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

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1745 commits)
  dp83640: free packet queues on remove
  dp83640: use proper function to free transmit time stamping packets
  ipv6: Do not use routes from locally generated RAs
  |PATCH net-next] tg3: add tx_dropped counter
  be2net: don't create multiple RX/TX rings in multi channel mode
  be2net: don't create multiple TXQs in BE2
  be2net: refactor VF setup/teardown code into be_vf_setup/clear()
  be2net: add vlan/rx-mode/flow-control config to be_setup()
  net_sched: cls_flow: use skb_header_pointer()
  ipv4: avoid useless call of the function check_peer_pmtu
  TCP: remove TCP_DEBUG
  net: Fix driver name for mdio-gpio.c
  ipv4: tcp: fix TOS value in ACK messages sent from TIME_WAIT
  rtnetlink: Add missing manual netlink notification in dev_change_net_namespaces
  ipv4: fix ipsec forward performance regression
  jme: fix irq storm after suspend/resume
  route: fix ICMP redirect validation
  net: hold sock reference while processing tx timestamps
  tcp: md5: add more const attributes
  Add ethtool -g support to virtio_net
  ...

Fix up conflicts in:
 - drivers/net/Kconfig:
	The split-up generated a trivial conflict with removal of a
	stale reference to Documentation/networking/net-modules.txt.
	Remove it from the new location instead.
 - fs/sysfs/dir.c:
	Fairly nasty conflicts with the sysfs rb-tree usage, conflicting
	with Eric Biederman's changes for tagged directories.

8 files changed, 19131 insertions, 0 deletions

diff --git a/drivers/net/ethernet/marvell/Kconfig b/drivers/net/ethernet/marvell/Kconfig
new file mode 100644
index 000000000000..0029934748bc
--- /dev/null
+++ b/drivers/net/ethernet/marvell/Kconfig
@@ -0,0 +1,111 @@
+#
+# Marvell device configuration
+#
+
+config NET_VENDOR_MARVELL
+	bool "Marvell devices"
+	default y
+	depends on PCI || CPU_PXA168 || MV64X60 || PPC32 || PLAT_ORION || INET
+	---help---
+	  If you have a network (Ethernet) card belonging to this class, say Y
+	  and read the Ethernet-HOWTO, available from
+	  <http://www.tldp.org/docs.html#howto>.
+
+	  Note that the answer to this question doesn't directly affect the
+	  kernel: saying N will just cause the configurator to skip all
+	  the questions about Marvell devices. If you say Y, you will be
+	  asked for your specific card in the following questions.
+
+if NET_VENDOR_MARVELL
+
+config MV643XX_ETH
+	tristate "Marvell Discovery (643XX) and Orion ethernet support"
+	depends on (MV64X60 || PPC32 || PLAT_ORION) && INET
+	select INET_LRO
+	select PHYLIB
+	---help---
+	  This driver supports the gigabit ethernet MACs in the
+	  Marvell Discovery PPC/MIPS chipset family (MV643XX) and
+	  in the Marvell Orion ARM SoC family.
+
+	  Some boards that use the Discovery chipset are the Momenco
+	  Ocelot C and Jaguar ATX and Pegasos II.
+
+config PXA168_ETH
+	tristate "Marvell pxa168 ethernet support"
+	depends on CPU_PXA168
+	select PHYLIB
+	---help---
+	  This driver supports the pxa168 Ethernet ports.
+
+	  To compile this driver as a module, choose M here. The module
+	  will be called pxa168_eth.
+
+config SKGE
+	tristate "Marvell Yukon Gigabit Ethernet support"
+	depends on PCI
+	select CRC32
+	---help---
+	  This driver support the Marvell Yukon or SysKonnect SK-98xx/SK-95xx
+	  and related Gigabit Ethernet adapters. It is a new smaller driver
+	  with better performance and more complete ethtool support.
+
+	  It does not support the link failover and network management
+	  features that "portable" vendor supplied sk98lin driver does.
+
+	  This driver supports adapters based on the original Yukon chipset:
+	  Marvell 88E8001, Belkin F5D5005, CNet GigaCard, DLink DGE-530T,
+	  Linksys EG1032/EG1064, 3Com 3C940/3C940B, SysKonnect SK-9871/9872.
+
+	  It does not support the newer Yukon2 chipset: a separate driver,
+	  sky2, is provided for these adapters.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called skge.  This is recommended.
+
+config SKGE_DEBUG
+	bool "Debugging interface"
+	depends on SKGE && DEBUG_FS
+	---help---
+	  This option adds the ability to dump driver state for debugging.
+	  The file /sys/kernel/debug/skge/ethX displays the state of the internal
+	  transmit and receive rings.
+
+	  If unsure, say N.
+
+config SKGE_GENESIS
+	bool "Support for older SysKonnect Genesis boards"
+	depends on SKGE
+	---help---
+	 This enables support for the older and uncommon SysKonnect Genesis
+	 chips, which support MII via an external transceiver, instead of
+	 an internal one. Disabling this option will save some memory
+	 by making code smaller. If unsure say Y.
+
+config SKY2
+	tristate "Marvell Yukon 2 support"
+	depends on PCI
+	select CRC32
+	---help---
+	  This driver supports Gigabit Ethernet adapters based on the
+	  Marvell Yukon 2 chipset:
+	  Marvell 88E8021/88E8022/88E8035/88E8036/88E8038/88E8050/88E8052/
+	  88E8053/88E8055/88E8061/88E8062, SysKonnect SK-9E21D/SK-9S21
+
+	  There is companion driver for the older Marvell Yukon and
+	  SysKonnect Genesis based adapters: skge.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called sky2.  This is recommended.
+
+config SKY2_DEBUG
+	bool "Debugging interface"
+	depends on SKY2 && DEBUG_FS
+	---help---
+	  This option adds the ability to dump driver state for debugging.
+	  The file /sys/kernel/debug/sky2/ethX displays the state of the internal
+	  transmit and receive rings.
+
+	  If unsure, say N.
+
+endif # NET_VENDOR_MARVELL
diff --git a/drivers/net/ethernet/marvell/Makefile b/drivers/net/ethernet/marvell/Makefile
new file mode 100644
index 000000000000..57e3234a37ba
--- /dev/null
+++ b/drivers/net/ethernet/marvell/Makefile
@@ -0,0 +1,8 @@
+#
+# Makefile for the Marvell device drivers.
+#
+
+obj-$(CONFIG_MV643XX_ETH) += mv643xx_eth.o
+obj-$(CONFIG_PXA168_ETH) += pxa168_eth.o
+obj-$(CONFIG_SKGE) += skge.o
+obj-$(CONFIG_SKY2) += sky2.o
diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c
new file mode 100644
index 000000000000..194a03113802
--- /dev/null
+++ b/drivers/net/ethernet/marvell/mv643xx_eth.c
@@ -0,0 +1,3019 @@
+/*
+ * Driver for Marvell Discovery (MV643XX) and Marvell Orion ethernet ports
+ * Copyright (C) 2002 Matthew Dharm <mdharm@momenco.com>
+ *
+ * Based on the 64360 driver from:
+ * Copyright (C) 2002 Rabeeh Khoury <rabeeh@galileo.co.il>
+ *		      Rabeeh Khoury <rabeeh@marvell.com>
+ *
+ * Copyright (C) 2003 PMC-Sierra, Inc.,
+ *	written by Manish Lachwani
+ *
+ * Copyright (C) 2003 Ralf Baechle <ralf@linux-mips.org>
+ *
+ * Copyright (C) 2004-2006 MontaVista Software, Inc.
+ *			   Dale Farnsworth <dale@farnsworth.org>
+ *
+ * Copyright (C) 2004 Steven J. Hill <sjhill1@rockwellcollins.com>
+ *				     <sjhill@realitydiluted.com>
+ *
+ * Copyright (C) 2007-2008 Marvell Semiconductor
+ *			   Lennert Buytenhek <buytenh@marvell.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/init.h>
+#include <linux/dma-mapping.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/etherdevice.h>
+#include <linux/delay.h>
+#include <linux/ethtool.h>
+#include <linux/platform_device.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+#include <linux/phy.h>
+#include <linux/mv643xx_eth.h>
+#include <linux/io.h>
+#include <linux/types.h>
+#include <linux/inet_lro.h>
+#include <linux/slab.h>
+#include <asm/system.h>
+
+static char mv643xx_eth_driver_name[] = "mv643xx_eth";
+static char mv643xx_eth_driver_version[] = "1.4";
+
+
+/*
+ * Registers shared between all ports.
+ */
+#define PHY_ADDR			0x0000
+#define SMI_REG				0x0004
+#define  SMI_BUSY			0x10000000
+#define  SMI_READ_VALID			0x08000000
+#define  SMI_OPCODE_READ		0x04000000
+#define  SMI_OPCODE_WRITE		0x00000000
+#define ERR_INT_CAUSE			0x0080
+#define  ERR_INT_SMI_DONE		0x00000010
+#define ERR_INT_MASK			0x0084
+#define WINDOW_BASE(w)			(0x0200 + ((w) << 3))
+#define WINDOW_SIZE(w)			(0x0204 + ((w) << 3))
+#define WINDOW_REMAP_HIGH(w)		(0x0280 + ((w) << 2))
+#define WINDOW_BAR_ENABLE		0x0290
+#define WINDOW_PROTECT(w)		(0x0294 + ((w) << 4))
+
+/*
+ * Main per-port registers.  These live at offset 0x0400 for
+ * port #0, 0x0800 for port #1, and 0x0c00 for port #2.
+ */
+#define PORT_CONFIG			0x0000
+#define  UNICAST_PROMISCUOUS_MODE	0x00000001
+#define PORT_CONFIG_EXT			0x0004
+#define MAC_ADDR_LOW			0x0014
+#define MAC_ADDR_HIGH			0x0018
+#define SDMA_CONFIG			0x001c
+#define  TX_BURST_SIZE_16_64BIT		0x01000000
+#define  TX_BURST_SIZE_4_64BIT		0x00800000
+#define  BLM_TX_NO_SWAP			0x00000020
+#define  BLM_RX_NO_SWAP			0x00000010
+#define  RX_BURST_SIZE_16_64BIT		0x00000008
+#define  RX_BURST_SIZE_4_64BIT		0x00000004
+#define PORT_SERIAL_CONTROL		0x003c
+#define  SET_MII_SPEED_TO_100		0x01000000
+#define  SET_GMII_SPEED_TO_1000		0x00800000
+#define  SET_FULL_DUPLEX_MODE		0x00200000
+#define  MAX_RX_PACKET_9700BYTE		0x000a0000
+#define  DISABLE_AUTO_NEG_SPEED_GMII	0x00002000
+#define  DO_NOT_FORCE_LINK_FAIL		0x00000400
+#define  SERIAL_PORT_CONTROL_RESERVED	0x00000200
+#define  DISABLE_AUTO_NEG_FOR_FLOW_CTRL	0x00000008
+#define  DISABLE_AUTO_NEG_FOR_DUPLEX	0x00000004
+#define  FORCE_LINK_PASS		0x00000002
+#define  SERIAL_PORT_ENABLE		0x00000001
+#define PORT_STATUS			0x0044
+#define  TX_FIFO_EMPTY			0x00000400
+#define  TX_IN_PROGRESS			0x00000080
+#define  PORT_SPEED_MASK		0x00000030
+#define  PORT_SPEED_1000		0x00000010
+#define  PORT_SPEED_100			0x00000020
+#define  PORT_SPEED_10			0x00000000
+#define  FLOW_CONTROL_ENABLED		0x00000008
+#define  FULL_DUPLEX			0x00000004
+#define  LINK_UP			0x00000002
+#define TXQ_COMMAND			0x0048
+#define TXQ_FIX_PRIO_CONF		0x004c
+#define TX_BW_RATE			0x0050
+#define TX_BW_MTU			0x0058
+#define TX_BW_BURST			0x005c
+#define INT_CAUSE			0x0060
+#define  INT_TX_END			0x07f80000
+#define  INT_TX_END_0			0x00080000
+#define  INT_RX				0x000003fc
+#define  INT_RX_0			0x00000004
+#define  INT_EXT			0x00000002
+#define INT_CAUSE_EXT			0x0064
+#define  INT_EXT_LINK_PHY		0x00110000
+#define  INT_EXT_TX			0x000000ff
+#define INT_MASK			0x0068
+#define INT_MASK_EXT			0x006c
+#define TX_FIFO_URGENT_THRESHOLD	0x0074
+#define TXQ_FIX_PRIO_CONF_MOVED		0x00dc
+#define TX_BW_RATE_MOVED		0x00e0
+#define TX_BW_MTU_MOVED			0x00e8
+#define TX_BW_BURST_MOVED		0x00ec
+#define RXQ_CURRENT_DESC_PTR(q)		(0x020c + ((q) << 4))
+#define RXQ_COMMAND			0x0280
+#define TXQ_CURRENT_DESC_PTR(q)		(0x02c0 + ((q) << 2))
+#define TXQ_BW_TOKENS(q)		(0x0300 + ((q) << 4))
+#define TXQ_BW_CONF(q)			(0x0304 + ((q) << 4))
+#define TXQ_BW_WRR_CONF(q)		(0x0308 + ((q) << 4))
+
+/*
+ * Misc per-port registers.
+ */
+#define MIB_COUNTERS(p)			(0x1000 + ((p) << 7))
+#define SPECIAL_MCAST_TABLE(p)		(0x1400 + ((p) << 10))
+#define OTHER_MCAST_TABLE(p)		(0x1500 + ((p) << 10))
+#define UNICAST_TABLE(p)		(0x1600 + ((p) << 10))
+
+
+/*
+ * SDMA configuration register default value.
+ */
+#if defined(__BIG_ENDIAN)
+#define PORT_SDMA_CONFIG_DEFAULT_VALUE		\
+		(RX_BURST_SIZE_4_64BIT	|	\
+		 TX_BURST_SIZE_4_64BIT)
+#elif defined(__LITTLE_ENDIAN)
+#define PORT_SDMA_CONFIG_DEFAULT_VALUE		\
+		(RX_BURST_SIZE_4_64BIT	|	\
+		 BLM_RX_NO_SWAP		|	\
+		 BLM_TX_NO_SWAP		|	\
+		 TX_BURST_SIZE_4_64BIT)
+#else
+#error One of __BIG_ENDIAN or __LITTLE_ENDIAN must be defined
+#endif
+
+
+/*
+ * Misc definitions.
+ */
+#define DEFAULT_RX_QUEUE_SIZE	128
+#define DEFAULT_TX_QUEUE_SIZE	256
+#define SKB_DMA_REALIGN		((PAGE_SIZE - NET_SKB_PAD) % SMP_CACHE_BYTES)
+
+
+/*
+ * RX/TX descriptors.
+ */
+#if defined(__BIG_ENDIAN)
+struct rx_desc {
+	u16 byte_cnt;		/* Descriptor buffer byte count		*/
+	u16 buf_size;		/* Buffer size				*/
+	u32 cmd_sts;		/* Descriptor command status		*/
+	u32 next_desc_ptr;	/* Next descriptor pointer		*/
+	u32 buf_ptr;		/* Descriptor buffer pointer		*/
+};
+
+struct tx_desc {
+	u16 byte_cnt;		/* buffer byte count			*/
+	u16 l4i_chk;		/* CPU provided TCP checksum		*/
+	u32 cmd_sts;		/* Command/status field			*/
+	u32 next_desc_ptr;	/* Pointer to next descriptor		*/
+	u32 buf_ptr;		/* pointer to buffer for this descriptor*/
+};
+#elif defined(__LITTLE_ENDIAN)
+struct rx_desc {
+	u32 cmd_sts;		/* Descriptor command status		*/
+	u16 buf_size;		/* Buffer size				*/
+	u16 byte_cnt;		/* Descriptor buffer byte count		*/
+	u32 buf_ptr;		/* Descriptor buffer pointer		*/
+	u32 next_desc_ptr;	/* Next descriptor pointer		*/
+};
+
+struct tx_desc {
+	u32 cmd_sts;		/* Command/status field			*/
+	u16 l4i_chk;		/* CPU provided TCP checksum		*/
+	u16 byte_cnt;		/* buffer byte count			*/
+	u32 buf_ptr;		/* pointer to buffer for this descriptor*/
+	u32 next_desc_ptr;	/* Pointer to next descriptor		*/
+};
+#else
+#error One of __BIG_ENDIAN or __LITTLE_ENDIAN must be defined
+#endif
+
+/* RX & TX descriptor command */
+#define BUFFER_OWNED_BY_DMA		0x80000000
+
+/* RX & TX descriptor status */
+#define ERROR_SUMMARY			0x00000001
+
+/* RX descriptor status */
+#define LAYER_4_CHECKSUM_OK		0x40000000
+#define RX_ENABLE_INTERRUPT		0x20000000
+#define RX_FIRST_DESC			0x08000000
+#define RX_LAST_DESC			0x04000000
+#define RX_IP_HDR_OK			0x02000000
+#define RX_PKT_IS_IPV4			0x01000000
+#define RX_PKT_IS_ETHERNETV2		0x00800000
+#define RX_PKT_LAYER4_TYPE_MASK		0x00600000
+#define RX_PKT_LAYER4_TYPE_TCP_IPV4	0x00000000
+#define RX_PKT_IS_VLAN_TAGGED		0x00080000
+
+/* TX descriptor command */
+#define TX_ENABLE_INTERRUPT		0x00800000
+#define GEN_CRC				0x00400000
+#define TX_FIRST_DESC			0x00200000
+#define TX_LAST_DESC			0x00100000
+#define ZERO_PADDING			0x00080000
+#define GEN_IP_V4_CHECKSUM		0x00040000
+#define GEN_TCP_UDP_CHECKSUM		0x00020000
+#define UDP_FRAME			0x00010000
+#define MAC_HDR_EXTRA_4_BYTES		0x00008000
+#define MAC_HDR_EXTRA_8_BYTES		0x00000200
+
+#define TX_IHL_SHIFT			11
+
+
+/* global *******************************************************************/
+struct mv643xx_eth_shared_private {
+	/*
+	 * Ethernet controller base address.
+	 */
+	void __iomem *base;
+
+	/*
+	 * Points at the right SMI instance to use.
+	 */
+	struct mv643xx_eth_shared_private *smi;
+
+	/*
+	 * Provides access to local SMI interface.
+	 */
+	struct mii_bus *smi_bus;
+
+	/*
+	 * If we have access to the error interrupt pin (which is
+	 * somewhat misnamed as it not only reflects internal errors
+	 * but also reflects SMI completion), use that to wait for
+	 * SMI access completion instead of polling the SMI busy bit.
+	 */
+	int err_interrupt;
+	wait_queue_head_t smi_busy_wait;
+
+	/*
+	 * Per-port MBUS window access register value.
+	 */
+	u32 win_protect;
+
+	/*
+	 * Hardware-specific parameters.
+	 */
+	unsigned int t_clk;
+	int extended_rx_coal_limit;
+	int tx_bw_control;
+	int tx_csum_limit;
+};
+
+#define TX_BW_CONTROL_ABSENT		0
+#define TX_BW_CONTROL_OLD_LAYOUT	1
+#define TX_BW_CONTROL_NEW_LAYOUT	2
+
+static int mv643xx_eth_open(struct net_device *dev);
+static int mv643xx_eth_stop(struct net_device *dev);
+
+
+/* per-port *****************************************************************/
+struct mib_counters {
+	u64 good_octets_received;
+	u32 bad_octets_received;
+	u32 internal_mac_transmit_err;
+	u32 good_frames_received;
+	u32 bad_frames_received;
+	u32 broadcast_frames_received;
+	u32 multicast_frames_received;
+	u32 frames_64_octets;
+	u32 frames_65_to_127_octets;
+	u32 frames_128_to_255_octets;
+	u32 frames_256_to_511_octets;
+	u32 frames_512_to_1023_octets;
+	u32 frames_1024_to_max_octets;
+	u64 good_octets_sent;
+	u32 good_frames_sent;
+	u32 excessive_collision;
+	u32 multicast_frames_sent;
+	u32 broadcast_frames_sent;
+	u32 unrec_mac_control_received;
+	u32 fc_sent;
+	u32 good_fc_received;
+	u32 bad_fc_received;
+	u32 undersize_received;
+	u32 fragments_received;
+	u32 oversize_received;
+	u32 jabber_received;
+	u32 mac_receive_error;
+	u32 bad_crc_event;
+	u32 collision;
+	u32 late_collision;
+};
+
+struct lro_counters {
+	u32 lro_aggregated;
+	u32 lro_flushed;
+	u32 lro_no_desc;
+};
+
+struct rx_queue {
+	int index;
+
+	int rx_ring_size;
+
+	int rx_desc_count;
+	int rx_curr_desc;
+	int rx_used_desc;
+
+	struct rx_desc *rx_desc_area;
+	dma_addr_t rx_desc_dma;
+	int rx_desc_area_size;
+	struct sk_buff **rx_skb;
+
+	struct net_lro_mgr lro_mgr;
+	struct net_lro_desc lro_arr[8];
+};
+
+struct tx_queue {
+	int index;
+
+	int tx_ring_size;
+
+	int tx_desc_count;
+	int tx_curr_desc;
+	int tx_used_desc;
+
+	struct tx_desc *tx_desc_area;
+	dma_addr_t tx_desc_dma;
+	int tx_desc_area_size;
+
+	struct sk_buff_head tx_skb;
+
+	unsigned long tx_packets;
+	unsigned long tx_bytes;
+	unsigned long tx_dropped;
+};
+
+struct mv643xx_eth_private {
+	struct mv643xx_eth_shared_private *shared;
+	void __iomem *base;
+	int port_num;
+
+	struct net_device *dev;
+
+	struct phy_device *phy;
+
+	struct timer_list mib_counters_timer;
+	spinlock_t mib_counters_lock;
+	struct mib_counters mib_counters;
+
+	struct lro_counters lro_counters;
+
+	struct work_struct tx_timeout_task;
+
+	struct napi_struct napi;
+	u32 int_mask;
+	u8 oom;
+	u8 work_link;
+	u8 work_tx;
+	u8 work_tx_end;
+	u8 work_rx;
+	u8 work_rx_refill;
+
+	int skb_size;
+	struct sk_buff_head rx_recycle;
+
+	/*
+	 * RX state.
+	 */
+	int rx_ring_size;
+	unsigned long rx_desc_sram_addr;
+	int rx_desc_sram_size;
+	int rxq_count;
+	struct timer_list rx_oom;
+	struct rx_queue rxq[8];
+
+	/*
+	 * TX state.
+	 */
+	int tx_ring_size;
+	unsigned long tx_desc_sram_addr;
+	int tx_desc_sram_size;
+	int txq_count;
+	struct tx_queue txq[8];
+};
+
+
+/* port register accessors **************************************************/
+static inline u32 rdl(struct mv643xx_eth_private *mp, int offset)
+{
+	return readl(mp->shared->base + offset);
+}
+
+static inline u32 rdlp(struct mv643xx_eth_private *mp, int offset)
+{
+	return readl(mp->base + offset);
+}
+
+static inline void wrl(struct mv643xx_eth_private *mp, int offset, u32 data)
+{
+	writel(data, mp->shared->base + offset);
+}
+
+static inline void wrlp(struct mv643xx_eth_private *mp, int offset, u32 data)
+{
+	writel(data, mp->base + offset);
+}
+
+
+/* rxq/txq helper functions *************************************************/
+static struct mv643xx_eth_private *rxq_to_mp(struct rx_queue *rxq)
+{
+	return container_of(rxq, struct mv643xx_eth_private, rxq[rxq->index]);
+}
+
+static struct mv643xx_eth_private *txq_to_mp(struct tx_queue *txq)
+{
+	return container_of(txq, struct mv643xx_eth_private, txq[txq->index]);
+}
+
+static void rxq_enable(struct rx_queue *rxq)
+{
+	struct mv643xx_eth_private *mp = rxq_to_mp(rxq);
+	wrlp(mp, RXQ_COMMAND, 1 << rxq->index);
+}
+
+static void rxq_disable(struct rx_queue *rxq)
+{
+	struct mv643xx_eth_private *mp = rxq_to_mp(rxq);
+	u8 mask = 1 << rxq->index;
+
+	wrlp(mp, RXQ_COMMAND, mask << 8);
+	while (rdlp(mp, RXQ_COMMAND) & mask)
+		udelay(10);
+}
+
+static void txq_reset_hw_ptr(struct tx_queue *txq)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	u32 addr;
+
+	addr = (u32)txq->tx_desc_dma;
+	addr += txq->tx_curr_desc * sizeof(struct tx_desc);
+	wrlp(mp, TXQ_CURRENT_DESC_PTR(txq->index), addr);
+}
+
+static void txq_enable(struct tx_queue *txq)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	wrlp(mp, TXQ_COMMAND, 1 << txq->index);
+}
+
+static void txq_disable(struct tx_queue *txq)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	u8 mask = 1 << txq->index;
+
+	wrlp(mp, TXQ_COMMAND, mask << 8);
+	while (rdlp(mp, TXQ_COMMAND) & mask)
+		udelay(10);
+}
+
+static void txq_maybe_wake(struct tx_queue *txq)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	struct netdev_queue *nq = netdev_get_tx_queue(mp->dev, txq->index);
+
+	if (netif_tx_queue_stopped(nq)) {
+		__netif_tx_lock(nq, smp_processor_id());
+		if (txq->tx_ring_size - txq->tx_desc_count >= MAX_SKB_FRAGS + 1)
+			netif_tx_wake_queue(nq);
+		__netif_tx_unlock(nq);
+	}
+}
+
+
+/* rx napi ******************************************************************/
+static int
+mv643xx_get_skb_header(struct sk_buff *skb, void **iphdr, void **tcph,
+		       u64 *hdr_flags, void *priv)
+{
+	unsigned long cmd_sts = (unsigned long)priv;
+
+	/*
+	 * Make sure that this packet is Ethernet II, is not VLAN
+	 * tagged, is IPv4, has a valid IP header, and is TCP.
+	 */
+	if ((cmd_sts & (RX_IP_HDR_OK | RX_PKT_IS_IPV4 |
+		       RX_PKT_IS_ETHERNETV2 | RX_PKT_LAYER4_TYPE_MASK |
+		       RX_PKT_IS_VLAN_TAGGED)) !=
+	    (RX_IP_HDR_OK | RX_PKT_IS_IPV4 |
+	     RX_PKT_IS_ETHERNETV2 | RX_PKT_LAYER4_TYPE_TCP_IPV4))
+		return -1;
+
+	skb_reset_network_header(skb);
+	skb_set_transport_header(skb, ip_hdrlen(skb));
+	*iphdr = ip_hdr(skb);
+	*tcph = tcp_hdr(skb);
+	*hdr_flags = LRO_IPV4 | LRO_TCP;
+
+	return 0;
+}
+
+static int rxq_process(struct rx_queue *rxq, int budget)
+{
+	struct mv643xx_eth_private *mp = rxq_to_mp(rxq);
+	struct net_device_stats *stats = &mp->dev->stats;
+	int lro_flush_needed;
+	int rx;
+
+	lro_flush_needed = 0;
+	rx = 0;
+	while (rx < budget && rxq->rx_desc_count) {
+		struct rx_desc *rx_desc;
+		unsigned int cmd_sts;
+		struct sk_buff *skb;
+		u16 byte_cnt;
+
+		rx_desc = &rxq->rx_desc_area[rxq->rx_curr_desc];
+
+		cmd_sts = rx_desc->cmd_sts;
+		if (cmd_sts & BUFFER_OWNED_BY_DMA)
+			break;
+		rmb();
+
+		skb = rxq->rx_skb[rxq->rx_curr_desc];
+		rxq->rx_skb[rxq->rx_curr_desc] = NULL;
+
+		rxq->rx_curr_desc++;
+		if (rxq->rx_curr_desc == rxq->rx_ring_size)
+			rxq->rx_curr_desc = 0;
+
+		dma_unmap_single(mp->dev->dev.parent, rx_desc->buf_ptr,
+				 rx_desc->buf_size, DMA_FROM_DEVICE);
+		rxq->rx_desc_count--;
+		rx++;
+
+		mp->work_rx_refill |= 1 << rxq->index;
+
+		byte_cnt = rx_desc->byte_cnt;
+
+		/*
+		 * Update statistics.
+		 *
+		 * Note that the descriptor byte count includes 2 dummy
+		 * bytes automatically inserted by the hardware at the
+		 * start of the packet (which we don't count), and a 4
+		 * byte CRC at the end of the packet (which we do count).
+		 */
+		stats->rx_packets++;
+		stats->rx_bytes += byte_cnt - 2;
+
+		/*
+		 * In case we received a packet without first / last bits
+		 * on, or the error summary bit is set, the packet needs
+		 * to be dropped.
+		 */
+		if ((cmd_sts & (RX_FIRST_DESC | RX_LAST_DESC | ERROR_SUMMARY))
+			!= (RX_FIRST_DESC | RX_LAST_DESC))
+			goto err;
+
+		/*
+		 * The -4 is for the CRC in the trailer of the
+		 * received packet
+		 */
+		skb_put(skb, byte_cnt - 2 - 4);
+
+		if (cmd_sts & LAYER_4_CHECKSUM_OK)
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+		skb->protocol = eth_type_trans(skb, mp->dev);
+
+		if (skb->dev->features & NETIF_F_LRO &&
+		    skb->ip_summed == CHECKSUM_UNNECESSARY) {
+			lro_receive_skb(&rxq->lro_mgr, skb, (void *)cmd_sts);
+			lro_flush_needed = 1;
+		} else
+			netif_receive_skb(skb);
+
+		continue;
+
+err:
+		stats->rx_dropped++;
+
+		if ((cmd_sts & (RX_FIRST_DESC | RX_LAST_DESC)) !=
+			(RX_FIRST_DESC | RX_LAST_DESC)) {
+			if (net_ratelimit())
+				netdev_err(mp->dev,
+					   "received packet spanning multiple descriptors\n");
+		}
+
+		if (cmd_sts & ERROR_SUMMARY)
+			stats->rx_errors++;
+
+		dev_kfree_skb(skb);
+	}
+
+	if (lro_flush_needed)
+		lro_flush_all(&rxq->lro_mgr);
+
+	if (rx < budget)
+		mp->work_rx &= ~(1 << rxq->index);
+
+	return rx;
+}
+
+static int rxq_refill(struct rx_queue *rxq, int budget)
+{
+	struct mv643xx_eth_private *mp = rxq_to_mp(rxq);
+	int refilled;
+
+	refilled = 0;
+	while (refilled < budget && rxq->rx_desc_count < rxq->rx_ring_size) {
+		struct sk_buff *skb;
+		int rx;
+		struct rx_desc *rx_desc;
+		int size;
+
+		skb = __skb_dequeue(&mp->rx_recycle);
+		if (skb == NULL)
+			skb = dev_alloc_skb(mp->skb_size);
+
+		if (skb == NULL) {
+			mp->oom = 1;
+			goto oom;
+		}
+
+		if (SKB_DMA_REALIGN)
+			skb_reserve(skb, SKB_DMA_REALIGN);
+
+		refilled++;
+		rxq->rx_desc_count++;
+
+		rx = rxq->rx_used_desc++;
+		if (rxq->rx_used_desc == rxq->rx_ring_size)
+			rxq->rx_used_desc = 0;
+
+		rx_desc = rxq->rx_desc_area + rx;
+
+		size = skb->end - skb->data;
+		rx_desc->buf_ptr = dma_map_single(mp->dev->dev.parent,
+						  skb->data, size,
+						  DMA_FROM_DEVICE);
+		rx_desc->buf_size = size;
+		rxq->rx_skb[rx] = skb;
+		wmb();
+		rx_desc->cmd_sts = BUFFER_OWNED_BY_DMA | RX_ENABLE_INTERRUPT;
+		wmb();
+
+		/*
+		 * The hardware automatically prepends 2 bytes of
+		 * dummy data to each received packet, so that the
+		 * IP header ends up 16-byte aligned.
+		 */
+		skb_reserve(skb, 2);
+	}
+
+	if (refilled < budget)
+		mp->work_rx_refill &= ~(1 << rxq->index);
+
+oom:
+	return refilled;
+}
+
+
+/* tx ***********************************************************************/
+static inline unsigned int has_tiny_unaligned_frags(struct sk_buff *skb)
+{
+	int frag;
+
+	for (frag = 0; frag < skb_shinfo(skb)->nr_frags; frag++) {
+		const skb_frag_t *fragp = &skb_shinfo(skb)->frags[frag];
+
+		if (skb_frag_size(fragp) <= 8 && fragp->page_offset & 7)
+			return 1;
+	}
+
+	return 0;
+}
+
+static void txq_submit_frag_skb(struct tx_queue *txq, struct sk_buff *skb)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	int nr_frags = skb_shinfo(skb)->nr_frags;
+	int frag;
+
+	for (frag = 0; frag < nr_frags; frag++) {
+		skb_frag_t *this_frag;
+		int tx_index;
+		struct tx_desc *desc;
+
+		this_frag = &skb_shinfo(skb)->frags[frag];
+		tx_index = txq->tx_curr_desc++;
+		if (txq->tx_curr_desc == txq->tx_ring_size)
+			txq->tx_curr_desc = 0;
+		desc = &txq->tx_desc_area[tx_index];
+
+		/*
+		 * The last fragment will generate an interrupt
+		 * which will free the skb on TX completion.
+		 */
+		if (frag == nr_frags - 1) {
+			desc->cmd_sts = BUFFER_OWNED_BY_DMA |
+					ZERO_PADDING | TX_LAST_DESC |
+					TX_ENABLE_INTERRUPT;
+		} else {
+			desc->cmd_sts = BUFFER_OWNED_BY_DMA;
+		}
+
+		desc->l4i_chk = 0;
+		desc->byte_cnt = skb_frag_size(this_frag);
+		desc->buf_ptr = skb_frag_dma_map(mp->dev->dev.parent,
+						 this_frag, 0,
+						 skb_frag_size(this_frag),
+						 DMA_TO_DEVICE);
+	}
+}
+
+static inline __be16 sum16_as_be(__sum16 sum)
+{
+	return (__force __be16)sum;
+}
+
+static int txq_submit_skb(struct tx_queue *txq, struct sk_buff *skb)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	int nr_frags = skb_shinfo(skb)->nr_frags;
+	int tx_index;
+	struct tx_desc *desc;
+	u32 cmd_sts;
+	u16 l4i_chk;
+	int length;
+
+	cmd_sts = TX_FIRST_DESC | GEN_CRC | BUFFER_OWNED_BY_DMA;
+	l4i_chk = 0;
+
+	if (skb->ip_summed == CHECKSUM_PARTIAL) {
+		int hdr_len;
+		int tag_bytes;
+
+		BUG_ON(skb->protocol != htons(ETH_P_IP) &&
+		       skb->protocol != htons(ETH_P_8021Q));
+
+		hdr_len = (void *)ip_hdr(skb) - (void *)skb->data;
+		tag_bytes = hdr_len - ETH_HLEN;
+		if (skb->len - hdr_len > mp->shared->tx_csum_limit ||
+		    unlikely(tag_bytes & ~12)) {
+			if (skb_checksum_help(skb) == 0)
+				goto no_csum;
+			kfree_skb(skb);
+			return 1;
+		}
+
+		if (tag_bytes & 4)
+			cmd_sts |= MAC_HDR_EXTRA_4_BYTES;
+		if (tag_bytes & 8)
+			cmd_sts |= MAC_HDR_EXTRA_8_BYTES;
+
+		cmd_sts |= GEN_TCP_UDP_CHECKSUM |
+			   GEN_IP_V4_CHECKSUM   |
+			   ip_hdr(skb)->ihl << TX_IHL_SHIFT;
+
+		switch (ip_hdr(skb)->protocol) {
+		case IPPROTO_UDP:
+			cmd_sts |= UDP_FRAME;
+			l4i_chk = ntohs(sum16_as_be(udp_hdr(skb)->check));
+			break;
+		case IPPROTO_TCP:
+			l4i_chk = ntohs(sum16_as_be(tcp_hdr(skb)->check));
+			break;
+		default:
+			BUG();
+		}
+	} else {
+no_csum:
+		/* Errata BTS #50, IHL must be 5 if no HW checksum */
+		cmd_sts |= 5 << TX_IHL_SHIFT;
+	}
+
+	tx_index = txq->tx_curr_desc++;
+	if (txq->tx_curr_desc == txq->tx_ring_size)
+		txq->tx_curr_desc = 0;
+	desc = &txq->tx_desc_area[tx_index];
+
+	if (nr_frags) {
+		txq_submit_frag_skb(txq, skb);
+		length = skb_headlen(skb);
+	} else {
+		cmd_sts |= ZERO_PADDING | TX_LAST_DESC | TX_ENABLE_INTERRUPT;
+		length = skb->len;
+	}
+
+	desc->l4i_chk = l4i_chk;
+	desc->byte_cnt = length;
+	desc->buf_ptr = dma_map_single(mp->dev->dev.parent, skb->data,
+				       length, DMA_TO_DEVICE);
+
+	__skb_queue_tail(&txq->tx_skb, skb);
+
+	skb_tx_timestamp(skb);
+
+	/* ensure all other descriptors are written before first cmd_sts */
+	wmb();
+	desc->cmd_sts = cmd_sts;
+
+	/* clear TX_END status */
+	mp->work_tx_end &= ~(1 << txq->index);
+
+	/* ensure all descriptors are written before poking hardware */
+	wmb();
+	txq_enable(txq);
+
+	txq->tx_desc_count += nr_frags + 1;
+
+	return 0;
+}
+
+static netdev_tx_t mv643xx_eth_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct mv643xx_eth_private *mp = netdev_priv(dev);
+	int length, queue;
+	struct tx_queue *txq;
+	struct netdev_queue *nq;
+
+	queue = skb_get_queue_mapping(skb);
+	txq = mp->txq + queue;
+	nq = netdev_get_tx_queue(dev, queue);
+
+	if (has_tiny_unaligned_frags(skb) && __skb_linearize(skb)) {
+		txq->tx_dropped++;
+		netdev_printk(KERN_DEBUG, dev,
+			      "failed to linearize skb with tiny unaligned fragment\n");
+		return NETDEV_TX_BUSY;
+	}
+
+	if (txq->tx_ring_size - txq->tx_desc_count < MAX_SKB_FRAGS + 1) {
+		if (net_ratelimit())
+			netdev_err(dev, "tx queue full?!\n");
+		kfree_skb(skb);
+		return NETDEV_TX_OK;
+	}
+
+	length = skb->len;
+
+	if (!txq_submit_skb(txq, skb)) {
+		int entries_left;
+
+		txq->tx_bytes += length;
+		txq->tx_packets++;
+
+		entries_left = txq->tx_ring_size - txq->tx_desc_count;
+		if (entries_left < MAX_SKB_FRAGS + 1)
+			netif_tx_stop_queue(nq);
+	}
+
+	return NETDEV_TX_OK;
+}
+
+
+/* tx napi ******************************************************************/
+static void txq_kick(struct tx_queue *txq)
+{
+	struct mv643xx_eth_private *mp = txq_to_mp(txq);
+	struct netdev_queue *nq = netdev_get_tx_queue(mp->dev, txq->index);
+	u32 hw_desc_ptr;
+	u32 expected_ptr;
+
+	__netif_tx_lock(nq, smp_processor_id());
+
+	if (rdlp(mp, TXQ_COMMAND) & (1 << txq->index))
+		goto out;
+
+	hw_desc_ptr = rdlp(mp, TXQ_CURRENT_DESC_PTR(txq->index));
+	expected_ptr = (u32)txq->tx_desc_dma +
+				txq->tx_curr_desc * sizeof(struct tx_desc);
+
+	if (hw_desc_ptr != expected_ptr)
+		txq_enable(txq);
+
+out:
+	__netif_tx_unlock(nq);
+
+	mp->work_tx_end &= ~(1 << txq->index);
+}
+
+static int txq_reclaim(struct tx_queue *txq, int budget, int force)
+{
+	struct