Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

MAINTAINERS

@@ -1696,11 +1696,13 @@ M: mtk-manpages@gmx.net
 W: ftp://ftp.kernel.org/pub/linux/docs/manpages
 S: Maintained
 
-MARVELL MV64340 ETHERNET DRIVER
+MARVELL MV643XX ETHERNET DRIVER
+P:	Dale Farnsworth
+M:	dale@farnsworth.org
 P:	Manish Lachwani
-L:	linux-mips@linux-mips.org
+M:	mlachwani@mvista.com
 L:	netdev@vger.kernel.org
-S:	Supported
+S:	Odd Fixes for 2.4; Maintained for 2.6.
 
 MATROX FRAMEBUFFER DRIVER
 P:	Petr Vandrovec

drivers/net/b44.c

@@ -2136,7 +2136,7 @@ static int __init b44_init(void)
 
 	/* Setup paramaters for syncing RX/TX DMA descriptors */
 	dma_desc_align_mask = ~(dma_desc_align_size - 1);
-	dma_desc_sync_size = max(dma_desc_align_size, sizeof(struct dma_desc));
+	dma_desc_sync_size = max_t(unsigned int, dma_desc_align_size, sizeof(struct dma_desc));
 
 	return pci_module_init(&b44_driver);
 }

drivers/net/e100.c

@@ -1,25 +1,25 @@
 /*******************************************************************************
 
-  
+
   Copyright(c) 1999 - 2005 Intel Corporation. All rights reserved.
-  
-  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) 
+
+  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 
+
+  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 
+  this program; if not, write to the Free Software Foundation, Inc., 59
   Temple Place - Suite 330, Boston, MA  02111-1307, USA.
-  
+
   The full GNU General Public License is included in this distribution in the
   file called LICENSE.
-  
+
   Contact Information:
   Linux NICS <linux.nics@intel.com>
   Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
@@ -160,7 +160,7 @@
 
 #define DRV_NAME		"e100"
 #define DRV_EXT		"-NAPI"
-#define DRV_VERSION		"3.4.14-k4"DRV_EXT
+#define DRV_VERSION		"3.5.10-k2"DRV_EXT
 #define DRV_DESCRIPTION		"Intel(R) PRO/100 Network Driver"
 #define DRV_COPYRIGHT		"Copyright(c) 1999-2005 Intel Corporation"
 #define PFX			DRV_NAME ": "
@@ -320,7 +320,7 @@ enum cuc_dump {
 	cuc_dump_complete       = 0x0000A005,
 	cuc_dump_reset_complete = 0x0000A007,
 };
-		
+
 enum port {
 	software_reset  = 0x0000,
 	selftest        = 0x0001,
@@ -715,10 +715,10 @@ static u16 e100_eeprom_read(struct nic *nic, u16 *addr_len, u16 addr)
 		ctrl = (cmd_addr_data & (1 << i)) ? eecs | eedi : eecs;
 		writeb(ctrl, &nic->csr->eeprom_ctrl_lo);
 		e100_write_flush(nic); udelay(4);
-		
+
 		writeb(ctrl | eesk, &nic->csr->eeprom_ctrl_lo);
 		e100_write_flush(nic); udelay(4);
-		
+
 		/* Eeprom drives a dummy zero to EEDO after receiving
 		 * complete address.  Use this to adjust addr_len. */
 		ctrl = readb(&nic->csr->eeprom_ctrl_lo);
@@ -726,7 +726,7 @@ static u16 e100_eeprom_read(struct nic *nic, u16 *addr_len, u16 addr)
 			*addr_len -= (i - 16);
 			i = 17;
 		}
-		
+
 		data = (data << 1) | (ctrl & eedo ? 1 : 0);
 	}
 
@@ -1170,7 +1170,7 @@ static void e100_configure(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 0x00000000, 0x00000000, 0x00000000, 0x00000000, \
 }
 
-static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
+static void e100_setup_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 {
 /* *INDENT-OFF* */
 	static struct {
@@ -1213,13 +1213,13 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 *  driver can change the algorithm.
 *
 *  INTDELAY - This loads the dead-man timer with its inital value.
-*    When this timer expires the interrupt is asserted, and the 
+*    When this timer expires the interrupt is asserted, and the
 *    timer is reset each time a new packet is received.  (see
 *    BUNDLEMAX below to set the limit on number of chained packets)
 *    The current default is 0x600 or 1536.  Experiments show that
 *    the value should probably stay within the 0x200 - 0x1000.
 *
-*  BUNDLEMAX - 
+*  BUNDLEMAX -
 *    This sets the maximum number of frames that will be bundled.  In
 *    some situations, such as the TCP windowing algorithm, it may be
 *    better to limit the growth of the bundle size than let it go as
@@ -1229,7 +1229,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 *    an interrupt for every frame received.  If you do not want to put
 *    a limit on the bundle size, set this value to xFFFF.
 *
-*  BUNDLESMALL - 
+*  BUNDLESMALL -
 *    This contains a bit-mask describing the minimum size frame that
 *    will be bundled.  The default masks the lower 7 bits, which means
 *    that any frame less than 128 bytes in length will not be bundled,
@@ -1244,7 +1244,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 *
 *    The current default is 0xFF80, which masks out the lower 7 bits.
 *    This means that any frame which is x7F (127) bytes or smaller
-*    will cause an immediate interrupt.  Because this value must be a 
+*    will cause an immediate interrupt.  Because this value must be a
 *    bit mask, there are only a few valid values that can be used.  To
 *    turn this feature off, the driver can write the value xFFFF to the
 *    lower word of this instruction (in the same way that the other
@@ -1253,7 +1253,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 *    standard Ethernet frames are <= 2047 bytes in length.
 *************************************************************************/
 
-/* if you wish to disable the ucode functionality, while maintaining the 
+/* if you wish to disable the ucode functionality, while maintaining the
  * workarounds it provides, set the following defines to:
  * BUNDLESMALL 0
  * BUNDLEMAX 1
@@ -1284,12 +1284,46 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb)
 
 		for (i = 0; i < UCODE_SIZE; i++)
 			cb->u.ucode[i] = cpu_to_le32(ucode[i]);
-		cb->command = cpu_to_le16(cb_ucode);
+		cb->command = cpu_to_le16(cb_ucode | cb_el);
 		return;
 	}
 
 noloaducode:
-	cb->command = cpu_to_le16(cb_nop);
+	cb->command = cpu_to_le16(cb_nop | cb_el);
+}
+
+static inline int e100_exec_cb_wait(struct nic *nic, struct sk_buff *skb,
+	void (*cb_prepare)(struct nic *, struct cb *, struct sk_buff *))
+{
+	int err = 0, counter = 50;
+	struct cb *cb = nic->cb_to_clean;
+
+	if ((err = e100_exec_cb(nic, NULL, e100_setup_ucode)))
+		DPRINTK(PROBE,ERR, "ucode cmd failed with error %d\n", err);
+
+	/* must restart cuc */
+	nic->cuc_cmd = cuc_start;
+
+	/* wait for completion */
+	e100_write_flush(nic);
+	udelay(10);
+
+	/* wait for possibly (ouch) 500ms */
+	while (!(cb->status & cpu_to_le16(cb_complete))) {
+		msleep(10);
+		if (!--counter) break;
+	}
+
+	/* ack any interupts, something could have been set */
+	writeb(~0, &nic->csr->scb.stat_ack);
+
+	/* if the command failed, or is not OK, notify and return */
+	if (!counter || !(cb->status & cpu_to_le16(cb_ok))) {
+		DPRINTK(PROBE,ERR, "ucode load failed\n");
+		err = -EPERM;
+	}
+
+	return err;
 }
 
 static void e100_setup_iaaddr(struct nic *nic, struct cb *cb,
@@ -1357,13 +1391,13 @@ static int e100_phy_init(struct nic *nic)
 		mdio_write(netdev, nic->mii.phy_id, MII_NSC_CONG, cong);
 	}
 
-	if((nic->mac >= mac_82550_D102) || ((nic->flags & ich) && 
+	if((nic->mac >= mac_82550_D102) || ((nic->flags & ich) &&
 	   (mdio_read(netdev, nic->mii.phy_id, MII_TPISTATUS) & 0x8000))) {
 		/* enable/disable MDI/MDI-X auto-switching.
 		   MDI/MDI-X auto-switching is disabled for 82551ER/QM chips */
 		if((nic->mac == mac_82551_E) || (nic->mac == mac_82551_F) ||
-		   (nic->mac == mac_82551_10) || (nic->mii.force_media) || 
-		   !(nic->eeprom[eeprom_cnfg_mdix] & eeprom_mdix_enabled)) 
+		   (nic->mac == mac_82551_10) || (nic->mii.force_media) ||
+		   !(nic->eeprom[eeprom_cnfg_mdix] & eeprom_mdix_enabled))
 			mdio_write(netdev, nic->mii.phy_id, MII_NCONFIG, 0);
 		else
 			mdio_write(netdev, nic->mii.phy_id, MII_NCONFIG, NCONFIG_AUTO_SWITCH);
@@ -1388,7 +1422,7 @@ static int e100_hw_init(struct nic *nic)
 		return err;
 	if((err = e100_exec_cmd(nic, ruc_load_base, 0)))
 		return err;
-	if((err = e100_exec_cb(nic, NULL, e100_load_ucode)))
+	if ((err = e100_exec_cb_wait(nic, NULL, e100_setup_ucode)))
 		return err;
 	if((err = e100_exec_cb(nic, NULL, e100_configure)))
 		return err;
@@ -1493,7 +1527,7 @@ static void e100_update_stats(struct nic *nic)
 		}
 	}
 
-	
+
 	if(e100_exec_cmd(nic, cuc_dump_reset, 0))
 		DPRINTK(TX_ERR, DEBUG, "exec cuc_dump_reset failed\n");
 }
@@ -1542,10 +1576,10 @@ static void e100_watchdog(unsigned long data)
 	mii_check_link(&nic->mii);
 
 	/* Software generated interrupt to recover from (rare) Rx
-	* allocation failure.
-	* Unfortunately have to use a spinlock to not re-enable interrupts
-	* accidentally, due to hardware that shares a register between the
-	* interrupt mask bit and the SW Interrupt generation bit */
+	 * allocation failure.
+	 * Unfortunately have to use a spinlock to not re-enable interrupts
+	 * accidentally, due to hardware that shares a register between the
+	 * interrupt mask bit and the SW Interrupt generation bit */
 	spin_lock_irq(&nic->cmd_lock);
 	writeb(readb(&nic->csr->scb.cmd_hi) | irq_sw_gen,&nic->csr->scb.cmd_hi);
 	spin_unlock_irq(&nic->cmd_lock);
@@ -1830,7 +1864,7 @@ static void e100_rx_clean(struct nic *nic, unsigned int *work_done,
 	struct rx *rx_to_start = NULL;
 
 	/* are we already rnr? then pay attention!!! this ensures that
-	 * the state machine progression never allows a start with a 
+	 * the state machine progression never allows a start with a
 	 * partially cleaned list, avoiding a race between hardware
 	 * and rx_to_clean when in NAPI mode */
 	if(RU_SUSPENDED == nic->ru_running)
@@ -2066,7 +2100,7 @@ static void e100_tx_timeout(struct net_device *netdev)
 {
 	struct nic *nic = netdev_priv(netdev);
 
-	/* Reset outside of interrupt context, to avoid request_irq 
+	/* Reset outside of interrupt context, to avoid request_irq
 	 * in interrupt context */
 	schedule_work(&nic->tx_timeout_task);
 }
@@ -2313,7 +2347,7 @@ static int e100_set_ringparam(struct net_device *netdev,
 	struct param_range *rfds = &nic->params.rfds;
 	struct param_range *cbs = &nic->params.cbs;
 
-	if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending)) 
+	if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending))
 		return -EINVAL;
 
 	if(netif_running(netdev))
@@ -2631,7 +2665,9 @@ static int __devinit e100_probe(struct pci_dev *pdev,
 		nic->flags |= wol_magic;
 
 	/* ack any pending wake events, disable PME */
-	pci_enable_wake(pdev, 0, 0);
+	err = pci_enable_wake(pdev, 0, 0);
+	if (err)
+		DPRINTK(PROBE, ERR, "Error clearing wake event\n");
 
 	strcpy(netdev->name, "eth%d");
 	if((err = register_netdev(netdev))) {
@@ -2682,6 +2718,7 @@ static int e100_suspend(struct pci_dev *pdev, pm_message_t state)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct nic *nic = netdev_priv(netdev);
+	int retval;
 
 	if(netif_running(netdev))
 		e100_down(nic);
@@ -2689,9 +2726,14 @@ static int e100_suspend(struct pci_dev *pdev, pm_message_t state)
 	netif_device_detach(netdev);
 
 	pci_save_state(pdev);
-	pci_enable_wake(pdev, pci_choose_state(pdev, state), nic->flags & (wol_magic | e100_asf(nic)));
+	retval = pci_enable_wake(pdev, pci_choose_state(pdev, state),
+	                         nic->flags & (wol_magic | e100_asf(nic)));
+	if (retval)
+		DPRINTK(PROBE,ERR, "Error enabling wake\n");
 	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
+	retval = pci_set_power_state(pdev, pci_choose_state(pdev, state));
+	if (retval)
+		DPRINTK(PROBE,ERR, "Error %d setting power state\n", retval);
 
 	return 0;
 }
@@ -2700,11 +2742,16 @@ static int e100_resume(struct pci_dev *pdev)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct nic *nic = netdev_priv(netdev);
+	int retval;
 
-	pci_set_power_state(pdev, PCI_D0);
+	retval = pci_set_power_state(pdev, PCI_D0);
+	if (retval)
+		DPRINTK(PROBE,ERR, "Error waking adapter\n");
 	pci_restore_state(pdev);
 	/* ack any pending wake events, disable PME */
-	pci_enable_wake(pdev, 0, 0);
+	retval = pci_enable_wake(pdev, 0, 0);
+	if (retval)
+		DPRINTK(PROBE,ERR, "Error clearing wake events\n");
 	if(e100_hw_init(nic))
 		DPRINTK(HW, ERR, "e100_hw_init failed\n");
 
@@ -2721,12 +2768,15 @@ static void e100_shutdown(struct pci_dev *pdev)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct nic *nic = netdev_priv(netdev);
+	int retval;
 
 #ifdef CONFIG_PM
-	pci_enable_wake(pdev, 0, nic->flags & (wol_magic | e100_asf(nic)));
+	retval = pci_enable_wake(pdev, 0, nic->flags & (wol_magic | e100_asf(nic)));
 #else
-	pci_enable_wake(pdev, 0, nic->flags & (wol_magic));
+	retval = pci_enable_wake(pdev, 0, nic->flags & (wol_magic));
 #endif
+	if (retval)
+		DPRINTK(PROBE,ERR, "Error enabling wake\n");
 }
 
 
@@ -2739,7 +2789,7 @@ static struct pci_driver e100_driver = {
 	.suspend =      e100_suspend,
 	.resume =       e100_resume,
 #endif
-	.shutdown =	e100_shutdown,
+	.shutdown =     e100_shutdown,
 };
 
 static int __init e100_init_module(void)

drivers/net/e1000/e1000.h

@@ -72,10 +72,6 @@
 #include <linux/mii.h>
 #include <linux/ethtool.h>
 #include <linux/if_vlan.h>
-#ifdef CONFIG_E1000_MQ
-#include <linux/cpu.h>
-#include <linux/smp.h>
-#endif
 
 #define BAR_0		0
 #define BAR_1		1
@@ -87,6 +83,10 @@
 struct e1000_adapter;
 
 #include "e1000_hw.h"
+#ifdef CONFIG_E1000_MQ
+#include <linux/cpu.h>
+#include <linux/smp.h>
+#endif
 
 #ifdef DBG
 #define E1000_DBG(args...) printk(KERN_DEBUG "e1000: " args)
@@ -169,6 +169,13 @@ struct e1000_buffer {
 	uint16_t next_to_watch;
 };
 
+#ifdef CONFIG_E1000_MQ
+struct e1000_queue_stats {
+	uint64_t packets;
+	uint64_t bytes;
+};
+#endif
+
 struct e1000_ps_page { struct page *ps_page[PS_PAGE_BUFFERS]; };
 struct e1000_ps_page_dma { uint64_t ps_page_dma[PS_PAGE_BUFFERS]; };
 
@@ -191,10 +198,12 @@ struct e1000_tx_ring {
 	spinlock_t tx_lock;
 	uint16_t tdh;
 	uint16_t tdt;
-	uint64_t pkt;
 
 	boolean_t last_tx_tso;
 
+#ifdef CONFIG_E1000_MQ
+	struct e1000_queue_stats tx_stats;
+#endif
 };
 
 struct e1000_rx_ring {
@@ -216,9 +225,17 @@ struct e1000_rx_ring {
 	struct e1000_ps_page *ps_page;
 	struct e1000_ps_page_dma *ps_page_dma;
 
+	struct sk_buff *rx_skb_top;
+	struct sk_buff *rx_skb_prev;
+
+	/* cpu for rx queue */
+	int cpu;
+
 	uint16_t rdh;
 	uint16_t rdt;
-	uint64_t pkt;
+#ifdef CONFIG_E1000_MQ
+	struct e1000_queue_stats rx_stats;
+#endif
 };
 
 #define E1000_DESC_UNUSED(R) \
@@ -251,6 +268,9 @@ struct e1000_adapter {
 	uint16_t link_speed;
 	uint16_t link_duplex;
 	spinlock_t stats_lock;
+#ifdef CONFIG_E1000_NAPI
+	spinlock_t tx_queue_lock;
+#endif
 	atomic_t irq_sem;
 	struct work_struct tx_timeout_task;
 	struct work_struct watchdog_task;
@@ -264,6 +284,7 @@ struct e1000_adapter {
 #ifdef CONFIG_E1000_MQ
 	struct e1000_tx_ring **cpu_tx_ring; /* per-cpu */
 #endif
+	unsigned long tx_queue_len;
 	uint32_t txd_cmd;
 	uint32_t tx_int_delay;
 	uint32_t tx_abs_int_delay;
@@ -271,9 +292,11 @@ struct e1000_adapter {
 	uint64_t gotcl_old;
 	uint64_t tpt_old;
 	uint64_t colc_old;
+	uint32_t tx_timeout_count;
 	uint32_t tx_fifo_head;
 	uint32_t tx_head_addr;
 	uint32_t tx_fifo_size;
+	uint8_t  tx_timeout_factor;
 	atomic_t tx_fifo_stall;
 	boolean_t pcix_82544;
 	boolean_t detect_tx_hung;
@@ -281,14 +304,15 @@ struct e1000_adapter {
 	/* RX */
 #ifdef CONFIG_E1000_NAPI
 	boolean_t (*clean_rx) (struct e1000_adapter *adapter,
-			       struct e1000_rx_ring *rx_ring,
-			       int *work_done, int work_to_do);
+						   struct e1000_rx_ring *rx_ring,
+						   int *work_done, int work_to_do);
 #else
 	boolean_t (*clean_rx) (struct e1000_adapter *adapter,
-			       struct e1000_rx_ring *rx_ring);
+						   struct e1000_rx_ring *rx_ring);
 #endif
 	void (*alloc_rx_buf) (struct e1000_adapter *adapter,
-			      struct e1000_rx_ring *rx_ring);
+						  struct e1000_rx_ring *rx_ring,
+						  int cleaned_count);
 	struct e1000_rx_ring *rx_ring;      /* One per active queue */
 #ifdef CONFIG_E1000_NAPI
 	struct net_device *polling_netdev;  /* One per active queue */
@@ -296,13 +320,15 @@ struct e1000_adapter {
 #ifdef CONFIG_E1000_MQ
 	struct net_device **cpu_netdev;     /* per-cpu */
 	struct call_async_data_struct rx_sched_call_data;
-	int cpu_for_queue[4];
+	cpumask_t cpumask;
 #endif
-	int num_queues;
+	int num_tx_queues;
+	int num_rx_queues;
 
 	uint64_t hw_csum_err;
 	uint64_t hw_csum_good;
 	uint64_t rx_hdr_split;
+	uint32_t alloc_rx_buff_failed;
 	uint32_t rx_int_delay;
 	uint32_t rx_abs_int_delay;
 	boolean_t rx_csum;
@@ -330,6 +356,7 @@ struct e1000_adapter {
 	struct e1000_rx_ring test_rx_ring;
 
 
+	u32 *config_space;
 	int msg_enable;
 #ifdef CONFIG_PCI_MSI
 	boolean_t have_msi;

drivers/net/e1000/e1000_ethtool.c

@@ -80,6 +80,7 @@ static const struct e1000_stats e1000_gstrings_stats[] = {
 	{ "tx_deferred_ok", E1000_STAT(stats.dc) },
 	{ "tx_single_coll_ok", E1000_STAT(stats.scc) },
 	{ "tx_multi_coll_ok", E1000_STAT(stats.mcc) },
+	{ "tx_timeout_count", E1000_STAT(tx_timeout_count) },
 	{ "rx_long_length_errors", E1000_STAT(stats.roc) },
 	{ "rx_short_length_errors", E1000_STAT(stats.ruc) },
 	{ "rx_align_errors", E1000_STAT(stats.algnerrc) },
@@ -93,9 +94,20 @@ static const struct e1000_stats e1000_gstrings_stats[] = {
 	{ "rx_csum_offload_good", E1000_STAT(hw_csum_good) },
 	{ "rx_csum_offload_errors", E1000_STAT(hw_csum_err) },
 	{ "rx_header_split", E1000_STAT(rx_hdr_split) },
+	{ "alloc_rx_buff_failed", E1000_STAT(alloc_rx_buff_failed) },
 };
-#define E1000_STATS_LEN	\
+
+#ifdef CONFIG_E1000_MQ
+#define E1000_QUEUE_STATS_LEN \
+	(((struct e1000_adapter *)netdev->priv)->num_tx_queues + \
+	 ((struct e1000_adapter *)netdev->priv)->num_rx_queues) \
+	* (sizeof(struct e1000_queue_stats) / sizeof(uint64_t))
+#else
+#define E1000_QUEUE_STATS_LEN 0
+#endif
+#define E1000_GLOBAL_STATS_LEN	\
 	sizeof(e1000_gstrings_stats) / sizeof(struct e1000_stats)
+#define E1000_STATS_LEN (E1000_GLOBAL_STATS_LEN + E1000_QUEUE_STATS_LEN)
 static const char e1000_gstrings_test[][ETH_GSTRING_LEN] = {
 	"Register test  (offline)", "Eeprom test    (offline)",
 	"Interrupt test (offline)", "Loopback test  (offline)",
@@ -183,7 +195,15 @@ e1000_set_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
 
-	if(ecmd->autoneg == AUTONEG_ENABLE) {
+	/* When SoL/IDER sessions are active, autoneg/speed/duplex
+	 * cannot be changed */
+	if (e1000_check_phy_reset_block(hw)) {
+		DPRINTK(DRV, ERR, "Cannot change link characteristics "
+		        "when SoL/IDER is active.\n");
+		return -EINVAL;
+	}
+
+	if (ecmd->autoneg == AUTONEG_ENABLE) {
 		hw->autoneg = 1;
 		if(hw->media_type == e1000_media_type_fiber)
 			hw->autoneg_advertised = ADVERTISED_1000baseT_Full |
@@ -567,21 +587,21 @@ e1000_get_drvinfo(struct net_device *netdev,
 
 	strncpy(drvinfo->driver,  e1000_driver_name, 32);
 	strncpy(drvinfo->version, e1000_driver_version, 32);
-	
-	/* EEPROM image version # is reported as firware version # for
+
+	/* EEPROM image version # is reported as firmware version # for
 	 * 8257{1|2|3} controllers */
 	e1000_read_eeprom(&adapter->hw, 5, 1, &eeprom_data);
 	switch (adapter->hw.mac_type) {
 	case e1000_82571:
 	case e1000_82572:
 	case e1000_82573:
-		sprintf(firmware_version, "%d.%d-%d", 
+		sprintf(firmware_version, "%d.%d-%d",
 			(eeprom_data & 0xF000) >> 12,
 			(eeprom_data & 0x0FF0) >> 4,
 			eeprom_data & 0x000F);
 		break;
 	default:
-		sprintf(firmware_version, "n/a");
+		sprintf(firmware_version, "N/A");
 	}
 
 	strncpy(drvinfo->fw_version, firmware_version, 32);
@@ -623,8 +643,8 @@ e1000_set_ringparam(struct net_device *netdev,
 	struct e1000_rx_ring *rxdr, *rx_old, *rx_new;
 	int i, err, tx_ring_size, rx_ring_size;
 
-	tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_queues;
-	rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_queues;
+	tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues;
+	rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues;
 
 	if (netif_running(adapter->netdev))
 		e1000_down(adapter);
@@ -663,10 +683,10 @@ e1000_set_ringparam(struct net_device *netdev,
 		E1000_MAX_TXD : E1000_MAX_82544_TXD));
 	E1000_ROUNDUP(txdr->count, REQ_TX_DESCRIPTOR_MULTIPLE); 
 
-	for (i = 0; i < adapter->num_queues; i++) {
+	for (i = 0; i < adapter->num_tx_queues; i++)
 		txdr[i].count = txdr->count;
+	for (i = 0; i < adapter->num_rx_queues; i++)
 		rxdr[i].count = rxdr->count;
-	}
 
 	if(netif_running(adapter->netdev)) {
 		/* Try to get new resources before deleting old */
@@ -979,18 +999,17 @@ e1000_free_desc_rings(struct e1000_adapter *adapter)
 		}
 	}
 
-	if(txdr->desc) {
+	if (txdr->desc) {
 		pci_free_consistent(pdev, txdr->size, txdr->desc, txdr->dma);
 		txdr->desc = NULL;
 	}
-	if(rxdr->desc) {
+	if (rxdr->desc) {
 		pci_free_consistent(pdev, rxdr->size, rxdr->desc, rxdr->dma);
 		rxdr->desc = NULL;
 	}
 
 	kfree(txdr->buffer_info);
 	txdr->buffer_info = NULL;
-
 	kfree(rxdr->buffer_info);
 	rxdr->buffer_info = NULL;
 
@@ -1327,11 +1346,11 @@ e1000_set_phy_loopback(struct e1000_adapter *adapter)
 static int
 e1000_setup_loopback_test(struct e1000_adapter *adapter)
 {
-	uint32_t rctl;
 	struct e1000_hw *hw = &adapter->hw;
+	uint32_t rctl;
 
 	if (hw->media_type == e1000_media_type_fiber ||
-	   hw->media_type == e1000_media_type_internal_serdes) {
+	    hw->media_type == e1000_media_type_internal_serdes) {
 		switch (hw->mac_type) {
 		case e1000_82545:
 		case e1000_82546:
@@ -1362,25 +1381,25 @@ e1000_setup_loopback_test(struct e1000_adapter *adapter)
 static void
 e1000_loopback_cleanup(struct e1000_adapter *adapter)
 {
+	struct e1000_hw *hw = &adapter->hw;
 	uint32_t rctl;
 	uint16_t phy_reg;
-	struct e1000_hw *hw = &adapter->hw;
 
-	rctl = E1000_READ_REG(&adapter->hw, RCTL);
+	rctl = E1000_READ_REG(hw, RCTL);
 	rctl &= ~(E1000_RCTL_LBM_TCVR | E1000_RCTL_LBM_MAC);
-	E1000_WRITE_REG(&adapter->hw, RCTL, rctl);
+	E1000_WRITE_REG(hw, RCTL, rctl);
 
 	switch (hw->mac_type) {
 	case e1000_82571:
 	case e1000_82572:
 		if (hw->media_type == e1000_media_type_fiber ||
-		   hw->media_type == e1000_media_type_internal_serdes){
+		    hw->media_type == e1000_media_type_internal_serdes) {
 #define E1000_SERDES_LB_OFF 0x400
 			E1000_WRITE_REG(hw, SCTL, E1000_SERDES_LB_OFF);
 			msec_delay(10);
 			break;
 		}
-		/* fall thru for Cu adapters */
+		/* Fall Through */
 	case e1000_82545:
 	case e1000_82546:
 	case e1000_82545_rev_3:
@@ -1401,7 +1420,7 @@ static void
 e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size)
 {
 	memset(skb->data, 0xFF, frame_size);
-	frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size;
+	frame_size &= ~1;
 	memset(&skb->data[frame_size / 2], 0xAA, frame_size / 2 - 1);
 	memset(&skb->data[frame_size / 2 + 10], 0xBE, 1);
 	memset(&skb->data[frame_size / 2 + 12], 0xAF, 1);
@@ -1410,7 +1429,7 @@ e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size)
 static int
 e1000_check_lbtest_frame(struct sk_buff *skb, unsigned int frame_size)
 {
-	frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size;
+	frame_size &= ~1;
 	if(*(skb->data + 3) == 0xFF) {
 		if((*(skb->data + frame_size / 2 + 10) == 0xBE) &&
 		   (*(skb->data + frame_size / 2 + 12) == 0xAF)) {
@@ -1488,14 +1507,25 @@ e1000_run_loopback_test(struct e1000_adapter *adapter)
 static int
 e1000_loopback_test(struct e1000_adapter *adapter, uint64_t *data)
 {
-	if((*data = e1000_setup_desc_rings(adapter))) goto err_loopback;
-	if((*data = e1000_setup_loopback_test(adapter)))
-		goto err_loopback_setup;
+	/* PHY loopback cannot be performed if SoL/IDER
+	 * sessions are active */
+	if (e1000_check_phy_reset_block(&adapter->hw)) {
+		DPRINTK(DRV, ERR, "Cannot do PHY loopback test "
+		        "when SoL/IDER is active.\n");
+		*data = 0;
+		goto out;
+	}
+
+	if ((*data = e1000_setup_desc_rings(adapter)))
+		goto out;
+	if ((*data = e1000_setup_loopback_test(adapter)))
+		goto err_loopback;
 	*data = e1000_run_loopback_test(adapter);
 	e1000_loopback_cleanup(adapter);
-err_loopback_setup:
-	e1000_free_desc_rings(adapter);
+
 err_loopback:
+	e1000_free_desc_rings(adapter);
+out:
 	return *data;
 }
 
@@ -1617,6 +1647,7 @@ e1000_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
 
 	case E1000_DEV_ID_82546EB_FIBER:
 	case E1000_DEV_ID_82546GB_FIBER:
+	case E1000_DEV_ID_82571EB_FIBER:
 		/* Wake events only supported on port A for dual fiber */
 		if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) {
 			wol->supported = 0;
@@ -1660,6 +1691,7 @@ e1000_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
 
 	case E1000_DEV_ID_82546EB_FIBER:
 	case E1000_DEV_ID_82546GB_FIBER:
+	case E1000_DEV_ID_82571EB_FIBER:
 		/* Wake events only supported on port A for dual fiber */
 		if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1)
 			return wol->wolopts ? -EOPNOTSUPP : 0;
@@ -1721,21 +1753,21 @@ e1000_phys_id(struct net_device *netdev, uint32_t data)
 		mod_timer(&adapter->blink_timer, jiffies);
 		msleep_interruptible(data * 1000);
 		del_timer_sync(&adapter->blink_timer);
-	}
-	else if(adapter->hw.mac_type < e1000_82573) {
-		E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE |
-			E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK |
-			(E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) |
-			(E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) |
-			(E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT)));
+	} else if (adapter->hw.mac_type < e1000_82573) {
+		E1000_WRITE_REG(&adapter->hw, LEDCTL,
+			(E1000_LEDCTL_LED2_BLINK_RATE |
+			 E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK |
+			 (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) |
+			 (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) |
+			 (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT)));
 		msleep_interruptible(data * 1000);
-	}
-	else {
-		E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE |
-			E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK | 
-			(E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) |
-			(E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) |
-			(E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT)));
+	} else {
+		E1000_WRITE_REG(&adapter->hw, LEDCTL,
+			(E1000_LEDCTL_LED2_BLINK_RATE |
+			 E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK |
+			 (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) |
+			 (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) |
+			 (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT)));
 		msleep_interruptible(data * 1000);
 	}
 
@@ -1768,19 +1800,43 @@ e1000_get_ethtool_stats(struct net_device *netdev,
 		struct ethtool_stats *stats, uint64_t *data)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
+#ifdef CONFIG_E1000_MQ
+	uint64_t *queue_stat;
+	int stat_count = sizeof(struct e1000_queue_stats) / sizeof(uint64_t);
+	int j, k;
+#endif
 	int i;
 
 	e1000_update_stats(adapter);
-	for(i = 0; i < E1000_STATS_LEN; i++) {
-		char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset;	
-		data[i] = (e1000_gstrings_stats[i].sizeof_stat == 
+	for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) {
+		char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset;
+		data[i] = (e1000_gstrings_stats[i].sizeof_stat ==
 			sizeof(uint64_t)) ? *(uint64_t *)p : *(uint32_t *)p;
 	}
+#ifdef CONFIG_E1000_MQ
+	for (j = 0; j < adapter->num_tx_queues; j++) {
+		queue_stat = (uint64_t *)&adapter->tx_ring[j].tx_stats;
+		for (k = 0; k < stat_count; k++)
+			data[i + k] = queue_stat[k];
+		i += k;
+	}
+	for (j = 0; j < adapter->num_rx_queues; j++) {
+		queue_stat = (uint64_t *)&adapter->rx_ring[j].rx_stats;
+		for (k = 0; k < stat_count; k++)
+			data[i + k] = queue_stat[k];
+		i += k;
+	}
+#endif
+/*	BUG_ON(i != E1000_STATS_LEN); */
 }
 
 static void 
 e1000_get_strings(struct net_device *netdev, uint32_t stringset, uint8_t *data)
 {
+#ifdef CONFIG_E1000_MQ
+	struct e1000_adapter *adapter = netdev_priv(netdev);
+#endif
+	uint8_t *p = data;
 	int i;
 
 	switch(stringset) {
@@ -1789,11 +1845,26 @@ e1000_get_strings(struct net_device *netdev, uint32_t stringset, uint8_t *data)
 			E1000_TEST_LEN*ETH_GSTRING_LEN);
 		break;
 	case ETH_SS_STATS:
-		for (i=0; i < E1000_STATS_LEN; i++) {
-			memcpy(data + i * ETH_GSTRING_LEN, 
-			e1000_gstrings_stats[i].stat_string,
-			ETH_GSTRING_LEN);
+		for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) {
+			memcpy(p, e1000_gstrings_stats[i].stat_string,
+			       ETH_GSTRING_LEN);
+			p += ETH_GSTRING_LEN;
+		}
+#ifdef CONFIG_E1000_MQ
+		for (i = 0; i < adapter->num_tx_queues; i++) {
+			sprintf(p, "tx_queue_%u_packets", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "tx_queue_%u_bytes", i);
+			p += ETH_GSTRING_LEN;
 		}
+		for (i = 0; i < adapter->num_rx_queues; i++) {
+			sprintf(p, "rx_queue_%u_packets", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "rx_queue_%u_bytes", i);
+			p += ETH_GSTRING_LEN;
+		}
+#endif
+/*		BUG_ON(p - data != E1000_STATS_LEN * ETH_GSTRING_LEN); */
 		break;
 	}
 }

drivers/net/e1000/e1000_hw.c

@@ -318,6 +318,8 @@ e1000_set_mac_type(struct e1000_hw *hw)
     case E1000_DEV_ID_82546GB_FIBER:
     case E1000_DEV_ID_82546GB_SERDES:
     case E1000_DEV_ID_82546GB_PCIE:
+    case E1000_DEV_ID_82546GB_QUAD_COPPER:
+    case E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3:
         hw->mac_type = e1000_82546_rev_3;
         break;
     case E1000_DEV_ID_82541EI:
@@ -639,6 +641,7 @@ e1000_init_hw(struct e1000_hw *hw)
     uint16_t cmd_mmrbc;
     uint16_t stat_mmrbc;
     uint32_t mta_size;
+    uint32_t ctrl_ext;
 
     DEBUGFUNC("e1000_init_hw");
 
@@ -735,7 +738,6 @@ e1000_init_hw(struct e1000_hw *hw)
             break;
         case e1000_82571:
         case e1000_82572:
-            ctrl |= (1 << 22);
         case e1000_82573:
             ctrl |= E1000_TXDCTL_COUNT_DESC;
             break;
@@ -775,6 +777,15 @@ e1000_init_hw(struct e1000_hw *hw)
      */
     e1000_clear_hw_cntrs(hw);
 
+    if (hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER ||
+        hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3) {
+        ctrl_ext = E1000_READ_REG(hw, CTRL_EXT);
+        /* Relaxed ordering must be disabled to avoid a parity
+         * error crash in a PCI slot. */
+        ctrl_ext |= E1000_CTRL_EXT_RO_DIS;
+        E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext);
+    }
+
     return ret_val;
 }
 
@@ -838,6 +849,11 @@ e1000_setup_link(struct e1000_hw *hw)
 
     DEBUGFUNC("e1000_setup_link");
 
+    /* In the case of the phy reset being blocked, we already have a link.
+     * We do not have to set it up again. */
+    if (e1000_check_phy_reset_block(hw))
+        return E1000_SUCCESS;
+
     /* Read and store word 0x0F of the EEPROM. This word contains bits
      * that determine the hardware's default PAUSE (flow control) mode,
      * a bit that determines whether the HW defaults to enabling or
@@ -1929,14 +1945,19 @@ e1000_phy_force_speed_duplex(struct e1000_hw *hw)
 void
 e1000_config_collision_dist(struct e1000_hw *hw)
 {
-    uint32_t tctl;
+    uint32_t tctl, coll_dist;
 
     DEBUGFUNC("e1000_config_collision_dist");
 
+    if (hw->mac_type < e1000_82543)
+        coll_dist = E1000_COLLISION_DISTANCE_82542;
+    else
+        coll_dist = E1000_COLLISION_DISTANCE;
+
     tctl = E1000_READ_REG(hw, TCTL);
 
     tctl &= ~E1000_TCTL_COLD;
-    tctl |= E1000_COLLISION_DISTANCE << E1000_COLD_SHIFT;
+    tctl |= coll_dist << E1000_COLD_SHIFT;
 
     E1000_WRITE_REG(hw, TCTL, tctl);
     E1000_WRITE_FLUSH(hw);
@@ -2982,6 +3003,8 @@ e1000_phy_hw_reset(struct e1000_hw *hw)
         
         if (hw->mac_type < e1000_82571) 
             msec_delay(10);
+        else
+            udelay(100);
         
         E1000_WRITE_REG(hw, CTRL, ctrl);
         E1000_WRITE_FLUSH(hw);
@@ -3881,14 +3904,16 @@ e1000_read_eeprom(struct e1000_hw *hw,
         return -E1000_ERR_EEPROM;
     }
 
-    /* FLASH reads without acquiring the semaphore are safe in 82573-based
-     * controllers.
-     */
-    if ((e1000_is_onboard_nvm_eeprom(hw) == TRUE) ||
-        (hw->mac_type != e1000_82573)) {
-        /* Prepare the EEPROM for reading  */
-        if(e1000_acquire_eeprom(hw) != E1000_SUCCESS)
-            return -E1000_ERR_EEPROM;
+    /* FLASH reads without acquiring the semaphore are safe */
+    if (e1000_is_onboard_nvm_eeprom(hw) == TRUE &&
+    hw->eeprom.use_eerd == FALSE) {
+        switch (hw->mac_type) {
+        default:
+            /* Prepare the EEPROM for reading  */
+            if (e1000_acquire_eeprom(hw) != E1000_SUCCESS)
+                return -E1000_ERR_EEPROM;
+            break;
+        }
     }
 
     if(eeprom->use_eerd == TRUE) {
@@ -6720,6 +6745,12 @@ e1000_get_phy_cfg_done(struct e1000_hw *hw)
         break;
     }
 
+    /* PHY configuration from NVM just starts after EECD_AUTO_RD sets to high.
+     * Need to wait for PHY configuration completion before accessing NVM
+     * and PHY. */
+    if (hw->mac_type == e1000_82573)
+        msec_delay(25);
+
     return E1000_SUCCESS;
 }
 

drivers/net/e1000/e1000_hw.h

@@ -439,6 +439,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw);
 #define E1000_DEV_ID_82546GB_FIBER       0x107A
 #define E1000_DEV_ID_82546GB_SERDES      0x107B
 #define E1000_DEV_ID_82546GB_PCIE        0x108A
+#define E1000_DEV_ID_82546GB_QUAD_COPPER 0x1099
 #define E1000_DEV_ID_82547EI             0x1019
 #define E1000_DEV_ID_82571EB_COPPER      0x105E
 #define E1000_DEV_ID_82571EB_FIBER       0x105F
@@ -449,6 +450,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw);
 #define E1000_DEV_ID_82573E              0x108B
 #define E1000_DEV_ID_82573E_IAMT         0x108C
 #define E1000_DEV_ID_82573L              0x109A
+#define E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3 0x10B5
 
 
 #define NODE_ADDRESS_SIZE 6
@@ -1497,6 +1499,7 @@ struct e1000_hw {
 #define E1000_CTRL_EXT_EE_RST    0x00002000 /* Reinitialize from EEPROM */
 #define E1000_CTRL_EXT_IPS       0x00004000 /* Invert Power State */
 #define E1000_CTRL_EXT_SPD_BYPS  0x00008000 /* Speed Select Bypass */
+#define E1000_CTRL_EXT_RO_DIS    0x00020000 /* Relaxed Ordering disable */
 #define E1000_CTRL_EXT_LINK_MODE_MASK 0x00C00000
 #define E1000_CTRL_EXT_LINK_MODE_GMII 0x00000000
 #define E1000_CTRL_EXT_LINK_MODE_TBI  0x00C00000
@@ -1954,6 +1957,23 @@ struct e1000_host_command_info {
 
 #define E1000_MDALIGN          4096
 
+/* PCI-Ex registers */
+
+/* PCI-Ex Control Register */
+#define E1000_GCR_RXD_NO_SNOOP			0x00000001
+#define E1000_GCR_RXDSCW_NO_SNOOP		0x00000002
+#define E1000_GCR_RXDSCR_NO_SNOOP		0x00000004
+#define E1000_GCR_TXD_NO_SNOOP			0x00000008
+#define E1000_GCR_TXDSCW_NO_SNOOP		0x00000010
+#define E1000_GCR_TXDSCR_NO_SNOOP		0x00000020
+
+#define PCI_EX_NO_SNOOP_ALL (E1000_GCR_RXD_NO_SNOOP		| \
+							 E1000_GCR_RXDSCW_NO_SNOOP	| \
+							 E1000_GCR_RXDSCR_NO_SNOOP	| \
+							 E1000_GCR TXD_NO_SNOOP		| \
+							 E1000_GCR_TXDSCW_NO_SNOOP	| \
+							 E1000_GCR_TXDSCR_NO_SNOOP)
+
 #define E1000_GCR_L1_ACT_WITHOUT_L0S_RX 0x08000000
 /* Function Active and Power State to MNG */
 #define E1000_FACTPS_FUNC0_POWER_STATE_MASK         0x00000003
@@ -2077,7 +2097,10 @@ struct e1000_host_command_info {
 /* Collision related configuration parameters */
 #define E1000_COLLISION_THRESHOLD       15
 #define E1000_CT_SHIFT                  4
-#define E1000_COLLISION_DISTANCE        64
+/* Collision distance is a 0-based value that applies to
+ * half-duplex-capable hardware only. */
+#define E1000_COLLISION_DISTANCE        63
+#define E1000_COLLISION_DISTANCE_82542  64
 #define E1000_FDX_COLLISION_DISTANCE    E1000_COLLISION_DISTANCE
 #define E1000_HDX_COLLISION_DISTANCE    E1000_COLLISION_DISTANCE
 #define E1000_COLD_SHIFT                12

drivers/net/e1000/e1000_main.c

@@ -43,7 +43,7 @@ static char e1000_driver_string[] = "Intel(R) PRO/1000 Network Driver";
 #else
 #define DRIVERNAPI "-NAPI"
 #endif
-#define DRV_VERSION "6.1.16-k2"DRIVERNAPI
+#define DRV_VERSION "6.3.9-k2"DRIVERNAPI
 char e1000_driver_version[] = DRV_VERSION;
 static char e1000_copyright[] = "Copyright (c) 1999-2005 Intel Corporation.";
 
@@ -97,7 +97,9 @@ static struct pci_device_id e1000_pci_tbl[] = {
 	INTEL_E1000_ETHERNET_DEVICE(0x108A),
 	INTEL_E1000_ETHERNET_DEVICE(0x108B),
 	INTEL_E1000_ETHERNET_DEVICE(0x108C),
+	INTEL_E1000_ETHERNET_DEVICE(0x1099),
 	INTEL_E1000_ETHERNET_DEVICE(0x109A),
+	INTEL_E1000_ETHERNET_DEVICE(0x10B5),
 	/* required last entry */
 	{0,}
 };
@@ -171,9 +173,11 @@ static boolean_t e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
                                        struct e1000_rx_ring *rx_ring);
 #endif
 static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
-                                   struct e1000_rx_ring *rx_ring);
+                                   struct e1000_rx_ring *rx_ring,
+				   int cleaned_count);
 static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter,
-                                      struct e1000_rx_ring *rx_ring);
+                                      struct e1000_rx_ring *rx_ring,
+				      int cleaned_count);
 static int e1000_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd);
 static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr,
 			   int cmd);
@@ -319,7 +323,75 @@ e1000_update_mng_vlan(struct e1000_adapter *adapter)
 		}
 	}
 }
-	
+
+/**
+ * e1000_release_hw_control - release control of the h/w to f/w
+ * @adapter: address of board private structure
+ *
+ * e1000_release_hw_control resets {CTRL_EXT|FWSM}:DRV_LOAD bit.
+ * For ASF and Pass Through versions of f/w this means that the
+ * driver is no longer loaded. For AMT version (only with 82573) i
+ * of the f/w this means that the netowrk i/f is closed.
+ * 
+ **/
+
+static inline void 
+e1000_release_hw_control(struct e1000_adapter *adapter)
+{
+	uint32_t ctrl_ext;
+	uint32_t swsm;
+
+	/* Let firmware taken over control of h/w */
+	switch (adapter->hw.mac_type) {
+	case e1000_82571:
+	case e1000_82572:
+		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
+		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
+				ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
+		break;
+	case e1000_82573:
+		swsm = E1000_READ_REG(&adapter->hw, SWSM);
+		E1000_WRITE_REG(&adapter->hw, SWSM,
+				swsm & ~E1000_SWSM_DRV_LOAD);
+	default:
+		break;
+	}
+}
+
+/**
+ * e1000_get_hw_control - get control of the h/w from f/w
+ * @adapter: address of board private structure
+ *
+ * e1000_get_hw_control sets {CTRL_EXT|FWSM}:DRV_LOAD bit.
+ * For ASF and Pass Through versions of f/w this means that 
+ * the driver is loaded. For AMT version (only with 82573) 
+ * of the f/w this means that the netowrk i/f is open.
+ * 
+ **/
+
+static inline void 
+e1000_get_hw_control(struct e1000_adapter *adapter)
+{
+	uint32_t ctrl_ext;
+	uint32_t swsm;
+	/* Let firmware know the driver has taken over */
+	switch (adapter->hw.mac_type) {
+	case e1000_82571:
+	case e1000_82572:
+		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
+		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
+				ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
+		break;
+	case e1000_82573:
+		swsm = E1000_READ_REG(&adapter->hw, SWSM);
+		E1000_WRITE_REG(&adapter->hw, SWSM,
+				swsm | E1000_SWSM_DRV_LOAD);
+		break;
+	default:
+		break;
+	}
+}
+
 int
 e1000_up(struct e1000_adapter *adapter)
 {
@@ -343,8 +415,14 @@ e1000_up(struct e1000_adapter *adapter)
 	e1000_configure_tx(adapter);
 	e1000_setup_rctl(adapter);
 	e1000_configure_rx(adapter);
-	for (i = 0; i < adapter->num_queues; i++)
-		adapter->alloc_rx_buf(adapter, &adapter->rx_ring[i]);
+	/* call E1000_DESC_UNUSED which always leaves
+	 * at least 1 descriptor unused to make sure
+	 * next_to_use != next_to_clean */
+	for (i = 0; i < adapter->num_rx_queues; i++) {
+		struct e1000_rx_ring *ring = &adapter->rx_ring[i];
+		adapter->alloc_rx_buf(adapter, ring,
+		                      E1000_DESC_UNUSED(ring));
+	}
 
 #ifdef CONFIG_PCI_MSI
 	if(adapter->hw.mac_type > e1000_82547_rev_2) {
@@ -364,6 +442,12 @@ e1000_up(struct e1000_adapter *adapter)
 		return err;
 	}
 
+#ifdef CONFIG_E1000_MQ
+	e1000_setup_queue_mapping(adapter);
+#endif
+
+	adapter->tx_queue_len = netdev->tx_queue_len;
+
 	mod_timer(&adapter->watchdog_timer, jiffies);
 
 #ifdef CONFIG_E1000_NAPI
@@ -378,6 +462,8 @@ void
 e1000_down(struct e1000_adapter *adapter)
 {
 	struct net_device *netdev = adapter->netdev;
+	boolean_t mng_mode_enabled = (adapter->hw.mac_type >= e1000_82571) &&
+				     e1000_check_mng_mode(&adapter->hw);
 
 	e1000_irq_disable(adapter);
 #ifdef CONFIG_E1000_MQ
@@ -396,6 +482,7 @@ e1000_down(struct e1000_adapter *adapter)
 #ifdef CONFIG_E1000_NAPI
 	netif_poll_disable(netdev);
 #endif
+	netdev->tx_queue_len = adapter->tx_queue_len;
 	adapter->link_speed = 0;
 	adapter->link_duplex = 0;
 	netif_carrier_off(netdev);
@@ -405,12 +492,16 @@ e1000_down(struct e1000_adapter *adapter)
 	e1000_clean_all_tx_rings(adapter);
 	e1000_clean_all_rx_rings(adapter);
 
-	/* If WoL is not enabled and management mode is not IAMT
-	 * Power down the PHY so no link is implied when interface is down */
-	if(!adapter->wol && adapter->hw.mac_type >= e1000_82540 &&
+	/* Power down the PHY so no link is implied when interface is down *
+	 * The PHY cannot be powered down if any of the following is TRUE *
+	 * (a) WoL is enabled
+	 * (b) AMT is active
+	 * (c) SoL/IDER session is active */
+	if (!adapter->wol && adapter->hw.mac_type >= e1000_82540 &&
 	   adapter->hw.media_type == e1000_media_type_copper &&
-	   !e1000_check_mng_mode(&adapter->hw) &&
-	   !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN)) {
+	   !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN) &&
+	   !mng_mode_enabled &&
+	   !e1000_check_phy_reset_block(&adapter->hw)) {
 		uint16_t mii_reg;
 		e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg);
 		mii_reg |= MII_CR_POWER_DOWN;
@@ -422,10 +513,8 @@ e1000_down(struct e1000_adapter *adapter)
 void
 e1000_reset(struct e1000_adapter *adapter)
 {
-	struct net_device *netdev = adapter->netdev;
 	uint32_t pba, manc;
 	uint16_t fc_high_water_mark = E1000_FC_HIGH_DIFF;
-	uint16_t fc_low_water_mark = E1000_FC_LOW_DIFF;
 
 	/* Repartition Pba for greater than 9k mtu
 	 * To take effect CTRL.RST is required.
@@ -449,15 +538,8 @@ e1000_reset(struct e1000_adapter *adapter)
 	}
 
 	if((adapter->hw.mac_type != e1000_82573) &&
-	   (adapter->rx_buffer_len > E1000_RXBUFFER_8192)) {
+	   (adapter->netdev->mtu > E1000_RXBUFFER_8192))
 		pba -= 8; /* allocate more FIFO for Tx */
-		/* send an XOFF when there is enough space in the
-		 * Rx FIFO to hold one extra full size Rx packet 
-		*/
-		fc_high_water_mark = netdev->mtu + ENET_HEADER_SIZE + 
-					ETHERNET_FCS_SIZE + 1;
-		fc_low_water_mark = fc_high_water_mark + 8;
-	}
 
 
 	if(adapter->hw.mac_type == e1000_82547) {
@@ -471,10 +553,12 @@ e1000_reset(struct e1000_adapter *adapter)
 	E1000_WRITE_REG(&adapter->hw, PBA, pba);
 
 	/* flow control settings */
-	adapter->hw.fc_high_water = (pba << E1000_PBA_BYTES_SHIFT) -
-				    fc_high_water_mark;
-	adapter->hw.fc_low_water = (pba << E1000_PBA_BYTES_SHIFT) -
-				   fc_low_water_mark;
+	/* Set the FC high water mark to 90% of the FIFO size.
+	 * Required to clear last 3 LSB */
+	fc_high_water_mark = ((pba * 9216)/10) & 0xFFF8;
+
+	adapter->hw.fc_high_water = fc_high_water_mark;
+	adapter->hw.fc_low_water = fc_high_water_mark - 8;
 	adapter->hw.fc_pause_time = E1000_FC_PAUSE_TIME;
 	adapter->hw.fc_send_xon = 1;
 	adapter->hw.fc = adapter->hw.original_fc;
@@ -517,8 +601,6 @@ e1000_probe(struct pci_dev *pdev,
 	struct net_device *netdev;
 	struct e1000_adapter *adapter;
 	unsigned long mmio_start, mmio_len;
-	uint32_t ctrl_ext;
-	uint32_t swsm;
 
 	static int cards_found = 0;
 	int i, err, pci_using_dac;
@@ -712,8 +794,7 @@ e1000_probe(struct pci_dev *pdev,
 	case e1000_82546:
 	case e1000_82546_rev_3:
 	case e1000_82571:
-		if((E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1)
-		   && (adapter->hw.media_type == e1000_media_type_copper)) {
+		if(E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1){
 			e1000_read_eeprom(&adapter->hw,
 				EEPROM_INIT_CONTROL3_PORT_B, 1, &eeprom_data);
 			break;
@@ -727,25 +808,36 @@ e1000_probe(struct pci_dev *pdev,
 	if(eeprom_data & eeprom_apme_mask)
 		adapter->wol |= E1000_WUFC_MAG;
 
+	/* print bus type/speed/width info */
+	{
+	struct e1000_hw *hw = &adapter->hw;
+	DPRINTK(PROBE, INFO, "(PCI%s:%s:%s) ",
+		((hw->bus_type == e1000_bus_type_pcix) ? "-X" :
+		 (hw->bus_type == e1000_bus_type_pci_express ? " Express":"")),
+		((hw->bus_speed == e1000_bus_speed_2500) ? "2.5Gb/s" :
+		 (hw->bus_speed == e1000_bus_speed_133) ? "133MHz" :
+		 (hw->bus_speed == e1000_bus_speed_120) ? "120MHz" :
+		 (hw->bus_speed == e1000_bus_speed_100) ? "100MHz" :
+		 (hw->bus_speed == e1000_bus_speed_66) ? "66MHz" : "33MHz"),
+		((hw->bus_width == e1000_bus_width_64) ? "64-bit" :
+		 (hw->bus_width == e1000_bus_width_pciex_4) ? "Width x4" :
+		 (hw->bus_width == e1000_bus_width_pciex_1) ? "Width x1" :
+		 "32-bit"));
+	}
+
+	for (i = 0; i < 6; i++)
+		printk("%2.2x%c", netdev->dev_addr[i], i == 5 ? '\n' : ':');
+
 	/* reset the hardware with the new settings */
 	e1000_reset(adapter);
 
-	/* Let firmware know the driver has taken over */
-	switch(adapter->hw.mac_type) {
-	case e1000_82571:
-	case e1000_82572:
-		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
-		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
-				ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
-		break;
-	case e1000_82573:
-		swsm = E1000_READ_REG(&adapter->hw, SWSM);
-		E1000_WRITE_REG(&adapter->hw, SWSM,
-				swsm | E1000_SWSM_DRV_LOAD);
-		break;
-	default:
-		break;
-	}
+	/* If the controller is 82573 and f/w is AMT, do not set
+	 * DRV_LOAD until the interface is up.  For all other cases,
+	 * let the f/w know that the h/w is now under the control
+	 * of the driver. */
+	if (adapter->hw.mac_type != e1000_82573 ||
+	    !e1000_check_mng_mode(&adapter->hw))
+		e1000_get_hw_control(adapter);
 
 	strcpy(netdev->name, "eth%d");
 	if((err = register_netdev(netdev)))
@@ -782,8 +874,7 @@ e1000_remove(struct pci_dev *pdev)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct e1000_adapter *adapter = netdev_priv(netdev);
-	uint32_t ctrl_ext;
-	uint32_t manc, swsm;
+	uint32_t manc;
 #ifdef CONFIG_E1000_NAPI
 	int i;
 #endif
@@ -799,26 +890,13 @@ e1000_remove(struct pci_dev *pdev)
 		}
 	}
 
-	switch(adapter->hw.mac_type) {
-	case e1000_82571:
-	case e1000_82572:
-		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
-		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
-				ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
-		break;
-	case e1000_82573:
-		swsm = E1000_READ_REG(&adapter->hw, SWSM);
-		E1000_WRITE_REG(&adapter->hw, SWSM,
-				swsm & ~E1000_SWSM_DRV_LOAD);
-		break;
-
-	default:
-		break;
-	}
+	/* Release control of h/w to f/w.  If f/w is AMT enabled, this
+	 * would have already happened in close and is redundant. */
+	e1000_release_hw_control(adapter);
 
 	unregister_netdev(netdev);
 #ifdef CONFIG_E1000_NAPI
-	for (i = 0; i < adapter->num_queues; i++)
+	for (i = 0; i < adapter->num_rx_queues; i++)
 		__dev_put(&adapter->polling_netdev[i]);
 #endif
 
@@ -923,15 +1001,34 @@ e1000_sw_init(struct e1000_adapter *adapter)
 	switch (hw->mac_type) {
 	case e1000_82571:
 	case e1000_82572:
-		adapter->num_queues = 2;
+		/* These controllers support 2 tx queues, but with a single
+		 * qdisc implementation, multiple tx queues aren't quite as
+		 * interesting.  If we can find a logical way of mapping
+		 * flows to a queue, then perhaps we can up the num_tx_queue
+		 * count back to its default.  Until then, we run the risk of
+		 * terrible performance due to SACK overload. */
+		adapter->num_tx_queues = 1;
+		adapter->num_rx_queues = 2;
 		break;
 	default:
-		adapter->num_queues = 1;
+		adapter->num_tx_queues = 1;
+		adapter->num_rx_queues = 1;
 		break;
 	}
-	adapter->num_queues = min(adapter->num_queues, num_online_cpus());
+	adapter->num_rx_queues = min(adapter->num_rx_queues, num_online_cpus());
+	adapter->num_tx_queues = min(adapter->num_tx_queues, num_online_cpus());
+	DPRINTK(DRV, INFO, "Multiqueue Enabled: Rx Queue count = %u %s\n",
+		adapter->num_rx_queues,
+		((adapter->num_rx_queues == 1)
+		 ? ((num_online_cpus() > 1)
+			? "(due to unsupported feature in current adapter)"
+			: "(due to unsupported system configuration)")
+		 : ""));
+	DPRINTK(DRV, INFO, "Multiqueue Enabled: Tx Queue count = %u\n",
+		adapter->num_tx_queues);
 #else
-	adapter->num_queues = 1;
+	adapter->num_tx_queues = 1;
+	adapter->num_rx_queues = 1;
 #endif
 
 	if (e1000_alloc_queues(adapter)) {
@@ -940,17 +1037,14 @@ e1000_sw_init(struct e1000_adapter *adapter)
 	}
 
 #ifdef CONFIG_E1000_NAPI
-	for (i = 0; i < adapter->num_queues; i++) {
+	for (i = 0; i < adapter->num_rx_queues; i++) {
 		adapter->polling_netdev[i].priv = adapter;
 		adapter->polling_netdev[i].poll = &e1000_clean;
 		adapter->polling_netdev[i].weight = 64;
 		dev_hold(&adapter->polling_netdev[i]);
 		set_bit(__LINK_STATE_START, &adapter->polling_netdev[i].state);
 	}
-#endif
-
-#ifdef CONFIG_E1000_MQ
-	e1000_setup_queue_mapping(adapter);
+	spin_lock_init(&adapter->tx_queue_lock);
 #endif
 
 	atomic_set(&adapter->irq_sem, 1);
@@ -973,13 +1067,13 @@ e1000_alloc_queues(struct e1000_adapter *adapter)
 {
 	int size;
 
-	size = sizeof(struct e1000_tx_ring) * adapter->num_queues;
+	size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues;
 	adapter->tx_ring = kmalloc(size, GFP_KERNEL);
 	if (!adapter->tx_ring)
 		return -ENOMEM;
 	memset(adapter->tx_ring, 0, size);
 
-	size = sizeof(struct e1000_rx_ring) * adapter->num_queues;
+	size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues;
 	adapter->rx_ring = kmalloc(size, GFP_KERNEL);
 	if (!adapter->rx_ring) {
 		kfree(adapter->tx_ring);
@@ -988,7 +1082,7 @@ e1000_alloc_queues(struct e1000_adapter *adapter)
 	memset(adapter->rx_ring, 0, size);
 
 #ifdef CONFIG_E1000_NAPI
-	size = sizeof(struct net_device) * adapter->num_queues;
+	size = sizeof(struct net_device) * adapter->num_rx_queues;
 	adapter->polling_netdev = kmalloc(size, GFP_KERNEL);
 	if (!adapter->polling_netdev) {
 		kfree(adapter->tx_ring);
@@ -998,6 +1092,14 @@ e1000_alloc_queues(struct e1000_adapter *adapter)
 	memset(adapter->polling_netdev, 0, size);
 #endif
 
+#ifdef CONFIG_E1000_MQ
+	adapter->rx_sched_call_data.func = e1000_rx_schedule;
+	adapter->rx_sched_call_data.info = adapter->netdev;
+
+	adapter->cpu_netdev = alloc_percpu(struct net_device *);
+	adapter->cpu_tx_ring = alloc_percpu(struct e1000_tx_ring *);
+#endif
+
 	return E1000_SUCCESS;
 }
 
@@ -1017,14 +1119,15 @@ e1000_setup_queue_mapping(struct e1000_adapter *adapter)
 	lock_cpu_hotplug();
 	i = 0;
 	for_each_online_cpu(cpu) {
-		*per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_queues];
+		*per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_tx_queues];
 		/* This is incomplete because we'd like to assign separate
 		 * physical cpus to these netdev polling structures and
 		 * avoid saturating a subset of cpus.
 		 */
-		if (i < adapter->num_queues) {
+		if (i < adapter->num_rx_queues) {
 			*per_cpu_ptr(adapter->cpu_netdev, cpu) = &adapter->polling_netdev[i];
-			adapter->cpu_for_queue[i] = cpu;
+			adapter->rx_ring[i].cpu = cpu;
+			cpu_set(cpu, adapter->cpumask);
 		} else
 			*per_cpu_ptr(adapter->cpu_netdev, cpu) = NULL;
 
@@ -1071,6 +1174,12 @@ e1000_open(struct net_device *netdev)
 		e1000_update_mng_vlan(adapter);
 	}
 
+	/* If AMT is enabled, let the firmware know that the network
+	 * interface is now open */
+	if (adapter->hw.mac_type == e1000_82573 &&
+	    e1000_check_mng_mode(&adapter->hw))
+		e1000_get_hw_control(adapter);
+
 	return E1000_SUCCESS;
 
 err_up:
@@ -1109,6 +1218,13 @@ e1000_close(struct net_device *netdev)
 			  E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) {
 		e1000_vlan_rx_kill_vid(netdev, adapter->mng_vlan_id);
 	}
+
+	/* If AMT is enabled, let the firmware know that the network
+	 * interface is now closed */
+	if (adapter->hw.mac_type == e1000_82573 &&
+	    e1000_check_mng_mode(&adapter->hw))
+		e1000_release_hw_control(adapter);
+
 	return 0;
 }
 
@@ -1229,7 +1345,7 @@ e1000_setup_all_tx_resources(struct e1000_adapter *adapter)
 {
 	int i, err = 0;
 
-	for (i = 0; i < adapter->num_queues; i++) {
+	for (i = 0; i < adapter->num_tx_queues; i++) {
 		err = e1000_setup_tx_resources(adapter, &adapter->tx_ring[i]);
 		if (err) {
 			DPRINTK(PROBE, ERR,
@@ -1254,10 +1370,11 @@ e1000_configure_tx(struct e1000_adapter *adapter)
 	uint64_t tdba;
 	struct e1000_hw *hw = &adapter->hw;
 	uint32_t tdlen, tctl, tipg, tarc;
+	uint32_t ipgr1, ipgr2;
 
 	/* Setup the HW Tx Head and Tail descriptor pointers */
 
-	switch (adapter->num_queues) {
+	switch (adapter->num_tx_queues) {
 	case 2:
 		tdba = adapter->tx_ring[1].dma;
 		tdlen = adapter->tx_ring[1].count *
@@ -1287,22 +1404,26 @@ e1000_configure_tx(struct e1000_adapter *adapter)
 
 	/* Set the default values for the Tx Inter Packet Gap timer */
 
+	if (hw->media_type == e1000_media_type_fiber ||
+	    hw->media_type == e1000_media_type_internal_serdes)
+		tipg = DEFAULT_82543_TIPG_IPGT_FIBER;
+	else
+		tipg = DEFAULT_82543_TIPG_IPGT_COPPER;
+
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
 		tipg = DEFAULT_82542_TIPG_IPGT;
-		tipg |= DEFAULT_82542_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT;
-		tipg |= DEFAULT_82542_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT;
+		ipgr1 = DEFAULT_82542_TIPG_IPGR1;
+		ipgr2 = DEFAULT_82542_TIPG_IPGR2;
 		break;
 	default:
-		if (hw->media_type == e1000_media_type_fiber ||
-		    hw->media_type == e1000_media_type_internal_serdes)
-			tipg = DEFAULT_82543_TIPG_IPGT_FIBER;
-		else
-			tipg = DEFAULT_82543_TIPG_IPGT_COPPER;
-		tipg |= DEFAULT_82543_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT;
-		tipg |= DEFAULT_82543_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT;
+		ipgr1 = DEFAULT_82543_TIPG_IPGR1;
+		ipgr2 = DEFAULT_82543_TIPG_IPGR2;
+		break;
 	}
+	tipg |= ipgr1 << E1000_TIPG_IPGR1_SHIFT;
+	tipg |= ipgr2 << E1000_TIPG_IPGR2_SHIFT;
 	E1000_WRITE_REG(hw, TIPG, tipg);
 
 	/* Set the Tx Interrupt Delay register */
@@ -1454,6 +1575,8 @@ setup_rx_desc_die:
 
 	rxdr->next_to_clean = 0;
 	rxdr->next_to_use = 0;
+	rxdr->rx_skb_top = NULL;
+	rxdr->rx_skb_prev = NULL;
 
 	return 0;
 }
@@ -1475,7 +1598,7 @@ e1000_setup_all_rx_resources(struct e1000_adapter *adapter)
 {
 	int i, err = 0;
 
-	for (i = 0; i < adapter->num_queues; i++) {
+	for (i = 0; i < adapter->num_rx_queues; i++) {
 		err = e1000_setup_rx_resources(adapter, &adapter->rx_ring[i]);
 		if (err) {
 			DPRINTK(PROBE, ERR,
@@ -1510,7 +1633,10 @@ e1000_setup_rctl(struct e1000_adapter *adapter)
 		E1000_RCTL_LBM_NO | E1000_RCTL_RDMTS_HALF |
 		(adapter->hw.mc_filter_type << E1000_RCTL_MO_SHIFT);
 
-	if(adapter->hw.tbi_compatibility_on == 1)
+	if (adapter->hw.mac_type > e1000_82543)
+		rctl |= E1000_RCTL_SECRC;
+
+	if (adapter->hw.tbi_compatibility_on == 1)
 		rctl |= E1000_RCTL_SBP;
 	else
 		rctl &= ~E1000_RCTL_SBP;
@@ -1638,16 +1764,21 @@ e1000_configure_rx(struct e1000_adapter *adapter)
 	}
 
 	if (hw->mac_type >= e1000_82571) {
-		/* Reset delay timers after every interrupt */
 		ctrl_ext = E1000_READ_REG(hw, CTRL_EXT);
+		/* Reset delay timers after every interrupt */
 		ctrl_ext |= E1000_CTRL_EXT_CANC;
+#ifdef CONFIG_E1000_NAPI
+		/* Auto-Mask interrupts upon ICR read. */
+		ctrl_ext |= E1000_CTRL_EXT_IAME;
+#endif
 		E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext);
+		E1000_WRITE_REG(hw, IAM, ~0);
 		E1000_WRITE_FLUSH(hw);
 	}
 
 	/* Setup the HW Rx Head and Tail Descriptor Pointers and
 	 * the Base and Length of the Rx Descriptor Ring */
-	switch (adapter->num_queues) {
+	switch (adapter->num_rx_queues) {
 #ifdef CONFIG_E1000_MQ
 	case 2:
 		rdba = adapter->rx_ring[1].dma;
@@ -1674,7 +1805,7 @@ e1000_configure_rx(struct e1000_adapter *adapter)
 	}
 
 #ifdef CONFIG_E1000_MQ
-	if (adapter->num_queues > 1) {
+	if (adapter->num_rx_queues > 1) {
 		uint32_t random[10];
 
 		get_random_bytes(&random[0], 40);
@@ -1684,7 +1815,7 @@ e1000_configure_rx(struct e1000_adapter *adapter)
 			E1000_WRITE_REG(hw, RSSIM, 0);
 		}
 
-		switch (adapter->num_queues) {
+		switch (adapter->num_rx_queues) {
 		case 2:
 		default:
 			reta = 0x00800080;
@@ -1776,7 +1907,7 @@ e1000_free_all_tx_resources(struct e1000_adapter *adapter)
 {
 	int i;
 
-	for (i = 0; i < adapter->num_queues; i++)
+	for (i = 0; i < adapter->num_tx_queues; i++)
 		e1000_free_tx_resources(adapter, &adapter->tx_ring[i]);
 }
 
@@ -1789,12 +1920,10 @@ e1000_unmap_and_free_tx_resource(struct e1000_adapter *adapter,
 				buffer_info->dma,
 				buffer_info->length,
 				PCI_DMA_TODEVICE);
-		buffer_info->dma = 0;
 	}
-	if(buffer_info->skb) {
+	if (buffer_info->skb)
 		dev_kfree_skb_any(buffer_info->skb);
-		buffer_info->skb = NULL;
-	}
+	memset(buffer_info, 0, sizeof(struct e1000_buffer));
 }
 
 /**
@@ -1843,7 +1972,7 @@ e1000_clean_all_tx_rings(struct e1000_adapter *adapter)
 {
 	int i;
 
-	for (i = 0; i < adapter->num_queues; i++)
+	for (i = 0; i < adapter->num_tx_queues; i++)
 		e1000_clean_tx_ring(adapter, &adapter->tx_ring[i]);
 }
 
@@ -1887,7 +2016,7 @@ e1000_free_all_rx_resources(struct e1000_adapter *adapter)
 {
 	int i;
 
-	for (i = 0; i < adapter->num_queues; i++)
+	for (i = 0; i < adapter->num_rx_queues; i++)
 		e1000_free_rx_resources(adapter, &adapter->rx_ring[i]);
 }
 
@@ -1913,8 +2042,6 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter,
 	for(i = 0; i < rx_ring->count; i++) {
 		buffer_info = &rx_ring->buffer_info[i];
 		if(buffer_info->skb) {
-			ps_page = &rx_ring->ps_page[i];
-			ps_page_dma = &rx_ring->ps_page_dma[i];
 			pci_unmap_single(pdev,
 					 buffer_info->dma,
 					 buffer_info->length,
@@ -1922,19 +2049,30 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter,
 
 			dev_kfree_skb(buffer_info->skb);
 			buffer_info->skb = NULL;
-
-			for(j = 0; j < adapter->rx_ps_pages; j++) {
-				if(!ps_page->ps_page[j]) break;
-				pci_unmap_single(pdev,
-						 ps_page_dma->ps_page_dma[j],
-						 PAGE_SIZE, PCI_DMA_FROMDEVICE);
-				ps_page_dma->ps_page_dma[j] = 0;
-				put_page(ps_page->ps_page[j]);
-				ps_page->ps_page[j] = NULL;
-			}
+		}
+		ps_page = &rx_ring->ps_page[i];
+		ps_page_dma = &rx_ring->ps_page_dma[i];
+		for (j = 0; j < adapter->rx_ps_pages; j++) {
+			if (!ps_page->ps_page[j]) break;
+			pci_unmap_page(pdev,
+				       ps_page_dma->ps_page_dma[j],
+				       PAGE_SIZE, PCI_DMA_FROMDEVICE);
+			ps_page_dma->ps_page_dma[j] = 0;
+			put_page(ps_page->ps_page[j]);
+			ps_page->ps_page[j] = NULL;
 		}
 	}
 
+	/* there also may be some cached data in our adapter */
+	if (rx_ring->rx_skb_top) {
+		dev_kfree_skb(rx_ring->rx_skb_top);
+
+		/* rx_skb_prev will be wiped out by rx_skb_top */
+		rx_ring->rx_skb_top = NULL;
+		rx_ring->rx_skb_prev = NULL;
+	}
+
+
 	size = sizeof(struct e1000_buffer) * rx_ring->count;
 	memset(rx_ring->buffer_info, 0, size);
 	size = sizeof(struct e1000_ps_page) * rx_ring->count;
@@ -1963,7 +2101,7 @@ e1000_clean_all_rx_rings(struct e1000_adapter *adapter)
 {
 	int i;
 
-	for (i = 0; i < adapter->num_queues; i++)
+	for (i = 0; i < adapter->num_rx_queues; i++)
 		e1000_clean_rx_ring(adapter, &adapter->rx_ring[i]);
 }
 
@@ -2005,7 +2143,9 @@ e1000_leave_82542_rst(struct e1000_adapter *adapter)
 
 	if(netif_running(netdev)) {
 		e1000_configure_rx(adapter);
-		e1000_alloc_rx_buffers(adapter, &adapter->rx_ring[0]);
+		/* No need to loop, because 82542 supports only 1 queue */
+		struct e1000_rx_ring *ring = &adapter->rx_ring[0];
+		adapter->alloc_rx_buf(adapter, ring, E1000_DESC_UNUSED(ring));
 	}
 }
 
@@ -2204,7 +2344,7 @@ static void
 e1000_watchdog_task(struct e1000_adapter *adapter)
 {
 	struct net_device *netdev = adapter->netdev;
-	struct e1000_tx_ring *txdr = &adapter->tx_ring[0];
+	struct e1000_tx_ring *txdr = adapter->tx_ring;
 	uint32_t link;
 
 	e1000_check_for_link(&adapter->hw);
@@ -2231,6 +2371,21 @@ e1000_watchdog_task(struct e1000_adapter *adapter)
 			       adapter->link_duplex == FULL_DUPLEX ?
 			       "Full Duplex" : "Half Duplex");
 
+			/* tweak tx_queue_len according to speed/duplex */
+			netdev->tx_queue_len = adapter->tx_queue_len;
+			adapter->tx_timeout_factor = 1;
+			if (adapter->link_duplex == HALF_DUPLEX) {
+				switch (adapter->link_speed) {
+				case SPEED_10:
+					netdev->tx_queue_len = 10;
+					adapter->tx_timeout_factor = 8;
+					break;
+				case SPEED_100:
+					netdev->tx_queue_len = 100;
+					break;
+				}
+			}
+
 			netif_carrier_on(netdev);
 			netif_wake_queue(netdev);
 			mod_timer(&adapter->phy_info_timer, jiffies + 2 * HZ);
@@ -2263,7 +2418,10 @@ e1000_watchdog_task(struct e1000_adapter *adapter)
 
 	e1000_update_adaptive(&adapter->hw);
 
-	if (adapter->num_queues == 1 && !netif_carrier_ok(netdev)) {
+#ifdef CONFIG_E1000_MQ
+	txdr = *per_cpu_ptr(adapter->cpu_tx_ring, smp_processor_id());
+#endif
+	if (!netif_carrier_ok(netdev)) {
 		if (E1000_DESC_UNUSED(txdr) + 1 < txdr->count) {
 			/* We've lost link, so the controller stops DMA,
 			 * but we've got queued Tx work that's never going
@@ -2314,6 +2472,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
 {
 #ifdef NETIF_F_TSO
 	struct e1000_context_desc *context_desc;
+	struct e1000_buffer *buffer_info;
 	unsigned int i;
 	uint32_t cmd_length = 0;
 	uint16_t ipcse = 0, tucse, mss;
@@ -2363,6 +2522,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
 
 		i = tx_ring->next_to_use;
 		context_desc = E1000_CONTEXT_DESC(*tx_ring, i);
+		buffer_info = &tx_ring->buffer_info[i];
 
 		context_desc->lower_setup.ip_fields.ipcss  = ipcss;
 		context_desc->lower_setup.ip_fields.ipcso  = ipcso;
@@ -2374,14 +2534,16 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
 		context_desc->tcp_seg_setup.fields.hdr_len = hdr_len;
 		context_desc->cmd_and_length = cpu_to_le32(cmd_length);
 
+		buffer_info->time_stamp = jiffies;
+
 		if (++i == tx_ring->count) i = 0;
 		tx_ring->next_to_use = i;
 
-		return 1;
+		return TRUE;
 	}
 #endif
 
-	return 0;
+	return FALSE;
 }
 
 static inline boolean_t
@@ -2389,6 +2551,7 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
               struct sk_buff *skb)
 {
 	struct e1000_context_desc *context_desc;
+	struct e1000_buffer *buffer_info;
 	unsigned int i;
 	uint8_t css;
 
@@ -2396,6 +2559,7 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
 		css = skb->h.raw - skb->data;
 
 		i = tx_ring->next_to_use;
+		buffer_info = &tx_ring->buffer_info[i];
 		context_desc = E1000_CONTEXT_DESC(*tx_ring, i);
 
 		context_desc->upper_setup.tcp_fields.tucss = css;
@@ -2404,6 +2568,8 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring,
 		context_desc->tcp_seg_setup.data = 0;
 		context_desc->cmd_and_length = cpu_to_le32(E1000_TXD_CMD_DEXT);
 
+		buffer_info->time_stamp = jiffies;
+
 		if (unlikely(++i == tx_ring->count)) i = 0;
 		tx_ring->next_to_use = i;
 
@@ -2688,11 +2854,30 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 	 * overrun the FIFO, adjust the max buffer len if mss
 	 * drops. */
 	if(mss) {
+		uint8_t hdr_len;
 		max_per_txd = min(mss << 2, max_per_txd);
 		max_txd_pwr = fls(max_per_txd) - 1;
+
+	/* TSO Workaround for 82571/2 Controllers -- if skb->data
+	 * points to just header, pull a few bytes of payload from
+	 * frags into skb->data */
+		hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2));
+		if (skb->data_len && (hdr_len == (skb->len - skb->data_len)) &&
+			(adapter->hw.mac_type == e1000_82571 ||
+			adapter->hw.mac_type == e1000_82572)) {
+			unsigned int pull_size;
+			pull_size = min((unsigned int)4, skb->data_len);
+			if (!__pskb_pull_tail(skb, pull_size)) {
+				printk(KERN_ERR "__pskb_pull_tail failed.\n");
+				dev_kfree_skb_any(skb);
+				return -EFAULT;
+			}
+			len = skb->len - skb->data_len;
+		}
 	}
 
 	if((mss) || (skb->ip_summed == CHECKSUM_HW))
+	/* reserve a descriptor for the offload context */
 		count++;
 	count++;
 #else
@@ -2726,27 +2911,6 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 	if(adapter->pcix_82544)
 		count += nr_frags;
 
-#ifdef NETIF_F_TSO
-	/* TSO Workaround for 82571/2 Controllers -- if skb->data
-	 * points to just header, pull a few bytes of payload from 
-	 * frags into skb->data */
-	if (skb_shinfo(skb)->tso_size) {
-		uint8_t hdr_len;
-		hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2));
-		if (skb->data_len && (hdr_len < (skb->len - skb->data_len)) && 
-			(adapter->hw.mac_type == e1000_82571 ||
-			adapter->hw.mac_type == e1000_82572)) {
-			unsigned int pull_size;
-			pull_size = min((unsigned int)4, skb->data_len);
-			if (!__pskb_pull_tail(skb, pull_size)) {
-				printk(KERN_ERR "__pskb_pull_tail failed.\n");
-				dev_kfree_skb_any(skb);
-				return -EFAULT;
-			}
-		}
-	}
-#endif
-
 	if(adapter->hw.tx_pkt_filtering && (adapter->hw.mac_type == e1000_82573) )
 		e1000_transfer_dhcp_info(adapter, skb);
 
@@ -2833,6 +2997,7 @@ e1000_tx_timeout_task(struct net_device *netdev)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 
+	adapter->tx_timeout_count++;
 	e1000_down(adapter);
 	e1000_up(adapter);
 }
@@ -2850,7 +3015,7 @@ e1000_get_stats(struct net_device *netdev)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 
-	e1000_update_stats(adapter);
+	/* only return the current stats */
 	return &adapter->net_stats;
 }
 
@@ -2871,50 +3036,51 @@ e1000_change_mtu(struct net_device *netdev, int new_mtu)
 	if((max_frame < MINIMUM_ETHERNET_FRAME_SIZE) ||
 		(max_frame > MAX_JUMBO_FRAME_SIZE)) {
 			DPRINTK(PROBE, ERR, "Invalid MTU setting\n");
-			return -EINVAL;
-	}
-
-#define MAX_STD_JUMBO_FRAME_SIZE 9234
-	/* might want this to be bigger enum check... */
-	/* 82571 controllers limit jumbo frame size to 10500 bytes */
-	if ((adapter->hw.mac_type == e1000_82571 || 
-	     adapter->hw.mac_type == e1000_82572) &&
-	    max_frame > MAX_STD_JUMBO_FRAME_SIZE) {
-		DPRINTK(PROBE, ERR, "MTU > 9216 bytes not supported "
-				    "on 82571 and 82572 controllers.\n");
 		return -EINVAL;
 	}
 
-	if(adapter->hw.mac_type == e1000_82573 &&
-	    max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) {
-		DPRINTK(PROBE, ERR, "Jumbo Frames not supported "
-				    "on 82573\n");
-		return -EINVAL;
-	}
-
-	if(adapter->hw.mac_type > e1000_82547_rev_2) {
-		adapter->rx_buffer_len = max_frame;
-		E1000_ROUNDUP(adapter->rx_buffer_len, 1024);
-	} else {
-		if(unlikely((adapter->hw.mac_type < e1000_82543) &&
-		   (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE))) {
-			DPRINTK(PROBE, ERR, "Jumbo Frames not supported "
-					    "on 82542\n");
+	/* Adapter-specific max frame size limits. */
+	switch (adapter->hw.mac_type) {
+	case e1000_82542_rev2_0:
+	case e1000_82542_rev2_1:
+	case e1000_82573:
+		if (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) {
+			DPRINTK(PROBE, ERR, "Jumbo Frames not supported.\n");
+			return -EINVAL;
+		}
+		break;
+	case e1000_82571:
+	case e1000_82572:
+#define MAX_STD_JUMBO_FRAME_SIZE 9234
+		if (max_frame > MAX_STD_JUMBO_FRAME_SIZE) {
+			DPRINTK(PROBE, ERR, "MTU > 9216 not supported.\n");
 			return -EINVAL;
-
-		} else {
-			if(max_frame <= E1000_RXBUFFER_2048) {
-				adapter->rx_buffer_len = E1000_RXBUFFER_2048;
-			} else if(max_frame <= E1000_RXBUFFER_4096) {
-				adapter->rx_buffer_len = E1000_RXBUFFER_4096;
-			} else if(max_frame <= E1000_RXBUFFER_8192) {
-				adapter->rx_buffer_len = E1000_RXBUFFER_8192;
-			} else if(max_frame <= E1000_RXBUFFER_16384) {
-				adapter->rx_buffer_len = E1000_RXBUFFER_16384;
-			}
 		}
+		break;
+	default:
+		/* Capable of supporting up to MAX_JUMBO_FRAME_SIZE limit. */
+		break;
 	}
 
+	/* since the driver code now supports splitting a packet across
+	 * multiple descriptors, most of the fifo related limitations on
+	 * jumbo frame traffic have gone away.
+	 * simply use 2k descriptors for everything.
+	 *
+	 * NOTE: dev_alloc_skb reserves 16 bytes, and typically NET_IP_ALIGN
+	 * means we reserve 2 more, this pushes us to allocate from the next
+	 * larger slab size
+	 * i.e. RXBUFFER_2048 --> size-4096 slab */
+
+	/* recent hardware supports 1KB granularity */
+	if (adapter->hw.mac_type > e1000_82547_rev_2) {
+		adapter->rx_buffer_len =
+		    ((max_frame < E1000_RXBUFFER_2048) ?
+		        max_frame : E1000_RXBUFFER_2048);
+		E1000_ROUNDUP(adapter->rx_buffer_len, 1024);
+	} else
+		adapter->rx_buffer_len = E1000_RXBUFFER_2048;
+
 	netdev->mtu = new_mtu;
 
 	if(netif_running(netdev)) {
@@ -3037,12 +3203,11 @@ e1000_update_stats(struct e1000_adapter *adapter)
 
 	adapter->net_stats.rx_errors = adapter->stats.rxerrc +
 		adapter->stats.crcerrs + adapter->stats.algnerrc +
-		adapter->stats.rlec + adapter->stats.mpc + 
-		adapter->stats.cexterr;
+		adapter->stats.rlec + adapter->stats.cexterr;
+	adapter->net_stats.rx_dropped = 0;
 	adapter->net_stats.rx_length_errors = adapter->stats.rlec;
 	adapter->net_stats.rx_crc_errors = adapter->stats.crcerrs;
 	adapter->net_stats.rx_frame_errors = adapter->stats.algnerrc;
-	adapter->net_stats.rx_fifo_errors = adapter->stats.mpc;
 	adapter->net_stats.rx_missed_errors = adapter->stats.mpc;
 
 	/* Tx Errors */
@@ -3110,12 +3275,24 @@ e1000_intr(int irq, void *data, struct pt_regs *regs)
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
 	uint32_t icr = E1000_READ_REG(hw, ICR);
-#if defined(CONFIG_E1000_NAPI) && defined(CONFIG_E1000_MQ) || !defined(CONFIG_E1000_NAPI)
+#ifndef CONFIG_E1000_NAPI
 	int i;
+#else
+	/* Interrupt Auto-Mask...upon reading ICR,
+	 * interrupts are masked.  No need for the
+	 * IMC write, but it does mean we should
+	 * account for it ASAP. */
+	if (likely(hw->mac_type >= e1000_82571))
+		atomic_inc(&adapter->irq_sem);
 #endif
 
-	if(unlikely(!icr))
+	if (unlikely(!icr)) {
+#ifdef CONFIG_E1000_NAPI
+		if (hw->mac_type >= e1000_82571)
+			e1000_irq_enable(adapter);
+#endif
 		return IRQ_NONE;  /* Not our interrupt */
+	}
 
 	if(unlikely(icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC))) {
 		hw->get_link_status = 1;
@@ -3123,19 +3300,19 @@ e1000_intr(int irq, void *data, struct pt_regs *regs)
 	}
 
 #ifdef CONFIG_E1000_NAPI
-	atomic_inc(&adapter->irq_sem);
-	E1000_WRITE_REG(hw, IMC, ~0);
-	E1000_WRITE_FLUSH(hw);
+	if (unlikely(hw->mac_type < e1000_82571)) {
+		atomic_inc(&adapter->irq_sem);
+		E1000_WRITE_REG(hw, IMC, ~0);
+		E1000_WRITE_FLUSH(hw);
+	}
 #ifdef CONFIG_E1000_MQ
 	if (atomic_read(&adapter->rx_sched_call_data.count) == 0) {
-		cpu_set(adapter->cpu_for_queue[0],
-			adapter->rx_sched_call_data.cpumask);
-		for (i = 1; i < adapter->num_queues; i++) {
-			cpu_set(adapter->cpu_for_queue[i],
-				adapter->rx_sched_call_data.cpumask);
-			atomic_inc(&adapter->irq_sem);
-		}
-		atomic_set(&adapter->rx_sched_call_data.count, i);
+		/* We must setup the cpumask once count == 0 since
+		 * each cpu bit is cleared when the work is done. */
+		adapter->rx_sched_call_data.cpumask = adapter->cpumask;
+		atomic_add(adapter->num_rx_queues - 1, &adapter->irq_sem);
+		atomic_set(&adapter->rx_sched_call_data.count,
+		           adapter->num_rx_queues);
 		smp_call_async_mask(&adapter->rx_sched_call_data);
 	} else {
 		printk("call_data.count == %u\n", atomic_read(&adapter->rx_sched_call_data.count));
@@ -3187,7 +3364,7 @@ e1000_clean(struct net_device *poll_dev, int *budget)
 {
 	struct e1000_adapter *adapter;
 	int work_to_do = min(*budget, poll_dev->quota);
-	int tx_cleaned, i = 0, work_done = 0;
+	int tx_cleaned = 0, i = 0, work_done = 0;
 
 	/* Must NOT use netdev_priv macro here. */
 	adapter = poll_dev->priv;
@@ -3198,11 +3375,23 @@ e1000_clean(struct net_device *poll_dev, int *budget)
 
 	while (poll_dev != &adapter->polling_netdev[i]) {
 		i++;
-		if (unlikely(i == adapter->num_queues))
+		if (unlikely(i == adapter->num_rx_queues))
 			BUG();
 	}
 
-	tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]);
+	if (likely(adapter->num_tx_queues == 1)) {
+		/* e1000_clean is called per-cpu.  This lock protects
+		 * tx_ring[0] from being cleaned by multiple cpus
+		 * simultaneously.  A failure obtaining the lock means
+		 * tx_ring[0] is currently being cleaned anyway. */
+		if (spin_trylock(&adapter->tx_queue_lock)) {
+			tx_cleaned = e1000_clean_tx_irq(adapter,
+							&adapter->tx_ring[0]);
+			spin_unlock(&adapter->tx_queue_lock);
+		}
+	} else
+		tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]);
+
 	adapter->clean_rx(adapter, &adapter->rx_ring[i],
 	                  &work_done, work_to_do);
 
@@ -3247,17 +3436,19 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter,
 			buffer_info = &tx_ring->buffer_info[i];
 			cleaned = (i == eop);
 
+#ifdef CONFIG_E1000_MQ
+			tx_ring->tx_stats.bytes += buffer_info->length;
+#endif
 			e1000_unmap_and_free_tx_resource(adapter, buffer_info);
-
-			tx_desc->buffer_addr = 0;
-			tx_desc->lower.data = 0;
-			tx_desc->upper.data = 0;
+			memset(tx_desc, 0, sizeof(struct e1000_tx_desc));
 
 			if(unlikely(++i == tx_ring->count)) i = 0;
 		}
 
-		tx_ring->pkt++;
-		
+#ifdef CONFIG_E1000_MQ
+		tx_ring->tx_stats.packets++;
+#endif
+
 		eop = tx_ring->buffer_info[i].next_to_watch;
 		eop_desc = E1000_TX_DESC(*tx_ring, eop);
 	}
@@ -3276,32 +3467,31 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter,
 		/* Detect a transmit hang in hardware, this serializes the
 		 * check with the clearing of time_stamp and movement of i */
 		adapter->detect_tx_hung = FALSE;
-		if (tx_ring->buffer_info[i].dma &&
-		    time_after(jiffies, tx_ring->buffer_info[i].time_stamp + HZ)
+		if (tx_ring->buffer_info[eop].dma &&
+		    time_after(jiffies, tx_ring->buffer_info[eop].time_stamp +
+		               adapter->tx_timeout_factor * HZ)
 		    && !(E1000_READ_REG(&adapter->hw, STATUS) &
-			E1000_STATUS_TXOFF)) {
+		         E1000_STATUS_TXOFF)) {
 
 			/* detected Tx unit hang */
-			i = tx_ring->next_to_clean;
-			eop = tx_ring->buffer_info[i].next_to_watch;
-			eop_desc = E1000_TX_DESC(*tx_ring, eop);
 			DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n"
+					"  Tx Queue             <%lu>\n"
 					"  TDH                  <%x>\n"
 					"  TDT                  <%x>\n"
 					"  next_to_use          <%x>\n"
 					"  next_to_clean        <%x>\n"
 					"buffer_info[next_to_clean]\n"
-					"  dma                  <%llx>\n"
 					"  time_stamp           <%lx>\n"
 					"  next_to_watch        <%x>\n"
 					"  jiffies              <%lx>\n"
 					"  next_to_watch.status <%x>\n",
+				(unsigned long)((tx_ring - adapter->tx_ring) /
+					sizeof(struct e1000_tx_ring)),
 				readl(adapter->hw.hw_addr + tx_ring->tdh),
 				readl(adapter->hw.hw_addr + tx_ring->tdt),
 				tx_ring->next_to_use,
-				i,
-				(unsigned long long)tx_ring->buffer_info[i].dma,
-				tx_ring->buffer_info[i].time_stamp,
+				tx_ring->next_to_clean,
+				tx_ring->buffer_info[eop].time_stamp,
 				eop,
 				jiffies,
 				eop_desc->upper.fields.status);
@@ -3386,20 +3576,23 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter,
 	uint32_t length;
 	uint8_t last_byte;
 	unsigned int i;
-	boolean_t cleaned = FALSE;
+	int cleaned_count = 0;
+	boolean_t cleaned = FALSE, multi_descriptor = FALSE;
 
 	i = rx_ring->next_to_clean;
 	rx_desc = E1000_RX_DESC(*rx_ring, i);
 
 	while(rx_desc->status & E1000_RXD_STAT_DD) {
 		buffer_info = &rx_ring->buffer_info[i];
+		u8 status;
 #ifdef CONFIG_E1000_NAPI
 		if(*work_done >= work_to_do)
 			break;
 		(*work_done)++;
 #endif
+		status = rx_desc->status;
 		cleaned = TRUE;
-
+		cleaned_count++;
 		pci_unmap_single(pdev,
 		                 buffer_info->dma,
 		                 buffer_info->length,
@@ -3433,18 +3626,40 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter,
 			}
 		}
 
-		/* Good Receive */
-		skb_put(skb, length - ETHERNET_FCS_SIZE);
+		/* code added for copybreak, this should improve
+		 * performance for small packets with large amounts
+		 * of reassembly being done in the stack */
+#define E1000_CB_LENGTH 256
+		if ((length < E1000_CB_LENGTH) &&
+		   !rx_ring->rx_skb_top &&
+		   /* or maybe (status & E1000_RXD_STAT_EOP) && */
+		   !multi_descriptor) {
+			struct sk_buff *new_skb =
+			    dev_alloc_skb(length + NET_IP_ALIGN);
+			if (new_skb) {
+				skb_reserve(new_skb, NET_IP_ALIGN);
+				new_skb->dev = netdev;
+				memcpy(new_skb->data - NET_IP_ALIGN,
+				       skb->data - NET_IP_ALIGN,
+				       length + NET_IP_ALIGN);
+				/* save the skb in buffer_info as good */
+				buffer_info->skb = skb;
+				skb = new_skb;
+				skb_put(skb, length);
+			}
+		}
+
+		/* end copybreak code */
 
 		/* Receive Checksum Offload */
 		e1000_rx_checksum(adapter,
-				  (uint32_t)(rx_desc->status) |
+				  (uint32_t)(status) |
 				  ((uint32_t)(rx_desc->errors) << 24),
 				  rx_desc->csum, skb);
 		skb->protocol = eth_type_trans(skb, netdev);
 #ifdef CONFIG_E1000_NAPI
 		if(unlikely(adapter->vlgrp &&
-			    (rx_desc->status & E1000_RXD_STAT_VP))) {
+			    (status & E1000_RXD_STAT_VP))) {
 			vlan_hwaccel_receive_skb(skb, adapter->vlgrp,
 						 le16_to_cpu(rx_desc->special) &
 						 E1000_RXD_SPC_VLAN_MASK);
@@ -3462,17 +3677,26 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter,
 		}
 #endif /* CONFIG_E1000_NAPI */
 		netdev->last_rx = jiffies;
-		rx_ring->pkt++;
+#ifdef CONFIG_E1000_MQ
+		rx_ring->rx_stats.packets++;
+		rx_ring->rx_stats.bytes += length;
+#endif
 
 next_desc:
 		rx_desc->status = 0;
-		buffer_info->skb = NULL;
-		if(unlikely(++i == rx_ring->count)) i = 0;
 
-		rx_desc = E1000_RX_DESC(*rx_ring, i);
+		/* return some buffers to hardware, one at a time is too slow */
+		if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) {
+			adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
+			cleaned_count = 0;
+		}
+
 	}
 	rx_ring->next_to_clean = i;
-	adapter->alloc_rx_buf(adapter, rx_ring);
+
+	cleaned_count = E1000_DESC_UNUSED(rx_ring);
+	if (cleaned_count)
+		adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
 
 	return cleaned;
 }
@@ -3501,6 +3725,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
 	struct sk_buff *skb;
 	unsigned int i, j;
 	uint32_t length, staterr;
+	int cleaned_count = 0;
 	boolean_t cleaned = FALSE;
 
 	i = rx_ring->next_to_clean;
@@ -3517,6 +3742,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
 		(*work_done)++;
 #endif
 		cleaned = TRUE;
+		cleaned_count++;
 		pci_unmap_single(pdev, buffer_info->dma,
 				 buffer_info->length,
 				 PCI_DMA_FROMDEVICE);
@@ -3593,18 +3819,28 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
 		}
 #endif /* CONFIG_E1000_NAPI */
 		netdev->last_rx = jiffies;
-		rx_ring->pkt++;
+#ifdef CONFIG_E1000_MQ
+		rx_ring->rx_stats.packets++;
+		rx_ring->rx_stats.bytes += length;
+#endif
 
 next_desc:
 		rx_desc->wb.middle.status_error &= ~0xFF;
 		buffer_info->skb = NULL;
-		if(unlikely(++i == rx_ring->count)) i = 0;
 
-		rx_desc = E1000_RX_DESC_PS(*rx_ring, i);
+		/* return some buffers to hardware, one at a time is too slow */
+		if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) {
+			adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
+			cleaned_count = 0;
+		}
+
 		staterr = le32_to_cpu(rx_desc->wb.middle.status_error);
 	}
 	rx_ring->next_to_clean = i;
-	adapter->alloc_rx_buf(adapter, rx_ring);
+
+	cleaned_count = E1000_DESC_UNUSED(rx_ring);
+	if (cleaned_count)
+		adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count);
 
 	return cleaned;
 }
@@ -3616,7 +3852,8 @@ next_desc:
 
 static void
 e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
-                       struct e1000_rx_ring *rx_ring)
+                       struct e1000_rx_ring *rx_ring,
+		       int cleaned_count)
 {
 	struct net_device *netdev = adapter->netdev;
 	struct pci_dev *pdev = adapter->pdev;
@@ -3629,11 +3866,18 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
 	i = rx_ring->next_to_use;
 	buffer_info = &rx_ring->buffer_info[i];
 
-	while(!buffer_info->skb) {
-		skb = dev_alloc_skb(bufsz);
+	while (cleaned_count--) {
+		if (!(skb = buffer_info->skb))
+			skb = dev_alloc_skb(bufsz);
+		else {
+			skb_trim(skb, 0);
+			goto map_skb;
+		}
+
 
 		if(unlikely(!skb)) {
 			/* Better luck next round */
+			adapter->alloc_rx_buff_failed++;
 			break;
 		}
 
@@ -3670,6 +3914,7 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
 
 		buffer_info->skb = skb;
 		buffer_info->length = adapter->rx_buffer_len;
+map_skb:
 		buffer_info->dma = pci_map_single(pdev,
 						  skb->data,
 						  adapter->rx_buffer_len,
@@ -3718,7 +3963,8 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter,
 
 static void
 e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter,
-                          struct e1000_rx_ring *rx_ring)
+                          struct e1000_rx_ring *rx_ring,
+			  int cleaned_count)
 {
 	struct net_device *netdev = adapter->netdev;
 	struct pci_dev *pdev = adapter->pdev;
@@ -3734,7 +3980,7 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter,
 	ps_page = &rx_ring->ps_page[i];
 	ps_page_dma = &rx_ring->ps_page_dma[i];
 
-	while(!buffer_info->skb) {
+	while (cleaned_count--) {
 		rx_desc = E1000_RX_DESC_PS(*rx_ring, i);
 
 		for(j = 0; j < PS_PAGE_BUFFERS; j++) {
@@ -4106,8 +4352,12 @@ e1000_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid)
 
 	if((adapter->hw.mng_cookie.status &
 		E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) &&
-		(vid == adapter->mng_vlan_id))
+	    (vid == adapter->mng_vlan_id)) {
+		/* release control to f/w */
+		e1000_release_hw_control(adapter);
 		return;
+	}
+
 	/* remove VID from filter table */
 	index = (vid >> 5) & 0x7F;
 	vfta = E1000_READ_REG_ARRAY(&adapter->hw, VFTA, index);
@@ -4173,8 +4423,9 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct e1000_adapter *adapter = netdev_priv(netdev);
-	uint32_t ctrl, ctrl_ext, rctl, manc, status, swsm;
+	uint32_t ctrl, ctrl_ext, rctl, manc, status;
 	uint32_t wufc = adapter->wol;
+	int retval = 0;
 
 	netif_device_detach(netdev);
 
@@ -4220,13 +4471,21 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state)
 
 		E1000_WRITE_REG(&adapter->hw, WUC, E1000_WUC_PME_EN);
 		E1000_WRITE_REG(&adapter->hw, WUFC, wufc);
-		pci_enable_wake(pdev, 3, 1);
-		pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */
+		retval = pci_enable_wake(pdev, PCI_D3hot, 1);
+		if (retval)
+			DPRINTK(PROBE, ERR, "Error enabling D3 wake\n");
+		retval = pci_enable_wake(pdev, PCI_D3cold, 1);
+		if (retval)
+			DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n");
 	} else {
 		E1000_WRITE_REG(&adapter->hw, WUC, 0);
 		E1000_WRITE_REG(&adapter->hw, WUFC, 0);
-		pci_enable_wake(pdev, 3, 0);
-		pci_enable_wake(pdev, 4, 0); /* 4 == D3 cold */
+		retval = pci_enable_wake(pdev, PCI_D3hot, 0);
+		if (retval)
+			DPRINTK(PROBE, ERR, "Error enabling D3 wake\n");
+		retval = pci_enable_wake(pdev, PCI_D3cold, 0); /* 4 == D3 cold */
+		if (retval)
+			DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n");
 	}
 
 	pci_save_state(pdev);
@@ -4237,29 +4496,24 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state)
 		if(manc & E1000_MANC_SMBUS_EN) {
 			manc |= E1000_MANC_ARP_EN;
 			E1000_WRITE_REG(&adapter->hw, MANC, manc);
-			pci_enable_wake(pdev, 3, 1);
-			pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */
+			retval = pci_enable_wake(pdev, PCI_D3hot, 1);
+			if (retval)
+				DPRINTK(PROBE, ERR, "Error enabling D3 wake\n");
+			retval = pci_enable_wake(pdev, PCI_D3cold, 1);
+			if (retval)
+				DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n");
 		}
 	}
 
-	switch(adapter->hw.mac_type) {
-	case e1000_82571:
-	case e1000_82572:
-		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
-		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
-				ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
-		break;
-	case e1000_82573:
-		swsm = E1000_READ_REG(&adapter->hw, SWSM);
-		E1000_WRITE_REG(&adapter->hw, SWSM,
-				swsm & ~E1000_SWSM_DRV_LOAD);
-		break;
-	default:
-		break;
-	}
+	/* Release control of h/w to f/w.  If f/w is AMT enabled, this
+	 * would have already happened in close and is redundant. */
+	e1000_release_hw_control(adapter);
 
 	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
+
+	retval = pci_set_power_state(pdev, pci_choose_state(pdev, state));
+	if (retval)
+		DPRINTK(PROBE, ERR, "Error in setting power state\n");
 
 	return 0;
 }
@@ -4269,16 +4523,21 @@ e1000_resume(struct pci_dev *pdev)
 {
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct e1000_adapter *adapter = netdev_priv(netdev);
-	uint32_t manc, ret_val, swsm;
-	uint32_t ctrl_ext;
+	int retval;
+	uint32_t manc, ret_val;
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
+	retval = pci_set_power_state(pdev, PCI_D0);
+	if (retval)
+		DPRINTK(PROBE, ERR, "Error in setting power state\n");
 	ret_val = pci_enable_device(pdev);
 	pci_set_master(pdev);
 
-	pci_enable_wake(pdev, PCI_D3hot, 0);
-	pci_enable_wake(pdev, PCI_D3cold, 0);
+	retval = pci_enable_wake(pdev, PCI_D3hot, 0);
+	if (retval)
+		DPRINTK(PROBE, ERR, "Error enabling D3 wake\n");
+	retval = pci_enable_wake(pdev, PCI_D3cold, 0);
+	if (retval)
+		DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n");
 
 	e1000_reset(adapter);
 	E1000_WRITE_REG(&adapter->hw, WUS, ~0);
@@ -4295,21 +4554,13 @@ e1000_resume(struct pci_dev *pdev)
 		E1000_WRITE_REG(&adapter->hw, MANC, manc);
 	}
 
-	switch(adapter->hw.mac_type) {
-	case e1000_82571:
-	case e1000_82572:
-		ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT);
-		E1000_WRITE_REG(&adapter->hw, CTRL_EXT,
-				ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
-		break;
-	case e1000_82573:
-		swsm = E1000_READ_REG(&adapter->hw, SWSM);
-		E1000_WRITE_REG(&adapter->hw, SWSM,
-				swsm | E1000_SWSM_DRV_LOAD);
-		break;
-	default:
-		break;
-	}
+	/* If the controller is 82573 and f/w is AMT, do not set
+	 * DRV_LOAD until the interface is up.  For all other cases,
+	 * let the f/w know that the h/w is now under the control
+	 * of the driver. */
+	if (adapter->hw.mac_type != e1000_82573 ||
+	    !e1000_check_mng_mode(&adapter->hw))
+		e1000_get_hw_control(adapter);
 
 	return 0;
 }
@@ -4327,6 +4578,9 @@ e1000_netpoll(struct net_device *netdev)
 	disable_irq(adapter->pdev->irq);
 	e1000_intr(adapter->pdev->irq, netdev, NULL);
 	e1000_clean_tx_irq(adapter, adapter->tx_ring);
+#ifndef CONFIG_E1000_NAPI
+	adapter->clean_rx(adapter, adapter->rx_ring);
+#endif
 	enable_irq(adapter->pdev->irq);
 }
 #endif

drivers/net/e1000/e1000_param.c

@@ -177,7 +177,7 @@ E1000_PARAM(RxAbsIntDelay, "Receive Absolute Interrupt Delay");
  *
  * Valid Range: 100-100000 (0=off, 1=dynamic)
  *
- * Default Value: 1
+ * Default Value: 8000
  */
 
 E1000_PARAM(InterruptThrottleRate, "Interrupt Throttling Rate");
@@ -320,7 +320,7 @@ e1000_check_options(struct e1000_adapter *adapter)
 		} else {
 			tx_ring->count = opt.def;
 		}
-		for (i = 0; i < adapter->num_queues; i++)
+		for (i = 0; i < adapter->num_tx_queues; i++)
 			tx_ring[i].count = tx_ring->count;
 	}
 	{ /* Receive Descriptor Count */
@@ -346,7 +346,7 @@ e1000_check_options(struct e1000_adapter *adapter)
 		} else {
 			rx_ring->count = opt.def;
 		}
-		for (i = 0; i < adapter->num_queues; i++)
+		for (i = 0; i < adapter->num_rx_queues; i++)
 			rx_ring[i].count = rx_ring->count;
 	}
 	{ /* Checksum Offload Enable/Disable */
@@ -388,7 +388,7 @@ e1000_check_options(struct e1000_adapter *adapter)
 			e1000_validate_option(&fc, &opt, adapter);
 			adapter->hw.fc = adapter->hw.original_fc = fc;
 		} else {
-			adapter->hw.fc = opt.def;
+			adapter->hw.fc = adapter->hw.original_fc = opt.def;
 		}
 	}
 	{ /* Transmit Interrupt Delay */
@@ -584,6 +584,12 @@ e1000_check_copper_options(struct e1000_adapter *adapter)
 					 .p = dplx_list }}
 		};
 
+		if (e1000_check_phy_reset_block(&adapter->hw)) {
+			DPRINTK(PROBE, INFO,
+				"Link active due to SoL/IDER Session. "
+			        "Speed/Duplex/AutoNeg parameter ignored.\n");
+			return;
+		}
 		if (num_Duplex > bd) {
 			dplx = Duplex[bd];
 			e1000_validate_option(&dplx, &opt, adapter);

drivers/net/mv643xx_eth.c

@@ -35,6 +35,8 @@
 #include <linux/tcp.h>
 #include <linux/udp.h>
 #include <linux/etherdevice.h>
+#include <linux/in.h>
+#include <linux/ip.h>
 
 #include <linux/bitops.h>
 #include <linux/delay.h>
@@ -55,13 +57,15 @@
 /* Constants */
 #define VLAN_HLEN		4
 #define FCS_LEN			4
-#define WRAP			NET_IP_ALIGN + ETH_HLEN + VLAN_HLEN + FCS_LEN
+#define DMA_ALIGN		8	/* hw requires 8-byte alignment */
+#define HW_IP_ALIGN		2	/* hw aligns IP header */
+#define WRAP			HW_IP_ALIGN + ETH_HLEN + VLAN_HLEN + FCS_LEN
 #define RX_SKB_SIZE		((dev->mtu + WRAP + 7) & ~0x7)
 
-#define INT_CAUSE_UNMASK_ALL		0x0007ffff
-#define INT_CAUSE_UNMASK_ALL_EXT	0x0011ffff
-#define INT_CAUSE_MASK_ALL		0x00000000
-#define INT_CAUSE_MASK_ALL_EXT		0x00000000
+#define INT_UNMASK_ALL			0x0007ffff
+#define INT_UNMASK_ALL_EXT		0x0011ffff
+#define INT_MASK_ALL			0x00000000
+#define INT_MASK_ALL_EXT		0x00000000
 #define INT_CAUSE_CHECK_BITS		INT_CAUSE_UNMASK_ALL
 #define INT_CAUSE_CHECK_BITS_EXT	INT_CAUSE_UNMASK_ALL_EXT
 
@@ -78,8 +82,9 @@
 static int eth_port_link_is_up(unsigned int eth_port_num);
 static void eth_port_uc_addr_get(struct net_device *dev,
 						unsigned char *MacAddr);
-static int mv643xx_eth_real_open(struct net_device *);
-static int mv643xx_eth_real_stop(struct net_device *);
+static void eth_port_set_multicast_list(struct net_device *);
+static int mv643xx_eth_open(struct net_device *);
+static int mv643xx_eth_stop(struct net_device *);
 static int mv643xx_eth_change_mtu(struct net_device *, int);
 static struct net_device_stats *mv643xx_eth_get_stats(struct net_device *);
 static void eth_port_init_mac_tables(unsigned int eth_port_num);
@@ -124,15 +129,8 @@ static inline void mv_write(int offset, u32 data)
  */
 static int mv643xx_eth_change_mtu(struct net_device *dev, int new_mtu)
 {
-	struct mv643xx_private *mp = netdev_priv(dev);
-	unsigned long flags;
-
-	spin_lock_irqsave(&mp->lock, flags);
-
-	if ((new_mtu > 9500) || (new_mtu < 64)) {
-		spin_unlock_irqrestore(&mp->lock, flags);
+	if ((new_mtu > 9500) || (new_mtu < 64))
 		return -EINVAL;
-	}
 
 	dev->mtu = new_mtu;
 	/*
@@ -142,17 +140,13 @@ static int mv643xx_eth_change_mtu(struct net_device *dev, int new_mtu)
 	 * to memory is full, which might fail the open function.
 	 */
 	if (netif_running(dev)) {
-		if (mv643xx_eth_real_stop(dev))
-			printk(KERN_ERR
-				"%s: Fatal error on stopping device\n",
-				dev->name);
-		if (mv643xx_eth_real_open(dev))
+		mv643xx_eth_stop(dev);
+		if (mv643xx_eth_open(dev))
 			printk(KERN_ERR
 				"%s: Fatal error on opening device\n",
 				dev->name);
 	}
 
-	spin_unlock_irqrestore(&mp->lock, flags);
 	return 0;
 }
 
@@ -170,15 +164,19 @@ static void mv643xx_eth_rx_task(void *data)
 	struct mv643xx_private *mp = netdev_priv(dev);
 	struct pkt_info pkt_info;
 	struct sk_buff *skb;
+	int unaligned;
 
 	if (test_and_set_bit(0, &mp->rx_task_busy))
 		panic("%s: Error in test_set_bit / clear_bit", dev->name);
 
 	while (mp->rx_ring_skbs < (mp->rx_ring_size - 5)) {
-		skb = dev_alloc_skb(RX_SKB_SIZE);
+		skb = dev_alloc_skb(RX_SKB_SIZE + DMA_ALIGN);
 		if (!skb)
 			break;
 		mp->rx_ring_skbs++;
+		unaligned = (u32)skb->data & (DMA_ALIGN - 1);
+		if (unaligned)
+			skb_reserve(skb, DMA_ALIGN - unaligned);
 		pkt_info.cmd_sts = ETH_RX_ENABLE_INTERRUPT;
 		pkt_info.byte_cnt = RX_SKB_SIZE;
 		pkt_info.buf_ptr = dma_map_single(NULL, skb->data, RX_SKB_SIZE,
@@ -189,7 +187,7 @@ static void mv643xx_eth_rx_task(void *data)
 				"%s: Error allocating RX Ring\n", dev->name);
 			break;
 		}
-		skb_reserve(skb, 2);
+		skb_reserve(skb, HW_IP_ALIGN);
 	}
 	clear_bit(0, &mp->rx_task_busy);
 	/*
@@ -207,7 +205,7 @@ static void mv643xx_eth_rx_task(void *data)
 	else {
 		/* Return interrupts */
 		mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(mp->port_num),
-							INT_CAUSE_UNMASK_ALL);
+							INT_UNMASK_ALL);
 	}
 #endif
 }
@@ -267,6 +265,8 @@ static void mv643xx_eth_set_rx_mode(struct net_device *dev)
 		mp->port_config &= ~(u32) MV643XX_ETH_UNICAST_PROMISCUOUS_MODE;
 
 	mv_write(MV643XX_ETH_PORT_CONFIG_REG(mp->port_num), mp->port_config);
+
+	eth_port_set_multicast_list(dev);
 }
 
 /*
@@ -342,8 +342,6 @@ static int mv643xx_eth_free_tx_queue(struct net_device *dev,
 	if (!(eth_int_cause_ext & (BIT0 | BIT8)))
 		return released;
 
-	spin_lock(&mp->lock);
-
 	/* Check only queue 0 */
 	while (eth_tx_return_desc(mp, &pkt_info) == ETH_OK) {
 		if (pkt_info.cmd_sts & BIT0) {
@@ -351,31 +349,21 @@ static int mv643xx_eth_free_tx_queue(struct net_device *dev,
 			stats->tx_errors++;
 		}
 
-		/*
-		 * If return_info is different than 0, release the skb.
-		 * The case where return_info is not 0 is only in case
-		 * when transmitted a scatter/gather packet, where only
-		 * last skb releases the whole chain.
-		 */
-		if (pkt_info.return_info) {
-			if (skb_shinfo(pkt_info.return_info)->nr_frags)
-				dma_unmap_page(NULL, pkt_info.buf_ptr,
-						pkt_info.byte_cnt,
-						DMA_TO_DEVICE);
-			else
-				dma_unmap_single(NULL, pkt_info.buf_ptr,
-						pkt_info.byte_cnt,
-						DMA_TO_DEVICE);
+		if (pkt_info.cmd_sts & ETH_TX_FIRST_DESC)
+			dma_unmap_single(NULL, pkt_info.buf_ptr,
+					pkt_info.byte_cnt,
+					DMA_TO_DEVICE);
+		else
+			dma_unmap_page(NULL, pkt_info.buf_ptr,
+					pkt_info.byte_cnt,
+					DMA_TO_DEVICE);
 
+		if (pkt_info.return_info) {
 			dev_kfree_skb_irq(pkt_info.return_info);
 			released = 0;
-		} else
-			dma_unmap_page(NULL, pkt_info.buf_ptr,
-					pkt_info.byte_cnt, DMA_TO_DEVICE);
+		}
 	}
 
-	spin_unlock(&mp->lock);
-
 	return released;
 }
 
@@ -482,12 +470,12 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id,
 
 	/* Read interrupt cause registers */
 	eth_int_cause = mv_read(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num)) &
-						INT_CAUSE_UNMASK_ALL;
+						INT_UNMASK_ALL;
 
 	if (eth_int_cause & BIT1)
 		eth_int_cause_ext = mv_read(
 			MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num)) &
-						INT_CAUSE_UNMASK_ALL_EXT;
+						INT_UNMASK_ALL_EXT;
 
 #ifdef MV643XX_NAPI
 	if (!(eth_int_cause & 0x0007fffd)) {
@@ -512,9 +500,10 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id,
 	} else {
 		if (netif_rx_schedule_prep(dev)) {
 			/* Mask all the interrupts */
-			mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), 0);
-			mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG
-								(port_num), 0);
+			mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
+								INT_MASK_ALL);
+			/* wait for previous write to complete */
+			mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num));
 			__netif_rx_schedule(dev);
 		}
 #else
@@ -527,9 +516,12 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id,
 		 * with skb's.
 		 */
 #ifdef MV643XX_RX_QUEUE_FILL_ON_TASK
-		/* Unmask all interrupts on ethernet port */
+		/* Mask all interrupts on ethernet port */
 		mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
-							INT_CAUSE_MASK_ALL);
+							INT_MASK_ALL);
+		/* wait for previous write to take effect */
+		mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num));
+
 		queue_task(&mp->rx_task, &tq_immediate);
 		mark_bh(IMMEDIATE_BH);
 #else
@@ -635,56 +627,6 @@ static unsigned int eth_port_set_tx_coal(unsigned int eth_port_num,
 	return coal;
 }
 
-/*
- * mv643xx_eth_open
- *
- * This function is called when openning the network device. The function
- * should initialize all the hardware, initialize cyclic Rx/Tx
- * descriptors chain and buffers and allocate an IRQ to the network
- * device.
- *
- * Input :	a pointer to the network device structure
- *
- * Output :	zero of success , nonzero if fails.
- */
-
-static int mv643xx_eth_open(struct net_device *dev)
-{
-	struct mv643xx_private *mp = netdev_priv(dev);
-	unsigned int port_num = mp->port_num;
-	int err;
-
-	spin_lock_irq(&mp->lock);
-
-	err = request_irq(dev->irq, mv643xx_eth_int_handler,
-			SA_SHIRQ | SA_SAMPLE_RANDOM, dev->name, dev);
-
-	if (err) {
-		printk(KERN_ERR "Can not assign IRQ number to MV643XX_eth%d\n",
-								port_num);
-		err = -EAGAIN;
-		goto out;
-	}
-
-	if (mv643xx_eth_real_open(dev)) {
-		printk("%s: Error opening interface\n", dev->name);
-		err = -EBUSY;
-		goto out_free;
-	}
-
-	spin_unlock_irq(&mp->lock);
-
-	return 0;
-
-out_free:
-	free_irq(dev->irq, dev);
-
-out:
-	spin_unlock_irq(&mp->lock);
-
-	return err;
-}
-
 /*
  * ether_init_rx_desc_ring - Curve a Rx chain desc list and buffer in memory.
  *
@@ -777,28 +719,37 @@ static void ether_init_tx_desc_ring(struct mv643xx_private *mp)
 	mp->port_tx_queue_command |= 1;
 }
 
-/* Helper function for mv643xx_eth_open */
-static int mv643xx_eth_real_open(struct net_device *dev)
+/*
+ * mv643xx_eth_open
+ *
+ * This function is called when openning the network device. The function
+ * should initialize all the hardware, initialize cyclic Rx/Tx
+ * descriptors chain and buffers and allocate an IRQ to the network
+ * device.
+ *
+ * Input :	a pointer to the network device structure
+ *
+ * Output :	zero of success , nonzero if fails.
+ */
+
+static int mv643xx_eth_open(struct net_device *dev)
 {
 	struct mv643xx_private *mp = netdev_priv(dev);
 	unsigned int port_num = mp->port_num;
 	unsigned int size;
+	int err;
+
+	err = request_irq(dev->irq, mv643xx_eth_int_handler,
+			SA_SHIRQ | SA_SAMPLE_RANDOM, dev->name, dev);
+	if (err) {
+		printk(KERN_ERR "Can not assign IRQ number to MV643XX_eth%d\n",
+								port_num);
+		return -EAGAIN;
+	}
 
 	/* Stop RX Queues */
 	mv_write(MV643XX_ETH_RECEIVE_QUEUE_COMMAND_REG(port_num), 0x0000ff00);
 
-	/* Clear the ethernet port interrupts */
-	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0);
-	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0);
-
-	/* Unmask RX buffer and TX end interrupt */
-	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
-						INT_CAUSE_UNMASK_ALL);
-
-	/* Unmask phy and link status changes interrupts */
-	mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num),
-						INT_CAUSE_UNMASK_ALL_EXT);
-
 	/* Set the MAC Address */
 	memcpy(mp->port_mac_addr, dev->dev_addr, 6);
 
@@ -818,14 +769,15 @@ static int mv643xx_eth_real_open(struct net_device *dev)
 								GFP_KERNEL);
 	if (!mp->rx_skb) {
 		printk(KERN_ERR "%s: Cannot allocate Rx skb ring\n", dev->name);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out_free_irq;
 	}
 	mp->tx_skb = kmalloc(sizeof(*mp->tx_skb) * mp->tx_ring_size,
 								GFP_KERNEL);
 	if (!mp->tx_skb) {
 		printk(KERN_ERR "%s: Cannot allocate Tx skb ring\n", dev->name);
-		kfree(mp->rx_skb);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out_free_rx_skb;
 	}
 
 	/* Allocate TX ring */
@@ -845,9 +797,8 @@ static int mv643xx_eth_real_open(struct net_device *dev)
 	if (!mp->p_tx_desc_area) {
 		printk(KERN_ERR "%s: Cannot allocate Tx Ring (size %d bytes)\n",
 							dev->name, size);
-		kfree(mp->rx_skb);
-		kfree(mp->tx_skb);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out_free_tx_skb;
 	}
 	BUG_ON((u32) mp->p_tx_desc_area & 0xf);	/* check 16-byte alignment */
 	memset((void *)mp->p_tx_desc_area, 0, mp->tx_desc_area_size);
@@ -874,13 +825,12 @@ static int mv643xx_eth_real_open(struct net_device *dev)
 		printk(KERN_ERR "%s: Freeing previously allocated TX queues...",
 							dev->name);
 		if (mp->rx_sram_size)
-			iounmap(mp->p_rx_desc_area);
+			iounmap(mp->p_tx_desc_area);
 		else
 			dma_free_coherent(NULL, mp->tx_desc_area_size,
 					mp->p_tx_desc_area, mp->tx_desc_dma);
-		kfree(mp->rx_skb);
-		kfree(mp->tx_skb);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out_free_tx_skb;
 	}
 	memset((void *)mp->p_rx_desc_area, 0, size);
 
@@ -900,9 +850,26 @@ static int mv643xx_eth_real_open(struct net_device *dev)
 	mp->tx_int_coal =
 		eth_port_set_tx_coal(port_num, 133000000, MV643XX_TX_COAL);
 
-	netif_start_queue(dev);
+	/* Clear any pending ethernet port interrupts */
+	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0);
+	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0);
+
+	/* Unmask phy and link status changes interrupts */
+	mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num),
+						INT_UNMASK_ALL_EXT);
 
+	/* Unmask RX buffer and TX end interrupt */
+	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_UNMASK_ALL);
 	return 0;
+
+out_free_tx_skb:
+	kfree(mp->tx_skb);
+out_free_rx_skb:
+	kfree(mp->rx_skb);
+out_free_irq:
+	free_irq(dev->irq, dev);
+
+	return err;
 }
 
 static void mv643xx_eth_free_tx_rings(struct net_device *dev)
@@ -910,14 +877,17 @@ static void mv643xx_eth_free_tx_rings(struct net_device *dev)
 	struct mv643xx_private *mp = netdev_priv(dev);
 	unsigned int port_num = mp->port_num;
 	unsigned int curr;
+	struct sk_buff *skb;
 
 	/* Stop Tx Queues */
 	mv_write(MV643XX_ETH_TRANSMIT_QUEUE_COMMAND_REG(port_num), 0x0000ff00);
 
 	/* Free outstanding skb's on TX rings */
 	for (curr = 0; mp->tx_ring_skbs && curr < mp->tx_ring_size; curr++) {
-		if (mp->tx_skb[curr]) {
-			dev_kfree_skb(mp->tx_skb[curr]);
+		skb = mp->tx_skb[curr];
+		if (skb) {
+			mp->tx_ring_skbs -= skb_shinfo(skb)->nr_frags;
+			dev_kfree_skb(skb);
 			mp->tx_ring_skbs--;
 		}
 	}
@@ -973,44 +943,32 @@ static void mv643xx_eth_free_rx_rings(struct net_device *dev)
  * Output :	zero if success , nonzero if fails
  */
 
-/* Helper function for mv643xx_eth_stop */
-
-static int mv643xx_eth_real_stop(struct net_device *dev)
+static int mv643xx_eth_stop(struct net_device *dev)
 {
 	struct mv643xx_private *mp = netdev_priv(dev);
 	unsigned int port_num = mp->port_num;
 
+	/* Mask all interrupts on ethernet port */
+	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_MASK_ALL);
+	/* wait for previous write to complete */
+	mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num));
+
+#ifdef MV643XX_NAPI
+	netif_poll_disable(dev);
+#endif
 	netif_carrier_off(dev);
 	netif_stop_queue(dev);
 
-	mv643xx_eth_free_tx_rings(dev);
-	mv643xx_eth_free_rx_rings(dev);
-
 	eth_port_reset(mp->port_num);
 
-	/* Disable ethernet port interrupts */
-	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0);
-	mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0);
-
-	/* Mask RX buffer and TX end interrupt */
-	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), 0);
-
-	/* Mask phy and link status changes interrupts */
-	mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), 0);
-
-	return 0;
-}
-
-static int mv643xx_eth_stop(struct net_device *dev)
-{
-	struct mv643xx_private *mp = netdev_priv(dev);
-
-	spin_lock_irq(&mp->lock);
+	mv643xx_eth_free_tx_rings(dev);
+	mv643xx_eth_free_rx_rings(dev);
 
-	mv643xx_eth_real_stop(dev);
+#ifdef MV643XX_NAPI
+	netif_poll_enable(dev);
+#endif
 
 	free_irq(dev->irq, dev);
-	spin_unlock_irq(&mp->lock);
 
 	return 0;
 }
@@ -1022,20 +980,17 @@ static void mv643xx_tx(struct net_device *dev)
 	struct pkt_info pkt_info;
 
 	while (eth_tx_return_desc(mp, &pkt_info) == ETH_OK) {
-		if (pkt_info.return_info) {
-			if (skb_shinfo(pkt_info.return_info)->nr_frags)
-				dma_unmap_page(NULL, pkt_info.buf_ptr,
-						pkt_info.byte_cnt,
-						DMA_TO_DEVICE);
-			else
-				dma_unmap_single(NULL, pkt_info.buf_ptr,
-						pkt_info.byte_cnt,
-						DMA_TO_DEVICE);
+		if (pkt_info.cmd_sts & ETH_TX_FIRST_DESC)
+			dma_unmap_single(NULL, pkt_info.buf_ptr,
+					pkt_info.byte_cnt,
+					DMA_TO_DEVICE);
+		else
+			dma_unmap_page(NULL, pkt_info.buf_ptr,
+					pkt_info.byte_cnt,
+					DMA_TO_DEVICE);
 
+		if (pkt_info.return_info)
 			dev_kfree_skb_irq(pkt_info.return_info);
-		} else
-			dma_unmap_page(NULL, pkt_info.buf_ptr,
-					pkt_info.byte_cnt, DMA_TO_DEVICE);
 	}
 
 	if (netif_queue_stopped(dev) &&
@@ -1053,14 +1008,11 @@ static int mv643xx_poll(struct net_device *dev, int *budget)
 	struct mv643xx_private *mp = netdev_priv(dev);
 	int done = 1, orig_budget, work_done;
 	unsigned int port_num = mp->port_num;
-	unsigned long flags;
 
 #ifdef MV643XX_TX_FAST_REFILL
 	if (++mp->tx_clean_threshold > 5) {
-		spin_lock_irqsave(&mp->lock, flags);
 		mv643xx_tx(dev);
 		mp->tx_clean_threshold = 0;
-		spin_unlock_irqrestore(&mp->lock, flags);
 	}
 #endif
 
@@ -1078,21 +1030,36 @@ static int mv643xx_poll(struct net_device *dev, int *budget)
 	}
 
 	if (done) {
-		spin_lock_irqsave(&mp->lock, flags);
-		__netif_rx_complete(dev);
+		netif_rx_complete(dev);
 		mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0);
 		mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0);
 		mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
-						INT_CAUSE_UNMASK_ALL);
-		mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num),
-						INT_CAUSE_UNMASK_ALL_EXT);
-		spin_unlock_irqrestore(&mp->lock, flags);
+						INT_UNMASK_ALL);
 	}
 
 	return done ? 0 : 1;
 }
 #endif
 
+/* Hardware can't handle unaligned fragments smaller than 9 bytes.
+ * This helper function detects that case.
+ */
+
+static inline unsigned int has_tiny_unaligned_frags(struct sk_buff *skb)
+{
+        unsigned int frag;
+        skb_frag_t *fragp;
+
+        for (frag = 0; frag < skb_shinfo(skb)->nr_frags; frag++) {
+                fragp = &skb_shinfo(skb)->frags[frag];
+                if (fragp->size <= 8 && fragp->page_offset & 0x7)
+                        return 1;
+
+        }
+        return 0;
+}
+
+
 /*
  * mv643xx_eth_start_xmit
  *
@@ -1136,12 +1103,19 @@ static int mv643xx_eth_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		return 1;
 	}
 
+#ifdef MV643XX_CHECKSUM_OFFLOAD_TX
+	if (has_tiny_unaligned_frags(skb)) {
+		if ((skb_linearize(skb, GFP_ATOMIC) != 0)) {
+			stats->tx_dropped++;
+			printk(KERN_DEBUG "%s: failed to linearize tiny "
+					"unaligned fragment\n", dev->name);
+			return 1;
+		}
+	}
+
 	spin_lock_irqsave(&mp->lock, flags);
 
-	/* Update packet info data structure -- DMA owned, first last */
-#ifdef MV643XX_CHECKSUM_OFFLOAD_TX
 	if (!skb_shinfo(skb)->nr_frags) {
-linear:
 		if (skb->ip_summed != CHECKSUM_HW) {
 			/* Errata BTS #50, IHL must be 5 if no HW checksum */
 			pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT |
@@ -1150,7 +1124,6 @@ linear:
 					   5 << ETH_TX_IHL_SHIFT;
 			pkt_info.l4i_chk = 0;
 		} else {
-
 			pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT |
 					   ETH_TX_FIRST_DESC |
 					   ETH_TX_LAST_DESC |
@@ -1158,14 +1131,16 @@ linear:
 					   ETH_GEN_IP_V_4_CHECKSUM |
 					   skb->nh.iph->ihl << ETH_TX_IHL_SHIFT;
 			/* CPU already calculated pseudo header checksum. */
-			if (skb->nh.iph->protocol == IPPROTO_UDP) {
+			if ((skb->protocol == ETH_P_IP) &&
+			    (skb->nh.iph->protocol == IPPROTO_UDP) ) {
 				pkt_info.cmd_sts |= ETH_UDP_FRAME;
 				pkt_info.l4i_chk = skb->h.uh->check;
-			} else if (skb->nh.iph->protocol == IPPROTO_TCP)
+			} else if ((skb->protocol == ETH_P_IP) &&
+				   (skb->nh.iph->protocol == IPPROTO_TCP))
 				pkt_info.l4i_chk = skb->h.th->check;
 			else {
 				printk(KERN_ERR
-					"%s: chksum proto != TCP or UDP\n",
+					"%s: chksum proto != IPv4 TCP or UDP\n",
 					dev->name);
 				spin_unlock_irqrestore(&mp->lock, flags);
 				return 1;
@@ -1183,26 +1158,6 @@ linear:
 	} else {
 		unsigned int frag;
 
-		/* Since hardware can't handle unaligned fragments smaller
-		 * than 9 bytes, if we find any, we linearize the skb
-		 * and start again.  When I've seen it, it's always been
-		 * the first frag (probably near the end of the page),
-		 * but we check all frags to be safe.
-		 */
-		for (frag = 0; frag < skb_shinfo(skb)->nr_frags; frag++) {
-			skb_frag_t *fragp;
-
-			fragp = &skb_shinfo(skb)->frags[frag];
-			if (fragp->size <= 8 && fragp->page_offset & 0x7) {
-				skb_linearize(skb, GFP_ATOMIC);
-				printk(KERN_DEBUG "%s: unaligned tiny fragment"
-						"%d of %d, fixed\n",
-						dev->name, frag,
-						skb_shinfo(skb)->nr_frags);
-				goto linear;
-			}
-		}
-
 		/* first frag which is skb header */
 		pkt_info.byte_cnt = skb_headlen(skb);
 		pkt_info.buf_ptr = dma_map_single(NULL, skb->data,
@@ -1221,14 +1176,16 @@ linear:
 					   ETH_GEN_IP_V_4_CHECKSUM |
 					   skb->nh.iph->ihl << ETH_TX_IHL_SHIFT;
 			/* CPU already calculated pseudo header checksum. */
-			if (skb->nh.iph->protocol == IPPROTO_UDP) {
+			if ((skb->protocol == ETH_P_IP) &&
+			    (skb->nh.iph->protocol == IPPROTO_UDP)) {
 				pkt_info.cmd_sts |= ETH_UDP_FRAME;
 				pkt_info.l4i_chk = skb->h.uh->check;
-			} else if (skb->nh.iph->protocol == IPPROTO_TCP)
+			} else if ((skb->protocol == ETH_P_IP) &&
+				   (skb->nh.iph->protocol == IPPROTO_TCP))
 				pkt_info.l4i_chk = skb->h.th->check;
 			else {
 				printk(KERN_ERR
-					"%s: chksum proto != TCP or UDP\n",
+					"%s: chksum proto != IPv4 TCP or UDP\n",
 					dev->name);
 				spin_unlock_irqrestore(&mp->lock, flags);
 				return 1;
@@ -1288,6 +1245,8 @@ linear:
 		}
 	}
 #else
+	spin_lock_irqsave(&mp->lock, flags);
+
 	pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT | ETH_TX_FIRST_DESC |
 							ETH_TX_LAST_DESC;
 	pkt_info.l4i_chk = 0;
@@ -1340,39 +1299,18 @@ static struct net_device_stats *mv643xx_eth_get_stats(struct net_device *dev)
 }
 
 #ifdef CONFIG_NET_POLL_CONTROLLER
-static inline void mv643xx_enable_irq(struct mv643xx_private *mp)
-{
-	int port_num = mp->port_num;
-	unsigned long flags;
-
-	spin_lock_irqsave(&mp->lock, flags);
-	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
-					INT_CAUSE_UNMASK_ALL);
-	mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num),
-					INT_CAUSE_UNMASK_ALL_EXT);
-	spin_unlock_irqrestore(&mp->lock, flags);
-}
-
-static inline void mv643xx_disable_irq(struct mv643xx_private *mp)
-{
-	int port_num = mp->port_num;
-	unsigned long flags;
-
-	spin_lock_irqsave(&mp->lock, flags);
-	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num),
-					INT_CAUSE_MASK_ALL);
-	mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num),
-					INT_CAUSE_MASK_ALL_EXT);
-	spin_unlock_irqrestore(&mp->lock, flags);
-}
-
 static void mv643xx_netpoll(struct net_device *netdev)
 {
 	struct mv643xx_private *mp = netdev_priv(netdev);
+	int port_num = mp->port_num;
+
+	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_MASK_ALL);
+	/* wait for previous write to complete */
+	mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num));
 
-	mv643xx_disable_irq(mp);
 	mv643xx_eth_int_handler(netdev->irq, netdev, NULL);
-	mv643xx_enable_irq(mp);
+
+	mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_UNMASK_ALL);
 }
 #endif
 
@@ -1441,7 +1379,7 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
 	 * Zero copy can only work if we use Discovery II memory. Else, we will
 	 * have to map the buffers to ISA memory which is only 16 MB
 	 */
-	dev->features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_HW_CSUM;
+	dev->features = NETIF_F_SG | NETIF_F_IP_CSUM;
 #endif
 #endif
 
@@ -2053,6 +1991,196 @@ static int eth_port_uc_addr(unsigned int eth_port_num, unsigned char uc_nibble,
 	return 1;
 }
 
+/*
+ * The entries in each table are indexed by a hash of a packet's MAC
+ * address.  One bit in each entry determines whether the packet is
+ * accepted.  There are 4 entries (each 8 bits wide) in each register
+ * of the table.  The bits in each entry are defined as follows:
+ *	0	Accept=1, Drop=0
+ *	3-1	Queue			(ETH_Q0=0)
+ *	7-4	Reserved = 0;
+ */
+static void eth_port_set_filter_table_entry(int table, unsigned char entry)
+{
+	unsigned int table_reg;
+	unsigned int tbl_offset;
+	unsigned int reg_offset;
+
+	tbl_offset = (entry / 4) * 4;	/* Register offset of DA table entry */
+	reg_offset = entry % 4;		/* Entry offset within the register */
+
+	/* Set "accepts frame bit" at specified table entry */
+	table_reg = mv_read(table + tbl_offset);
+	table_reg |= 0x01 << (8 * reg_offset);
+	mv_write(table + tbl_offset, table_reg);
+}
+
+/*
+ * eth_port_mc_addr - Multicast address settings.
+ *
+ * The MV device supports multicast using two tables:
+ * 1) Special Multicast Table for MAC addresses of the form
+ *    0x01-00-5E-00-00-XX (where XX is between 0x00 and 0x_FF).
+ *    The MAC DA[7:0] bits are used as a pointer to the Special Multicast
+ *    Table entries in the DA-Filter table.
+ * 2) Other Multicast Table for multicast of another type. A CRC-8bit
+ *    is used as an index to the Other Multicast Table entries in the
+ *    DA-Filter table.  This function calculates the CRC-8bit value.
+ * In either case, eth_port_set_filter_table_entry() is then called
+ * to set to set the actual table entry.
+ */
+static void eth_port_mc_addr(unsigned int eth_port_num, unsigned char *p_addr)
+{
+	unsigned int mac_h;
+	unsigned int mac_l;
+	unsigned char crc_result = 0;
+	int table;
+	int mac_array[48];
+	int crc[8];
+	int i;
+
+	if ((p_addr[0] == 0x01) && (p_addr[1] == 0x00) &&
+	    (p_addr[2] == 0x5E) && (p_addr[3] == 0x00) && (p_addr[4] == 0x00)) {
+		table = MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE
+					(eth_port_num);
+		eth_port_set_filter_table_entry(table, p_addr[5]);
+		return;
+	}
+
+	/* Calculate CRC-8 out of the given address */
+	mac_h = (p_addr[0] << 8) | (p_addr[1]);
+	mac_l = (p_addr[2] << 24) | (p_addr[3] << 16) |
+			(p_addr[4] << 8) | (p_addr[5] << 0);
+
+	for (i = 0; i < 32; i++)
+		mac_array[i] = (mac_l >> i) & 0x1;
+	for (i = 32; i < 48; i++)
+		mac_array[i] = (mac_h >> (i - 32)) & 0x1;
+
+	crc[0] = mac_array[45] ^ mac_array[43] ^ mac_array[40] ^ mac_array[39] ^
+		 mac_array[35] ^ mac_array[34] ^ mac_array[31] ^ mac_array[30] ^
+		 mac_array[28] ^ mac_array[23] ^ mac_array[21] ^ mac_array[19] ^
+		 mac_array[18] ^ mac_array[16] ^ mac_array[14] ^ mac_array[12] ^
+		 mac_array[8]  ^ mac_array[7]  ^ mac_array[6]  ^ mac_array[0];
+
+	crc[1] = mac_array[46] ^ mac_array[45] ^ mac_array[44] ^ mac_array[43] ^
+		 mac_array[41] ^ mac_array[39] ^ mac_array[36] ^ mac_array[34] ^
+		 mac_array[32] ^ mac_array[30] ^ mac_array[29] ^ mac_array[28] ^
+		 mac_array[24] ^ mac_array[23] ^ mac_array[22] ^ mac_array[21] ^
+		 mac_array[20] ^ mac_array[18] ^ mac_array[17] ^ mac_array[16] ^
+		 mac_array[15] ^ mac_array[14] ^ mac_array[13] ^ mac_array[12] ^
+		 mac_array[9]  ^ mac_array[6]  ^ mac_array[1]  ^ mac_array[0];
+
+	crc[2] = mac_array[47] ^ mac_array[46] ^ mac_array[44] ^ mac_array[43] ^
+		 mac_array[42] ^ mac_array[39] ^ mac_array[37] ^ mac_array[34] ^
+		 mac_array[33] ^ mac_array[29] ^ mac_array[28] ^ mac_array[25] ^
+		 mac_array[24] ^ mac_array[22] ^ mac_array[17] ^ mac_array[15] ^
+		 mac_array[13] ^ mac_array[12] ^ mac_array[10] ^ mac_array[8]  ^
+		 mac_array[6]  ^ mac_array[2]  ^ mac_array[1]  ^ mac_array[0];
+
+	crc[3] = mac_array[47] ^ mac_array[45] ^ mac_array[44] ^ mac_array[43] ^
+		 mac_array[40] ^ mac_array[38] ^ mac_array[35] ^ mac_array[34] ^
+		 mac_array[30] ^ mac_array[29] ^ mac_array[26] ^ mac_array[25] ^
+		 mac_array[23] ^ mac_array[18] ^ mac_array[16] ^ mac_array[14] ^
+		 mac_array[13] ^ mac_array[11] ^ mac_array[9]  ^ mac_array[7]  ^
+		 mac_array[3]  ^ mac_array[2]  ^ mac_array[1];
+
+	crc[4] = mac_array[46] ^ mac_array[45] ^ mac_array[44] ^ mac_array[41] ^
+		 mac_array[39] ^ mac_array[36] ^ mac_array[35] ^ mac_array[31] ^
+		 mac_array[30] ^ mac_array[27] ^ mac_array[26] ^ mac_array[24] ^
+		 mac_array[19] ^ mac_array[17] ^ mac_array[15] ^ mac_array[14] ^
+		 mac_array[12] ^ mac_array[10] ^ mac_array[8]  ^ mac_array[4]  ^
+		 mac_array[3]  ^ mac_array[2];
+
+	crc[5] = mac_array[47] ^ mac_array[46] ^ mac_array[45] ^ mac_array[42] ^
+		 mac_array[40] ^ mac_array[37] ^ mac_array[36] ^ mac_array[32] ^
+		 mac_array[31] ^ mac_array[28] ^ mac_array[27] ^ mac_array[25] ^
+		 mac_array[20] ^ mac_array[18] ^ mac_array[16] ^ mac_array[15] ^
+		 mac_array[13] ^ mac_array[11] ^ mac_array[9]  ^ mac_array[5]  ^
+		 mac_array[4]  ^ mac_array[3];
+
+	crc[6] = mac_array[47] ^ mac_array[46] ^ mac_array[43] ^ mac_array[41] ^
+		 mac_array[38] ^ mac_array[37] ^ mac_array[33] ^ mac_array[32] ^
+		 mac_array[29] ^ mac_array[28] ^ mac_array[26] ^ mac_array[21] ^
+		 mac_array[19] ^ mac_array[17] ^ mac_array[16] ^ mac_array[14] ^
+		 mac_array[12] ^ mac_array[10] ^ mac_array[6]  ^ mac_array[5]  ^
+		 mac_array[4];
+
+	crc[7] = mac_array[47] ^ mac_array[44] ^ mac_array[42] ^ mac_array[39] ^
+		 mac_array[38] ^ mac_array[34] ^ mac_array[33] ^ mac_array[30] ^
+		 mac_array[29] ^ mac_array[27] ^ mac_array[22] ^ mac_array[20] ^
+		 mac_array[18] ^ mac_array[17] ^ mac_array[15] ^ mac_array[13] ^
+		 mac_array[11] ^ mac_array[7]  ^ mac_array[6]  ^ mac_array[5];
+
+	for (i = 0; i < 8; i++)
+		crc_result = crc_result | (crc[i] << i);
+
+	table = MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE(eth_port_num);
+	eth_port_set_filter_table_entry(table, crc_result);
+}
+
+/*
+ * Set the entire multicast list based on dev->mc_list.
+ */
+static void eth_port_set_multicast_list(struct net_device *dev)
+{
+
+	struct dev_mc_list	*mc_list;
+	int			i;
+	int			table_index;
+	struct mv643xx_private	*mp = netdev_priv(dev);
+	unsigned int		eth_port_num = mp->port_num;
+
+	/* If the device is in promiscuous mode or in all multicast mode,
+	 * we will fully populate both multicast tables with accept.
+	 * This is guaranteed to yield a match on all multicast addresses...
+	 */
+	if ((dev->flags & IFF_PROMISC) || (dev->flags & IFF_ALLMULTI)) {
+		for (table_index = 0; table_index <= 0xFC; table_index += 4) {
+			 /* Set all entries in DA filter special multicast
+			  * table (Ex_dFSMT)
+			  * Set for ETH_Q0 for now
+			  * Bits
+			  * 0	  Accept=1, Drop=0
+			  * 3-1  Queue	 ETH_Q0=0
+			  * 7-4  Reserved = 0;
+			  */
+			 mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE(eth_port_num) + table_index, 0x01010101);
+
+			 /* Set all entries in DA filter other multicast
+			  * table (Ex_dFOMT)
+			  * Set for ETH_Q0 for now
+			  * Bits
+			  * 0	  Accept=1, Drop=0
+			  * 3-1  Queue	 ETH_Q0=0
+			  * 7-4  Reserved = 0;
+			  */
+			 mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE(eth_port_num) + table_index, 0x01010101);
+       	}
+		return;
+	}
+
+	/* We will clear out multicast tables every time we get the list.
+	 * Then add the entire new list...
+	 */
+	for (table_index = 0; table_index <= 0xFC; table_index += 4) {
+		/* Clear DA filter special multicast table (Ex_dFSMT) */
+		mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE
+				(eth_port_num) + table_index, 0);
+
+		/* Clear DA filter other multicast table (Ex_dFOMT) */
+		mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE
+				(eth_port_num) + table_index, 0);
+	}
+
+	/* Get pointer to net_device multicast list and add each one... */
+	for (i = 0, mc_list = dev->mc_list;
+			(i < 256) && (mc_list != NULL) && (i < dev->mc_count);
+			i++, mc_list = mc_list->next)
+		if (mc_list->dmi_addrlen == 6)
+			eth_port_mc_addr(eth_port_num, mc_list->dmi_addr);
+}
+
 /*
  * eth_port_init_mac_tables - Clear all entrance in the UC, SMC and OMC tables
  *
@@ -2080,11 +2208,11 @@ static void eth_port_init_mac_tables(unsigned int eth_port_num)
 
 	for (table_index = 0; table_index <= 0xFC; table_index += 4) {
 		/* Clear DA filter special multicast table (Ex_dFSMT) */
-		mv_write((MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE
-					(eth_port_num) + table_index), 0);
+		mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE
+					(eth_port_num) + table_index, 0);
 		/* Clear DA filter other multicast table (Ex_dFOMT) */
-		mv_write((MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE
-					(eth_port_num) + table_index), 0);
+		mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE
+					(eth_port_num) + table_index, 0);
 	}
 }
 
@@ -2489,6 +2617,7 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
 	struct eth_tx_desc *current_descriptor;
 	struct eth_tx_desc *first_descriptor;
 	u32 command;
+	unsigned long flags;
 
 	/* Do not process Tx ring in case of Tx ring resource error */
 	if (mp->tx_resource_err)
@@ -2505,6 +2634,8 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
 		return ETH_ERROR;
 	}
 
+	spin_lock_irqsave(&mp->lock, flags);
+
 	mp->tx_ring_skbs++;
 	BUG_ON(mp->tx_ring_skbs > mp->tx_ring_size);
 
@@ -2554,11 +2685,15 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
 		mp->tx_resource_err = 1;
 		mp->tx_curr_desc_q = tx_first_desc;
 
+		spin_unlock_irqrestore(&mp->lock, flags);
+
 		return ETH_QUEUE_LAST_RESOURCE;
 	}
 
 	mp->tx_curr_desc_q = tx_next_desc;
 
+	spin_unlock_irqrestore(&mp->lock, flags);
+
 	return ETH_OK;
 }
 #else
@@ -2569,11 +2704,14 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
 	int tx_desc_used;
 	struct eth_tx_desc *current_descriptor;
 	unsigned int command_status;
+	unsigned long flags;
 
 	/* Do not process Tx ring in case of Tx ring resource error */
 	if (mp->tx_resource_err)
 		return ETH_QUEUE_FULL;
 
+	spin_lock_irqsave(&mp->lock, flags);
+
 	mp->tx_ring_skbs++;
 	BUG_ON(mp->tx_ring_skbs > mp->tx_ring_size);
 
@@ -2604,9 +2742,12 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
 	/* Check for ring index overlap in the Tx desc ring */
 	if (tx_desc_curr == tx_desc_used) {
 		mp->tx_resource_err = 1;
+
+		spin_unlock_irqrestore(&mp->lock, flags);
 		return ETH_QUEUE_LAST_RESOURCE;
 	}
 
+	spin_unlock_irqrestore(&mp->lock, flags);
 	return ETH_OK;
 }
 #endif
@@ -2629,23 +2770,27 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp,
  *	Tx ring 'first' and 'used' indexes are updated.
  *
  * RETURN:
- *	ETH_ERROR in case the routine can not access Tx desc ring.
- *	ETH_RETRY in case there is transmission in process.
- *	ETH_END_OF_JOB if the routine has nothing to release.
- *	ETH_OK otherwise.
+ *	ETH_OK on success
+ *	ETH_ERROR otherwise.
  *
  */
 static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp,
 						struct pkt_info *p_pkt_info)
 {
 	int tx_desc_used;
+	int tx_busy_desc;
+	struct eth_tx_desc *p_tx_desc_used;
+	unsigned int command_status;
+	unsigned long flags;
+	int err = ETH_OK;
+
+	spin_lock_irqsave(&mp->lock, flags);
+
 #ifdef MV643XX_CHECKSUM_OFFLOAD_TX
-	int tx_busy_desc = mp->tx_first_desc_q;
+	tx_busy_desc = mp->tx_first_desc_q;
 #else
-	int tx_busy_desc = mp->tx_curr_desc_q;
+	tx_busy_desc = mp->tx_curr_desc_q;
 #endif
-	struct eth_tx_desc *p_tx_desc_used;
-	unsigned int command_status;
 
 	/* Get the Tx Desc ring indexes */
 	tx_desc_used = mp->tx_used_desc_q;
@@ -2653,22 +2798,30 @@ static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp,
 	p_tx_desc_used = &mp->p_tx_desc_area[tx_desc_used];
 
 	/* Sanity check */
-	if (p_tx_desc_used == NULL)
-		return ETH_ERROR;
+	if (p_tx_desc_used == NULL) {
+		err = ETH_ERROR;
+		goto out;
+	}
 
 	/* Stop release. About to overlap the current available Tx descriptor */
-	if (tx_desc_used == tx_busy_desc && !mp->tx_resource_err)
-		return ETH_END_OF_JOB;
+	if (tx_desc_used == tx_busy_desc && !mp->tx_resource_err) {
+		err = ETH_ERROR;
+		goto out;
+	}
 
 	command_status = p_tx_desc_used->cmd_sts;
 
 	/* Still transmitting... */
-	if (command_status & (ETH_BUFFER_OWNED_BY_DMA))
-		return ETH_RETRY;
+	if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) {
+		err = ETH_ERROR;
+		goto out;
+	}
 
 	/* Pass the packet information to the caller */
 	p_pkt_info->cmd_sts = command_status;
 	p_pkt_info->return_info = mp->tx_skb[tx_desc_used];
+	p_pkt_info->buf_ptr = p_tx_desc_used->buf_ptr;
+	p_pkt_info->byte_cnt = p_tx_desc_used->byte_cnt;
 	mp->tx_skb[tx_desc_used] = NULL;
 
 	/* Update the next descriptor to release. */
@@ -2680,7 +2833,10 @@ static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp,
 	BUG_ON(mp->tx_ring_skbs == 0);
 	mp->tx_ring_skbs--;
 
-	return ETH_OK;
+out:
+	spin_unlock_irqrestore(&mp->lock, flags);
+
+	return err;
 }
 
 /*
@@ -2712,11 +2868,14 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp,
 	int rx_next_curr_desc, rx_curr_desc, rx_used_desc;
 	volatile struct eth_rx_desc *p_rx_desc;
 	unsigned int command_status;
+	unsigned long flags;
 
 	/* Do not process Rx ring in case of Rx ring resource error */
 	if (mp->rx_resource_err)
 		return ETH_QUEUE_FULL;
 
+	spin_lock_irqsave(&mp->lock, flags);
+
 	/* Get the Rx Desc ring 'curr and 'used' indexes */
 	rx_curr_desc = mp->rx_curr_desc_q;
 	rx_used_desc = mp->rx_used_desc_q;
@@ -2728,8 +2887,10 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp,
 	rmb();
 
 	/* Nothing to receive... */
-	if (command_status & (ETH_BUFFER_OWNED_BY_DMA))
+	if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) {
+		spin_unlock_irqrestore(&mp->lock, flags);
 		return ETH_END_OF_JOB;
+	}
 
 	p_pkt_info->byte_cnt = (p_rx_desc->byte_cnt) - RX_BUF_OFFSET;
 	p_pkt_info->cmd_sts = command_status;
@@ -2749,6 +2910,8 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp,
 	if (rx_next_curr_desc == rx_used_desc)
 		mp->rx_resource_err = 1;
 
+	spin_unlock_irqrestore(&mp->lock, flags);
+
 	return ETH_OK;
 }
 
@@ -2777,6 +2940,9 @@ static ETH_FUNC_RET_STATUS eth_rx_return_buff(struct mv643xx_private *mp,
 {
 	int used_rx_desc;	/* Where to return Rx resource */
 	volatile struct eth_rx_desc *p_used_rx_desc;
+	unsigned long flags;
+
+	spin_lock_irqsave(&mp->lock, flags);
 
 	/* Get 'used' Rx descriptor */
 	used_rx_desc = mp->rx_used_desc_q;
@@ -2800,6 +2966,8 @@ static ETH_FUNC_RET_STATUS eth_rx_return_buff(struct mv643xx_private *mp,
 	/* Any Rx return cancels the Rx resource error status */
 	mp->rx_resource_err = 0;
 
+	spin_unlock_irqrestore(&mp->lock, flags);
+
 	return ETH_OK;
 }
 

drivers/net/skge.c

@@ -3243,12 +3243,22 @@ static int __devinit skge_probe(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	if (!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK)))
+	if (sizeof(dma_addr_t) > sizeof(u32) &&
+	    !(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) {
 		using_dac = 1;
-	else if (!(err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) {
-		printk(KERN_ERR PFX "%s no usable DMA configuration\n",
-		       pci_name(pdev));
-		goto err_out_free_regions;
+		err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK);
+		if (err < 0) {
+			printk(KERN_ERR PFX "%s unable to obtain 64 bit DMA "
+			       "for consistent allocations\n", pci_name(pdev));
+			goto err_out_free_regions;
+		}
+	} else {
+		err = pci_set_dma_mask(pdev, DMA_32BIT_MASK);
+		if (err) {
+			printk(KERN_ERR PFX "%s no usable DMA configuration\n",
+			       pci_name(pdev));
+			goto err_out_free_regions;
+		}
 	}
 
 #ifdef __BIG_ENDIAN

drivers/net/sky2.c

@@ -57,7 +57,7 @@
 #include "sky2.h"
 
 #define DRV_NAME		"sky2"
-#define DRV_VERSION		"0.11"
+#define DRV_VERSION		"0.13"
 #define PFX			DRV_NAME " "
 
 /*
@@ -75,6 +75,7 @@
 #define RX_LE_BYTES		(RX_LE_SIZE*sizeof(struct sky2_rx_le))
 #define RX_MAX_PENDING		(RX_LE_SIZE/2 - 2)
 #define RX_DEF_PENDING		RX_MAX_PENDING
+#define RX_SKB_ALIGN		8
 
 #define TX_RING_SIZE		512
 #define TX_DEF_PENDING		(TX_RING_SIZE - 1)
@@ -91,7 +92,7 @@
 static const u32 default_msg =
     NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK
     | NETIF_MSG_TIMER | NETIF_MSG_TX_ERR | NETIF_MSG_RX_ERR
-    | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN | NETIF_MSG_INTR;
+    | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN;
 
 static int debug = -1;		/* defaults above */
 module_param(debug, int, 0);
@@ -624,13 +625,16 @@ static void sky2_mac_init(struct sky2_hw *hw, unsigned port)
 
 }
 
-static void sky2_ramset(struct sky2_hw *hw, u16 q, u32 start, size_t len)
+/* Assign Ram Buffer allocation.
+ * start and end are in units of 4k bytes
+ * ram registers are in units of 64bit words
+ */
+static void sky2_ramset(struct sky2_hw *hw, u16 q, u8 startk, u8 endk)
 {
-	u32 end;
+	u32 start, end;
 
-	start /= 8;
-	len /= 8;
-	end = start + len - 1;
+	start = startk * 4096/8;
+	end = (endk * 4096/8) - 1;
 
 	sky2_write8(hw, RB_ADDR(q, RB_CTRL), RB_RST_CLR);
 	sky2_write32(hw, RB_ADDR(q, RB_START), start);
@@ -639,14 +643,19 @@ static void sky2_ramset(struct sky2_hw *hw, u16 q, u32 start, size_t len)
 	sky2_write32(hw, RB_ADDR(q, RB_RP), start);
 
 	if (q == Q_R1 || q == Q_R2) {
-		u32 rxup, rxlo;
+		u32 space = (endk - startk) * 4096/8;
+		u32 tp = space - space/4;
 
-		rxlo = len/2;
-		rxup = rxlo + len/4;
+		/* On receive queue's set the thresholds
+		 * give receiver priority when > 3/4 full
+		 * send pause when down to 2K
+		 */
+		sky2_write32(hw, RB_ADDR(q, RB_RX_UTHP), tp);
+		sky2_write32(hw, RB_ADDR(q, RB_RX_LTHP), space/2);
 
-		/* Set thresholds on receive queue's */
-		sky2_write32(hw, RB_ADDR(q, RB_RX_UTPP), rxup);
-		sky2_write32(hw, RB_ADDR(q, RB_RX_LTPP), rxlo);
+		tp = space - 2048/8;
+		sky2_write32(hw, RB_ADDR(q, RB_RX_UTPP), tp);
+		sky2_write32(hw, RB_ADDR(q, RB_RX_LTPP), space/4);
 	} else {
 		/* Enable store & forward on Tx queue's because
 		 * Tx FIFO is only 1K on Yukon
@@ -695,9 +704,10 @@ static inline struct sky2_tx_le *get_tx_le(struct sky2_port *sky2)
  * This is a workaround code taken from SysKonnect sk98lin driver
  * to deal with chip bug on Yukon EC rev 0 in the wraparound case.
  */
-static inline void sky2_put_idx(struct sky2_hw *hw, unsigned q,
+static void sky2_put_idx(struct sky2_hw *hw, unsigned q,
 				u16 idx, u16 *last, u16 size)
 {
+	wmb();
 	if (is_ec_a1(hw) && idx < *last) {
 		u16 hwget = sky2_read16(hw, Y2_QADDR(q, PREF_UNIT_GET_IDX));
 
@@ -721,6 +731,7 @@ setnew:
 		sky2_write16(hw, Y2_QADDR(q, PREF_UNIT_PUT_IDX), idx);
 	}
 	*last = idx;
+	mmiowb();
 }
 
 
@@ -734,11 +745,11 @@ static inline struct sky2_rx_le *sky2_next_rx(struct sky2_port *sky2)
 /* Return high part of DMA address (could be 32 or 64 bit) */
 static inline u32 high32(dma_addr_t a)
 {
-	return (a >> 16) >> 16;
+	return sizeof(a) > sizeof(u32) ? (a >> 16) >> 16 : 0;
 }
 
 /* Build description to hardware about buffer */
-static inline void sky2_rx_add(struct sky2_port *sky2, dma_addr_t map)
+static void sky2_rx_add(struct sky2_port *sky2, dma_addr_t map)
 {
 	struct sky2_rx_le *le;
 	u32 hi = high32(map);
@@ -878,13 +889,13 @@ static void sky2_vlan_rx_register(struct net_device *dev, struct vlan_group *grp
 	struct sky2_hw *hw = sky2->hw;
 	u16 port = sky2->port;
 
-	spin_lock(&sky2->tx_lock);
+	spin_lock_bh(&sky2->tx_lock);
 
 	sky2_write32(hw, SK_REG(port, RX_GMF_CTRL_T), RX_VLAN_STRIP_ON);
 	sky2_write32(hw, SK_REG(port, TX_GMF_CTRL_T), TX_VLAN_TAG_ON);
 	sky2->vlgrp = grp;
 
-	spin_unlock(&sky2->tx_lock);
+	spin_unlock_bh(&sky2->tx_lock);
 }
 
 static void sky2_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid)
@@ -893,27 +904,42 @@ static void sky2_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid)
 	struct sky2_hw *hw = sky2->hw;
 	u16 port = sky2->port;
 
-	spin_lock(&sky2->tx_lock);
+	spin_lock_bh(&sky2->tx_lock);
 
 	sky2_write32(hw, SK_REG(port, RX_GMF_CTRL_T), RX_VLAN_STRIP_OFF);
 	sky2_write32(hw, SK_REG(port, TX_GMF_CTRL_T), TX_VLAN_TAG_OFF);
 	if (sky2->vlgrp)
 		sky2->vlgrp->vlan_devices[vid] = NULL;
 
-	spin_unlock(&sky2->tx_lock);
+	spin_unlock_bh(&sky2->tx_lock);
 }
 #endif
 
+/*
+ * It appears the hardware has a bug in the FIFO logic that
+ * cause it to hang if the FIFO gets overrun and the receive buffer
+ * is not aligned. ALso alloc_skb() won't align properly if slab
+ * debugging is enabled.
+ */
+static inline struct sk_buff *sky2_alloc_skb(unsigned int size, gfp_t gfp_mask)
+{
+	struct sk_buff *skb;
+
+	skb = alloc_skb(size + RX_SKB_ALIGN, gfp_mask);
+	if (likely(skb)) {
+		unsigned long p	= (unsigned long) skb->data;
+		skb_reserve(skb,
+			((p + RX_SKB_ALIGN - 1) & ~(RX_SKB_ALIGN - 1)) - p);
+	}
+
+	return skb;
+}
+
 /*
  * Allocate and setup receiver buffer pool.
  * In case of 64 bit dma, there are 2X as many list elements
  * available as ring entries
  * and need to reserve one list element so we don't wrap around.
- *
- * It appears the hardware has a bug in the FIFO logic that
- * cause it to hang if the FIFO gets overrun and the receive buffer
- * is not aligned.  This means we can't use skb_reserve to align
- * the IP header.
  */
 static int sky2_rx_start(struct sky2_port *sky2)
 {
@@ -929,7 +955,7 @@ static int sky2_rx_start(struct sky2_port *sky2)
 	for (i = 0; i < sky2->rx_pending; i++) {
 		struct ring_info *re = sky2->rx_ring + i;
 
-		re->skb = dev_alloc_skb(sky2->rx_bufsize);
+		re->skb = sky2_alloc_skb(sky2->rx_bufsize, GFP_KERNEL);
 		if (!re->skb)
 			goto nomem;
 
@@ -986,19 +1012,19 @@ static int sky2_up(struct net_device *dev)
 
 	sky2_mac_init(hw, port);
 
-	/* Configure RAM buffers */
-	if (hw->chip_id == CHIP_ID_YUKON_FE ||
-	    (hw->chip_id == CHIP_ID_YUKON_EC && hw->chip_rev == 2))
-		ramsize = 4096;
-	else {
-		u8 e0 = sky2_read8(hw, B2_E_0);
-		ramsize = (e0 == 0) ? (128 * 1024) : (e0 * 4096);
-	}
+	/* Determine available ram buffer space (in 4K blocks).
+	 * Note: not sure about the FE setting below yet
+	 */
+	if (hw->chip_id == CHIP_ID_YUKON_FE)
+		ramsize = 4;
+	else
+		ramsize = sky2_read8(hw, B2_E_0);
+
+	/* Give transmitter one third (rounded up) */
+	rxspace = ramsize - (ramsize + 2) / 3;
 
-	/* 2/3 for Rx */
-	rxspace = (2 * ramsize) / 3;
 	sky2_ramset(hw, rxqaddr[port], 0, rxspace);
-	sky2_ramset(hw, txqaddr[port], rxspace, ramsize - rxspace);
+	sky2_ramset(hw, txqaddr[port], rxspace, ramsize);
 
 	/* Make sure SyncQ is disabled */
 	sky2_write8(hw, RB_ADDR(port == 0 ? Q_XS1 : Q_XS2, RB_CTRL),
@@ -1054,7 +1080,7 @@ static inline int tx_avail(const struct sky2_port *sky2)
 }
 
 /* Estimate of number of transmit list elements required */
-static inline unsigned tx_le_req(const struct sk_buff *skb)
+static unsigned tx_le_req(const struct sk_buff *skb)
 {
 	unsigned count;
 
@@ -1090,6 +1116,10 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev)
 	u16 mss;
 	u8 ctrl;
 
+	/* No BH disabling for tx_lock here.  We are running in BH disabled
+	 * context and TX reclaim runs via poll inside of a software
+	 * interrupt, and no related locks in IRQ processing.
+	 */
 	if (!spin_trylock(&sky2->tx_lock))
 		return NETDEV_TX_LOCKED;
 
@@ -1099,8 +1129,9 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev)
 		 */
 		if (!netif_queue_stopped(dev)) {
 			netif_stop_queue(dev);
-			printk(KERN_WARNING PFX "%s: ring full when queue awake!\n",
-			       dev->name);
+			if (net_ratelimit())
+				printk(KERN_WARNING PFX "%s: ring full when queue awake!\n",
+				       dev->name);
 		}
 		spin_unlock(&sky2->tx_lock);
 
@@ -1199,7 +1230,7 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev)
 
 		mapping = pci_map_page(hw->pdev, frag->page, frag->page_offset,
 				       frag->size, PCI_DMA_TODEVICE);
-		addr64 = (mapping >> 16) >> 16;
+		addr64 = high32(mapping);
 		if (addr64 != sky2->tx_addr64) {
 			le = get_tx_le(sky2);
 			le->tx.addr = cpu_to_le32(addr64);
@@ -1229,7 +1260,6 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev)
 		netif_stop_queue(dev);
 
 out_unlock:
-	mmiowb();
 	spin_unlock(&sky2->tx_lock);
 
 	dev->trans_start = jiffies;
@@ -1282,17 +1312,17 @@ static void sky2_tx_complete(struct sky2_port *sky2, u16 done)
 		dev_kfree_skb_any(skb);
 	}
 
-	spin_lock(&sky2->tx_lock);
 	sky2->tx_cons = put;
 	if (netif_queue_stopped(dev) && tx_avail(sky2) > MAX_SKB_TX_LE)
 		netif_wake_queue(dev);
-	spin_unlock(&sky2->tx_lock);
 }
 
 /* Cleanup all untransmitted buffers, assume transmitter not running */
 static void sky2_tx_clean(struct sky2_port *sky2)
 {
+	spin_lock_bh(&sky2->tx_lock);
 	sky2_tx_complete(sky2, sky2->tx_prod);
+	spin_unlock_bh(&sky2->tx_lock);
 }
 
 /* Network shutdown */
@@ -1582,28 +1612,40 @@ out:
 	local_irq_enable();
 }
 
+
+/* Transmit timeout is only called if we are running, carries is up
+ * and tx queue is full (stopped).
+ */
 static void sky2_tx_timeout(struct net_device *dev)
 {
 	struct sky2_port *sky2 = netdev_priv(dev);
 	struct sky2_hw *hw = sky2->hw;
 	unsigned txq = txqaddr[sky2->port];
+	u16 ridx;
+
+	/* Maybe we just missed an status interrupt */
+	spin_lock(&sky2->tx_lock);
+	ridx = sky2_read16(hw,
+			   sky2->port == 0 ? STAT_TXA1_RIDX : STAT_TXA2_RIDX);
+	sky2_tx_complete(sky2, ridx);
+	spin_unlock(&sky2->tx_lock);
+
+	if (!netif_queue_stopped(dev)) {
+		if (net_ratelimit())
+			pr_info(PFX "transmit interrupt missed? recovered\n");
+		return;
+	}
 
 	if (netif_msg_timer(sky2))
 		printk(KERN_ERR PFX "%s: tx timeout\n", dev->name);
 
-	netif_stop_queue(dev);
-
 	sky2_write32(hw, Q_ADDR(txq, Q_CSR), BMU_STOP);
-	sky2_read32(hw, Q_ADDR(txq, Q_CSR));
-
 	sky2_write32(hw, Y2_QADDR(txq, PREF_UNIT_CTRL), PREF_UNIT_RST_SET);
 
 	sky2_tx_clean(sky2);
 
 	sky2_qset(hw, txq);
 	sky2_prefetch_init(hw, txq, sky2->tx_le_map, TX_RING_SIZE - 1);
-
-	netif_wake_queue(dev);
 }
 
 
@@ -1713,7 +1755,7 @@ static struct sk_buff *sky2_receive(struct sky2_port *sky2,
 	} else {
 		struct sk_buff *nskb;
 
-		nskb = dev_alloc_skb(sky2->rx_bufsize);
+		nskb = sky2_alloc_skb(sky2->rx_bufsize, GFP_ATOMIC);
 		if (!nskb)
 			goto resubmit;
 
@@ -1745,7 +1787,7 @@ oversize:
 error:
 	++sky2->net_stats.rx_errors;
 
-	if (netif_msg_rx_err(sky2))
+	if (netif_msg_rx_err(sky2) && net_ratelimit())
 		printk(KERN_INFO PFX "%s: rx error, status 0x%x length %d\n",
 		       sky2->netdev->name, status, length);
 
@@ -1766,13 +1808,16 @@ error:
  */
 #define TX_NO_STATUS	0xffff
 
-static inline void sky2_tx_check(struct sky2_hw *hw, int port, u16 last)
+static void sky2_tx_check(struct sky2_hw *hw, int port, u16 last)
 {
 	if (last != TX_NO_STATUS) {
 		struct net_device *dev = hw->dev[port];
 		if (dev && netif_running(dev)) {
 			struct sky2_port *sky2 = netdev_priv(dev);
+
+			spin_lock(&sky2->tx_lock);
 			sky2_tx_complete(sky2, last);
+			spin_unlock(&sky2->tx_lock);
 		}
 	}
 }
@@ -1800,7 +1845,6 @@ static int sky2_poll(struct net_device *dev0, int *budget)
 		struct sk_buff *skb;
 		u32 status;
 		u16 length;
-		u8 op;
 
 		le = hw->st_le + hw->st_idx;
 		hw->st_idx = (hw->st_idx + 1) % STATUS_RING_SIZE;
@@ -1814,10 +1858,8 @@ static int sky2_poll(struct net_device *dev0, int *budget)
 		sky2 = netdev_priv(dev);
 		status = le32_to_cpu(le->status);
 		length = le16_to_cpu(le->length);
-		op = le->opcode & ~HW_OWNER;
-		le->opcode = 0;
 
-		switch (op) {
+		switch (le->opcode & ~HW_OWNER) {
 		case OP_RXSTAT:
 			skb = sky2_receive(sky2, length, status);
 			if (!skb)
@@ -1865,14 +1907,13 @@ static int sky2_poll(struct net_device *dev0, int *budget)
 		default:
 			if (net_ratelimit())
 				printk(KERN_WARNING PFX
-				       "unknown status opcode 0x%x\n", op);
+				       "unknown status opcode 0x%x\n", le->opcode);
 			break;
 		}
 	}
 
 exit_loop:
 	sky2_write32(hw, STAT_CTRL, SC_STAT_CLR_IRQ);
-	mmiowb();
 
 	sky2_tx_check(hw, 0, tx_done[0]);
 	sky2_tx_check(hw, 1, tx_done[1]);
@@ -1887,7 +1928,6 @@ exit_loop:
 		netif_rx_complete(dev0);
 		hw->intr_mask |= Y2_IS_STAT_BMU;
 		sky2_write32(hw, B0_IMSK, hw->intr_mask);
-		mmiowb();
 		return 0;
 	} else {
 		*budget -= work_done;
@@ -1900,35 +1940,42 @@ static void sky2_hw_error(struct sky2_hw *hw, unsigned port, u32 status)
 {
 	struct net_device *dev = hw->dev[port];
 
-	printk(KERN_INFO PFX "%s: hw error interrupt status 0x%x\n",
-	       dev->name, status);
+	if (net_ratelimit())
+		printk(KERN_INFO PFX "%s: hw error interrupt status 0x%x\n",
+		       dev->name, status);
 
 	if (status & Y2_IS_PAR_RD1) {
-		printk(KERN_ERR PFX "%s: ram data read parity error\n",
-		       dev->name);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: ram data read parity error\n",
+			       dev->name);
 		/* Clear IRQ */
 		sky2_write16(hw, RAM_BUFFER(port, B3_RI_CTRL), RI_CLR_RD_PERR);
 	}
 
 	if (status & Y2_IS_PAR_WR1) {
-		printk(KERN_ERR PFX "%s: ram data write parity error\n",
-		       dev->name);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: ram data write parity error\n",
+			       dev->name);
 
 		sky2_write16(hw, RAM_BUFFER(port, B3_RI_CTRL), RI_CLR_WR_PERR);
 	}
 
 	if (status & Y2_IS_PAR_MAC1) {
-		printk(KERN_ERR PFX "%s: MAC parity error\n", dev->name);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: MAC parity error\n", dev->name);
 		sky2_write8(hw, SK_REG(port, TX_GMF_CTRL_T), GMF_CLI_TX_PE);
 	}
 
 	if (status & Y2_IS_PAR_RX1) {
-		printk(KERN_ERR PFX "%s: RX parity error\n", dev->name);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: RX parity error\n", dev->name);
 		sky2_write32(hw, Q_ADDR(rxqaddr[port], Q_CSR), BMU_CLR_IRQ_PAR);
 	}
 
 	if (status & Y2_IS_TCP_TXA1) {
-		printk(KERN_ERR PFX "%s: TCP segmentation error\n", dev->name);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: TCP segmentation error\n",
+			       dev->name);
 		sky2_write32(hw, Q_ADDR(txqaddr[port], Q_CSR), BMU_CLR_IRQ_TCP);
 	}
 }
@@ -1944,8 +1991,9 @@ static void sky2_hw_intr(struct sky2_hw *hw)
 		u16 pci_err;
 
 		pci_read_config_word(hw->pdev, PCI_STATUS, &pci_err);
-		printk(KERN_ERR PFX "%s: pci hw error (0x%x)\n",
-		       pci_name(hw->pdev), pci_err);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: pci hw error (0x%x)\n",
+			       pci_name(hw->pdev), pci_err);
 
 		sky2_write8(hw, B2_TST_CTRL1, TST_CFG_WRITE_ON);
 		pci_write_config_word(hw->pdev, PCI_STATUS,
@@ -1959,8 +2007,9 @@ static void sky2_hw_intr(struct sky2_hw *hw)
 
 		pci_read_config_dword(hw->pdev, PEX_UNC_ERR_STAT, &pex_err);
 
-		printk(KERN_ERR PFX "%s: pci express error (0x%x)\n",
-		       pci_name(hw->pdev), pex_err);
+		if (net_ratelimit())
+			printk(KERN_ERR PFX "%s: pci express error (0x%x)\n",
+			       pci_name(hw->pdev), pex_err);
 
 		/* clear the interrupt */
 		sky2_write32(hw, B2_TST_CTRL1, TST_CFG_WRITE_ON);
@@ -2250,7 +2299,7 @@ static int sky2_reset(struct sky2_hw *hw)
 	return 0;
 }
 
-static inline u32 sky2_supported_modes(const struct sky2_hw *hw)
+static u32 sky2_supported_modes(const struct sky2_hw *hw)
 {
 	u32 modes;
 	if (hw->copper) {
@@ -2995,7 +3044,7 @@ static __devinit struct net_device *sky2_init_netdev(struct sky2_hw *hw,
 	return dev;
 }
 
-static inline void sky2_show_addr(struct net_device *dev)
+static void __devinit sky2_show_addr(struct net_device *dev)
 {
 	const struct sky2_port *sky2 = netdev_priv(dev);
 
@@ -3038,13 +3087,17 @@ static int __devinit sky2_probe(struct pci_dev *pdev,
 		goto err_out_free_regions;
 	}
 
-	if (sizeof(dma_addr_t) > sizeof(u32)) {
-		err = pci_set_dma_mask(pdev, DMA_64BIT_MASK);
-		if (!err)
-			using_dac = 1;
-	}
+	if (sizeof(dma_addr_t) > sizeof(u32) &&
+	    !(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) {
+		using_dac = 1;
+		err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK);
+		if (err < 0) {
+			printk(KERN_ERR PFX "%s unable to obtain 64 bit DMA "
+			       "for consistent allocations\n", pci_name(pdev));
+			goto err_out_free_regions;
+		}
 
-	if (!using_dac) {
+	} else {
 		err = pci_set_dma_mask(pdev, DMA_32BIT_MASK);
 		if (err) {
 			printk(KERN_ERR PFX "%s no usable DMA configuration\n",
@@ -3052,6 +3105,7 @@ static int __devinit sky2_probe(struct pci_dev *pdev,
 			goto err_out_free_regions;
 		}
 	}
+
 #ifdef __BIG_ENDIAN
 	/* byte swap descriptors in hardware */
 	{
@@ -3064,14 +3118,13 @@ static int __devinit sky2_probe(struct pci_dev *pdev,
 #endif
 
 	err = -ENOMEM;
-	hw = kmalloc(sizeof(*hw), GFP_KERNEL);
+	hw = kzalloc(sizeof(*hw), GFP_KERNEL);
 	if (!hw) {
 		printk(KERN_ERR PFX "%s: cannot allocate hardware struct\n",
 		       pci_name(pdev));
 		goto err_out_free_regions;
 	}
 
-	memset(hw, 0, sizeof(*hw));
 	hw->pdev = pdev;
 
 	hw->regs = ioremap_nocache(pci_resource_start(pdev, 0), 0x4000);

drivers/net/spider_net.c

@@ -22,7 +22,6 @@
  */
 
 #include <linux/config.h>
-
 #include <linux/compiler.h>
 #include <linux/crc32.h>
 #include <linux/delay.h>
@@ -30,6 +29,7 @@
 #include <linux/ethtool.h>
 #include <linux/firmware.h>
 #include <linux/if_vlan.h>
+#include <linux/in.h>
 #include <linux/init.h>
 #include <linux/ioport.h>
 #include <linux/ip.h>
@@ -43,6 +43,7 @@
 #include <linux/slab.h>
 #include <linux/tcp.h>
 #include <linux/types.h>
+#include <linux/vmalloc.h>
 #include <linux/wait.h>
 #include <linux/workqueue.h>
 #include <asm/bitops.h>
@@ -108,42 +109,6 @@ spider_net_write_reg(struct spider_net_card *card, u32 reg, u32 value)
 	writel(value, card->regs + reg);
 }
 
-/**
- * spider_net_write_reg_sync - writes to an SMMIO register of a card
- * @card: device structure
- * @reg: register to write to
- * @value: value to write into the specified SMMIO register
- *
- * Unlike spider_net_write_reg, this will also make sure the
- * data arrives on the card by reading the reg again.
- */
-static void
-spider_net_write_reg_sync(struct spider_net_card *card, u32 reg, u32 value)
-{
-	value = cpu_to_le32(value);
-	writel(value, card->regs + reg);
-	(void)readl(card->regs + reg);
-}
-
-/**
- * spider_net_rx_irq_off - switch off rx irq on this spider card
- * @card: device structure
- *
- * switches off rx irq by masking them out in the GHIINTnMSK register
- */
-static void
-spider_net_rx_irq_off(struct spider_net_card *card)
-{
-	u32 regvalue;
-	unsigned long flags;
-
-	spin_lock_irqsave(&card->intmask_lock, flags);
-	regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK);
-	regvalue &= ~SPIDER_NET_RXINT;
-	spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue);
-	spin_unlock_irqrestore(&card->intmask_lock, flags);
-}
-
 /** spider_net_write_phy - write to phy register
  * @netdev: adapter to be written to
  * @mii_id: id of MII
@@ -199,60 +164,33 @@ spider_net_read_phy(struct net_device *netdev, int mii_id, int reg)
 }
 
 /**
- * spider_net_rx_irq_on - switch on rx irq on this spider card
- * @card: device structure
- *
- * switches on rx irq by enabling them in the GHIINTnMSK register
- */
-static void
-spider_net_rx_irq_on(struct spider_net_card *card)
-{
-	u32 regvalue;
-	unsigned long flags;
-
-	spin_lock_irqsave(&card->intmask_lock, flags);
-	regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK);
-	regvalue |= SPIDER_NET_RXINT;
-	spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue);
-	spin_unlock_irqrestore(&card->intmask_lock, flags);
-}
-
-/**
- * spider_net_tx_irq_off - switch off tx irq on this spider card
+ * spider_net_rx_irq_off - switch off rx irq on this spider card
  * @card: device structure
  *
- * switches off tx irq by masking them out in the GHIINTnMSK register
+ * switches off rx irq by masking them out in the GHIINTnMSK register
  */
 static void
-spider_net_tx_irq_off(struct spider_net_card *card)
+spider_net_rx_irq_off(struct spider_net_card *card)
 {
 	u32 regvalue;
-	unsigned long flags;
 
-	spin_lock_irqsave(&card->intmask_lock, flags);
-	regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK);
-	regvalue &= ~SPIDER_NET_TXINT;
-	spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue);
-	spin_unlock_irqrestore(&card->intmask_lock, flags);
+	regvalue = SPIDER_NET_INT0_MASK_VALUE & (~SPIDER_NET_RXINT);
+	spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, regvalue);
 }
 
 /**
- * spider_net_tx_irq_on - switch on tx irq on this spider card
+ * spider_net_rx_irq_on - switch on rx irq on this spider card
  * @card: device structure
  *
- * switches on tx irq by enabling them in the GHIINTnMSK register
+ * switches on rx irq by enabling them in the GHIINTnMSK register
  */
 static void
-spider_net_tx_irq_on(struct spider_net_card *card)
+spider_net_rx_irq_on(struct spider_net_card *card)
 {
 	u32 regvalue;
-	unsigned long flags;
 
-	spin_lock_irqsave(&card->intmask_lock, flags);
-	regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK);
-	regvalue |= SPIDER_NET_TXINT;
-	spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue);
-	spin_unlock_irqrestore(&card->intmask_lock, flags);
+	regvalue = SPIDER_NET_INT0_MASK_VALUE | SPIDER_NET_RXINT;
+	spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, regvalue);
 }
 
 /**
@@ -326,9 +264,8 @@ static enum spider_net_descr_status
 spider_net_get_descr_status(struct spider_net_descr *descr)
 {
 	u32 cmd_status;
-	rmb();
+
 	cmd_status = descr->dmac_cmd_status;
-	rmb();
 	cmd_status >>= SPIDER_NET_DESCR_IND_PROC_SHIFT;
 	/* no need to mask out any bits, as cmd_status is 32 bits wide only
 	 * (and unsigned) */
@@ -349,7 +286,6 @@ spider_net_set_descr_status(struct spider_net_descr *descr,
 {
 	u32 cmd_status;
 	/* read the status */
-	mb();
 	cmd_status = descr->dmac_cmd_status;
 	/* clean the upper 4 bits */
 	cmd_status &= SPIDER_NET_DESCR_IND_PROC_MASKO;
@@ -357,7 +293,6 @@ spider_net_set_descr_status(struct spider_net_descr *descr,
 	cmd_status |= ((u32)status)<<SPIDER_NET_DESCR_IND_PROC_SHIFT;
 	/* and write it back */
 	descr->dmac_cmd_status = cmd_status;
-	wmb();
 }
 
 /**
@@ -398,8 +333,9 @@ spider_net_init_chain(struct spider_net_card *card,
 {
 	int i;
 	struct spider_net_descr *descr;
+	dma_addr_t buf;
 
-	spin_lock_init(&card->chain_lock);
+	atomic_set(&card->rx_chain_refill,0);
 
 	descr = start_descr;
 	memset(descr, 0, sizeof(*descr) * no);
@@ -408,14 +344,14 @@ spider_net_init_chain(struct spider_net_card *card,
 	for (i=0; i<no; i++, descr++) {
 		spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE);
 
-		descr->bus_addr =
-			pci_map_single(card->pdev, descr,
-				       SPIDER_NET_DESCR_SIZE,
-				       PCI_DMA_BIDIRECTIONAL);
+		buf = pci_map_single(card->pdev, descr,
+				     SPIDER_NET_DESCR_SIZE,
+				     PCI_DMA_BIDIRECTIONAL);
 
-		if (descr->bus_addr == DMA_ERROR_CODE)
+		if (buf == DMA_ERROR_CODE)
 			goto iommu_error;
 
+		descr->bus_addr = buf;
 		descr->next = descr + 1;
 		descr->prev = descr - 1;
 
@@ -439,7 +375,8 @@ iommu_error:
 	for (i=0; i < no; i++, descr++)
 		if (descr->bus_addr)
 			pci_unmap_single(card->pdev, descr->bus_addr,
-					 SPIDER_NET_DESCR_SIZE, PCI_DMA_BIDIRECTIONAL);
+					 SPIDER_NET_DESCR_SIZE,
+					 PCI_DMA_BIDIRECTIONAL);
 	return -ENOMEM;
 }
 
@@ -459,7 +396,7 @@ spider_net_free_rx_chain_contents(struct spider_net_card *card)
 		if (descr->skb) {
 			dev_kfree_skb(descr->skb);
 			pci_unmap_single(card->pdev, descr->buf_addr,
-					 SPIDER_NET_MAX_MTU,
+					 SPIDER_NET_MAX_FRAME,
 					 PCI_DMA_BIDIRECTIONAL);
 		}
 		descr = descr->next;
@@ -480,12 +417,13 @@ static int
 spider_net_prepare_rx_descr(struct spider_net_card *card,
 			    struct spider_net_descr *descr)
 {
+	dma_addr_t buf;
 	int error = 0;
 	int offset;
 	int bufsize;
 
 	/* we need to round up the buffer size to a multiple of 128 */
-	bufsize = (SPIDER_NET_MAX_MTU + SPIDER_NET_RXBUF_ALIGN - 1) &
+	bufsize = (SPIDER_NET_MAX_FRAME + SPIDER_NET_RXBUF_ALIGN - 1) &
 		(~(SPIDER_NET_RXBUF_ALIGN - 1));
 
 	/* and we need to have it 128 byte aligned, therefore we allocate a
@@ -493,10 +431,8 @@ spider_net_prepare_rx_descr(struct spider_net_card *card,
 	/* allocate an skb */
 	descr->skb = dev_alloc_skb(bufsize + SPIDER_NET_RXBUF_ALIGN - 1);
 	if (!descr->skb) {
-		if (net_ratelimit())
-			if (netif_msg_rx_err(card))
-				pr_err("Not enough memory to allocate "
-					"rx buffer\n");
+		if (netif_msg_rx_err(card) && net_ratelimit())
+			pr_err("Not enough memory to allocate rx buffer\n");
 		return -ENOMEM;
 	}
 	descr->buf_size = bufsize;
@@ -510,12 +446,12 @@ spider_net_prepare_rx_descr(struct spider_net_card *card,
 	if (offset)
 		skb_reserve(descr->skb, SPIDER_NET_RXBUF_ALIGN - offset);
 	/* io-mmu-map the skb */
-	descr->buf_addr = pci_map_single(card->pdev, descr->skb->data,
-					 SPIDER_NET_MAX_MTU,
-					 PCI_DMA_BIDIRECTIONAL);
-	if (descr->buf_addr == DMA_ERROR_CODE) {
+	buf = pci_map_single(card->pdev, descr->skb->data,
+			     SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL);
+	descr->buf_addr = buf;
+	if (buf == DMA_ERROR_CODE) {
 		dev_kfree_skb_any(descr->skb);
-		if (netif_msg_rx_err(card))
+		if (netif_msg_rx_err(card) && net_ratelimit())
 			pr_err("Could not iommu-map rx buffer\n");
 		spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE);
 	} else {
@@ -526,10 +462,10 @@ spider_net_prepare_rx_descr(struct spider_net_card *card,
 }
 
 /**
- * spider_net_enable_rxctails - sets RX dmac chain tail addresses
+ * spider_net_enable_rxchtails - sets RX dmac chain tail addresses
  * @card: card structure
  *
- * spider_net_enable_rxctails sets the RX DMAC chain tail adresses in the
+ * spider_net_enable_rxchtails sets the RX DMAC chain tail adresses in the
  * chip by writing to the appropriate register. DMA is enabled in
  * spider_net_enable_rxdmac.
  */
@@ -551,6 +487,7 @@ spider_net_enable_rxchtails(struct spider_net_card *card)
 static void
 spider_net_enable_rxdmac(struct spider_net_card *card)
 {
+	wmb();
 	spider_net_write_reg(card, SPIDER_NET_GDADMACCNTR,
 			     SPIDER_NET_DMA_RX_VALUE);
 }
@@ -559,32 +496,28 @@ spider_net_enable_rxdmac(struct spider_net_card *card)
  * spider_net_refill_rx_chain - refills descriptors/skbs in the rx chains
  * @card: card structure
  *
- * refills descriptors in all chains (last used chain first): allocates skbs
- * and iommu-maps them.
+ * refills descriptors in the rx chain: allocates skbs and iommu-maps them.
  */
 static void
 spider_net_refill_rx_chain(struct spider_net_card *card)
 {
 	struct spider_net_descr_chain *chain;
-	int count = 0;
-	unsigned long flags;
 
 	chain = &card->rx_chain;
 
-	spin_lock_irqsave(&card->chain_lock, flags);
-	while (spider_net_get_descr_status(chain->head) ==
-				SPIDER_NET_DESCR_NOT_IN_USE) {
-		if (spider_net_prepare_rx_descr(card, chain->head))
-			break;
-		count++;
-		chain->head = chain->head->next;
-	}
-	spin_unlock_irqrestore(&card->chain_lock, flags);
+	/* one context doing the refill (and a second context seeing that
+	 * and omitting it) is ok. If called by NAPI, we'll be called again
+	 * as spider_net_decode_one_descr is called several times. If some
+	 * interrupt calls us, the NAPI is about to clean up anyway. */
+	if (atomic_inc_return(&card->rx_chain_refill) == 1)
+		while (spider_net_get_descr_status(chain->head) ==
+		       SPIDER_NET_DESCR_NOT_IN_USE) {
+			if (spider_net_prepare_rx_descr(card, chain->head))
+				break;
+			chain->head = chain->head->next;
+		}
 
-	/* could be optimized, only do that, if we know the DMA processing
-	 * has terminated */
-	if (count)
-		spider_net_enable_rxdmac(card);
+	atomic_dec(&card->rx_chain_refill);
 }
 
 /**
@@ -613,6 +546,7 @@ spider_net_alloc_rx_skbs(struct spider_net_card *card)
 	/* this will allocate the rest of the rx buffers; if not, it's
 	 * business as usual later on */
 	spider_net_refill_rx_chain(card);
+	spider_net_enable_rxdmac(card);
 	return 0;
 
 error:
@@ -649,24 +583,30 @@ spider_net_release_tx_descr(struct spider_net_card *card,
  * @card: adapter structure
  * @brutal: if set, don't care about whether descriptor seems to be in use
  *
- * releases the tx descriptors that spider has finished with (if non-brutal)
- * or simply release tx descriptors (if brutal)
+ * returns 0 if the tx ring is empty, otherwise 1.
+ *
+ * spider_net_release_tx_chain releases the tx descriptors that spider has
+ * finished with (if non-brutal) or simply release tx descriptors (if brutal).
+ * If some other context is calling this function, we return 1 so that we're
+ * scheduled again (if we were scheduled) and will not loose initiative.
  */
-static void
+static int
 spider_net_release_tx_chain(struct spider_net_card *card, int brutal)
 {
 	struct spider_net_descr_chain *tx_chain = &card->tx_chain;
 	enum spider_net_descr_status status;
 
-	spider_net_tx_irq_off(card);
+	if (atomic_inc_return(&card->tx_chain_release) != 1) {
+		atomic_dec(&card->tx_chain_release);
+		return 1;
+	}
 
-	/* no lock for chain needed, if this is only executed once at a time */
-again:
 	for (;;) {
 		status = spider_net_get_descr_status(tx_chain->tail);
 		switch (status) {
 		case SPIDER_NET_DESCR_CARDOWNED:
-			if (!brutal) goto out;
+			if (!brutal)
+				goto out;
 			/* fallthrough, if we release the descriptors
 			 * brutally (then we don't care about
 			 * SPIDER_NET_DESCR_CARDOWNED) */
@@ -693,25 +633,30 @@ again:
 		tx_chain->tail = tx_chain->tail->next;
 	}
 out:
+	atomic_dec(&card->tx_chain_release);
+
 	netif_wake_queue(card->netdev);
 
-	if (!brutal) {
-		/* switch on tx irqs (while we are still in the interrupt
-		 * handler, so we don't get an interrupt), check again
-		 * for done descriptors. This results in fewer interrupts */
-		spider_net_tx_irq_on(card);
-		status = spider_net_get_descr_status(tx_chain->tail);
-		switch (status) {
-			case SPIDER_NET_DESCR_RESPONSE_ERROR:
-			case SPIDER_NET_DESCR_PROTECTION_ERROR:
-			case SPIDER_NET_DESCR_FORCE_END:
-			case SPIDER_NET_DESCR_COMPLETE:
-				goto again;
-			default:
-				break;
-		}
-	}
+	if (status == SPIDER_NET_DESCR_CARDOWNED)
+		return 1;
+	return 0;
+}
 
+/**
+ * spider_net_cleanup_tx_ring - cleans up the TX ring
+ * @card: card structure
+ *
+ * spider_net_cleanup_tx_ring is called by the tx_timer (as we don't use
+ * interrupts to cleanup our TX ring) and returns sent packets to the stack
+ * by freeing them
+ */
+static void
+spider_net_cleanup_tx_ring(struct spider_net_card *card)
+{
+	if ( (spider_net_release_tx_chain(card, 0)) &&
+	      (card->netdev->flags & IFF_UP) ) {
+		mod_timer(&card->tx_timer, jiffies + SPIDER_NET_TX_TIMER);
+	}
 }
 
 /**
@@ -726,16 +671,22 @@ out:
 static u8
 spider_net_get_multicast_hash(struct net_device *netdev, __u8 *addr)
 {
-	/* FIXME: an addr of 01:00:5e:00:00:01 must result in 0xa9,
-	 * ff:ff:ff:ff:ff:ff must result in 0xfd */
 	u32 crc;
 	u8 hash;
+	char addr_for_crc[ETH_ALEN] = { 0, };
+	int i, bit;
 
-	crc = crc32_be(~0, addr, netdev->addr_len);
+	for (i = 0; i < ETH_ALEN * 8; i++) {
+		bit = (addr[i / 8] >> (i % 8)) & 1;
+		addr_for_crc[ETH_ALEN - 1 - i / 8] += bit << (7 - (i % 8));
+	}
+
+	crc = crc32_be(~0, addr_for_crc, netdev->addr_len);
 
 	hash = (crc >> 27);
 	hash <<= 3;
 	hash |= crc & 7;
+	hash &= 0xff;
 
 	return hash;
 }
@@ -821,9 +772,11 @@ spider_net_stop(struct net_device *netdev)
 {
 	struct spider_net_card *card = netdev_priv(netdev);
 
+	tasklet_kill(&card->rxram_full_tl);
 	netif_poll_disable(netdev);
 	netif_carrier_off(netdev);
 	netif_stop_queue(netdev);
+	del_timer_sync(&card->tx_timer);
 
 	/* disable/mask all interrupts */
 	spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, 0);
@@ -872,13 +825,15 @@ spider_net_get_next_tx_descr(struct spider_net_card *card)
  * @skb: packet to consider
  *
  * fills out the command and status field of the descriptor structure,
- * depending on hardware checksum settings. This function assumes a wmb()
- * has executed before.
+ * depending on hardware checksum settings.
  */
 static void
 spider_net_set_txdescr_cmdstat(struct spider_net_descr *descr,
 			       struct sk_buff *skb)
 {
+	/* make sure the other fields in the descriptor are written */
+	wmb();
+
 	if (skb->ip_summed != CHECKSUM_HW) {
 		descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_NOCS;
 		return;
@@ -887,14 +842,13 @@ spider_net_set_txdescr_cmdstat(struct spider_net_descr *descr,
 	/* is packet ip?
 	 * if yes: tcp? udp? */
 	if (skb->protocol == htons(ETH_P_IP)) {
-		if (skb->nh.iph->protocol == IPPROTO_TCP) {
+		if (skb->nh.iph->protocol == IPPROTO_TCP)
 			descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_TCPCS;
-		} else if (skb->nh.iph->protocol == IPPROTO_UDP) {
+		else if (skb->nh.iph->protocol == IPPROTO_UDP)
 			descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_UDPCS;
-		} else { /* the stack should checksum non-tcp and non-udp
-			    packets on his own: NETIF_F_IP_CSUM */
+		else /* the stack should checksum non-tcp and non-udp
+			packets on his own: NETIF_F_IP_CSUM */
 			descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_NOCS;
-		}
 	}
 }
 
@@ -914,23 +868,22 @@ spider_net_prepare_tx_descr(struct spider_net_card *card,
 			    struct spider_net_descr *descr,
 			    struct sk_buff *skb)
 {
-	descr->buf_addr = pci_map_single(card->pdev, skb->data,
-					 skb->len, PCI_DMA_BIDIRECTIONAL);
-	if (descr->buf_addr == DMA_ERROR_CODE) {
-		if (netif_msg_tx_err(card))
+	dma_addr_t buf;
+
+	buf = pci_map_single(card->pdev, skb->data,
+			     skb->len, PCI_DMA_BIDIRECTIONAL);
+	if (buf == DMA_ERROR_CODE) {
+		if (netif_msg_tx_err(card) && net_ratelimit())
 			pr_err("could not iommu-map packet (%p, %i). "
 				  "Dropping packet\n", skb->data, skb->len);
 		return -ENOMEM;
 	}
 
+	descr->buf_addr = buf;
 	descr->buf_size = skb->len;
 	descr->skb = skb;
 	descr->data_status = 0;
 
-	/* make sure the above values are in memory before we change the
-	 * status */
-	wmb();
-
 	spider_net_set_txdescr_cmdstat(descr,skb);
 
 	return 0;
@@ -972,17 +925,12 @@ spider_net_xmit(struct sk_buff *skb, struct net_device *netdev)
 	struct spider_net_descr *descr;
 	int result;
 
-	descr = spider_net_get_next_tx_descr(card);
+	spider_net_release_tx_chain(card, 0);
 
-	if (!descr) {
-		netif_stop_queue(netdev);
+	descr = spider_net_get_next_tx_descr(card);
 
-		descr = spider_net_get_next_tx_descr(card);
-		if (!descr)
-			goto error;
-		else
-			netif_start_queue(netdev);
-	}
+	if (!descr)
+		goto error;
 
 	result = spider_net_prepare_tx_descr(card, descr, skb);
 	if (result)
@@ -990,19 +938,25 @@ spider_net_xmit(struct sk_buff *skb, struct net_device *netdev)
 
 	card->tx_chain.head = card->tx_chain.head->next;
 
-	/* make sure the status from spider_net_prepare_tx_descr is in
-	 * memory before we check out the previous descriptor */
-	wmb();
-
 	if (spider_net_get_descr_status(descr->prev) !=
-	    SPIDER_NET_DESCR_CARDOWNED)
-		spider_net_kick_tx_dma(card, descr);
+	    SPIDER_NET_DESCR_CARDOWNED) {
+		/* make sure the current descriptor is in memory. Then
+		 * kicking it on again makes sense, if the previous is not
+		 * card-owned anymore. Check the previous descriptor twice
+		 * to omit an mb() in heavy traffic cases */
+		mb();
+		if (spider_net_get_descr_status(descr->prev) !=
+		    SPIDER_NET_DESCR_CARDOWNED)
+			spider_net_kick_tx_dma(card, descr);
+	}
+
+	mod_timer(&card->tx_timer, jiffies + SPIDER_NET_TX_TIMER);
 
 	return NETDEV_TX_OK;
 
 error:
 	card->netdev_stats.tx_dropped++;
-	return NETDEV_TX_LOCKED;
+	return NETDEV_TX_BUSY;
 }
 
 /**
@@ -1027,6 +981,7 @@ spider_net_do_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
  * spider_net_pass_skb_up - takes an skb from a descriptor and passes it on
  * @descr: descriptor to process
  * @card: card structure
+ * @napi: whether caller is in NAPI context
  *
  * returns 1 on success, 0 if no packet was passed to the stack
  *
@@ -1035,7 +990,7 @@ spider_net_do_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
  */
 static int
 spider_net_pass_skb_up(struct spider_net_descr *descr,
-		       struct spider_net_card *card)
+		       struct spider_net_card *card, int napi)
 {
 	struct sk_buff *skb;
 	struct net_device *netdev;
@@ -1046,22 +1001,20 @@ spider_net_pass_skb_up(struct spider_net_descr *descr,
 
 	netdev = card->netdev;
 
-	/* check for errors in the data_error flag */
-	if ((data_error & SPIDER_NET_DATA_ERROR_MASK) &&
-	    netif_msg_rx_err(card))
-		pr_err("error in received descriptor found, "
-		       "data_status=x%08x, data_error=x%08x\n",
-		       data_status, data_error);
-
-	/* prepare skb, unmap descriptor */
-	skb = descr->skb;
-	pci_unmap_single(card->pdev, descr->buf_addr, SPIDER_NET_MAX_MTU,
+	/* unmap descriptor */
+	pci_unmap_single(card->pdev, descr->buf_addr, SPIDER_NET_MAX_FRAME,
 			 PCI_DMA_BIDIRECTIONAL);
 
 	/* the cases we'll throw away the packet immediately */
-	if (data_error & SPIDER_NET_DESTROY_RX_FLAGS)
+	if (data_error & SPIDER_NET_DESTROY_RX_FLAGS) {
+		if (netif_msg_rx_err(card))
+			pr_err("error in received descriptor found, "
+			       "data_status=x%08x, data_error=x%08x\n",
+			       data_status, data_error);
 		return 0;
+	}
 
+	skb = descr->skb;
 	skb->dev = netdev;
 	skb_put(skb, descr->valid_size);
 
@@ -1073,14 +1026,14 @@ spider_net_pass_skb_up(struct spider_net_descr *descr,
 
 	/* checksum offload */
 	if (card->options.rx_csum) {
-		if ( (data_status & SPIDER_NET_DATA_STATUS_CHK_MASK) &&
-		     (!(data_error & SPIDER_NET_DATA_ERROR_CHK_MASK)) )
+		if ( ( (data_status & SPIDER_NET_DATA_STATUS_CKSUM_MASK) ==
+		       SPIDER_NET_DATA_STATUS_CKSUM_MASK) &&
+		     !(data_error & SPIDER_NET_DATA_ERR_CKSUM_MASK))
 			skb->ip_summed = CHECKSUM_UNNECESSARY;
 		else
 			skb->ip_summed = CHECKSUM_NONE;
-	} else {
+	} else
 		skb->ip_summed = CHECKSUM_NONE;
-	}
 
 	if (data_status & SPIDER_NET_VLAN_PACKET) {
 		/* further enhancements: HW-accel VLAN
@@ -1089,7 +1042,10 @@ spider_net_pass_skb_up(struct spider_net_descr *descr,
 	}
 
 	/* pass skb up to stack */
-	netif_receive_skb(skb);
+	if (napi)
+		netif_receive_skb(skb);
+	else
+		netif_rx_ni(skb);
 
 	/* update netdevice statistics */
 	card->netdev_stats.rx_packets++;
@@ -1099,16 +1055,18 @@ spider_net_pass_skb_up(struct spider_net_descr *descr,
 }
 
 /**
- * spider_net_decode_descr - processes an rx descriptor
+ * spider_net_decode_one_descr - processes an rx descriptor
  * @card: card structure
+ * @napi: whether caller is in NAPI context
  *
  * returns 1 if a packet has been sent to the stack, otherwise 0
  *
  * processes an rx descriptor by iommu-unmapping the data buffer and passing
- * the packet up to the stack
+ * the packet up to the stack. This function is called in softirq
+ * context, e.g. either bottom half from interrupt or NAPI polling context
  */
 static int
-spider_net_decode_one_descr(struct spider_net_card *card)
+spider_net_decode_one_descr(struct spider_net_card *card, int napi)
 {
 	enum spider_net_descr_status status;
 	struct spider_net_descr *descr;
@@ -1122,17 +1080,19 @@ spider_net_decode_one_descr(struct spider_net_card *card)
 
 	if (status == SPIDER_NET_DESCR_CARDOWNED) {
 		/* nothing in the descriptor yet */
-		return 0;
+		result=0;
+		goto out;
 	}
 
 	if (status == SPIDER_NET_DESCR_NOT_IN_USE) {
-		/* not initialized yet, I bet chain->tail == chain->head
-		 * and the ring is empty */
+		/* not initialized yet, the ring must be empty */
 		spider_net_refill_rx_chain(card);
-		return 0;
+		spider_net_enable_rxdmac(card);
+		result=0;
+		goto out;
 	}
 
-	/* descriptor definitively used -- move on head */
+	/* descriptor definitively used -- move on tail */
 	chain->tail = descr->next;
 
 	result = 0;
@@ -1143,6 +1103,9 @@ spider_net_decode_one_descr(struct spider_net_card *card)
 			pr_err("%s: dropping RX descriptor with state %d\n",
 			       card->netdev->name, status);
 		card->netdev_stats.rx_dropped++;
+		pci_unmap_single(card->pdev, descr->buf_addr,
+				 SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL);
+		dev_kfree_skb_irq(descr->skb);
 		goto refill;
 	}
 
@@ -1155,12 +1118,13 @@ spider_net_decode_one_descr(struct spider_net_card *card)
 	}
 
 	/* ok, we've got a packet in descr */
-	result = spider_net_pass_skb_up(descr, card);
+	result = spider_net_pass_skb_up(descr, card, napi);
 refill:
 	spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE);
 	/* change the descriptor state: */
-	spider_net_refill_rx_chain(card);
-
+	if (!napi)
+		spider_net_refill_rx_chain(card);
+out:
 	return result;
 }
 
@@ -1186,7 +1150,7 @@ spider_net_poll(struct net_device *netdev, int *budget)
 	packets_to_do = min(*budget, netdev->quota);
 
 	while (packets_to_do) {
-		if (spider_net_decode_one_descr(card)) {
+		if (spider_net_decode_one_descr(card, 1)) {
 			packets_done++;
 			packets_to_do--;
 		} else {
@@ -1198,6 +1162,7 @@ spider_net_poll(struct net_device *netdev, int *budget)
 
 	netdev->quota -= packets_done;
 	*budget -= packets_done;
+	spider_net_refill_rx_chain(card);
 
 	/* if all packets are in the stack, enable interrupts and return 0 */
 	/* if not, return 1 */
@@ -1341,6 +1306,24 @@ spider_net_enable_txdmac(struct spider_net_card *card)
 			     card->tx_chain.tail->bus_addr);
 }
 
+/**
+ * spider_net_handle_rxram_full - cleans up RX ring upon RX RAM full interrupt
+ * @card: card structure
+ *
+ * spider_net_handle_rxram_full empties the RX ring so that spider can put
+ * more packets in it and empty its RX RAM. This is called in bottom half
+ * context
+ */
+static void
+spider_net_handle_rxram_full(struct spider_net_card *card)
+{
+	while (spider_net_decode_one_descr(card, 0))
+		;
+	spider_net_enable_rxchtails(card);
+	spider_net_enable_rxdmac(card);
+	netif_rx_schedule(card->netdev);
+}
+
 /**
  * spider_net_handle_error_irq - handles errors raised by an interrupt
  * @card: card structure
@@ -1449,17 +1432,21 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg)
 				switch (i)
 	{
 	case SPIDER_NET_GTMFLLINT:
-		if (netif_msg_intr(card))
+		if (netif_msg_intr(card) && net_ratelimit())
 			pr_err("Spider TX RAM full\n");
 		show_error = 0;
 		break;
+	case SPIDER_NET_GRFDFLLINT: /* fallthrough */
+	case SPIDER_NET_GRFCFLLINT: /* fallthrough */
+	case SPIDER_NET_GRFBFLLINT: /* fallthrough */
+	case SPIDER_NET_GRFAFLLINT: /* fallthrough */
 	case SPIDER_NET_GRMFLLINT:
-		if (netif_msg_intr(card))
+		if (netif_msg_intr(card) && net_ratelimit())
 			pr_err("Spider RX RAM full, incoming packets "
-			       "might be discarded !\n");
-		netif_rx_schedule(card->netdev);
-		spider_net_enable_rxchtails(card);
-		spider_net_enable_rxdmac(card);
+			       "might be discarded!\n");
+		spider_net_rx_irq_off(card);
+		tasklet_schedule(&card->rxram_full_tl);
+		show_error = 0;
 		break;
 
 	/* case SPIDER_NET_GTMSHTINT: problem, print a message */
@@ -1467,10 +1454,6 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg)
 		/* allrighty. tx from previous descr ok */
 		show_error = 0;
 		break;
-	/* case SPIDER_NET_GRFDFLLINT: print a message down there */
-	/* case SPIDER_NET_GRFCFLLINT: print a message down there */
-	/* case SPIDER_NET_GRFBFLLINT: print a message down there */
-	/* case SPIDER_NET_GRFAFLLINT: print a message down there */
 
 	/* chain end */
 	case SPIDER_NET_GDDDCEINT: /* fallthrough */
@@ -1482,6 +1465,7 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg)
 			       "restarting DMAC %c.\n",
 			       'D'+i-SPIDER_NET_GDDDCEINT);
 		spider_net_refill_rx_chain(card);
+		spider_net_enable_rxdmac(card);
 		show_error = 0;
 		break;
 
@@ -1492,6 +1476,7 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg)
 	case SPIDER_NET_GDAINVDINT:
 		/* could happen when rx chain is full */
 		spider_net_refill_rx_chain(card);
+		spider_net_enable_rxdmac(card);
 		show_error = 0;
 		break;
 
@@ -1580,17 +1565,13 @@ spider_net_interrupt(int irq, void *ptr, struct pt_regs *regs)
 	if (!status_reg)
 		return IRQ_NONE;
 
-	if (status_reg & SPIDER_NET_TXINT)
-		spider_net_release_tx_chain(card, 0);
-
 	if (status_reg & SPIDER_NET_RXINT ) {
 		spider_net_rx_irq_off(card);
 		netif_rx_schedule(netdev);
 	}
 
-	/* we do this after rx and tx processing, as we want the tx chain
-	 * processed to see, whether we should restart tx dma processing */
-	spider_net_handle_error_irq(card, status_reg);
+	if (status_reg & SPIDER_NET_ERRINT )
+		spider_net_handle_error_irq(card, status_reg);
 
 	/* clear interrupt sources */
 	spider_net_write_reg(card, SPIDER_NET_GHIINT0STS, status_reg);
@@ -1831,34 +1812,40 @@ spider_net_setup_phy(struct spider_net_card *card)
 /**
  * spider_net_download_firmware - loads firmware into the adapter
  * @card: card structure
- * @firmware: firmware pointer
+ * @firmware_ptr: pointer to firmware data
  *
- * spider_net_download_firmware loads the firmware opened by
- * spider_net_init_firmware into the adapter.
+ * spider_net_download_firmware loads the firmware data into the
+ * adapter. It assumes the length etc. to be allright.
  */
-static void
+static int
 spider_net_download_firmware(struct spider_net_card *card,
-			     const struct firmware *firmware)
+			     u8 *firmware_ptr)
 {
 	int sequencer, i;
-	u32 *fw_ptr = (u32 *)firmware->data;
+	u32 *fw_ptr = (u32 *)firmware_ptr;
 
 	/* stop sequencers */
 	spider_net_write_reg(card, SPIDER_NET_GSINIT,
 			     SPIDER_NET_STOP_SEQ_VALUE);
 
-	for (sequencer = 0; sequencer < 6; sequencer++) {
+	for (sequencer = 0; sequencer < SPIDER_NET_FIRMWARE_SEQS;
+	     sequencer++) {
 		spider_net_write_reg(card,
 				     SPIDER_NET_GSnPRGADR + sequencer * 8, 0);
-		for (i = 0; i < SPIDER_NET_FIRMWARE_LEN; i++) {
+		for (i = 0; i < SPIDER_NET_FIRMWARE_SEQWORDS; i++) {
 			spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT +
 					     sequencer * 8, *fw_ptr);
 			fw_ptr++;
 		}
 	}
 
+	if (spider_net_read_reg(card, SPIDER_NET_GSINIT))
+		return -EIO;
+
 	spider_net_write_reg(card, SPIDER_NET_GSINIT,
 			     SPIDER_NET_RUN_SEQ_VALUE);
+
+	return 0;
 }
 
 /**
@@ -1890,31 +1877,53 @@ spider_net_download_firmware(struct spider_net_card *card,
 static int
 spider_net_init_firmware(struct spider_net_card *card)
 {
-	const struct firmware *firmware;
-	int err = -EIO;
+	struct firmware *firmware = NULL;
+	struct device_node *dn;
+	u8 *fw_prop = NULL;
+	int err = -ENOENT;
+	int fw_size;
+
+	if (request_firmware((const struct firmware **)&firmware,
+			     SPIDER_NET_FIRMWARE_NAME, &card->pdev->dev) == 0) {
+		if ( (firmware->size != SPIDER_NET_FIRMWARE_LEN) &&
+		     netif_msg_probe(card) ) {
+			pr_err("Incorrect size of spidernet firmware in " \
+			       "filesystem. Looking in host firmware...\n");
+			goto try_host_fw;
+		}
+		err = spider_net_download_firmware(card, firmware->data);
 
-	if (request_firmware(&firmware,
-			     SPIDER_NET_FIRMWARE_NAME, &card->pdev->dev) < 0) {
-		if (netif_msg_probe(card))
-			pr_err("Couldn't read in sequencer data file %s.\n",
-			       SPIDER_NET_FIRMWARE_NAME);
-		firmware = NULL;
-		goto out;
-	}
+		release_firmware(firmware);
+		if (err)
+			goto try_host_fw;
 
-	if (firmware->size != 6 * SPIDER_NET_FIRMWARE_LEN * sizeof(u32)) {
-		if (netif_msg_probe(card))
-			pr_err("Invalid size of sequencer data file %s.\n",
-			       SPIDER_NET_FIRMWARE_NAME);
-		goto out;
+		goto done;
 	}
 
-	spider_net_download_firmware(card, firmware);
+try_host_fw:
+	dn = pci_device_to_OF_node(card->pdev);
+	if (!dn)
+		goto out_err;
 
-	err = 0;
-out:
-	release_firmware(firmware);
+	fw_prop = (u8 *)get_property(dn, "firmware", &fw_size);
+	if (!fw_prop)
+		goto out_err;
+
+	if ( (fw_size != SPIDER_NET_FIRMWARE_LEN) &&
+	     netif_msg_probe(card) ) {
+		pr_err("Incorrect size of spidernet firmware in " \
+		       "host firmware\n");
+		goto done;
+	}
 
+	err = spider_net_download_firmware(card, fw_prop);
+
+done:
+	return err;
+out_err:
+	if (netif_msg_probe(card))
+		pr_err("Couldn't find spidernet firmware in filesystem " \
+		       "or host firmware\n");
 	return err;
 }
 
@@ -1934,10 +1943,11 @@ spider_net_workaround_rxramfull(struct spider_net_card *card)
 			     SPIDER_NET_CKRCTRL_RUN_VALUE);
 
 	/* empty sequencer data */
-	for (sequencer = 0; sequencer < 6; sequencer++) {
+	for (sequencer = 0; sequencer < SPIDER_NET_FIRMWARE_SEQS;
+	     sequencer++) {
 		spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT +
 				     sequencer * 8, 0x0);
-		for (i = 0; i < SPIDER_NET_FIRMWARE_LEN; i++) {
+		for (i = 0; i < SPIDER_NET_FIRMWARE_SEQWORDS; i++) {
 			spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT +
 					     sequencer * 8, 0x0);
 		}
@@ -2061,7 +2071,15 @@ spider_net_setup_netdev(struct spider_net_card *card)
 	SET_NETDEV_DEV(netdev, &card->pdev->dev);
 
 	pci_set_drvdata(card->pdev, netdev);
-	spin_lock_init(&card->intmask_lock);
+
+	atomic_set(&card->tx_chain_release,0);
+	card->rxram_full_tl.data = (unsigned long) card;
+	card->rxram_full_tl.func =
+		(void (*)(unsigned long)) spider_net_handle_rxram_full;
+	init_timer(&card->tx_timer);
+	card->tx_timer.function =
+		(void (*)(unsigned long)) spider_net_cleanup_tx_ring;
+	card->tx_timer.data = (unsigned long) card;
 	netdev->irq = card->pdev->irq;
 
 	card->options.rx_csum = SPIDER_NET_RX_CSUM_DEFAULT;

drivers/net/spider_net.h

@@ -33,25 +33,32 @@ extern struct ethtool_ops spider_net_ethtool_ops;
 
 extern char spider_net_driver_name[];
 
-#define SPIDER_NET_MAX_MTU			2308
+#define SPIDER_NET_MAX_FRAME			2312
+#define SPIDER_NET_MAX_MTU			2294
 #define SPIDER_NET_MIN_MTU			64
 
 #define SPIDER_NET_RXBUF_ALIGN			128
 
-#define SPIDER_NET_RX_DESCRIPTORS_DEFAULT	64
+#define SPIDER_NET_RX_DESCRIPTORS_DEFAULT	256
 #define SPIDER_NET_RX_DESCRIPTORS_MIN		16
-#define SPIDER_NET_RX_DESCRIPTORS_MAX		256
+#define SPIDER_NET_RX_DESCRIPTORS_MAX		512
 
-#define SPIDER_NET_TX_DESCRIPTORS_DEFAULT	64
+#define SPIDER_NET_TX_DESCRIPTORS_DEFAULT	256
 #define SPIDER_NET_TX_DESCRIPTORS_MIN		16
-#define SPIDER_NET_TX_DESCRIPTORS_MAX		256
+#define SPIDER_NET_TX_DESCRIPTORS_MAX		512
+
+#define SPIDER_NET_TX_TIMER			20
 
 #define SPIDER_NET_RX_CSUM_DEFAULT		1
 
-#define SPIDER_NET_WATCHDOG_TIMEOUT 5*HZ
-#define SPIDER_NET_NAPI_WEIGHT 64
+#define SPIDER_NET_WATCHDOG_TIMEOUT		50*HZ
+#define SPIDER_NET_NAPI_WEIGHT			64
 
-#define SPIDER_NET_FIRMWARE_LEN		1024
+#define SPIDER_NET_FIRMWARE_SEQS	6
+#define SPIDER_NET_FIRMWARE_SEQWORDS	1024
+#define SPIDER_NET_FIRMWARE_LEN		(SPIDER_NET_FIRMWARE_SEQS * \
+					 SPIDER_NET_FIRMWARE_SEQWORDS * \
+					 sizeof(u32))
 #define SPIDER_NET_FIRMWARE_NAME	"spider_fw.bin"
 
 /** spider_net SMMIO registers */
@@ -142,14 +149,12 @@ extern char spider_net_driver_name[];
 /** SCONFIG registers */
 #define SPIDER_NET_SCONFIG_IOACTE	0x00002810
 
-/** hardcoded register values */
-#define SPIDER_NET_INT0_MASK_VALUE	0x3f7fe3ff
-#define SPIDER_NET_INT1_MASK_VALUE	0xffffffff
+/** interrupt mask registers */
+#define SPIDER_NET_INT0_MASK_VALUE	0x3f7fe2c7
+#define SPIDER_NET_INT1_MASK_VALUE	0xffff7ff7
 /* no MAC aborts -> auto retransmission */
-#define SPIDER_NET_INT2_MASK_VALUE	0xfffffff1
+#define SPIDER_NET_INT2_MASK_VALUE	0xffef7ff1
 
-/* clear counter when interrupt sources are cleared
-#define SPIDER_NET_FRAMENUM_VALUE	0x0001f001 */
 /* we rely on flagged descriptor interrupts */
 #define SPIDER_NET_FRAMENUM_VALUE	0x00000000
 /* set this first, then the FRAMENUM_VALUE */
@@ -168,7 +173,7 @@ extern char spider_net_driver_name[];
 #if 0
 #define SPIDER_NET_WOL_VALUE		0x00000000
 #endif
-#define SPIDER_NET_IPSECINIT_VALUE	0x00f000f8
+#define SPIDER_NET_IPSECINIT_VALUE	0x6f716f71
 
 /* pause frames: automatic, no upper retransmission count */
 /* outside loopback mode: ETOMOD signal dont matter, not connected */
@@ -318,6 +323,10 @@ enum spider_net_int2_status {
 #define SPIDER_NET_RXINT	( (1 << SPIDER_NET_GDAFDCINT) | \
 				  (1 << SPIDER_NET_GRMFLLINT) )
 
+#define SPIDER_NET_ERRINT	( 0xffffffff & \
+				  (~SPIDER_NET_TXINT) & \
+				  (~SPIDER_NET_RXINT) )
+
 #define SPIDER_NET_GPREXEC		0x80000000
 #define SPIDER_NET_GPRDAT_MASK		0x0000ffff
 
@@ -358,9 +367,6 @@ enum spider_net_int2_status {
 /* descr ready, descr is in middle of chain, get interrupt on completion */
 #define SPIDER_NET_DMAC_RX_CARDOWNED	0xa0800000
 
-/* multicast is no problem */
-#define SPIDER_NET_DATA_ERROR_MASK	0xffffbfff
-
 enum spider_net_descr_status {
 	SPIDER_NET_DESCR_COMPLETE		= 0x00, /* used in rx and tx */
 	SPIDER_NET_DESCR_RESPONSE_ERROR		= 0x01, /* used in rx and tx */
@@ -373,9 +379,9 @@ enum spider_net_descr_status {
 
 struct spider_net_descr {
 	/* as defined by the hardware */
-	dma_addr_t buf_addr;
+	u32 buf_addr;
 	u32 buf_size;
-	dma_addr_t next_descr_addr;
+	u32 next_descr_addr;
 	u32 dmac_cmd_status;
 	u32 result_size;
 	u32 valid_size;	/* all zeroes for tx */
@@ -384,7 +390,7 @@ struct spider_net_descr {
 
 	/* used in the driver */
 	struct sk_buff *skb;
-	dma_addr_t bus_addr;
+	u32 bus_addr;
 	struct spider_net_descr *next;
 	struct spider_net_descr *prev;
 } __attribute__((aligned(32)));
@@ -396,21 +402,21 @@ struct spider_net_descr_chain {
 };
 
 /* descriptor data_status bits */
-#define SPIDER_NET_RXIPCHK		29
-#define SPIDER_NET_TCPUDPIPCHK		28
-#define SPIDER_NET_DATA_STATUS_CHK_MASK	(1 << SPIDER_NET_RXIPCHK | \
-					 1 << SPIDER_NET_TCPUDPIPCHK)
-
+#define SPIDER_NET_RX_IPCHK		29
+#define SPIDER_NET_RX_TCPCHK		28
 #define SPIDER_NET_VLAN_PACKET		21
+#define SPIDER_NET_DATA_STATUS_CKSUM_MASK ( (1 << SPIDER_NET_RX_IPCHK) | \
+					  (1 << SPIDER_NET_RX_TCPCHK) )
 
 /* descriptor data_error bits */
-#define SPIDER_NET_RXIPCHKERR		27
-#define SPIDER_NET_RXTCPCHKERR		26
-#define SPIDER_NET_DATA_ERROR_CHK_MASK	(1 << SPIDER_NET_RXIPCHKERR | \
-					 1 << SPIDER_NET_RXTCPCHKERR)
+#define SPIDER_NET_RX_IPCHKERR		27
+#define SPIDER_NET_RX_RXTCPCHKERR	28
+
+#define SPIDER_NET_DATA_ERR_CKSUM_MASK	(1 << SPIDER_NET_RX_IPCHKERR)
 
-/* the cases we don't pass the packet to the stack */
-#define SPIDER_NET_DESTROY_RX_FLAGS	0x70138000
+/* the cases we don't pass the packet to the stack.
+ * 701b8000 would be correct, but every packets gets that flag */
+#define SPIDER_NET_DESTROY_RX_FLAGS	0x700b8000
 
 #define SPIDER_NET_DESCR_SIZE		32
 
@@ -445,13 +451,16 @@ struct spider_net_card {
 
 	struct spider_net_descr_chain tx_chain;
 	struct spider_net_descr_chain rx_chain;
-	spinlock_t chain_lock;
+	atomic_t rx_chain_refill;
+	atomic_t tx_chain_release;
 
 	struct net_device_stats netdev_stats;
 
 	struct spider_net_options options;
 
 	spinlock_t intmask_lock;
+	struct tasklet_struct rxram_full_tl;
+	struct timer_list tx_timer;
 
 	struct work_struct tx_timeout_task;
 	atomic_t tx_timeout_task_counter;

drivers/net/spider_net_ethtool.c

@@ -113,6 +113,23 @@ spider_net_ethtool_set_rx_csum(struct net_device *netdev, u32 n)
 	return 0;
 }
 
+static uint32_t
+spider_net_ethtool_get_tx_csum(struct net_device *netdev)
+{
+        return (netdev->features & NETIF_F_HW_CSUM) != 0;
+}
+
+static int
+spider_net_ethtool_set_tx_csum(struct net_device *netdev, uint32_t data)
+{
+        if (data)
+                netdev->features |= NETIF_F_HW_CSUM;
+        else
+                netdev->features &= ~NETIF_F_HW_CSUM;
+
+        return 0;
+}
+
 struct ethtool_ops spider_net_ethtool_ops = {
 	.get_settings		= spider_net_ethtool_get_settings,
 	.get_drvinfo		= spider_net_ethtool_get_drvinfo,
@@ -122,5 +139,7 @@ struct ethtool_ops spider_net_ethtool_ops = {
 	.nway_reset		= spider_net_ethtool_nway_reset,
 	.get_rx_csum		= spider_net_ethtool_get_rx_csum,
 	.set_rx_csum		= spider_net_ethtool_set_rx_csum,
+	.get_tx_csum		= spider_net_ethtool_get_tx_csum,
+	.set_tx_csum		= spider_net_ethtool_set_tx_csum,
 };
 

drivers/net/wireless/airo.c

@@ -5668,13 +5668,13 @@ static int airo_set_freq(struct net_device *dev,
 		int channel = fwrq->m;
 		/* We should do a better check than that,
 		 * based on the card capability !!! */
-		if((channel < 1) || (channel > 16)) {
+		if((channel < 1) || (channel > 14)) {
 			printk(KERN_DEBUG "%s: New channel value of %d is invalid!\n", dev->name, fwrq->m);
 			rc = -EINVAL;
 		} else {
 			readConfigRid(local, 1);
 			/* Yes ! We can set it !!! */
-			local->config.channelSet = (u16)(channel - 1);
+			local->config.channelSet = (u16) channel;
 			set_bit (FLAG_COMMIT, &local->flags);
 		}
 	}
@@ -5692,6 +5692,7 @@ static int airo_get_freq(struct net_device *dev,
 {
 	struct airo_info *local = dev->priv;
 	StatusRid status_rid;		/* Card status info */
+	int ch;
 
 	readConfigRid(local, 1);
 	if ((local->config.opmode & 0xFF) == MODE_STA_ESS)
@@ -5699,16 +5700,14 @@ static int airo_get_freq(struct net_device *dev,
 	else
 		readStatusRid(local, &status_rid, 1);
 
-#ifdef WEXT_USECHANNELS
-	fwrq->m = ((int)status_rid.channel) + 1;
-	fwrq->e = 0;
-#else
-	{
-		int f = (int)status_rid.channel;
-		fwrq->m = frequency_list[f] * 100000;
+	ch = (int)status_rid.channel;
+	if((ch > 0) && (ch < 15)) {
+		fwrq->m = frequency_list[ch - 1] * 100000;
 		fwrq->e = 1;
+	} else {
+		fwrq->m = ch;
+		fwrq->e = 0;
 	}
-#endif
 
 	return 0;
 }
@@ -5783,7 +5782,7 @@ static int airo_get_essid(struct net_device *dev,
 	/* If none, we may want to get the one that was set */
 
 	/* Push it out ! */
-	dwrq->length = status_rid.SSIDlen + 1;
+	dwrq->length = status_rid.SSIDlen;
 	dwrq->flags = 1; /* active */
 
 	return 0;

drivers/net/wireless/atmel.c

@@ -1718,11 +1718,11 @@ static int atmel_get_essid(struct net_device *dev,
 	if (priv->new_SSID_size != 0) {
 		memcpy(extra, priv->new_SSID, priv->new_SSID_size);
 		extra[priv->new_SSID_size] = '\0';
-		dwrq->length = priv->new_SSID_size + 1;
+		dwrq->length = priv->new_SSID_size;
 	} else {
 		memcpy(extra, priv->SSID, priv->SSID_size);
 		extra[priv->SSID_size] = '\0';
-		dwrq->length = priv->SSID_size + 1;
+		dwrq->length = priv->SSID_size;
 	}
 
 	dwrq->flags = !priv->connect_to_any_BSS; /* active */

drivers/net/wireless/hostap/Kconfig

@@ -26,11 +26,25 @@ config HOSTAP_FIRMWARE
 	depends on HOSTAP
 	---help---
 	Configure Host AP driver to include support for firmware image
-	download. Current version supports only downloading to volatile, i.e.,
-	RAM memory. Flash upgrade is not yet supported.
+	download. This option by itself only enables downloading to the
+	volatile memory, i.e. the card RAM. This option is required to
+	support cards that don't have firmware in flash, such as D-Link
+	DWL-520 rev E and D-Link DWL-650 rev P.
 
-	Firmware image downloading needs user space tool, prism2_srec. It is
-	available from http://hostap.epitest.fi/.
+	Firmware image downloading needs a user space tool, prism2_srec.
+	It is available from http://hostap.epitest.fi/.
+
+config HOSTAP_FIRMWARE_NVRAM
+	bool "Support for non-volatile firmware download"
+	depends on HOSTAP_FIRMWARE
+	---help---
+	Allow Host AP driver to write firmware images to the non-volatile
+	card memory, i.e. flash memory that survives power cycling.
+	Enable this option if you want to be able to change card firmware
+	permanently.
+
+	Firmware image downloading needs a user space tool, prism2_srec.
+	It is available from http://hostap.epitest.fi/.
 
 config HOSTAP_PLX
 	tristate "Host AP driver for Prism2/2.5/3 in PLX9052 PCI adaptors"

drivers/net/wireless/hostap/Makefile

@@ -1,4 +1,5 @@
-hostap-y := hostap_main.o
+hostap-y := hostap_80211_rx.o hostap_80211_tx.o hostap_ap.o hostap_info.o \
+            hostap_ioctl.o hostap_main.o hostap_proc.o 
 obj-$(CONFIG_HOSTAP) += hostap.o
 
 obj-$(CONFIG_HOSTAP_CS) += hostap_cs.o

drivers/net/wireless/hostap/hostap.h

@@ -1,6 +1,15 @@
 #ifndef HOSTAP_H
 #define HOSTAP_H
 
+#include <linux/ethtool.h>
+
+#include "hostap_wlan.h"
+#include "hostap_ap.h"
+
+static const long freq_list[] = { 2412, 2417, 2422, 2427, 2432, 2437, 2442,
+				  2447, 2452, 2457, 2462, 2467, 2472, 2484 };
+#define FREQ_COUNT (sizeof(freq_list) / sizeof(freq_list[0]))
+
 /* hostap.c */
 
 extern struct proc_dir_entry *hostap_proc;
@@ -40,6 +49,26 @@ int prism2_update_comms_qual(struct net_device *dev);
 int prism2_sta_send_mgmt(local_info_t *local, u8 *dst, u16 stype,
 			 u8 *body, size_t bodylen);
 int prism2_sta_deauth(local_info_t *local, u16 reason);
+int prism2_wds_add(local_info_t *local, u8 *remote_addr,
+		   int rtnl_locked);
+int prism2_wds_del(local_info_t *local, u8 *remote_addr,
+		   int rtnl_locked, int do_not_remove);
+
+
+/* hostap_ap.c */
+
+int ap_control_add_mac(struct mac_restrictions *mac_restrictions, u8 *mac);
+int ap_control_del_mac(struct mac_restrictions *mac_restrictions, u8 *mac);
+void ap_control_flush_macs(struct mac_restrictions *mac_restrictions);
+int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, u8 *mac);
+void ap_control_kickall(struct ap_data *ap);
+void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent,
+			 struct ieee80211_crypt_data ***crypt);
+int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[],
+			   struct iw_quality qual[], int buf_size,
+			   int aplist);
+int prism2_ap_translate_scan(struct net_device *dev, char *buffer);
+int prism2_hostapd(struct ap_data *ap, struct prism2_hostapd_param *param);
 
 
 /* hostap_proc.c */
@@ -54,4 +83,12 @@ void hostap_info_init(local_info_t *local);
 void hostap_info_process(local_info_t *local, struct sk_buff *skb);
 
 
+/* hostap_ioctl.c */
+
+extern const struct iw_handler_def hostap_iw_handler_def;
+extern struct ethtool_ops prism2_ethtool_ops;
+
+int hostap_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
+
+
 #endif /* HOSTAP_H */

drivers/net/wireless/hostap/hostap_80211.h

@@ -1,6 +1,9 @@
 #ifndef HOSTAP_80211_H
 #define HOSTAP_80211_H
 
+#include <linux/types.h>
+#include <net/ieee80211_crypt.h>
+
 struct hostap_ieee80211_mgmt {
 	u16 frame_control;
 	u16 duration;

drivers/net/wireless/hostap/hostap_80211_rx.c

@@ -1,7 +1,18 @@
 #include <linux/etherdevice.h>
+#include <net/ieee80211_crypt.h>
 
 #include "hostap_80211.h"
 #include "hostap.h"
+#include "hostap_ap.h"
+
+/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */
+/* Ethernet-II snap header (RFC1042 for most EtherTypes) */
+static unsigned char rfc1042_header[] =
+{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 };
+/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */
+static unsigned char bridge_tunnel_header[] =
+{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 };
+/* No encapsulation header if EtherType < 0x600 (=length) */
 
 void hostap_dump_rx_80211(const char *name, struct sk_buff *skb,
 			  struct hostap_80211_rx_status *rx_stats)

drivers/net/wireless/hostap/hostap_80211_tx.c

@@ -1,3 +1,18 @@
+#include "hostap_80211.h"
+#include "hostap_common.h"
+#include "hostap_wlan.h"
+#include "hostap.h"
+#include "hostap_ap.h"
+
+/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */
+/* Ethernet-II snap header (RFC1042 for most EtherTypes) */
+static unsigned char rfc1042_header[] =
+{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 };
+/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */
+static unsigned char bridge_tunnel_header[] =
+{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 };
+/* No encapsulation header if EtherType < 0x600 (=length) */
+
 void hostap_dump_tx_80211(const char *name, struct sk_buff *skb)
 {
 	struct ieee80211_hdr_4addr *hdr;

drivers/net/wireless/hostap/hostap_ap.c

@@ -16,6 +16,14 @@
  *   (8802.11: 5.5)
  */
 
+#include <linux/proc_fs.h>
+#include <linux/delay.h>
+#include <linux/random.h>
+
+#include "hostap_wlan.h"
+#include "hostap.h"
+#include "hostap_ap.h"
+
 static int other_ap_policy[MAX_PARM_DEVICES] = { AP_OTHER_AP_SKIP_ALL,
 						 DEF_INTS };
 module_param_array(other_ap_policy, int, NULL, 0444);
@@ -360,8 +368,7 @@ static int ap_control_proc_read(char *page, char **start, off_t off,
 }
 
 
-static int ap_control_add_mac(struct mac_restrictions *mac_restrictions,
-			      u8 *mac)
+int ap_control_add_mac(struct mac_restrictions *mac_restrictions, u8 *mac)
 {
 	struct mac_entry *entry;
 
@@ -380,8 +387,7 @@ static int ap_control_add_mac(struct mac_restrictions *mac_restrictions,
 }
 
 
-static int ap_control_del_mac(struct mac_restrictions *mac_restrictions,
-			      u8 *mac)
+int ap_control_del_mac(struct mac_restrictions *mac_restrictions, u8 *mac)
 {
 	struct list_head *ptr;
 	struct mac_entry *entry;
@@ -433,7 +439,7 @@ static int ap_control_mac_deny(struct mac_restrictions *mac_restrictions,
 }
 
 
-static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions)
+void ap_control_flush_macs(struct mac_restrictions *mac_restrictions)
 {
 	struct list_head *ptr, *n;
 	struct mac_entry *entry;
@@ -454,8 +460,7 @@ static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions)
 }
 
 
-static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev,
-			       u8 *mac)
+int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, u8 *mac)
 {
 	struct sta_info *sta;
 	u16 resp;
@@ -486,7 +491,7 @@ static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev,
 #endif /* PRISM2_NO_KERNEL_IEEE80211_MGMT */
 
 
-static void ap_control_kickall(struct ap_data *ap)
+void ap_control_kickall(struct ap_data *ap)
 {
 	struct list_head *ptr, *n;
 	struct sta_info *sta;
@@ -2321,9 +2326,9 @@ static void schedule_packet_send(local_info_t *local, struct sta_info *sta)
 }
 
 
-static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[],
-				  struct iw_quality qual[], int buf_size,
-				  int aplist)
+int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[],
+			   struct iw_quality qual[], int buf_size,
+			   int aplist)
 {
 	struct ap_data *ap = local->ap;
 	struct list_head *ptr;
@@ -2363,7 +2368,7 @@ static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[],
 
 /* Translate our list of Access Points & Stations to a card independant
  * format that the Wireless Tools will understand - Jean II */
-static int prism2_ap_translate_scan(struct net_device *dev, char *buffer)
+int prism2_ap_translate_scan(struct net_device *dev, char *buffer)
 {
 	struct hostap_interface *iface;
 	local_info_t *local;
@@ -2608,8 +2613,7 @@ static int prism2_hostapd_sta_clear_stats(struct ap_data *ap,
 }
 
 
-static int prism2_hostapd(struct ap_data *ap,
-			  struct prism2_hostapd_param *param)
+int prism2_hostapd(struct ap_data *ap, struct prism2_hostapd_param *param)
 {
 	switch (param->cmd) {
 	case PRISM2_HOSTAPD_FLUSH:
@@ -3207,8 +3211,8 @@ void hostap_update_rates(local_info_t *local)
 }
 
 
-static void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent,
-				struct ieee80211_crypt_data ***crypt)
+void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent,
+			 struct ieee80211_crypt_data ***crypt)
 {
 	struct sta_info *sta;
 

drivers/net/wireless/hostap/hostap_ap.h

@@ -1,6 +1,8 @@
 #ifndef HOSTAP_AP_H
 #define HOSTAP_AP_H
 
+#include "hostap_80211.h"
+
 /* AP data structures for STAs */
 
 /* maximum number of frames to buffer per STA */

drivers/net/wireless/hostap/hostap_common.h

@@ -1,6 +1,9 @@
 #ifndef HOSTAP_COMMON_H
 #define HOSTAP_COMMON_H
 
+#include <linux/types.h>
+#include <linux/if_ether.h>
+
 #define BIT(x) (1 << (x))
 
 #define MAC2STR(a) (a)[0], (a)[1], (a)[2], (a)[3], (a)[4], (a)[5]

drivers/net/wireless/hostap/hostap_config.h

@@ -21,15 +21,10 @@
 #define PRISM2_DOWNLOAD_SUPPORT
 #endif
 
-#ifdef PRISM2_DOWNLOAD_SUPPORT
-/* Allow writing firmware images into flash, i.e., to non-volatile storage.
- * Before you enable this option, you should make absolutely sure that you are
- * using prism2_srec utility that comes with THIS version of the driver!
- * In addition, please note that it is possible to kill your card with
- * non-volatile download if you are using incorrect image. This feature has not
- * been fully tested, so please be careful with it. */
-/* #define PRISM2_NON_VOLATILE_DOWNLOAD */
-#endif /* PRISM2_DOWNLOAD_SUPPORT */
+/* Allow kernel configuration to enable non-volatile download support. */
+#ifdef CONFIG_HOSTAP_FIRMWARE_NVRAM
+#define PRISM2_NON_VOLATILE_DOWNLOAD
+#endif
 
 /* Save low-level I/O for debugging. This should not be enabled in normal use.
  */

drivers/net/wireless/hostap/hostap_info.c

@@ -1,5 +1,8 @@
 /* Host AP driver Info Frame processing (part of hostap.o module) */
 
+#include "hostap_wlan.h"
+#include "hostap.h"
+#include "hostap_ap.h"
 
 /* Called only as a tasklet (software IRQ) */
 static void prism2_info_commtallies16(local_info_t *local, unsigned char *buf,

drivers/net/wireless/hostap/hostap_ioctl.c

@@ -1,11 +1,13 @@
 /* ioctl() (mostly Linux Wireless Extensions) routines for Host AP driver */
 
-#ifdef in_atomic
-/* Get kernel_locked() for in_atomic() */
+#include <linux/types.h>
 #include <linux/smp_lock.h>
-#endif
 #include <linux/ethtool.h>
+#include <net/ieee80211_crypt.h>
 
+#include "hostap_wlan.h"
+#include "hostap.h"
+#include "hostap_ap.h"
 
 static struct iw_statistics *hostap_get_wireless_stats(struct net_device *dev)
 {
@@ -3910,7 +3912,7 @@ static void prism2_get_drvinfo(struct net_device *dev,
 		 local->sta_fw_ver & 0xff);
 }
 
-static struct ethtool_ops prism2_ethtool_ops = {
+struct ethtool_ops prism2_ethtool_ops = {
 	.get_drvinfo = prism2_get_drvinfo
 };
 
@@ -3985,7 +3987,7 @@ static const iw_handler prism2_private_handler[] =
 	(iw_handler) prism2_ioctl_priv_readmif,		/* 3 */
 };
 
-static const struct iw_handler_def hostap_iw_handler_def =
+const struct iw_handler_def hostap_iw_handler_def =
 {
 	.num_standard	= sizeof(prism2_handler) / sizeof(iw_handler),
 	.num_private	= sizeof(prism2_private_handler) / sizeof(iw_handler),

drivers/net/wireless/hostap/hostap_main.c

@@ -24,6 +24,7 @@
 #include <linux/kmod.h>
 #include <linux/rtnetlink.h>
 #include <linux/wireless.h>
+#include <linux/etherdevice.h>
 #include <net/iw_handler.h>
 #include <net/ieee80211.h>
 #include <net/ieee80211_crypt.h>
@@ -47,57 +48,6 @@ MODULE_VERSION(PRISM2_VERSION);
 #define PRISM2_MAX_MTU (PRISM2_MAX_FRAME_SIZE - (6 /* LLC */ + 8 /* WEP */))
 
 
-/* hostap.c */
-static int prism2_wds_add(local_info_t *local, u8 *remote_addr,
-			  int rtnl_locked);
-static int prism2_wds_del(local_info_t *local, u8 *remote_addr,
-			  int rtnl_locked, int do_not_remove);
-
-/* hostap_ap.c */
-static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[],
-				  struct iw_quality qual[], int buf_size,
-				  int aplist);
-static int prism2_ap_translate_scan(struct net_device *dev, char *buffer);
-static int prism2_hostapd(struct ap_data *ap,
-			  struct prism2_hostapd_param *param);
-static void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent,
-				struct ieee80211_crypt_data ***crypt);
-static void ap_control_kickall(struct ap_data *ap);
-#ifndef PRISM2_NO_KERNEL_IEEE80211_MGMT
-static int ap_control_add_mac(struct mac_restrictions *mac_restrictions,
-			      u8 *mac);
-static int ap_control_del_mac(struct mac_restrictions *mac_restrictions,
-			      u8 *mac);
-static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions);
-static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev,
-			       u8 *mac);
-#endif /* !PRISM2_NO_KERNEL_IEEE80211_MGMT */
-
-
-static const long freq_list[] = { 2412, 2417, 2422, 2427, 2432, 2437, 2442,
-				  2447, 2452, 2457, 2462, 2467, 2472, 2484 };
-#define FREQ_COUNT (sizeof(freq_list) / sizeof(freq_list[0]))
-
-
-/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */
-/* Ethernet-II snap header (RFC1042 for most EtherTypes) */
-static unsigned char rfc1042_header[] =
-{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 };
-/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */
-static unsigned char bridge_tunnel_header[] =
-{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 };
-/* No encapsulation header if EtherType < 0x600 (=length) */
-
-
-/* FIX: these could be compiled separately and linked together to hostap.o */
-#include "hostap_ap.c"
-#include "hostap_info.c"
-#include "hostap_ioctl.c"
-#include "hostap_proc.c"
-#include "hostap_80211_rx.c"
-#include "hostap_80211_tx.c"
-
-
 struct net_device * hostap_add_interface(struct local_info *local,
 					 int type, int rtnl_locked,
 					 const char *prefix,
@@ -196,8 +146,8 @@ static inline int prism2_wds_special_addr(u8 *addr)
 }
 
 
-static int prism2_wds_add(local_info_t *local, u8 *remote_addr,
-			  int rtnl_locked)
+int prism2_wds_add(local_info_t *local, u8 *remote_addr,
+		   int rtnl_locked)
 {
 	struct net_device *dev;
 	struct list_head *ptr;
@@ -258,8 +208,8 @@ static int prism2_wds_add(local_info_t *local, u8 *remote_addr,
 }
 
 
-static int prism2_wds_del(local_info_t *local, u8 *remote_addr,
-			  int rtnl_locked, int do_not_remove)
+int prism2_wds_del(local_info_t *local, u8 *remote_addr,
+		   int rtnl_locked, int do_not_remove)
 {
 	unsigned long flags;
 	struct list_head *ptr;

drivers/net/wireless/hostap/hostap_proc.c

@@ -1,5 +1,12 @@
 /* /proc routines for Host AP driver */
 
+#include <linux/types.h>
+#include <linux/proc_fs.h>
+#include <net/ieee80211_crypt.h>
+
+#include "hostap_wlan.h"
+#include "hostap.h"
+
 #define PROC_LIMIT (PAGE_SIZE - 80)
 
 

drivers/net/wireless/hostap/hostap_wlan.h

@@ -1,6 +1,10 @@
 #ifndef HOSTAP_WLAN_H
 #define HOSTAP_WLAN_H
 
+#include <linux/wireless.h>
+#include <linux/netdevice.h>
+#include <net/iw_handler.h>
+
 #include "hostap_config.h"
 #include "hostap_common.h"
 

drivers/net/wireless/ipw2100.c

@@ -5735,70 +5735,6 @@ static struct net_device_stats *ipw2100_stats(struct net_device *dev)
 	return &priv->ieee->stats;
 }
 
-#if WIRELESS_EXT < 18
-/* Support for wpa_supplicant before WE-18, deprecated. */
-
-/* following definitions must match definitions in driver_ipw.c */
-
-#define IPW2100_IOCTL_WPA_SUPPLICANT		SIOCIWFIRSTPRIV+30
-
-#define IPW2100_CMD_SET_WPA_PARAM		1
-#define	IPW2100_CMD_SET_WPA_IE			2
-#define IPW2100_CMD_SET_ENCRYPTION		3
-#define IPW2100_CMD_MLME			4
-
-#define IPW2100_PARAM_WPA_ENABLED		1
-#define IPW2100_PARAM_TKIP_COUNTERMEASURES	2
-#define IPW2100_PARAM_DROP_UNENCRYPTED		3
-#define IPW2100_PARAM_PRIVACY_INVOKED		4
-#define IPW2100_PARAM_AUTH_ALGS			5
-#define IPW2100_PARAM_IEEE_802_1X		6
-
-#define IPW2100_MLME_STA_DEAUTH			1
-#define IPW2100_MLME_STA_DISASSOC		2
-
-#define IPW2100_CRYPT_ERR_UNKNOWN_ALG		2
-#define IPW2100_CRYPT_ERR_UNKNOWN_ADDR		3
-#define IPW2100_CRYPT_ERR_CRYPT_INIT_FAILED	4
-#define IPW2100_CRYPT_ERR_KEY_SET_FAILED	5
-#define IPW2100_CRYPT_ERR_TX_KEY_SET_FAILED	6
-#define IPW2100_CRYPT_ERR_CARD_CONF_FAILED	7
-
-#define	IPW2100_CRYPT_ALG_NAME_LEN		16
-
-struct ipw2100_param {
-	u32 cmd;
-	u8 sta_addr[ETH_ALEN];
-	union {
-		struct {
-			u8 name;
-			u32 value;
-		} wpa_param;
-		struct {
-			u32 len;
-			u8 reserved[32];
-			u8 data[0];
-		} wpa_ie;
-		struct {
-			u32 command;
-			u32 reason_code;
-		} mlme;
-		struct {
-			u8 alg[IPW2100_CRYPT_ALG_NAME_LEN];
-			u8 set_tx;
-			u32 err;
-			u8 idx;
-			u8 seq[8];	/* sequence counter (set: RX, get: TX) */
-			u16 key_len;
-			u8 key[0];
-		} crypt;
-
-	} u;
-};
-
-/* end of driver_ipw.c code */
-#endif				/* WIRELESS_EXT < 18 */
-
 static int ipw2100_wpa_enable(struct ipw2100_priv *priv, int value)
 {
 	/* This is called when wpa_supplicant loads and closes the driver
@@ -5807,11 +5743,6 @@ static int ipw2100_wpa_enable(struct ipw2100_priv *priv, int value)
 	return 0;
 }
 
-#if WIRELESS_EXT < 18
-#define IW_AUTH_ALG_OPEN_SYSTEM			0x1
-#define IW_AUTH_ALG_SHARED_KEY			0x2
-#endif
-
 static int ipw2100_wpa_set_auth_algs(struct ipw2100_priv *priv, int value)
 {
 
@@ -5855,360 +5786,6 @@ void ipw2100_wpa_assoc_frame(struct ipw2100_priv *priv,
 	ipw2100_set_wpa_ie(priv, &frame, 0);
 }
 
-#if WIRELESS_EXT < 18
-static int ipw2100_wpa_set_param(struct net_device *dev, u8 name, u32 value)
-{
-	struct ipw2100_priv *priv = ieee80211_priv(dev);
-	struct ieee80211_crypt_data *crypt;
-	unsigned long flags;
-	int ret = 0;
-
-	switch (name) {
-	case IPW2100_PARAM_WPA_ENABLED:
-		ret = ipw2100_wpa_enable(priv, value);
-		break;
-
-	case IPW2100_PARAM_TKIP_COUNTERMEASURES:
-		crypt = priv->ieee->crypt[priv->ieee->tx_keyidx];
-		if (!crypt || !crypt->ops->set_flags || !crypt->ops->get_flags)
-			break;
-
-		flags = crypt->ops->get_flags(crypt->priv);
-
-		if (value)
-			flags |= IEEE80211_CRYPTO_TKIP_COUNTERMEASURES;
-		else
-			flags &= ~IEEE80211_CRYPTO_TKIP_COUNTERMEASURES;
-
-		crypt->ops->set_flags(flags, crypt->priv);
-
-		break;
-
-	case IPW2100_PARAM_DROP_UNENCRYPTED:{
-			/* See IW_AUTH_DROP_UNENCRYPTED handling for details */
-			struct ieee80211_security sec = {
-				.flags = SEC_ENABLED,
-				.enabled = value,
-			};
-			priv->ieee->drop_unencrypted = value;
-			/* We only change SEC_LEVEL for open mode. Others
-			 * are set by ipw_wpa_set_encryption.
-			 */
-			if (!value) {
-				sec.flags |= SEC_LEVEL;
-				sec.level = SEC_LEVEL_0;
-			} else {
-				sec.flags |= SEC_LEVEL;
-				sec.level = SEC_LEVEL_1;
-			}
-			if (priv->ieee->set_security)
-				priv->ieee->set_security(priv->ieee->dev, &sec);
-			break;
-		}
-
-	case IPW2100_PARAM_PRIVACY_INVOKED:
-		priv->ieee->privacy_invoked = value;
-		break;
-
-	case IPW2100_PARAM_AUTH_ALGS:
-		ret = ipw2100_wpa_set_auth_algs(priv, value);
-		break;
-
-	case IPW2100_PARAM_IEEE_802_1X:
-		priv->ieee->ieee802_1x = value;
-		break;
-
-	default:
-		printk(KERN_ERR DRV_NAME ": %s: Unknown WPA param: %d\n",
-		       dev->name, name);
-		ret = -EOPNOTSUPP;
-	}
-
-	return ret;
-}
-
-static int ipw2100_wpa_mlme(struct net_device *dev, int command, int reason)
-{
-
-	struct ipw2100_priv *priv = ieee80211_priv(dev);
-	int ret = 0;
-
-	switch (command) {
-	case IPW2100_MLME_STA_DEAUTH:
-		// silently ignore
-		break;
-
-	case IPW2100_MLME_STA_DISASSOC:
-		ipw2100_disassociate_bssid(priv);
-		break;
-
-	default:
-		printk(KERN_ERR DRV_NAME ": %s: Unknown MLME request: %d\n",
-		       dev->name, command);
-		ret = -EOPNOTSUPP;
-	}
-
-	return ret;
-}
-
-static int ipw2100_wpa_set_wpa_ie(struct net_device *dev,
-				  struct ipw2100_param *param, int plen)
-{
-
-	struct ipw2100_priv *priv = ieee80211_priv(dev);
-	struct ieee80211_device *ieee = priv->ieee;
-	u8 *buf;
-
-	if (!ieee->wpa_enabled)
-		return -EOPNOTSUPP;
-
-	if (param->u.wpa_ie.len > MAX_WPA_IE_LEN ||
-	    (param->u.wpa_ie.len && param->u.wpa_ie.data == NULL))
-		return -EINVAL;
-
-	if (param->u.wpa_ie.len) {
-		buf = kmalloc(param->u.wpa_ie.len, GFP_KERNEL);
-		if (buf == NULL)
-			return -ENOMEM;
-
-		memcpy(buf, param->u.wpa_ie.data, param->u.wpa_ie.len);
-
-		kfree(ieee->wpa_ie);
-		ieee->wpa_ie = buf;
-		ieee->wpa_ie_len = param->u.wpa_ie.len;
-
-	} else {
-		kfree(ieee->wpa_ie);
-		ieee->wpa_ie = NULL;
-		ieee->wpa_ie_len = 0;
-	}
-
-	ipw2100_wpa_assoc_frame(priv, ieee->wpa_ie, ieee->wpa_ie_len);
-
-	return 0;
-}
-
-/* implementation borrowed from hostap driver */
-
-static int ipw2100_wpa_set_encryption(struct net_device *dev,
-				      struct ipw2100_param *param,
-				      int param_len)
-{
-	int ret = 0;
-	struct ipw2100_priv *priv = ieee80211_priv(dev);
-	struct ieee80211_device *ieee = priv->ieee;
-	struct ieee80211_crypto_ops *ops;
-	struct ieee80211_crypt_data **crypt;
-
-	struct ieee80211_security sec = {
-		.flags = 0,
-	};
-
-	param->u.crypt.err = 0;
-	param->u.crypt.alg[IPW2100_CRYPT_ALG_NAME_LEN - 1] = '\0';
-
-	if (param_len !=
-	    (int)((char *)param->u.crypt.key - (char *)param) +
-	    param->u.crypt.key_len) {
-		IPW_DEBUG_INFO("Len mismatch %d, %d\n", param_len,
-			       param->u.crypt.key_len);
-		return -EINVAL;
-	}
-	if (param->sta_addr[0] == 0xff && param->sta_addr[1] == 0xff &&
-	    param->sta_addr[2] == 0xff && param->sta_addr[3] == 0xff &&
-	    param->sta_addr[4] == 0xff && param->sta_addr[5] == 0xff) {
-		if (param->u.crypt.idx >= WEP_KEYS)
-			return -EINVAL;
-		crypt = &ieee->crypt[param->u.crypt.idx];
-	} else {
-		return -EINVAL;
-	}
-
-	sec.flags |= SEC_ENABLED | SEC_ENCRYPT;
-	if (strcmp(param->u.crypt.alg, "none") == 0) {
-		if (crypt) {
-			sec.enabled = 0;
-			sec.encrypt = 0;
-			sec.level = SEC_LEVEL_0;
-			sec.flags |= SEC_LEVEL;
-			ieee80211_crypt_delayed_deinit(ieee, crypt);
-		}
-		goto done;
-	}
-	sec.enabled = 1;
-	sec.encrypt = 1;
-
-	ops = ieee80211_get_crypto_ops(param->u.crypt.alg);
-	if (ops == NULL && strcmp(param->u.crypt.alg, "WEP") == 0) {
-		request_module("ieee80211_crypt_wep");
-		ops = ieee80211_get_crypto_ops(param->u.crypt.alg);
-	} else if (ops == NULL && strcmp(param->u.crypt.alg, "TKIP") == 0) {
-		request_module("ieee80211_crypt_tkip");
-		ops = ieee80211_get_crypto_ops(param->u.crypt.alg);
-	} else if (ops == NULL && strcmp(param->u.crypt.alg, "CCMP") == 0) {
-		request_module("ieee80211_crypt_ccmp");
-		ops = ieee80211_get_crypto_ops(param->u.crypt.alg);
-	}
-	if (ops == NULL) {
-		IPW_DEBUG_INFO("%s: unknown crypto alg '%s'\n",
-			       dev->name, param->u.crypt.alg);
-		param->u.crypt.err = IPW2100_CRYPT_ERR_UNKNOWN_ALG;
-		ret = -EINVAL;
-		goto done;
-	}
-
-	if (*crypt == NULL || (*crypt)->ops != ops) {
-		struct ieee80211_crypt_data *new_crypt;
-
-		ieee80211_crypt_delayed_deinit(ieee, crypt);
-
-		new_crypt = kzalloc(sizeof(struct ieee80211_crypt_data), GFP_KERNEL);
-		if (new_crypt == NULL) {
-			ret = -ENOMEM;
-			goto done;
-		}
-		new_crypt->ops = ops;
-		if (new_crypt->ops && try_module_get(new_crypt->ops->owner))
-			new_crypt->priv =
-			    new_crypt->ops->init(param->u.crypt.idx);
-
-		if (new_crypt->priv == NULL) {
-			kfree(new_crypt);
-			param->u.crypt.err =
-			    IPW2100_CRYPT_ERR_CRYPT_INIT_FAILED;
-			ret = -EINVAL;
-			goto done;
-		}
-
-		*crypt = new_crypt;
-	}
-
-	if (param->u.crypt.key_len > 0 && (*crypt)->ops->set_key &&
-	    (*crypt)->ops->set_key(param->u.crypt.key,
-				   param->u.crypt.key_len, param->u.crypt.seq,
-				   (*crypt)->priv) < 0) {
-		IPW_DEBUG_INFO("%s: key setting failed\n", dev->name);
-		param->u.crypt.err = IPW2100_CRYPT_ERR_KEY_SET_FAILED;
-		ret = -EINVAL;
-		goto done;
-	}
-
-	if (param->u.crypt.set_tx) {
-		ieee->tx_keyidx = param->u.crypt.idx;
-		sec.active_key = param->u.crypt.idx;
-		sec.flags |= SEC_ACTIVE_KEY;
-	}
-
-	if (ops->name != NULL) {
-
-		if (strcmp(ops->name, "WEP") == 0) {
-			memcpy(sec.keys[param->u.crypt.idx],
-			       param->u.crypt.key, param->u.crypt.key_len);
-			sec.key_sizes[param->u.crypt.idx] =
-			    param->u.crypt.key_len;
-			sec.flags |= (1 << param->u.crypt.idx);
-			sec.flags |= SEC_LEVEL;
-			sec.level = SEC_LEVEL_1;
-		} else if (strcmp(ops->name, "TKIP") == 0) {
-			sec.flags |= SEC_LEVEL;
-			sec.level = SEC_LEVEL_2;
-		} else if (strcmp(ops->name, "CCMP") == 0) {
-			sec.flags |= SEC_LEVEL;
-			sec.level = SEC_LEVEL_3;
-		}
-	}
-      done:
-	if (ieee->set_security)
-		ieee->set_security(ieee->dev, &sec);
-
-	/* Do not reset port if card is in Managed mode since resetting will
-	 * generate new IEEE 802.11 authentication which may end up in looping
-	 * with IEEE 802.1X.  If your hardware requires a reset after WEP
-	 * configuration (for example... Prism2), implement the reset_port in
-	 * the callbacks structures used to initialize the 802.11 stack. */
-	if (ieee->reset_on_keychange &&
-	    ieee->iw_mode != IW_MODE_INFRA &&
-	    ieee->reset_port && ieee->reset_port(dev)) {
-		IPW_DEBUG_INFO("%s: reset_port failed\n", dev->name);
-		param->u.crypt.err = IPW2100_CRYPT_ERR_CARD_CONF_FAILED;
-		return -EINVAL;
-	}
-
-	return ret;
-}
-
-static int ipw2100_wpa_supplicant(struct net_device *dev, struct iw_point *p)
-{
-
-	struct ipw2100_param *param;
-	int ret = 0;
-
-	IPW_DEBUG_IOCTL("wpa_supplicant: len=%d\n", p->length);
-
-	if (p->length < sizeof(struct ipw2100_param) || !p->pointer)
-		return -EINVAL;
-
-	param = (struct ipw2100_param *)kmalloc(p->length, GFP_KERNEL);
-	if (param == NULL)
-		return -ENOMEM;
-
-	if (copy_from_user(param, p->pointer, p->length)) {
-		kfree(param);
-		return -EFAULT;
-	}
-
-	switch (param->cmd) {
-
-	case IPW2100_CMD_SET_WPA_PARAM:
-		ret = ipw2100_wpa_set_param(dev, param->u.wpa_param.name,
-					    param->u.wpa_param.value);
-		break;
-
-	case IPW2100_CMD_SET_WPA_IE:
-		ret = ipw2100_wpa_set_wpa_ie(dev, param, p->length);
-		break;
-
-	case IPW2100_CMD_SET_ENCRYPTION:
-		ret = ipw2100_wpa_set_encryption(dev, param, p->length);
-		break;
-
-	case IPW2100_CMD_MLME:
-		ret = ipw2100_wpa_mlme(dev, param->u.mlme.command,
-				       param->u.mlme.reason_code);
-		break;
-
-	default:
-		printk(KERN_ERR DRV_NAME
-		       ": %s: Unknown WPA supplicant request: %d\n", dev->name,
-		       param->cmd);
-		ret = -EOPNOTSUPP;
-
-	}
-
-	if (ret == 0 && copy_to_user(p->pointer, param, p->length))
-		ret = -EFAULT;
-
-	kfree(param);
-	return ret;
-}
-
-static int ipw2100_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
-{
-	struct iwreq *wrq = (struct iwreq *)rq;
-	int ret = -1;
-	switch (cmd) {
-	case IPW2100_IOCTL_WPA_SUPPLICANT:
-		ret = ipw2100_wpa_supplicant(dev, &wrq->u.data);
-		return ret;
-
-	default:
-		return -EOPNOTSUPP;
-	}
-
-	return -EOPNOTSUPP;
-}
-#endif				/* WIRELESS_EXT < 18 */
-
 static void ipw_ethtool_get_drvinfo(struct net_device *dev,
 				    struct ethtool_drvinfo *info)
 {
@@ -6337,9 +5914,6 @@ static struct net_device *ipw2100_alloc_device(struct pci_dev *pci_dev,
 	dev->open = ipw2100_open;
 	dev->stop = ipw2100_close;
 	dev->init = ipw2100_net_init;
-#if WIRELESS_EXT < 18
-	dev->do_ioctl = ipw2100_ioctl;
-#endif
 	dev->get_stats = ipw2100_stats;
 	dev->ethtool_ops = &ipw2100_ethtool_ops;
 	dev->tx_timeout = ipw2100_tx_timeout;
@@ -7855,7 +7429,6 @@ static int ipw2100_wx_get_power(struct net_device *dev,
 	return 0;
 }
 
-#if WIRELESS_EXT > 17
 /*
  * WE-18 WPA support
  */
@@ -8117,7 +7690,6 @@ static int ipw2100_wx_set_mlme(struct net_device *dev,
 	}
 	return 0;
 }
-#endif				/* WIRELESS_EXT > 17 */
 
 /*
  *
@@ -8350,11 +7922,7 @@ static iw_handler ipw2100_wx_handlers[] = {
 	NULL,			/* SIOCWIWTHRSPY */
 	ipw2100_wx_set_wap,	/* SIOCSIWAP */
 	ipw2100_wx_get_wap,	/* SIOCGIWAP */
-#if WIRELESS_EXT > 17
 	ipw2100_wx_set_mlme,	/* SIOCSIWMLME */
-#else
-	NULL,			/* -- hole -- */
-#endif
 	NULL,			/* SIOCGIWAPLIST -- deprecated */
 	ipw2100_wx_set_scan,	/* SIOCSIWSCAN */
 	ipw2100_wx_get_scan,	/* SIOCGIWSCAN */
@@ -8378,7 +7946,6 @@ static iw_handler ipw2100_wx_handlers[] = {
 	ipw2100_wx_get_encode,	/* SIOCGIWENCODE */
 	ipw2100_wx_set_power,	/* SIOCSIWPOWER */
 	ipw2100_wx_get_power,	/* SIOCGIWPOWER */
-#if WIRELESS_EXT > 17
 	NULL,			/* -- hole -- */
 	NULL,			/* -- hole -- */
 	ipw2100_wx_set_genie,	/* SIOCSIWGENIE */
@@ -8388,7 +7955,6 @@ static iw_handler ipw2100_wx_handlers[] = {
 	ipw2100_wx_set_encodeext,	/* SIOCSIWENCODEEXT */
 	ipw2100_wx_get_encodeext,	/* SIOCGIWENCODEEXT */
 	NULL,			/* SIOCSIWPMKSA */
-#endif
 };
 
 #define IPW2100_PRIV_SET_MONITOR	SIOCIWFIRSTPRIV

drivers/net/wireless/ipw2200.c

@@ -8936,14 +8936,12 @@ static int ipw_request_direct_scan(struct ipw_priv *priv, char *essid,
 	IPW_DEBUG_HC("starting request direct scan!\n");
 
 	if (priv->status & (STATUS_SCANNING | STATUS_SCAN_ABORTING)) {
-		err = wait_event_interruptible(priv->wait_state,
-					       !(priv->
-						 status & (STATUS_SCANNING |
-							   STATUS_SCAN_ABORTING)));
-		if (err) {
-			IPW_DEBUG_HC("aborting direct scan");
-			goto done;
-		}
+		/* We should not sleep here; otherwise we will block most
+		 * of the system (for instance, we hold rtnl_lock when we
+		 * get here).
+		 */
+		err = -EAGAIN;
+		goto done;
 	}
 	memset(&scan, 0, sizeof(scan));
 

drivers/net/wireless/prism54/isl_ioctl.c

@@ -748,7 +748,7 @@ prism54_get_essid(struct net_device *ndev, struct iw_request_info *info,
 	if (essid->length) {
 		dwrq->flags = 1;	/* set ESSID to ON for Wireless Extensions */
 		/* if it is to big, trunk it */
-		dwrq->length = min(IW_ESSID_MAX_SIZE, essid->length + 1);
+		dwrq->length = min(IW_ESSID_MAX_SIZE, essid->length);
 	} else {
 		dwrq->flags = 0;
 		dwrq->length = 0;

drivers/net/wireless/prism54/islpci_eth.c

@@ -177,7 +177,7 @@ islpci_eth_transmit(struct sk_buff *skb, struct net_device *ndev)
 #endif
 
 			newskb->dev = skb->dev;
-			dev_kfree_skb(skb);
+			dev_kfree_skb_irq(skb);
 			skb = newskb;
 		}
 	}

drivers/net/wireless/ray_cs.c

@@ -1256,7 +1256,7 @@ static int ray_get_essid(struct net_device *dev,
 	extra[IW_ESSID_MAX_SIZE] = '\0';
 
 	/* Push it out ! */
-	dwrq->length = strlen(extra) + 1;
+	dwrq->length = strlen(extra);
 	dwrq->flags = 1; /* active */
 
 	return 0;

drivers/net/wireless/wavelan_cs.c

@@ -2280,7 +2280,7 @@ static int wavelan_get_essid(struct net_device *dev,
 	extra[IW_ESSID_MAX_SIZE] = '\0';
 
 	/* Set the length */
-	wrqu->data.length = strlen(extra) + 1;
+	wrqu->data.length = strlen(extra);
 
 	return 0;
 }

include/net/ieee80211_crypt.h

@@ -25,6 +25,7 @@
 
 #include <linux/types.h>
 #include <linux/list.h>
+#include <net/ieee80211.h>
 #include <asm/atomic.h>
 
 enum {

include/net/iw_handler.h

@@ -327,7 +327,7 @@ struct iw_handler_def
 	__u16			num_private_args;
 
 	/* Array of handlers for standard ioctls
-	 * We will call dev->wireless_handlers->standard[ioctl - SIOCSIWNAME]
+	 * We will call dev->wireless_handlers->standard[ioctl - SIOCSIWCOMMIT]
 	 */
 	const iw_handler *	standard;