Merge branch 'master' of /home/davem/src/GIT/linux-2.6/

Conflicts:
	drivers/firmware/iscsi_ibft.c

23 files changed, 725 insertions, 887 deletions

diff --git a/drivers/staging/go7007/s2250-board.c b/drivers/staging/go7007/s2250-board.c
index 8cf7f2750b3..c324f6ea002 100644
--- a/drivers/staging/go7007/s2250-board.c
+++ b/drivers/staging/go7007/s2250-board.c
@@ -159,7 +159,7 @@ static int write_reg(struct i2c_client *client, u8 reg, u8 value)
 	struct go7007 *go = i2c_get_adapdata(client->adapter);
 	struct go7007_usb *usb;
 	int rc;
-	int dev_addr = client->addr;
+	int dev_addr = client->addr << 1;  /* firmware wants 8-bit address */
 	u8 *buf;
 
 	if (go == NULL)
diff --git a/drivers/staging/octeon/Makefile b/drivers/staging/octeon/Makefile
index c0a583cc222..87447c102fa 100644
--- a/drivers/staging/octeon/Makefile
+++ b/drivers/staging/octeon/Makefile
@@ -14,7 +14,6 @@ obj-${CONFIG_OCTEON_ETHERNET} :=  octeon-ethernet.o
 octeon-ethernet-objs := ethernet.o
 octeon-ethernet-objs += ethernet-mdio.o
 octeon-ethernet-objs += ethernet-mem.o
-octeon-ethernet-objs += ethernet-proc.o
 octeon-ethernet-objs += ethernet-rgmii.o
 octeon-ethernet-objs += ethernet-rx.o
 octeon-ethernet-objs += ethernet-sgmii.o
diff --git a/drivers/staging/octeon/ethernet-defines.h b/drivers/staging/octeon/ethernet-defines.h
index f13131b03c3..6a2cd50a17d 100644
--- a/drivers/staging/octeon/ethernet-defines.h
+++ b/drivers/staging/octeon/ethernet-defines.h
@@ -41,17 +41,10 @@
  *      Tells the driver to populate the packet buffers with kernel skbuffs.
  *      This allows the driver to receive packets without copying them. It also
  *      means that 32bit userspace can't access the packet buffers.
- *  USE_32BIT_SHARED
- *      This define tells the driver to allocate memory for buffers from the
- *      32bit sahred region instead of the kernel memory space.
  *  USE_HW_TCPUDP_CHECKSUM
  *      Controls if the Octeon TCP/UDP checksum engine is used for packet
  *      output. If this is zero, the kernel will perform the checksum in
  *      software.
- *  USE_MULTICORE_RECEIVE
- *      Process receive interrupts on multiple cores. This spreads the network
- *      load across the first 8 processors. If ths is zero, only one core
- *      processes incomming packets.
  *  USE_ASYNC_IOBDMA
  *      Use asynchronous IO access to hardware. This uses Octeon's asynchronous
  *      IOBDMAs to issue IO accesses without stalling. Set this to zero
@@ -75,29 +68,15 @@
 #define CONFIG_CAVIUM_RESERVE32 0
 #endif
 
-#if CONFIG_CAVIUM_RESERVE32
-#define USE_32BIT_SHARED            1
-#define USE_SKBUFFS_IN_HW           0
-#define REUSE_SKBUFFS_WITHOUT_FREE  0
-#else
-#define USE_32BIT_SHARED            0
 #define USE_SKBUFFS_IN_HW           1
 #ifdef CONFIG_NETFILTER
 #define REUSE_SKBUFFS_WITHOUT_FREE  0
 #else
 #define REUSE_SKBUFFS_WITHOUT_FREE  1
 #endif
-#endif
-
-/* Max interrupts per second per core */
-#define INTERRUPT_LIMIT             10000
 
-/* Don't limit the number of interrupts */
-/*#define INTERRUPT_LIMIT             0     */
 #define USE_HW_TCPUDP_CHECKSUM      1
 
-#define USE_MULTICORE_RECEIVE       1
-
 /* Enable Random Early Dropping under load */
 #define USE_RED                     1
 #define USE_ASYNC_IOBDMA            (CONFIG_CAVIUM_OCTEON_CVMSEG_SIZE > 0)
@@ -115,21 +94,12 @@
 /* Use this to not have FPA frees control L2 */
 /*#define DONT_WRITEBACK(x)         0   */
 
-/* Maximum number of packets to process per interrupt. */
-#define MAX_RX_PACKETS 120
 /* Maximum number of SKBs to try to free per xmit packet. */
-#define MAX_SKB_TO_FREE 10
 #define MAX_OUT_QUEUE_DEPTH 1000
 
-#ifndef CONFIG_SMP
-#undef USE_MULTICORE_RECEIVE
-#define USE_MULTICORE_RECEIVE 0
-#endif
-
-#define IP_PROTOCOL_TCP             6
-#define IP_PROTOCOL_UDP             0x11
+#define FAU_TOTAL_TX_TO_CLEAN (CVMX_FAU_REG_END - sizeof(uint32_t))
+#define FAU_NUM_PACKET_BUFFERS_TO_FREE (FAU_TOTAL_TX_TO_CLEAN - sizeof(uint32_t))
 
-#define FAU_NUM_PACKET_BUFFERS_TO_FREE (CVMX_FAU_REG_END - sizeof(uint32_t))
 #define TOTAL_NUMBER_OF_PORTS       (CVMX_PIP_NUM_INPUT_PORTS+1)
 
 
diff --git a/drivers/staging/octeon/ethernet-mdio.c b/drivers/staging/octeon/ethernet-mdio.c
index 05a5cc0f43e..7e0be8d00dc 100644
--- a/drivers/staging/octeon/ethernet-mdio.c
+++ b/drivers/staging/octeon/ethernet-mdio.c
@@ -96,11 +96,11 @@ const struct ethtool_ops cvm_oct_ethtool_ops = {
 };
 
 /**
- * IOCTL support for PHY control
- *
+ * cvm_oct_ioctl - IOCTL support for PHY control
  * @dev:    Device to change
  * @rq:     the request
  * @cmd:    the command
+ *
  * Returns Zero on success
  */
 int cvm_oct_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
@@ -153,7 +153,7 @@ static void cvm_oct_adjust_link(struct net_device *dev)
 
 
 /**
- * Setup the PHY
+ * cvm_oct_phy_setup_device - setup the PHY
  *
  * @dev:    Device to setup
  *
diff --git a/drivers/staging/octeon/ethernet-mdio.h b/drivers/staging/octeon/ethernet-mdio.h
index 55d0614a7cd..a417d4fce12 100644
--- a/drivers/staging/octeon/ethernet-mdio.h
+++ b/drivers/staging/octeon/ethernet-mdio.h
@@ -32,7 +32,6 @@
 #include <linux/ip.h>
 #include <linux/string.h>
 #include <linux/ethtool.h>
-#include <linux/mii.h>
 #include <linux/seq_file.h>
 #include <linux/proc_fs.h>
 #include <net/dst.h>
diff --git a/drivers/staging/octeon/ethernet-mem.c b/drivers/staging/octeon/ethernet-mem.c
index b595903e2af..00cc91df6b4 100644
--- a/drivers/staging/octeon/ethernet-mem.c
+++ b/drivers/staging/octeon/ethernet-mem.c
@@ -4,7 +4,7 @@
  * Contact: support@caviumnetworks.com
  * This file is part of the OCTEON SDK
  *
- * Copyright (c) 2003-2007 Cavium Networks
+ * Copyright (c) 2003-2010 Cavium Networks
  *
  * This file is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License, Version 2, as
@@ -26,8 +26,6 @@
 **********************************************************************/
 #include <linux/kernel.h>
 #include <linux/netdevice.h>
-#include <linux/mii.h>
-#include <net/dst.h>
 
 #include <asm/octeon/octeon.h>
 
@@ -36,18 +34,19 @@
 #include "cvmx-fpa.h"
 
 /**
- * Fill the supplied hardware pool with skbuffs
- *
+ * cvm_oct_fill_hw_skbuff - fill the supplied hardware pool with skbuffs
  * @pool:     Pool to allocate an skbuff for
  * @size:     Size of the buffer needed for the pool
  * @elements: Number of buffers to allocate
+ *
+ * Returns the actual number of buffers allocated.
  */
 static int cvm_oct_fill_hw_skbuff(int pool, int size, int elements)
 {
 	int freed = elements;
 	while (freed) {
 
-		struct sk_buff *skb = dev_alloc_skb(size + 128);
+		struct sk_buff *skb = dev_alloc_skb(size + 256);
 		if (unlikely(skb == NULL)) {
 			pr_warning
 			    ("Failed to allocate skb for hardware pool %d\n",
@@ -55,7 +54,7 @@ static int cvm_oct_fill_hw_skbuff(int pool, int size, int elements)
 			break;
 		}
 
-		skb_reserve(skb, 128 - (((unsigned long)skb->data) & 0x7f));
+		skb_reserve(skb, 256 - (((unsigned long)skb->data) & 0x7f));
 		*(struct sk_buff **)(skb->data - sizeof(void *)) = skb;
 		cvmx_fpa_free(skb->data, pool, DONT_WRITEBACK(size / 128));
 		freed--;
@@ -64,8 +63,7 @@ static int cvm_oct_fill_hw_skbuff(int pool, int size, int elements)
 }
 
 /**
- * Free the supplied hardware pool of skbuffs
- *
+ * cvm_oct_free_hw_skbuff- free hardware pool skbuffs
  * @pool:     Pool to allocate an skbuff for
  * @size:     Size of the buffer needed for the pool
  * @elements: Number of buffers to allocate
@@ -93,96 +91,76 @@ static void cvm_oct_free_hw_skbuff(int pool, int size, int elements)
 }
 
 /**
- * This function fills a hardware pool with memory. Depending
- * on the config defines, this memory might come from the
- * kernel or global 32bit memory allocated with
- * cvmx_bootmem_alloc.
- *
+ * cvm_oct_fill_hw_memory - fill a hardware pool with memory.
  * @pool:     Pool to populate
  * @size:     Size of each buffer in the pool
  * @elements: Number of buffers to allocate
+ *
+ * Returns the actual number of buffers allocated.
  */
 static int cvm_oct_fill_hw_memory(int pool, int size, int elements)
 {
 	char *memory;
+	char *fpa;
 	int freed = elements;
 
-	if (USE_32BIT_SHARED) {
-		extern uint64_t octeon_reserve32_memory;
-
-		memory =
-		    cvmx_bootmem_alloc_range(elements * size, 128,
-					     octeon_reserve32_memory,
-					     octeon_reserve32_memory +
-					     (CONFIG_CAVIUM_RESERVE32 << 20) -
-					     1);
-		if (memory == NULL)
-			panic("Unable to allocate %u bytes for FPA pool %d\n",
-			      elements * size, pool);
-
-		pr_notice("Memory range %p - %p reserved for "
-			  "hardware\n", memory,
-			  memory + elements * size - 1);
-
-		while (freed) {
-			cvmx_fpa_free(memory, pool, 0);
-			memory += size;
-			freed--;
-		}
-	} else {
-		while (freed) {
-			/* We need to force alignment to 128 bytes here */
-			memory = kmalloc(size + 127, GFP_ATOMIC);
-			if (unlikely(memory == NULL)) {
-				pr_warning("Unable to allocate %u bytes for "
-					   "FPA pool %d\n",
-				     elements * size, pool);
-				break;
-			}
-			memory = (char *)(((unsigned long)memory + 127) & -128);
-			cvmx_fpa_free(memory, pool, 0);
-			freed--;
+	while (freed) {
+		/*
+		 * FPA memory must be 128 byte aligned.  Since we are
+		 * aligning we need to save the original pointer so we
+		 * can feed it to kfree when the memory is returned to
+		 * the kernel.
+		 *
+		 * We allocate an extra 256 bytes to allow for
+		 * alignment and space for the original pointer saved
+		 * just before the block.
+		 */
+		memory = kmalloc(size + 256, GFP_ATOMIC);
+		if (unlikely(memory == NULL)) {
+			pr_warning("Unable to allocate %u bytes for FPA pool %d\n",
+				   elements * size, pool);
+			break;
 		}
+		fpa = (char *)(((unsigned long)memory + 256) & ~0x7fUL);
+		*((char **)fpa - 1) = memory;
+		cvmx_fpa_free(fpa, pool, 0);
+		freed--;
 	}
 	return elements - freed;
 }
 
 /**
- * Free memory previously allocated with cvm_oct_fill_hw_memory
- *
+ * cvm_oct_free_hw_memory - Free memory allocated by cvm_oct_fill_hw_memory
  * @pool:     FPA pool to free
  * @size:     Size of each buffer in the pool
  * @elements: Number of buffers that should be in the pool
  */
 static void cvm_oct_free_hw_memory(int pool, int size, int elements)
 {
-	if (USE_32BIT_SHARED) {
-		pr_warning("Warning: 32 shared memory is not freeable\n");
-	} else {
-		char *memory;
-		do {
-			memory = cvmx_fpa_alloc(pool);
-			if (memory) {
-				elements--;
-				kfree(phys_to_virt(cvmx_ptr_to_phys(memory)));
-			}
-		} while (memory);
+	char *memory;
+	char *fpa;
+	do {
+		fpa = cvmx_fpa_alloc(pool);
+		if (fpa) {
+			elements--;
+			fpa = (char *)phys_to_virt(cvmx_ptr_to_phys(fpa));
+			memory = *((char **)fpa - 1);
+			kfree(memory);
+		}
+	} while (fpa);
 
-		if (elements < 0)
-			pr_warning("Freeing of pool %u had too many "
-				   "buffers (%d)\n",
-			       pool, elements);
-		else if (elements > 0)
-			pr_warning("Warning: Freeing of pool %u is "
-				"missing %d buffers\n",
-			     pool, elements);
-	}
+	if (elements < 0)
+		pr_warning("Freeing of pool %u had too many buffers (%d)\n",
+			pool, elements);
+	else if (elements > 0)
+		pr_warning("Warning: Freeing of pool %u is missing %d buffers\n",
+			pool, elements);
 }
 
 int cvm_oct_mem_fill_fpa(int pool, int size, int elements)
 {
 	int freed;
-	if (USE_SKBUFFS_IN_HW)
+	if (USE_SKBUFFS_IN_HW && pool == CVMX_FPA_PACKET_POOL)
 		freed = cvm_oct_fill_hw_skbuff(pool, size, elements);
 	else
 		freed = cvm_oct_fill_hw_memory(pool, size, elements);
@@ -191,7 +169,7 @@ int cvm_oct_mem_fill_fpa(int pool, int size, int elements)
 
 void cvm_oct_mem_empty_fpa(int pool, int size, int elements)
 {
-	if (USE_SKBUFFS_IN_HW)
+	if (USE_SKBUFFS_IN_HW && pool == CVMX_FPA_PACKET_POOL)
 		cvm_oct_free_hw_skbuff(pool, size, elements);
 	else
 		cvm_oct_free_hw_memory(pool, size, elements);
diff --git a/drivers/staging/octeon/ethernet-proc.c b/drivers/staging/octeon/ethernet-proc.c
deleted file mode 100644
index 16308d484d3..00000000000
--- a/drivers/staging/octeon/ethernet-proc.c
+++ /dev/null
@@ -1,144 +0,0 @@
-/**********************************************************************
- * Author: Cavium Networks
- *
- * Contact: support@caviumnetworks.com
- * This file is part of the OCTEON SDK
- *
- * Copyright (c) 2003-2007 Cavium Networks
- *
- * This file is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License, Version 2, as
- * published by the Free Software Foundation.
- *
- * This file is distributed in the hope that it will be useful, but
- * AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or
- * NONINFRINGEMENT.  See the GNU General Public License for more
- * details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this file; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- * or visit http://www.gnu.org/licenses/.
- *
- * This file may also be available under a different license from Cavium.
- * Contact Cavium Networks for more information
-**********************************************************************/
-#include <linux/kernel.h>
-#include <linux/seq_file.h>
-#include <linux/proc_fs.h>
-#include <net/dst.h>
-
-#include <asm/octeon/octeon.h>
-
-#include "octeon-ethernet.h"
-#include "ethernet-defines.h"
-
-#include "cvmx-helper.h"
-#include "cvmx-pip.h"
-
-/**
- * User is reading /proc/octeon_ethernet_stats
- *
- * @m:
- * @v:
- * Returns
- */
-static int cvm_oct_stats_show(struct seq_file *m, void *v)
-{
-	struct octeon_ethernet *priv;
-	int port;
-
-	for (port = 0; port < TOTAL_NUMBER_OF_PORTS; port++) {
-
-		if (cvm_oct_device[port]) {
-			priv = netdev_priv(cvm_oct_device[port]);
-
-			seq_printf(m, "\nOcteon Port %d (%s)\n", port,
-				   cvm_oct_device[port]->name);
-			seq_printf(m,
-				   "rx_packets:             %12lu\t"
-				   "tx_packets:             %12lu\n",
-				   priv->stats.rx_packets,
-				   priv->stats.tx_packets);
-			seq_printf(m,
-				   "rx_bytes:               %12lu\t"
-				   "tx_bytes:               %12lu\n",
-				   priv->stats.rx_bytes, priv->stats.tx_bytes);
-			seq_printf(m,
-				   "rx_errors:              %12lu\t"
-				   "tx_errors:              %12lu\n",
-				   priv->stats.rx_errors,
-				   priv->stats.tx_errors);
-			seq_printf(m,
-				   "rx_dropped:             %12lu\t"
-				   "tx_dropped:             %12lu\n",
-				   priv->stats.rx_dropped,
-				   priv->stats.tx_dropped);
-			seq_printf(m,
-				   "rx_length_errors:       %12lu\t"
-				   "tx_aborted_errors:      %12lu\n",
-				   priv->stats.rx_length_errors,
-				   priv->stats.tx_aborted_errors);
-			seq_printf(m,
-				   "rx_over_errors:         %12lu\t"
-				   "tx_carrier_errors:      %12lu\n",
-				   priv->stats.rx_over_errors,
-				   priv->stats.tx_carrier_errors);
-			seq_printf(m,
-				   "rx_crc_errors:          %12lu\t"
-				   "tx_fifo_errors:         %12lu\n",
-				   priv->stats.rx_crc_errors,
-				   priv->stats.tx_fifo_errors);
-			seq_printf(m,
-				   "rx_frame_errors:        %12lu\t"
-				   "tx_heartbeat_errors:    %12lu\n",
-				   priv->stats.rx_frame_errors,
-				   priv->stats.tx_heartbeat_errors);
-			seq_printf(m,
-				   "rx_fifo_errors:         %12lu\t"
-				   "tx_window_errors:       %12lu\n",
-				   priv->stats.rx_fifo_errors,
-				   priv->stats.tx_window_errors);
-			seq_printf(m,
-				   "rx_missed_errors:       %12lu\t"
-				   "multicast:              %12lu\n",
-				   priv->stats.rx_missed_errors,
-				   priv->stats.multicast);
-		}
-	}
-
-	return 0;
-}
-
-/**
- * /proc/octeon_ethernet_stats was openned. Use the single_open iterator
- *
- * @inode:
- * @file:
- * Returns
- */
-static int cvm_oct_stats_open(struct inode *inode, struct file *file)
-{
-	return single_open(file, cvm_oct_stats_show, NULL);
-}
-
-static const struct file_operations cvm_oct_stats_operations = {
-	.open = cvm_oct_stats_open,
-	.read = seq_read,
-	.llseek = seq_lseek,
-	.release = single_release,
-};
-
-void cvm_oct_proc_initialize(void)
-{
-	struct proc_dir_entry *entry =
-	    create_proc_entry("octeon_ethernet_stats", 0, NULL);
-	if (entry)
-		entry->proc_fops = &cvm_oct_stats_operations;
-}
-
-void cvm_oct_proc_shutdown(void)
-{
-	remove_proc_entry("octeon_ethernet_stats", NULL);
-}
diff --git a/drivers/staging/octeon/ethernet-proc.h b/drivers/staging/octeon/ethernet-proc.h
deleted file mode 100644
index 82c7d9f78bc..00000000000
--- a/drivers/staging/octeon/ethernet-proc.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*********************************************************************
- * Author: Cavium Networks
- *
- * Contact: support@caviumnetworks.com
- * This file is part of the OCTEON SDK
- *
- * Copyright (c) 2003-2007 Cavium Networks
- *
- * This file is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License, Version 2, as
- * published by the Free Software Foundation.
- *
- * This file is distributed in the hope that it will be useful, but
- * AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or
- * NONINFRINGEMENT.  See the GNU General Public License for more
- * details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this file; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- * or visit http://www.gnu.org/licenses/.
- *
- * This file may also be available under a different license from Cavium.
- * Contact Cavium Networks for more information
-*********************************************************************/
-
-void cvm_oct_proc_initialize(void);
-void cvm_oct_proc_shutdown(void);
diff --git a/drivers/staging/octeon/ethernet-rgmii.c b/drivers/staging/octeon/ethernet-rgmii.c
index 3820f1ec11d..a0d4d4b98bd 100644
--- a/drivers/staging/octeon/ethernet-rgmii.c
+++ b/drivers/staging/octeon/ethernet-rgmii.c
@@ -26,7 +26,7 @@
 **********************************************************************/
 #include <linux/kernel.h>
 #include <linux/netdevice.h>
-#include <linux/mii.h>
+#include <linux/phy.h>
 #include <net/dst.h>
 
 #include <asm/octeon/octeon.h>
@@ -48,14 +48,20 @@ static int number_rgmii_ports;
 static void cvm_oct_rgmii_poll(struct net_device *dev)
 {
 	struct octeon_ethernet *priv = netdev_priv(dev);
-	unsigned long flags;
+	unsigned long flags = 0;
 	cvmx_helper_link_info_t link_info;
+	int use_global_register_lock = (priv->phydev == NULL);
 
-	/*
-	 * Take the global register lock since we are going to touch
-	 * registers that affect more than one port.
-	 */
-	spin_lock_irqsave(&global_register_lock, flags);
+	BUG_ON(in_interrupt());
+	if (use_global_register_lock) {
+		/*
+		 * Take the global register lock since we are going to
+		 * touch registers that affect more than one port.
+		 */
+		spin_lock_irqsave(&global_register_lock, flags);
+	} else {
+		mutex_lock(&priv->phydev->bus->mdio_lock);
+	}
 
 	link_info = cvmx_helper_link_get(priv->port);
 	if (link_info.u64 == priv->link_info) {
@@ -115,7 +121,11 @@ static void cvm_oct_rgmii_poll(struct net_device *dev)
 				     dev->name);
 			}
 		}
-		spin_unlock_irqrestore(&global_register_lock, flags);
+
+		if (use_global_register_lock)
+			spin_unlock_irqrestore(&global_register_lock, flags);
+		else
+			mutex_unlock(&priv->phydev->bus->mdio_lock);
 		return;
 	}
 
@@ -151,7 +161,12 @@ static void cvm_oct_rgmii_poll(struct net_device *dev)
 		link_info = cvmx_helper_link_autoconf(priv->port);
 		priv->link_info = link_info.u64;
 	}
-	spin_unlock_irqrestore(&global_register_lock, flags);
+
+	if (use_global_register_lock)
+		spin_unlock_irqrestore(&global_register_lock, flags);
+	else {
+		mutex_unlock(&priv->phydev->bus->mdio_lock);
+	}
 
 	if (priv->phydev == NULL) {
 		/* Tell core. */
@@ -213,8 +228,11 @@ static irqreturn_t cvm_oct_rgmii_rml_interrupt(int cpl, void *dev_id)
 				struct net_device *dev =
 				    cvm_oct_device[cvmx_helper_get_ipd_port
 						   (interface, index)];
-				if (dev)
-					cvm_oct_rgmii_poll(dev);
+				struct octeon_ethernet *priv = netdev_priv(dev);
+
+				if (dev && !atomic_read(&cvm_oct_poll_queue_stopping))
+					queue_work(cvm_oct_poll_queue, &priv->port_work);
+
 				gmx_rx_int_reg.u64 = 0;
 				gmx_rx_int_reg.s.phy_dupx = 1;
 				gmx_rx_int_reg.s.phy_link = 1;
@@ -252,8 +270,11 @@ static irqreturn_t cvm_oct_rgmii_rml_interrupt(int cpl, void *dev_id)
 				struct net_device *dev =
 				    cvm_oct_device[cvmx_helper_get_ipd_port
 						   (interface, index)];
-				if (dev)
-					cvm_oct_rgmii_poll(dev);
+				struct octeon_ethernet *priv = netdev_priv(dev);
+
+				if (dev && !atomic_read(&cvm_oct_poll_queue_stopping))
+					queue_work(cvm_oct_poll_queue, &priv->port_work);
+
 				gmx_rx_int_reg.u64 = 0;
 				gmx_rx_int_reg.s.phy_dupx = 1;
 				gmx_rx_int_reg.s.phy_link = 1;
@@ -302,6 +323,12 @@ int cvm_oct_rgmii_stop(struct net_device *dev)
 	return 0;
 }
 
+static void cvm_oct_rgmii_immediate_poll(struct work_struct *work)
+{
+	struct octeon_ethernet *priv = container_of(work, struct octeon_ethernet, port_work);
+	cvm_oct_rgmii_poll(cvm_oct_device[priv->port]);
+}
+
 int cvm_oct_rgmii_init(struct net_device *dev)
 {
 	struct octeon_ethernet *priv = netdev_priv(dev);
@@ -309,7 +336,7 @@ int cvm_oct_rgmii_init(struct net_device *dev)
 
 	cvm_oct_common_init(dev);
 	dev->netdev_ops->ndo_stop(dev);
-
+	INIT_WORK(&priv->port_work, cvm_oct_rgmii_immediate_poll);
 	/*
 	 * Due to GMX errata in CN3XXX series chips, it is necessary
 	 * to take the link down immediately when the PHY changes
@@ -397,4 +424,5 @@ void cvm_oct_rgmii_uninit(struct net_device *dev)
 	number_rgmii_ports--;
 	if (number_rgmii_ports == 0)
 		free_irq(OCTEON_IRQ_RML, &number_rgmii_ports);
+	cancel_work_sync(&priv->port_work);
 }
diff --git a/drivers/staging/octeon/ethernet-rx.c b/drivers/staging/octeon/ethernet-rx.c
index 1b237b7e689..cb38f9eb2cc 100644
--- a/drivers/staging/octeon/ethernet-rx.c
+++ b/drivers/staging/octeon/ethernet-rx.c
@@ -4,7 +4,7 @@
  * Contact: support@caviumnetworks.com
  * This file is part of the OCTEON SDK
  *
- * Copyright (c) 2003-2007 Cavium Networks
+ * Copyright (c) 2003-2010 Cavium Networks
  *
  * This file is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License, Version 2, as
@@ -27,16 +27,14 @@
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/cache.h>
+#include <linux/cpumask.h>
 #include <linux/netdevice.h>
 #include <linux/init.h>
 #include <linux/etherdevice.h>
 #include <linux/ip.h>
 #include <linux/string.h>
 #include <linux/prefetch.h>
-#include <linux/ethtool.h>
-#include <linux/mii.h>
-#include <linux/seq_file.h>
-#include <linux/proc_fs.h>
+#include <linux/smp.h>
 #include <net/dst.h>
 #ifdef CONFIG_XFRM
 #include <linux/xfrm.h>
@@ -48,8 +46,9 @@
 #include <asm/octeon/octeon.h>
 
 #include "ethernet-defines.h"
-#include "octeon-ethernet.h"
 #include "ethernet-mem.h"
+#include "ethernet-rx.h"
+#include "octeon-ethernet.h"
 #include "ethernet-util.h"
 
 #include "cvmx-helper.h"
@@ -61,62 +60,88 @@
 
 #include "cvmx-gmxx-defs.h"
 
-struct cvm_tasklet_wrapper {
-	struct tasklet_struct t;
-};
+struct cvm_napi_wrapper {
+	struct napi_struct napi;
+} ____cacheline_aligned_in_smp;
 
-/*
- * Aligning the tasklet_struct on cachline boundries seems to decrease
- * throughput even though in theory it would reduce contantion on the
- * cache lines containing the locks.
- */
+static struct cvm_napi_wrapper cvm_oct_napi[NR_CPUS] __cacheline_aligned_in_smp;
 
-static struct cvm_tasklet_wrapper cvm_oct_tasklet[NR_CPUS];
+struct cvm_oct_core_state {
+	int baseline_cores;
+	/*
+	 * The number of additional cores that could be processing
+	 * input packtes.
+	 */
+	atomic_t available_cores;
+	cpumask_t cpu_state;
+} ____cacheline_aligned_in_smp;
 
-/**
- * Interrupt handler. The interrupt occurs whenever the POW
- * transitions from 0->1 packets in our group.
- *
- * @cpl:
- * @dev_id:
- * @regs:
- * Returns
- */
-irqreturn_t cvm_oct_do_interrupt(int cpl, void *dev_id)
+static struct cvm_oct_core_state core_state __cacheline_aligned_in_smp;
+
+static void cvm_oct_enable_napi(void *_)
 {
-	/* Acknowledge the interrupt */
-	if (INTERRUPT_LIMIT)
-		cvmx_write_csr(CVMX_POW_WQ_INT, 1 << pow_receive_group);
-	else
-		cvmx_write_csr(CVMX_POW_WQ_INT, 0x10001 << pow_receive_group);
-	preempt_disable();
-	tasklet_schedule(&cvm_oct_tasklet[smp_processor_id()].t);
-	preempt_enable();
-	return IRQ_HANDLED;
+	int cpu = smp_processor_id();
+	napi_schedule(&cvm_oct_napi[cpu].napi);
+}
+
+static void cvm_oct_enable_one_cpu(void)
+{
+	int v;
+	int cpu;
+
+	/* Check to see if more CPUs are available for receive processing... */
+	v = atomic_sub_if_positive(1, &core_state.available_cores);
+	if (v < 0)
+		return;
+
+	/* ... if a CPU is available, Turn on NAPI polling for that CPU.  */
+	for_each_online_cpu(cpu) {
+		if (!cpu_test_and_set(cpu, core_state.cpu_state)) {
+			v = smp_call_function_single(cpu, cvm_oct_enable_napi,
+						     NULL, 0);
+			if (v)
+				panic("Can't enable NAPI.");
+			break;
+		}
+	}
+}
+
+static void cvm_oct_no_more_work(void)
+{
+	int cpu = smp_processor_id();
+
+	/*
+	 * CPU zero is special.  It always has the irq enabled when
+	 * waiting for incoming packets.
+	 */
+	if (cpu == 0) {
+		enable_irq(OCTEON_IRQ_WORKQ0 + pow_receive_group);
+		return;
+	}
+
+	cpu_clear(cpu, core_state.cpu_state);
+	atomic_add(1, &core_state.available_cores);
 }
 
-#ifdef CONFIG_NET_POLL_CONTROLLER
 /**
- * This is called when the kernel needs to manually poll the
- * device. For Octeon, this is simply calling the interrupt
- * handler. We actually poll all the devices, not just the
- * one supplied.
+ * cvm_oct_do_interrupt - interrupt handler.
+ *
+ * The interrupt occurs whenever the POW has packets in our group.
  *
- * @dev:    Device to poll. Unused
  */
-void cvm_oct_poll_controller(struct net_device *dev)
+static irqreturn_t cvm_oct_do_interrupt(int cpl, void *dev_id)
 {
-	preempt_disable();
-	tasklet_schedule(&cvm_oct_tasklet[smp_processor_id()].t);
-	preempt_enable();
+	/* Disable the IRQ and start napi_poll. */
+	disable_irq_nosync(OCTEON_IRQ_WORKQ0 + pow_receive_group);
+	cvm_oct_enable_napi(NULL);
+
+	return IRQ_HANDLED;
 }
-#endif
 
 /**
- * This is called on receive errors, and determines if the packet
- * can be dropped early-on in cvm_oct_tasklet_rx().
- *
+ * cvm_oct_check_rcv_error - process receive errors
  * @work: Work queue entry pointing to the packet.
+ *
  * Returns Non-zero if the packet can be dropped, zero otherwise.
  */
 static inline int cvm_oct_check_rcv_error(cvmx_wqe_t *work)
@@ -199,19 +224,20 @@ static inline int cvm_oct_check_rcv_error(cvmx_wqe_t *work)
 }
 
 /**
- * Tasklet function that is scheduled on a core when an interrupt occurs.
+ * cvm_oct_napi_poll - the NAPI poll function.
+ * @napi: The NAPI instance, or null if called from cvm_oct_poll_controller
+ * @budget: Maximum number of packets to receive.
  *
- * @unused:
+ * Returns the number of packets processed.
  */
-void cvm_oct_tasklet_rx(unsigned long unused)
+static int cvm_oct_napi_poll(struct napi_struct *napi, int budget)
 {
-	const int coreid = cvmx_get_core_num();
-	uint64_t old_group_mask;
-	uint64_t old_scratch;
-	int rx_count = 0;
-	int number_to_free;
-	int num_freed;
-	int packet_not_copied;
+	const int	coreid = cvmx_get_core_num();
+	uint64_t	old_group_mask;
+	uint64_t	old_scratch;
+	int		rx_count = 0;
+	int		did_work_request = 0;
+	int		packet_not_copied;
 
 	/* Prefetch cvm_oct_