be added.
+ */
+static inline unsigned int calc_tx_flits_ofld(const struct sk_buff *skb)
+{
+	unsigned int flits, cnt;
+
+	if (is_ofld_imm(skb))
+		return DIV_ROUND_UP(skb->len, 8);
+
+	flits = skb_transport_offset(skb) / 8U;   /* headers */
+	cnt = skb_shinfo(skb)->nr_frags;
+	if (skb->tail != skb->transport_header)
+		cnt++;
+	return flits + sgl_len(cnt);
+}
+
+/**
+ *	txq_stop_maperr - stop a Tx queue due to I/O MMU exhaustion
+ *	@adap: the adapter
+ *	@q: the queue to stop
+ *
+ *	Mark a Tx queue stopped due to I/O MMU exhaustion and resulting
+ *	inability to map packets.  A periodic timer attempts to restart
+ *	queues so marked.
+ */
+static void txq_stop_maperr(struct sge_ofld_txq *q)
+{
+	q->mapping_err++;
+	q->q.stops++;
+	set_bit(q->q.cntxt_id, q->adap->sge.txq_maperr);
+}
+
+/**
+ *	ofldtxq_stop - stop an offload Tx queue that has become full
+ *	@q: the queue to stop
+ *	@skb: the packet causing the queue to become full
+ *
+ *	Stops an offload Tx queue that has become full and modifies the packet
+ *	being written to request a wakeup.
+ */
+static void ofldtxq_stop(struct sge_ofld_txq *q, struct sk_buff *skb)
+{
+	struct fw_wr_hdr *wr = (struct fw_wr_hdr *)skb->data;
+
+	wr->lo |= htonl(FW_WR_EQUEQ | FW_WR_EQUIQ);
+	q->q.stops++;
+	q->full = 1;
+}
+
+/**
+ *	service_ofldq - restart a suspended offload queue
+ *	@q: the offload queue
+ *
+ *	Services an offload Tx queue by moving packets from its packet queue
+ *	to the HW Tx ring.  The function starts and ends with the queue locked.
+ */
+static void service_ofldq(struct sge_ofld_txq *q)
+{
+	u64 *pos;
+	int credits;
+	struct sk_buff *skb;
+	unsigned int written = 0;
+	unsigned int flits, ndesc;
+
+	while ((skb = skb_peek(&q->sendq)) != NULL && !q->full) {
+		/*
+		 * We drop the lock but leave skb on sendq, thus retaining
+		 * exclusive access to the state of the queue.
+		 */
+		spin_unlock(&q->sendq.lock);
+
+		reclaim_completed_tx(q->adap, &q->q, false);
+
+		flits = skb->priority;                /* previously saved */
+		ndesc = flits_to_desc(flits);
+		credits = txq_avail(&q->q) - ndesc;
+		BUG_ON(credits < 0);
+		if (unlikely(credits < TXQ_STOP_THRES))
+			ofldtxq_stop(q, skb);
+
+		pos = (u64 *)&q->q.desc[q->q.pidx];
+		if (is_ofld_imm(skb))
+			inline_tx_skb(skb, &q->q, pos);
+		else if (map_skb(q->adap->pdev_dev, skb,
+				 (dma_addr_t *)skb->head)) {
+			txq_stop_maperr(q);
+			spin_lock(&q->sendq.lock);
+			break;
+		} else {
+			int last_desc, hdr_len = skb_transport_offset(skb);
+
+			memcpy(pos, skb->data, hdr_len);
+			write_sgl(skb, &q->q, (void *)pos + hdr_len,
+				  pos + flits, hdr_len,
+				  (dma_addr_t *)skb->head);
+#ifdef CONFIG_NEED_DMA_MAP_STATE
+			skb->dev = q->adap->port[0];
+			skb->destructor = deferred_unmap_destructor;
+#endif
+			last_desc = q->q.pidx + ndesc - 1;
+			if (last_desc >= q->q.size)
+				last_desc -= q->q.size;
+			q->q.sdesc[last_desc].skb = skb;
+		}
+
+		txq_advance(&q->q, ndesc);
+		written += ndesc;
+		if (unlikely(written > 32)) {
+			ring_tx_db(q->adap, &q->q, written);
+			written = 0;
+		}
+
+		spin_lock(&q->sendq.lock);
+		__skb_unlink(skb, &q->sendq);
+		if (is_ofld_imm(skb))
+			kfree_skb(skb);
+	}
+	if (likely(written))
+		ring_tx_db(q->adap, &q->q, written);
+}
+
+/**
+ *	ofld_xmit - send a packet through an offload queue
+ *	@q: the Tx offload queue
+ *	@skb: the packet
+ *
+ *	Send an offload packet through an SGE offload queue.
+ */
+static int ofld_xmit(struct sge_ofld_txq *q, struct sk_buff *skb)
+{
+	skb->priority = calc_tx_flits_ofld(skb);       /* save for restart */
+	spin_lock(&q->sendq.lock);
+	__skb_queue_tail(&q->sendq, skb);
+	if (q->sendq.qlen == 1)
+		service_ofldq(q);
+	spin_unlock(&q->sendq.lock);
+	return NET_XMIT_SUCCESS;
+}
+
+/**
+ *	restart_ofldq - restart a suspended offload queue
+ *	@data: the offload queue to restart
+ *
+ *	Resumes transmission on a suspended Tx offload queue.
+ */
+static void restart_ofldq(unsigned long data)
+{
+	struct sge_ofld_txq *q = (struct sge_ofld_txq *)data;
+
+	spin_lock(&q->sendq.lock);
+	q->full = 0;            /* the queue actually is completely empty now */
+	service_ofldq(q);
+	spin_unlock(&q->sendq.lock);
+}
+
+/**
+ *	skb_txq - return the Tx queue an offload packet should use
+ *	@skb: the packet
+ *
+ *	Returns the Tx queue an offload packet should use as indicated by bits
+ *	1-15 in the packet's queue_mapping.
+ */
+static inline unsigned int skb_txq(const struct sk_buff *skb)
+{
+	return skb->queue_mapping >> 1;
+}
+
+/**
+ *	is_ctrl_pkt - return whether an offload packet is a control packet
+ *	@skb: the packet
+ *
+ *	Returns whether an offload packet should use an OFLD or a CTRL
+ *	Tx queue as indicated by bit 0 in the packet's queue_mapping.
+ */
+static inline unsigned int is_ctrl_pkt(const struct sk_buff *skb)
+{
+	return skb->queue_mapping & 1;
+}
+
+static inline int ofld_send(struct adapter *adap, struct sk_buff *skb)
+{
+	unsigned int idx = skb_txq(skb);
+
+	if (unlikely(is_ctrl_pkt(skb)))
+		return ctrl_xmit(&adap->sge.ctrlq[idx], skb);
+	return ofld_xmit(&adap->sge.ofldtxq[idx], skb);
+}
+
+/**
+ *	t4_ofld_send - send an offload packet
+ *	@adap: the adapter
+ *	@skb: the packet
+ *
+ *	Sends an offload packet.  We use the packet queue_mapping to select the
+ *	appropriate Tx queue as follows: bit 0 indicates whether the packet
+ *	should be sent as regular or control, bits 1-15 select the queue.
+ */
+int t4_ofld_send(struct adapter *adap, struct sk_buff *skb)
+{
+	int ret;
+
+	local_bh_disable();
+	ret = ofld_send(adap, skb);
+	local_bh_enable();
+	return ret;
+}
+
+/**
+ *	cxgb4_ofld_send - send an offload packet
+ *	@dev: the net device
+ *	@skb: the packet
+ *
+ *	Sends an offload packet.  This is an exported version of @t4_ofld_send,
+ *	intended for ULDs.
+ */
+int cxgb4_ofld_send(struct net_device *dev, struct sk_buff *skb)
+{
+	return t4_ofld_send(netdev2adap(dev), skb);
+}
+EXPORT_SYMBOL(cxgb4_ofld_send);
+
+static inline void copy_frags(struct skb_shared_info *ssi,
+			      const struct pkt_gl *gl, unsigned int offset)
+{
+	unsigned int n;
+
+	/* usually there's just one frag */
+	ssi->frags[0].page = gl->frags[0].page;
+	ssi->frags[0].page_offset = gl->frags[0].page_offset + offset;
+	ssi->frags[0].size = gl->frags[0].size - offset;
+	ssi->nr_frags = gl->nfrags;
+	n = gl->nfrags - 1;
+	if (n)
+		memcpy(&ssi->frags[1], &gl->frags[1], n * sizeof(skb_frag_t));
+
+	/* get a reference to the last page, we don't own it */
+	get_page(gl->frags[n].page);
+}
+
+/**
+ *	cxgb4_pktgl_to_skb - build an sk_buff from a packet gather list
+ *	@gl: the gather list
+ *	@skb_len: size of sk_buff main body if it carries fragments
+ *	@pull_len: amount of data to move to the sk_buff's main body
+ *
+ *	Builds an sk_buff from the given packet gather list.  Returns the
+ *	sk_buff or %NULL if sk_buff allocation failed.
+ */
+struct sk_buff *cxgb4_pktgl_to_skb(const struct pkt_gl *gl,
+				   unsigned int skb_len, unsigned int pull_len)
+{
+	struct sk_buff *skb;
+
+	/*
+	 * Below we rely on RX_COPY_THRES being less than the smallest Rx buffer
+	 * size, which is expected since buffers are at least PAGE_SIZEd.
+	 * In this case packets up to RX_COPY_THRES have only one fragment.
+	 */
+	if (gl->tot_len <= RX_COPY_THRES) {
+		skb = dev_alloc_skb(gl->tot_len);
+		if (unlikely(!skb))
+			goto out;
+		__skb_put(skb, gl->tot_len);
+		skb_copy_to_linear_data(skb, gl->va, gl->tot_len);
+	} else {
+		skb = dev_alloc_skb(skb_len);
+		if (unlikely(!skb))
+			goto out;
+		__skb_put(skb, pull_len);
+		skb_copy_to_linear_data(skb, gl->va, pull_len);
+
+		copy_frags(skb_shinfo(skb), gl, pull_len);
+		skb->len = gl->tot_len;
+		skb->data_len = skb->len - pull_len;
+		skb->truesize += skb->data_len;
+	}
+out:	return skb;
+}
+EXPORT_SYMBOL(cxgb4_pktgl_to_skb);
+
+/**
+ *	t4_pktgl_free - free a packet gather list
+ *	@gl: the gather list
+ *
+ *	Releases the pages of a packet gather list.  We do not own the last
+ *	page on the list and do not free it.
+ */
+void t4_pktgl_free(const struct pkt_gl *gl)
+{
+	int n;
+	const skb_frag_t *p;
+
+	for (p = gl->frags, n = gl->nfrags - 1; n--; p++)
+		put_page(p->page);
+}
+
+/*
+ * Process an MPS trace packet.  Give it an unused protocol number so it won't
+ * be delivered to anyone and send it to the stack for capture.
+ */
+static noinline int handle_trace_pkt(struct adapter *adap,
+				     const struct pkt_gl *gl)
+{
+	struct sk_buff *skb;
+	struct cpl_trace_pkt *p;
+
+	skb = cxgb4_pktgl_to_skb(gl, RX_PULL_LEN, RX_PULL_LEN);
+	if (unlikely(!skb)) {
+		t4_pktgl_free(gl);
+		return 0;
+	}
+
+	p = (struct cpl_trace_pkt *)skb->data;
+	__skb_pull(skb, sizeof(*p));
+	skb_reset_mac_header(skb);
+	skb->protocol = htons(0xffff);
+	skb->dev = adap->port[0];
+	netif_receive_skb(skb);
+	return 0;
+}
+
+static void do_gro(struct sge_eth_rxq *rxq, const struct pkt_gl *gl,
+		   const struct cpl_rx_pkt *pkt)
+{
+	int ret;
+	struct sk_buff *skb;
+
+	skb = napi_get_frags(&rxq->rspq.napi);
+	if (unlikely(!skb)) {
+		t4_pktgl_free(gl);
+		rxq->stats.rx_drops++;
+		return;
+	}
+
+	copy_frags(skb_shinfo(skb), gl, RX_PKT_PAD);
+	skb->len = gl->tot_len - RX_PKT_PAD;
+	skb->data_len = skb->len;
+	skb->truesize += skb->data_len;
+	skb->ip_summed = CHECKSUM_UNNECESSARY;
+	skb_record_rx_queue(skb, rxq->rspq.idx);
+
+	if (unlikely(pkt->vlan_ex)) {
+		struct port_info *pi = netdev_priv(rxq->rspq.netdev);
+		struct vlan_group *grp = pi->vlan_grp;
+
+		rxq->stats.vlan_ex++;
+		if (likely(grp)) {
+			ret = vlan_gro_frags(&rxq->rspq.napi, grp,
+					     ntohs(pkt->vlan));
+			goto stats;
+		}
+	}
+	ret = napi_gro_frags(&rxq->rspq.napi);
+stats:	if (ret == GRO_HELD)
+		rxq->stats.lro_pkts++;
+	else if (ret == GRO_MERGED || ret == GRO_MERGED_FREE)
+		rxq->stats.lro_merged++;
+	rxq->stats.pkts++;
+	rxq->stats.rx_cso++;
+}
+
+/**
+ *	t4_ethrx_handler - process an ingress ethernet packet
+ *	@q: the response queue that received the packet
+ *	@rsp: the response queue descriptor holding the RX_PKT message
+ *	@si: the gather list of packet fragments
+ *
+ *	Process an ingress ethernet packet and deliver it to the stack.
+ */
+int t4_ethrx_handler(struct sge_rspq *q, const __be64 *rsp,
+		     const struct pkt_gl *si)
+{
+	bool csum_ok;
+	struct sk_buff *skb;
+	struct port_info *pi;
+	const struct cpl_rx_pkt *pkt;
+	struct sge_eth_rxq *rxq = container_of(q, struct sge_eth_rxq, rspq);
+
+	if (unlikely(*(u8 *)rsp == CPL_TRACE_PKT))
+		return handle_trace_pkt(q->adap, si);
+
+	pkt = (void *)&rsp[1];
+	csum_ok = pkt->csum_calc && !pkt->err_vec;
+	if ((pkt->l2info & htonl(RXF_TCP)) &&
+	    (q->netdev->features & NETIF_F_GRO) && csum_ok && !pkt->ip_frag) {
+		do_gro(rxq, si, pkt);
+		return 0;
+	}
+
+	skb = cxgb4_pktgl_to_skb(si, RX_PKT_SKB_LEN, RX_PULL_LEN);
+	if (unlikely(!skb)) {
+		t4_pktgl_free(si);
+		rxq->stats.rx_drops++;
+		return 0;
+	}
+
+	__skb_pull(skb, RX_PKT_PAD);      /* remove ethernet header padding */
+	skb->protocol = eth_type_trans(skb, q->netdev);
+	skb_record_rx_queue(skb, q->idx);
+	pi = netdev_priv(skb->dev);
+	rxq->stats.pkts++;
+
+	if (csum_ok && (pi->rx_offload & RX_CSO) &&
+	    (pkt->l2info & htonl(RXF_UDP | RXF_TCP))) {
+		if (!pkt->ip_frag)
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+		else {
+			__sum16 c = (__force __sum16)pkt->csum;
+			skb->csum = csum_unfold(c);
+			skb->ip_summed = CHECKSUM_COMPLETE;
+		}
+		rxq->stats.rx_cso++;
+	} else
+		skb->ip_summed = CHECKSUM_NONE;
+
+	if (unlikely(pkt->vlan_ex)) {
+		struct vlan_group *grp = pi->vlan_grp;
+
+		rxq->stats.vlan_ex++;
+		if (likely(grp))
+			vlan_hwaccel_receive_skb(skb, grp, ntohs(pkt->vlan));
+		else
+			dev_kfree_skb_any(skb);
+	} else
+		netif_receive_skb(skb);
+
+	return 0;
+}
+
+/**
+ *	restore_rx_bufs - put back a packet's Rx buffers
+ *	@si: the packet gather list
+ *	@q: the SGE free list
+ *	@frags: number of FL buffers to restore
+ *
+ *	Puts back on an FL the Rx buffers associated with @si.  The buffers
+ *	have already been unmapped and are left unmapped, we mark them so to
+ *	prevent further unmapping attempts.
+ *
+ *	This function undoes a series of @unmap_rx_buf calls when we find out
+ *	that the current packet can't be processed right away afterall and we
+ *	need to come back to it later.  This is a very rare event and there's
+ *	no effort to make this particularly efficient.
+ */
+static void restore_rx_bufs(const struct pkt_gl *si, struct sge_fl *q,
+			    int frags)
+{
+	struct rx_sw_desc *d;
+
+	while (frags--) {
+		if (q->cidx == 0)
+			q->cidx = q->size - 1;
+		else
+			q->cidx--;
+		d = &q->sdesc[q->cidx];
+		d->page = si->frags[frags].page;
+		d->dma_addr |= RX_UNMAPPED_BUF;
+		q->avail++;
+	}
+}
+
+/**
+ *	is_new_response - check if a response is newly written
+ *	@r: the response descriptor
+ *	@q: the response queue
+ *
+ *	Returns true if a response descriptor contains a yet unprocessed
+ *	response.
+ */
+static inline bool is_new_response(const struct rsp_ctrl *r,
+				   const struct sge_rspq *q)
+{
+	return RSPD_GEN(r->type_gen) == q->gen;
+}
+
+/**
+ *	rspq_next - advance to the next entry in a response queue
+ *	@q: the queue
+ *
+ *	Updates the state of a response queue to advance it to the next entry.
+ */
+static inline void rspq_next(struct sge_rspq *q)
+{
+	q->cur_desc = (void *)q->cur_desc + q->iqe_len;
+	if (unlikely(++q->cidx == q->size)) {
+		q->cidx = 0;
+		q->gen ^= 1;
+		q->cur_desc = q->desc;
+	}
+}
+
+/**
+ *	process_responses - process responses from an SGE response queue
+ *	@q: the ingress queue to process
+ *	@budget: how many responses can be processed in this round
+ *
+ *	Process responses from an SGE response queue up to the supplied budget.
+ *	Responses include received packets as well as control messages from FW
+ *	or HW.
+ *
+ *	Additionally choose the interrupt holdoff time for the next interrupt
+ *	on this queue.  If the system is under memory shortage use a fairly
+ *	long delay to help recovery.
+ */
+static int process_responses(struct sge_rspq *q, int budget)
+{
+	int ret, rsp_type;
+	int budget_left = budget;
+	const struct rsp_ctrl *rc;
+	struct sge_eth_rxq *rxq = container_of(q, struct sge_eth_rxq, rspq);
+
+	while (likely(budget_left)) {
+		rc = (void *)q->cur_desc + (q->iqe_len - sizeof(*rc));
+		if (!is_new_response(rc, q))
+			break;
+
+		rmb();
+		rsp_type = RSPD_TYPE(rc->type_gen);
+		if (likely(rsp_type == RSP_TYPE_FLBUF)) {
+			skb_frag_t *fp;
+			struct pkt_gl si;
+			const struct rx_sw_desc *rsd;
+			u32 len = ntohl(rc->pldbuflen_qid), bufsz, frags;
+
+			if (len & RSPD_NEWBUF) {
+				if (likely(q->offset > 0)) {
+					free_rx_bufs(q->adap, &rxq->fl, 1);
+					q->offset = 0;
+				}
+				len &= RSPD_LEN;
+			}
+			si.tot_len = len;
+
+			/* gather packet fragments */
+			for (frags = 0, fp = si.frags; ; frags++, fp++) {
+				rsd = &rxq->fl.sdesc[rxq->fl.cidx];
+				bufsz = get_buf_size(rsd);
+				fp->page = rsd->page;
+				fp->page_offset = q->offset;
+				fp->size = min(bufsz, len);
+				len -= fp->size;
+				if (!len)
+					break;
+				unmap_rx_buf(q->adap, &rxq->fl);
+			}
+
+			/*
+			 * Last buffer remains mapped so explicitly make it
+			 * coherent for CPU access.
+			 */
+			dma_sync_single_for_cpu(q->adap->pdev_dev,
+						get_buf_addr(rsd),
+						fp->size, DMA_FROM_DEVICE);
+
+			si.va = page_address(si.frags[0].page) +
+				si.frags[0].page_offset;
+			prefetch(si.va);
+
+			si.nfrags = frags + 1;
+			ret = q->handler(q, q->cur_desc, &si);
+			if (likely(ret == 0))
+				q->offset += ALIGN(fp->size, FL_ALIGN);
+			else
+				restore_rx_bufs(&si, &rxq->fl, frags);
+		} else if (likely(rsp_type == RSP_TYPE_CPL)) {
+			ret = q->handler(q, q->cur_desc, NULL);
+		} else {
+			ret = q->handler(q, (const __be64 *)rc, CXGB4_MSG_AN);
+		}
+
+		if (unlikely(ret)) {
+			/* couldn't process descriptor, back off for recovery */
+			q->next_intr_params = QINTR_TIMER_IDX(NOMEM_TMR_IDX);
+			break;
+		}
+
+		rspq_next(q);
+		budget_left--;
+	}
+
+	if (q->offset >= 0 && rxq->fl.size - rxq->fl.avail >= 16)
+		__refill_fl(q->adap, &rxq->fl);
+	return budget - budget_left;
+}
+
+/**
+ *	napi_rx_handler - the NAPI handler for Rx processing
+ *	@napi: the napi instance
+ *	@budget: how many packets we can process in this round
+ *
+ *	Handler for new data events when using NAPI.  This does not need any
+ *	locking or protection from interrupts as data interrupts are off at
+ *	this point and other adapter interrupts do not interfere (the latter
+ *	in not a concern at all with MSI-X as non-data interrupts then have
+ *	a separate handler).
+ */
+static int napi_rx_handler(struct napi_struct *napi, int budget)
+{
+	unsigned int params;
+	struct sge_rspq *q = container_of(napi, struct sge_rspq, napi);
+	int work_done = process_responses(q, budget);
+
+	if (likely(work_done < budget)) {
+		napi_complete(napi);
+		params = q->next_intr_params;
+		q->next_intr_params = q->intr_params;
+	} else
+		params = QINTR_TIMER_IDX(7);
+
+	t4_write_reg(q->adap, MYPF_REG(SGE_PF_GTS), CIDXINC(work_done) |
+		     INGRESSQID((u32)q->cntxt_id) | SEINTARM(params));
+	return work_done;
+}
+
+/*
+ * The MSI-X interrupt handler for an SGE response queue.
+ */
+irqreturn_t t4_sge_intr_msix(int irq, void *cookie)
+{
+	struct sge_rspq *q = cookie;
+
+	napi_schedule(&q->napi);
+	return IRQ_HANDLED;
+}
+
+/*
+ * Process the indirect interrupt entries in the interrupt queue and kick off
+ * NAPI for each queue that has generated an entry.
+ */
+static unsigned int process_intrq(struct adapter *adap)
+{
+	unsigned int credits;
+	const struct rsp_ctrl *rc;
+	struct sge_rspq *q = &adap->sge.intrq;
+
+	spin_lock(&adap->sge.intrq_lock);
+	for (credits = 0; ; credits++) {
+		rc = (void *)q->cur_desc + (q->iqe_len - sizeof(*rc));
+		if (!is_new_response(rc, q))
+			break;
+
+		rmb();
+		if (RSPD_TYPE(rc->type_gen) == RSP_TYPE_INTR) {
+			unsigned int qid = ntohl(rc->pldbuflen_qid);
+
+			napi_schedule(&adap->sge.ingr_map[qid]->napi);
+		}
+
+		rspq_next(q);
+	}
+
+	t4_write_reg(adap, MYPF_REG(SGE_PF_GTS), CIDXINC(credits) |
+		     INGRESSQID(q->cntxt_id) | SEINTARM(q->intr_params));
+	spin_unlock(&adap->sge.intrq_lock);
+	return credits;
+}
+
+/*
+ * The MSI interrupt handler, which handles data events from SGE response queues
+ * as well as error and other async events as they all use the same MSI vector.
+ */
+static irqreturn_t t4_intr_msi(int irq, void *cookie)
+{
+	struct adapter *adap = cookie;
+
+	t4_slow_intr_handler(adap);
+	process_intrq(adap);
+	return IRQ_HANDLED;
+}
+
+/*
+ * Interrupt handler for legacy INTx interrupts.
+ * Handles data events from SGE response queues as well as error and other
+ * async events as they all use the same interrupt line.
+ */
+static irqreturn_t t4_intr_intx(int irq, void *cookie)
+{
+	struct adapter *adap = cookie;
+
+	t4_write_reg(adap, MYPF_REG(PCIE_PF_CLI), 0);
+	if (t4_slow_intr_handler(adap) | process_intrq(adap))
+		return IRQ_HANDLED;
+	return IRQ_NONE;             /* probably shared interrupt */
+}
+
+/**
+ *	t4_intr_handler - select the top-level interrupt handler
+ *	@adap: the adapter
+ *
+ *	Selects the top-level interrupt handler based on the type of interrupts
+ *	(MSI-X, MSI, or INTx).
+ */
+irq_handler_t t4_intr_handler(struct adapter *adap)
+{
+	if (adap->flags & USING_MSIX)
+		return t4_sge_intr_msix;
+	if (adap->flags & USING_MSI)
+		return t4_intr_msi;
+	return t4_intr_intx;
+}
+
+static void sge_rx_timer_cb(unsigned long data)
+{
+	unsigned long m;
+	unsigned int i, cnt[2];
+	struct adapter *adap = (struct adapter *)data;
+	struct sge *s = &adap->sge;
+
+	for (i = 0; i < ARRAY_SIZE(s->starving_fl); i++)
+		for (m = s->starving_fl[i]; m; m &= m - 1) {
+			struct sge_eth_rxq *rxq;
+			unsigned int id = __ffs(m) + i * BITS_PER_LONG;
+			struct sge_fl *fl = s->egr_map[id];
+
+			clear_bit(id, s->starving_fl);
+			smp_mb__after_clear_bit();
+
+			if (fl_starving(fl)) {
+				rxq = container_of(fl, struct sge_eth_rxq, fl);
+				if (napi_reschedule(&rxq->rspq.napi))
+					fl->starving++;
+				else
+					set_bit(id, s->starving_fl);
+			}
+		}
+
+	t4_write_reg(adap, SGE_DEBUG_INDEX, 13);
+	cnt[0] = t4_read_reg(adap, SGE_DEBUG_DATA_HIGH);
+	cnt[1] = t4_read_reg(adap, SGE_DEBUG_DATA_LOW);
+
+	for (i = 0; i < 2; i++)
+		if (cnt[i] >= s->starve_thres) {
+			if (s->idma_state[i] || cnt[i] == 0xffffffff)
+				continue;
+			s->idma_state[i] = 1;
+			t4_write_reg(adap, SGE_DEBUG_INDEX, 11);
+			m = t4_read_reg(adap, SGE_DEBUG_DATA_LOW) >> (i * 16);
+			dev_warn(adap->pdev_dev,
+				 "SGE idma%u starvation detected for "
+				 "queue %lu\n", i, m & 0xffff);
+		} else if (s->idma_state[i])
+			s->idma_state[i] = 0;
+
+	mod_timer(&s->rx_timer, jiffies + RX_QCHECK_PERIOD);
+}
+
+static void sge_tx_timer_cb(unsigned long data)
+{
+	unsigned long m;
+	unsigned int i, budget;
+	struct adapter *adap = (struct adapter *)data;
+	struct sge *s = &adap->sge;
+
+	for (i = 0; i < ARRAY_SIZE(s->txq_maperr); i++)
+		for (m = s->txq_maperr[i]; m; m &= m - 1) {
+			unsigned long id = __ffs(m) + i * BITS_PER_LONG;
+			struct sge_ofld_txq *txq = s->egr_map[id];
+
+			clear_bit(id, s->txq_maperr);
+			tasklet_schedule(&txq->qresume_tsk);
+		}
+
+	budget = MAX_TIMER_TX_RECLAIM;
+	i = s->ethtxq_rover;
+	do {
+		struct sge_eth_txq *q = &s->ethtxq[i];
+
+		if (q->q.in_use &&
+		    time_after_eq(jiffies, q->txq->trans_start + HZ / 100) &&
+		    __netif_tx_trylock(q->txq)) {
+			int avail = reclaimable(&q->q);
+
+			if (avail) {
+				if (avail > budget)
+					avail = budget;
+
+				free_tx_desc(adap, &q->q, avail, true);
+				q->q.in_use -= avail;
+				budget -= avail;
+			}
+			__netif_tx_unlock(q->txq);
+		}
+
+		if (++i >= s->ethqsets)
+			i = 0;
+	} while (budget && i != s->ethtxq_rover);
+	s->ethtxq_rover = i;
+	mod_timer(&s->tx_timer, jiffies + (budget ? TX_QCHECK_PERIOD : 2));
+}
+
+int t4_sge_alloc_rxq(struct adapter *adap, struct sge_rspq *iq, bool fwevtq,
+		     struct net_device *dev, int intr_idx,
+		     struct sge_fl *fl, rspq_handler_t hnd)
+{
+	int ret, flsz = 0;
+	struct fw_iq_cmd c;
+	struct port_info *pi = netdev_priv(dev);
+
+	/* Size needs to be multiple of 16, including status entry. */
+	iq->size = roundup(iq->size, 16);
+
+	iq->desc = alloc_ring(adap->pdev_dev, iq->size, iq->iqe_len, 0,
+			      &iq->phys_addr, NULL, 0);
+	if (!iq->desc)
+		return -ENOMEM;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_IQ_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_CMD_EXEC |
+			    FW_IQ_CMD_PFN(0) | FW_IQ_CMD_VFN(0));
+	c.alloc_to_len16 = htonl(FW_IQ_CMD_ALLOC | FW_IQ_CMD_IQSTART(1) |
+				 FW_LEN16(c));
+	c.type_to_iqandstindex = htonl(FW_IQ_CMD_TYPE(FW_IQ_TYPE_FL_INT_CAP) |
+		FW_IQ_CMD_IQASYNCH(fwevtq) | FW_IQ_CMD_VIID(pi->viid) |
+		FW_IQ_CMD_IQANDST(intr_idx < 0) | FW_IQ_CMD_IQANUD(1) |
+		FW_IQ_CMD_IQANDSTINDEX(intr_idx >= 0 ? intr_idx :
+							-intr_idx - 1));
+	c.iqdroprss_to_iqesize = htons(FW_IQ_CMD_IQPCIECH(pi->tx_chan) |
+		FW_IQ_CMD_IQGTSMODE |
+		FW_IQ_CMD_IQINTCNTTHRESH(iq->pktcnt_idx) |
+		FW_IQ_CMD_IQESIZE(ilog2(iq->iqe_len) - 4));
+	c.iqsize = htons(iq->size);
+	c.iqaddr = cpu_to_be64(iq->phys_addr);
+
+	if (fl) {
+		fl->size = roundup(fl->size, 8);
+		fl->desc = alloc_ring(adap->pdev_dev, fl->size, sizeof(__be64),
+				      sizeof(struct rx_sw_desc), &fl->addr,
+				      &fl->sdesc, STAT_LEN);
+		if (!fl->desc)
+			goto fl_nomem;
+
+		flsz = fl->size / 8 + STAT_LEN / sizeof(struct tx_desc);
+		c.iqns_to_fl0congen = htonl(FW_IQ_CMD_FL0PACKEN |
+					    FW_IQ_CMD_FL0PADEN);
+		c.fl0dcaen_to_fl0cidxfthresh = htons(FW_IQ_CMD_FL0FBMIN(2) |
+				FW_IQ_CMD_FL0FBMAX(3));
+		c.fl0size = htons(flsz);
+		c.fl0addr = cpu_to_be64(fl->addr);
+	}
+
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+	if (ret)
+		goto err;
+
+	netif_napi_add(dev, &iq->napi, napi_rx_handler, 64);
+	iq->cur_desc = iq->desc;
+	iq->cidx = 0;
+	iq->gen = 1;
+	iq->next_intr_params = iq->intr_params;
+	iq->cntxt_id = ntohs(c.iqid);
+	iq->abs_id = ntohs(c.physiqid);
+	iq->size--;                           /* subtract status entry */
+	iq->adap = adap;
+	iq->netdev = dev;
+	iq->handler = hnd;
+
+	/* set offset to -1 to distinguish ingress queues without FL */
+	iq->offset = fl ? 0 : -1;
+
+	adap->sge.ingr_map[iq->cntxt_id] = iq;
+
+	if (fl) {
+		fl->cntxt_id = htons(c.fl0id);
+		fl->avail = fl->pend_cred = 0;
+		fl->pidx = fl->cidx = 0;
+		fl->alloc_failed = fl->large_alloc_failed = fl->starving = 0;
+		adap->sge.egr_map[fl->cntxt_id] = fl;
+		refill_fl(adap, fl, fl_cap(fl), GFP_KERNEL);
+	}
+	return 0;
+
+fl_nomem:
+	ret = -ENOMEM;
+err:
+	if (iq->desc) {
+		dma_free_coherent(adap->pdev_dev, iq->size * iq->iqe_len,
+				  iq->desc, iq->phys_addr);
+		iq->desc = NULL;
+	}
+	if (fl && fl->desc) {
+		kfree(fl->sdesc);
+		fl->sdesc = NULL;
+		dma_free_coherent(adap->pdev_dev, flsz * sizeof(struct tx_desc),
+				  fl->desc, fl->addr);
+		fl->desc = NULL;
+	}
+	return ret;
+}
+
+static void init_txq(struct adapter *adap, struct sge_txq *q, unsigned int id)
+{
+	q->in_use = 0;
+	q->cidx = q->pidx = 0;
+	q->stops = q->restarts = 0;
+	q->stat = (void *)&q->desc[q->size];
+	q->cntxt_id = id;
+	adap->sge.egr_map[id] = q;
+}
+
+int t4_sge_alloc_eth_txq(struct adapter *adap, struct sge_eth_txq *txq,
+			 struct net_device *dev, struct netdev_queue *netdevq,
+			 unsigned int iqid)
+{
+	int ret, nentries;
+	struct fw_eq_eth_cmd c;
+	struct port_info *pi = netdev_priv(dev);
+
+	/* Add status entries */
+	nentries = txq->q.size + STAT_LEN / sizeof(struct tx_desc);
+
+	txq->q.desc = alloc_ring(adap->pdev_dev, txq->q.size,
+			sizeof(struct tx_desc), sizeof(struct tx_sw_desc),
+			&txq->q.phys_addr, &txq->q.sdesc, STAT_LEN);
+	if (!txq->q.desc)
+		return -ENOMEM;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_ETH_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_CMD_EXEC |
+			    FW_EQ_ETH_CMD_PFN(0) | FW_EQ_ETH_CMD_VFN(0));
+	c.alloc_to_len16 = htonl(FW_EQ_ETH_CMD_ALLOC |
+				 FW_EQ_ETH_CMD_EQSTART | FW_LEN16(c));
+	c.viid_pkd = htonl(FW_EQ_ETH_CMD_VIID(pi->viid));
+	c.fetchszm_to_iqid = htonl(FW_EQ_ETH_CMD_HOSTFCMODE(2) |
+				   FW_EQ_ETH_CMD_PCIECHN(pi->tx_chan) |
+				   FW_EQ_ETH_CMD_IQID(iqid));
+	c.dcaen_to_eqsize = htonl(FW_EQ_ETH_CMD_FBMIN(2) |
+				  FW_EQ_ETH_CMD_FBMAX(3) |
+				  FW_EQ_ETH_CMD_CIDXFTHRESH(5) |
+				  FW_EQ_ETH_CMD_EQSIZE(nentries));
+	c.eqaddr = cpu_to_be64(txq->q.phys_addr);
+
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+	if (ret) {
+		kfree(txq->q.sdesc);
+		txq->q.sdesc = NULL;
+		dma_free_coherent(adap->pdev_dev,
+				  nentries * sizeof(struct tx_desc),
+				  txq->q.desc, txq->q.phys_addr);
+		txq->q.desc = NULL;
+		return ret;
+	}
+
+	init_txq(adap, &txq->q, FW_EQ_ETH_CMD_EQID_GET(ntohl(c.eqid_pkd)));
+	txq->txq = netdevq;
+	txq->tso = txq->tx_cso = txq->vlan_ins = 0;
+	txq->mapping_err = 0;
+	return 0;
+}
+
+int t4_sge_alloc_ctrl_txq(struct adapter *adap, struct sge_ctrl_txq *txq,
+			  struct net_device *dev, unsigned int iqid,
+			  unsigned int cmplqid)
+{
+	int ret, nentries;
+	struct fw_eq_ctrl_cmd c;
+	struct port_info *pi = netdev_priv(dev);
+
+	/* Add status entries */
+	nentries = txq->q.size + STAT_LEN / sizeof(struct tx_desc);
+
+	txq->q.desc = alloc_ring(adap->pdev_dev, nentries,
+				 sizeof(struct tx_desc), 0, &txq->q.phys_addr,
+				 NULL, 0);
+	if (!txq->q.desc)
+		return -ENOMEM;
+
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_CTRL_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_CMD_EXEC |
+			    FW_EQ_CTRL_CMD_PFN(0) | FW_EQ_CTRL_CMD_VFN(0));
+	c.alloc_to_len16 = htonl(FW_EQ_CTRL_CMD_ALLOC |
+				 FW_EQ_CTRL_CMD_EQSTART | FW_LEN16(c));
+	c.cmpliqid_eqid = htonl(FW_EQ_CTRL_CMD_CMPLIQID(cmplqid));
+	c.physeqid_pkd = htonl(0);
+	c.fetchszm_to_iqid = htonl(FW_EQ_CTRL_CMD_HOSTFCMODE(2) |
+				   FW_EQ_CTRL_CMD_PCIECHN(pi->tx_chan) |
+				   FW_EQ_CTRL_CMD_IQID(iqid));
+	c.dcaen_to_eqsize = htonl(FW_EQ_CTRL_CMD_FBMIN(2) |
+				  FW_EQ_CTRL_CMD_FBMAX(3) |
+				  FW_EQ_CTRL_CMD_CIDXFTHRESH(5) |
+				  FW_EQ_CTRL_CMD_EQSIZE(nentries));
+	c.eqaddr = cpu_to_be64(txq->q.phys_addr);
+
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+	if (ret) {
+		dma_free_coherent(adap->pdev_dev,
+				  nentries * sizeof(struct tx_desc),
+				  txq->q.desc, txq->q.phys_addr);
+		txq->q.desc = NULL;
+		return ret;
+	}
+
+	init_txq(adap, &txq->q, FW_EQ_CTRL_CMD_EQID_GET(ntohl(c.cmpliqid_eqid)));
+	txq->adap = adap;
+	skb_queue_head_init(&txq->sendq);
+	tasklet_init(&txq->qresume_tsk, restart_ctrlq, (unsigned long)txq);
+	txq->full = 0;
+	return 0;
+}
+
+int t4_sge_alloc_ofld_txq(struct adapter *adap, struct sge_ofld_txq *txq,
+			  struct net_device *dev, unsigned int iqid)
+{
+	int ret, nentries;
+	struct fw_eq_ofld_cmd c;
+	struct port_info *pi = netdev_priv(dev);
+
+	/* Add status entries */
+	nentries = txq->q.size + STAT_LEN / sizeof(struct tx_desc);
+
+	txq->q.desc = alloc_ring(adap->pdev_dev, txq->q.size,
+			sizeof(struct tx_desc), sizeof(struct tx_sw_desc),
+			&txq->q.phys_addr, &txq->q.sdesc, STAT_LEN);
+	if (!txq->q.desc)
+		return -ENOMEM;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_OFLD_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_CMD_EXEC |
+			    FW_EQ_OFLD_CMD_PFN(0) | FW_EQ_OFLD_CMD_VFN(0));
+	c.alloc_to_len16 = htonl(FW_EQ_OFLD_CMD_ALLOC |
+				 FW_EQ_OFLD_CMD_EQSTART | FW_LEN16(c));
+	c.fetchszm_to_iqid = htonl(FW_EQ_OFLD_CMD_HOSTFCMODE(2) |
+				   FW_EQ_OFLD_CMD_PCIECHN(pi->tx_chan) |
+				   FW_EQ_OFLD_CMD_IQID(iqid));
+	c.dcaen_to_eqsize = htonl(FW_EQ_OFLD_CMD_FBMIN(2) |
+				  FW_EQ_OFLD_CMD_FBMAX(3) |
+				  FW_EQ_OFLD_CMD_CIDXFTHRESH(5) |
+				  FW_EQ_OFLD_CMD_EQSIZE(nentries));
+	c.eqaddr = cpu_to_be64(txq->q.phys_addr);
+
+	ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+	if (ret) {
+		kfree(txq->q.sdesc);
+		txq->q.sdesc = NULL;
+		dma_free_coherent(adap->pdev_dev,
+				  nentries * sizeof(struct tx_desc),
+				  txq->q.desc, txq->q.phys_addr);
+		txq->q.desc = NULL;
+		return ret;
+	}
+
+	init_txq(adap, &txq->q, FW_EQ_OFLD_CMD_EQID_GET(ntohl(c.eqid_pkd)));
+	txq->adap = adap;
+	skb_queue_head_init(&txq->sendq);
+	tasklet_init(&txq->qresume_tsk, restart_ofldq, (unsigned long)txq);
+	txq->full = 0;
+	txq->mapping_err = 0;
+	return 0;
+}
+
+static void free_txq(struct adapter *adap, struct sge_txq *q)
+{
+	dma_free_coherent(adap->pdev_dev,
+			  q->size * sizeof(struct tx_desc) + STAT_LEN,
+			  q->desc, q->phys_addr);
+	q->cntxt_id = 0;
+	q->sdesc = NULL;
+	q->desc = NULL;
+}
+
+static void free_rspq_fl(struct adapter *adap, struct sge_rspq *rq,
+			 struct sge_fl *fl)
+{
+	unsigned int fl_id = fl ? fl->cntxt_id : 0xffff;
+
+	adap->sge.ingr_map[rq->cntxt_id] = NULL;
+	t4_iq_free(adap, 0, 0, 0, FW_IQ_TYPE_FL_INT_CAP, rq->cntxt_id, fl_id,
+		   0xffff);
+	dma_free_coherent(adap->pdev_dev, (rq->size + 1) * rq->iqe_len,
+			  rq->desc, rq->phys_addr);
+	netif_napi_del(&rq->napi);
+	rq->netdev = NULL;
+	rq->cntxt_id = rq->abs_id = 0;
+	rq->desc = NULL;
+
+	if (fl) {
+		free_rx_bufs(adap, fl, fl->avail);
+		dma_free_coherent(adap->pdev_dev, fl->size * 8 + STAT_LEN,
+				  fl->desc, fl->addr);
+		kfree(fl->sdesc);
+		fl->sdesc = NULL;
+		fl->cntxt_id = 0;
+		fl->desc = NULL;
+	}
+}
+
+/**
+ *	t4_free_sge_resources - free SGE resources
+ *	@adap: the adapter
+ *
+ *	Frees resources used by the SGE queue sets.
+ */
+void t4_free_sge_resources(struct adapter *adap)
+{
+	int i;
+	struct sge_eth_rxq *eq = adap->sge.ethrxq;
+	struct sge_eth_txq *etq = adap->sge.ethtxq;
+	struct sge_ofld_rxq *oq = adap->sge.ofldrxq;
+
+	/* clean up Ethernet Tx/Rx queues */
+	for (i = 0; i < adap->sge.ethqsets; i++, eq++, etq++) {
+		if (eq->rspq.desc)
+			free_rspq_fl(adap, &eq->rspq, &eq->fl);
+		if (etq->q.desc) {
+			t4_eth_eq_free(adap, 0, 0, 0, etq->q.cntxt_id);
+			free_tx_desc(adap, &etq->q, etq->q.in_use, true);
+			kfree(etq->q.sdesc);
+			free_txq(adap, &etq->q);
+		}
+	}
+
+	/* clean up RDMA and iSCSI Rx queues */
+	for (i = 0; i < adap->sge.ofldqsets; i++, oq++) {
+		if (oq->rspq.desc)
+			free_rspq_fl(adap, &oq->rspq, &oq->fl);
+	}
+	for (i = 0, oq = adap->sge.rdmarxq; i < adap->sge.rdmaqs; i++, oq++) {
+		if (oq->rspq.desc)
+			free_rspq_fl(adap, &oq->rspq, &oq->fl);
+	}
+
+	/* clean up offload Tx queues */
+	for (i = 0; i < ARRAY_SIZE(adap->sge.ofldtxq); i++) {
+		struct sge_ofld_txq *q = &adap->sge.ofldtxq[i];
+
+		if (q->q.desc) {
+			tasklet_kill(&q->qresume_tsk);
+			t4_ofld_eq_free(adap, 0, 0, 0, q->q.cntxt_id);
+			free_tx_desc(adap, &q->q, q->q.in_use, false);
+			kfree(q->q.sdesc);
+			__skb_queue_purge(&q->sendq);
+			free_txq(adap, &q->q);
+		}
+	}
+
+	/* clean up control Tx queues */
+	for (i = 0; i < ARRAY_SIZE(adap->sge.ctrlq); i++) {
+		struct sge_ctrl_txq *cq = &adap->sge.ctrlq[i];
+
+		if (cq->q.desc) {
+			tasklet_kill(&cq->qresume_tsk);
+			t4_ctrl_eq_free(adap, 0, 0, 0, cq->q.cntxt_id);
+			__skb_queue_purge(&cq->sendq);
+			free_txq(adap, &cq->q);
+		}
+	}
+
+	if (adap->sge.fw_evtq.desc)
+		free_rspq_fl(adap, &adap->sge.fw_evtq, NULL);
+
+	if (adap->sge.intrq.desc)
+		free_rspq_fl(adap, &adap->sge.intrq, NULL);
+
+	/* clear the reverse egress queue map */
+	memset(adap->sge.egr_map, 0, sizeof(adap->sge.egr_map));
+}
+
+void t4_sge_start(struct adapter *adap)
+{
+	adap->sge.ethtxq_rover = 0;
+	mod_timer(&adap->sge.rx_timer, jiffies + RX_QCHECK_PERIOD);
+	mod_timer(&adap->sge.tx_timer, jiffies + TX_QCHECK_PERIOD);
+}
+
+/**
+ *	t4_sge_stop - disable SGE operation
+ *	@adap: the adapter
+ *
+ *	Stop tasklets and timers associated with the DMA engine.  Note that
+ *	this is effective only if measures have been taken to disable any HW
+ *	events that may restart them.
+ */
+void t4_sge_stop(struct adapter *adap)
+{
+	int i;
+	struct sge *s = &adap->sge;
+
+	if (in_interrupt())  /* actions below require waiting */
+		return;
+
+	if (s->rx_timer.function)
+		del_timer_sync(&s->rx_timer);
+	if (s->tx_timer.function)
+		del_timer_sync(&s->tx_timer);
+
+	for (i = 0; i < ARRAY_SIZE(s->ofldtxq); i++) {
+		struct sge_ofld_txq *q = &s->ofldtxq[i];
+
+		if (q->q.desc)
+			tasklet_kill(&q->qresume_tsk);
+	}
+	for (i = 0; i < ARRAY_SIZE(s->ctrlq); i++) {
+		struct sge_ctrl_txq *cq = &s->ctrlq[i];
+
+		if (cq->q.desc)
+			tasklet_kill(&cq->qresume_tsk);
+	}
+}
+
+/**
+ *	t4_sge_init - initialize SGE
+ *	@adap: the adapter
+ *
+ *	Performs SGE initialization needed every time after a chip reset.
+ *	We do not initialize any of the queues here, instead the driver
+ *	top-level must request them individually.
+ */
+void t4_sge_init(struct adapter *adap)
+{
+	struct sge *s = &adap->sge;
+	unsigned int fl_align_log = ilog2(FL_ALIGN);
+
+	t4_set_reg_field(adap, SGE_CONTROL, PKTSHIFT_MASK |
+			 INGPADBOUNDARY_MASK | EGRSTATUSPAGESIZE,
+			 INGPADBOUNDARY(fl_align_log - 5) | PKTSHIFT(2) |
+			 RXPKTCPLMODE |
+			 (STAT_LEN == 128 ? EGRSTATUSPAGESIZE : 0));
+	t4_set_reg_field(adap, SGE_HOST_PAGE_SIZE, HOSTPAGESIZEPF0_MASK,
+			 HOSTPAGESIZEPF0(PAGE_SHIFT - 10));
+	t4_write_reg(adap, SGE_FL_BUFFER_SIZE0, PAGE_SIZE);
+#if FL_PG_ORDER > 0
+	t4_write_reg(adap, SGE_FL_BUFFER_SIZE1, PAGE_SIZE << FL_PG_ORDER);
+#endif
+	t4_write_reg(adap, SGE_INGRESS_RX_THRESHOLD,
+		     THRESHOLD_0(s->counter_val[0]) |
+		     THRESHOLD_1(s->counter_val[1]) |
+		     THRESHOLD_2(s->counter_val[2]) |
+		     THRESHOLD_3(s->counter_val[3]));
+	t4_write_reg(adap, SGE_TIMER_VALUE_0_AND_1,
+		     TIMERVALUE0(us_to_core_ticks(adap, s->timer_val[0])) |
+		     TIMERVALUE1(us_to_core_ticks(adap, s->timer_val[1])));
+	t4_write_reg(adap, SGE_TIMER_VALUE_2_AND_3,
+		     TIMERVALUE0(us_to_core_ticks(adap, s->timer_val[2])) |
+		     TIMERVALUE1(us_to_core_ticks(adap, s->timer_val[3])));
+	t4_write_reg(adap, SGE_TIMER_VALUE_4_AND_5,
+		     TIMERVALUE0(us_to_core_ticks(adap, s->timer_val[4])) |
+		     TIMERVALUE1(us_to_core_ticks(adap, s->timer_val[5])));
+	setup_timer(&s->rx_timer, sge_rx_timer_cb, (unsigned long)adap);
+	setup_timer(&s->tx_timer, sge_tx_timer_cb, (unsigned long)adap);
+	s->starve_thres = core_ticks_per_usec(adap) * 1000000;  /* 1 s */
+	s->idma_state[0] = s->idma_state[1] = 0;
+	spin_lock_init(&s->intrq_lock);
+}
diff --git a/drivers/net/cxgb4/t4_hw.c b/drivers/net/cxgb4/t4_hw.c
new file mode 100644
index 000000000000..a814a3afe123
--- /dev/null
+++ b/drivers/net/cxgb4/t4_hw.c
@@ -0,0 +1,3131 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/init.h>
+#include <linux/delay.h>
+#include "cxgb4.h"
+#include "t4_regs.h"
+#include "t4fw_api.h"
+
+/**
+ *	t4_wait_op_done_val - wait until an operation is completed
+ *	@adapter: the adapter performing the operation
+ *	@reg: the register to check for completion
+ *	@mask: a single-bit field within @reg that indicates completion
+ *	@polarity: the value of the field when the operation is completed
+ *	@attempts: number of check iterations
+ *	@delay: delay in usecs between iterations
+ *	@valp: where to store the value of the register at completion time
+ *
+ *	Wait until an operation is completed by checking a bit in a register
+ *	up to @attempts times.  If @valp is not NULL the value of the register
+ *	at the time it indicated completion is stored there.  Returns 0 if the
+ *	operation completes and	-EAGAIN	otherwise.
+ */
+int t4_wait_op_done_val(struct adapter *adapter, int reg, u32 mask,
+			int polarity, int attempts, int delay, u32 *valp)
+{
+	while (1) {
+		u32 val = t4_read_reg(adapter, reg);
+
+		if (!!(val & mask) == polarity) {
+			if (valp)
+				*valp = val;
+			return 0;
+		}
+		if (--attempts == 0)
+			return -EAGAIN;
+		if (delay)
+			udelay(delay);
+	}
+}
+
+static inline int t4_wait_op_done(struct adapter *adapter, int reg, u32 mask,
+				  int polarity, int attempts, int delay)
+{
+	return t4_wait_op_done_val(adapter, reg, mask, polarity, attempts,
+				   delay, NULL);
+}
+
+/**
+ *	t4_set_reg_field - set a register field to a value
+ *	@adapter: the adapter to program
+ *	@addr: the register address
+ *	@mask: specifies the portion of the register to modify
+ *	@val: the new value for the register field
+ *
+ *	Sets a register field specified by the supplied mask to the
+ *	given value.
+ */
+void t4_set_reg_field(struct adapter *adapter, unsigned int addr, u32 mask,
+		      u32 val)
+{
+	u32 v = t4_read_reg(adapter, addr) & ~mask;
+
+	t4_write_reg(adapter, addr, v | val);
+	(void) t4_read_reg(adapter, addr);      /* flush */
+}
+
+/**
+ *	t4_read_indirect - read indirectly addressed registers
+ *	@adap: the adapter
+ *	@addr_reg: register holding the indirect address
+ *	@data_reg: register holding the value of the indirect register
+ *	@vals: where the read register values are stored
+ *	@nregs: how many indirect registers to read
+ *	@start_idx: index of first indirect register to read
+ *
+ *	Reads registers that are accessed indirectly through an address/data
+ *	register pair.
+ */
+void t4_read_indirect(struct adapter *adap, unsigned int addr_reg,
+		      unsigned int data_reg, u32 *vals, unsigned int nregs,
+		      unsigned int start_idx)
+{
+	while (nregs--) {
+		t4_write_reg(adap, addr_reg, start_idx);
+		*vals++ = t4_read_reg(adap, data_reg);
+		start_idx++;
+	}
+}
+
+/**
+ *	t4_write_indirect - write indirectly addressed registers
+ *	@adap: the adapter
+ *	@addr_reg: register holding the indirect addresses
+ *	@data_reg: register holding the value for the indirect registers
+ *	@vals: values to write
+ *	@nregs: how many indirect registers to write
+ *	@start_idx: address of first indirect register to write
+ *
+ *	Writes a sequential block of registers that are accessed indirectly
+ *	through an address/data register pair.
+ */
+void t4_write_indirect(struct adapter *adap, unsigned int addr_reg,
+		       unsigned int data_reg, const u32 *vals,
+		       unsigned int nregs, unsigned int start_idx)
+{
+	while (nregs--) {
+		t4_write_reg(adap, addr_reg, start_idx++);
+		t4_write_reg(adap, data_reg, *vals++);
+	}
+}
+
+/*
+ * Get the reply to a mailbox command and store it in @rpl in big-endian order.
+ */
+static void get_mbox_rpl(struct adapter *adap, __be64 *rpl, int nflit,
+			 u32 mbox_addr)
+{
+	for ( ; nflit; nflit--, mbox_addr += 8)
+		*rpl++ = cpu_to_be64(t4_read_reg64(adap, mbox_addr));
+}
+
+/*
+ * Handle a FW assertion reported in a mailbox.
+ */
+static void fw_asrt(struct adapter *adap, u32 mbox_addr)
+{
+	struct fw_debug_cmd asrt;
+
+	get_mbox_rpl(adap, (__be64 *)&asrt, sizeof(asrt) / 8, mbox_addr);
+	dev_alert(adap->pdev_dev,
+		  "FW assertion at %.16s:%u, val0 %#x, val1 %#x\n",
+		  asrt.u.assert.filename_0_7, ntohl(asrt.u.assert.line),
+		  ntohl(asrt.u.assert.x), ntohl(asrt.u.assert.y));
+}
+
+static void dump_mbox(struct adapter *adap, int mbox, u32 data_reg)
+{
+	dev_err(adap->pdev_dev,
+		"mbox %d: %llx %llx %llx %llx %llx %llx %llx %llx\n", mbox,
+		(unsigned long long)t4_read_reg64(adap, data_reg),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 8),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 16),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 24),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 32),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 40),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 48),
+		(unsigned long long)t4_read_reg64(adap, data_reg + 56));
+}
+
+/**
+ *	t4_wr_mbox_meat - send a command to FW through the given mailbox
+ *	@adap: the adapter
+ *	@mbox: index of the mailbox to use
+ *	@cmd: the command to write
+ *	@size: command length in bytes
+ *	@rpl: where to optionally store the reply
+ *	@sleep_ok: if true we may sleep while awaiting command completion
+ *
+ *	Sends the given command to FW through the selected mailbox and waits
+ *	for the FW to execute the command.  If @rpl is not %NULL it is used to
+ *	store the FW's reply to the command.  The command and its optional
+ *	reply are of the same length.  FW can take up to %FW_CMD_MAX_TIMEOUT ms
+ *	to respond.  @sleep_ok determines whether we may sleep while awaiting
+ *	the response.  If sleeping is allowed we use progressive backoff
+ *	otherwise we spin.
+ *
+ *	The return value is 0 on success or a negative errno on failure.  A
+ *	failure can happen either because we are not able to execute the
+ *	command or FW executes it but signals an error.  In the latter case
+ *	the return value is the error code indicated by FW (negated).
+ */
+int t4_wr_mbox_meat(struct adapter *adap, int mbox, const void *cmd, int size,
+		    void *rpl, bool sleep_ok)
+{
+	static int delay[] = {
+		1, 1, 3, 5, 10, 10, 20, 50, 100, 200
+	};
+
+	u32 v;
+	u64 res;
+	int i, ms, delay_idx;
+	const __be64 *p = cmd;
+	u32 data_reg = PF_REG(mbox, CIM_PF_MAILBOX_DATA);
+	u32 ctl_reg = PF_REG(mbox, CIM_PF_MAILBOX_CTRL);
+
+	if ((size & 15) || size > MBOX_LEN)
+		return -EINVAL;
+
+	v = MBOWNER_GET(t4_read_reg(adap, ctl_reg));
+	for (i = 0; v == MBOX_OWNER_NONE && i < 3; i++)
+		v = MBOWNER_GET(t4_read_reg(adap, ctl_reg));
+
+	if (v != MBOX_OWNER_DRV)
+		return v ? -EBUSY : -ETIMEDOUT;
+
+	for (i = 0; i < size; i += 8)
+		t4_write_reg64(adap, data_reg + i, be64_to_cpu(*p++));
+
+	t4_write_reg(adap, ctl_reg, MBMSGVALID | MBOWNER(MBOX_OWNER_FW));
+	t4_read_reg(adap, ctl_reg);          /* flush write */
+
+	delay_idx = 0;
+	ms = delay[0];
+
+	for (i = 0; i < FW_CMD_MAX_TIMEOUT; i += ms) {
+		if (sleep_ok) {
+			ms = delay[delay_idx];  /* last element may repeat */
+			if (delay_idx < ARRAY_SIZE(delay) - 1)
+				delay_idx++;
+			msleep(ms);
+		} else
+			mdelay(ms);
+
+		v = t4_read_reg(adap, ctl_reg);
+		if (MBOWNER_GET(v) == MBOX_OWNER_DRV) {
+			if (!(v & MBMSGVALID)) {
+				t4_write_reg(adap, ctl_reg, 0);
+				continue;
+			}
+
+			res = t4_read_reg64(adap, data_reg);
+			if (FW_CMD_OP_GET(res >> 32) == FW_DEBUG_CMD) {
+				fw_asrt(adap, data_reg);
+				res = FW_CMD_RETVAL(EIO);
+			} else if (rpl)
+				get_mbox_rpl(adap, rpl, size / 8, data_reg);
+
+			if (FW_CMD_RETVAL_GET((int)res))
+				dump_mbox(adap, mbox, data_reg);
+			t4_write_reg(adap, ctl_reg, 0);
+			return -FW_CMD_RETVAL_GET((int)res);
+		}
+	}
+
+	dump_mbox(adap, mbox, data_reg);
+	dev_err(adap->pdev_dev, "command %#x in mailbox %d timed out\n",
+		*(const u8 *)cmd, mbox);
+	return -ETIMEDOUT;
+}
+
+/**
+ *	t4_mc_read - read from MC through backdoor accesses
+ *	@adap: the adapter
+ *	@addr: address of first byte requested
+ *	@data: 64 bytes of data containing the requested address
+ *	@ecc: where to store the corresponding 64-bit ECC word
+ *
+ *	Read 64 bytes of data from MC starting at a 64-byte-aligned address
+ *	that covers the requested address @addr.  If @parity is not %NULL it
+ *	is assigned the 64-bit ECC word for the read data.
+ */
+int t4_mc_read(struct adapter *adap, u32 addr, __be32 *data, u64 *ecc)
+{
+	int i;
+
+	if (t4_read_reg(adap, MC_BIST_CMD) & START_BIST)
+		return -EBUSY;
+	t4_write_reg(adap, MC_BIST_CMD_ADDR, addr & ~0x3fU);
+	t4_write_reg(adap, MC_BIST_CMD_LEN, 64);
+	t4_write_reg(adap, MC_BIST_DATA_PATTERN, 0xc);
+	t4_write_reg(adap, MC_BIST_CMD, BIST_OPCODE(1) | START_BIST |
+		     BIST_CMD_GAP(1));
+	i = t4_wait_op_done(adap, MC_BIST_CMD, START_BIST, 0, 10, 1);
+	if (i)
+		return i;
+
+#define MC_DATA(i) MC_BIST_STATUS_REG(MC_BIST_STATUS_RDATA, i)
+
+	for (i = 15; i >= 0; i--)
+		*data++ = htonl(t4_read_reg(adap, MC_DATA(i)));
+	if (ecc)
+		*ecc = t4_read_reg64(adap, MC_DATA(16));
+#undef MC_DATA
+	return 0;
+}
+
+/**
+ *	t4_edc_read - read from EDC through backdoor accesses
+ *	@adap: the adapter
+ *	@idx: which EDC to access
+ *	@addr: address of first byte requested
+ *	@data: 64 bytes of data containing the requested address
+ *	@ecc: where to store the corresponding 64-bit ECC word
+ *
+ *	Read 64 bytes of data from EDC starting at a 64-byte-aligned address
+ *	that covers the requested address @addr.  If @parity is not %NULL it
+ *	is assigned the 64-bit ECC word for the read data.
+ */
+int t4_edc_read(struct adapter *adap, int idx, u32 addr, __be32 *data, u64 *ecc)
+{
+	int i;
+
+	idx *= EDC_STRIDE;
+	if (t4_read_reg(adap, EDC_BIST_CMD + idx) & START_BIST)
+		return -EBUSY;
+	t4_write_reg(adap, EDC_BIST_CMD_ADDR + idx, addr & ~0x3fU);
+	t4_write_reg(adap, EDC_BIST_CMD_LEN + idx, 64);
+	t4_write_reg(adap, EDC_BIST_DATA_PATTERN + idx, 0xc);
+	t4_write_reg(adap, EDC_BIST_CMD + idx,
+		     BIST_OPCODE(1) | BIST_CMD_GAP(1) | START_BIST);
+	i = t4_wait_op_done(adap, EDC_BIST_CMD + idx, START_BIST, 0, 10, 1);
+	if (i)
+		return i;
+
+#define EDC_DATA(i) (EDC_BIST_STATUS_REG(EDC_BIST_STATUS_RDATA, i) + idx)
+
+	for (i = 15; i >= 0; i--)
+		*data++ = htonl(t4_read_reg(adap, EDC_DATA(i)));
+	if (ecc)
+		*ecc = t4_read_reg64(adap, EDC_DATA(16));
+#undef EDC_DATA
+	return 0;
+}
+
+#define VPD_ENTRY(name, len) \
+	u8 name##_kword[2]; u8 name##_len; u8 name##_data[len]
+
+/*
+ * Partial EEPROM Vital Product Data structure.  Includes only the ID and
+ * VPD-R sections.
+ */
+struct t4_vpd {
+	u8  id_tag;
+	u8  id_len[2];
+	u8  id_data[ID_LEN];
+	u8  vpdr_tag;
+	u8  vpdr_len[2];
+	VPD_ENTRY(pn, 16);                     /* part number */
+	VPD_ENTRY(ec, EC_LEN);                 /* EC level */
+	VPD_ENTRY(sn, SERNUM_LEN);             /* serial number */
+	VPD_ENTRY(na, 12);                     /* MAC address base */
+	VPD_ENTRY(port_type, 8);               /* port types */
+	VPD_ENTRY(gpio, 14);                   /* GPIO usage */
+	VPD_ENTRY(cclk, 6);                    /* core clock */
+	VPD_ENTRY(port_addr, 8);               /* port MDIO addresses */
+	VPD_ENTRY(rv, 1);                      /* csum */
+	u32 pad;                  /* for multiple-of-4 sizing and alignment */
+};
+
+#define EEPROM_STAT_ADDR   0x7bfc
+#define VPD_BASE           0
+
+/**
+ *	t4_seeprom_wp - enable/disable EEPROM write protection
+ *	@adapter: the adapter
+ *	@enable: whether to enable or disable write protection
+ *
+ *	Enables or disables write protection on the serial EEPROM.
+ */
+int t4_seeprom_wp(struct adapter *adapter, bool enable)
+{
+	unsigned int v = enable ? 0xc : 0;
+	int ret = pci_write_vpd(adapter->pdev, EEPROM_STAT_ADDR, 4, &v);
+	return ret < 0 ? ret : 0;
+}
+
+/**
+ *	get_vpd_params - read VPD parameters from VPD EEPROM
+ *	@adapter: adapter to read
+ *	@p: where to store the parameters
+ *
+ *	Reads card parameters stored in VPD EEPROM.
+ */
+static int get_vpd_params(struct adapter *adapter, struct vpd_params *p)
+{
+	int ret;
+	struct t4_vpd vpd;
+	u8 *q = (u8 *)&vpd, csum;
+
+	ret = pci_read_vpd(adapter->pdev, VPD_BASE, sizeof(vpd), &vpd);
+	if (ret < 0)
+		return ret;
+
+	for (csum = 0; q <= vpd.rv_data; q++)
+		csum += *q;
+
+	if (csum) {
+		dev_err(adapter->pdev_dev,
+			"corrupted VPD EEPROM, actual csum %u\n", csum);
+		return -EINVAL;
+	}
+
+	p->cclk = simple_strtoul(vpd.cclk_data, NULL, 10);
+	memcpy(p->id, vpd.id_data, sizeof(vpd.id_data));
+	strim(p->id);
+	memcpy(p->ec, vpd.ec_data, sizeof(vpd.ec_data));
+	strim(p->ec);
+	memcpy(p->sn, vpd.sn_data, sizeof(vpd.sn_data));
+	strim(p->sn);
+	return 0;
+}
+
+/* serial flash and firmware constants */
+enum {
+	SF_ATTEMPTS = 10,             /* max retries for SF operations */
+
+	/* flash command opcodes */
+	SF_PROG_PAGE    = 2,          /* program page */
+	SF_WR_DISABLE   = 4,          /* disable writes */
+	SF_RD_STATUS    = 5,          /* read status register */
+	SF_WR_ENABLE    = 6,          /* enable writes */
+	SF_RD_DATA_FAST = 0xb,        /* read flash */
+	SF_ERASE_SECTOR = 0xd8,       /* erase sector */
+
+	FW_START_SEC = 8,             /* first flash sector for FW */
+	FW_END_SEC = 15,              /* last flash sector for FW */
+	FW_IMG_START = FW_START_SEC * SF_SEC_SIZE,
+	FW_MAX_SIZE = (FW_END_SEC - FW_START_SEC + 1) * SF_SEC_SIZE,
+};
+
+/**
+ *	sf1_read - read data from the serial flash
+ *	@adapter: the adapter
+ *	@byte_cnt: number of bytes to read
+ *	@cont: whether another operation will be chained
+ *	@lock: whether to lock SF for PL access only
+ *	@valp: where to store the read data
+ *
+ *	Reads up to 4 bytes of data from the serial flash.  The location of
+ *	the read needs to be specified prior to calling this by issuing the
+ *	appropriate commands to the serial flash.
+ */
+static int sf1_read(struct adapter *adapter, unsigned int byte_cnt, int cont,
+		    int lock, u32 *valp)
+{
+	int ret;
+
+	if (!byte_cnt || byte_cnt > 4)
+		return -EINVAL;
+	if (t4_read_reg(adapter, SF_OP) & BUSY)
+		return -EBUSY;
+	cont = cont ? SF_CONT : 0;
+	lock = lock ? SF_LOCK : 0;
+	t4_write_reg(adapter, SF_OP, lock | cont | BYTECNT(byte_cnt - 1));
+	ret = t4_wait_op_done(adapter, SF_OP, BUSY, 0, SF_ATTEMPTS, 5);
+	if (!ret)
+		*valp = t4_read_reg(adapter, SF_DATA);
+	return ret;
+}
+
+/**
+ *	sf1_write - write data to the serial flash
+ *	@adapter: the adapter
+ *	@byte_cnt: number of bytes to write
+ *	@cont: whether another operation will be chained
+ *	@lock: whether to lock SF for PL access only
+ *	@val: value to write
+ *
+ *	Writes up to 4 bytes of data to the serial flash.  The location of
+ *	the write needs to be specified prior to calling this by issuing the
+ *	appropriate commands to the serial flash.
+ */
+static int sf1_write(struct adapter *adapter, unsigned int byte_cnt, int cont,
+		     int lock, u32 val)
+{
+	if (!byte_cnt || byte_cnt > 4)
+		return -EINVAL;
+	if (t4_read_reg(adapter, SF_OP) & BUSY)
+		return -EBUSY;
+	cont = cont ? SF_CONT : 0;
+	lock = lock ? SF_LOCK : 0;
+	t4_write_reg(adapter, SF_DATA, val);
+	t4_write_reg(adapter, SF_OP, lock |
+		     cont | BYTECNT(byte_cnt - 1) | OP_WR);
+	return t4_wait_op_done(adapter, SF_OP, BUSY, 0, SF_ATTEMPTS, 5);
+}
+
+/**
+ *	flash_wait_op - wait for a flash operation to complete
+ *	@adapter: the adapter
+ *	@attempts: max number of polls of the status register
+ *	@delay: delay between polls in ms
+ *
+ *	Wait for a flash operation to complete by polling the status register.
+ */
+static int flash_wait_op(struct adapter *adapter, int attempts, int delay)
+{
+	int ret;
+	u32 status;
+
+	while (1) {
+		if ((ret = sf1_write(adapter, 1, 1, 1, SF_RD_STATUS)) != 0 ||
+		    (ret = sf1_read(adapter, 1, 0, 1, &status)) != 0)
+			return ret;
+		if (!(status & 1))
+			return 0;
+		if (--attempts == 0)
+			return -EAGAIN;
+		if (delay)
+			msleep(delay);
+	}
+}
+
+/**
+ *	t4_read_flash - read words from serial flash
+ *	@adapter: the adapter
+ *	@addr: the start address for the read
+ *	@nwords: how many 32-bit words to read
+ *	@data: where to store the read data
+ *	@byte_oriented: whether to store data as bytes or as words
+ *
+ *	Read the specified number of 32-bit words from the serial flash.
+ *	If @byte_oriented is set the read data is stored as a byte array
+ *	(i.e., big-endian), otherwise as 32-bit words in the platform's
+ *	natural endianess.
+ */
+int t4_read_flash(struct adapter *adapter, unsigned int addr,
+		  unsigned int nwords, u32 *data, int byte_oriented)
+{
+	int ret;
+
+	if (addr + nwords * sizeof(u32) > SF_SIZE || (addr & 3))
+		return -EINVAL;
+
+	addr = swab32(addr) | SF_RD_DATA_FAST;
+
+	if ((ret = sf1_write(adapter, 4, 1, 0, addr)) != 0 ||
+	    (ret = sf1_read(adapter, 1, 1, 0, data)) != 0)
+		return ret;
+
+	for ( ; nwords; nwords--, data++) {
+		ret = sf1_read(adapter, 4, nwords > 1, nwords == 1, data);
+		if (nwords == 1)
+			t4_write_reg(adapter, SF_OP, 0);    /* unlock SF */
+		if (ret)
+			return ret;
+		if (byte_oriented)
+			*data = htonl(*data);
+	}
+	return 0;
+}
+
+/**
+ *	t4_write_flash - write up to a page of data to the serial flash
+ *	@adapter: the adapter
+ *	@addr: the start address to write
+ *	@n: length of data to write in bytes
+ *	@data: the data to write
+ *
+ *	Writes up to a page of data (256 bytes) to the serial flash starting
+ *	at the given address.  All the data must be written to the same page.
+ */
+static int t4_write_flash(struct adapter *adapter, unsigned int addr,
+			  unsigned int n, const u8 *data)
+{
+	int ret;
+	u32 buf[64];
+	unsigned int i, c, left, val, offset = addr & 0xff;
+
+	if (addr >= SF_SIZE || offset + n > SF_PAGE_SIZE)
+		return -EINVAL;
+
+	val = swab32(addr) | SF_PROG_PAGE;
+
+	if ((ret = sf1_write(adapter, 1, 0, 1, SF_WR_ENABLE)) != 0 ||
+	    (ret = sf1_write(adapter, 4, 1, 1, val)) != 0)
+		goto unlock;
+
+	for (left = n; left; left -= c) {
+		c = min(left, 4U);
+		for (val = 0, i = 0; i < c; ++i)
+			val = (val << 8) + *data++;
+
+		ret = sf1_write(adapter, c, c != left, 1, val);
+		if (ret)
+			goto unlock;
+	}
+	ret = flash_wait_op(adapter, 5, 1);
+	if (ret)
+		goto unlock;
+
+	t4_write_reg(adapter, SF_OP, 0);    /* unlock SF */
+
+	/* Read the page to verify the write succeeded */
+	ret = t4_read_flash(adapter, addr & ~0xff, ARRAY_SIZE(buf), buf, 1);
+	if (ret)
+		return ret;
+
+	if (memcmp(data - n, (u8 *)buf + offset, n)) {
+		dev_err(adapter->pdev_dev,
+			"failed to correctly write the flash page at %#x\n",
+			addr);
+		return -EIO;
+	}
+	return 0;
+
+unlock:
+	t4_write_reg(adapter, SF_OP, 0);    /* unlock SF */
+	return ret;
+}
+
+/**
+ *	get_fw_version - read the firmware version
+ *	@adapter: the adapter
+ *	@vers: where to place the version
+ *
+ *	Reads the FW version from flash.
+ */
+static int get_fw_version(struct adapter *adapter, u32 *vers)
+{
+	return t4_read_flash(adapter,
+			     FW_IMG_START + offsetof(struct fw_hdr, fw_ver), 1,
+			     vers, 0);
+}
+
+/**
+ *	get_tp_version - read the TP microcode version
+ *	@adapter: the adapter
+ *	@vers: where to place the version
+ *
+ *	Reads the TP microcode version from flash.
+ */
+static int get_tp_version(struct adapter *adapter, u32 *vers)
+{
+	return t4_read_flash(adapter, FW_IMG_START + offsetof(struct fw_hdr,
+							      tp_microcode_ver),
+			     1, vers, 0);
+}
+
+/**
+ *	t4_check_fw_version - check if the FW is compatible with this driver
+ *	@adapter: the adapter
+ *
+ *	Checks if an adapter's FW is compatible with the driver.  Returns 0
+ *	if there's exact match, a negative error if the version could not be
+ *	read or there's a major version mismatch, and a positive value if the
+ *	expected major version is found but there's a minor version mismatch.
+ */
+int t4_check_fw_version(struct adapter *adapter)
+{
+	u32 api_vers[2];
+	int ret, major, minor, micro;
+
+	ret = get_fw_version(adapter, &adapter->params.fw_vers);
+	if (!ret)
+		ret = get_tp_version(adapter, &adapter->params.tp_vers);
+	if (!ret)
+		ret = t4_read_flash(adapter,
+			FW_IMG_START + offsetof(struct fw_hdr, intfver_nic),
+			2, api_vers, 1);
+	if (ret)
+		return ret;
+
+	major = FW_HDR_FW_VER_MAJOR_GET(adapter->params.fw_vers);
+	minor = FW_HDR_FW_VER_MINOR_GET(adapter->params.fw_vers);
+	micro = FW_HDR_FW_VER_MICRO_GET(adapter->params.fw_vers);
+	memcpy(adapter->params.api_vers, api_vers,
+	       sizeof(adapter->params.api_vers));
+
+	if (major != FW_VERSION_MAJOR) {            /* major mismatch - fail */
+		dev_err(adapter->pdev_dev,
+			"card FW has major version %u, driver wants %u\n",
+			major, FW_VERSION_MAJOR);
+		return -EINVAL;
+	}
+
+	if (minor == FW_VERSION_MINOR && micro == FW_VERSION_MICRO)
+		return 0;                                   /* perfect match */
+
+	/* Minor/micro version mismatch.  Report it but often it's OK. */
+	return 1;
+}
+
+/**
+ *	t4_flash_erase_sectors - erase a range of flash sectors
+ *	@adapter: the adapter
+ *	@start: the first sector to erase
+ *	@end: the last sector to erase
+ *
+ *	Erases the sectors in the given inclusive range.
+ */
+static int t4_flash_erase_sectors(struct adapter *adapter, int start, int end)
+{
+	int ret = 0;
+
+	while (start <= end) {
+		if ((ret = sf1_write(adapter, 1, 0, 1, SF_WR_ENABLE)) != 0 ||
+		    (ret = sf1_write(adapter, 4, 0, 1,
+				     SF_ERASE_SECTOR | (start << 8))) != 0 ||
+		    (ret = flash_wait_op(adapter, 5, 500)) != 0) {
+			dev_err(adapter->pdev_dev,
+				"erase of flash sector %d failed, error %d\n",
+				start, ret);
+			break;
+		}
+		start++;
+	}
+	t4_write_reg(adapter, SF_OP, 0);    /* unlock SF */
+	return ret;
+}
+
+/**
+ *	t4_load_fw - download firmware
+ *	@adap: the adapter
+ *	@fw_data: the firmware image to write
+ *	@size: image size
+ *
+ *	Write the supplied firmware image to the card's serial flash.
+ */
+int t4_load_fw(struct adapter *adap, const u8 *fw_data, unsigned int size)
+{
+	u32 csum;
+	int ret, addr;
+	unsigned int i;
+	u8 first_page[SF_PAGE_SIZE];
+	const u32 *p = (const u32 *)fw_data;
+	const struct fw_hdr *hdr = (const struct fw_hdr *)fw_data;
+
+	if (!size) {
+		dev_err(adap->pdev_dev, "FW image has no data\n");
+		return -EINVAL;
+	}
+	if (size & 511) {
+		dev_err(adap->pdev_dev,
+			"FW image size not multiple of 512 bytes\n");
+		return -EINVAL;
+	}
+	if (ntohs(hdr->len512) * 512 != size) {
+		dev_err(adap->pdev_dev,
+			"FW image size differs from size in FW header\n");
+		return -EINVAL;
+	}
+	if (size > FW_MAX_SIZE) {
+		dev_err(adap->pdev_dev, "FW image too large, max is %u bytes\n",
+			FW_MAX_SIZE);
+		return -EFBIG;
+	}
+
+	for (csum = 0, i = 0; i < size / sizeof(csum); i++)
+		csum += ntohl(p[i]);
+
+	if (csum != 0xffffffff) {
+		dev_err(adap->pdev_dev,
+			"corrupted firmware image, checksum %#x\n", csum);
+		return -EINVAL;
+	}
+
+	i = DIV_ROUND_UP(size, SF_SEC_SIZE);        /* # of sectors spanned */
+	ret = t4_flash_erase_sectors(adap, FW_START_SEC, FW_START_SEC + i - 1);
+	if (ret)
+		goto out;
+
+	/*
+	 * We write the correct version at the end so the driver can see a bad
+	 * version if the FW write fails.  Start by writing a copy of the
+	 * first page with a bad version.
+	 */
+	memcpy(first_page, fw_data, SF_PAGE_SIZE);
+	((struct fw_hdr *)first_page)->fw_ver = htonl(0xffffffff);
+	ret = t4_write_flash(adap, FW_IMG_START, SF_PAGE_SIZE, first_page);
+	if (ret)
+		goto out;
+
+	addr = FW_IMG_START;
+	for (size -= SF_PAGE_SIZE; size; size -= SF_PAGE_SIZE) {
+		addr += SF_PAGE_SIZE;
+		fw_data += SF_PAGE_SIZE;
+		ret = t4_write_flash(adap, addr, SF_PAGE_SIZE, fw_data);
+		if (ret)
+			goto out;
+	}
+
+	ret = t4_write_flash(adap,
+			     FW_IMG_START + offsetof(struct fw_hdr, fw_ver),
+			     sizeof(hdr->fw_ver), (const u8 *)&hdr->fw_ver);
+out:
+	if (ret)
+		dev_err(adap->pdev_dev, "firmware download failed, error %d\n",
+			ret);
+	return ret;
+}
+
+#define ADVERT_MASK (FW_PORT_CAP_SPEED_100M | FW_PORT_CAP_SPEED_1G |\
+		     FW_PORT_CAP_SPEED_10G | FW_PORT_CAP_ANEG)
+
+/**
+ *	t4_link_start - apply link configuration to MAC/PHY
+ *	@phy: the PHY to setup
+ *	@mac: the MAC to setup
+ *	@lc: the requested link configuration
+ *
+ *	Set up a port's MAC and PHY according to a desired link configuration.
+ *	- If the PHY can auto-negotiate first decide what to advertise, then
+ *	  enable/disable auto-negotiation as desired, and reset.
+ *	- If the PHY does not auto-negotiate just reset it.
+ *	- If auto-negotiation is off set the MAC to the proper speed/duplex/FC,
+ *	  otherwise do it later based on the outcome of auto-negotiation.
+ */
+int t4_link_start(struct adapter *adap, unsigned int mbox, unsigned int port,
+		  struct link_config *lc)
+{
+	struct fw_port_cmd c;
+	unsigned int fc = 0, mdi = FW_PORT_MDI(FW_PORT_MDI_AUTO);
+
+	lc->link_ok = 0;
+	if (lc->requested_fc & PAUSE_RX)
+		fc |= FW_PORT_CAP_FC_RX;
+	if (lc->requested_fc & PAUSE_TX)
+		fc |= FW_PORT_CAP_FC_TX;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_portid = htonl(FW_CMD_OP(FW_PORT_CMD) | FW_CMD_REQUEST |
+			       FW_CMD_EXEC | FW_PORT_CMD_PORTID(port));
+	c.action_to_len16 = htonl(FW_PORT_CMD_ACTION(FW_PORT_ACTION_L1_CFG) |
+				  FW_LEN16(c));
+
+	if (!(lc->supported & FW_PORT_CAP_ANEG)) {
+		c.u.l1cfg.rcap = htonl((lc->supported & ADVERT_MASK) | fc);
+		lc->fc = lc->requested_fc & (PAUSE_RX | PAUSE_TX);
+	} else if (lc->autoneg == AUTONEG_DISABLE) {
+		c.u.l1cfg.rcap = htonl(lc->requested_speed | fc | mdi);
+		lc->fc = lc->requested_fc & (PAUSE_RX | PAUSE_TX);
+	} else
+		c.u.l1cfg.rcap = htonl(lc->advertising | fc | mdi);
+
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_restart_aneg - restart autonegotiation
+ *	@adap: the adapter
+ *	@mbox: mbox to use for the FW command
+ *	@port: the port id
+ *
+ *	Restarts autonegotiation for the selected port.
+ */
+int t4_restart_aneg(struct adapter *adap, unsigned int mbox, unsigned int port)
+{
+	struct fw_port_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_portid = htonl(FW_CMD_OP(FW_PORT_CMD) | FW_CMD_REQUEST |
+			       FW_CMD_EXEC | FW_PORT_CMD_PORTID(port));
+	c.action_to_len16 = htonl(FW_PORT_CMD_ACTION(FW_PORT_ACTION_L1_CFG) |
+				  FW_LEN16(c));
+	c.u.l1cfg.rcap = htonl(FW_PORT_CAP_ANEG);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_set_vlan_accel - configure HW VLAN extraction
+ *	@adap: the adapter
+ *	@ports: bitmap of adapter ports to operate on
+ *	@on: enable (1) or disable (0) HW VLAN extraction
+ *
+ *	Enables or disables HW extraction of VLAN tags for the ports specified
+ *	by @ports.  @ports is a bitmap with the ith bit designating the port
+ *	associated with the ith adapter channel.
+ */
+void t4_set_vlan_accel(struct adapter *adap, unsigned int ports, int on)
+{
+	ports <<= VLANEXTENABLE_SHIFT;
+	t4_set_reg_field(adap, TP_OUT_CONFIG, ports, on ? ports : 0);
+}
+
+struct intr_info {
+	unsigned int mask;       /* bits to check in interrupt status */
+	const char *msg;         /* message to print or NULL */
+	short stat_idx;          /* stat counter to increment or -1 */
+	unsigned short fatal;    /* whether the condition reported is fatal */
+};
+
+/**
+ *	t4_handle_intr_status - table driven interrupt handler
+ *	@adapter: the adapter that generated the interrupt
+ *	@reg: the interrupt status register to process
+ *	@acts: table of interrupt actions
+ *
+ *	A table driven interrupt handler that applies a set of masks to an
+ *	interrupt status word and performs the corresponding actions if the
+ *	interrupts described by the mask have occured.  The actions include
+ *	optionally emitting a warning or alert message.  The table is terminated
+ *	by an entry specifying mask 0.  Returns the number of fatal interrupt
+ *	conditions.
+ */
+static int t4_handle_intr_status(struct adapter *adapter, unsigned int reg,
+				 const struct intr_info *acts)
+{
+	int fatal = 0;
+	unsigned int mask = 0;
+	unsigned int status = t4_read_reg(adapter, reg);
+
+	for ( ; acts->mask; ++acts) {
+		if (!(status & acts->mask))
+			continue;
+		if (acts->fatal) {
+			fatal++;
+			dev_alert(adapter->pdev_dev, "%s (0x%x)\n", acts->msg,
+				  status & acts->mask);
+		} else if (acts->msg && printk_ratelimit())
+			dev_warn(adapter->pdev_dev, "%s (0x%x)\n", acts->msg,
+				 status & acts->mask);
+		mask |= acts->mask;
+	}
+	status &= mask;
+	if (status)                           /* clear processed interrupts */
+		t4_write_reg(adapter, reg, status);
+	return fatal;
+}
+
+/*
+ * Interrupt handler for the PCIE module.
+ */
+static void pcie_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info sysbus_intr_info[] = {
+		{ RNPP, "RXNP array parity error", -1, 1 },
+		{ RPCP, "RXPC array parity error", -1, 1 },
+		{ RCIP, "RXCIF array parity error", -1, 1 },
+		{ RCCP, "Rx completions control array parity error", -1, 1 },
+		{ RFTP, "RXFT array parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info pcie_port_intr_info[] = {
+		{ TPCP, "TXPC array parity error", -1, 1 },
+		{ TNPP, "TXNP array parity error", -1, 1 },
+		{ TFTP, "TXFT array parity error", -1, 1 },
+		{ TCAP, "TXCA array parity error", -1, 1 },
+		{ TCIP, "TXCIF array parity error", -1, 1 },
+		{ RCAP, "RXCA array parity error", -1, 1 },
+		{ OTDD, "outbound request TLP discarded", -1, 1 },
+		{ RDPE, "Rx data parity error", -1, 1 },
+		{ TDUE, "Tx uncorrectable data error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info pcie_intr_info[] = {
+		{ MSIADDRLPERR, "MSI AddrL parity error", -1, 1 },
+		{ MSIADDRHPERR, "MSI AddrH parity error", -1, 1 },
+		{ MSIDATAPERR, "MSI data parity error", -1, 1 },
+		{ MSIXADDRLPERR, "MSI-X AddrL parity error", -1, 1 },
+		{ MSIXADDRHPERR, "MSI-X AddrH parity error", -1, 1 },
+		{ MSIXDATAPERR, "MSI-X data parity error", -1, 1 },
+		{ MSIXDIPERR, "MSI-X DI parity error", -1, 1 },
+		{ PIOCPLPERR, "PCI PIO completion FIFO parity error", -1, 1 },
+		{ PIOREQPERR, "PCI PIO request FIFO parity error", -1, 1 },
+		{ TARTAGPERR, "PCI PCI target tag FIFO parity error", -1, 1 },
+		{ CCNTPERR, "PCI CMD channel count parity error", -1, 1 },
+		{ CREQPERR, "PCI CMD channel request parity error", -1, 1 },
+		{ CRSPPERR, "PCI CMD channel response parity error", -1, 1 },
+		{ DCNTPERR, "PCI DMA channel count parity error", -1, 1 },
+		{ DREQPERR, "PCI DMA channel request parity error", -1, 1 },
+		{ DRSPPERR, "PCI DMA channel response parity error", -1, 1 },
+		{ HCNTPERR, "PCI HMA channel count parity error", -1, 1 },
+		{ HREQPERR, "PCI HMA channel request parity error", -1, 1 },
+		{ HRSPPERR, "PCI HMA channel response parity error", -1, 1 },
+		{ CFGSNPPERR, "PCI config snoop FIFO parity error", -1, 1 },
+		{ FIDPERR, "PCI FID parity error", -1, 1 },
+		{ INTXCLRPERR, "PCI INTx clear parity error", -1, 1 },
+		{ MATAGPERR, "PCI MA tag parity error", -1, 1 },
+		{ PIOTAGPERR, "PCI PIO tag parity error", -1, 1 },
+		{ RXCPLPERR, "PCI Rx completion parity error", -1, 1 },
+		{ RXWRPERR, "PCI Rx write parity error", -1, 1 },
+		{ RPLPERR, "PCI replay buffer parity error", -1, 1 },
+		{ PCIESINT, "PCI core secondary fault", -1, 1 },
+		{ PCIEPINT, "PCI core primary fault", -1, 1 },
+		{ UNXSPLCPLERR, "PCI unexpected split completion error", -1, 0 },
+		{ 0 }
+	};
+
+	int fat;
+
+	fat = t4_handle_intr_status(adapter,
+				    PCIE_CORE_UTL_SYSTEM_BUS_AGENT_STATUS,
+				    sysbus_intr_info) +
+	      t4_handle_intr_status(adapter,
+				    PCIE_CORE_UTL_PCI_EXPRESS_PORT_STATUS,
+				    pcie_port_intr_info) +
+	      t4_handle_intr_status(adapter, PCIE_INT_CAUSE, pcie_intr_info);
+	if (fat)
+		t4_fatal_err(adapter);
+}
+
+/*
+ * TP interrupt handler.
+ */
+static void tp_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info tp_intr_info[] = {
+		{ 0x3fffffff, "TP parity error", -1, 1 },
+		{ FLMTXFLSTEMPTY, "TP out of Tx pages", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, TP_INT_CAUSE, tp_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * SGE interrupt handler.
+ */
+static void sge_intr_handler(struct adapter *adapter)
+{
+	u64 v;
+
+	static struct intr_info sge_intr_info[] = {
+		{ ERR_CPL_EXCEED_IQE_SIZE,
+		  "SGE received CPL exceeding IQE size", -1, 1 },
+		{ ERR_INVALID_CIDX_INC,
+		  "SGE GTS CIDX increment too large", -1, 0 },
+		{ ERR_CPL_OPCODE_0, "SGE received 0-length CPL", -1, 0 },
+		{ ERR_DROPPED_DB, "SGE doorbell dropped", -1, 0 },
+		{ ERR_DATA_CPL_ON_HIGH_QID1 | ERR_DATA_CPL_ON_HIGH_QID0,
+		  "SGE IQID > 1023 received CPL for FL", -1, 0 },
+		{ ERR_BAD_DB_PIDX3, "SGE DBP 3 pidx increment too large", -1,
+		  0 },
+		{ ERR_BAD_DB_PIDX2, "SGE DBP 2 pidx increment too large", -1,
+		  0 },
+		{ ERR_BAD_DB_PIDX1, "SGE DBP 1 pidx increment too large", -1,
+		  0 },
+		{ ERR_BAD_DB_PIDX0, "SGE DBP 0 pidx increment too large", -1,
+		  0 },
+		{ ERR_ING_CTXT_PRIO,
+		  "SGE too many priority ingress contexts", -1, 0 },
+		{ ERR_EGR_CTXT_PRIO,
+		  "SGE too many priority egress contexts", -1, 0 },
+		{ INGRESS_SIZE_ERR, "SGE illegal ingress QID", -1, 0 },
+		{ EGRESS_SIZE_ERR, "SGE illegal egress QID", -1, 0 },
+		{ 0 }
+	};
+
+	v = (u64)t4_read_reg(adapter, SGE_INT_CAUSE1) |
+	    ((u64)t4_read_reg(adapter, SGE_INT_CAUSE2) << 32);
+	if (v) {
+		dev_alert(adapter->pdev_dev, "SGE parity error (%#llx)\n",
+			 (unsigned long long)v);
+		t4_write_reg(adapter, SGE_INT_CAUSE1, v);
+		t4_write_reg(adapter, SGE_INT_CAUSE2, v >> 32);
+	}
+
+	if (t4_handle_intr_status(adapter, SGE_INT_CAUSE3, sge_intr_info) ||
+	    v != 0)
+		t4_fatal_err(adapter);
+}
+
+/*
+ * CIM interrupt handler.
+ */
+static void cim_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info cim_intr_info[] = {
+		{ PREFDROPINT, "CIM control register prefetch drop", -1, 1 },
+		{ OBQPARERR, "CIM OBQ parity error", -1, 1 },
+		{ IBQPARERR, "CIM IBQ parity error", -1, 1 },
+		{ MBUPPARERR, "CIM mailbox uP parity error", -1, 1 },
+		{ MBHOSTPARERR, "CIM mailbox host parity error", -1, 1 },
+		{ TIEQINPARERRINT, "CIM TIEQ outgoing parity error", -1, 1 },
+		{ TIEQOUTPARERRINT, "CIM TIEQ incoming parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info cim_upintr_info[] = {
+		{ RSVDSPACEINT, "CIM reserved space access", -1, 1 },
+		{ ILLTRANSINT, "CIM illegal transaction", -1, 1 },
+		{ ILLWRINT, "CIM illegal write", -1, 1 },
+		{ ILLRDINT, "CIM illegal read", -1, 1 },
+		{ ILLRDBEINT, "CIM illegal read BE", -1, 1 },
+		{ ILLWRBEINT, "CIM illegal write BE", -1, 1 },
+		{ SGLRDBOOTINT, "CIM single read from boot space", -1, 1 },
+		{ SGLWRBOOTINT, "CIM single write to boot space", -1, 1 },
+		{ BLKWRBOOTINT, "CIM block write to boot space", -1, 1 },
+		{ SGLRDFLASHINT, "CIM single read from flash space", -1, 1 },
+		{ SGLWRFLASHINT, "CIM single write to flash space", -1, 1 },
+		{ BLKWRFLASHINT, "CIM block write to flash space", -1, 1 },
+		{ SGLRDEEPROMINT, "CIM single EEPROM read", -1, 1 },
+		{ SGLWREEPROMINT, "CIM single EEPROM write", -1, 1 },
+		{ BLKRDEEPROMINT, "CIM block EEPROM read", -1, 1 },
+		{ BLKWREEPROMINT, "CIM block EEPROM write", -1, 1 },
+		{ SGLRDCTLINT , "CIM single read from CTL space", -1, 1 },
+		{ SGLWRCTLINT , "CIM single write to CTL space", -1, 1 },
+		{ BLKRDCTLINT , "CIM block read from CTL space", -1, 1 },
+		{ BLKWRCTLINT , "CIM block write to CTL space", -1, 1 },
+		{ SGLRDPLINT , "CIM single read from PL space", -1, 1 },
+		{ SGLWRPLINT , "CIM single write to PL space", -1, 1 },
+		{ BLKRDPLINT , "CIM block read from PL space", -1, 1 },
+		{ BLKWRPLINT , "CIM block write to PL space", -1, 1 },
+		{ REQOVRLOOKUPINT , "CIM request FIFO overwrite", -1, 1 },
+		{ RSPOVRLOOKUPINT , "CIM response FIFO overwrite", -1, 1 },
+		{ TIMEOUTINT , "CIM PIF timeout", -1, 1 },
+		{ TIMEOUTMAINT , "CIM PIF MA timeout", -1, 1 },
+		{ 0 }
+	};
+
+	int fat;
+
+	fat = t4_handle_intr_status(adapter, CIM_HOST_INT_CAUSE,
+				    cim_intr_info) +
+	      t4_handle_intr_status(adapter, CIM_HOST_UPACC_INT_CAUSE,
+				    cim_upintr_info);
+	if (fat)
+		t4_fatal_err(adapter);
+}
+
+/*
+ * ULP RX interrupt handler.
+ */
+static void ulprx_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info ulprx_intr_info[] = {
+		{ 0x7fffff, "ULPRX parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, ULP_RX_INT_CAUSE, ulprx_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * ULP TX interrupt handler.
+ */
+static void ulptx_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info ulptx_intr_info[] = {
+		{ PBL_BOUND_ERR_CH3, "ULPTX channel 3 PBL out of bounds", -1,
+		  0 },
+		{ PBL_BOUND_ERR_CH2, "ULPTX channel 2 PBL out of bounds", -1,
+		  0 },
+		{ PBL_BOUND_ERR_CH1, "ULPTX channel 1 PBL out of bounds", -1,
+		  0 },
+		{ PBL_BOUND_ERR_CH0, "ULPTX channel 0 PBL out of bounds", -1,
+		  0 },
+		{ 0xfffffff, "ULPTX parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, ULP_TX_INT_CAUSE, ulptx_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * PM TX interrupt handler.
+ */
+static void pmtx_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info pmtx_intr_info[] = {
+		{ PCMD_LEN_OVFL0, "PMTX channel 0 pcmd too large", -1, 1 },
+		{ PCMD_LEN_OVFL1, "PMTX channel 1 pcmd too large", -1, 1 },
+		{ PCMD_LEN_OVFL2, "PMTX channel 2 pcmd too large", -1, 1 },
+		{ ZERO_C_CMD_ERROR, "PMTX 0-length pcmd", -1, 1 },
+		{ PMTX_FRAMING_ERROR, "PMTX framing error", -1, 1 },
+		{ OESPI_PAR_ERROR, "PMTX oespi parity error", -1, 1 },
+		{ DB_OPTIONS_PAR_ERROR, "PMTX db_options parity error", -1, 1 },
+		{ ICSPI_PAR_ERROR, "PMTX icspi parity error", -1, 1 },
+		{ C_PCMD_PAR_ERROR, "PMTX c_pcmd parity error", -1, 1},
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, PM_TX_INT_CAUSE, pmtx_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * PM RX interrupt handler.
+ */
+static void pmrx_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info pmrx_intr_info[] = {
+		{ ZERO_E_CMD_ERROR, "PMRX 0-length pcmd", -1, 1 },
+		{ PMRX_FRAMING_ERROR, "PMRX framing error", -1, 1 },
+		{ OCSPI_PAR_ERROR, "PMRX ocspi parity error", -1, 1 },
+		{ DB_OPTIONS_PAR_ERROR, "PMRX db_options parity error", -1, 1 },
+		{ IESPI_PAR_ERROR, "PMRX iespi parity error", -1, 1 },
+		{ E_PCMD_PAR_ERROR, "PMRX e_pcmd parity error", -1, 1},
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, PM_RX_INT_CAUSE, pmrx_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * CPL switch interrupt handler.
+ */
+static void cplsw_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info cplsw_intr_info[] = {
+		{ CIM_OP_MAP_PERR, "CPLSW CIM op_map parity error", -1, 1 },
+		{ CIM_OVFL_ERROR, "CPLSW CIM overflow", -1, 1 },
+		{ TP_FRAMING_ERROR, "CPLSW TP framing error", -1, 1 },
+		{ SGE_FRAMING_ERROR, "CPLSW SGE framing error", -1, 1 },
+		{ CIM_FRAMING_ERROR, "CPLSW CIM framing error", -1, 1 },
+		{ ZERO_SWITCH_ERROR, "CPLSW no-switch error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adapter, CPL_INTR_CAUSE, cplsw_intr_info))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * LE interrupt handler.
+ */
+static void le_intr_handler(struct adapter *adap)
+{
+	static struct intr_info le_intr_info[] = {
+		{ LIPMISS, "LE LIP miss", -1, 0 },
+		{ LIP0, "LE 0 LIP error", -1, 0 },
+		{ PARITYERR, "LE parity error", -1, 1 },
+		{ UNKNOWNCMD, "LE unknown command", -1, 1 },
+		{ REQQPARERR, "LE request queue parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adap, LE_DB_INT_CAUSE, le_intr_info))
+		t4_fatal_err(adap);
+}
+
+/*
+ * MPS interrupt handler.
+ */
+static void mps_intr_handler(struct adapter *adapter)
+{
+	static struct intr_info mps_rx_intr_info[] = {
+		{ 0xffffff, "MPS Rx parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_tx_intr_info[] = {
+		{ TPFIFO, "MPS Tx TP FIFO parity error", -1, 1 },
+		{ NCSIFIFO, "MPS Tx NC-SI FIFO parity error", -1, 1 },
+		{ TXDATAFIFO, "MPS Tx data FIFO parity error", -1, 1 },
+		{ TXDESCFIFO, "MPS Tx desc FIFO parity error", -1, 1 },
+		{ BUBBLE, "MPS Tx underflow", -1, 1 },
+		{ SECNTERR, "MPS Tx SOP/EOP error", -1, 1 },
+		{ FRMERR, "MPS Tx framing error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_trc_intr_info[] = {
+		{ FILTMEM, "MPS TRC filter parity error", -1, 1 },
+		{ PKTFIFO, "MPS TRC packet FIFO parity error", -1, 1 },
+		{ MISCPERR, "MPS TRC misc parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_stat_sram_intr_info[] = {
+		{ 0x1fffff, "MPS statistics SRAM parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_stat_tx_intr_info[] = {
+		{ 0xfffff, "MPS statistics Tx FIFO parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_stat_rx_intr_info[] = {
+		{ 0xffffff, "MPS statistics Rx FIFO parity error", -1, 1 },
+		{ 0 }
+	};
+	static struct intr_info mps_cls_intr_info[] = {
+		{ MATCHSRAM, "MPS match SRAM parity error", -1, 1 },
+		{ MATCHTCAM, "MPS match TCAM parity error", -1, 1 },
+		{ HASHSRAM, "MPS hash SRAM parity error", -1, 1 },
+		{ 0 }
+	};
+
+	int fat;
+
+	fat = t4_handle_intr_status(adapter, MPS_RX_PERR_INT_CAUSE,
+				    mps_rx_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_TX_INT_CAUSE,
+				    mps_tx_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_TRC_INT_CAUSE,
+				    mps_trc_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_STAT_PERR_INT_CAUSE_SRAM,
+				    mps_stat_sram_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_STAT_PERR_INT_CAUSE_TX_FIFO,
+				    mps_stat_tx_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_STAT_PERR_INT_CAUSE_RX_FIFO,
+				    mps_stat_rx_intr_info) +
+	      t4_handle_intr_status(adapter, MPS_CLS_INT_CAUSE,
+				    mps_cls_intr_info);
+
+	t4_write_reg(adapter, MPS_INT_CAUSE, CLSINT | TRCINT |
+		     RXINT | TXINT | STATINT);
+	t4_read_reg(adapter, MPS_INT_CAUSE);                    /* flush */
+	if (fat)
+		t4_fatal_err(adapter);
+}
+
+#define MEM_INT_MASK (PERR_INT_CAUSE | ECC_CE_INT_CAUSE | ECC_UE_INT_CAUSE)
+
+/*
+ * EDC/MC interrupt handler.
+ */
+static void mem_intr_handler(struct adapter *adapter, int idx)
+{
+	static const char name[3][5] = { "EDC0", "EDC1", "MC" };
+
+	unsigned int addr, cnt_addr, v;
+
+	if (idx <= MEM_EDC1) {
+		addr = EDC_REG(EDC_INT_CAUSE, idx);
+		cnt_addr = EDC_REG(EDC_ECC_STATUS, idx);
+	} else {
+		addr = MC_INT_CAUSE;
+		cnt_addr = MC_ECC_STATUS;
+	}
+
+	v = t4_read_reg(adapter, addr) & MEM_INT_MASK;
+	if (v & PERR_INT_CAUSE)
+		dev_alert(adapter->pdev_dev, "%s FIFO parity error\n",
+			  name[idx]);
+	if (v & ECC_CE_INT_CAUSE) {
+		u32 cnt = ECC_CECNT_GET(t4_read_reg(adapter, cnt_addr));
+
+		t4_write_reg(adapter, cnt_addr, ECC_CECNT_MASK);
+		if (printk_ratelimit())
+			dev_warn(adapter->pdev_dev,
+				 "%u %s correctable ECC data error%s\n",
+				 cnt, name[idx], cnt > 1 ? "s" : "");
+	}
+	if (v & ECC_UE_INT_CAUSE)
+		dev_alert(adapter->pdev_dev,
+			  "%s uncorrectable ECC data error\n", name[idx]);
+
+	t4_write_reg(adapter, addr, v);
+	if (v & (PERR_INT_CAUSE | ECC_UE_INT_CAUSE))
+		t4_fatal_err(adapter);
+}
+
+/*
+ * MA interrupt handler.
+ */
+static void ma_intr_handler(struct adapter *adap)
+{
+	u32 v, status = t4_read_reg(adap, MA_INT_CAUSE);
+
+	if (status & MEM_PERR_INT_CAUSE)
+		dev_alert(adap->pdev_dev,
+			  "MA parity error, parity status %#x\n",
+			  t4_read_reg(adap, MA_PARITY_ERROR_STATUS));
+	if (status & MEM_WRAP_INT_CAUSE) {
+		v = t4_read_reg(adap, MA_INT_WRAP_STATUS);
+		dev_alert(adap->pdev_dev, "MA address wrap-around error by "
+			  "client %u to address %#x\n",
+			  MEM_WRAP_CLIENT_NUM_GET(v),
+			  MEM_WRAP_ADDRESS_GET(v) << 4);
+	}
+	t4_write_reg(adap, MA_INT_CAUSE, status);
+	t4_fatal_err(adap);
+}
+
+/*
+ * SMB interrupt handler.
+ */
+static void smb_intr_handler(struct adapter *adap)
+{
+	static struct intr_info smb_intr_info[] = {
+		{ MSTTXFIFOPARINT, "SMB master Tx FIFO parity error", -1, 1 },
+		{ MSTRXFIFOPARINT, "SMB master Rx FIFO parity error", -1, 1 },
+		{ SLVFIFOPARINT, "SMB slave FIFO parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adap, SMB_INT_CAUSE, smb_intr_info))
+		t4_fatal_err(adap);
+}
+
+/*
+ * NC-SI interrupt handler.
+ */
+static void ncsi_intr_handler(struct adapter *adap)
+{
+	static struct intr_info ncsi_intr_info[] = {
+		{ CIM_DM_PRTY_ERR, "NC-SI CIM parity error", -1, 1 },
+		{ MPS_DM_PRTY_ERR, "NC-SI MPS parity error", -1, 1 },
+		{ TXFIFO_PRTY_ERR, "NC-SI Tx FIFO parity error", -1, 1 },
+		{ RXFIFO_PRTY_ERR, "NC-SI Rx FIFO parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adap, NCSI_INT_CAUSE, ncsi_intr_info))
+		t4_fatal_err(adap);
+}
+
+/*
+ * XGMAC interrupt handler.
+ */
+static void xgmac_intr_handler(struct adapter *adap, int port)
+{
+	u32 v = t4_read_reg(adap, PORT_REG(port, XGMAC_PORT_INT_CAUSE));
+
+	v &= TXFIFO_PRTY_ERR | RXFIFO_PRTY_ERR;
+	if (!v)
+		return;
+
+	if (v & TXFIFO_PRTY_ERR)
+		dev_alert(adap->pdev_dev, "XGMAC %d Tx FIFO parity error\n",
+			  port);
+	if (v & RXFIFO_PRTY_ERR)
+		dev_alert(adap->pdev_dev, "XGMAC %d Rx FIFO parity error\n",
+			  port);
+	t4_write_reg(adap, PORT_REG(port, XGMAC_PORT_INT_CAUSE), v);
+	t4_fatal_err(adap);
+}
+
+/*
+ * PL interrupt handler.
+ */
+static void pl_intr_handler(struct adapter *adap)
+{
+	static struct intr_info pl_intr_info[] = {
+		{ FATALPERR, "T4 fatal parity error", -1, 1 },
+		{ PERRVFID, "PL VFID_MAP parity error", -1, 1 },
+		{ 0 }
+	};
+
+	if (t4_handle_intr_status(adap, PL_PL_INT_CAUSE, pl_intr_info))
+		t4_fatal_err(adap);
+}
+
+#define PF_INTR_MASK (PFSW | PFCIM)
+#define GLBL_INTR_MASK (CIM | MPS | PL | PCIE | MC | EDC0 | \
+		EDC1 | LE | TP | MA | PM_TX | PM_RX | ULP_RX | \
+		CPL_SWITCH | SGE | ULP_TX)
+
+/**
+ *	t4_slow_intr_handler - control path interrupt handler
+ *	@adapter: the adapter
+ *
+ *	T4 interrupt handler for non-data global interrupt events, e.g., errors.
+ *	The designation 'slow' is because it involves register reads, while
+ *	data interrupts typically don't involve any MMIOs.
+ */
+int t4_slow_intr_handler(struct adapter *adapter)
+{
+	u32 cause = t4_read_reg(adapter, PL_INT_CAUSE);
+
+	if (!(cause & GLBL_INTR_MASK))
+		return 0;
+	if (cause & CIM)
+		cim_intr_handler(adapter);
+	if (cause & MPS)
+		mps_intr_handler(adapter);
+	if (cause & NCSI)
+		ncsi_intr_handler(adapter);
+	if (cause & PL)
+		pl_intr_handler(adapter);
+	if (cause & SMB)
+		smb_intr_handler(adapter);
+	if (cause & XGMAC0)
+		xgmac_intr_handler(adapter, 0);
+	if (cause & XGMAC1)
+		xgmac_intr_handler(adapter, 1);
+	if (cause & XGMAC_KR0)
+		xgmac_intr_handler(adapter, 2);
+	if (cause & XGMAC_KR1)
+		xgmac_intr_handler(adapter, 3);
+	if (cause & PCIE)
+		pcie_intr_handler(adapter);
+	if (cause & MC)
+		mem_intr_handler(adapter, MEM_MC);
+	if (cause & EDC0)
+		mem_intr_handler(adapter, MEM_EDC0);
+	if (cause & EDC1)
+		mem_intr_handler(adapter, MEM_EDC1);
+	if (cause & LE)
+		le_intr_handler(adapter);
+	if (cause & TP)
+		tp_intr_handler(adapter);
+	if (cause & MA)
+		ma_intr_handler(adapter);
+	if (cause & PM_TX)
+		pmtx_intr_handler(adapter);
+	if (cause & PM_RX)
+		pmrx_intr_handler(adapter);
+	if (cause & ULP_RX)
+		ulprx_intr_handler(adapter);
+	if (cause & CPL_SWITCH)
+		cplsw_intr_handler(adapter);
+	if (cause & SGE)
+		sge_intr_handler(adapter);
+	if (cause & ULP_TX)
+		ulptx_intr_handler(adapter);
+
+	/* Clear the interrupts just processed for which we are the master. */
+	t4_write_reg(adapter, PL_INT_CAUSE, cause & GLBL_INTR_MASK);
+	(void) t4_read_reg(adapter, PL_INT_CAUSE); /* flush */
+	return 1;
+}
+
+/**
+ *	t4_intr_enable - enable interrupts
+ *	@adapter: the adapter whose interrupts should be enabled
+ *
+ *	Enable PF-specific interrupts for the calling function and the top-level
+ *	interrupt concentrator for global interrupts.  Interrupts are already
+ *	enabled at each module,	here we just enable the roots of the interrupt
+ *	hierarchies.
+ *
+ *	Note: this function should be called only when the driver manages
+ *	non PF-specific interrupts from the various HW modules.  Only one PCI
+ *	function at a time should be doing this.
+ */
+void t4_intr_enable(struct adapter *adapter)
+{
+	u32 pf = SOURCEPF_GET(t4_read_reg(adapter, PL_WHOAMI));
+
+	t4_write_reg(adapter, SGE_INT_ENABLE3, ERR_CPL_EXCEED_IQE_SIZE |
+		     ERR_INVALID_CIDX_INC | ERR_CPL_OPCODE_0 |
+		     ERR_DROPPED_DB | ERR_DATA_CPL_ON_HIGH_QID1 |
+		     ERR_DATA_CPL_ON_HIGH_QID0 | ERR_BAD_DB_PIDX3 |
+		     ERR_BAD_DB_PIDX2 | ERR_BAD_DB_PIDX1 |
+		     ERR_BAD_DB_PIDX0 | ERR_ING_CTXT_PRIO |
+		     ERR_EGR_CTXT_PRIO | INGRESS_SIZE_ERR |
+		     EGRESS_SIZE_ERR);
+	t4_write_reg(adapter, MYPF_REG(PL_PF_INT_ENABLE), PF_INTR_MASK);
+	t4_set_reg_field(adapter, PL_INT_MAP0, 0, 1 << pf);
+}
+
+/**
+ *	t4_intr_disable - disable interrupts
+ *	@adapter: the adapter whose interrupts should be disabled
+ *
+ *	Disable interrupts.  We only disable the top-level interrupt
+ *	concentrators.  The caller must be a PCI function managing global
+ *	interrupts.
+ */
+void t4_intr_disable(struct adapter *adapter)
+{
+	u32 pf = SOURCEPF_GET(t4_read_reg(adapter, PL_WHOAMI));
+
+	t4_write_reg(adapter, MYPF_REG(PL_PF_INT_ENABLE), 0);
+	t4_set_reg_field(adapter, PL_INT_MAP0, 1 << pf, 0);
+}
+
+/**
+ *	t4_intr_clear - clear all interrupts
+ *	@adapter: the adapter whose interrupts should be cleared
+ *
+ *	Clears all interrupts.  The caller must be a PCI function managing
+ *	global interrupts.
+ */
+void t4_intr_clear(struct adapter *adapter)
+{
+	static const unsigned int cause_reg[] = {
+		SGE_INT_CAUSE1, SGE_INT_CAUSE2, SGE_INT_CAUSE3,
+		PCIE_CORE_UTL_SYSTEM_BUS_AGENT_STATUS,
+		PCIE_CORE_UTL_PCI_EXPRESS_PORT_STATUS,
+		PCIE_NONFAT_ERR, PCIE_INT_CAUSE,
+		MC_INT_CAUSE,
+		MA_INT_WRAP_STATUS, MA_PARITY_ERROR_STATUS, MA_INT_CAUSE,
+		EDC_INT_CAUSE, EDC_REG(EDC_INT_CAUSE, 1),
+		CIM_HOST_INT_CAUSE, CIM_HOST_UPACC_INT_CAUSE,
+		MYPF_REG(CIM_PF_HOST_INT_CAUSE),
+		TP_INT_CAUSE,
+		ULP_RX_INT_CAUSE, ULP_TX_INT_CAUSE,
+		PM_RX_INT_CAUSE, PM_TX_INT_CAUSE,
+		MPS_RX_PERR_INT_CAUSE,
+		CPL_INTR_CAUSE,
+		MYPF_REG(PL_PF_INT_CAUSE),
+		PL_PL_INT_CAUSE,
+		LE_DB_INT_CAUSE,
+	};
+
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(cause_reg); ++i)
+		t4_write_reg(adapter, cause_reg[i], 0xffffffff);
+
+	t4_write_reg(adapter, PL_INT_CAUSE, GLBL_INTR_MASK);
+	(void) t4_read_reg(adapter, PL_INT_CAUSE);          /* flush */
+}
+
+/**
+ *	hash_mac_addr - return the hash value of a MAC address
+ *	@addr: the 48-bit Ethernet MAC address
+ *
+ *	Hashes a MAC address according to the hash function used by HW inexact
+ *	(hash) address matching.
+ */
+static int hash_mac_addr(const u8 *addr)
+{
+	u32 a = ((u32)addr[0] << 16) | ((u32)addr[1] << 8) | addr[2];
+	u32 b = ((u32)addr[3] << 16) | ((u32)addr[4] << 8) | addr[5];
+	a ^= b;
+	a ^= (a >> 12);
+	a ^= (a >> 6);
+	return a & 0x3f;
+}
+
+/**
+ *	t4_config_rss_range - configure a portion of the RSS mapping table
+ *	@adapter: the adapter
+ *	@mbox: mbox to use for the FW command
+ *	@viid: virtual interface whose RSS subtable is to be written
+ *	@start: start entry in the table to write
+ *	@n: how many table entries to write
+ *	@rspq: values for the response queue lookup table
+ *	@nrspq: number of values in @rspq
+ *
+ *	Programs the selected part of the VI's RSS mapping table with the
+ *	provided values.  If @nrspq < @n the supplied values are used repeatedly
+ *	until the full table range is populated.
+ *
+ *	The caller must ensure the values in @rspq are in the range allowed for
+ *	@viid.
+ */
+int t4_config_rss_range(struct adapter *adapter, int mbox, unsigned int viid,
+			int start, int n, const u16 *rspq, unsigned int nrspq)
+{
+	int ret;
+	const u16 *rsp = rspq;
+	const u16 *rsp_end = rspq + nrspq;
+	struct fw_rss_ind_tbl_cmd cmd;
+
+	memset(&cmd, 0, sizeof(cmd));
+	cmd.op_to_viid = htonl(FW_CMD_OP(FW_RSS_IND_TBL_CMD) |
+			       FW_CMD_REQUEST | FW_CMD_WRITE |
+			       FW_RSS_IND_TBL_CMD_VIID(viid));
+	cmd.retval_len16 = htonl(FW_LEN16(cmd));
+
+	/* each fw_rss_ind_tbl_cmd takes up to 32 entries */
+	while (n > 0) {
+		int nq = min(n, 32);
+		__be32 *qp = &cmd.iq0_to_iq2;
+
+		cmd.niqid = htons(nq);
+		cmd.startidx = htons(start);
+
+		start += nq;
+		n -= nq;
+
+		while (nq > 0) {
+			unsigned int v;
+
+			v = FW_RSS_IND_TBL_CMD_IQ0(*rsp);
+			if (++rsp >= rsp_end)
+				rsp = rspq;
+			v |= FW_RSS_IND_TBL_CMD_IQ1(*rsp);
+			if (++rsp >= rsp_end)
+				rsp = rspq;
+			v |= FW_RSS_IND_TBL_CMD_IQ2(*rsp);
+			if (++rsp >= rsp_end)
+				rsp = rspq;
+
+			*qp++ = htonl(v);
+			nq -= 3;
+		}
+
+		ret = t4_wr_mbox(adapter, mbox, &cmd, sizeof(cmd), NULL);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+/**
+ *	t4_config_glbl_rss - configure the global RSS mode
+ *	@adapter: the adapter
+ *	@mbox: mbox to use for the FW command
+ *	@mode: global RSS mode
+ *	@flags: mode-specific flags
+ *
+ *	Sets the global RSS mode.
+ */
+int t4_config_glbl_rss(struct adapter *adapter, int mbox, unsigned int mode,
+		       unsigned int flags)
+{
+	struct fw_rss_glb_config_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_write = htonl(FW_CMD_OP(FW_RSS_GLB_CONFIG_CMD) |
+			      FW_CMD_REQUEST | FW_CMD_WRITE);
+	c.retval_len16 = htonl(FW_LEN16(c));
+	if (mode == FW_RSS_GLB_CONFIG_CMD_MODE_MANUAL) {
+		c.u.manual.mode_pkd = htonl(FW_RSS_GLB_CONFIG_CMD_MODE(mode));
+	} else if (mode == FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL) {
+		c.u.basicvirtual.mode_pkd =
+			htonl(FW_RSS_GLB_CONFIG_CMD_MODE(mode));
+		c.u.basicvirtual.synmapen_to_hashtoeplitz = htonl(flags);
+	} else
+		return -EINVAL;
+	return t4_wr_mbox(adapter, mbox, &c, sizeof(c), NULL);
+}
+
+/* Read an RSS table row */
+static int rd_rss_row(struct adapter *adap, int row, u32 *val)
+{
+	t4_write_reg(adap, TP_RSS_LKP_TABLE, 0xfff00000 | row);
+	return t4_wait_op_done_val(adap, TP_RSS_LKP_TABLE, LKPTBLROWVLD, 1,
+				   5, 0, val);
+}
+
+/**
+ *	t4_read_rss - read the contents of the RSS mapping table
+ *	@adapter: the adapter
+ *	@map: holds the contents of the RSS mapping table
+ *
+ *	Reads the contents of the RSS hash->queue mapping table.
+ */
+int t4_read_rss(struct adapter *adapter, u16 *map)
+{
+	u32 val;
+	int i, ret;
+
+	for (i = 0; i < RSS_NENTRIES / 2; ++i) {
+		ret = rd_rss_row(adapter, i, &val);
+		if (ret)
+			return ret;
+		*map++ = LKPTBLQUEUE0_GET(val);
+		*map++ = LKPTBLQUEUE1_GET(val);
+	}
+	return 0;
+}
+
+/**
+ *	t4_tp_get_tcp_stats - read TP's TCP MIB counters
+ *	@adap: the adapter
+ *	@v4: holds the TCP/IP counter values
+ *	@v6: holds the TCP/IPv6 counter values
+ *
+ *	Returns the values of TP's TCP/IP and TCP/IPv6 MIB counters.
+ *	Either @v4 or @v6 may be %NULL to skip the corresponding stats.
+ */
+void t4_tp_get_tcp_stats(struct adapter *adap, struct tp_tcp_stats *v4,
+			 struct tp_tcp_stats *v6)
+{
+	u32 val[TP_MIB_TCP_RXT_SEG_LO - TP_MIB_TCP_OUT_RST + 1];
+
+#define STAT_IDX(x) ((TP_MIB_TCP_##x) - TP_MIB_TCP_OUT_RST)
+#define STAT(x)     val[STAT_IDX(x)]
+#define STAT64(x)   (((u64)STAT(x##_HI) << 32) | STAT(x##_LO))
+
+	if (v4) {
+		t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, val,
+				 ARRAY_SIZE(val), TP_MIB_TCP_OUT_RST);
+		v4->tcpOutRsts = STAT(OUT_RST);
+		v4->tcpInSegs  = STAT64(IN_SEG);
+		v4->tcpOutSegs = STAT64(OUT_SEG);
+		v4->tcpRetransSegs = STAT64(RXT_SEG);
+	}
+	if (v6) {
+		t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, val,
+				 ARRAY_SIZE(val), TP_MIB_TCP_V6OUT_RST);
+		v6->tcpOutRsts = STAT(OUT_RST);
+		v6->tcpInSegs  = STAT64(IN_SEG);
+		v6->tcpOutSegs = STAT64(OUT_SEG);
+		v6->tcpRetransSegs = STAT64(RXT_SEG);
+	}
+#undef STAT64
+#undef STAT
+#undef STAT_IDX
+}
+
+/**
+ *	t4_tp_get_err_stats - read TP's error MIB counters
+ *	@adap: the adapter
+ *	@st: holds the counter values
+ *
+ *	Returns the values of TP's error counters.
+ */
+void t4_tp_get_err_stats(struct adapter *adap, struct tp_err_stats *st)
+{
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, st->macInErrs,
+			 12, TP_MIB_MAC_IN_ERR_0);
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, st->tnlCongDrops,
+			 8, TP_MIB_TNL_CNG_DROP_0);
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, st->tnlTxDrops,
+			 4, TP_MIB_TNL_DROP_0);
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, st->ofldVlanDrops,
+			 4, TP_MIB_OFD_VLN_DROP_0);
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, st->tcp6InErrs,
+			 4, TP_MIB_TCP_V6IN_ERR_0);
+	t4_read_indirect(adap, TP_MIB_INDEX, TP_MIB_DATA, &st->ofldNoNeigh,
+			 2, TP_MIB_OFD_ARP_DROP);
+}
+
+/**
+ *	t4_read_mtu_tbl - returns the values in the HW path MTU table
+ *	@adap: the adapter
+ *	@mtus: where to store the MTU values
+ *	@mtu_log: where to store the MTU base-2 log (may be %NULL)
+ *
+ *	Reads the HW path MTU table.
+ */
+void t4_read_mtu_tbl(struct adapter *adap, u16 *mtus, u8 *mtu_log)
+{
+	u32 v;
+	int i;
+
+	for (i = 0; i < NMTUS; ++i) {
+		t4_write_reg(adap, TP_MTU_TABLE,
+			     MTUINDEX(0xff) | MTUVALUE(i));
+		v = t4_read_reg(adap, TP_MTU_TABLE);
+		mtus[i] = MTUVALUE_GET(v);
+		if (mtu_log)
+			mtu_log[i] = MTUWIDTH_GET(v);
+	}
+}
+
+/**
+ *	init_cong_ctrl - initialize congestion control parameters
+ *	@a: the alpha values for congestion control
+ *	@b: the beta values for congestion control
+ *
+ *	Initialize the congestion control parameters.
+ */
+static void __devinit init_cong_ctrl(unsigned short *a, unsigned short *b)
+{
+	a[0] = a[1] = a[2] = a[3] = a[4] = a[5] = a[6] = a[7] = a[8] = 1;
+	a[9] = 2;
+	a[10] = 3;
+	a[11] = 4;
+	a[12] = 5;
+	a[13] = 6;
+	a[14] = 7;
+	a[15] = 8;
+	a[16] = 9;
+	a[17] = 10;
+	a[18] = 14;
+	a[19] = 17;
+	a[20] = 21;
+	a[21] = 25;
+	a[22] = 30;
+	a[23] = 35;
+	a[24] = 45;
+	a[25] = 60;
+	a[26] = 80;
+	a[27] = 100;
+	a[28] = 200;
+	a[29] = 300;
+	a[30] = 400;
+	a[31] = 500;
+
+	b[0] = b[1] = b[2] = b[3] = b[4] = b[5] = b[6] = b[7] = b[8] = 0;
+	b[9] = b[10] = 1;
+	b[11] = b[12] = 2;
+	b[13] = b[14] = b[15] = b[16] = 3;
+	b[17] = b[18] = b[19] = b[20] = b[21] = 4;
+	b[22] = b[23] = b[24] = b[25] = b[26] = b[27] = 5;
+	b[28] = b[29] = 6;
+	b[30] = b[31] = 7;
+}
+
+/* The minimum additive increment value for the congestion control table */
+#define CC_MIN_INCR 2U
+
+/**
+ *	t4_load_mtus - write the MTU and congestion control HW tables
+ *	@adap: the adapter
+ *	@mtus: the values for the MTU table
+ *	@alpha: the values for the congestion control alpha parameter
+ *	@beta: the values for the congestion control beta parameter
+ *
+ *	Write the HW MTU table with the supplied MTUs and the high-speed
+ *	congestion control table with the supplied alpha, beta, and MTUs.
+ *	We write the two tables together because the additive increments
+ *	depend on the MTUs.
+ */
+void t4_load_mtus(struct adapter *adap, const unsigned short *mtus,
+		  const unsigned short *alpha, const unsigned short *beta)
+{
+	static const unsigned int avg_pkts[NCCTRL_WIN] = {
+		2, 6, 10, 14, 20, 28, 40, 56, 80, 112, 160, 224, 320, 448, 640,
+		896, 1281, 1792, 2560, 3584, 5120, 7168, 10240, 14336, 20480,
+		28672, 40960, 57344, 81920, 114688, 163840, 229376
+	};
+
+	unsigned int i, w;
+
+	for (i = 0; i < NMTUS; ++i) {
+		unsigned int mtu = mtus[i];
+		unsigned int log2 = fls(mtu);
+
+		if (!(mtu & ((1 << log2) >> 2)))     /* round */
+			log2--;
+		t4_write_reg(adap, TP_MTU_TABLE, MTUINDEX(i) |
+			     MTUWIDTH(log2) | MTUVALUE(mtu));
+
+		for (w = 0; w < NCCTRL_WIN; ++w) {
+			unsigned int inc;
+
+			inc = max(((mtu - 40) * alpha[w]) / avg_pkts[w],
+				  CC_MIN_INCR);
+
+			t4_write_reg(adap, TP_CCTRL_TABLE, (i << 21) |
+				     (w << 16) | (beta[w] << 13) | inc);
+		}
+	}
+}
+
+/**
+ *	t4_set_trace_filter - configure one of the tracing filters
+ *	@adap: the adapter
+ *	@tp: the desired trace filter parameters
+ *	@idx: which filter to configure
+ *	@enable: whether to enable or disable the filter
+ *
+ *	Configures one of the tracing filters available in HW.  If @enable is
+ *	%0 @tp is not examined and may be %NULL.
+ */
+int t4_set_trace_filter(struct adapter *adap, const struct trace_params *tp,
+			int idx, int enable)
+{
+	int i, ofst = idx * 4;
+	u32 data_reg, mask_reg, cfg;
+	u32 multitrc = TRCMULTIFILTER;
+
+	if (!enable) {
+		t4_write_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + ofst, 0);
+		goto out;
+	}
+
+	if (tp->port > 11 || tp->invert > 1 || tp->skip_len > 0x1f ||
+	    tp->skip_ofst > 0x1f || tp->min_len > 0x1ff ||
+	    tp->snap_len > 9600 || (idx && tp->snap_len > 256))
+		return -EINVAL;
+
+	if (tp->snap_len > 256) {            /* must be tracer 0 */
+		if ((t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + 4) |
+		     t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + 8) |
+		     t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + 12)) & TFEN)
+			return -EINVAL;  /* other tracers are enabled */
+		multitrc = 0;
+	} else if (idx) {
+		i = t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_B);
+		if (TFCAPTUREMAX_GET(i) > 256 &&
+		    (t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A) & TFEN))
+			return -EINVAL;
+	}
+
+	/* stop the tracer we'll be changing */
+	t4_write_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + ofst, 0);
+
+	/* disable tracing globally if running in the wrong single/multi mode */
+	cfg = t4_read_reg(adap, MPS_TRC_CFG);
+	if ((cfg & TRCEN) && multitrc != (cfg & TRCMULTIFILTER)) {
+		t4_write_reg(adap, MPS_TRC_CFG, cfg ^ TRCEN);
+		t4_read_reg(adap, MPS_TRC_CFG);                  /* flush */
+		msleep(1);
+		if (!(t4_read_reg(adap, MPS_TRC_CFG) & TRCFIFOEMPTY))
+			return -ETIMEDOUT;
+	}
+	/*
+	 * At this point either the tracing is enabled and in the right mode or
+	 * disabled.
+	 */
+
+	idx *= (MPS_TRC_FILTER1_MATCH - MPS_TRC_FILTER0_MATCH);
+	data_reg = MPS_TRC_FILTER0_MATCH + idx;
+	mask_reg = MPS_TRC_FILTER0_DONT_CARE + idx;
+
+	for (i = 0; i < TRACE_LEN / 4; i++, data_reg += 4, mask_reg += 4) {
+		t4_write_reg(adap, data_reg, tp->data[i]);
+		t4_write_reg(adap, mask_reg, ~tp->mask[i]);
+	}
+	t4_write_reg(adap, MPS_TRC_FILTER_MATCH_CTL_B + ofst,
+		     TFCAPTUREMAX(tp->snap_len) |
+		     TFMINPKTSIZE(tp->min_len));
+	t4_write_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + ofst,
+		     TFOFFSET(tp->skip_ofst) | TFLENGTH(tp->skip_len) |
+		     TFPORT(tp->port) | TFEN |
+		     (tp->invert ? TFINVERTMATCH : 0));
+
+	cfg &= ~TRCMULTIFILTER;
+	t4_write_reg(adap, MPS_TRC_CFG, cfg | TRCEN | multitrc);
+out:	t4_read_reg(adap, MPS_TRC_CFG);  /* flush */
+	return 0;
+}
+
+/**
+ *	t4_get_trace_filter - query one of the tracing filters
+ *	@adap: the adapter
+ *	@tp: the current trace filter parameters
+ *	@idx: which trace filter to query
+ *	@enabled: non-zero if the filter is enabled
+ *
+ *	Returns the current settings of one of the HW tracing filters.
+ */
+void t4_get_trace_filter(struct adapter *adap, struct trace_params *tp, int idx,
+			 int *enabled)
+{
+	u32 ctla, ctlb;
+	int i, ofst = idx * 4;
+	u32 data_reg, mask_reg;
+
+	ctla = t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_A + ofst);
+	ctlb = t4_read_reg(adap, MPS_TRC_FILTER_MATCH_CTL_B + ofst);
+
+	*enabled = !!(ctla & TFEN);
+	tp->snap_len = TFCAPTUREMAX_GET(ctlb);
+	tp->min_len = TFMINPKTSIZE_GET(ctlb);
+	tp->skip_ofst = TFOFFSET_GET(ctla);
+	tp->skip_len = TFLENGTH_GET(ctla);
+	tp->invert = !!(ctla & TFINVERTMATCH);
+	tp->port = TFPORT_GET(ctla);
+
+	ofst = (MPS_TRC_FILTER1_MATCH - MPS_TRC_FILTER0_MATCH) * idx;
+	data_reg = MPS_TRC_FILTER0_MATCH + ofst;
+	mask_reg = MPS_TRC_FILTER0_DONT_CARE + ofst;
+
+	for (i = 0; i < TRACE_LEN / 4; i++, data_reg += 4, mask_reg += 4) {
+		tp->mask[i] = ~t4_read_reg(adap, mask_reg);
+		tp->data[i] = t4_read_reg(adap, data_reg) & tp->mask[i];
+	}
+}
+
+/**
+ *	get_mps_bg_map - return the buffer groups associated with a port
+ *	@adap: the adapter
+ *	@idx: the port index
+ *
+ *	Returns a bitmap indicating which MPS buffer groups are associated
+ *	with the given port.  Bit i is set if buffer group i is used by the
+ *	port.
+ */
+static unsigned int get_mps_bg_map(struct adapter *adap, int idx)
+{
+	u32 n = NUMPORTS_GET(t4_read_reg(adap, MPS_CMN_CTL));
+
+	if (n == 0)
+		return idx == 0 ? 0xf : 0;
+	if (n == 1)
+		return idx < 2 ? (3 << (2 * idx)) : 0;
+	return 1 << idx;
+}
+
+/**
+ *	t4_get_port_stats - collect port statistics
+ *	@adap: the adapter
+ *	@idx: the port index
+ *	@p: the stats structure to fill
+ *
+ *	Collect statistics related to the given port from HW.
+ */
+void t4_get_port_stats(struct adapter *adap, int idx, struct port_stats *p)
+{
+	u32 bgmap = get_mps_bg_map(adap, idx);
+
+#define GET_STAT(name) \
+	t4_read_reg64(adap, PORT_REG(idx, MPS_PORT_STAT_##name##_L))
+#define GET_STAT_COM(name) t4_read_reg64(adap, MPS_STAT_##name##_L)
+
+	p->tx_octets           = GET_STAT(TX_PORT_BYTES);
+	p->tx_frames           = GET_STAT(TX_PORT_FRAMES);
+	p->tx_bcast_frames     = GET_STAT(TX_PORT_BCAST);
+	p->tx_mcast_frames     = GET_STAT(TX_PORT_MCAST);
+	p->tx_ucast_frames     = GET_STAT(TX_PORT_UCAST);
+	p->tx_error_frames     = GET_STAT(TX_PORT_ERROR);
+	p->tx_frames_64        = GET_STAT(TX_PORT_64B);
+	p->tx_frames_65_127    = GET_STAT(TX_PORT_65B_127B);
+	p->tx_frames_128_255   = GET_STAT(TX_PORT_128B_255B);
+	p->tx_frames_256_511   = GET_STAT(TX_PORT_256B_511B);
+	p->tx_frames_512_1023  = GET_STAT(TX_PORT_512B_1023B);
+	p->tx_frames_1024_1518 = GET_STAT(TX_PORT_1024B_1518B);
+	p->tx_frames_1519_max  = GET_STAT(TX_PORT_1519B_MAX);
+	p->tx_drop             = GET_STAT(TX_PORT_DROP);
+	p->tx_pause            = GET_STAT(TX_PORT_PAUSE);
+	p->tx_ppp0             = GET_STAT(TX_PORT_PPP0);
+	p->tx_ppp1             = GET_STAT(TX_PORT_PPP1);
+	p->tx_ppp2             = GET_STAT(TX_PORT_PPP2);
+	p->tx_ppp3             = GET_STAT(TX_PORT_PPP3);
+	p->tx_ppp4             = GET_STAT(TX_PORT_PPP4);
+	p->tx_ppp5             = GET_STAT(TX_PORT_PPP5);
+	p->tx_ppp6             = GET_STAT(TX_PORT_PPP6);
+	p->tx_ppp7             = GET_STAT(TX_PORT_PPP7);
+
+	p->rx_octets           = GET_STAT(RX_PORT_BYTES);
+	p->rx_frames           = GET_STAT(RX_PORT_FRAMES);
+	p->rx_bcast_frames     = GET_STAT(RX_PORT_BCAST);
+	p->rx_mcast_frames     = GET_STAT(RX_PORT_MCAST);
+	p->rx_ucast_frames     = GET_STAT(RX_PORT_UCAST);
+	p->rx_too_long         = GET_STAT(RX_PORT_MTU_ERROR);
+	p->rx_jabber           = GET_STAT(RX_PORT_MTU_CRC_ERROR);
+	p->rx_fcs_err          = GET_STAT(RX_PORT_CRC_ERROR);
+	p->rx_len_err          = GET_STAT(RX_PORT_LEN_ERROR);
+	p->rx_symbol_err       = GET_STAT(RX_PORT_SYM_ERROR);
+	p->rx_runt             = GET_STAT(RX_PORT_LESS_64B);
+	p->rx_frames_64        = GET_STAT(RX_PORT_64B);
+	p->rx_frames_65_127    = GET_STAT(RX_PORT_65B_127B);
+	p->rx_frames_128_255   = GET_STAT(RX_PORT_128B_255B);
+	p->rx_frames_256_511   = GET_STAT(RX_PORT_256B_511B);
+	p->rx_frames_512_1023  = GET_STAT(RX_PORT_512B_1023B);
+	p->rx_frames_1024_1518 = GET_STAT(RX_PORT_1024B_1518B);
+	p->rx_frames_1519_max  = GET_STAT(RX_PORT_1519B_MAX);
+	p->rx_pause            = GET_STAT(RX_PORT_PAUSE);
+	p->rx_ppp0             = GET_STAT(RX_PORT_PPP0);
+	p->rx_ppp1             = GET_STAT(RX_PORT_PPP1);
+	p->rx_ppp2             = GET_STAT(RX_PORT_PPP2);
+	p->rx_ppp3             = GET_STAT(RX_PORT_PPP3);
+	p->rx_ppp4             = GET_STAT(RX_PORT_PPP4);
+	p->rx_ppp5             = GET_STAT(RX_PORT_PPP5);
+	p->rx_ppp6             = GET_STAT(RX_PORT_PPP6);
+	p->rx_ppp7             = GET_STAT(RX_PORT_PPP7);
+
+	p->rx_ovflow0 = (bgmap & 1) ? GET_STAT_COM(RX_BG_0_MAC_DROP_FRAME) : 0;
+	p->rx_ovflow1 = (bgmap & 2) ? GET_STAT_COM(RX_BG_1_MAC_DROP_FRAME) : 0;
+	p->rx_ovflow2 = (bgmap & 4) ? GET_STAT_COM(RX_BG_2_MAC_DROP_FRAME) : 0;
+	p->rx_ovflow3 = (bgmap & 8) ? GET_STAT_COM(RX_BG_3_MAC_DROP_FRAME) : 0;
+	p->rx_trunc0 = (bgmap & 1) ? GET_STAT_COM(RX_BG_0_MAC_TRUNC_FRAME) : 0;
+	p->rx_trunc1 = (bgmap & 2) ? GET_STAT_COM(RX_BG_1_MAC_TRUNC_FRAME) : 0;
+	p->rx_trunc2 = (bgmap & 4) ? GET_STAT_COM(RX_BG_2_MAC_TRUNC_FRAME) : 0;
+	p->rx_trunc3 = (bgmap & 8) ? GET_STAT_COM(RX_BG_3_MAC_TRUNC_FRAME) : 0;
+
+#undef GET_STAT
+#undef GET_STAT_COM
+}
+
+/**
+ *	t4_get_lb_stats - collect loopback port statistics
+ *	@adap: the adapter
+ *	@idx: the loopback port index
+ *	@p: the stats structure to fill
+ *
+ *	Return HW statistics for the given loopback port.
+ */
+void t4_get_lb_stats(struct adapter *adap, int idx, struct lb_port_stats *p)
+{
+	u32 bgmap = get_mps_bg_map(adap, idx);
+
+#define GET_STAT(name) \
+	t4_read_reg64(adap, PORT_REG(idx, MPS_PORT_STAT_LB_PORT_##name##_L))
+#define GET_STAT_COM(name) t4_read_reg64(adap, MPS_STAT_##name##_L)
+
+	p->octets           = GET_STAT(BYTES);
+	p->frames           = GET_STAT(FRAMES);
+	p->bcast_frames     = GET_STAT(BCAST);
+	p->mcast_frames     = GET_STAT(MCAST);
+	p->ucast_frames     = GET_STAT(UCAST);
+	p->error_frames     = GET_STAT(ERROR);
+
+	p->frames_64        = GET_STAT(64B);
+	p->frames_65_127    = GET_STAT(65B_127B);
+	p->frames_128_255   = GET_STAT(128B_255B);
+	p->frames_256_511   = GET_STAT(256B_511B);
+	p->frames_512_1023  = GET_STAT(512B_1023B);
+	p->frames_1024_1518 = GET_STAT(1024B_1518B);
+	p->frames_1519_max  = GET_STAT(1519B_MAX);
+	p->drop             = t4_read_reg(adap, PORT_REG(idx,
+					  MPS_PORT_STAT_LB_PORT_DROP_FRAMES));
+
+	p->ovflow0 = (bgmap & 1) ? GET_STAT_COM(RX_BG_0_LB_DROP_FRAME) : 0;
+	p->ovflow1 = (bgmap & 2) ? GET_STAT_COM(RX_BG_1_LB_DROP_FRAME) : 0;
+	p->ovflow2 = (bgmap & 4) ? GET_STAT_COM(RX_BG_2_LB_DROP_FRAME) : 0;
+	p->ovflow3 = (bgmap & 8) ? GET_STAT_COM(RX_BG_3_LB_DROP_FRAME) : 0;
+	p->trunc0 = (bgmap & 1) ? GET_STAT_COM(RX_BG_0_LB_TRUNC_FRAME) : 0;
+	p->trunc1 = (bgmap & 2) ? GET_STAT_COM(RX_BG_1_LB_TRUNC_FRAME) : 0;
+	p->trunc2 = (bgmap & 4) ? GET_STAT_COM(RX_BG_2_LB_TRUNC_FRAME) : 0;
+	p->trunc3 = (bgmap & 8) ? GET_STAT_COM(RX_BG_3_LB_TRUNC_FRAME) : 0;
+
+#undef GET_STAT
+#undef GET_STAT_COM
+}
+
+/**
+ *	t4_wol_magic_enable - enable/disable magic packet WoL
+ *	@adap: the adapter
+ *	@port: the physical port index
+ *	@addr: MAC address expected in magic packets, %NULL to disable
+ *
+ *	Enables/disables magic packet wake-on-LAN for the selected port.
+ */
+void t4_wol_magic_enable(struct adapter *adap, unsigned int port,
+			 const u8 *addr)
+{
+	if (addr) {
+		t4_write_reg(adap, PORT_REG(port, XGMAC_PORT_MAGIC_MACID_LO),
+			     (addr[2] << 24) | (addr[3] << 16) |
+			     (addr[4] << 8) | addr[5]);
+		t4_write_reg(adap, PORT_REG(port, XGMAC_PORT_MAGIC_MACID_HI),
+			     (addr[0] << 8) | addr[1]);
+	}
+	t4_set_reg_field(adap, PORT_REG(port, XGMAC_PORT_CFG2), MAGICEN,
+			 addr ? MAGICEN : 0);
+}
+
+/**
+ *	t4_wol_pat_enable - enable/disable pattern-based WoL
+ *	@adap: the adapter
+ *	@port: the physical port index
+ *	@map: bitmap of which HW pattern filters to set
+ *	@mask0: byte mask for bytes 0-63 of a packet
+ *	@mask1: byte mask for bytes 64-127 of a packet
+ *	@crc: Ethernet CRC for selected bytes
+ *	@enable: enable/disable switch
+ *
+ *	Sets the pattern filters indicated in @map to mask out the bytes
+ *	specified in @mask0/@mask1 in received packets and compare the CRC of
+ *	the resulting packet against @crc.  If @enable is %true pattern-based
+ *	WoL is enabled, otherwise disabled.
+ */
+int t4_wol_pat_enable(struct adapter *adap, unsigned int port, unsigned int map,
+		      u64 mask0, u64 mask1, unsigned int crc, bool enable)
+{
+	int i;
+
+	if (!enable) {
+		t4_set_reg_field(adap, PORT_REG(port, XGMAC_PORT_CFG2),
+				 PATEN, 0);
+		return 0;
+	}
+	if (map > 0xff)
+		return -EINVAL;
+
+#define EPIO_REG(name) PORT_REG(port, XGMAC_PORT_EPIO_##name)
+
+	t4_write_reg(adap, EPIO_REG(DATA1), mask0 >> 32);
+	t4_write_reg(adap, EPIO_REG(DATA2), mask1);
+	t4_write_reg(adap, EPIO_REG(DATA3), mask1 >> 32);
+
+	for (i = 0; i < NWOL_PAT; i++, map >>= 1) {
+		if (!(map & 1))
+			continue;
+
+		/* write byte masks */
+		t4_write_reg(adap, EPIO_REG(DATA0), mask0);
+		t4_write_reg(adap, EPIO_REG(OP), ADDRESS(i) | EPIOWR);
+		t4_read_reg(adap, EPIO_REG(OP));                /* flush */
+		if (t4_read_reg(adap, EPIO_REG(OP)) & BUSY)
+			return -ETIMEDOUT;
+
+		/* write CRC */
+		t4_write_reg(adap, EPIO_REG(DATA0), crc);
+		t4_write_reg(adap, EPIO_REG(OP), ADDRESS(i + 32) | EPIOWR);
+		t4_read_reg(adap, EPIO_REG(OP));                /* flush */
+		if (t4_read_reg(adap, EPIO_REG(OP)) & BUSY)
+			return -ETIMEDOUT;
+	}
+#undef EPIO_REG
+
+	t4_set_reg_field(adap, PORT_REG(port, XGMAC_PORT_CFG2), 0, PATEN);
+	return 0;
+}
+
+#define INIT_CMD(var, cmd, rd_wr) do { \
+	(var).op_to_write = htonl(FW_CMD_OP(FW_##cmd##_CMD) | \
+				  FW_CMD_REQUEST | FW_CMD_##rd_wr); \
+	(var).retval_len16 = htonl(FW_LEN16(var)); \
+} while (0)
+
+/**
+ *	t4_mdio_rd - read a PHY register through MDIO
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@phy_addr: the PHY address
+ *	@mmd: the PHY MMD to access (0 for clause 22 PHYs)
+ *	@reg: the register to read
+ *	@valp: where to store the value
+ *
+ *	Issues a FW command through the given mailbox to read a PHY register.
+ */
+int t4_mdio_rd(struct adapter *adap, unsigned int mbox, unsigned int phy_addr,
+	       unsigned int mmd, unsigned int reg, u16 *valp)
+{
+	int ret;
+	struct fw_ldst_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_addrspace = htonl(FW_CMD_OP(FW_LDST_CMD) | FW_CMD_REQUEST |
+		FW_CMD_READ | FW_LDST_CMD_ADDRSPACE(FW_LDST_ADDRSPC_MDIO));
+	c.cycles_to_len16 = htonl(FW_LEN16(c));
+	c.u.mdio.paddr_mmd = htons(FW_LDST_CMD_PADDR(phy_addr) |
+				   FW_LDST_CMD_MMD(mmd));
+	c.u.mdio.raddr = htons(reg);
+
+	ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+	if (ret == 0)
+		*valp = ntohs(c.u.mdio.rval);
+	return ret;
+}
+
+/**
+ *	t4_mdio_wr - write a PHY register through MDIO
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@phy_addr: the PHY address
+ *	@mmd: the PHY MMD to access (0 for clause 22 PHYs)
+ *	@reg: the register to write
+ *	@valp: value to write
+ *
+ *	Issues a FW command through the given mailbox to write a PHY register.
+ */
+int t4_mdio_wr(struct adapter *adap, unsigned int mbox, unsigned int phy_addr,
+	       unsigned int mmd, unsigned int reg, u16 val)
+{
+	struct fw_ldst_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_addrspace = htonl(FW_CMD_OP(FW_LDST_CMD) | FW_CMD_REQUEST |
+		FW_CMD_WRITE | FW_LDST_CMD_ADDRSPACE(FW_LDST_ADDRSPC_MDIO));
+	c.cycles_to_len16 = htonl(FW_LEN16(c));
+	c.u.mdio.paddr_mmd = htons(FW_LDST_CMD_PADDR(phy_addr) |
+				   FW_LDST_CMD_MMD(mmd));
+	c.u.mdio.raddr = htons(reg);
+	c.u.mdio.rval = htons(val);
+
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_fw_hello - establish communication with FW
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@evt_mbox: mailbox to receive async FW events
+ *	@master: specifies the caller's willingness to be the device master
+ *	@state: returns the current device state
+ *
+ *	Issues a command to establish communication with FW.
+ */
+int t4_fw_hello(struct adapter *adap, unsigned int mbox, unsigned int evt_mbox,
+		enum dev_master master, enum dev_state *state)
+{
+	int ret;
+	struct fw_hello_cmd c;
+
+	INIT_CMD(c, HELLO, WRITE);
+	c.err_to_mbasyncnot = htonl(
+		FW_HELLO_CMD_MASTERDIS(master == MASTER_CANT) |
+		FW_HELLO_CMD_MASTERFORCE(master == MASTER_MUST) |
+		FW_HELLO_CMD_MBMASTER(master == MASTER_MUST ? mbox : 0xff) |
+		FW_HELLO_CMD_MBASYNCNOT(evt_mbox));
+
+	ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+	if (ret == 0 && state) {
+		u32 v = ntohl(c.err_to_mbasyncnot);
+		if (v & FW_HELLO_CMD_INIT)
+			*state = DEV_STATE_INIT;
+		else if (v & FW_HELLO_CMD_ERR)
+			*state = DEV_STATE_ERR;
+		else
+			*state = DEV_STATE_UNINIT;
+	}
+	return ret;
+}
+
+/**
+ *	t4_fw_bye - end communication with FW
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *
+ *	Issues a command to terminate communication with FW.
+ */
+int t4_fw_bye(struct adapter *adap, unsigned int mbox)
+{
+	struct fw_bye_cmd c;
+
+	INIT_CMD(c, BYE, WRITE);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_init_cmd - ask FW to initialize the device
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *
+ *	Issues a command to FW to partially initialize the device.  This
+ *	performs initialization that generally doesn't depend on user input.
+ */
+int t4_early_init(struct adapter *adap, unsigned int mbox)
+{
+	struct fw_initialize_cmd c;
+
+	INIT_CMD(c, INITIALIZE, WRITE);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_fw_reset - issue a reset to FW
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@reset: specifies the type of reset to perform
+ *
+ *	Issues a reset command of the specified type to FW.
+ */
+int t4_fw_reset(struct adapter *adap, unsigned int mbox, int reset)
+{
+	struct fw_reset_cmd c;
+
+	INIT_CMD(c, RESET, WRITE);
+	c.val = htonl(reset);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_query_params - query FW or device parameters
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF
+ *	@vf: the VF
+ *	@nparams: the number of parameters
+ *	@params: the parameter names
+ *	@val: the parameter values
+ *
+ *	Reads the value of FW or device parameters.  Up to 7 parameters can be
+ *	queried at once.
+ */
+int t4_query_params(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int nparams, const u32 *params,
+		    u32 *val)
+{
+	int i, ret;
+	struct fw_params_cmd c;
+	__be32 *p = &c.param[0].mnem;
+
+	if (nparams > 7)
+		return -EINVAL;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_PARAMS_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_READ | FW_PARAMS_CMD_PFN(pf) |
+			    FW_PARAMS_CMD_VFN(vf));
+	c.retval_len16 = htonl(FW_LEN16(c));
+	for (i = 0; i < nparams; i++, p += 2)
+		*p = htonl(*params++);
+
+	ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+	if (ret == 0)
+		for (i = 0, p = &c.param[0].val; i < nparams; i++, p += 2)
+			*val++ = ntohl(*p);
+	return ret;
+}
+
+/**
+ *	t4_set_params - sets FW or device parameters
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF
+ *	@vf: the VF
+ *	@nparams: the number of parameters
+ *	@params: the parameter names
+ *	@val: the parameter values
+ *
+ *	Sets the value of FW or device parameters.  Up to 7 parameters can be
+ *	specified at once.
+ */
+int t4_set_params(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		  unsigned int vf, unsigned int nparams, const u32 *params,
+		  const u32 *val)
+{
+	struct fw_params_cmd c;
+	__be32 *p = &c.param[0].mnem;
+
+	if (nparams > 7)
+		return -EINVAL;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_PARAMS_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_PARAMS_CMD_PFN(pf) |
+			    FW_PARAMS_CMD_VFN(vf));
+	c.retval_len16 = htonl(FW_LEN16(c));
+	while (nparams--) {
+		*p++ = htonl(*params++);
+		*p++ = htonl(*val++);
+	}
+
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_cfg_pfvf - configure PF/VF resource limits
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF being configured
+ *	@vf: the VF being configured
+ *	@txq: the max number of egress queues
+ *	@txq_eth_ctrl: the max number of egress Ethernet or control queues
+ *	@rxqi: the max number of interrupt-capable ingress queues
+ *	@rxq: the max number of interruptless ingress queues
+ *	@tc: the PCI traffic class
+ *	@vi: the max number of virtual interfaces
+ *	@cmask: the channel access rights mask for the PF/VF
+ *	@pmask: the port access rights mask for the PF/VF
+ *	@nexact: the maximum number of exact MPS filters
+ *	@rcaps: read capabilities
+ *	@wxcaps: write/execute capabilities
+ *
+ *	Configures resource limits and capabilities for a physical or virtual
+ *	function.
+ */
+int t4_cfg_pfvf(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		unsigned int vf, unsigned int txq, unsigned int txq_eth_ctrl,
+		unsigned int rxqi, unsigned int rxq, unsigned int tc,
+		unsigned int vi, unsigned int cmask, unsigned int pmask,
+		unsigned int nexact, unsigned int rcaps, unsigned int wxcaps)
+{
+	struct fw_pfvf_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_PFVF_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_PFVF_CMD_PFN(pf) |
+			    FW_PFVF_CMD_VFN(vf));
+	c.retval_len16 = htonl(FW_LEN16(c));
+	c.niqflint_niq = htonl(FW_PFVF_CMD_NIQFLINT(rxqi) |
+			       FW_PFVF_CMD_NIQ(rxq));
+	c.cmask_to_neq = htonl(FW_PFVF_CMD_CMASK(cmask) |
+			       FW_PFVF_CMD_PMASK(pmask) |
+			       FW_PFVF_CMD_NEQ(txq));
+	c.tc_to_nexactf = htonl(FW_PFVF_CMD_TC(tc) | FW_PFVF_CMD_NVI(vi) |
+				FW_PFVF_CMD_NEXACTF(nexact));
+	c.r_caps_to_nethctrl = htonl(FW_PFVF_CMD_R_CAPS(rcaps) |
+				     FW_PFVF_CMD_WX_CAPS(wxcaps) |
+				     FW_PFVF_CMD_NETHCTRL(txq_eth_ctrl));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_alloc_vi - allocate a virtual interface
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@port: physical port associated with the VI
+ *	@pf: the PF owning the VI
+ *	@vf: the VF owning the VI
+ *	@nmac: number of MAC addresses needed (1 to 5)
+ *	@mac: the MAC addresses of the VI
+ *	@rss_size: size of RSS table slice associated with this VI
+ *
+ *	Allocates a virtual interface for the given physical port.  If @mac is
+ *	not %NULL it contains the MAC addresses of the VI as assigned by FW.
+ *	@mac should be large enough to hold @nmac Ethernet addresses, they are
+ *	stored consecutively so the space needed is @nmac * 6 bytes.
+ *	Returns a negative error number or the non-negative VI id.
+ */
+int t4_alloc_vi(struct adapter *adap, unsigned int mbox, unsigned int port,
+		unsigned int pf, unsigned int vf, unsigned int nmac, u8 *mac,
+		unsigned int *rss_size)
+{
+	int ret;
+	struct fw_vi_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_VI_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_WRITE | FW_CMD_EXEC |
+			    FW_VI_CMD_PFN(pf) | FW_VI_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_VI_CMD_ALLOC | FW_LEN16(c));
+	c.portid_pkd = FW_VI_CMD_PORTID(port);
+	c.nmac = nmac - 1;
+
+	ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+	if (ret)
+		return ret;
+
+	if (mac) {
+		memcpy(mac, c.mac, sizeof(c.mac));
+		switch (nmac) {
+		case 5:
+			memcpy(mac + 24, c.nmac3, sizeof(c.nmac3));
+		case 4:
+			memcpy(mac + 18, c.nmac2, sizeof(c.nmac2));
+		case 3:
+			memcpy(mac + 12, c.nmac1, sizeof(c.nmac1));
+		case 2:
+			memcpy(mac + 6,  c.nmac0, sizeof(c.nmac0));
+		}
+	}
+	if (rss_size)
+		*rss_size = FW_VI_CMD_RSSSIZE_GET(ntohs(c.rsssize_pkd));
+	return ntohs(c.viid_pkd);
+}
+
+/**
+ *	t4_free_vi - free a virtual interface
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF owning the VI
+ *	@vf: the VF owning the VI
+ *	@viid: virtual interface identifiler
+ *
+ *	Free a previously allocated virtual interface.
+ */
+int t4_free_vi(struct adapter *adap, unsigned int mbox, unsigned int pf,
+	       unsigned int vf, unsigned int viid)
+{
+	struct fw_vi_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_VI_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_VI_CMD_PFN(pf) |
+			    FW_VI_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_VI_CMD_FREE | FW_LEN16(c));
+	c.viid_pkd = htons(FW_VI_CMD_VIID(viid));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+}
+
+/**
+ *	t4_set_rxmode - set Rx properties of a virtual interface
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@mtu: the new MTU or -1
+ *	@promisc: 1 to enable promiscuous mode, 0 to disable it, -1 no change
+ *	@all_multi: 1 to enable all-multi mode, 0 to disable it, -1 no change
+ *	@bcast: 1 to enable broadcast Rx, 0 to disable it, -1 no change
+ *	@sleep_ok: if true we may sleep while awaiting command completion
+ *
+ *	Sets Rx properties of a virtual interface.
+ */
+int t4_set_rxmode(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		  int mtu, int promisc, int all_multi, int bcast, bool sleep_ok)
+{
+	struct fw_vi_rxmode_cmd c;
+
+	/* convert to FW values */
+	if (mtu < 0)
+		mtu = FW_RXMODE_MTU_NO_CHG;
+	if (promisc < 0)
+		promisc = FW_VI_RXMODE_CMD_PROMISCEN_MASK;
+	if (all_multi < 0)
+		all_multi = FW_VI_RXMODE_CMD_ALLMULTIEN_MASK;
+	if (bcast < 0)
+		bcast = FW_VI_RXMODE_CMD_BROADCASTEN_MASK;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_RXMODE_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_WRITE | FW_VI_RXMODE_CMD_VIID(viid));
+	c.retval_len16 = htonl(FW_LEN16(c));
+	c.mtu_to_broadcasten = htonl(FW_VI_RXMODE_CMD_MTU(mtu) |
+				     FW_VI_RXMODE_CMD_PROMISCEN(promisc) |
+				     FW_VI_RXMODE_CMD_ALLMULTIEN(all_multi) |
+				     FW_VI_RXMODE_CMD_BROADCASTEN(bcast));
+	return t4_wr_mbox_meat(adap, mbox, &c, sizeof(c), NULL, sleep_ok);
+}
+
+/**
+ *	t4_alloc_mac_filt - allocates exact-match filters for MAC addresses
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@free: if true any existing filters for this VI id are first removed
+ *	@naddr: the number of MAC addresses to allocate filters for (up to 7)
+ *	@addr: the MAC address(es)
+ *	@idx: where to store the index of each allocated filter
+ *	@hash: pointer to hash address filter bitmap
+ *	@sleep_ok: call is allowed to sleep
+ *
+ *	Allocates an exact-match filter for each of the supplied addresses and
+ *	sets it to the corresponding address.  If @idx is not %NULL it should
+ *	have at least @naddr entries, each of which will be set to the index of
+ *	the filter allocated for the corresponding MAC address.  If a filter
+ *	could not be allocated for an address its index is set to 0xffff.
+ *	If @hash is not %NULL addresses that fail to allocate an exact filter
+ *	are hashed and update the hash filter bitmap pointed at by @hash.
+ *
+ *	Returns a negative error number or the number of filters allocated.
+ */
+int t4_alloc_mac_filt(struct adapter *adap, unsigned int mbox,
+		      unsigned int viid, bool free, unsigned int naddr,
+		      const u8 **addr, u16 *idx, u64 *hash, bool sleep_ok)
+{
+	int i, ret;
+	struct fw_vi_mac_cmd c;
+	struct fw_vi_mac_exact *p;
+
+	if (naddr > 7)
+		return -EINVAL;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_MAC_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_WRITE | (free ? FW_CMD_EXEC : 0) |
+			     FW_VI_MAC_CMD_VIID(viid));
+	c.freemacs_to_len16 = htonl(FW_VI_MAC_CMD_FREEMACS(free) |
+				    FW_CMD_LEN16((naddr + 2) / 2));
+
+	for (i = 0, p = c.u.exact; i < naddr; i++, p++) {
+		p->valid_to_idx = htons(FW_VI_MAC_CMD_VALID |
+				      FW_VI_MAC_CMD_IDX(FW_VI_MAC_ADD_MAC));
+		memcpy(p->macaddr, addr[i], sizeof(p->macaddr));
+	}
+
+	ret = t4_wr_mbox_meat(adap, mbox, &c, sizeof(c), &c, sleep_ok);
+	if (ret)
+		return ret;
+
+	for (i = 0, p = c.u.exact; i < naddr; i++, p++) {
+		u16 index = FW_VI_MAC_CMD_IDX_GET(ntohs(p->valid_to_idx));
+
+		if (idx)
+			idx[i] = index >= NEXACT_MAC ? 0xffff : index;
+		if (index < NEXACT_MAC)
+			ret++;
+		else if (hash)
+			*hash |= (1 << hash_mac_addr(addr[i]));
+	}
+	return ret;
+}
+
+/**
+ *	t4_change_mac - modifies the exact-match filter for a MAC address
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@idx: index of existing filter for old value of MAC address, or -1
+ *	@addr: the new MAC address value
+ *	@persist: whether a new MAC allocation should be persistent
+ *	@add_smt: if true also add the address to the HW SMT
+ *
+ *	Modifies an exact-match filter and sets it to the new MAC address.
+ *	Note that in general it is not possible to modify the value of a given
+ *	filter so the generic way to modify an address filter is to free the one
+ *	being used by the old address value and allocate a new filter for the
+ *	new address value.  @idx can be -1 if the address is a new addition.
+ *
+ *	Returns a negative error number or the index of the filter with the new
+ *	MAC value.
+ */
+int t4_change_mac(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		  int idx, const u8 *addr, bool persist, bool add_smt)
+{
+	int ret, mode;
+	struct fw_vi_mac_cmd c;
+	struct fw_vi_mac_exact *p = c.u.exact;
+
+	if (idx < 0)                             /* new allocation */
+		idx = persist ? FW_VI_MAC_ADD_PERSIST_MAC : FW_VI_MAC_ADD_MAC;
+	mode = add_smt ? FW_VI_MAC_SMT_AND_MPSTCAM : FW_VI_MAC_MPS_TCAM_ENTRY;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_MAC_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_WRITE | FW_VI_MAC_CMD_VIID(viid));
+	c.freemacs_to_len16 = htonl(FW_CMD_LEN16(1));
+	p->valid_to_idx = htons(FW_VI_MAC_CMD_VALID |
+				FW_VI_MAC_CMD_SMAC_RESULT(mode) |
+				FW_VI_MAC_CMD_IDX(idx));
+	memcpy(p->macaddr, addr, sizeof(p->macaddr));
+
+	ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+	if (ret == 0) {
+		ret = FW_VI_MAC_CMD_IDX_GET(ntohs(p->valid_to_idx));
+		if (ret >= NEXACT_MAC)
+			ret = -ENOMEM;
+	}
+	return ret;
+}
+
+/**
+ *	t4_set_addr_hash - program the MAC inexact-match hash filter
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@ucast: whether the hash filter should also match unicast addresses
+ *	@vec: the value to be written to the hash filter
+ *	@sleep_ok: call is allowed to sleep
+ *
+ *	Sets the 64-bit inexact-match hash filter for a virtual interface.
+ */
+int t4_set_addr_hash(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		     bool ucast, u64 vec, bool sleep_ok)
+{
+	struct fw_vi_mac_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_MAC_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_WRITE | FW_VI_ENABLE_CMD_VIID(viid));
+	c.freemacs_to_len16 = htonl(FW_VI_MAC_CMD_HASHVECEN |
+				    FW_VI_MAC_CMD_HASHUNIEN(ucast) |
+				    FW_CMD_LEN16(1));
+	c.u.hash.hashvec = cpu_to_be64(vec);
+	return t4_wr_mbox_meat(adap, mbox, &c, sizeof(c), NULL, sleep_ok);
+}
+
+/**
+ *	t4_enable_vi - enable/disable a virtual interface
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@rx_en: 1=enable Rx, 0=disable Rx
+ *	@tx_en: 1=enable Tx, 0=disable Tx
+ *
+ *	Enables/disables a virtual interface.
+ */
+int t4_enable_vi(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		 bool rx_en, bool tx_en)
+{
+	struct fw_vi_enable_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_ENABLE_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_EXEC | FW_VI_ENABLE_CMD_VIID(viid));
+	c.ien_to_len16 = htonl(FW_VI_ENABLE_CMD_IEN(rx_en) |
+			       FW_VI_ENABLE_CMD_EEN(tx_en) | FW_LEN16(c));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_identify_port - identify a VI's port by blinking its LED
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@viid: the VI id
+ *	@nblinks: how many times to blink LED at 2.5 Hz
+ *
+ *	Identifies a VI's port by blinking its LED.
+ */
+int t4_identify_port(struct adapter *adap, unsigned int mbox, unsigned int viid,
+		     unsigned int nblinks)
+{
+	struct fw_vi_enable_cmd c;
+
+	c.op_to_viid = htonl(FW_CMD_OP(FW_VI_ENABLE_CMD) | FW_CMD_REQUEST |
+			     FW_CMD_EXEC | FW_VI_ENABLE_CMD_VIID(viid));
+	c.ien_to_len16 = htonl(FW_VI_ENABLE_CMD_LED | FW_LEN16(c));
+	c.blinkdur = htons(nblinks);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_iq_start_stop - enable/disable an ingress queue and its FLs
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@start: %true to enable the queues, %false to disable them
+ *	@pf: the PF owning the queues
+ *	@vf: the VF owning the queues
+ *	@iqid: ingress queue id
+ *	@fl0id: FL0 queue id or 0xffff if no attached FL0
+ *	@fl1id: FL1 queue id or 0xffff if no attached FL1
+ *
+ *	Starts or stops an ingress queue and its associated FLs, if any.
+ */
+int t4_iq_start_stop(struct adapter *adap, unsigned int mbox, bool start,
+		     unsigned int pf, unsigned int vf, unsigned int iqid,
+		     unsigned int fl0id, unsigned int fl1id)
+{
+	struct fw_iq_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_IQ_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_IQ_CMD_PFN(pf) |
+			    FW_IQ_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_IQ_CMD_IQSTART(start) |
+				 FW_IQ_CMD_IQSTOP(!start) | FW_LEN16(c));
+	c.iqid = htons(iqid);
+	c.fl0id = htons(fl0id);
+	c.fl1id = htons(fl1id);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_iq_free - free an ingress queue and its FLs
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF owning the queues
+ *	@vf: the VF owning the queues
+ *	@iqtype: the ingress queue type
+ *	@iqid: ingress queue id
+ *	@fl0id: FL0 queue id or 0xffff if no attached FL0
+ *	@fl1id: FL1 queue id or 0xffff if no attached FL1
+ *
+ *	Frees an ingress queue and its associated FLs, if any.
+ */
+int t4_iq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+	       unsigned int vf, unsigned int iqtype, unsigned int iqid,
+	       unsigned int fl0id, unsigned int fl1id)
+{
+	struct fw_iq_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_IQ_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_IQ_CMD_PFN(pf) |
+			    FW_IQ_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_IQ_CMD_FREE | FW_LEN16(c));
+	c.type_to_iqandstindex = htonl(FW_IQ_CMD_TYPE(iqtype));
+	c.iqid = htons(iqid);
+	c.fl0id = htons(fl0id);
+	c.fl1id = htons(fl1id);
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_eth_eq_free - free an Ethernet egress queue
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF owning the queue
+ *	@vf: the VF owning the queue
+ *	@eqid: egress queue id
+ *
+ *	Frees an Ethernet egress queue.
+ */
+int t4_eth_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		   unsigned int vf, unsigned int eqid)
+{
+	struct fw_eq_eth_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_ETH_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_EQ_ETH_CMD_PFN(pf) |
+			    FW_EQ_ETH_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_EQ_ETH_CMD_FREE | FW_LEN16(c));
+	c.eqid_pkd = htonl(FW_EQ_ETH_CMD_EQID(eqid));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_ctrl_eq_free - free a control egress queue
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF owning the queue
+ *	@vf: the VF owning the queue
+ *	@eqid: egress queue id
+ *
+ *	Frees a control egress queue.
+ */
+int t4_ctrl_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int eqid)
+{
+	struct fw_eq_ctrl_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_CTRL_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_EQ_CTRL_CMD_PFN(pf) |
+			    FW_EQ_CTRL_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_EQ_CTRL_CMD_FREE | FW_LEN16(c));
+	c.cmpliqid_eqid = htonl(FW_EQ_CTRL_CMD_EQID(eqid));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_ofld_eq_free - free an offload egress queue
+ *	@adap: the adapter
+ *	@mbox: mailbox to use for the FW command
+ *	@pf: the PF owning the queue
+ *	@vf: the VF owning the queue
+ *	@eqid: egress queue id
+ *
+ *	Frees a control egress queue.
+ */
+int t4_ofld_eq_free(struct adapter *adap, unsigned int mbox, unsigned int pf,
+		    unsigned int vf, unsigned int eqid)
+{
+	struct fw_eq_ofld_cmd c;
+
+	memset(&c, 0, sizeof(c));
+	c.op_to_vfn = htonl(FW_CMD_OP(FW_EQ_OFLD_CMD) | FW_CMD_REQUEST |
+			    FW_CMD_EXEC | FW_EQ_OFLD_CMD_PFN(pf) |
+			    FW_EQ_OFLD_CMD_VFN(vf));
+	c.alloc_to_len16 = htonl(FW_EQ_OFLD_CMD_FREE | FW_LEN16(c));
+	c.eqid_pkd = htonl(FW_EQ_OFLD_CMD_EQID(eqid));
+	return t4_wr_mbox(adap, mbox, &c, sizeof(c), NULL);
+}
+
+/**
+ *	t4_handle_fw_rpl - process a FW reply message
+ *	@adap: the adapter
+ *	@rpl: start of the FW message
+ *
+ *	Processes a FW message, such as link state change messages.
+ */
+int t4_handle_fw_rpl(struct adapter *adap, const __be64 *rpl)
+{
+	u8 opcode = *(const u8 *)rpl;
+
+	if (opcode == FW_PORT_CMD) {    /* link/module state change message */
+		int speed = 0, fc = 0;
+		const struct fw_port_cmd *p = (void *)rpl;
+		int chan = FW_PORT_CMD_PORTID_GET(ntohl(p->op_to_portid));
+		int port = adap->chan_map[chan];
+		struct port_info *pi = adap2pinfo(adap, port);
+		struct link_config *lc = &pi->link_cfg;
+		u32 stat = ntohl(p->u.info.lstatus_to_modtype);
+		int link_ok = (stat & FW_PORT_CMD_LSTATUS) != 0;
+		u32 mod = FW_PORT_CMD_MODTYPE_GET(stat);
+
+		if (stat & FW_PORT_CMD_RXPAUSE)
+			fc |= PAUSE_RX;
+		if (stat & FW_PORT_CMD_TXPAUSE)
+			fc |= PAUSE_TX;
+		if (stat & FW_PORT_CMD_LSPEED(FW_PORT_CAP_SPEED_100M))
+			speed = SPEED_100;
+		else if (stat & FW_PORT_CMD_LSPEED(FW_PORT_CAP_SPEED_1G))
+			speed = SPEED_1000;
+		else if (stat & FW_PORT_CMD_LSPEED(FW_PORT_CAP_SPEED_10G))
+			speed = SPEED_10000;
+
+		if (link_ok != lc->link_ok || speed != lc->speed ||
+		    fc != lc->fc) {                    /* something changed */
+			lc->link_ok = link_ok;
+			lc->speed = speed;
+			lc->fc = fc;
+			t4_os_link_changed(adap, port, link_ok);
+		}
+		if (mod != pi->mod_type) {
+			pi->mod_type = mod;
+			t4_os_portmod_changed(adap, port);
+		}
+	}
+	return 0;
+}
+
+static void __devinit get_pci_mode(struct adapter *adapter,
+				   struct pci_params *p)
+{
+	u16 val;
+	u32 pcie_cap = pci_pcie_cap(adapter->pdev);
+
+	if (pcie_cap) {
+		pci_read_config_word(adapter->pdev, pcie_cap + PCI_EXP_LNKSTA,
+				     &val);
+		p->speed = val & PCI_EXP_LNKSTA_CLS;
+		p->width = (val & PCI_EXP_LNKSTA_NLW) >> 4;
+	}
+}
+
+/**
+ *	init_link_config - initialize a link's SW state
+ *	@lc: structure holding the link state
+ *	@caps: link capabilities
+ *
+ *	Initializes the SW state maintained for each link, including the link's
+ *	capabilities and default speed/flow-control/autonegotiation settings.
+ */
+static void __devinit init_link_config(struct link_config *lc,
+				       unsigned int caps)
+{
+	lc->supported = caps;
+	lc->requested_speed = 0;
+	lc->speed = 0;
+	lc->requested_fc = lc->fc = PAUSE_RX | PAUSE_TX;
+	if (lc->supported & FW_PORT_CAP_ANEG) {
+		lc->advertising = lc->supported & ADVERT_MASK;
+		lc->autoneg = AUTONEG_ENABLE;
+		lc->requested_fc |= PAUSE_AUTONEG;
+	} else {
+		lc->advertising = 0;
+		lc->autoneg = AUTONEG_DISABLE;
+	}
+}
+
+static int __devinit wait_dev_ready(struct adapter *adap)
+{
+	if (t4_read_reg(adap, PL_WHOAMI) != 0xffffffff)
+		return 0;
+	msleep(500);
+	return t4_read_reg(adap, PL_WHOAMI) != 0xffffffff ? 0 : -EIO;
+}
+
+/**
+ *	t4_prep_adapter - prepare SW and HW for operation
+ *	@adapter: the adapter
+ *	@reset: if true perform a HW reset
+ *
+ *	Initialize adapter SW state for the various HW modules, set initial
+ *	values for some adapter tunables, take PHYs out of reset, and
+ *	initialize the MDIO interface.
+ */
+int __devinit t4_prep_adapter(struct adapter *adapter)
+{
+	int ret;
+
+	ret = wait_dev_ready(adapter);
+	if (ret < 0)
+		return ret;
+
+	get_pci_mode(adapter, &adapter->params.pci);
+	adapter->params.rev = t4_read_reg(adapter, PL_REV);
+
+	ret = get_vpd_params(adapter, &adapter->params.vpd);
+	if (ret < 0)
+		return ret;
+
+	init_cong_ctrl(adapter->params.a_wnd, adapter->params.b_wnd);
+
+	/*
+	 * Default port for debugging in case we can't reach FW.
+	 */
+	adapter->params.nports = 1;
+	adapter->params.portvec = 1;
+	return 0;
+}
+
+int __devinit t4_port_init(struct adapter *adap, int mbox, int pf, int vf)
+{
+	u8 addr[6];
+	int ret, i, j = 0;
+	struct fw_port_cmd c;
+
+	memset(&c, 0, sizeof(c));
+
+	for_each_port(adap, i) {
+		unsigned int rss_size;
+		struct port_info *p = adap2pinfo(adap, i);
+
+		while ((adap->params.portvec & (1 << j)) == 0)
+			j++;
+
+		c.op_to_portid = htonl(FW_CMD_OP(FW_PORT_CMD) |
+				       FW_CMD_REQUEST | FW_CMD_READ |
+				       FW_PORT_CMD_PORTID(j));
+		c.action_to_len16 = htonl(
+			FW_PORT_CMD_ACTION(FW_PORT_ACTION_GET_PORT_INFO) |
+			FW_LEN16(c));
+		ret = t4_wr_mbox(adap, mbox, &c, sizeof(c), &c);
+		if (ret)
+			return ret;
+
+		ret = t4_alloc_vi(adap, mbox, j, pf, vf, 1, addr, &rss_size);
+		if (ret < 0)
+			return ret;
+
+		p->viid = ret;
+		p->tx_chan = j;
+		p->lport = j;
+		p->rss_size = rss_size;
+		memcpy(adap->port[i]->dev_addr, addr, ETH_ALEN);
+		memcpy(adap->port[i]->perm_addr, addr, ETH_ALEN);
+
+		ret = ntohl(c.u.info.lstatus_to_modtype);
+		p->mdio_addr = (ret & FW_PORT_CMD_MDIOCAP) ?
+			FW_PORT_CMD_MDIOADDR_GET(ret) : -1;
+		p->port_type = FW_PORT_CMD_PTYPE_GET(ret);
+		p->mod_type = FW_PORT_CMD_MODTYPE_GET(ret);
+
+		init_link_config(&p->link_cfg, ntohs(c.u.info.pcap));
+		j++;
+	}
+	return 0;
+}
diff --git a/drivers/net/cxgb4/t4_hw.h b/drivers/net/cxgb4/t4_hw.h
new file mode 100644
index 000000000000..025623285c93
--- /dev/null
+++ b/drivers/net/cxgb4/t4_hw.h
@@ -0,0 +1,100 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __T4_HW_H
+#define __T4_HW_H
+
+#include <linux/types.h>
+
+enum {
+	NCHAN          = 4,     /* # of HW channels */
+	MAX_MTU        = 9600,  /* max MAC MTU, excluding header + FCS */
+	EEPROMSIZE     = 17408, /* Serial EEPROM physical size */
+	EEPROMVSIZE    = 32768, /* Serial EEPROM virtual address space size */
+	RSS_NENTRIES   = 2048,  /* # of entries in RSS mapping table */
+	TCB_SIZE       = 128,   /* TCB size */
+	NMTUS          = 16,    /* size of MTU table */
+	NCCTRL_WIN     = 32,    /* # of congestion control windows */
+	NEXACT_MAC     = 336,   /* # of exact MAC address filters */
+	L2T_SIZE       = 4096,  /* # of L2T entries */
+	MBOX_LEN       = 64,    /* mailbox size in bytes */
+	TRACE_LEN      = 112,   /* length of trace data and mask */
+	FILTER_OPT_LEN = 36,    /* filter tuple width for optional components */
+	NWOL_PAT       = 8,     /* # of WoL patterns */
+	WOL_PAT_LEN    = 128,   /* length of WoL patterns */
+};
+
+enum {
+	SF_PAGE_SIZE = 256,           /* serial flash page size */
+	SF_SEC_SIZE = 64 * 1024,      /* serial flash sector size */
+	SF_SIZE = SF_SEC_SIZE * 16,   /* serial flash size */
+};
+
+enum { RSP_TYPE_FLBUF, RSP_TYPE_CPL, RSP_TYPE_INTR }; /* response entry types */
+
+enum { MBOX_OWNER_NONE, MBOX_OWNER_FW, MBOX_OWNER_DRV };    /* mailbox owners */
+
+enum {
+	SGE_MAX_WR_LEN = 512,     /* max WR size in bytes */
+	SGE_NTIMERS = 6,          /* # of interrupt holdoff timer values */
+	SGE_NCOUNTERS = 4,        /* # of interrupt packet counter values */
+};
+
+struct sge_qstat {                /* data written to SGE queue status entries */
+	__be32 qid;
+	__be16 cidx;
+	__be16 pidx;
+};
+
+/*
+ * Structure for last 128 bits of response descriptors
+ */
+struct rsp_ctrl {
+	__be32 hdrbuflen_pidx;
+	__be32 pldbuflen_qid;
+	union {
+		u8 type_gen;
+		__be64 last_flit;
+	};
+};
+
+#define RSPD_NEWBUF 0x80000000U
+#define RSPD_LEN    0x7fffffffU
+
+#define RSPD_GEN(x)  ((x) >> 7)
+#define RSPD_TYPE(x) (((x) >> 4) & 3)
+
+#define QINTR_CNT_EN       0x1
+#define QINTR_TIMER_IDX(x) ((x) << 1)
+#endif /* __T4_HW_H */
diff --git a/drivers/net/cxgb4/t4_msg.h b/drivers/net/cxgb4/t4_msg.h
new file mode 100644
index 000000000000..fdb117443144
--- /dev/null
+++ b/drivers/net/cxgb4/t4_msg.h
@@ -0,0 +1,664 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __T4_MSG_H
+#define __T4_MSG_H
+
+#include <linux/types.h>
+
+enum {
+	CPL_PASS_OPEN_REQ     = 0x1,
+	CPL_PASS_ACCEPT_RPL   = 0x2,
+	CPL_ACT_OPEN_REQ      = 0x3,
+	CPL_SET_TCB_FIELD     = 0x5,
+	CPL_GET_TCB           = 0x6,
+	CPL_CLOSE_CON_REQ     = 0x8,
+	CPL_CLOSE_LISTSRV_REQ = 0x9,
+	CPL_ABORT_REQ         = 0xA,
+	CPL_ABORT_RPL         = 0xB,
+	CPL_RX_DATA_ACK       = 0xD,
+	CPL_TX_PKT            = 0xE,
+	CPL_L2T_WRITE_REQ     = 0x12,
+	CPL_TID_RELEASE       = 0x1A,
+
+	CPL_CLOSE_LISTSRV_RPL = 0x20,
+	CPL_L2T_WRITE_RPL     = 0x23,
+	CPL_PASS_OPEN_RPL     = 0x24,
+	CPL_ACT_OPEN_RPL      = 0x25,
+	CPL_PEER_CLOSE        = 0x26,
+	CPL_ABORT_REQ_RSS     = 0x2B,
+	CPL_ABORT_RPL_RSS     = 0x2D,
+
+	CPL_CLOSE_CON_RPL     = 0x32,
+	CPL_ISCSI_HDR         = 0x33,
+	CPL_RDMA_CQE          = 0x35,
+	CPL_RDMA_CQE_READ_RSP = 0x36,
+	CPL_RDMA_CQE_ERR      = 0x37,
+	CPL_RX_DATA           = 0x39,
+	CPL_SET_TCB_RPL       = 0x3A,
+	CPL_RX_PKT            = 0x3B,
+	CPL_RX_DDP_COMPLETE   = 0x3F,
+
+	CPL_ACT_ESTABLISH     = 0x40,
+	CPL_PASS_ESTABLISH    = 0x41,
+	CPL_RX_DATA_DDP       = 0x42,
+	CPL_PASS_ACCEPT_REQ   = 0x44,
+
+	CPL_RDMA_READ_REQ     = 0x60,
+
+	CPL_PASS_OPEN_REQ6    = 0x81,
+	CPL_ACT_OPEN_REQ6     = 0x83,
+
+	CPL_RDMA_TERMINATE    = 0xA2,
+	CPL_RDMA_WRITE        = 0xA4,
+	CPL_SGE_EGR_UPDATE    = 0xA5,
+
+	CPL_TRACE_PKT         = 0xB0,
+
+	CPL_FW4_MSG           = 0xC0,
+	CPL_FW4_PLD           = 0xC1,
+	CPL_FW4_ACK           = 0xC3,
+
+	CPL_FW6_MSG           = 0xE0,
+	CPL_FW6_PLD           = 0xE1,
+	CPL_TX_PKT_LSO        = 0xED,
+	CPL_TX_PKT_XT         = 0xEE,
+
+	NUM_CPL_CMDS
+};
+
+enum CPL_error {
+	CPL_ERR_NONE               = 0,
+	CPL_ERR_TCAM_FULL          = 3,
+	CPL_ERR_BAD_LENGTH         = 15,
+	CPL_ERR_BAD_ROUTE          = 18,
+	CPL_ERR_CONN_RESET         = 20,
+	CPL_ERR_CONN_EXIST_SYNRECV = 21,
+	CPL_ERR_CONN_EXIST         = 22,
+	CPL_ERR_ARP_MISS           = 23,
+	CPL_ERR_BAD_SYN            = 24,
+	CPL_ERR_CONN_TIMEDOUT      = 30,
+	CPL_ERR_XMIT_TIMEDOUT      = 31,
+	CPL_ERR_PERSIST_TIMEDOUT   = 32,
+	CPL_ERR_FINWAIT2_TIMEDOUT  = 33,
+	CPL_ERR_KEEPALIVE_TIMEDOUT = 34,
+	CPL_ERR_RTX_NEG_ADVICE     = 35,
+	CPL_ERR_PERSIST_NEG_ADVICE = 36,
+	CPL_ERR_ABORT_FAILED       = 42,
+	CPL_ERR_IWARP_FLM          = 50,
+};
+
+enum {
+	ULP_MODE_NONE          = 0,
+	ULP_MODE_ISCSI         = 2,
+	ULP_MODE_RDMA          = 4,
+	ULP_MODE_FCOE          = 6,
+};
+
+enum {
+	ULP_CRC_HEADER = 1 << 0,
+	ULP_CRC_DATA   = 1 << 1
+};
+
+enum {
+	CPL_ABORT_SEND_RST = 0,
+	CPL_ABORT_NO_RST,
+};
+
+enum {                     /* TX_PKT_XT checksum types */
+	TX_CSUM_TCP    = 0,
+	TX_CSUM_UDP    = 1,
+	TX_CSUM_CRC16  = 4,
+	TX_CSUM_CRC32  = 5,
+	TX_CSUM_CRC32C = 6,
+	TX_CSUM_FCOE   = 7,
+	TX_CSUM_TCPIP  = 8,
+	TX_CSUM_UDPIP  = 9,
+	TX_CSUM_TCPIP6 = 10,
+	TX_CSUM_UDPIP6 = 11,
+	TX_CSUM_IP     = 12,
+};
+
+union opcode_tid {
+	__be32 opcode_tid;
+	u8 opcode;
+};
+
+#define CPL_OPCODE(x) ((x) << 24)
+#define MK_OPCODE_TID(opcode, tid) (CPL_OPCODE(opcode) | (tid))
+#define OPCODE_TID(cmd) ((cmd)->ot.opcode_tid)
+#define GET_TID(cmd) (ntohl(OPCODE_TID(cmd)) & 0xFFFFFF)
+
+/* partitioning of TID fields that also carry a queue id */
+#define GET_TID_TID(x) ((x) & 0x3fff)
+#define GET_TID_QID(x) (((x) >> 14) & 0x3ff)
+#define TID_QID(x)     ((x) << 14)
+
+struct rss_header {
+	u8 opcode;
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	u8 channel:2;
+	u8 filter_hit:1;
+	u8 filter_tid:1;
+	u8 hash_type:2;
+	u8 ipv6:1;
+	u8 send2fw:1;
+#else
+	u8 send2fw:1;
+	u8 ipv6:1;
+	u8 hash_type:2;
+	u8 filter_tid:1;
+	u8 filter_hit:1;
+	u8 channel:2;
+#endif
+	__be16 qid;
+	__be32 hash_val;
+};
+
+struct work_request_hdr {
+	__be32 wr_hi;
+	__be32 wr_mid;
+	__be64 wr_lo;
+};
+
+#define WR_HDR struct work_request_hdr wr
+
+struct cpl_pass_open_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 local_port;
+	__be16 peer_port;
+	__be32 local_ip;
+	__be32 peer_ip;
+	__be64 opt0;
+#define TX_CHAN(x)    ((x) << 2)
+#define DELACK(x)     ((x) << 5)
+#define ULP_MODE(x)   ((x) << 8)
+#define RCV_BUFSIZ(x) ((x) << 12)
+#define DSCP(x)       ((x) << 22)
+#define SMAC_SEL(x)   ((u64)(x) << 28)
+#define L2T_IDX(x)    ((u64)(x) << 36)
+#define NAGLE(x)      ((u64)(x) << 49)
+#define WND_SCALE(x)  ((u64)(x) << 50)
+#define KEEP_ALIVE(x) ((u64)(x) << 54)
+#define MSS_IDX(x)    ((u64)(x) << 60)
+	__be64 opt1;
+#define SYN_RSS_ENABLE   (1 << 0)
+#define SYN_RSS_QUEUE(x) ((x) << 2)
+#define CONN_POLICY_ASK  (1 << 22)
+};
+
+struct cpl_pass_open_req6 {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 local_port;
+	__be16 peer_port;
+	__be64 local_ip_hi;
+	__be64 local_ip_lo;
+	__be64 peer_ip_hi;
+	__be64 peer_ip_lo;
+	__be64 opt0;
+	__be64 opt1;
+};
+
+struct cpl_pass_open_rpl {
+	union opcode_tid ot;
+	u8 rsvd[3];
+	u8 status;
+};
+
+struct cpl_pass_accept_rpl {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 opt2;
+#define RSS_QUEUE(x)         ((x) << 0)
+#define RSS_QUEUE_VALID      (1 << 10)
+#define RX_COALESCE_VALID(x) ((x) << 11)
+#define RX_COALESCE(x)       ((x) << 12)
+#define TX_QUEUE(x)          ((x) << 23)
+#define RX_CHANNEL(x)        ((x) << 26)
+#define WND_SCALE_EN(x)      ((x) << 28)
+#define TSTAMPS_EN(x)        ((x) << 29)
+#define SACK_EN(x)           ((x) << 30)
+	__be64 opt0;
+};
+
+struct cpl_act_open_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 local_port;
+	__be16 peer_port;
+	__be32 local_ip;
+	__be32 peer_ip;
+	__be64 opt0;
+	__be32 params;
+	__be32 opt2;
+};
+
+struct cpl_act_open_req6 {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 local_port;
+	__be16 peer_port;
+	__be64 local_ip_hi;
+	__be64 local_ip_lo;
+	__be64 peer_ip_hi;
+	__be64 peer_ip_lo;
+	__be64 opt0;
+	__be32 params;
+	__be32 opt2;
+};
+
+struct cpl_act_open_rpl {
+	union opcode_tid ot;
+	__be32 atid_status;
+#define GET_AOPEN_STATUS(x) ((x) & 0xff)
+#define GET_AOPEN_ATID(x)   (((x) >> 8) & 0xffffff)
+};
+
+struct cpl_pass_establish {
+	union opcode_tid ot;
+	__be32 rsvd;
+	__be32 tos_stid;
+#define GET_POPEN_TID(x) ((x) & 0xffffff)
+#define GET_POPEN_TOS(x) (((x) >> 24) & 0xff)
+	__be16 mac_idx;
+	__be16 tcp_opt;
+#define GET_TCPOPT_WSCALE_OK(x)  (((x) >> 5) & 1)
+#define GET_TCPOPT_SACK(x)       (((x) >> 6) & 1)
+#define GET_TCPOPT_TSTAMP(x)     (((x) >> 7) & 1)
+#define GET_TCPOPT_SND_WSCALE(x) (((x) >> 8) & 0xf)
+#define GET_TCPOPT_MSS(x)        (((x) >> 12) & 0xf)
+	__be32 snd_isn;
+	__be32 rcv_isn;
+};
+
+struct cpl_act_establish {
+	union opcode_tid ot;
+	__be32 rsvd;
+	__be32 tos_atid;
+	__be16 mac_idx;
+	__be16 tcp_opt;
+	__be32 snd_isn;
+	__be32 rcv_isn;
+};
+
+struct cpl_get_tcb {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 reply_ctrl;
+#define QUEUENO(x)    ((x) << 0)
+#define REPLY_CHAN(x) ((x) << 14)
+#define NO_REPLY(x)   ((x) << 15)
+	__be16 cookie;
+};
+
+struct cpl_set_tcb_field {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 reply_ctrl;
+	__be16 word_cookie;
+#define TCB_WORD(x)   ((x) << 0)
+#define TCB_COOKIE(x) ((x) << 5)
+	__be64 mask;
+	__be64 val;
+};
+
+struct cpl_set_tcb_rpl {
+	union opcode_tid ot;
+	__be16 rsvd;
+	u8 cookie;
+	u8 status;
+	__be64 oldval;
+};
+
+struct cpl_close_con_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 rsvd;
+};
+
+struct cpl_close_con_rpl {
+	union opcode_tid ot;
+	u8 rsvd[3];
+	u8 status;
+	__be32 snd_nxt;
+	__be32 rcv_nxt;
+};
+
+struct cpl_close_listsvr_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 reply_ctrl;
+#define LISTSVR_IPV6 (1 << 14)
+	__be16 rsvd;
+};
+
+struct cpl_close_listsvr_rpl {
+	union opcode_tid ot;
+	u8 rsvd[3];
+	u8 status;
+};
+
+struct cpl_abort_req_rss {
+	union opcode_tid ot;
+	u8 rsvd[3];
+	u8 status;
+};
+
+struct cpl_abort_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 rsvd0;
+	u8 rsvd1;
+	u8 cmd;
+	u8 rsvd2[6];
+};
+
+struct cpl_abort_rpl_rss {
+	union opcode_tid ot;
+	u8 rsvd[3];
+	u8 status;
+};
+
+struct cpl_abort_rpl {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 rsvd0;
+	u8 rsvd1;
+	u8 cmd;
+	u8 rsvd2[6];
+};
+
+struct cpl_peer_close {
+	union opcode_tid ot;
+	__be32 rcv_nxt;
+};
+
+struct cpl_tid_release {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 rsvd;
+};
+
+struct cpl_tx_pkt_core {
+	__be32 ctrl0;
+#define TXPKT_VF(x)        ((x) << 0)
+#define TXPKT_PF(x)        ((x) << 8)
+#define TXPKT_VF_VLD       (1 << 11)
+#define TXPKT_OVLAN_IDX(x) ((x) << 12)
+#define TXPKT_INTF(x)      ((x) << 16)
+#define TXPKT_INS_OVLAN    (1 << 21)
+#define TXPKT_OPCODE(x)    ((x) << 24)
+	__be16 pack;
+	__be16 len;
+	__be64 ctrl1;
+#define TXPKT_CSUM_END(x)   ((x) << 12)
+#define TXPKT_CSUM_START(x) ((x) << 20)
+#define TXPKT_IPHDR_LEN(x)  ((u64)(x) << 20)
+#define TXPKT_CSUM_LOC(x)   ((u64)(x) << 30)
+#define TXPKT_ETHHDR_LEN(x) ((u64)(x) << 34)
+#define TXPKT_CSUM_TYPE(x)  ((u64)(x) << 40)
+#define TXPKT_VLAN(x)       ((u64)(x) << 44)
+#define TXPKT_VLAN_VLD      (1ULL << 60)
+#define TXPKT_IPCSUM_DIS    (1ULL << 62)
+#define TXPKT_L4CSUM_DIS    (1ULL << 63)
+};
+
+struct cpl_tx_pkt {
+	WR_HDR;
+	struct cpl_tx_pkt_core c;
+};
+
+#define cpl_tx_pkt_xt cpl_tx_pkt
+
+struct cpl_tx_pkt_lso {
+	WR_HDR;
+	__be32 lso_ctrl;
+#define LSO_TCPHDR_LEN(x) ((x) << 0)
+#define LSO_IPHDR_LEN(x)  ((x) << 4)
+#define LSO_ETHHDR_LEN(x) ((x) << 16)
+#define LSO_IPV6(x)       ((x) << 20)
+#define LSO_LAST_SLICE    (1 << 22)
+#define LSO_FIRST_SLICE   (1 << 23)
+#define LSO_OPCODE(x)     ((x) << 24)
+	__be16 ipid_ofst;
+	__be16 mss;
+	__be32 seqno_offset;
+	__be32 len;
+	/* encapsulated CPL (TX_PKT, TX_PKT_XT or TX_DATA) follows here */
+};
+
+struct cpl_iscsi_hdr {
+	union opcode_tid ot;
+	__be16 pdu_len_ddp;
+#define ISCSI_PDU_LEN(x) ((x) & 0x7FFF)
+#define ISCSI_DDP        (1 << 15)
+	__be16 len;
+	__be32 seq;
+	__be16 urg;
+	u8 rsvd;
+	u8 status;
+};
+
+struct cpl_rx_data {
+	union opcode_tid ot;
+	__be16 rsvd;
+	__be16 len;
+	__be32 seq;
+	__be16 urg;
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	u8 dack_mode:2;
+	u8 psh:1;
+	u8 heartbeat:1;
+	u8 ddp_off:1;
+	u8 :3;
+#else
+	u8 :3;
+	u8 ddp_off:1;
+	u8 heartbeat:1;
+	u8 psh:1;
+	u8 dack_mode:2;
+#endif
+	u8 status;
+};
+
+struct cpl_rx_data_ack {
+	WR_HDR;
+	union opcode_tid ot;
+	__be32 credit_dack;
+#define RX_CREDITS(x)   ((x) << 0)
+#define RX_FORCE_ACK(x) ((x) << 28)
+};
+
+struct cpl_rx_pkt {
+	u8 opcode;
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	u8 iff:4;
+	u8 csum_calc:1;
+	u8 ipmi_pkt:1;
+	u8 vlan_ex:1;
+	u8 ip_frag:1;
+#else
+	u8 ip_frag:1;
+	u8 vlan_ex:1;
+	u8 ipmi_pkt:1;
+	u8 csum_calc:1;
+	u8 iff:4;
+#endif
+	__be16 csum;
+	__be16 vlan;
+	__be16 len;
+	__be32 l2info;
+#define RXF_UDP (1 << 22)
+#define RXF_TCP (1 << 23)
+	__be16 hdr_len;
+	__be16 err_vec;
+};
+
+struct cpl_trace_pkt {
+	u8 opcode;
+	u8 intf;
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	u8 runt:4;
+	u8 filter_hit:4;
+	u8 :6;
+	u8 err:1;
+	u8 trunc:1;
+#else
+	u8 filter_hit:4;
+	u8 runt:4;
+	u8 trunc:1;
+	u8 err:1;
+	u8 :6;
+#endif
+	__be16 rsvd;
+	__be16 len;
+	__be64 tstamp;
+};
+
+struct cpl_l2t_write_req {
+	WR_HDR;
+	union opcode_tid ot;
+	__be16 params;
+#define L2T_W_INFO(x)    ((x) << 2)
+#define L2T_W_PORT(x)    ((x) << 8)
+#define L2T_W_NOREPLY(x) ((x) << 15)
+	__be16 l2t_idx;
+	__be16 vlan;
+	u8 dst_mac[6];
+};
+
+struct cpl_l2t_write_rpl {
+	union opcode_tid ot;
+	u8 status;
+	u8 rsvd[3];
+};
+
+struct cpl_rdma_terminate {
+	union opcode_tid ot;
+	__be16 rsvd;
+	__be16 len;
+};
+
+struct cpl_sge_egr_update {
+	__be32 opcode_qid;
+#define EGR_QID(x) ((x) & 0x1FFFF)
+	__be16 cidx;
+	__be16 pidx;
+};
+
+struct cpl_fw4_pld {
+	u8 opcode;
+	u8 rsvd0[3];
+	u8 type;
+	u8 rsvd1;
+	__be16 len;
+	__be64 data;
+	__be64 rsvd2;
+};
+
+struct cpl_fw6_pld {
+	u8 opcode;
+	u8 rsvd[5];
+	__be16 len;
+	__be64 data[4];
+};
+
+struct cpl_fw4_msg {
+	u8 opcode;
+	u8 type;
+	__be16 rsvd0;
+	__be32 rsvd1;
+	__be64 data[2];
+};
+
+struct cpl_fw4_ack {
+	union opcode_tid ot;
+	u8 credits;
+	u8 rsvd0[2];
+	u8 seq_vld;
+	__be32 snd_nxt;
+	__be32 snd_una;
+	__be64 rsvd1;
+};
+
+struct cpl_fw6_msg {
+	u8 opcode;
+	u8 type;
+	__be16 rsvd0;
+	__be32 rsvd1;
+	__be64 data[4];
+};
+
+enum {
+	ULP_TX_MEM_READ = 2,
+	ULP_TX_MEM_WRITE = 3,
+	ULP_TX_PKT = 4
+};
+
+enum {
+	ULP_TX_SC_NOOP = 0x80,
+	ULP_TX_SC_IMM  = 0x81,
+	ULP_TX_SC_DSGL = 0x82,
+	ULP_TX_SC_ISGL = 0x83
+};
+
+struct ulptx_sge_pair {
+	__be32 len[2];
+	__be64 addr[2];
+};
+
+struct ulptx_sgl {
+	__be32 cmd_nsge;
+#define ULPTX_CMD(x) ((x) << 24)
+#define ULPTX_NSGE(x) ((x) << 0)
+	__be32 len0;
+	__be64 addr0;
+	struct ulptx_sge_pair sge[0];
+};
+
+struct ulp_mem_io {
+	WR_HDR;
+	__be32 cmd;
+#define ULP_MEMIO_ORDER(x) ((x) << 23)
+	__be32 len16;             /* command length */
+	__be32 dlen;              /* data length in 32-byte units */
+#define ULP_MEMIO_DATA_LEN(x) ((x) << 0)
+	__be32 lock_addr;
+#define ULP_MEMIO_ADDR(x) ((x) << 0)
+#define ULP_MEMIO_LOCK(x) ((x) << 31)
+};
+
+#endif  /* __T4_MSG_H */
diff --git a/drivers/net/cxgb4/t4_regs.h b/drivers/net/cxgb4/t4_regs.h
new file mode 100644
index 000000000000..5ed56483cbc2
--- /dev/null
+++ b/drivers/net/cxgb4/t4_regs.h
@@ -0,0 +1,878 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __T4_REGS_H
+#define __T4_REGS_H
+
+#define MYPF_BASE 0x1b000
+#define MYPF_REG(reg_addr) (MYPF_BASE + (reg_addr))
+
+#define PF0_BASE 0x1e000
+#define PF0_REG(reg_addr) (PF0_BASE + (reg_addr))
+
+#define PF_STRIDE 0x400
+#define PF_BASE(idx) (PF0_BASE + (idx) * PF_STRIDE)
+#define PF_REG(idx, reg) (PF_BASE(idx) + (reg))
+
+#define MYPORT_BASE 0x1c000
+#define MYPORT_REG(reg_addr) (MYPORT_BASE + (reg_addr))
+
+#define PORT0_BASE 0x20000
+#define PORT0_REG(reg_addr) (PORT0_BASE + (reg_addr))
+
+#define PORT_STRIDE 0x2000
+#define PORT_BASE(idx) (PORT0_BASE + (idx) * PORT_STRIDE)
+#define PORT_REG(idx, reg) (PORT_BASE(idx) + (reg))
+
+#define EDC_STRIDE (EDC_1_BASE_ADDR - EDC_0_BASE_ADDR)
+#define EDC_REG(reg, idx) (reg + EDC_STRIDE * idx)
+
+#define PCIE_MEM_ACCESS_REG(reg_addr, idx) ((reg_addr) + (idx) * 8)
+#define PCIE_MAILBOX_REG(reg_addr, idx) ((reg_addr) + (idx) * 8)
+#define MC_BIST_STATUS_REG(reg_addr, idx) ((reg_addr) + (idx) * 4)
+#define EDC_BIST_STATUS_REG(reg_addr, idx) ((reg_addr) + (idx) * 4)
+
+#define SGE_PF_KDOORBELL 0x0
+#define  QID_MASK    0xffff8000U
+#define  QID_SHIFT   15
+#define  QID(x)      ((x) << QID_SHIFT)
+#define  DBPRIO      0x00004000U
+#define  PIDX_MASK   0x00003fffU
+#define  PIDX_SHIFT  0
+#define  PIDX(x)     ((x) << PIDX_SHIFT)
+
+#define SGE_PF_GTS 0x4
+#define  INGRESSQID_MASK   0xffff0000U
+#define  INGRESSQID_SHIFT  16
+#define  INGRESSQID(x)     ((x) << INGRESSQID_SHIFT)
+#define  TIMERREG_MASK     0x0000e000U
+#define  TIMERREG_SHIFT    13
+#define  TIMERREG(x)       ((x) << TIMERREG_SHIFT)
+#define  SEINTARM_MASK     0x00001000U
+#define  SEINTARM_SHIFT    12
+#define  SEINTARM(x)       ((x) << SEINTARM_SHIFT)
+#define  CIDXINC_MASK      0x00000fffU
+#define  CIDXINC_SHIFT     0
+#define  CIDXINC(x)        ((x) << CIDXINC_SHIFT)
+
+#define SGE_CONTROL 0x1008
+#define  DCASYSTYPE             0x00080000U
+#define  RXPKTCPLMODE           0x00040000U
+#define  EGRSTATUSPAGESIZE      0x00020000U
+#define  PKTSHIFT_MASK          0x00001c00U
+#define  PKTSHIFT_SHIFT         10
+#define  PKTSHIFT(x)            ((x) << PKTSHIFT_SHIFT)
+#define  INGPCIEBOUNDARY_MASK   0x00000380U
+#define  INGPCIEBOUNDARY_SHIFT  7
+#define  INGPCIEBOUNDARY(x)     ((x) << INGPCIEBOUNDARY_SHIFT)
+#define  INGPADBOUNDARY_MASK    0x00000070U
+#define  INGPADBOUNDARY_SHIFT   4
+#define  INGPADBOUNDARY(x)      ((x) << INGPADBOUNDARY_SHIFT)
+#define  EGRPCIEBOUNDARY_MASK   0x0000000eU
+#define  EGRPCIEBOUNDARY_SHIFT  1
+#define  EGRPCIEBOUNDARY(x)     ((x) << EGRPCIEBOUNDARY_SHIFT)
+#define  GLOBALENABLE           0x00000001U
+
+#define SGE_HOST_PAGE_SIZE 0x100c
+#define  HOSTPAGESIZEPF0_MASK   0x0000000fU
+#define  HOSTPAGESIZEPF0_SHIFT  0
+#define  HOSTPAGESIZEPF0(x)     ((x) << HOSTPAGESIZEPF0_SHIFT)
+
+#define SGE_EGRESS_QUEUES_PER_PAGE_PF 0x1010
+#define  QUEUESPERPAGEPF0_MASK   0x0000000fU
+#define  QUEUESPERPAGEPF0_GET(x) ((x) & QUEUESPERPAGEPF0_MASK)
+
+#define SGE_INT_CAUSE1 0x1024
+#define SGE_INT_CAUSE2 0x1030
+#define SGE_INT_CAUSE3 0x103c
+#define  ERR_FLM_DBP               0x80000000U
+#define  ERR_FLM_IDMA1             0x40000000U
+#define  ERR_FLM_IDMA0             0x20000000U
+#define  ERR_FLM_HINT              0x10000000U
+#define  ERR_PCIE_ERROR3           0x08000000U
+#define  ERR_PCIE_ERROR2           0x04000000U
+#define  ERR_PCIE_ERROR1           0x02000000U
+#define  ERR_PCIE_ERROR0           0x01000000U
+#define  ERR_TIMER_ABOVE_MAX_QID   0x00800000U
+#define  ERR_CPL_EXCEED_IQE_SIZE   0x00400000U
+#define  ERR_INVALID_CIDX_INC      0x00200000U
+#define  ERR_ITP_TIME_PAUSED       0x00100000U
+#define  ERR_CPL_OPCODE_0          0x00080000U
+#define  ERR_DROPPED_DB            0x00040000U
+#define  ERR_DATA_CPL_ON_HIGH_QID1 0x00020000U
+#define  ERR_DATA_CPL_ON_HIGH_QID0 0x00010000U
+#define  ERR_BAD_DB_PIDX3          0x00008000U
+#define  ERR_BAD_DB_PIDX2          0x00004000U
+#define  ERR_BAD_DB_PIDX1          0x00002000U
+#define  ERR_BAD_DB_PIDX0          0x00001000U
+#define  ERR_ING_PCIE_CHAN         0x00000800U
+#define  ERR_ING_CTXT_PRIO         0x00000400U
+#define  ERR_EGR_CTXT_PRIO         0x00000200U
+#define  DBFIFO_HP_INT             0x00000100U
+#define  DBFIFO_LP_INT             0x00000080U
+#define  REG_ADDRESS_ERR           0x00000040U
+#define  INGRESS_SIZE_ERR          0x00000020U
+#define  EGRESS_SIZE_ERR           0x00000010U
+#define  ERR_INV_CTXT3             0x00000008U
+#define  ERR_INV_CTXT2             0x00000004U
+#define  ERR_INV_CTXT1             0x00000002U
+#define  ERR_INV_CTXT0             0x00000001U
+
+#define SGE_INT_ENABLE3 0x1040
+#define SGE_FL_BUFFER_SIZE0 0x1044
+#define SGE_FL_BUFFER_SIZE1 0x1048
+#define SGE_INGRESS_RX_THRESHOLD 0x10a0
+#define  THRESHOLD_0_MASK   0x3f000000U
+#define  THRESHOLD_0_SHIFT  24
+#define  THRESHOLD_0(x)     ((x) << THRESHOLD_0_SHIFT)
+#define  THRESHOLD_0_GET(x) (((x) & THRESHOLD_0_MASK) >> THRESHOLD_0_SHIFT)
+#define  THRESHOLD_1_MASK   0x003f0000U
+#define  THRESHOLD_1_SHIFT  16
+#define  THRESHOLD_1(x)     ((x) << THRESHOLD_1_SHIFT)
+#define  THRESHOLD_1_GET(x) (((x) & THRESHOLD_1_MASK) >> THRESHOLD_1_SHIFT)
+#define  THRESHOLD_2_MASK   0x00003f00U
+#define  THRESHOLD_2_SHIFT  8
+#define  THRESHOLD_2(x)     ((x) << THRESHOLD_2_SHIFT)
+#define  THRESHOLD_2_GET(x) (((x) & THRESHOLD_2_MASK) >> THRESHOLD_2_SHIFT)
+#define  THRESHOLD_3_MASK   0x0000003fU
+#define  THRESHOLD_3_SHIFT  0
+#define  THRESHOLD_3(x)     ((x) << THRESHOLD_3_SHIFT)
+#define  THRESHOLD_3_GET(x) (((x) & THRESHOLD_3_MASK) >> THRESHOLD_3_SHIFT)
+
+#define SGE_TIMER_VALUE_0_AND_1 0x10b8
+#define  TIMERVALUE0_MASK   0xffff0000U
+#define  TIMERVALUE0_SHIFT  16
+#define  TIMERVALUE0(x)     ((x) << TIMERVALUE0_SHIFT)
+#define  TIMERVALUE0_GET(x) (((x) & TIMERVALUE0_MASK) >> TIMERVALUE0_SHIFT)
+#define  TIMERVALUE1_MASK   0x0000ffffU
+#define  TIMERVALUE1_SHIFT  0
+#define  TIMERVALUE1(x)     ((x) << TIMERVALUE1_SHIFT)
+#define  TIMERVALUE1_GET(x) (((x) & TIMERVALUE1_MASK) >> TIMERVALUE1_SHIFT)
+
+#define SGE_TIMER_VALUE_2_AND_3 0x10bc
+#define SGE_TIMER_VALUE_4_AND_5 0x10c0
+#define SGE_DEBUG_INDEX 0x10cc
+#define SGE_DEBUG_DATA_HIGH 0x10d0
+#define SGE_DEBUG_DATA_LOW 0x10d4
+#define SGE_INGRESS_QUEUES_PER_PAGE_PF 0x10f4
+
+#define PCIE_PF_CLI 0x44
+#define PCIE_INT_CAUSE 0x3004
+#define  UNXSPLCPLERR  0x20000000U
+#define  PCIEPINT      0x10000000U
+#define  PCIESINT      0x08000000U
+#define  RPLPERR       0x04000000U
+#define  RXWRPERR      0x02000000U
+#define  RXCPLPERR     0x01000000U
+#define  PIOTAGPERR    0x00800000U
+#define  MATAGPERR     0x00400000U
+#define  INTXCLRPERR   0x00200000U
+#define  FIDPERR       0x00100000U
+#define  CFGSNPPERR    0x00080000U
+#define  HRSPPERR      0x00040000U
+#define  HREQPERR      0x00020000U
+#define  HCNTPERR      0x00010000U
+#define  DRSPPERR      0x00008000U
+#define  DREQPERR      0x00004000U
+#define  DCNTPERR      0x00002000U
+#define  CRSPPERR      0x00001000U
+#define  CREQPERR      0x00000800U
+#define  CCNTPERR      0x00000400U
+#define  TARTAGPERR    0x00000200U
+#define  PIOREQPERR    0x00000100U
+#define  PIOCPLPERR    0x00000080U
+#define  MSIXDIPERR    0x00000040U
+#define  MSIXDATAPERR  0x00000020U
+#define  MSIXADDRHPERR 0x00000010U
+#define  MSIXADDRLPERR 0x00000008U
+#define  MSIDATAPERR   0x00000004U
+#define  MSIADDRHPERR  0x00000002U
+#define  MSIADDRLPERR  0x00000001U
+
+#define PCIE_NONFAT_ERR 0x3010
+#define PCIE_MEM_ACCESS_BASE_WIN 0x3068
+#define  PCIEOFST_MASK   0xfffffc00U
+#define  BIR_MASK        0x00000300U
+#define  BIR_SHIFT       8
+#define  BIR(x)          ((x) << BIR_SHIFT)
+#define  WINDOW_MASK     0x000000ffU
+#define  WINDOW_SHIFT    0
+#define  WINDOW(x)       ((x) << WINDOW_SHIFT)
+
+#define PCIE_CORE_UTL_SYSTEM_BUS_AGENT_STATUS 0x5908
+#define  RNPP 0x80000000U
+#define  RPCP 0x20000000U
+#define  RCIP 0x08000000U
+#define  RCCP 0x04000000U
+#define  RFTP 0x00800000U
+#define  PTRP 0x00100000U
+
+#define PCIE_CORE_UTL_PCI_EXPRESS_PORT_STATUS 0x59a4
+#define  TPCP 0x40000000U
+#define  TNPP 0x20000000U
+#define  TFTP 0x10000000U
+#define  TCAP 0x08000000U
+#define  TCIP 0x04000000U
+#define  RCAP 0x02000000U
+#define  PLUP 0x00800000U
+#define  PLDN 0x00400000U
+#define  OTDD 0x00200000U
+#define  GTRP 0x00100000U
+#define  RDPE 0x00040000U
+#define  TDCE 0x00020000U
+#define  TDUE 0x00010000U
+
+#define MC_INT_CAUSE 0x7518
+#define  ECC_UE_INT_CAUSE 0x00000004U
+#define  ECC_CE_INT_CAUSE 0x00000002U
+#define  PERR_INT_CAUSE   0x00000001U
+
+#define MC_ECC_STATUS 0x751c
+#define  ECC_CECNT_MASK   0xffff0000U
+#define  ECC_CECNT_SHIFT  16
+#define  ECC_CECNT(x)     ((x) << ECC_CECNT_SHIFT)
+#define  ECC_CECNT_GET(x) (((x) & ECC_CECNT_MASK) >> ECC_CECNT_SHIFT)
+#define  ECC_UECNT_MASK   0x0000ffffU
+#define  ECC_UECNT_SHIFT  0
+#define  ECC_UECNT(x)     ((x) << ECC_UECNT_SHIFT)
+#define  ECC_UECNT_GET(x) (((x) & ECC_UECNT_MASK) >> ECC_UECNT_SHIFT)
+
+#define MC_BIST_CMD 0x7600
+#define  START_BIST          0x80000000U
+#define  BIST_CMD_GAP_MASK   0x0000ff00U
+#define  BIST_CMD_GAP_SHIFT  8
+#define  BIST_CMD_GAP(x)     ((x) << BIST_CMD_GAP_SHIFT)
+#define  BIST_OPCODE_MASK    0x00000003U
+#define  BIST_OPCODE_SHIFT   0
+#define  BIST_OPCODE(x)      ((x) << BIST_OPCODE_SHIFT)
+
+#define MC_BIST_CMD_ADDR 0x7604
+#define MC_BIST_CMD_LEN 0x7608
+#define MC_BIST_DATA_PATTERN 0x760c
+#define  BIST_DATA_TYPE_MASK   0x0000000fU
+#define  BIST_DATA_TYPE_SHIFT  0
+#define  BIST_DATA_TYPE(x)     ((x) << BIST_DATA_TYPE_SHIFT)
+
+#define MC_BIST_STATUS_RDATA 0x7688
+
+#define MA_EXT_MEMORY_BAR 0x77c8
+#define  EXT_MEM_SIZE_MASK   0x00000fffU
+#define  EXT_MEM_SIZE_SHIFT  0
+#define  EXT_MEM_SIZE_GET(x) (((x) & EXT_MEM_SIZE_MASK) >> EXT_MEM_SIZE_SHIFT)
+
+#define MA_TARGET_MEM_ENABLE 0x77d8
+#define  EXT_MEM_ENABLE 0x00000004U
+#define  EDRAM1_ENABLE  0x00000002U
+#define  EDRAM0_ENABLE  0x00000001U
+
+#define MA_INT_CAUSE 0x77e0
+#define  MEM_PERR_INT_CAUSE 0x00000002U
+#define  MEM_WRAP_INT_CAUSE 0x00000001U
+
+#define MA_INT_WRAP_STATUS 0x77e4
+#define  MEM_WRAP_ADDRESS_MASK   0xfffffff0U
+#define  MEM_WRAP_ADDRESS_SHIFT  4
+#define  MEM_WRAP_ADDRESS_GET(x) (((x) & MEM_WRAP_ADDRESS_MASK) >> MEM_WRAP_ADDRESS_SHIFT)
+#define  MEM_WRAP_CLIENT_NUM_MASK   0x0000000fU
+#define  MEM_WRAP_CLIENT_NUM_SHIFT  0
+#define  MEM_WRAP_CLIENT_NUM_GET(x) (((x) & MEM_WRAP_CLIENT_NUM_MASK) >> MEM_WRAP_CLIENT_NUM_SHIFT)
+
+#define MA_PARITY_ERROR_STATUS 0x77f4
+
+#define EDC_0_BASE_ADDR 0x7900
+
+#define EDC_BIST_CMD 0x7904
+#define EDC_BIST_CMD_ADDR 0x7908
+#define EDC_BIST_CMD_LEN 0x790c
+#define EDC_BIST_DATA_PATTERN 0x7910
+#define EDC_BIST_STATUS_RDATA 0x7928
+#define EDC_INT_CAUSE 0x7978
+#define  ECC_UE_PAR     0x00000020U
+#define  ECC_CE_PAR     0x00000010U
+#define  PERR_PAR_CAUSE 0x00000008U
+
+#define EDC_ECC_STATUS 0x797c
+
+#define EDC_1_BASE_ADDR 0x7980
+
+#define CIM_PF_MAILBOX_DATA 0x240
+#define CIM_PF_MAILBOX_CTRL 0x280
+#define  MBMSGVALID     0x00000008U
+#define  MBINTREQ       0x00000004U
+#define  MBOWNER_MASK   0x00000003U
+#define  MBOWNER_SHIFT  0
+#define  MBOWNER(x)     ((x) << MBOWNER_SHIFT)
+#define  MBOWNER_GET(x) (((x) & MBOWNER_MASK) >> MBOWNER_SHIFT)
+
+#define CIM_PF_HOST_INT_CAUSE 0x28c
+#define  MBMSGRDYINT 0x00080000U
+
+#define CIM_HOST_INT_CAUSE 0x7b2c
+#define  TIEQOUTPARERRINT  0x00100000U
+#define  TIEQINPARERRINT   0x00080000U
+#define  MBHOSTPARERR      0x00040000U
+#define  MBUPPARERR        0x00020000U
+#define  IBQPARERR         0x0001f800U
+#define  IBQTP0PARERR      0x00010000U
+#define  IBQTP1PARERR      0x00008000U
+#define  IBQULPPARERR      0x00004000U
+#define  IBQSGELOPARERR    0x00002000U
+#define  IBQSGEHIPARERR    0x00001000U
+#define  IBQNCSIPARERR     0x00000800U
+#define  OBQPARERR         0x000007e0U
+#define  OBQULP0PARERR     0x00000400U
+#define  OBQULP1PARERR     0x00000200U
+#define  OBQULP2PARERR     0x00000100U
+#define  OBQULP3PARERR     0x00000080U
+#define  OBQSGEPARERR      0x00000040U
+#define  OBQNCSIPARERR     0x00000020U
+#define  PREFDROPINT       0x00000002U
+#define  UPACCNONZERO      0x00000001U
+
+#define CIM_HOST_UPACC_INT_CAUSE 0x7b34
+#define  EEPROMWRINT      0x40000000U
+#define  TIMEOUTMAINT     0x20000000U
+#define  TIMEOUTINT       0x10000000U
+#define  RSPOVRLOOKUPINT  0x08000000U
+#define  REQOVRLOOKUPINT  0x04000000U
+#define  BLKWRPLINT       0x02000000U
+#define  BLKRDPLINT       0x01000000U
+#define  SGLWRPLINT       0x00800000U
+#define  SGLRDPLINT       0x00400000U
+#define  BLKWRCTLINT      0x00200000U
+#define  BLKRDCTLINT      0x00100000U
+#define  SGLWRCTLINT      0x00080000U
+#define  SGLRDCTLINT      0x00040000U
+#define  BLKWREEPROMINT   0x00020000U
+#define  BLKRDEEPROMINT   0x00010000U
+#define  SGLWREEPROMINT   0x00008000U
+#define  SGLRDEEPROMINT   0x00004000U
+#define  BLKWRFLASHINT    0x00002000U
+#define  BLKRDFLASHINT    0x00001000U
+#define  SGLWRFLASHINT    0x00000800U
+#define  SGLRDFLASHINT    0x00000400U
+#define  BLKWRBOOTINT     0x00000200U
+#define  BLKRDBOOTINT     0x00000100U
+#define  SGLWRBOOTINT     0x00000080U
+#define  SGLRDBOOTINT     0x00000040U
+#define  ILLWRBEINT       0x00000020U
+#define  ILLRDBEINT       0x00000010U
+#define  ILLRDINT         0x00000008U
+#define  ILLWRINT         0x00000004U
+#define  ILLTRANSINT      0x00000002U
+#define  RSVDSPACEINT     0x00000001U
+
+#define TP_OUT_CONFIG 0x7d04
+#define  VLANEXTENABLE_MASK  0x0000f000U
+#define  VLANEXTENABLE_SHIFT 12
+
+#define TP_PARA_REG2 0x7d68
+#define  MAXRXDATA_MASK    0xffff0000U
+#define  MAXRXDATA_SHIFT   16
+#define  MAXRXDATA_GET(x) (((x) & MAXRXDATA_MASK) >> MAXRXDATA_SHIFT)
+
+#define TP_TIMER_RESOLUTION 0x7d90
+#define  TIMERRESOLUTION_MASK   0x00ff0000U
+#define  TIMERRESOLUTION_SHIFT  16
+#define  TIMERRESOLUTION_GET(x) (((x) & TIMERRESOLUTION_MASK) >> TIMERRESOLUTION_SHIFT)
+
+#define TP_SHIFT_CNT 0x7dc0
+
+#define TP_CCTRL_TABLE 0x7ddc
+#define TP_MTU_TABLE 0x7de4
+#define  MTUINDEX_MASK   0xff000000U
+#define  MTUINDEX_SHIFT  24
+#define  MTUINDEX(x)     ((x) << MTUINDEX_SHIFT)
+#define  MTUWIDTH_MASK   0x000f0000U
+#define  MTUWIDTH_SHIFT  16
+#define  MTUWIDTH(x)     ((x) << MTUWIDTH_SHIFT)
+#define  MTUWIDTH_GET(x) (((x) & MTUWIDTH_MASK) >> MTUWIDTH_SHIFT)
+#define  MTUVALUE_MASK   0x00003fffU
+#define  MTUVALUE_SHIFT  0
+#define  MTUVALUE(x)     ((x) << MTUVALUE_SHIFT)
+#define  MTUVALUE_GET(x) (((x) & MTUVALUE_MASK) >> MTUVALUE_SHIFT)
+
+#define TP_RSS_LKP_TABLE 0x7dec
+#define  LKPTBLROWVLD        0x80000000U
+#define  LKPTBLQUEUE1_MASK   0x000ffc00U
+#define  LKPTBLQUEUE1_SHIFT  10
+#define  LKPTBLQUEUE1(x)     ((x) << LKPTBLQUEUE1_SHIFT)
+#define  LKPTBLQUEUE1_GET(x) (((x) & LKPTBLQUEUE1_MASK) >> LKPTBLQUEUE1_SHIFT)
+#define  LKPTBLQUEUE0_MASK   0x000003ffU
+#define  LKPTBLQUEUE0_SHIFT  0
+#define  LKPTBLQUEUE0(x)     ((x) << LKPTBLQUEUE0_SHIFT)
+#define  LKPTBLQUEUE0_GET(x) (((x) & LKPTBLQUEUE0_MASK) >> LKPTBLQUEUE0_SHIFT)
+
+#define TP_PIO_ADDR 0x7e40
+#define TP_PIO_DATA 0x7e44
+#define TP_MIB_INDEX 0x7e50
+#define TP_MIB_DATA 0x7e54
+#define TP_INT_CAUSE 0x7e74
+#define  FLMTXFLSTEMPTY 0x40000000U
+
+#define TP_INGRESS_CONFIG 0x141
+#define  VNIC                0x00000800U
+#define  CSUM_HAS_PSEUDO_HDR 0x00000400U
+#define  RM_OVLAN            0x00000200U
+#define  LOOKUPEVERYPKT      0x00000100U
+
+#define TP_MIB_MAC_IN_ERR_0 0x0
+#define TP_MIB_TCP_OUT_RST 0xc
+#define TP_MIB_TCP_IN_SEG_HI 0x10
+#define TP_MIB_TCP_IN_SEG_LO 0x11
+#define TP_MIB_TCP_OUT_SEG_HI 0x12
+#define TP_MIB_TCP_OUT_SEG_LO 0x13
+#define TP_MIB_TCP_RXT_SEG_HI 0x14
+#define TP_MIB_TCP_RXT_SEG_LO 0x15
+#define TP_MIB_TNL_CNG_DROP_0 0x18
+#define TP_MIB_TCP_V6IN_ERR_0 0x28
+#define TP_MIB_TCP_V6OUT_RST 0x2c
+#define TP_MIB_OFD_ARP_DROP 0x36
+#define TP_MIB_TNL_DROP_0 0x44
+#define TP_MIB_OFD_VLN_DROP_0 0x58
+
+#define ULP_TX_INT_CAUSE 0x8dcc
+#define  PBL_BOUND_ERR_CH3 0x80000000U
+#define  PBL_BOUND_ERR_CH2 0x40000000U
+#define  PBL_BOUND_ERR_CH1 0x20000000U
+#define  PBL_BOUND_ERR_CH0 0x10000000U
+
+#define PM_RX_INT_CAUSE 0x8fdc
+#define  ZERO_E_CMD_ERROR     0x00400000U
+#define  PMRX_FRAMING_ERROR   0x003ffff0U
+#define  OCSPI_PAR_ERROR      0x00000008U
+#define  DB_OPTIONS_PAR_ERROR 0x00000004U
+#define  IESPI_PAR_ERROR      0x00000002U
+#define  E_PCMD_PAR_ERROR     0x00000001U
+
+#define PM_TX_INT_CAUSE 0x8ffc
+#define  PCMD_LEN_OVFL0     0x80000000U
+#define  PCMD_LEN_OVFL1     0x40000000U
+#define  PCMD_LEN_OVFL2     0x20000000U
+#define  ZERO_C_CMD_ERROR   0x10000000U
+#define  PMTX_FRAMING_ERROR 0x0ffffff0U
+#define  OESPI_PAR_ERROR    0x00000008U
+#define  ICSPI_PAR_ERROR    0x00000002U
+#define  C_PCMD_PAR_ERROR   0x00000001U
+
+#define MPS_PORT_STAT_TX_PORT_BYTES_L 0x400
+#define MPS_PORT_STAT_TX_PORT_BYTES_H 0x404
+#define MPS_PORT_STAT_TX_PORT_FRAMES_L 0x408
+#define MPS_PORT_STAT_TX_PORT_FRAMES_H 0x40c
+#define MPS_PORT_STAT_TX_PORT_BCAST_L 0x410
+#define MPS_PORT_STAT_TX_PORT_BCAST_H 0x414
+#define MPS_PORT_STAT_TX_PORT_MCAST_L 0x418
+#define MPS_PORT_STAT_TX_PORT_MCAST_H 0x41c
+#define MPS_PORT_STAT_TX_PORT_UCAST_L 0x420
+#define MPS_PORT_STAT_TX_PORT_UCAST_H 0x424
+#define MPS_PORT_STAT_TX_PORT_ERROR_L 0x428
+#define MPS_PORT_STAT_TX_PORT_ERROR_H 0x42c
+#define MPS_PORT_STAT_TX_PORT_64B_L 0x430
+#define MPS_PORT_STAT_TX_PORT_64B_H 0x434
+#define MPS_PORT_STAT_TX_PORT_65B_127B_L 0x438
+#define MPS_PORT_STAT_TX_PORT_65B_127B_H 0x43c
+#define MPS_PORT_STAT_TX_PORT_128B_255B_L 0x440
+#define MPS_PORT_STAT_TX_PORT_128B_255B_H 0x444
+#define MPS_PORT_STAT_TX_PORT_256B_511B_L 0x448
+#define MPS_PORT_STAT_TX_PORT_256B_511B_H 0x44c
+#define MPS_PORT_STAT_TX_PORT_512B_1023B_L 0x450
+#define MPS_PORT_STAT_TX_PORT_512B_1023B_H 0x454
+#define MPS_PORT_STAT_TX_PORT_1024B_1518B_L 0x458
+#define MPS_PORT_STAT_TX_PORT_1024B_1518B_H 0x45c
+#define MPS_PORT_STAT_TX_PORT_1519B_MAX_L 0x460
+#define MPS_PORT_STAT_TX_PORT_1519B_MAX_H 0x464
+#define MPS_PORT_STAT_TX_PORT_DROP_L 0x468
+#define MPS_PORT_STAT_TX_PORT_DROP_H 0x46c
+#define MPS_PORT_STAT_TX_PORT_PAUSE_L 0x470
+#define MPS_PORT_STAT_TX_PORT_PAUSE_H 0x474
+#define MPS_PORT_STAT_TX_PORT_PPP0_L 0x478
+#define MPS_PORT_STAT_TX_PORT_PPP0_H 0x47c
+#define MPS_PORT_STAT_TX_PORT_PPP1_L 0x480
+#define MPS_PORT_STAT_TX_PORT_PPP1_H 0x484
+#define MPS_PORT_STAT_TX_PORT_PPP2_L 0x488
+#define MPS_PORT_STAT_TX_PORT_PPP2_H 0x48c
+#define MPS_PORT_STAT_TX_PORT_PPP3_L 0x490
+#define MPS_PORT_STAT_TX_PORT_PPP3_H 0x494
+#define MPS_PORT_STAT_TX_PORT_PPP4_L 0x498
+#define MPS_PORT_STAT_TX_PORT_PPP4_H 0x49c
+#define MPS_PORT_STAT_TX_PORT_PPP5_L 0x4a0
+#define MPS_PORT_STAT_TX_PORT_PPP5_H 0x4a4
+#define MPS_PORT_STAT_TX_PORT_PPP6_L 0x4a8
+#define MPS_PORT_STAT_TX_PORT_PPP6_H 0x4ac
+#define MPS_PORT_STAT_TX_PORT_PPP7_L 0x4b0
+#define MPS_PORT_STAT_TX_PORT_PPP7_H 0x4b4
+#define MPS_PORT_STAT_LB_PORT_BYTES_L 0x4c0
+#define MPS_PORT_STAT_LB_PORT_BYTES_H 0x4c4
+#define MPS_PORT_STAT_LB_PORT_FRAMES_L 0x4c8
+#define MPS_PORT_STAT_LB_PORT_FRAMES_H 0x4cc
+#define MPS_PORT_STAT_LB_PORT_BCAST_L 0x4d0
+#define MPS_PORT_STAT_LB_PORT_BCAST_H 0x4d4
+#define MPS_PORT_STAT_LB_PORT_MCAST_L 0x4d8
+#define MPS_PORT_STAT_LB_PORT_MCAST_H 0x4dc
+#define MPS_PORT_STAT_LB_PORT_UCAST_L 0x4e0
+#define MPS_PORT_STAT_LB_PORT_UCAST_H 0x4e4
+#define MPS_PORT_STAT_LB_PORT_ERROR_L 0x4e8
+#define MPS_PORT_STAT_LB_PORT_ERROR_H 0x4ec
+#define MPS_PORT_STAT_LB_PORT_64B_L 0x4f0
+#define MPS_PORT_STAT_LB_PORT_64B_H 0x4f4
+#define MPS_PORT_STAT_LB_PORT_65B_127B_L 0x4f8
+#define MPS_PORT_STAT_LB_PORT_65B_127B_H 0x4fc
+#define MPS_PORT_STAT_LB_PORT_128B_255B_L 0x500
+#define MPS_PORT_STAT_LB_PORT_128B_255B_H 0x504
+#define MPS_PORT_STAT_LB_PORT_256B_511B_L 0x508
+#define MPS_PORT_STAT_LB_PORT_256B_511B_H 0x50c
+#define MPS_PORT_STAT_LB_PORT_512B_1023B_L 0x510
+#define MPS_PORT_STAT_LB_PORT_512B_1023B_H 0x514
+#define MPS_PORT_STAT_LB_PORT_1024B_1518B_L 0x518
+#define MPS_PORT_STAT_LB_PORT_1024B_1518B_H 0x51c
+#define MPS_PORT_STAT_LB_PORT_1519B_MAX_L 0x520
+#define MPS_PORT_STAT_LB_PORT_1519B_MAX_H 0x524
+#define MPS_PORT_STAT_LB_PORT_DROP_FRAMES 0x528
+#define MPS_PORT_STAT_RX_PORT_BYTES_L 0x540
+#define MPS_PORT_STAT_RX_PORT_BYTES_H 0x544
+#define MPS_PORT_STAT_RX_PORT_FRAMES_L 0x548
+#define MPS_PORT_STAT_RX_PORT_FRAMES_H 0x54c
+#define MPS_PORT_STAT_RX_PORT_BCAST_L 0x550
+#define MPS_PORT_STAT_RX_PORT_BCAST_H 0x554
+#define MPS_PORT_STAT_RX_PORT_MCAST_L 0x558
+#define MPS_PORT_STAT_RX_PORT_MCAST_H 0x55c
+#define MPS_PORT_STAT_RX_PORT_UCAST_L 0x560
+#define MPS_PORT_STAT_RX_PORT_UCAST_H 0x564
+#define MPS_PORT_STAT_RX_PORT_MTU_ERROR_L 0x568
+#define MPS_PORT_STAT_RX_PORT_MTU_ERROR_H 0x56c
+#define MPS_PORT_STAT_RX_PORT_MTU_CRC_ERROR_L 0x570
+#define MPS_PORT_STAT_RX_PORT_MTU_CRC_ERROR_H 0x574
+#define MPS_PORT_STAT_RX_PORT_CRC_ERROR_L 0x578
+#define MPS_PORT_STAT_RX_PORT_CRC_ERROR_H 0x57c
+#define MPS_PORT_STAT_RX_PORT_LEN_ERROR_L 0x580
+#define MPS_PORT_STAT_RX_PORT_LEN_ERROR_H 0x584
+#define MPS_PORT_STAT_RX_PORT_SYM_ERROR_L 0x588
+#define MPS_PORT_STAT_RX_PORT_SYM_ERROR_H 0x58c
+#define MPS_PORT_STAT_RX_PORT_64B_L 0x590
+#define MPS_PORT_STAT_RX_PORT_64B_H 0x594
+#define MPS_PORT_STAT_RX_PORT_65B_127B_L 0x598
+#define MPS_PORT_STAT_RX_PORT_65B_127B_H 0x59c
+#define MPS_PORT_STAT_RX_PORT_128B_255B_L 0x5a0
+#define MPS_PORT_STAT_RX_PORT_128B_255B_H 0x5a4
+#define MPS_PORT_STAT_RX_PORT_256B_511B_L 0x5a8
+#define MPS_PORT_STAT_RX_PORT_256B_511B_H 0x5ac
+#define MPS_PORT_STAT_RX_PORT_512B_1023B_L 0x5b0
+#define MPS_PORT_STAT_RX_PORT_512B_1023B_H 0x5b4
+#define MPS_PORT_STAT_RX_PORT_1024B_1518B_L 0x5b8
+#define MPS_PORT_STAT_RX_PORT_1024B_1518B_H 0x5bc
+#define MPS_PORT_STAT_RX_PORT_1519B_MAX_L 0x5c0
+#define MPS_PORT_STAT_RX_PORT_1519B_MAX_H 0x5c4
+#define MPS_PORT_STAT_RX_PORT_PAUSE_L 0x5c8
+#define MPS_PORT_STAT_RX_PORT_PAUSE_H 0x5cc
+#define MPS_PORT_STAT_RX_PORT_PPP0_L 0x5d0
+#define MPS_PORT_STAT_RX_PORT_PPP0_H 0x5d4
+#define MPS_PORT_STAT_RX_PORT_PPP1_L 0x5d8
+#define MPS_PORT_STAT_RX_PORT_PPP1_H 0x5dc
+#define MPS_PORT_STAT_RX_PORT_PPP2_L 0x5e0
+#define MPS_PORT_STAT_RX_PORT_PPP2_H 0x5e4
+#define MPS_PORT_STAT_RX_PORT_PPP3_L 0x5e8
+#define MPS_PORT_STAT_RX_PORT_PPP3_H 0x5ec
+#define MPS_PORT_STAT_RX_PORT_PPP4_L 0x5f0
+#define MPS_PORT_STAT_RX_PORT_PPP4_H 0x5f4
+#define MPS_PORT_STAT_RX_PORT_PPP5_L 0x5f8
+#define MPS_PORT_STAT_RX_PORT_PPP5_H 0x5fc
+#define MPS_PORT_STAT_RX_PORT_PPP6_L 0x600
+#define MPS_PORT_STAT_RX_PORT_PPP6_H 0x604
+#define MPS_PORT_STAT_RX_PORT_PPP7_L 0x608
+#define MPS_PORT_STAT_RX_PORT_PPP7_H 0x60c
+#define MPS_PORT_STAT_RX_PORT_LESS_64B_L 0x610
+#define MPS_PORT_STAT_RX_PORT_LESS_64B_H 0x614
+#define MPS_CMN_CTL 0x9000
+#define  NUMPORTS_MASK   0x00000003U
+#define  NUMPORTS_SHIFT  0
+#define  NUMPORTS_GET(x) (((x) & NUMPORTS_MASK) >> NUMPORTS_SHIFT)
+
+#define MPS_INT_CAUSE 0x9008
+#define  STATINT 0x00000020U
+#define  TXINT   0x00000010U
+#define  RXINT   0x00000008U
+#define  TRCINT  0x00000004U
+#define  CLSINT  0x00000002U
+#define  PLINT   0x00000001U
+
+#define MPS_TX_INT_CAUSE 0x9408
+#define  PORTERR    0x00010000U
+#define  FRMERR     0x00008000U
+#define  SECNTERR   0x00004000U
+#define  BUBBLE     0x00002000U
+#define  TXDESCFIFO 0x00001e00U
+#define  TXDATAFIFO 0x000001e0U
+#define  NCSIFIFO   0x00000010U
+#define  TPFIFO     0x0000000fU
+
+#define MPS_STAT_PERR_INT_CAUSE_SRAM 0x9614
+#define MPS_STAT_PERR_INT_CAUSE_TX_FIFO 0x9620
+#define MPS_STAT_PERR_INT_CAUSE_RX_FIFO 0x962c
+
+#define MPS_STAT_RX_BG_0_MAC_DROP_FRAME_L 0x9640
+#define MPS_STAT_RX_BG_0_MAC_DROP_FRAME_H 0x9644
+#define MPS_STAT_RX_BG_1_MAC_DROP_FRAME_L 0x9648
+#define MPS_STAT_RX_BG_1_MAC_DROP_FRAME_H 0x964c
+#define MPS_STAT_RX_BG_2_MAC_DROP_FRAME_L 0x9650
+#define MPS_STAT_RX_BG_2_MAC_DROP_FRAME_H 0x9654
+#define MPS_STAT_RX_BG_3_MAC_DROP_FRAME_L 0x9658
+#define MPS_STAT_RX_BG_3_MAC_DROP_FRAME_H 0x965c
+#define MPS_STAT_RX_BG_0_LB_DROP_FRAME_L 0x9660
+#define MPS_STAT_RX_BG_0_LB_DROP_FRAME_H 0x9664
+#define MPS_STAT_RX_BG_1_LB_DROP_FRAME_L 0x9668
+#define MPS_STAT_RX_BG_1_LB_DROP_FRAME_H 0x966c
+#define MPS_STAT_RX_BG_2_LB_DROP_FRAME_L 0x9670
+#define MPS_STAT_RX_BG_2_LB_DROP_FRAME_H 0x9674
+#define MPS_STAT_RX_BG_3_LB_DROP_FRAME_L 0x9678
+#define MPS_STAT_RX_BG_3_LB_DROP_FRAME_H 0x967c
+#define MPS_STAT_RX_BG_0_MAC_TRUNC_FRAME_L 0x9680
+#define MPS_STAT_RX_BG_0_MAC_TRUNC_FRAME_H 0x9684
+#define MPS_STAT_RX_BG_1_MAC_TRUNC_FRAME_L 0x9688
+#define MPS_STAT_RX_BG_1_MAC_TRUNC_FRAME_H 0x968c
+#define MPS_STAT_RX_BG_2_MAC_TRUNC_FRAME_L 0x9690
+#define MPS_STAT_RX_BG_2_MAC_TRUNC_FRAME_H 0x9694
+#define MPS_STAT_RX_BG_3_MAC_TRUNC_FRAME_L 0x9698
+#define MPS_STAT_RX_BG_3_MAC_TRUNC_FRAME_H 0x969c
+#define MPS_STAT_RX_BG_0_LB_TRUNC_FRAME_L 0x96a0
+#define MPS_STAT_RX_BG_0_LB_TRUNC_FRAME_H 0x96a4
+#define MPS_STAT_RX_BG_1_LB_TRUNC_FRAME_L 0x96a8
+#define MPS_STAT_RX_BG_1_LB_TRUNC_FRAME_H 0x96ac
+#define MPS_STAT_RX_BG_2_LB_TRUNC_FRAME_L 0x96b0
+#define MPS_STAT_RX_BG_2_LB_TRUNC_FRAME_H 0x96b4
+#define MPS_STAT_RX_BG_3_LB_TRUNC_FRAME_L 0x96b8
+#define MPS_STAT_RX_BG_3_LB_TRUNC_FRAME_H 0x96bc
+#define MPS_TRC_CFG 0x9800
+#define  TRCFIFOEMPTY       0x00000010U
+#define  TRCIGNOREDROPINPUT 0x00000008U
+#define  TRCKEEPDUPLICATES  0x00000004U
+#define  TRCEN              0x00000002U
+#define  TRCMULTIFILTER     0x00000001U
+
+#define MPS_TRC_RSS_CONTROL 0x9808
+#define  RSSCONTROL_MASK    0x00ff0000U
+#define  RSSCONTROL_SHIFT   16
+#define  RSSCONTROL(x)      ((x) << RSSCONTROL_SHIFT)
+#define  QUEUENUMBER_MASK   0x0000ffffU
+#define  QUEUENUMBER_SHIFT  0
+#define  QUEUENUMBER(x)     ((x) << QUEUENUMBER_SHIFT)
+
+#define MPS_TRC_FILTER_MATCH_CTL_A 0x9810
+#define  TFINVERTMATCH   0x01000000U
+#define  TFPKTTOOLARGE   0x00800000U
+#define  TFEN            0x00400000U
+#define  TFPORT_MASK     0x003c0000U
+#define  TFPORT_SHIFT    18
+#define  TFPORT(x)       ((x) << TFPORT_SHIFT)
+#define  TFPORT_GET(x)   (((x) & TFPORT_MASK) >> TFPORT_SHIFT)
+#define  TFDROP          0x00020000U
+#define  TFSOPEOPERR     0x00010000U
+#define  TFLENGTH_MASK   0x00001f00U
+#define  TFLENGTH_SHIFT  8
+#define  TFLENGTH(x)     ((x) << TFLENGTH_SHIFT)
+#define  TFLENGTH_GET(x) (((x) & TFLENGTH_MASK) >> TFLENGTH_SHIFT)
+#define  TFOFFSET_MASK   0x0000001fU
+#define  TFOFFSET_SHIFT  0
+#define  TFOFFSET(x)     ((x) << TFOFFSET_SHIFT)
+#define  TFOFFSET_GET(x) (((x) & TFOFFSET_MASK) >> TFOFFSET_SHIFT)
+
+#define MPS_TRC_FILTER_MATCH_CTL_B 0x9820
+#define  TFMINPKTSIZE_MASK   0x01ff0000U
+#define  TFMINPKTSIZE_SHIFT  16
+#define  TFMINPKTSIZE(x)     ((x) << TFMINPKTSIZE_SHIFT)
+#define  TFMINPKTSIZE_GET(x) (((x) & TFMINPKTSIZE_MASK) >> TFMINPKTSIZE_SHIFT)
+#define  TFCAPTUREMAX_MASK   0x00003fffU
+#define  TFCAPTUREMAX_SHIFT  0
+#define  TFCAPTUREMAX(x)     ((x) << TFCAPTUREMAX_SHIFT)
+#define  TFCAPTUREMAX_GET(x) (((x) & TFCAPTUREMAX_MASK) >> TFCAPTUREMAX_SHIFT)
+
+#define MPS_TRC_INT_CAUSE 0x985c
+#define  MISCPERR 0x00000100U
+#define  PKTFIFO  0x000000f0U
+#define  FILTMEM  0x0000000fU
+
+#define MPS_TRC_FILTER0_MATCH 0x9c00
+#define MPS_TRC_FILTER0_DONT_CARE 0x9c80
+#define MPS_TRC_FILTER1_MATCH 0x9d00
+#define MPS_CLS_INT_CAUSE 0xd028
+#define  PLERRENB  0x00000008U
+#define  HASHSRAM  0x00000004U
+#define  MATCHTCAM 0x00000002U
+#define  MATCHSRAM 0x00000001U
+
+#define MPS_RX_PERR_INT_CAUSE 0x11074
+
+#define CPL_INTR_CAUSE 0x19054
+#define  CIM_OP_MAP_PERR   0x00000020U
+#define  CIM_OVFL_ERROR    0x00000010U
+#define  TP_FRAMING_ERROR  0x00000008U
+#define  SGE_FRAMING_ERROR 0x00000004U
+#define  CIM_FRAMING_ERROR 0x00000002U
+#define  ZERO_SWITCH_ERROR 0x00000001U
+
+#define SMB_INT_CAUSE 0x19090
+#define  MSTTXFIFOPARINT 0x00200000U
+#define  MSTRXFIFOPARINT 0x00100000U
+#define  SLVFIFOPARINT   0x00080000U
+
+#define ULP_RX_INT_CAUSE 0x19158
+#define ULP_RX_ISCSI_TAGMASK 0x19164
+#define ULP_RX_ISCSI_PSZ 0x19168
+#define  HPZ3_MASK   0x0f000000U
+#define  HPZ3_SHIFT  24
+#define  HPZ3(x)     ((x) << HPZ3_SHIFT)
+#define  HPZ2_MASK   0x000f0000U
+#define  HPZ2_SHIFT  16
+#define  HPZ2(x)     ((x) << HPZ2_SHIFT)
+#define  HPZ1_MASK   0x00000f00U
+#define  HPZ1_SHIFT  8
+#define  HPZ1(x)     ((x) << HPZ1_SHIFT)
+#define  HPZ0_MASK   0x0000000fU
+#define  HPZ0_SHIFT  0
+#define  HPZ0(x)     ((x) << HPZ0_SHIFT)
+
+#define ULP_RX_TDDP_PSZ 0x19178
+
+#define SF_DATA 0x193f8
+#define SF_OP 0x193fc
+#define  BUSY          0x80000000U
+#define  SF_LOCK       0x00000010U
+#define  SF_CONT       0x00000008U
+#define  BYTECNT_MASK  0x00000006U
+#define  BYTECNT_SHIFT 1
+#define  BYTECNT(x)    ((x) << BYTECNT_SHIFT)
+#define  OP_WR         0x00000001U
+
+#define PL_PF_INT_CAUSE 0x3c0
+#define  PFSW  0x00000008U
+#define  PFSGE 0x00000004U
+#define  PFCIM 0x00000002U
+#define  PFMPS 0x00000001U
+
+#define PL_PF_INT_ENABLE 0x3c4
+#define PL_PF_CTL 0x3c8
+#define  SWINT 0x00000001U
+
+#define PL_WHOAMI 0x19400
+#define  SOURCEPF_MASK   0x00000700U
+#define  SOURCEPF_SHIFT  8
+#define  SOURCEPF(x)     ((x) << SOURCEPF_SHIFT)
+#define  SOURCEPF_GET(x) (((x) & SOURCEPF_MASK) >> SOURCEPF_SHIFT)
+#define  ISVF            0x00000080U
+#define  VFID_MASK       0x0000007fU
+#define  VFID_SHIFT      0
+#define  VFID(x)         ((x) << VFID_SHIFT)
+#define  VFID_GET(x)     (((x) & VFID_MASK) >> VFID_SHIFT)
+
+#define PL_INT_CAUSE 0x1940c
+#define  ULP_TX     0x08000000U
+#define  SGE        0x04000000U
+#define  HMA        0x02000000U
+#define  CPL_SWITCH 0x01000000U
+#define  ULP_RX     0x00800000U
+#define  PM_RX      0x00400000U
+#define  PM_TX      0x00200000U
+#define  MA         0x00100000U
+#define  TP         0x00080000U
+#define  LE         0x00040000U
+#define  EDC1       0x00020000U
+#define  EDC0       0x00010000U
+#define  MC         0x00008000U
+#define  PCIE       0x00004000U
+#define  PMU        0x00002000U
+#define  XGMAC_KR1  0x00001000U
+#define  XGMAC_KR0  0x00000800U
+#define  XGMAC1     0x00000400U
+#define  XGMAC0     0x00000200U
+#define  SMB        0x00000100U
+#define  SF         0x00000080U
+#define  PL         0x00000040U
+#define  NCSI       0x00000020U
+#define  MPS        0x00000010U
+#define  MI         0x00000008U
+#define  DBG        0x00000004U
+#define  I2CM       0x00000002U
+#define  CIM        0x00000001U
+
+#define PL_INT_MAP0 0x19414
+#define PL_RST 0x19428
+#define  PIORST     0x00000002U
+#define  PIORSTMODE 0x00000001U
+
+#define PL_PL_INT_CAUSE 0x19430
+#define  FATALPERR 0x00000010U
+#define  PERRVFID  0x00000001U
+
+#define PL_REV 0x1943c
+
+#define LE_DB_CONFIG 0x19c04
+#define  HASHEN 0x00100000U
+
+#define LE_DB_SERVER_INDEX 0x19c18
+#define LE_DB_ACT_CNT_IPV4 0x19c20
+#define LE_DB_ACT_CNT_IPV6 0x19c24
+
+#define LE_DB_INT_CAUSE 0x19c3c
+#define  REQQPARERR 0x00010000U
+#define  UNKNOWNCMD 0x00008000U
+#define  PARITYERR  0x00000040U
+#define  LIPMISS    0x00000020U
+#define  LIP0       0x00000010U
+
+#define LE_DB_TID_HASHBASE 0x19df8
+
+#define NCSI_INT_CAUSE 0x1a0d8
+#define  CIM_DM_PRTY_ERR 0x00000100U
+#define  MPS_DM_PRTY_ERR 0x00000080U
+#define  TXFIFO_PRTY_ERR 0x00000002U
+#define  RXFIFO_PRTY_ERR 0x00000001U
+
+#define XGMAC_PORT_CFG2 0x1018
+#define  PATEN   0x00040000U
+#define  MAGICEN 0x00020000U
+
+#define XGMAC_PORT_MAGIC_MACID_LO 0x1024
+#define XGMAC_PORT_MAGIC_MACID_HI 0x1028
+
+#define XGMAC_PORT_EPIO_DATA0 0x10c0
+#define XGMAC_PORT_EPIO_DATA1 0x10c4
+#define XGMAC_PORT_EPIO_DATA2 0x10c8
+#define XGMAC_PORT_EPIO_DATA3 0x10cc
+#define XGMAC_PORT_EPIO_OP 0x10d0
+#define  EPIOWR         0x00000100U
+#define  ADDRESS_MASK   0x000000ffU
+#define  ADDRESS_SHIFT  0
+#define  ADDRESS(x)     ((x) << ADDRESS_SHIFT)
+
+#define XGMAC_PORT_INT_CAUSE 0x10dc
+#endif /* __T4_REGS_H */
diff --git a/drivers/net/cxgb4/t4fw_api.h b/drivers/net/cxgb4/t4fw_api.h
new file mode 100644
index 000000000000..3393d05a388a
--- /dev/null
+++ b/drivers/net/cxgb4/t4fw_api.h
@@ -0,0 +1,1580 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2009-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _T4FW_INTERFACE_H_
+#define _T4FW_INTERFACE_H_
+
+#define FW_T4VF_SGE_BASE_ADDR      0x0000
+#define FW_T4VF_MPS_BASE_ADDR      0x0100
+#define FW_T4VF_PL_BASE_ADDR       0x0200
+#define FW_T4VF_MBDATA_BASE_ADDR   0x0240
+#define FW_T4VF_CIM_BASE_ADDR      0x0300
+
+enum fw_wr_opcodes {
+	FW_FILTER_WR                   = 0x02,
+	FW_ULPTX_WR                    = 0x04,
+	FW_TP_WR                       = 0x05,
+	FW_ETH_TX_PKT_WR               = 0x08,
+	FW_FLOWC_WR                    = 0x0a,
+	FW_OFLD_TX_DATA_WR             = 0x0b,
+	FW_CMD_WR                      = 0x10,
+	FW_ETH_TX_PKT_VM_WR            = 0x11,
+	FW_RI_RES_WR                   = 0x0c,
+	FW_RI_INIT_WR                  = 0x0d,
+	FW_RI_RDMA_WRITE_WR            = 0x14,
+	FW_RI_SEND_WR                  = 0x15,
+	FW_RI_RDMA_READ_WR             = 0x16,
+	FW_RI_RECV_WR                  = 0x17,
+	FW_RI_BIND_MW_WR               = 0x18,
+	FW_RI_FR_NSMR_WR               = 0x19,
+	FW_RI_INV_LSTAG_WR             = 0x1a,
+	FW_LASTC2E_WR                  = 0x40
+};
+
+struct fw_wr_hdr {
+	__be32 hi;
+	__be32 lo;
+};
+
+#define FW_WR_OP(x)	 ((x) << 24)
+#define FW_WR_ATOMIC(x)	 ((x) << 23)
+#define FW_WR_FLUSH(x)   ((x) << 22)
+#define FW_WR_COMPL(x)   ((x) << 21)
+#define FW_WR_IMMDLEN(x) ((x) << 0)
+
+#define FW_WR_EQUIQ	(1U << 31)
+#define FW_WR_EQUEQ	(1U << 30)
+#define FW_WR_FLOWID(x)	((x) << 8)
+#define FW_WR_LEN16(x)	((x) << 0)
+
+struct fw_ulptx_wr {
+	__be32 op_to_compl;
+	__be32 flowid_len16;
+	u64 cookie;
+};
+
+struct fw_tp_wr {
+	__be32 op_to_immdlen;
+	__be32 flowid_len16;
+	u64 cookie;
+};
+
+struct fw_eth_tx_pkt_wr {
+	__be32 op_immdlen;
+	__be32 equiq_to_len16;
+	__be64 r3;
+};
+
+enum fw_flowc_mnem {
+	FW_FLOWC_MNEM_PFNVFN,		/* PFN [15:8] VFN [7:0] */
+	FW_FLOWC_MNEM_CH,
+	FW_FLOWC_MNEM_PORT,
+	FW_FLOWC_MNEM_IQID,
+	FW_FLOWC_MNEM_SNDNXT,
+	FW_FLOWC_MNEM_RCVNXT,
+	FW_FLOWC_MNEM_SNDBUF,
+	FW_FLOWC_MNEM_MSS,
+};
+
+struct fw_flowc_mnemval {
+	u8 mnemonic;
+	u8 r4[3];
+	__be32 val;
+};
+
+struct fw_flowc_wr {
+	__be32 op_to_nparams;
+#define FW_FLOWC_WR_NPARAMS(x)	((x) << 0)
+	__be32 flowid_len16;
+	struct fw_flowc_mnemval mnemval[0];
+};
+
+struct fw_ofld_tx_data_wr {
+	__be32 op_to_immdlen;
+	__be32 flowid_len16;
+	__be32 plen;
+	__be32 tunnel_to_proxy;
+#define FW_OFLD_TX_DATA_WR_TUNNEL(x)	 ((x) << 19)
+#define FW_OFLD_TX_DATA_WR_SAVE(x)	 ((x) << 18)
+#define FW_OFLD_TX_DATA_WR_FLUSH(x)	 ((x) << 17)
+#define FW_OFLD_TX_DATA_WR_URGENT(x)	 ((x) << 16)
+#define FW_OFLD_TX_DATA_WR_MORE(x)	 ((x) << 15)
+#define FW_OFLD_TX_DATA_WR_SHOVE(x)	 ((x) << 14)
+#define FW_OFLD_TX_DATA_WR_ULPMODE(x)	 ((x) << 10)
+#define FW_OFLD_TX_DATA_WR_ULPSUBMODE(x) ((x) << 6)
+};
+
+struct fw_cmd_wr {
+	__be32 op_dma;
+#define FW_CMD_WR_DMA (1U << 17)
+	__be32 len16_pkd;
+	__be64 cookie_daddr;
+};
+
+struct fw_eth_tx_pkt_vm_wr {
+	__be32 op_immdlen;
+	__be32 equiq_to_len16;
+	__be32 r3[2];
+	u8 ethmacdst[6];
+	u8 ethmacsrc[6];
+	__be16 ethtype;
+	__be16 vlantci;
+};
+
+#define FW_CMD_MAX_TIMEOUT 3000
+
+enum fw_cmd_opcodes {
+	FW_LDST_CMD                    = 0x01,
+	FW_RESET_CMD                   = 0x03,
+	FW_HELLO_CMD                   = 0x04,
+	FW_BYE_CMD                     = 0x05,
+	FW_INITIALIZE_CMD              = 0x06,
+	FW_CAPS_CONFIG_CMD             = 0x07,
+	FW_PARAMS_CMD                  = 0x08,
+	FW_PFVF_CMD                    = 0x09,
+	FW_IQ_CMD                      = 0x10,
+	FW_EQ_MNGT_CMD                 = 0x11,
+	FW_EQ_ETH_CMD                  = 0x12,
+	FW_EQ_CTRL_CMD                 = 0x13,
+	FW_EQ_OFLD_CMD                 = 0x21,
+	FW_VI_CMD                      = 0x14,
+	FW_VI_MAC_CMD                  = 0x15,
+	FW_VI_RXMODE_CMD               = 0x16,
+	FW_VI_ENABLE_CMD               = 0x17,
+	FW_ACL_MAC_CMD                 = 0x18,
+	FW_ACL_VLAN_CMD                = 0x19,
+	FW_VI_STATS_CMD                = 0x1a,
+	FW_PORT_CMD                    = 0x1b,
+	FW_PORT_STATS_CMD              = 0x1c,
+	FW_PORT_LB_STATS_CMD           = 0x1d,
+	FW_PORT_TRACE_CMD              = 0x1e,
+	FW_PORT_TRACE_MMAP_CMD         = 0x1f,
+	FW_RSS_IND_TBL_CMD             = 0x20,
+	FW_RSS_GLB_CONFIG_CMD          = 0x22,
+	FW_RSS_VI_CONFIG_CMD           = 0x23,
+	FW_LASTC2E_CMD                 = 0x40,
+	FW_ERROR_CMD                   = 0x80,
+	FW_DEBUG_CMD                   = 0x81,
+};
+
+enum fw_cmd_cap {
+	FW_CMD_CAP_PF                  = 0x01,
+	FW_CMD_CAP_DMAQ                = 0x02,
+	FW_CMD_CAP_PORT                = 0x04,
+	FW_CMD_CAP_PORTPROMISC         = 0x08,
+	FW_CMD_CAP_PORTSTATS           = 0x10,
+	FW_CMD_CAP_VF                  = 0x80,
+};
+
+/*
+ * Generic command header flit0
+ */
+struct fw_cmd_hdr {
+	__be32 hi;
+	__be32 lo;
+};
+
+#define FW_CMD_OP(x)		((x) << 24)
+#define FW_CMD_OP_GET(x)        (((x) >> 24) & 0xff)
+#define FW_CMD_REQUEST          (1U << 23)
+#define FW_CMD_READ		(1U << 22)
+#define FW_CMD_WRITE		(1U << 21)
+#define FW_CMD_EXEC		(1U << 20)
+#define FW_CMD_RAMASK(x)	((x) << 20)
+#define FW_CMD_RETVAL(x)	((x) << 8)
+#define FW_CMD_RETVAL_GET(x)	(((x) >> 8) & 0xff)
+#define FW_CMD_LEN16(x)         ((x) << 0)
+
+enum fw_ldst_addrspc {
+	FW_LDST_ADDRSPC_FIRMWARE  = 0x0001,
+	FW_LDST_ADDRSPC_SGE_EGRC  = 0x0008,
+	FW_LDST_ADDRSPC_SGE_INGC  = 0x0009,
+	FW_LDST_ADDRSPC_SGE_FLMC  = 0x000a,
+	FW_LDST_ADDRSPC_SGE_CONMC = 0x000b,
+	FW_LDST_ADDRSPC_TP_PIO    = 0x0010,
+	FW_LDST_ADDRSPC_TP_TM_PIO = 0x0011,
+	FW_LDST_ADDRSPC_TP_MIB    = 0x0012,
+	FW_LDST_ADDRSPC_MDIO      = 0x0018,
+	FW_LDST_ADDRSPC_MPS       = 0x0020,
+	FW_LDST_ADDRSPC_FUNC      = 0x0028
+};
+
+enum fw_ldst_mps_fid {
+	FW_LDST_MPS_ATRB,
+	FW_LDST_MPS_RPLC
+};
+
+enum fw_ldst_func_access_ctl {
+	FW_LDST_FUNC_ACC_CTL_VIID,
+	FW_LDST_FUNC_ACC_CTL_FID
+};
+
+enum fw_ldst_func_mod_index {
+	FW_LDST_FUNC_MPS
+};
+
+struct fw_ldst_cmd {
+	__be32 op_to_addrspace;
+#define FW_LDST_CMD_ADDRSPACE(x) ((x) << 0)
+	__be32 cycles_to_len16;
+	union fw_ldst {
+		struct fw_ldst_addrval {
+			__be32 addr;
+			__be32 val;
+		} addrval;
+		struct fw_ldst_idctxt {
+			__be32 physid;
+			__be32 msg_pkd;
+			__be32 ctxt_data7;
+			__be32 ctxt_data6;
+			__be32 ctxt_data5;
+			__be32 ctxt_data4;
+			__be32 ctxt_data3;
+			__be32 ctxt_data2;
+			__be32 ctxt_data1;
+			__be32 ctxt_data0;
+		} idctxt;
+		struct fw_ldst_mdio {
+			__be16 paddr_mmd;
+			__be16 raddr;
+			__be16 vctl;
+			__be16 rval;
+		} mdio;
+		struct fw_ldst_mps {
+			__be16 fid_ctl;
+			__be16 rplcpf_pkd;
+			__be32 rplc127_96;
+			__be32 rplc95_64;
+			__be32 rplc63_32;
+			__be32 rplc31_0;
+			__be32 atrb;
+			__be16 vlan[16];
+		} mps;
+		struct fw_ldst_func {
+			u8 access_ctl;
+			u8 mod_index;
+			__be16 ctl_id;
+			__be32 offset;
+			__be64 data0;
+			__be64 data1;
+		} func;
+	} u;
+};
+
+#define FW_LDST_CMD_MSG(x)	((x) << 31)
+#define FW_LDST_CMD_PADDR(x)	((x) << 8)
+#define FW_LDST_CMD_MMD(x)	((x) << 0)
+#define FW_LDST_CMD_FID(x)	((x) << 15)
+#define FW_LDST_CMD_CTL(x)	((x) << 0)
+#define FW_LDST_CMD_RPLCPF(x)	((x) << 0)
+
+struct fw_reset_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	__be32 val;
+	__be32 r3;
+};
+
+struct fw_hello_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	__be32 err_to_mbasyncnot;
+#define FW_HELLO_CMD_ERR	    (1U << 31)
+#define FW_HELLO_CMD_INIT	    (1U << 30)
+#define FW_HELLO_CMD_MASTERDIS(x)   ((x) << 29)
+#define FW_HELLO_CMD_MASTERFORCE(x) ((x) << 28)
+#define FW_HELLO_CMD_MBMASTER(x)    ((x) << 24)
+#define FW_HELLO_CMD_MBASYNCNOT(x)  ((x) << 20)
+	__be32 fwrev;
+};
+
+struct fw_bye_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	__be64 r3;
+};
+
+struct fw_initialize_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	__be64 r3;
+};
+
+enum fw_caps_config_hm {
+	FW_CAPS_CONFIG_HM_PCIE		= 0x00000001,
+	FW_CAPS_CONFIG_HM_PL		= 0x00000002,
+	FW_CAPS_CONFIG_HM_SGE		= 0x00000004,
+	FW_CAPS_CONFIG_HM_CIM		= 0x00000008,
+	FW_CAPS_CONFIG_HM_ULPTX		= 0x00000010,
+	FW_CAPS_CONFIG_HM_TP		= 0x00000020,
+	FW_CAPS_CONFIG_HM_ULPRX		= 0x00000040,
+	FW_CAPS_CONFIG_HM_PMRX		= 0x00000080,
+	FW_CAPS_CONFIG_HM_PMTX		= 0x00000100,
+	FW_CAPS_CONFIG_HM_MC		= 0x00000200,
+	FW_CAPS_CONFIG_HM_LE		= 0x00000400,
+	FW_CAPS_CONFIG_HM_MPS		= 0x00000800,
+	FW_CAPS_CONFIG_HM_XGMAC		= 0x00001000,
+	FW_CAPS_CONFIG_HM_CPLSWITCH	= 0x00002000,
+	FW_CAPS_CONFIG_HM_T4DBG		= 0x00004000,
+	FW_CAPS_CONFIG_HM_MI		= 0x00008000,
+	FW_CAPS_CONFIG_HM_I2CM		= 0x00010000,
+	FW_CAPS_CONFIG_HM_NCSI		= 0x00020000,
+	FW_CAPS_CONFIG_HM_SMB		= 0x00040000,
+	FW_CAPS_CONFIG_HM_MA		= 0x00080000,
+	FW_CAPS_CONFIG_HM_EDRAM		= 0x00100000,
+	FW_CAPS_CONFIG_HM_PMU		= 0x00200000,
+	FW_CAPS_CONFIG_HM_UART		= 0x00400000,
+	FW_CAPS_CONFIG_HM_SF		= 0x00800000,
+};
+
+enum fw_caps_config_nbm {
+	FW_CAPS_CONFIG_NBM_IPMI		= 0x00000001,
+	FW_CAPS_CONFIG_NBM_NCSI		= 0x00000002,
+};
+
+enum fw_caps_config_link {
+	FW_CAPS_CONFIG_LINK_PPP		= 0x00000001,
+	FW_CAPS_CONFIG_LINK_QFC		= 0x00000002,
+	FW_CAPS_CONFIG_LINK_DCBX	= 0x00000004,
+};
+
+enum fw_caps_config_switch {
+	FW_CAPS_CONFIG_SWITCH_INGRESS	= 0x00000001,
+	FW_CAPS_CONFIG_SWITCH_EGRESS	= 0x00000002,
+};
+
+enum fw_caps_config_nic {
+	FW_CAPS_CONFIG_NIC		= 0x00000001,
+	FW_CAPS_CONFIG_NIC_VM		= 0x00000002,
+};
+
+enum fw_caps_config_ofld {
+	FW_CAPS_CONFIG_OFLD		= 0x00000001,
+};
+
+enum fw_caps_config_rdma {
+	FW_CAPS_CONFIG_RDMA_RDDP	= 0x00000001,
+	FW_CAPS_CONFIG_RDMA_RDMAC	= 0x00000002,
+};
+
+enum fw_caps_config_iscsi {
+	FW_CAPS_CONFIG_ISCSI_INITIATOR_PDU = 0x00000001,
+	FW_CAPS_CONFIG_ISCSI_TARGET_PDU = 0x00000002,
+	FW_CAPS_CONFIG_ISCSI_INITIATOR_CNXOFLD = 0x00000004,
+	FW_CAPS_CONFIG_ISCSI_TARGET_CNXOFLD = 0x00000008,
+};
+
+enum fw_caps_config_fcoe {
+	FW_CAPS_CONFIG_FCOE_INITIATOR	= 0x00000001,
+	FW_CAPS_CONFIG_FCOE_TARGET	= 0x00000002,
+};
+
+struct fw_caps_config_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	__be32 r2;
+	__be32 hwmbitmap;
+	__be16 nbmcaps;
+	__be16 linkcaps;
+	__be16 switchcaps;
+	__be16 r3;
+	__be16 niccaps;
+	__be16 ofldcaps;
+	__be16 rdmacaps;
+	__be16 r4;
+	__be16 iscsicaps;
+	__be16 fcoecaps;
+	__be32 r5;
+	__be64 r6;
+};
+
+/*
+ * params command mnemonics
+ */
+enum fw_params_mnem {
+	FW_PARAMS_MNEM_DEV		= 1,	/* device params */
+	FW_PARAMS_MNEM_PFVF		= 2,	/* function params */
+	FW_PARAMS_MNEM_REG		= 3,	/* limited register access */
+	FW_PARAMS_MNEM_DMAQ		= 4,	/* dma queue params */
+	FW_PARAMS_MNEM_LAST
+};
+
+/*
+ * device parameters
+ */
+enum fw_params_param_dev {
+	FW_PARAMS_PARAM_DEV_CCLK	= 0x00, /* chip core clock in khz */
+	FW_PARAMS_PARAM_DEV_PORTVEC	= 0x01, /* the port vector */
+	FW_PARAMS_PARAM_DEV_NTID	= 0x02, /* reads the number of TIDs
+						 * allocated by the device's
+						 * Lookup Engine
+						 */
+	FW_PARAMS_PARAM_DEV_FLOWC_BUFFIFO_SZ = 0x03,
+	FW_PARAMS_PARAM_DEV_INTVER_NIC	= 0x04,
+	FW_PARAMS_PARAM_DEV_INTVER_VNIC = 0x05,
+	FW_PARAMS_PARAM_DEV_INTVER_OFLD = 0x06,
+	FW_PARAMS_PARAM_DEV_INTVER_RI	= 0x07,
+	FW_PARAMS_PARAM_DEV_INTVER_ISCSIPDU = 0x08,
+	FW_PARAMS_PARAM_DEV_INTVER_ISCSI = 0x09,
+	FW_PARAMS_PARAM_DEV_INTVER_FCOE = 0x0A
+};
+
+/*
+ * physical and virtual function parameters
+ */
+enum fw_params_param_pfvf {
+	FW_PARAMS_PARAM_PFVF_RWXCAPS	= 0x00,
+	FW_PARAMS_PARAM_PFVF_ROUTE_START = 0x01,
+	FW_PARAMS_PARAM_PFVF_ROUTE_END = 0x02,
+	FW_PARAMS_PARAM_PFVF_CLIP_START = 0x03,
+	FW_PARAMS_PARAM_PFVF_CLIP_END = 0x04,
+	FW_PARAMS_PARAM_PFVF_FILTER_START = 0x05,
+	FW_PARAMS_PARAM_PFVF_FILTER_END = 0x06,
+	FW_PARAMS_PARAM_PFVF_SERVER_START = 0x07,
+	FW_PARAMS_PARAM_PFVF_SERVER_END = 0x08,
+	FW_PARAMS_PARAM_PFVF_TDDP_START = 0x09,
+	FW_PARAMS_PARAM_PFVF_TDDP_END = 0x0A,
+	FW_PARAMS_PARAM_PFVF_ISCSI_START = 0x0B,
+	FW_PARAMS_PARAM_PFVF_ISCSI_END = 0x0C,
+	FW_PARAMS_PARAM_PFVF_STAG_START = 0x0D,
+	FW_PARAMS_PARAM_PFVF_STAG_END = 0x0E,
+	FW_PARAMS_PARAM_PFVF_RQ_START = 0x1F,
+	FW_PARAMS_PARAM_PFVF_RQ_END	= 0x10,
+	FW_PARAMS_PARAM_PFVF_PBL_START = 0x11,
+	FW_PARAMS_PARAM_PFVF_PBL_END	= 0x12,
+	FW_PARAMS_PARAM_PFVF_L2T_START = 0x13,
+	FW_PARAMS_PARAM_PFVF_L2T_END = 0x14,
+	FW_PARAMS_PARAM_PFVF_SCHEDCLASS_ETH = 0x20,
+};
+
+/*
+ * dma queue parameters
+ */
+enum fw_params_param_dmaq {
+	FW_PARAMS_PARAM_DMAQ_IQ_DCAEN_DCACPU = 0x00,
+	FW_PARAMS_PARAM_DMAQ_IQ_INTCNTTHRESH = 0x01,
+	FW_PARAMS_PARAM_DMAQ_EQ_CMPLIQID_MNGT = 0x10,
+	FW_PARAMS_PARAM_DMAQ_EQ_CMPLIQID_CTRL = 0x11,
+	FW_PARAMS_PARAM_DMAQ_EQ_SCHEDCLASS_ETH = 0x12,
+};
+
+#define FW_PARAMS_MNEM(x)      ((x) << 24)
+#define FW_PARAMS_PARAM_X(x)   ((x) << 16)
+#define FW_PARAMS_PARAM_Y(x)   ((x) << 8)
+#define FW_PARAMS_PARAM_Z(x)   ((x) << 0)
+#define FW_PARAMS_PARAM_XYZ(x) ((x) << 0)
+#define FW_PARAMS_PARAM_YZ(x)  ((x) << 0)
+
+struct fw_params_cmd {
+	__be32 op_to_vfn;
+	__be32 retval_len16;
+	struct fw_params_param {
+		__be32 mnem;
+		__be32 val;
+	} param[7];
+};
+
+#define FW_PARAMS_CMD_PFN(x) ((x) << 8)
+#define FW_PARAMS_CMD_VFN(x) ((x) << 0)
+
+struct fw_pfvf_cmd {
+	__be32 op_to_vfn;
+	__be32 retval_len16;
+	__be32 niqflint_niq;
+	__be32 cmask_to_neq;
+	__be32 tc_to_nexactf;
+	__be32 r_caps_to_nethctrl;
+	__be16 nricq;
+	__be16 nriqp;
+	__be32 r4;
+};
+
+#define FW_PFVF_CMD_PFN(x) ((x) << 8)
+#define FW_PFVF_CMD_VFN(x) ((x) << 0)
+
+#define FW_PFVF_CMD_NIQFLINT(x) ((x) << 20)
+#define FW_PFVF_CMD_NIQFLINT_GET(x) (((x) >> 20) & 0xfff)
+
+#define FW_PFVF_CMD_NIQ(x) ((x) << 0)
+#define FW_PFVF_CMD_NIQ_GET(x) (((x) >> 0) & 0xfffff)
+
+#define FW_PFVF_CMD_CMASK(x) ((x) << 24)
+#define FW_PFVF_CMD_CMASK_GET(x) (((x) >> 24) & 0xf)
+
+#define FW_PFVF_CMD_PMASK(x) ((x) << 20)
+#define FW_PFVF_CMD_PMASK_GET(x) (((x) >> 20) & 0xf)
+
+#define FW_PFVF_CMD_NEQ(x) ((x) << 0)
+#define FW_PFVF_CMD_NEQ_GET(x) (((x) >> 0) & 0xfffff)
+
+#define FW_PFVF_CMD_TC(x) ((x) << 24)
+#define FW_PFVF_CMD_TC_GET(x) (((x) >> 24) & 0xff)
+
+#define FW_PFVF_CMD_NVI(x) ((x) << 16)
+#define FW_PFVF_CMD_NVI_GET(x) (((x) >> 16) & 0xff)
+
+#define FW_PFVF_CMD_NEXACTF(x) ((x) << 0)
+#define FW_PFVF_CMD_NEXACTF_GET(x) (((x) >> 0) & 0xffff)
+
+#define FW_PFVF_CMD_R_CAPS(x) ((x) << 24)
+#define FW_PFVF_CMD_R_CAPS_GET(x) (((x) >> 24) & 0xff)
+
+#define FW_PFVF_CMD_WX_CAPS(x) ((x) << 16)
+#define FW_PFVF_CMD_WX_CAPS_GET(x) (((x) >> 16) & 0xff)
+
+#define FW_PFVF_CMD_NETHCTRL(x) ((x) << 0)
+#define FW_PFVF_CMD_NETHCTRL_GET(x) (((x) >> 0) & 0xffff)
+
+enum fw_iq_type {
+	FW_IQ_TYPE_FL_INT_CAP,
+	FW_IQ_TYPE_NO_FL_INT_CAP
+};
+
+struct fw_iq_cmd {
+	__be32 op_to_vfn;
+	__be32 alloc_to_len16;
+	__be16 physiqid;
+	__be16 iqid;
+	__be16 fl0id;
+	__be16 fl1id;
+	__be32 type_to_iqandstindex;
+	__be16 iqdroprss_to_iqesize;
+	__be16 iqsize;
+	__be64 iqaddr;
+	__be32 iqns_to_fl0congen;
+	__be16 fl0dcaen_to_fl0cidxfthresh;
+	__be16 fl0size;
+	__be64 fl0addr;
+	__be32 fl1cngchmap_to_fl1congen;
+	__be16 fl1dcaen_to_fl1cidxfthresh;
+	__be16 fl1size;
+	__be64 fl1addr;
+};
+
+#define FW_IQ_CMD_PFN(x) ((x) << 8)
+#define FW_IQ_CMD_VFN(x) ((x) << 0)
+
+#define FW_IQ_CMD_ALLOC (1U << 31)
+#define FW_IQ_CMD_FREE (1U << 30)
+#define FW_IQ_CMD_MODIFY (1U << 29)
+#define FW_IQ_CMD_IQSTART(x) ((x) << 28)
+#define FW_IQ_CMD_IQSTOP(x) ((x) << 27)
+
+#define FW_IQ_CMD_TYPE(x) ((x) << 29)
+#define FW_IQ_CMD_IQASYNCH(x) ((x) << 28)
+#define FW_IQ_CMD_VIID(x) ((x) << 16)
+#define FW_IQ_CMD_IQANDST(x) ((x) << 15)
+#define FW_IQ_CMD_IQANUS(x) ((x) << 14)
+#define FW_IQ_CMD_IQANUD(x) ((x) << 12)
+#define FW_IQ_CMD_IQANDSTINDEX(x) ((x) << 0)
+
+#define FW_IQ_CMD_IQDROPRSS (1U << 15)
+#define FW_IQ_CMD_IQGTSMODE (1U << 14)
+#define FW_IQ_CMD_IQPCIECH(x) ((x) << 12)
+#define FW_IQ_CMD_IQDCAEN(x) ((x) << 11)
+#define FW_IQ_CMD_IQDCACPU(x) ((x) << 6)
+#define FW_IQ_CMD_IQINTCNTTHRESH(x) ((x) << 4)
+#define FW_IQ_CMD_IQO (1U << 3)
+#define FW_IQ_CMD_IQCPRIO(x) ((x) << 2)
+#define FW_IQ_CMD_IQESIZE(x) ((x) << 0)
+
+#define FW_IQ_CMD_IQNS(x) ((x) << 31)
+#define FW_IQ_CMD_IQRO(x) ((x) << 30)
+#define FW_IQ_CMD_IQFLINTIQHSEN(x) ((x) << 28)
+#define FW_IQ_CMD_IQFLINTCONGEN(x) ((x) << 27)
+#define FW_IQ_CMD_IQFLINTISCSIC(x) ((x) << 26)
+#define FW_IQ_CMD_FL0CNGCHMAP(x) ((x) << 20)
+#define FW_IQ_CMD_FL0CACHELOCK(x) ((x) << 15)
+#define FW_IQ_CMD_FL0DBP(x) ((x) << 14)
+#define FW_IQ_CMD_FL0DATANS(x) ((x) << 13)
+#define FW_IQ_CMD_FL0DATARO(x) ((x) << 12)
+#define FW_IQ_CMD_FL0CONGCIF(x) ((x) << 11)
+#define FW_IQ_CMD_FL0ONCHIP(x) ((x) << 10)
+#define FW_IQ_CMD_FL0STATUSPGNS(x) ((x) << 9)
+#define FW_IQ_CMD_FL0STATUSPGRO(x) ((x) << 8)
+#define FW_IQ_CMD_FL0FETCHNS(x) ((x) << 7)
+#define FW_IQ_CMD_FL0FETCHRO(x) ((x) << 6)
+#define FW_IQ_CMD_FL0HOSTFCMODE(x) ((x) << 4)
+#define FW_IQ_CMD_FL0CPRIO(x) ((x) << 3)
+#define FW_IQ_CMD_FL0PADEN (1U << 2)
+#define FW_IQ_CMD_FL0PACKEN (1U << 1)
+#define FW_IQ_CMD_FL0CONGEN (1U << 0)
+
+#define FW_IQ_CMD_FL0DCAEN(x) ((x) << 15)
+#define FW_IQ_CMD_FL0DCACPU(x) ((x) << 10)
+#define FW_IQ_CMD_FL0FBMIN(x) ((x) << 7)
+#define FW_IQ_CMD_FL0FBMAX(x) ((x) << 4)
+#define FW_IQ_CMD_FL0CIDXFTHRESHO (1U << 3)
+#define FW_IQ_CMD_FL0CIDXFTHRESH(x) ((x) << 0)
+
+#define FW_IQ_CMD_FL1CNGCHMAP(x) ((x) << 20)
+#define FW_IQ_CMD_FL1CACHELOCK(x) ((x) << 15)
+#define FW_IQ_CMD_FL1DBP(x) ((x) << 14)
+#define FW_IQ_CMD_FL1DATANS(x) ((x) << 13)
+#define FW_IQ_CMD_FL1DATARO(x) ((x) << 12)
+#define FW_IQ_CMD_FL1CONGCIF(x) ((x) << 11)
+#define FW_IQ_CMD_FL1ONCHIP(x) ((x) << 10)
+#define FW_IQ_CMD_FL1STATUSPGNS(x) ((x) << 9)
+#define FW_IQ_CMD_FL1STATUSPGRO(x) ((x) << 8)
+#define FW_IQ_CMD_FL1FETCHNS(x) ((x) << 7)
+#define FW_IQ_CMD_FL1FETCHRO(x) ((x) << 6)
+#define FW_IQ_CMD_FL1HOSTFCMODE(x) ((x) << 4)
+#define FW_IQ_CMD_FL1CPRIO(x) ((x) << 3)
+#define FW_IQ_CMD_FL1PADEN (1U << 2)
+#define FW_IQ_CMD_FL1PACKEN (1U << 1)
+#define FW_IQ_CMD_FL1CONGEN (1U << 0)
+
+#define FW_IQ_CMD_FL1DCAEN(x) ((x) << 15)
+#define FW_IQ_CMD_FL1DCACPU(x) ((x) << 10)
+#define FW_IQ_CMD_FL1FBMIN(x) ((x) << 7)
+#define FW_IQ_CMD_FL1FBMAX(x) ((x) << 4)
+#define FW_IQ_CMD_FL1CIDXFTHRESHO (1U << 3)
+#define FW_IQ_CMD_FL1CIDXFTHRESH(x) ((x) << 0)
+
+struct fw_eq_eth_cmd {
+	__be32 op_to_vfn;
+	__be32 alloc_to_len16;
+	__be32 eqid_pkd;
+	__be32 physeqid_pkd;
+	__be32 fetchszm_to_iqid;
+	__be32 dcaen_to_eqsize;
+	__be64 eqaddr;
+	__be32 viid_pkd;
+	__be32 r8_lo;
+	__be64 r9;
+};
+
+#define FW_EQ_ETH_CMD_PFN(x) ((x) << 8)
+#define FW_EQ_ETH_CMD_VFN(x) ((x) << 0)
+#define FW_EQ_ETH_CMD_ALLOC (1U << 31)
+#define FW_EQ_ETH_CMD_FREE (1U << 30)
+#define FW_EQ_ETH_CMD_MODIFY (1U << 29)
+#define FW_EQ_ETH_CMD_EQSTART (1U << 28)
+#define FW_EQ_ETH_CMD_EQSTOP (1U << 27)
+
+#define FW_EQ_ETH_CMD_EQID(x) ((x) << 0)
+#define FW_EQ_ETH_CMD_EQID_GET(x) (((x) >> 0) & 0xfffff)
+#define FW_EQ_ETH_CMD_PHYSEQID(x) ((x) << 0)
+
+#define FW_EQ_ETH_CMD_FETCHSZM(x) ((x) << 26)
+#define FW_EQ_ETH_CMD_STATUSPGNS(x) ((x) << 25)
+#define FW_EQ_ETH_CMD_STATUSPGRO(x) ((x) << 24)
+#define FW_EQ_ETH_CMD_FETCHNS(x) ((x) << 23)
+#define FW_EQ_ETH_CMD_FETCHRO(x) ((x) << 22)
+#define FW_EQ_ETH_CMD_HOSTFCMODE(x) ((x) << 20)
+#define FW_EQ_ETH_CMD_CPRIO(x) ((x) << 19)
+#define FW_EQ_ETH_CMD_ONCHIP(x) ((x) << 18)
+#define FW_EQ_ETH_CMD_PCIECHN(x) ((x) << 16)
+#define FW_EQ_ETH_CMD_IQID(x) ((x) << 0)
+
+#define FW_EQ_ETH_CMD_DCAEN(x) ((x) << 31)
+#define FW_EQ_ETH_CMD_DCACPU(x) ((x) << 26)
+#define FW_EQ_ETH_CMD_FBMIN(x) ((x) << 23)
+#define FW_EQ_ETH_CMD_FBMAX(x) ((x) << 20)
+#define FW_EQ_ETH_CMD_CIDXFTHRESHO(x) ((x) << 19)
+#define FW_EQ_ETH_CMD_CIDXFTHRESH(x) ((x) << 16)
+#define FW_EQ_ETH_CMD_EQSIZE(x) ((x) << 0)
+
+#define FW_EQ_ETH_CMD_VIID(x) ((x) << 16)
+
+struct fw_eq_ctrl_cmd {
+	__be32 op_to_vfn;
+	__be32 alloc_to_len16;
+	__be32 cmpliqid_eqid;
+	__be32 physeqid_pkd;
+	__be32 fetchszm_to_iqid;
+	__be32 dcaen_to_eqsize;
+	__be64 eqaddr;
+};
+
+#define FW_EQ_CTRL_CMD_PFN(x) ((x) << 8)
+#define FW_EQ_CTRL_CMD_VFN(x) ((x) << 0)
+
+#define FW_EQ_CTRL_CMD_ALLOC (1U << 31)
+#define FW_EQ_CTRL_CMD_FREE (1U << 30)
+#define FW_EQ_CTRL_CMD_MODIFY (1U << 29)
+#define FW_EQ_CTRL_CMD_EQSTART (1U << 28)
+#define FW_EQ_CTRL_CMD_EQSTOP (1U << 27)
+
+#define FW_EQ_CTRL_CMD_CMPLIQID(x) ((x) << 20)
+#define FW_EQ_CTRL_CMD_EQID(x) ((x) << 0)
+#define FW_EQ_CTRL_CMD_EQID_GET(x) (((x) >> 0) & 0xfffff)
+#define FW_EQ_CTRL_CMD_PHYSEQID_GET(x) (((x) >> 0) & 0xfffff)
+
+#define FW_EQ_CTRL_CMD_FETCHSZM (1U << 26)
+#define FW_EQ_CTRL_CMD_STATUSPGNS (1U << 25)
+#define FW_EQ_CTRL_CMD_STATUSPGRO (1U << 24)
+#define FW_EQ_CTRL_CMD_FETCHNS (1U << 23)
+#define FW_EQ_CTRL_CMD_FETCHRO (1U << 22)
+#define FW_EQ_CTRL_CMD_HOSTFCMODE(x) ((x) << 20)
+#define FW_EQ_CTRL_CMD_CPRIO(x) ((x) << 19)
+#define FW_EQ_CTRL_CMD_ONCHIP(x) ((x) << 18)
+#define FW_EQ_CTRL_CMD_PCIECHN(x) ((x) << 16)
+#define FW_EQ_CTRL_CMD_IQID(x) ((x) << 0)
+
+#define FW_EQ_CTRL_CMD_DCAEN(x) ((x) << 31)
+#define FW_EQ_CTRL_CMD_DCACPU(x) ((x) << 26)
+#define FW_EQ_CTRL_CMD_FBMIN(x) ((x) << 23)
+#define FW_EQ_CTRL_CMD_FBMAX(x) ((x) << 20)
+#define FW_EQ_CTRL_CMD_CIDXFTHRESHO(x) ((x) << 19)
+#define FW_EQ_CTRL_CMD_CIDXFTHRESH(x) ((x) << 16)
+#define FW_EQ_CTRL_CMD_EQSIZE(x) ((x) << 0)
+
+struct fw_eq_ofld_cmd {
+	__be32 op_to_vfn;
+	__be32 alloc_to_len16;
+	__be32 eqid_pkd;
+	__be32 physeqid_pkd;
+	__be32 fetchszm_to_iqid;
+	__be32 dcaen_to_eqsize;
+	__be64 eqaddr;
+};
+
+#define FW_EQ_OFLD_CMD_PFN(x) ((x) << 8)
+#define FW_EQ_OFLD_CMD_VFN(x) ((x) << 0)
+
+#define FW_EQ_OFLD_CMD_ALLOC (1U << 31)
+#define FW_EQ_OFLD_CMD_FREE (1U << 30)
+#define FW_EQ_OFLD_CMD_MODIFY (1U << 29)
+#define FW_EQ_OFLD_CMD_EQSTART (1U << 28)
+#define FW_EQ_OFLD_CMD_EQSTOP (1U << 27)
+
+#define FW_EQ_OFLD_CMD_EQID(x) ((x) << 0)
+#define FW_EQ_OFLD_CMD_EQID_GET(x) (((x) >> 0) & 0xfffff)
+#define FW_EQ_OFLD_CMD_PHYSEQID_GET(x) (((x) >> 0) & 0xfffff)
+
+#define FW_EQ_OFLD_CMD_FETCHSZM(x) ((x) << 26)
+#define FW_EQ_OFLD_CMD_STATUSPGNS(x) ((x) << 25)
+#define FW_EQ_OFLD_CMD_STATUSPGRO(x) ((x) << 24)
+#define FW_EQ_OFLD_CMD_FETCHNS(x) ((x) << 23)
+#define FW_EQ_OFLD_CMD_FETCHRO(x) ((x) << 22)
+#define FW_EQ_OFLD_CMD_HOSTFCMODE(x) ((x) << 20)
+#define FW_EQ_OFLD_CMD_CPRIO(x) ((x) << 19)
+#define FW_EQ_OFLD_CMD_ONCHIP(x) ((x) << 18)
+#define FW_EQ_OFLD_CMD_PCIECHN(x) ((x) << 16)
+#define FW_EQ_OFLD_CMD_IQID(x) ((x) << 0)
+
+#define FW_EQ_OFLD_CMD_DCAEN(x) ((x) << 31)
+#define FW_EQ_OFLD_CMD_DCACPU(x) ((x) << 26)
+#define FW_EQ_OFLD_CMD_FBMIN(x) ((x) << 23)
+#define FW_EQ_OFLD_CMD_FBMAX(x) ((x) << 20)
+#define FW_EQ_OFLD_CMD_CIDXFTHRESHO(x) ((x) << 19)
+#define FW_EQ_OFLD_CMD_CIDXFTHRESH(x) ((x) << 16)
+#define FW_EQ_OFLD_CMD_EQSIZE(x) ((x) << 0)
+
+/*
+ * Macros for VIID parsing:
+ * VIID - [10:8] PFN, [7] VI Valid, [6:0] VI number
+ */
+#define FW_VIID_PFN_GET(x) (((x) >> 8) & 0x7)
+#define FW_VIID_VIVLD_GET(x) (((x) >> 7) & 0x1)
+#define FW_VIID_VIN_GET(x) (((x) >> 0) & 0x7F)
+
+struct fw_vi_cmd {
+	__be32 op_to_vfn;
+	__be32 alloc_to_len16;
+	__be16 viid_pkd;
+	u8 mac[6];
+	u8 portid_pkd;
+	u8 nmac;
+	u8 nmac0[6];
+	__be16 rsssize_pkd;
+	u8 nmac1[6];
+	__be16 r7;
+	u8 nmac2[6];
+	__be16 r8;
+	u8 nmac3[6];
+	__be64 r9;
+	__be64 r10;
+};
+
+#define FW_VI_CMD_PFN(x) ((x) << 8)
+#define FW_VI_CMD_VFN(x) ((x) << 0)
+#define FW_VI_CMD_ALLOC (1U << 31)
+#define FW_VI_CMD_FREE (1U << 30)
+#define FW_VI_CMD_VIID(x) ((x) << 0)
+#define FW_VI_CMD_PORTID(x) ((x) << 4)
+#define FW_VI_CMD_RSSSIZE_GET(x) (((x) >> 0) & 0x7ff)
+
+/* Special VI_MAC command index ids */
+#define FW_VI_MAC_ADD_MAC		0x3FF
+#define FW_VI_MAC_ADD_PERSIST_MAC	0x3FE
+#define FW_VI_MAC_MAC_BASED_FREE	0x3FD
+
+enum fw_vi_mac_smac {
+	FW_VI_MAC_MPS_TCAM_ENTRY,
+	FW_VI_MAC_MPS_TCAM_ONLY,
+	FW_VI_MAC_SMT_ONLY,
+	FW_VI_MAC_SMT_AND_MPSTCAM
+};
+
+enum fw_vi_mac_result {
+	FW_VI_MAC_R_SUCCESS,
+	FW_VI_MAC_R_F_NONEXISTENT_NOMEM,
+	FW_VI_MAC_R_SMAC_FAIL,
+	FW_VI_MAC_R_F_ACL_CHECK
+};
+
+struct fw_vi_mac_cmd {
+	__be32 op_to_viid;
+	__be32 freemacs_to_len16;
+	union fw_vi_mac {
+		struct fw_vi_mac_exact {
+			__be16 valid_to_idx;
+			u8 macaddr[6];
+		} exact[7];
+		struct fw_vi_mac_hash {
+			__be64 hashvec;
+		} hash;
+	} u;
+};
+
+#define FW_VI_MAC_CMD_VIID(x) ((x) << 0)
+#define FW_VI_MAC_CMD_FREEMACS(x) ((x) << 31)
+#define FW_VI_MAC_CMD_HASHVECEN (1U << 23)
+#define FW_VI_MAC_CMD_HASHUNIEN(x) ((x) << 22)
+#define FW_VI_MAC_CMD_VALID (1U << 15)
+#define FW_VI_MAC_CMD_PRIO(x) ((x) << 12)
+#define FW_VI_MAC_CMD_SMAC_RESULT(x) ((x) << 10)
+#define FW_VI_MAC_CMD_SMAC_RESULT_GET(x) (((x) >> 10) & 0x3)
+#define FW_VI_MAC_CMD_IDX(x) ((x) << 0)
+#define FW_VI_MAC_CMD_IDX_GET(x) (((x) >> 0) & 0x3ff)
+
+#define FW_RXMODE_MTU_NO_CHG	65535
+
+struct fw_vi_rxmode_cmd {
+	__be32 op_to_viid;
+	__be32 retval_len16;
+	__be32 mtu_to_broadcasten;
+	__be32 r4_lo;
+};
+
+#define FW_VI_RXMODE_CMD_VIID(x) ((x) << 0)
+#define FW_VI_RXMODE_CMD_MTU(x) ((x) << 16)
+#define FW_VI_RXMODE_CMD_PROMISCEN_MASK 0x3
+#define FW_VI_RXMODE_CMD_PROMISCEN(x) ((x) << 14)
+#define FW_VI_RXMODE_CMD_ALLMULTIEN_MASK 0x3
+#define FW_VI_RXMODE_CMD_ALLMULTIEN(x) ((x) << 12)
+#define FW_VI_RXMODE_CMD_BROADCASTEN_MASK 0x3
+#define FW_VI_RXMODE_CMD_BROADCASTEN(x) ((x) << 10)
+
+struct fw_vi_enable_cmd {
+	__be32 op_to_viid;
+	__be32 ien_to_len16;
+	__be16 blinkdur;
+	__be16 r3;
+	__be32 r4;
+};
+
+#define FW_VI_ENABLE_CMD_VIID(x) ((x) << 0)
+#define FW_VI_ENABLE_CMD_IEN(x) ((x) << 31)
+#define FW_VI_ENABLE_CMD_EEN(x) ((x) << 30)
+#define FW_VI_ENABLE_CMD_LED (1U << 29)
+
+/* VI VF stats offset definitions */
+#define VI_VF_NUM_STATS	16
+enum fw_vi_stats_vf_index {
+	FW_VI_VF_STAT_TX_BCAST_BYTES_IX,
+	FW_VI_VF_STAT_TX_BCAST_FRAMES_IX,
+	FW_VI_VF_STAT_TX_MCAST_BYTES_IX,
+	FW_VI_VF_STAT_TX_MCAST_FRAMES_IX,
+	FW_VI_VF_STAT_TX_UCAST_BYTES_IX,
+	FW_VI_VF_STAT_TX_UCAST_FRAMES_IX,
+	FW_VI_VF_STAT_TX_DROP_FRAMES_IX,
+	FW_VI_VF_STAT_TX_OFLD_BYTES_IX,
+	FW_VI_VF_STAT_TX_OFLD_FRAMES_IX,
+	FW_VI_VF_STAT_RX_BCAST_BYTES_IX,
+	FW_VI_VF_STAT_RX_BCAST_FRAMES_IX,
+	FW_VI_VF_STAT_RX_MCAST_BYTES_IX,
+	FW_VI_VF_STAT_RX_MCAST_FRAMES_IX,
+	FW_VI_VF_STAT_RX_UCAST_BYTES_IX,
+	FW_VI_VF_STAT_RX_UCAST_FRAMES_IX,
+	FW_VI_VF_STAT_RX_ERR_FRAMES_IX
+};
+
+/* VI PF stats offset definitions */
+#define VI_PF_NUM_STATS	17
+enum fw_vi_stats_pf_index {
+	FW_VI_PF_STAT_TX_BCAST_BYTES_IX,
+	FW_VI_PF_STAT_TX_BCAST_FRAMES_IX,
+	FW_VI_PF_STAT_TX_MCAST_BYTES_IX,
+	FW_VI_PF_STAT_TX_MCAST_FRAMES_IX,
+	FW_VI_PF_STAT_TX_UCAST_BYTES_IX,
+	FW_VI_PF_STAT_TX_UCAST_FRAMES_IX,
+	FW_VI_PF_STAT_TX_OFLD_BYTES_IX,
+	FW_VI_PF_STAT_TX_OFLD_FRAMES_IX,
+	FW_VI_PF_STAT_RX_BYTES_IX,
+	FW_VI_PF_STAT_RX_FRAMES_IX,
+	FW_VI_PF_STAT_RX_BCAST_BYTES_IX,
+	FW_VI_PF_STAT_RX_BCAST_FRAMES_IX,
+	FW_VI_PF_STAT_RX_MCAST_BYTES_IX,
+	FW_VI_PF_STAT_RX_MCAST_FRAMES_IX,
+	FW_VI_PF_STAT_RX_UCAST_BYTES_IX,
+	FW_VI_PF_STAT_RX_UCAST_FRAMES_IX,
+	FW_VI_PF_STAT_RX_ERR_FRAMES_IX
+};
+
+struct fw_vi_stats_cmd {
+	__be32 op_to_viid;
+	__be32 retval_len16;
+	union fw_vi_stats {
+		struct fw_vi_stats_ctl {
+			__be16 nstats_ix;
+			__be16 r6;
+			__be32 r7;
+			__be64 stat0;
+			__be64 stat1;
+			__be64 stat2;
+			__be64 stat3;
+			__be64 stat4;
+			__be64 stat5;
+		} ctl;
+		struct fw_vi_stats_pf {
+			__be64 tx_bcast_bytes;
+			__be64 tx_bcast_frames;
+			__be64 tx_mcast_bytes;
+			__be64 tx_mcast_frames;
+			__be64 tx_ucast_bytes;
+			__be64 tx_ucast_frames;
+			__be64 tx_offload_bytes;
+			__be64 tx_offload_frames;
+			__be64 rx_pf_bytes;
+			__be64 rx_pf_frames;
+			__be64 rx_bcast_bytes;
+			__be64 rx_bcast_frames;
+			__be64 rx_mcast_bytes;
+			__be64 rx_mcast_frames;
+			__be64 rx_ucast_bytes;
+			__be64 rx_ucast_frames;
+			__be64 rx_err_frames;
+		} pf;
+		struct fw_vi_stats_vf {
+			__be64 tx_bcast_bytes;
+			__be64 tx_bcast_frames;
+			__be64 tx_mcast_bytes;
+			__be64 tx_mcast_frames;
+			__be64 tx_ucast_bytes;
+			__be64 tx_ucast_frames;
+			__be64 tx_drop_frames;
+			__be64 tx_offload_bytes;
+			__be64 tx_offload_frames;
+			__be64 rx_bcast_bytes;
+			__be64 rx_bcast_frames;
+			__be64 rx_mcast_bytes;
+			__be64 rx_mcast_frames;
+			__be64 rx_ucast_bytes;
+			__be64 rx_ucast_frames;
+			__be64 rx_err_frames;
+		} vf;
+	} u;
+};
+
+#define FW_VI_STATS_CMD_VIID(x) ((x) << 0)
+#define FW_VI_STATS_CMD_NSTATS(x) ((x) << 12)
+#define FW_VI_STATS_CMD_IX(x) ((x) << 0)
+
+struct fw_acl_mac_cmd {
+	__be32 op_to_vfn;
+	__be32 en_to_len16;
+	u8 nmac;
+	u8 r3[7];
+	__be16 r4;
+	u8 macaddr0[6];
+	__be16 r5;
+	u8 macaddr1[6];
+	__be16 r6;
+	u8 macaddr2[6];
+	__be16 r7;
+	u8 macaddr3[6];
+};
+
+#define FW_ACL_MAC_CMD_PFN(x) ((x) << 8)
+#define FW_ACL_MAC_CMD_VFN(x) ((x) << 0)
+#define FW_ACL_MAC_CMD_EN(x) ((x) << 31)
+
+struct fw_acl_vlan_cmd {
+	__be32 op_to_vfn;
+	__be32 en_to_len16;
+	u8 nvlan;
+	u8 dropnovlan_fm;
+	u8 r3_lo[6];
+	__be16 vlanid[16];
+};
+
+#define FW_ACL_VLAN_CMD_PFN(x) ((x) << 8)
+#define FW_ACL_VLAN_CMD_VFN(x) ((x) << 0)
+#define FW_ACL_VLAN_CMD_EN(x) ((x) << 31)
+#define FW_ACL_VLAN_CMD_DROPNOVLAN(x) ((x) << 7)
+#define FW_ACL_VLAN_CMD_FM(x) ((x) << 6)
+
+enum fw_port_cap {
+	FW_PORT_CAP_SPEED_100M		= 0x0001,
+	FW_PORT_CAP_SPEED_1G		= 0x0002,
+	FW_PORT_CAP_SPEED_2_5G		= 0x0004,
+	FW_PORT_CAP_SPEED_10G		= 0x0008,
+	FW_PORT_CAP_SPEED_40G		= 0x0010,
+	FW_PORT_CAP_SPEED_100G		= 0x0020,
+	FW_PORT_CAP_FC_RX		= 0x0040,
+	FW_PORT_CAP_FC_TX		= 0x0080,
+	FW_PORT_CAP_ANEG		= 0x0100,
+	FW_PORT_CAP_MDI_0		= 0x0200,
+	FW_PORT_CAP_MDI_1		= 0x0400,
+	FW_PORT_CAP_BEAN		= 0x0800,
+	FW_PORT_CAP_PMA_LPBK		= 0x1000,
+	FW_PORT_CAP_PCS_LPBK		= 0x2000,
+	FW_PORT_CAP_PHYXS_LPBK		= 0x4000,
+	FW_PORT_CAP_FAR_END_LPBK	= 0x8000,
+};
+
+enum fw_port_mdi {
+	FW_PORT_MDI_UNCHANGED,
+	FW_PORT_MDI_AUTO,
+	FW_PORT_MDI_F_STRAIGHT,
+	FW_PORT_MDI_F_CROSSOVER
+};
+
+#define FW_PORT_MDI(x) ((x) << 9)
+
+enum fw_port_action {
+	FW_PORT_ACTION_L1_CFG		= 0x0001,
+	FW_PORT_ACTION_L2_CFG		= 0x0002,
+	FW_PORT_ACTION_GET_PORT_INFO	= 0x0003,
+	FW_PORT_ACTION_L2_PPP_CFG	= 0x0004,
+	FW_PORT_ACTION_L2_DCB_CFG	= 0x0005,
+	FW_PORT_ACTION_LOW_PWR_TO_NORMAL = 0x0010,
+	FW_PORT_ACTION_L1_LOW_PWR_EN	= 0x0011,
+	FW_PORT_ACTION_L2_WOL_MODE_EN	= 0x0012,
+	FW_PORT_ACTION_LPBK_TO_NORMAL	= 0x0020,
+	FW_PORT_ACTION_L1_LPBK		= 0x0021,
+	FW_PORT_ACTION_L1_PMA_LPBK	= 0x0022,
+	FW_PORT_ACTION_L1_PCS_LPBK	= 0x0023,
+	FW_PORT_ACTION_L1_PHYXS_CSIDE_LPBK = 0x0024,
+	FW_PORT_ACTION_L1_PHYXS_ESIDE_LPBK = 0x0025,
+	FW_PORT_ACTION_PHY_RESET	= 0x0040,
+	FW_PORT_ACTION_PMA_RESET	= 0x0041,
+	FW_PORT_ACTION_PCS_RESET	= 0x0042,
+	FW_PORT_ACTION_PHYXS_RESET	= 0x0043,
+	FW_PORT_ACTION_DTEXS_REEST	= 0x0044,
+	FW_PORT_ACTION_AN_RESET		= 0x0045
+};
+
+enum fw_port_l2cfg_ctlbf {
+	FW_PORT_L2_CTLBF_OVLAN0	= 0x01,
+	FW_PORT_L2_CTLBF_OVLAN1	= 0x02,
+	FW_PORT_L2_CTLBF_OVLAN2	= 0x04,
+	FW_PORT_L2_CTLBF_OVLAN3	= 0x08,
+	FW_PORT_L2_CTLBF_IVLAN	= 0x10,
+	FW_PORT_L2_CTLBF_TXIPG	= 0x20
+};
+
+enum fw_port_dcb_cfg {
+	FW_PORT_DCB_CFG_PG	= 0x01,
+	FW_PORT_DCB_CFG_PFC	= 0x02,
+	FW_PORT_DCB_CFG_APPL	= 0x04
+};
+
+enum fw_port_dcb_cfg_rc {
+	FW_PORT_DCB_CFG_SUCCESS	= 0x0,
+	FW_PORT_DCB_CFG_ERROR	= 0x1
+};
+
+struct fw_port_cmd {
+	__be32 op_to_portid;
+	__be32 action_to_len16;
+	union fw_port {
+		struct fw_port_l1cfg {
+			__be32 rcap;
+			__be32 r;
+		} l1cfg;
+		struct fw_port_l2cfg {
+			__be16 ctlbf_to_ivlan0;
+			__be16 ivlantype;
+			__be32 txipg_pkd;
+			__be16 ovlan0mask;
+			__be16 ovlan0type;
+			__be16 ovlan1mask;
+			__be16 ovlan1type;
+			__be16 ovlan2mask;
+			__be16 ovlan2type;
+			__be16 ovlan3mask;
+			__be16 ovlan3type;
+		} l2cfg;
+		struct fw_port_info {
+			__be32 lstatus_to_modtype;
+			__be16 pcap;
+			__be16 acap;
+		} info;
+		struct fw_port_ppp {
+			__be32 pppen_to_ncsich;
+			__be32 r11;
+		} ppp;
+		struct fw_port_dcb {
+			__be16 cfg;
+			u8 up_map;
+			u8 sf_cfgrc;
+			__be16 prot_ix;
+			u8 pe7_to_pe0;
+			u8 numTCPFCs;
+			__be32 pgid0_to_pgid7;
+			__be32 numTCs_oui;
+			u8 pgpc[8];
+		} dcb;
+	} u;
+};
+
+#define FW_PORT_CMD_READ (1U << 22)
+
+#define FW_PORT_CMD_PORTID(x) ((x) << 0)
+#define FW_PORT_CMD_PORTID_GET(x) (((x) >> 0) & 0xf)
+
+#define FW_PORT_CMD_ACTION(x) ((x) << 16)
+
+#define FW_PORT_CMD_CTLBF(x) ((x) << 10)
+#define FW_PORT_CMD_OVLAN3(x) ((x) << 7)
+#define FW_PORT_CMD_OVLAN2(x) ((x) << 6)
+#define FW_PORT_CMD_OVLAN1(x) ((x) << 5)
+#define FW_PORT_CMD_OVLAN0(x) ((x) << 4)
+#define FW_PORT_CMD_IVLAN0(x) ((x) << 3)
+
+#define FW_PORT_CMD_TXIPG(x) ((x) << 19)
+
+#define FW_PORT_CMD_LSTATUS (1U << 31)
+#define FW_PORT_CMD_LSPEED(x) ((x) << 24)
+#define FW_PORT_CMD_LSPEED_GET(x) (((x) >> 24) & 0x3f)
+#define FW_PORT_CMD_TXPAUSE (1U << 23)
+#define FW_PORT_CMD_RXPAUSE (1U << 22)
+#define FW_PORT_CMD_MDIOCAP (1U << 21)
+#define FW_PORT_CMD_MDIOADDR_GET(x) (((x) >> 16) & 0x1f)
+#define FW_PORT_CMD_LPTXPAUSE (1U << 15)
+#define FW_PORT_CMD_LPRXPAUSE (1U << 14)
+#define FW_PORT_CMD_PTYPE_MASK 0x1f
+#define FW_PORT_CMD_PTYPE_GET(x) (((x) >> 8) & FW_PORT_CMD_PTYPE_MASK)
+#define FW_PORT_CMD_MODTYPE_MASK 0x1f
+#define FW_PORT_CMD_MODTYPE_GET(x) (((x) >> 0) & FW_PORT_CMD_MODTYPE_MASK)
+
+#define FW_PORT_CMD_PPPEN(x) ((x) << 31)
+#define FW_PORT_CMD_TPSRC(x) ((x) << 28)
+#define FW_PORT_CMD_NCSISRC(x) ((x) << 24)
+
+#define FW_PORT_CMD_CH0(x) ((x) << 20)
+#define FW_PORT_CMD_CH1(x) ((x) << 16)
+#define FW_PORT_CMD_CH2(x) ((x) << 12)
+#define FW_PORT_CMD_CH3(x) ((x) << 8)
+#define FW_PORT_CMD_NCSICH(x) ((x) << 4)
+
+enum fw_port_type {
+	FW_PORT_TYPE_FIBER,
+	FW_PORT_TYPE_KX4,
+	FW_PORT_TYPE_BT_SGMII,
+	FW_PORT_TYPE_KX,
+	FW_PORT_TYPE_BT_XAUI,
+	FW_PORT_TYPE_KR,
+	FW_PORT_TYPE_CX4,
+	FW_PORT_TYPE_TWINAX,
+
+	FW_PORT_TYPE_NONE = FW_PORT_CMD_PTYPE_MASK
+};
+
+enum fw_port_module_type {
+	FW_PORT_MOD_TYPE_NA,
+	FW_PORT_MOD_TYPE_LR,
+	FW_PORT_MOD_TYPE_SR,
+	FW_PORT_MOD_TYPE_ER,
+
+	FW_PORT_MOD_TYPE_NONE = FW_PORT_CMD_MODTYPE_MASK
+};
+
+/* port stats */
+#define FW_NUM_PORT_STATS 50
+#define FW_NUM_PORT_TX_STATS 23
+#define FW_NUM_PORT_RX_STATS 27
+
+enum fw_port_stats_tx_index {
+	FW_STAT_TX_PORT_BYTES_IX,
+	FW_STAT_TX_PORT_FRAMES_IX,
+	FW_STAT_TX_PORT_BCAST_IX,
+	FW_STAT_TX_PORT_MCAST_IX,
+	FW_STAT_TX_PORT_UCAST_IX,
+	FW_STAT_TX_PORT_ERROR_IX,
+	FW_STAT_TX_PORT_64B_IX,
+	FW_STAT_TX_PORT_65B_127B_IX,
+	FW_STAT_TX_PORT_128B_255B_IX,
+	FW_STAT_TX_PORT_256B_511B_IX,
+	FW_STAT_TX_PORT_512B_1023B_IX,
+	FW_STAT_TX_PORT_1024B_1518B_IX,
+	FW_STAT_TX_PORT_1519B_MAX_IX,
+	FW_STAT_TX_PORT_DROP_IX,
+	FW_STAT_TX_PORT_PAUSE_IX,
+	FW_STAT_TX_PORT_PPP0_IX,
+	FW_STAT_TX_PORT_PPP1_IX,
+	FW_STAT_TX_PORT_PPP2_IX,
+	FW_STAT_TX_PORT_PPP3_IX,
+	FW_STAT_TX_PORT_PPP4_IX,
+	FW_STAT_TX_PORT_PPP5_IX,
+	FW_STAT_TX_PORT_PPP6_IX,
+	FW_STAT_TX_PORT_PPP7_IX
+};
+
+enum fw_port_stat_rx_index {
+	FW_STAT_RX_PORT_BYTES_IX,
+	FW_STAT_RX_PORT_FRAMES_IX,
+	FW_STAT_RX_PORT_BCAST_IX,
+	FW_STAT_RX_PORT_MCAST_IX,
+	FW_STAT_RX_PORT_UCAST_IX,
+	FW_STAT_RX_PORT_MTU_ERROR_IX,
+	FW_STAT_RX_PORT_MTU_CRC_ERROR_IX,
+	FW_STAT_RX_PORT_CRC_ERROR_IX,
+	FW_STAT_RX_PORT_LEN_ERROR_IX,
+	FW_STAT_RX_PORT_SYM_ERROR_IX,
+	FW_STAT_RX_PORT_64B_IX,
+	FW_STAT_RX_PORT_65B_127B_IX,
+	FW_STAT_RX_PORT_128B_255B_IX,
+	FW_STAT_RX_PORT_256B_511B_IX,
+	FW_STAT_RX_PORT_512B_1023B_IX,
+	FW_STAT_RX_PORT_1024B_1518B_IX,
+	FW_STAT_RX_PORT_1519B_MAX_IX,
+	FW_STAT_RX_PORT_PAUSE_IX,
+	FW_STAT_RX_PORT_PPP0_IX,
+	FW_STAT_RX_PORT_PPP1_IX,
+	FW_STAT_RX_PORT_PPP2_IX,
+	FW_STAT_RX_PORT_PPP3_IX,
+	FW_STAT_RX_PORT_PPP4_IX,
+	FW_STAT_RX_PORT_PPP5_IX,
+	FW_STAT_RX_PORT_PPP6_IX,
+	FW_STAT_RX_PORT_PPP7_IX,
+	FW_STAT_RX_PORT_LESS_64B_IX
+};
+
+struct fw_port_stats_cmd {
+	__be32 op_to_portid;
+	__be32 retval_len16;
+	union fw_port_stats {
+		struct fw_port_stats_ctl {
+			u8 nstats_bg_bm;
+			u8 tx_ix;
+			__be16 r6;
+			__be32 r7;
+			__be64 stat0;
+			__be64 stat1;
+			__be64 stat2;
+			__be64 stat3;
+			__be64 stat4;
+			__be64 stat5;
+		} ctl;
+		struct fw_port_stats_all {
+			__be64 tx_bytes;
+			__be64 tx_frames;
+			__be64 tx_bcast;
+			__be64 tx_mcast;
+			__be64 tx_ucast;
+			__be64 tx_error;
+			__be64 tx_64b;
+			__be64 tx_65b_127b;
+			__be64 tx_128b_255b;
+			__be64 tx_256b_511b;
+			__be64 tx_512b_1023b;
+			__be64 tx_1024b_1518b;
+			__be64 tx_1519b_max;
+			__be64 tx_drop;
+			__be64 tx_pause;
+			__be64 tx_ppp0;
+			__be64 tx_ppp1;
+			__be64 tx_ppp2;
+			__be64 tx_ppp3;
+			__be64 tx_ppp4;
+			__be64 tx_ppp5;
+			__be64 tx_ppp6;
+			__be64 tx_ppp7;
+			__be64 rx_bytes;
+			__be64 rx_frames;
+			__be64 rx_bcast;
+			__be64 rx_mcast;
+			__be64 rx_ucast;
+			__be64 rx_mtu_error;
+			__be64 rx_mtu_crc_error;
+			__be64 rx_crc_error;
+			__be64 rx_len_error;
+			__be64 rx_sym_error;
+			__be64 rx_64b;
+			__be64 rx_65b_127b;
+			__be64 rx_128b_255b;
+			__be64 rx_256b_511b;
+			__be64 rx_512b_1023b;
+			__be64 rx_1024b_1518b;
+			__be64 rx_1519b_max;
+			__be64 rx_pause;
+			__be64 rx_ppp0;
+			__be64 rx_ppp1;
+			__be64 rx_ppp2;
+			__be64 rx_ppp3;
+			__be64 rx_ppp4;
+			__be64 rx_ppp5;
+			__be64 rx_ppp6;
+			__be64 rx_ppp7;
+			__be64 rx_less_64b;
+			__be64 rx_bg_drop;
+			__be64 rx_bg_trunc;
+		} all;
+	} u;
+};
+
+#define FW_PORT_STATS_CMD_NSTATS(x) ((x) << 4)
+#define FW_PORT_STATS_CMD_BG_BM(x) ((x) << 0)
+#define FW_PORT_STATS_CMD_TX(x) ((x) << 7)
+#define FW_PORT_STATS_CMD_IX(x) ((x) << 0)
+
+/* port loopback stats */
+#define FW_NUM_LB_STATS 16
+enum fw_port_lb_stats_index {
+	FW_STAT_LB_PORT_BYTES_IX,
+	FW_STAT_LB_PORT_FRAMES_IX,
+	FW_STAT_LB_PORT_BCAST_IX,
+	FW_STAT_LB_PORT_MCAST_IX,
+	FW_STAT_LB_PORT_UCAST_IX,
+	FW_STAT_LB_PORT_ERROR_IX,
+	FW_STAT_LB_PORT_64B_IX,
+	FW_STAT_LB_PORT_65B_127B_IX,
+	FW_STAT_LB_PORT_128B_255B_IX,
+	FW_STAT_LB_PORT_256B_511B_IX,
+	FW_STAT_LB_PORT_512B_1023B_IX,
+	FW_STAT_LB_PORT_1024B_1518B_IX,
+	FW_STAT_LB_PORT_1519B_MAX_IX,
+	FW_STAT_LB_PORT_DROP_FRAMES_IX
+};
+
+struct fw_port_lb_stats_cmd {
+	__be32 op_to_lbport;
+	__be32 retval_len16;
+	union fw_port_lb_stats {
+		struct fw_port_lb_stats_ctl {
+			u8 nstats_bg_bm;
+			u8 ix_pkd;
+			__be16 r6;
+			__be32 r7;
+			__be64 stat0;
+			__be64 stat1;
+			__be64 stat2;
+			__be64 stat3;
+			__be64 stat4;
+			__be64 stat5;
+		} ctl;
+		struct fw_port_lb_stats_all {
+			__be64 tx_bytes;
+			__be64 tx_frames;
+			__be64 tx_bcast;
+			__be64 tx_mcast;
+			__be64 tx_ucast;
+			__be64 tx_error;
+			__be64 tx_64b;
+			__be64 tx_65b_127b;
+			__be64 tx_128b_255b;
+			__be64 tx_256b_511b;
+			__be64 tx_512b_1023b;
+			__be64 tx_1024b_1518b;
+			__be64 tx_1519b_max;
+			__be64 rx_lb_drop;
+			__be64 rx_lb_trunc;
+		} all;
+	} u;
+};
+
+#define FW_PORT_LB_STATS_CMD_LBPORT(x) ((x) << 0)
+#define FW_PORT_LB_STATS_CMD_NSTATS(x) ((x) << 4)
+#define FW_PORT_LB_STATS_CMD_BG_BM(x) ((x) << 0)
+#define FW_PORT_LB_STATS_CMD_IX(x) ((x) << 0)
+
+struct fw_rss_ind_tbl_cmd {
+	__be32 op_to_viid;
+#define FW_RSS_IND_TBL_CMD_VIID(x) ((x) << 0)
+	__be32 retval_len16;
+	__be16 niqid;
+	__be16 startidx;
+	__be32 r3;
+	__be32 iq0_to_iq2;
+#define FW_RSS_IND_TBL_CMD_IQ0(x) ((x) << 20)
+#define FW_RSS_IND_TBL_CMD_IQ1(x) ((x) << 10)
+#define FW_RSS_IND_TBL_CMD_IQ2(x) ((x) << 0)
+	__be32 iq3_to_iq5;
+	__be32 iq6_to_iq8;
+	__be32 iq9_to_iq11;
+	__be32 iq12_to_iq14;
+	__be32 iq15_to_iq17;
+	__be32 iq18_to_iq20;
+	__be32 iq21_to_iq23;
+	__be32 iq24_to_iq26;
+	__be32 iq27_to_iq29;
+	__be32 iq30_iq31;
+	__be32 r15_lo;
+};
+
+struct fw_rss_glb_config_cmd {
+	__be32 op_to_write;
+	__be32 retval_len16;
+	union fw_rss_glb_config {
+		struct fw_rss_glb_config_manual {
+			__be32 mode_pkd;
+			__be32 r3;
+			__be64 r4;
+			__be64 r5;
+		} manual;
+		struct fw_rss_glb_config_basicvirtual {
+			__be32 mode_pkd;
+			__be32 synmapen_to_hashtoeplitz;
+#define FW_RSS_GLB_CONFIG_CMD_SYNMAPEN      (1U << 8)
+#define FW_RSS_GLB_CONFIG_CMD_SYN4TUPENIPV6 (1U << 7)
+#define FW_RSS_GLB_CONFIG_CMD_SYN2TUPENIPV6 (1U << 6)
+#define FW_RSS_GLB_CONFIG_CMD_SYN4TUPENIPV4 (1U << 5)
+#define FW_RSS_GLB_CONFIG_CMD_SYN2TUPENIPV4 (1U << 4)
+#define FW_RSS_GLB_CONFIG_CMD_OFDMAPEN      (1U << 3)
+#define FW_RSS_GLB_CONFIG_CMD_TNLMAPEN      (1U << 2)
+#define FW_RSS_GLB_CONFIG_CMD_TNLALLLKP     (1U << 1)
+#define FW_RSS_GLB_CONFIG_CMD_HASHTOEPLITZ  (1U << 0)
+			__be64 r8;
+			__be64 r9;
+		} basicvirtual;
+	} u;
+};
+
+#define FW_RSS_GLB_CONFIG_CMD_MODE(x)	((x) << 28)
+
+#define FW_RSS_GLB_CONFIG_CMD_MODE_MANUAL	0
+#define FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL	1
+
+struct fw_rss_vi_config_cmd {
+	__be32 op_to_viid;
+#define FW_RSS_VI_CONFIG_CMD_VIID(x) ((x) << 0)
+	__be32 retval_len16;
+	union fw_rss_vi_config {
+		struct fw_rss_vi_config_manual {
+			__be64 r3;
+			__be64 r4;
+			__be64 r5;
+		} manual;
+		struct fw_rss_vi_config_basicvirtual {
+			__be32 r6;
+			__be32 defaultq_to_ip4udpen;
+#define FW_RSS_VI_CONFIG_CMD_DEFAULTQ(x)  ((x) << 16)
+#define FW_RSS_VI_CONFIG_CMD_IP6FOURTUPEN (1U << 4)
+#define FW_RSS_VI_CONFIG_CMD_IP6TWOTUPEN  (1U << 3)
+#define FW_RSS_VI_CONFIG_CMD_IP4FOURTUPEN (1U << 2)
+#define FW_RSS_VI_CONFIG_CMD_IP4TWOTUPEN  (1U << 1)
+#define FW_RSS_VI_CONFIG_CMD_IP4UDPEN     (1U << 0)
+			__be64 r9;
+			__be64 r10;
+		} basicvirtual;
+	} u;
+};
+
+enum fw_error_type {
+	FW_ERROR_TYPE_EXCEPTION		= 0x0,
+	FW_ERROR_TYPE_HWMODULE		= 0x1,
+	FW_ERROR_TYPE_WR		= 0x2,
+	FW_ERROR_TYPE_ACL		= 0x3,
+};
+
+struct fw_error_cmd {
+	__be32 op_to_type;
+	__be32 len16_pkd;
+	union fw_error {
+		struct fw_error_exception {
+			__be32 info[6];
+		} exception;
+		struct fw_error_hwmodule {
+			__be32 regaddr;
+			__be32 regval;
+		} hwmodule;
+		struct fw_error_wr {
+			__be16 cidx;
+			__be16 pfn_vfn;
+			__be32 eqid;
+			u8 wrhdr[16];
+		} wr;
+		struct fw_error_acl {
+			__be16 cidx;
+			__be16 pfn_vfn;
+			__be32 eqid;
+			__be16 mv_pkd;
+			u8 val[6];
+			__be64 r4;
+		} acl;
+	} u;
+};
+
+struct fw_debug_cmd {
+	__be32 op_type;
+#define FW_DEBUG_CMD_TYPE_GET(x) ((x) & 0xff)
+	__be32 len16_pkd;
+	union fw_debug {
+		struct fw_debug_assert {
+			__be32 fcid;
+			__be32 line;
+			__be32 x;
+			__be32 y;
+			u8 filename_0_7[8];
+			u8 filename_8_15[8];
+			__be64 r3;
+		} assert;
+		struct fw_debug_prt {
+			__be16 dprtstridx;
+			__be16 r3[3];
+			__be32 dprtstrparam0;
+			__be32 dprtstrparam1;
+			__be32 dprtstrparam2;
+			__be32 dprtstrparam3;
+		} prt;
+	} u;
+};
+
+struct fw_hdr {
+	u8 ver;
+	u8 reserved1;
+	__be16	len512;			/* bin length in units of 512-bytes */
+	__be32	fw_ver;			/* firmware version */
+	__be32	tp_microcode_ver;
+	u8 intfver_nic;
+	u8 intfver_vnic;
+	u8 intfver_ofld;
+	u8 intfver_ri;
+	u8 intfver_iscsipdu;
+	u8 intfver_iscsi;
+	u8 intfver_fcoe;
+	u8 reserved2;
+	__be32  reserved3[27];
+};
+
+#define FW_HDR_FW_VER_MAJOR_GET(x) (((x) >> 24) & 0xff)
+#define FW_HDR_FW_VER_MINOR_GET(x) (((x) >> 16) & 0xff)
+#define FW_HDR_FW_VER_MICRO_GET(x) (((x) >> 8) & 0xff)
+#define FW_HDR_FW_VER_BUILD_GET(x) (((x) >> 0) & 0xff)
+#endif /* _T4FW_INTERFACE_H_ */
diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c
index 669de028d44f..080d1cea5b26 100644
--- a/drivers/net/gianfar.c
+++ b/drivers/net/gianfar.c
@@ -676,7 +676,7 @@ static int gfar_of_init(struct of_device *ofdev, struct net_device **pdev)
 		priv->rx_queue[i] = NULL;
 
 	for (i = 0; i < priv->num_tx_queues; i++) {
-		priv->tx_queue[i] =  (struct gfar_priv_tx_q *)kmalloc(
+		priv->tx_queue[i] =  (struct gfar_priv_tx_q *)kzalloc(
 				sizeof (struct gfar_priv_tx_q), GFP_KERNEL);
 		if (!priv->tx_queue[i]) {
 			err = -ENOMEM;
@@ -689,7 +689,7 @@ static int gfar_of_init(struct of_device *ofdev, struct net_device **pdev)
 	}
 
 	for (i = 0; i < priv->num_rx_queues; i++) {
-		priv->rx_queue[i] = (struct gfar_priv_rx_q *)kmalloc(
+		priv->rx_queue[i] = (struct gfar_priv_rx_q *)kzalloc(
 					sizeof (struct gfar_priv_rx_q), GFP_KERNEL);
 		if (!priv->rx_queue[i]) {
 			err = -ENOMEM;
@@ -1120,10 +1120,10 @@ static int gfar_probe(struct of_device *ofdev,
 	/* provided which set of benchmarks. */
 	printk(KERN_INFO "%s: Running with NAPI enabled\n", dev->name);
 	for (i = 0; i < priv->num_rx_queues; i++)
-		printk(KERN_INFO "%s: :RX BD ring size for Q[%d]: %d\n",
+		printk(KERN_INFO "%s: RX BD ring size for Q[%d]: %d\n",
 			dev->name, i, priv->rx_queue[i]->rx_ring_size);
 	for(i = 0; i < priv->num_tx_queues; i++)
-		 printk(KERN_INFO "%s:TX BD ring size for Q[%d]: %d\n",
+		 printk(KERN_INFO "%s: TX BD ring size for Q[%d]: %d\n",
 			dev->name, i, priv->tx_queue[i]->tx_ring_size);
 
 	return 0;
@@ -1638,13 +1638,13 @@ static void free_skb_resources(struct gfar_private *priv)
 	/* Go through all the buffer descriptors and free their data buffers */
 	for (i = 0; i < priv->num_tx_queues; i++) {
 		tx_queue = priv->tx_queue[i];
-		if(!tx_queue->tx_skbuff)
+		if(tx_queue->tx_skbuff)
 			free_skb_tx_queue(tx_queue);
 	}
 
 	for (i = 0; i < priv->num_rx_queues; i++) {
 		rx_queue = priv->rx_queue[i];
-		if(!rx_queue->rx_skbuff)
+		if(rx_queue->rx_skbuff)
 			free_skb_rx_queue(rx_queue);
 	}
 
diff --git a/drivers/net/pcmcia/smc91c92_cs.c b/drivers/net/pcmcia/smc91c92_cs.c
index 5adc662c4bfb..ff7eb9116b6a 100644
--- a/drivers/net/pcmcia/smc91c92_cs.c
+++ b/drivers/net/pcmcia/smc91c92_cs.c
@@ -493,13 +493,14 @@ static int pcmcia_get_versmac(struct pcmcia_device *p_dev,
 {
 	struct net_device *dev = priv;
 	cisparse_t parse;
+	u8 *buf;
 
 	if (pcmcia_parse_tuple(tuple, &parse))
 		return -EINVAL;
 
-	if ((parse.version_1.ns > 3) &&
-	    (cvt_ascii_address(dev,
-			       (parse.version_1.str + parse.version_1.ofs[3]))))
+	buf = parse.version_1.str + parse.version_1.ofs[3];
+
+	if ((parse.version_1.ns > 3) && (cvt_ascii_address(dev, buf) == 0))
 		return 0;
 
 	return -EINVAL;
@@ -528,7 +529,7 @@ static int mhz_setup(struct pcmcia_device *link)
     len = pcmcia_get_tuple(link, 0x81, &buf);
     if (buf && len >= 13) {
 	    buf[12] = '\0';
-	    if (cvt_ascii_address(dev, buf))
+	    if (cvt_ascii_address(dev, buf) == 0)
 		    rc = 0;
     }
     kfree(buf);
@@ -910,7 +911,7 @@ static int smc91c92_config(struct pcmcia_device *link)
 
     if (i != 0) {
 	printk(KERN_NOTICE "smc91c92_cs: Unable to find hardware address.\n");
-	goto config_undo;
+	goto config_failed;
     }
 
     smc->duplex = 0;
@@ -998,6 +999,7 @@ config_undo:
     unregister_netdev(dev);
 config_failed:
     smc91c92_release(link);
+    free_netdev(dev);
     return -ENODEV;
 } /* smc91c92_config */
 
diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c
index 96740051cdcc..dbb1f5a1824c 100644
--- a/drivers/net/r8169.c
+++ b/drivers/net/r8169.c
@@ -3227,8 +3227,8 @@ static void rtl8169_set_rxbufsize(struct rtl8169_private *tp,
 	unsigned int max_frame = mtu + VLAN_ETH_HLEN + ETH_FCS_LEN;
 
 	if (max_frame != 16383)
-		printk(KERN_WARNING "WARNING! Changing of MTU on this NIC"
-			"May lead to frame reception errors!\n");
+		printk(KERN_WARNING PFX "WARNING! Changing of MTU on this "
+			"NIC may lead to frame reception errors!\n");
 
 	tp->rx_buf_sz = (max_frame > RX_BUF_SIZE) ? max_frame : RX_BUF_SIZE;
 }
diff --git a/drivers/net/sgiseeq.c b/drivers/net/sgiseeq.c
index beb537dbe9a9..c8fc896fc460 100644
--- a/drivers/net/sgiseeq.c
+++ b/drivers/net/sgiseeq.c
@@ -593,8 +593,10 @@ static int sgiseeq_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	/* Setup... */
 	len = skb->len;
 	if (len < ETH_ZLEN) {
-		if (skb_padto(skb, ETH_ZLEN))
+		if (skb_padto(skb, ETH_ZLEN)) {
+			spin_unlock_irqrestore(&sp->tx_lock, flags);
 			return NETDEV_TX_OK;
+		}
 		len = ETH_ZLEN;
 	}
 
diff --git a/drivers/net/stmmac/Kconfig b/drivers/net/stmmac/Kconfig
index fb287649a305..eb63d44748a7 100644
--- a/drivers/net/stmmac/Kconfig
+++ b/drivers/net/stmmac/Kconfig
@@ -2,6 +2,7 @@ config STMMAC_ETH
 	tristate "STMicroelectronics 10/100/1000 Ethernet driver"
 	select MII
 	select PHYLIB
+	select CRC32
 	depends on NETDEVICES && CPU_SUBTYPE_ST40
 	help
 	  This is the driver for the Ethernet IPs are built around a
diff --git a/drivers/net/wireless/ath/ar9170/usb.c b/drivers/net/wireless/ath/ar9170/usb.c
index 0b0d2dc2f38c..99a6da464bd3 100644
--- a/drivers/net/wireless/ath/ar9170/usb.c
+++ b/drivers/net/wireless/ath/ar9170/usb.c
@@ -95,6 +95,8 @@ static struct usb_device_id ar9170_usb_ids[] = {
 	{ USB_DEVICE(0x04bb, 0x093f) },
 	/* AVM FRITZ!WLAN USB Stick N */
 	{ USB_DEVICE(0x057C, 0x8401) },
+	/* NEC WL300NU-G */
+	{ USB_DEVICE(0x0409, 0x0249) },
 	/* AVM FRITZ!WLAN USB Stick N 2.4 */
 	{ USB_DEVICE(0x057C, 0x8402), .driver_info = AR9170_REQ_FW1_ONLY },
 
@@ -417,7 +419,7 @@ static int ar9170_usb_exec_cmd(struct ar9170 *ar, enum ar9170_cmd cmd,
 	spin_unlock_irqrestore(&aru->common.cmdlock, flags);
 
 	usb_fill_int_urb(urb, aru->udev,
-			 usb_sndbulkpipe(aru->udev, AR9170_EP_CMD),
+			 usb_sndintpipe(aru->udev, AR9170_EP_CMD),
 			 aru->common.cmdbuf, plen + 4,
 			 ar9170_usb_tx_urb_complete, NULL, 1);
 
diff --git a/drivers/net/wireless/iwlwifi/iwl-4965.c b/drivers/net/wireless/iwlwifi/iwl-4965.c
index 1bd2cd836026..83c52a682622 100644
--- a/drivers/net/wireless/iwlwifi/iwl-4965.c
+++ b/drivers/net/wireless/iwlwifi/iwl-4965.c
@@ -2041,16 +2041,14 @@ static void iwl4965_rx_reply_tx(struct iwl_priv *priv,
 				   tx_resp->failure_frame);
 
 		freed = iwl_tx_queue_reclaim(priv, txq_id, index);
-		if (qc && likely(sta_id != IWL_INVALID_STATION))
-			priv->stations[sta_id].tid[tid].tfds_in_queue -= freed;
+		iwl_free_tfds_in_queue(priv, sta_id, tid, freed);
 
 		if (priv->mac80211_registered &&
 		    (iwl_queue_space(&txq->q) > txq->q.low_mark))
 			iwl_wake_queue(priv, txq_id);
 	}
 
-	if (qc && likely(sta_id != IWL_INVALID_STATION))
-		iwl_txq_check_empty(priv, sta_id, tid, txq_id);
+	iwl_txq_check_empty(priv, sta_id, tid, txq_id);
 
 	if (iwl_check_bits(status, TX_ABORT_REQUIRED_MSK))
 		IWL_ERR(priv, "TODO:  Implement Tx ABORT REQUIRED!!!\n");
diff --git a/drivers/net/wireless/iwlwifi/iwl-agn.c b/drivers/net/wireless/iwlwifi/iwl-agn.c
index 5e0c6bf3fbb1..8b8e3e1cbb44 100644
--- a/drivers/net/wireless/iwlwifi/iwl-agn.c
+++ b/drivers/net/wireless/iwlwifi/iwl-agn.c
@@ -1259,7 +1259,15 @@ static void iwl_irq_tasklet(struct iwl_priv *priv)
 	/* Ack/clear/reset pending uCode interrupts.
 	 * Note:  Some bits in CSR_INT are "OR" of bits in CSR_FH_INT_STATUS,
 	 */
-	iwl_write32(priv, CSR_INT, priv->inta);
+	/* There is a hardware bug in the interrupt mask function that some
+	 * interrupts (i.e. CSR_INT_BIT_SCD) can still be generated even if
+	 * they are disabled in the CSR_INT_MASK register. Furthermore the
+	 * ICT interrupt handling mechanism has another bug that might cause
+	 * these unmasked interrupts fail to be detected. We workaround the
+	 * hardware bugs here by ACKing all the possible interrupts so that
+	 * interrupt coalescing can still be achieved.
+	 */
+	iwl_write32(priv, CSR_INT, priv->inta | ~priv->inta_mask);
 
 	inta = priv->inta;
 
@@ -2645,7 +2653,7 @@ static int iwl_mac_setup_register(struct iwl_priv *priv)
 		BIT(NL80211_IFTYPE_STATION) |
 		BIT(NL80211_IFTYPE_ADHOC);
 
-	hw->wiphy->flags |= WIPHY_FLAG_STRICT_REGULATORY |
+	hw->wiphy->flags |= WIPHY_FLAG_CUSTOM_REGULATORY |
 			    WIPHY_FLAG_DISABLE_BEACON_HINTS;
 
 	/*
diff --git a/drivers/net/wireless/iwlwifi/iwl3945-base.c b/drivers/net/wireless/iwlwifi/iwl3945-base.c
index 1eaa0052c11b..b55e4f39a9e1 100644
--- a/drivers/net/wireless/iwlwifi/iwl3945-base.c
+++ b/drivers/net/wireless/iwlwifi/iwl3945-base.c
@@ -1956,7 +1956,7 @@ static void iwl3945_init_hw_rates(struct iwl_priv *priv,
 {
 	int i;
 
-	for (i = 0; i < IWL_RATE_COUNT; i++) {
+	for (i = 0; i < IWL_RATE_COUNT_LEGACY; i++) {
 		rates[i].bitrate = iwl3945_rates[i].ieee * 5;
 		rates[i].hw_value = i; /* Rate scaling will work on indexes */
 		rates[i].hw_value_short = i;
@@ -3922,7 +3922,7 @@ static int iwl3945_setup_mac(struct iwl_priv *priv)
 		BIT(NL80211_IFTYPE_STATION) |
 		BIT(NL80211_IFTYPE_ADHOC);
 
-	hw->wiphy->flags |= WIPHY_FLAG_STRICT_REGULATORY |
+	hw->wiphy->flags |= WIPHY_FLAG_CUSTOM_REGULATORY |
 			    WIPHY_FLAG_DISABLE_BEACON_HINTS;
 
 	hw->wiphy->max_scan_ssids = PROBE_OPTION_MAX_3945;
diff --git a/drivers/net/wireless/libertas/cfg.c b/drivers/net/wireless/libertas/cfg.c
index e196b84914db..ce7bec402a33 100644
--- a/drivers/net/wireless/libertas/cfg.c
+++ b/drivers/net/wireless/libertas/cfg.c
@@ -173,6 +173,8 @@ int lbs_cfg_register(struct lbs_private *priv)
 	if (ret < 0)
 		lbs_pr_err("cannot register wiphy device\n");
 
+	priv->wiphy_registered = true;
+
 	ret = register_netdev(priv->dev);
 	if (ret)
 		lbs_pr_err("cannot register network device\n");
@@ -191,9 +193,11 @@ void lbs_cfg_free(struct lbs_private *priv)
 	if (!wdev)
 		return;
 
-	if (wdev->wiphy) {
+	if (priv->wiphy_registered)
 		wiphy_unregister(wdev->wiphy);
+
+	if (wdev->wiphy)
 		wiphy_free(wdev->wiphy);
-	}
+
 	kfree(wdev);
 }
diff --git a/drivers/net/wireless/libertas/dev.h b/drivers/net/wireless/libertas/dev.h
index 6977ee820214..6875e1498bd5 100644
--- a/drivers/net/wireless/libertas/dev.h
+++ b/drivers/net/wireless/libertas/dev.h
@@ -36,6 +36,7 @@ struct lbs_private {
 
 	/* CFG80211 */
 	struct wireless_dev *wdev;
+	bool wiphy_registered;
 
 	/* Mesh */
 	struct net_device *mesh_dev; /* Virtual device */
diff --git a/drivers/net/wireless/mwl8k.c b/drivers/net/wireless/mwl8k.c
index 89354c29f088..12fdcb25fd38 100644
--- a/drivers/net/wireless/mwl8k.c
+++ b/drivers/net/wireless/mwl8k.c
@@ -3852,6 +3852,7 @@ MODULE_FIRMWARE("mwl8k/helper_8366.fw");
 MODULE_FIRMWARE("mwl8k/fmimage_8366.fw");
 
 static DEFINE_PCI_DEVICE_TABLE(mwl8k_pci_id_table) = {
+	{ PCI_VDEVICE(MARVELL, 0x2a0a), .driver_data = MWL8363, },
 	{ PCI_VDEVICE(MARVELL, 0x2a0c), .driver_data = MWL8363, },
 	{ PCI_VDEVICE(MARVELL, 0x2a24), .driver_data = MWL8363, },
 	{ PCI_VDEVICE(MARVELL, 0x2a2b), .driver_data = MWL8687, },
diff --git a/drivers/net/wireless/p54/p54usb.c b/drivers/net/wireless/p54/p54usb.c
index 762952d688e2..743a6c68b29d 100644
--- a/drivers/net/wireless/p54/p54usb.c
+++ b/drivers/net/wireless/p54/p54usb.c
@@ -36,6 +36,7 @@ MODULE_FIRMWARE("isl3887usb");
 static struct usb_device_id p54u_table[] __devinitdata = {
 	/* Version 1 devices (pci chip + net2280) */
 	{USB_DEVICE(0x0506, 0x0a11)},	/* 3COM 3CRWE254G72 */
+	{USB_DEVICE(0x06b9, 0x0120)},	/* Thomson SpeedTouch 120g */
 	{USB_DEVICE(0x0707, 0xee06)},	/* SMC 2862W-G */
 	{USB_DEVICE(0x07aa, 0x001c)},	/* Corega CG-WLUSB2GT */
 	{USB_DEVICE(0x083a, 0x4501)},	/* Accton 802.11g WN4501 USB */
diff --git a/drivers/net/wireless/rt2x00/rt2500usb.c b/drivers/net/wireless/rt2x00/rt2500usb.c
index d2cc4458477f..8ebb705fe106 100644
--- a/drivers/net/wireless/rt2x00/rt2500usb.c
+++ b/drivers/net/wireless/rt2x00/rt2500usb.c
@@ -1644,6 +1644,11 @@ static int rt2500usb_probe_hw_mode(struct rt2x00_dev *rt2x00dev)
 	unsigned int i;
 
 	/*
+	 * Disable powersaving as default.
+	 */
+	rt2x00dev->hw->wiphy->flags &= ~WIPHY_FLAG_PS_ON_BY_DEFAULT;
+
+	/*
 	 * Initialize all hw fields.
 	 */
 	rt2x00dev->hw->flags =
diff --git a/drivers/net/wireless/rt2x00/rt2800lib.c b/drivers/net/wireless/rt2x00/rt2800lib.c
index 58c7f218019d..c015ce9fdd09 100644
--- a/drivers/net/wireless/rt2x00/rt2800lib.c
+++ b/drivers/net/wireless/rt2x00/rt2800lib.c
@@ -813,9 +813,9 @@ static void rt2800_config_channel_rt3x(struct rt2x00_dev *rt2x00dev,
 	rt2800_rfcsr_write(rt2x00dev, 24,
 			      rt2x00dev->calibration[conf_is_ht40(conf)]);
 
-	rt2800_rfcsr_read(rt2x00dev, 23, &rfcsr);
+	rt2800_rfcsr_read(rt2x00dev, 7, &rfcsr);
 	rt2x00_set_field8(&rfcsr, RFCSR7_RF_TUNING, 1);
-	rt2800_rfcsr_write(rt2x00dev, 23, rfcsr);
+	rt2800_rfcsr_write(rt2x00dev, 7, rfcsr);
 }
 
 static void rt2800_config_channel(struct rt2x00_dev *rt2x00dev,
diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c
index 7794a2e2adce..99d68c34e4f1 100644
--- a/net/bluetooth/l2cap.c
+++ b/net/bluetooth/l2cap.c
@@ -1002,7 +1002,8 @@ static int l2cap_sock_connect(struct socket *sock, struct sockaddr *addr, int al
 
 	BT_DBG("sk %p", sk);
 
-	if (!addr || addr->sa_family != AF_BLUETOOTH)
+	if (!addr || alen < sizeof(addr->sa_family) ||
+	    addr->sa_family != AF_BLUETOOTH)
 		return -EINVAL;
 
 	memset(&la, 0, sizeof(la));
diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c
index 7f439765403d..8ed3c37684fa 100644
--- a/net/bluetooth/rfcomm/sock.c
+++ b/net/bluetooth/rfcomm/sock.c
@@ -397,7 +397,8 @@ static int rfcomm_sock_connect(struct socket *sock, struct sockaddr *addr, int a
 
 	BT_DBG("sk %p", sk);
 
-	if (addr->sa_family != AF_BLUETOOTH || alen < sizeof(struct sockaddr_rc))
+	if (alen < sizeof(struct sockaddr_rc) ||
+	    addr->sa_family != AF_BLUETOOTH)
 		return -EINVAL;
 
 	lock_sock(sk);
diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c
index e5b16b76b22e..ca6b2ad1c3fc 100644
--- a/net/bluetooth/sco.c
+++ b/net/bluetooth/sco.c
@@ -499,7 +499,8 @@ static int sco_sock_connect(struct socket *sock, struct sockaddr *addr, int alen
 
 	BT_DBG("sk %p", sk);
 
-	if (addr->sa_family != AF_BLUETOOTH || alen < sizeof(struct sockaddr_sco))
+	if (alen < sizeof(struct sockaddr_sco) ||
+	    addr->sa_family != AF_BLUETOOTH)
 		return -EINVAL;
 
 	if (sk->sk_state != BT_OPEN && sk->sk_state != BT_BOUND)
diff --git a/net/can/bcm.c b/net/can/bcm.c
index a2dee522b43e..907dc871fac8 100644
--- a/net/can/bcm.c
+++ b/net/can/bcm.c
@@ -1479,6 +1479,9 @@ static int bcm_connect(struct socket *sock, struct sockaddr *uaddr, int len,
 	struct sock *sk = sock->sk;
 	struct bcm_sock *bo = bcm_sk(sk);
 
+	if (len < sizeof(*addr))
+		return -EINVAL;
+
 	if (bo->bound)
 		return -EISCONN;
 
diff --git a/net/ieee802154/af_ieee802154.c b/net/ieee802154/af_ieee802154.c
index 79886d546bab..c7da600750bb 100644
--- a/net/ieee802154/af_ieee802154.c
+++ b/net/ieee802154/af_ieee802154.c
@@ -127,6 +127,9 @@ static int ieee802154_sock_connect(struct socket *sock, struct sockaddr *uaddr,
 {
 	struct sock *sk = sock->sk;
 
+	if (addr_len < sizeof(uaddr->sa_family))
+		return -EINVAL;
+
 	if (uaddr->sa_family == AF_UNSPEC)
 		return sk->sk_prot->disconnect(sk, flags);
 
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 2ed85714540f..f71357422380 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -531,6 +531,8 @@ int inet_dgram_connect(struct socket *sock, struct sockaddr * uaddr,
 {
 	struct sock *sk = sock->sk;
 
+	if (addr_len < sizeof(uaddr->sa_family))
+		return -EINVAL;
 	if (uaddr->sa_family == AF_UNSPEC)
 		return sk->sk_prot->disconnect(sk, flags);
 
@@ -574,6 +576,9 @@ int inet_stream_connect(struct socket *sock, struct sockaddr *uaddr,
 	int err;
 	long timeo;
 
+	if (addr_len < sizeof(uaddr->sa_family))
+		return -EINVAL;
+
 	lock_sock(sk);
 
 	if (uaddr->sa_family == AF_UNSPEC) {
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 7a1f1d78893f..0f8caf64caa3 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1369,6 +1369,7 @@ int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
 		sk_eat_skb(sk, skb, 0);
 		if (!desc->count)
 			break;
+		tp->copied_seq = seq;
 	}
 	tp->copied_seq = seq;
 
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 122c11380ffe..fefc45c4b4e8 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -392,7 +392,7 @@ static u32 hwmp_route_info_get(struct ieee80211_sub_if_data *sdata,
 				if (SN_GT(mpath->sn, orig_sn) ||
 				    (mpath->sn == orig_sn &&
 				     action == MPATH_PREQ &&
-				     new_metric > mpath->metric)) {
+				     new_metric >= mpath->metric)) {
 					process = false;
 					fresh_info = false;
 				}
@@ -612,7 +612,7 @@ static void hwmp_prep_frame_process(struct ieee80211_sub_if_data *sdata,
 
 	mesh_path_sel_frame_tx(MPATH_PREP, flags, orig_addr,
 		cpu_to_le32(orig_sn), 0, target_addr,
-		cpu_to_le32(target_sn), mpath->next_hop->sta.addr, hopcount,
+		cpu_to_le32(target_sn), next_hop, hopcount,
 		ttl, cpu_to_le32(lifetime), cpu_to_le32(metric),
 		0, sdata);
 	rcu_read_unlock();
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index cbe53ed4fb0b..cfc473e1b050 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -1991,6 +1991,7 @@ static bool ieee80211_tx_pending_skb(struct ieee80211_local *local,
 void ieee80211_tx_pending(unsigned long data)
 {
 	struct ieee80211_local *local = (struct ieee80211_local *)data;
+	struct ieee80211_sub_if_data *sdata;
 	unsigned long flags;
 	int i;
 	bool txok;
@@ -2029,6 +2030,11 @@ void ieee80211_tx_pending(unsigned long data)
 			if (!txok)
 				break;
 		}
+
+		if (skb_queue_empty(&local->pending[i]))
+			list_for_each_entry_rcu(sdata, &local->interfaces, list)
+				netif_tx_wake_queue(
+					netdev_get_tx_queue(sdata->dev, i));
 	}
 	spin_unlock_irqrestore(&local->queue_stop_reason_lock, flags);
 
diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index c453226f06b2..53af57047435 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -279,13 +279,13 @@ static void __ieee80211_wake_queue(struct ieee80211_hw *hw, int queue,
 		/* someone still has this queue stopped */
 		return;
 
-	if (!skb_queue_empty(&local->pending[queue]))
+	if (skb_queue_empty(&local->pending[queue])) {
+		rcu_read_lock();
+		list_for_each_entry_rcu(sdata, &local->interfaces, list)
+			netif_tx_wake_queue(netdev_get_tx_queue(sdata->dev, queue));
+		rcu_read_unlock();
+	} else
 		tasklet_schedule(&local->tx_pending_tasklet);
-
-	rcu_read_lock();
-	list_for_each_entry_rcu(sdata, &local->interfaces, list)
-		netif_tx_wake_queue(netdev_get_tx_queue(sdata->dev, queue));
-	rcu_read_unlock();
 }
 
 void ieee80211_wake_queue_by_reason(struct ieee80211_hw *hw, int queue,
@@ -1097,9 +1097,9 @@ int ieee80211_reconfig(struct ieee80211_local *local)
 		 */
 		res = drv_start(local);
 		if (res) {
-			WARN(local->suspended, "Harware became unavailable "
-			     "upon resume. This is could be a software issue"
-			     "prior to suspend or a hardware issue\n");
+			WARN(local->suspended, "Hardware became unavailable "
+			     "upon resume. This could be a software issue "
+			     "prior to suspend or a hardware issue.\n");
 			return res;
 		}
 
diff --git a/net/netlabel/netlabel_domainhash.c b/net/netlabel/netlabel_domainhash.c
index 016ab9c75ebd..d37b7f80fa37 100644
--- a/net/netlabel/netlabel_domainhash.c
+++ b/net/netlabel/netlabel_domainhash.c
@@ -51,9 +51,12 @@ struct netlbl_domhsh_tbl {
 };
 
 /* Domain hash table */
-/* XXX - updates should be so rare that having one spinlock for the entire
- * hash table should be okay */
+/* updates should be so rare that having one spinlock for the entire hash table
+ * should be okay */
 static DEFINE_SPINLOCK(netlbl_domhsh_lock);
+#define netlbl_domhsh_rcu_deref(p) \
+	rcu_dereference_check(p, rcu_read_lock_held() || \
+				 lockdep_is_held(&netlbl_domhsh_lock))
 static struct netlbl_domhsh_tbl *netlbl_domhsh = NULL;
 static struct netlbl_dom_map *netlbl_domhsh_def = NULL;
 
@@ -107,7 +110,8 @@ static void netlbl_domhsh_free_entry(struct rcu_head *entry)
  * Description:
  * This is the hashing function for the domain hash table, it returns the
  * correct bucket number for the domain.  The caller is responsibile for
- * calling the rcu_read_[un]lock() functions.
+ * ensuring that the hash table is protected with either a RCU read lock or the
+ * hash table lock.
  *
  */
 static u32 netlbl_domhsh_hash(const char *key)
@@ -121,7 +125,7 @@ static u32 netlbl_domhsh_hash(const char *key)
 
 	for (iter = 0, val = 0, len = strlen(key); iter < len; iter++)
 		val = (val << 4 | (val >> (8 * sizeof(u32) - 4))) ^ key[iter];
-	return val & (rcu_dereference(netlbl_domhsh)->size - 1);
+	return val & (netlbl_domhsh_rcu_deref(netlbl_domhsh)->size - 1);
 }
 
 /**
@@ -131,7 +135,8 @@ static u32 netlbl_domhsh_hash(const char *key)
  * Description:
  * Searches the domain hash table and returns a pointer to the hash table
  * entry if found, otherwise NULL is returned.  The caller is responsibile for
- * the rcu hash table locks (i.e. the caller much call rcu_read_[un]lock()).
+ * ensuring that the hash table is protected with either a RCU read lock or the
+ * hash table lock.
  *
  */
 static struct netlbl_dom_map *netlbl_domhsh_search(const char *domain)
@@ -142,7 +147,7 @@ static struct netlbl_dom_map *netlbl_domhsh_search(const char *domain)
 
 	if (domain != NULL) {
 		bkt = netlbl_domhsh_hash(domain);
-		bkt_list = &rcu_dereference(netlbl_domhsh)->tbl[bkt];
+		bkt_list = &netlbl_domhsh_rcu_deref(netlbl_domhsh)->tbl[bkt];
 		list_for_each_entry_rcu(iter, bkt_list, list)
 			if (iter->valid && strcmp(iter->domain, domain) == 0)
 				return iter;
@@ -160,8 +165,8 @@ static struct netlbl_dom_map *netlbl_domhsh_search(const char *domain)
  * Searches the domain hash table and returns a pointer to the hash table
  * entry if an exact match is found, if an exact match is not present in the
  * hash table then the default entry is returned if valid otherwise NULL is
- * returned.  The caller is responsibile for the rcu hash table locks
- * (i.e. the caller much call rcu_read_[un]lock()).
+ * returned.  The caller is responsibile ensuring that the hash table is
+ * protected with either a RCU read lock or the hash table lock.
  *
  */
 static struct netlbl_dom_map *netlbl_domhsh_search_def(const char *domain)
@@ -170,7 +175,7 @@ static struct netlbl_dom_map *netlbl_domhsh_search_def(const char *domain)
 
 	entry = netlbl_domhsh_search(domain);
 	if (entry == NULL) {
-		entry = rcu_dereference(netlbl_domhsh_def);
+		entry = netlbl_domhsh_rcu_deref(netlbl_domhsh_def);
 		if (entry != NULL && !entry->valid)
 			entry = NULL;
 	}
@@ -307,8 +312,11 @@ int netlbl_domhsh_add(struct netlbl_dom_map *entry,
 	struct netlbl_af6list *tmp6;
 #endif /* IPv6 */
 
+	/* XXX - we can remove this RCU read lock as the spinlock protects the
+	 *       entire function, but before we do we need to fixup the
+	 *       netlbl_af[4,6]list RCU functions to do "the right thing" with
+	 *       respect to rcu_dereference() when only a spinlock is held. */
 	rcu_read_lock();
-
 	spin_lock(&netlbl_domhsh_lock);
 	if (entry->domain != NULL)
 		entry_old = netlbl_domhsh_search(entry->domain);
diff --git a/net/netlabel/netlabel_unlabeled.c b/net/netlabel/netlabel_unlabeled.c
index d7ea2cf390b7..a3d64aabe2f7 100644
--- a/net/netlabel/netlabel_unlabeled.c
+++ b/net/netlabel/netlabel_unlabeled.c
@@ -115,6 +115,9 @@ struct netlbl_unlhsh_walk_arg {
 /* updates should be so rare that having one spinlock for the entire
  * hash table should be okay */
 static DEFINE_SPINLOCK(netlbl_unlhsh_lock);
+#define netlbl_unlhsh_rcu_deref(p) \
+	rcu_dereference_check(p, rcu_read_lock_held() || \
+				 lockdep_is_held(&netlbl_unlhsh_lock))
 static struct netlbl_unlhsh_tbl *netlbl_unlhsh = NULL;
 static struct netlbl_unlhsh_iface *netlbl_unlhsh_def = NULL;
 
@@ -236,15 +239,13 @@ static void netlbl_unlhsh_free_iface(struct rcu_head *entry)
  * Description:
  * This is the hashing function for the unlabeled hash table, it returns the
  * bucket number for the given device/interface.  The caller is responsible for
- * calling the rcu_read_[un]lock() functions.
+ * ensuring that the hash table is protected with either a RCU read lock or
+ * the hash table lock.
  *
  */
 static u32 netlbl_unlhsh_hash(int ifindex)
 {
-	/* this is taken _almost_ directly from
-	 * security/selinux/netif.c:sel_netif_hasfn() as they do pretty much
-	 * the same thing */
-	return ifindex & (rcu_dereference(netlbl_unlhsh)->size - 1);
+	return ifindex & (netlbl_unlhsh_rcu_deref(netlbl_unlhsh)->size - 1);
 }
 
 /**
@@ -254,7 +255,8 @@ static u32 netlbl_unlhsh_hash(int ifindex)
  * Description:
  * Searches the unlabeled connection hash table and returns a pointer to the
  * interface entry which matches @ifindex, otherwise NULL is returned.  The
- * caller is responsible for calling the rcu_read_[un]lock() functions.
+ * caller is responsible for ensuring that the hash table is protected with
+ * either a RCU read lock or the hash table lock.
  *
  */
 static struct netlbl_unlhsh_iface *netlbl_unlhsh_search_iface(int ifindex)
@@ -264,7 +266,7 @@ static struct netlbl_unlhsh_iface *netlbl_unlhsh_search_iface(int ifindex)
 	struct netlbl_unlhsh_iface *iter;
 
 	bkt = netlbl_unlhsh_hash(ifindex);
-	bkt_list = &rcu_dereference(netlbl_unlhsh)->tbl[bkt];
+	bkt_list = &netlbl_unlhsh_rcu_deref(netlbl_unlhsh)->tbl[bkt];
 	list_for_each_entry_rcu(iter, bkt_list, list)
 		if (iter->valid && iter->ifindex == ifindex)
 			return iter;
@@ -273,33 +275,6 @@ static struct netlbl_unlhsh_iface *netlbl_unlhsh_search_iface(int ifindex)
 }
 
 /**
- * netlbl_unlhsh_search_iface_def - Search for a matching interface entry
- * @ifindex: the network interface
- *
- * Description:
- * Searches the unlabeled connection hash table and returns a pointer to the
- * interface entry which matches @ifindex.  If an exact match can not be found
- * and there is a valid default entry, the default entry is returned, otherwise
- * NULL is returned.  The caller is responsible for calling the
- * rcu_read_[un]lock() functions.
- *
- */
-static struct netlbl_unlhsh_iface *netlbl_unlhsh_search_iface_def(int ifindex)
-{
-	struct netlbl_unlhsh_iface *entry;
-
-	entry = netlbl_unlhsh_search_iface(ifindex);
-	if (entry != NULL)
-		return entry;
-
-	entry = rcu_dereference(netlbl_unlhsh_def);
-	if (entry != NULL && entry->valid)
-		return entry;
-
-	return NULL;
-}
-
-/**
  * netlbl_unlhsh_add_addr4 - Add a new IPv4 address entry to the hash table
  * @iface: the associated interface entry
  * @addr: IPv4 address in network byte order
@@ -309,8 +284,7 @@ static struct netlbl_unlhsh_iface *netlbl_unlhsh_search_iface_def(int ifindex)
  * Description:
  * Add a new address entry into the unlabeled connection hash table using the
  * interface entry specified by @iface.  On success zero is returned, otherwise
- * a negative value is returned.  The caller is responsible for calling the
- * rcu_read_[un]lock() functions.
+ * a negative value is returned.
  *
  */
 static int netlbl_unlhsh_add_addr4(struct netlbl_unlhsh_iface *iface,
@@ -350,8 +324,7 @@ static int netlbl_unlhsh_add_addr4(struct netlbl_unlhsh_iface *iface,
  * Description:
  * Add a new address entry into the unlabeled connection hash table using the
  * interface entry specified by @iface.  On success zero is returned, otherwise
- * a negative value is returned.  The caller is responsible for calling the
- * rcu_read_[un]lock() functions.
+ * a negative value is returned.
  *
  */
 static int netlbl_unlhsh_add_addr6(struct netlbl_unlhsh_iface *iface,
@@ -392,8 +365,7 @@ static int netlbl_unlhsh_add_addr6(struct netlbl_unlhsh_iface *iface,
  * Description:
  * Add a new, empty, interface entry into the unlabeled connection hash table.
  * On success a pointer to the new interface entry is returned, on failure NULL
- * is returned.  The caller is responsible for calling the rcu_read_[un]lock()
- * functions.
+ * is returned.
  *
  */
 static struct netlbl_unlhsh_iface *netlbl_unlhsh_add_iface(int ifindex)
@@ -416,10 +388,10 @@ static struct netlbl_unlhsh_iface *netlbl_unlhsh_add_iface(int ifindex)
 		if (netlbl_unlhsh_search_iface(ifindex) != NULL)
 			goto add_iface_failure;
 		list_add_tail_rcu(&iface->list,
-				  &rcu_dereference(netlbl_unlhsh)->tbl[bkt]);
+			     &netlbl_unlhsh_rcu_deref(netlbl_unlhsh)->tbl[bkt]);
 	} else {
 		INIT_LIST_HEAD(&iface->list);
-		if (rcu_dereference(netlbl_unlhsh_def) != NULL)
+		if (netlbl_unlhsh_rcu_deref(netlbl_unlhsh_def) != NULL)
 			goto add_iface_failure;
 		rcu_assign_pointer(netlbl_unlhsh_def, iface);
 	}
@@ -549,8 +521,7 @@ unlhsh_add_return:
  *
  * Description:
  * Remove an IP address entry from the unlabeled connection hash table.
- * Returns zero on success, negative values on failure.  The caller is
- * responsible for calling the rcu_read_[un]lock() functions.
+ * Returns zero on success, negative values on failure.
  *
  */
 static int netlbl_unlhsh_remove_addr4(struct net *net,
@@ -612,8 +583,7 @@ static int netlbl_unlhsh_remove_addr4(struct net *net,
  *
  * Description:
  * Remove an IP address entry from the unlabeled connection hash table.
- * Returns zero on success, negative values on failure.  The caller is
- * responsible for calling the rcu_read_[un]lock() functions.
+ * Returns zero on success, negative values on failure.
  *
  */
 static int netlbl_unlhsh_remove_addr6(struct net *net,
@@ -1548,8 +1518,10 @@ int netlbl_unlabel_getattr(const struct sk_buff *skb,
 	struct netlbl_unlhsh_iface *iface;
 
 	rcu_read_lock();
-	iface = netlbl_unlhsh_search_iface_def(skb->skb_iif);
+	iface = netlbl_unlhsh_search_iface(skb->skb_iif);
 	if (iface == NULL)
+		iface = rcu_dereference(netlbl_unlhsh_def);
+	if (iface == NULL || !iface->valid)
 		goto unlabel_getattr_nolabel;
 	switch (family) {
 	case PF_INET: {
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index acbbae1e89b5..795424396aff 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -683,6 +683,9 @@ static int netlink_connect(struct socket *sock, struct sockaddr *addr,
 	struct netlink_sock *nlk = nlk_sk(sk);
 	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
 
+	if (alen < sizeof(addr->sa_family))
+		return -EINVAL;
+
 	if (addr->sa_family == AF_UNSPEC) {
 		sk->sk_state	= NETLINK_UNCONNECTED;
 		nlk->dst_pid	= 0;
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index b7604b823f46..422da20d1e5b 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -325,7 +325,7 @@ struct reg_regdb_search_request {
 };
 
 static LIST_HEAD(reg_regdb_search_list);
-static DEFINE_SPINLOCK(reg_regdb_search_lock);
+static DEFINE_MUTEX(reg_regdb_search_mutex);
 
 static void reg_regdb_search(struct work_struct *work)
 {
@@ -333,7 +333,7 @@ static void reg_regdb_search(struct work_struct *work)
 	const struct ieee80211_regdomain *curdom, *regdom;
 	int i, r;
 
-	spin_lock(&reg_regdb_search_lock);
+	mutex_lock(&reg_regdb_search_mutex);
 	while (!list_empty(&reg_regdb_search_list)) {
 		request = list_first_entry(&reg_regdb_search_list,
 					   struct reg_regdb_search_request,
@@ -347,18 +347,16 @@ static void reg_regdb_search(struct work_struct *work)
 				r = reg_copy_regd(&regdom, curdom);
 				if (r)
 					break;
-				spin_unlock(&reg_regdb_search_lock);
 				mutex_lock(&cfg80211_mutex);
 				set_regdom(regdom);
 				mutex_unlock(&cfg80211_mutex);
-				spin_lock(&reg_regdb_search_lock);
 				break;
 			}
 		}
 
 		kfree(request);
 	}
-	spin_unlock(&reg_regdb_search_lock);
+	mutex_unlock(&reg_regdb_search_mutex);
 }
 
 static DECLARE_WORK(reg_regdb_work, reg_regdb_search);
@@ -376,9 +374,9 @@ static void reg_regdb_query(const char *alpha2)
 
 	memcpy(request->alpha2, alpha2, 2);
 
-	spin_lock(&reg_regdb_search_lock);
+	mutex_lock(&reg_regdb_search_mutex);
 	list_add_tail(&request->list, &reg_regdb_search_list);
-	spin_unlock(&reg_regdb_search_lock);
+	mutex_unlock(&reg_regdb_search_mutex);
 
 	schedule_work(&reg_regdb_work);
 }
author	Linus Torvalds <torvalds@linux-foundation.org>	2010-04-06 19:34:06 +0400
committer	Linus Torvalds <torvalds@linux-foundation.org>	2010-04-06 19:34:06 +0400
commit	cb4361c1dc29cd870f664c004b1817106fbce0fa (patch)
tree	c31533210d738c9074b21b8f31fe0399b39545f0
parent	309361e09ca9e9670dc8664e5d14125bf82078af (diff)
parent	fb9e2d887243499b8d28efcf80821c4f6a092395 (diff)
download	linux-cb4361c1dc29cd870f664c004b1817106fbce0fa.tar.xz