kernel / drivers / net / ns83820.c

#define VERSION "0.22"
/* ns83820.c by Benjamin LaHaise with contributions.
 *
 * Questions/comments/discussion to [e-mail].
 *
 * $Revision: 1.34.2.23 $
 *
 * Copyright 2001 Benjamin LaHaise.
 * Copyright 2001, 2002 Red Hat.
 *
 * Mmmm, chocolate vanilla mocha...
 *
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 *
 * ChangeLog
 * =========
 *	20010414	0.1 - created
 *	20010622	0.2 - basic rx and tx.
 *	20010711	0.3 - added duplex and link state detection support.
 *	20010713	0.4 - zero copy, no hangs.
 *			0.5 - 64 bit dma support (davem will hate me for this)
 *			    - disable jumbo frames to avoid tx hangs
 *			    - work around tx deadlocks on my 1.02 card via
 *			      fiddling with TXCFG
 *	20010810	0.6 - use pci dma api for ringbuffers, work on ia64
 *	20010816	0.7 - misc cleanups
 *	20010826	0.8 - fix critical zero copy bugs
 *			0.9 - internal experiment
 *	20010827	0.10 - fix ia64 unaligned access.
 *	20010906	0.11 - accept all packets with checksum errors as
 *			       otherwise fragments get lost
 *			     - fix >> 32 bugs
 *			0.12 - add statistics counters
 *			     - add allmulti/promisc support
 *	20011009	0.13 - hotplug support, other smaller pci api cleanups
 *	20011204	0.13a - optical transceiver support added
 *				by Michael Clark <[e-mail]>
 *	20011205	0.13b - call register_netdev earlier in initialization
 *				suppress duplicate link status messages
 *	20011117 	0.14 - ethtool GDRVINFO, GLINK support from jgarzik
 *	20011204 	0.15	get ppc (big endian) working
 *	20011218	0.16	various cleanups
 *	20020310	0.17	speedups
 *	20020610	0.18 -	actually use the pci dma api for highmem
 *			     -	remove pci latency register fiddling
 *			0.19 -	better bist support
 *			     -	add ihr and reset_phy parameters
 *			     -	gmii bus probing
 *			     -	fix missed txok introduced during performance
 *				tuning
 *			0.20 -	fix stupid RFEN thinko.  i am such a smurf.
 *	20040828	0.21 -	add hardware vlan accleration
 *				by Neil Horman <[e-mail]>
 *	20050406	0.22 -	improved DAC ifdefs from Andi Kleen
 *			     -	removal of dead code from Adrian Bunk
 *			     -	fix half duplex collision behaviour
 * Driver Overview
 * ===============
 *
 * This driver was originally written for the National Semiconductor
 * 83820 chip, a 10/100/1000 Mbps 64 bit PCI ethernet NIC.  Hopefully
 * this code will turn out to be a) clean, b) correct, and c) fast.
 * With that in mind, I'm aiming to split the code up as much as
 * reasonably possible.  At present there are X major sections that
 * break down into a) packet receive, b) packet transmit, c) link
 * management, d) initialization and configuration.  Where possible,
 * these code paths are designed to run in parallel.
 *
 * This driver has been tested and found to work with the following
 * cards (in no particular order):
 *
 *	Cameo		SOHO-GA2000T	SOHO-GA2500T
 *	D-Link		DGE-500T
 *	PureData	PDP8023Z-TG
 *	SMC		SMC9452TX	SMC9462TX
 *	Netgear		GA621
 *
 * Special thanks to SMC for providing hardware to test this driver on.
 *
 * Reports of success or failure would be greatly appreciated.
 */
//#define dprintk		printk
#define dprintk(x...)		do { } while (0)

#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/types.h>
#include <linux/pci.h>
#include <linux/dma-mapping.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/delay.h>
#include <linux/smp_lock.h>
#include <linux/workqueue.h>
#include <linux/init.h>
#include <linux/ip.h>	/* for iph */
#include <linux/in.h>	/* for IPPROTO_... */
#include <linux/compiler.h>
#include <linux/prefetch.h>
#include <linux/ethtool.h>
#include <linux/timer.h>
#include <linux/if_vlan.h>
#include <linux/rtnetlink.h>
#include <linux/jiffies.h>

#include <asm/io.h>
#include <asm/uaccess.h>
#include <asm/system.h>

#define DRV_NAME "ns83820"

/* Global parameters.  See module_param near the bottom. */
static int ihr = 2;
static int reset_phy = 0;
static int lnksts = 0;		/* CFG_LNKSTS bit polarity */

/* Dprintk is used for more interesting debug events */
#undef Dprintk
#define	Dprintk			dprintk

/* tunables */
#define RX_BUF_SIZE	1500	/* 8192 */
#if defined(CONFIG_VLAN_8021Q) || defined(CONFIG_VLAN_8021Q_MODULE)
#define NS83820_VLAN_ACCEL_SUPPORT
#endif

/* Must not exceed ~65000. */
#define NR_RX_DESC	64
#define NR_TX_DESC	128

/* not tunable */
#define REAL_RX_BUF_SIZE (RX_BUF_SIZE + 14)	/* rx/tx mac addr + type */

#define MIN_TX_DESC_FREE	8

/* register defines */
#define CFGCS		0x04

#define CR_TXE		0x00000001
#define CR_TXD		0x00000002
/* Ramit : Here's a tip, don't do a RXD immediately followed by an RXE
 * The Receive engine skips one descriptor and moves
 * onto the next one!! */
#define CR_RXE		0x00000004
#define CR_RXD		0x00000008
#define CR_TXR		0x00000010
#define CR_RXR		0x00000020
#define CR_SWI		0x00000080
#define CR_RST		0x00000100

#define PTSCR_EEBIST_FAIL       0x00000001
#define PTSCR_EEBIST_EN         0x00000002
#define PTSCR_EELOAD_EN         0x00000004
#define PTSCR_RBIST_FAIL        0x000001b8
#define PTSCR_RBIST_DONE        0x00000200
#define PTSCR_RBIST_EN          0x00000400
#define PTSCR_RBIST_RST         0x00002000

#define MEAR_EEDI		0x00000001
#define MEAR_EEDO		0x00000002
#define MEAR_EECLK		0x00000004
#define MEAR_EESEL		0x00000008
#define MEAR_MDIO		0x00000010
#define MEAR_MDDIR		0x00000020
#define MEAR_MDC		0x00000040

#define ISR_TXDESC3	0x40000000
#define ISR_TXDESC2	0x20000000
#define ISR_TXDESC1	0x10000000
#define ISR_TXDESC0	0x08000000
#define ISR_RXDESC3	0x04000000
#define ISR_RXDESC2	0x02000000
#define ISR_RXDESC1	0x01000000
#define ISR_RXDESC0	0x00800000
#define ISR_TXRCMP	0x00400000
#define ISR_RXRCMP	0x00200000
#define ISR_DPERR	0x00100000
#define ISR_SSERR	0x00080000
#define ISR_RMABT	0x00040000
#define ISR_RTABT	0x00020000
#define ISR_RXSOVR	0x00010000
#define ISR_HIBINT	0x00008000
#define ISR_PHY		0x00004000
#define ISR_PME		0x00002000
#define ISR_SWI		0x00001000
#define ISR_MIB		0x00000800
#define ISR_TXURN	0x00000400
#define ISR_TXIDLE	0x00000200
#define ISR_TXERR	0x00000100
#define ISR_TXDESC	0x00000080
#define ISR_TXOK	0x00000040
#define ISR_RXORN	0x00000020
#define ISR_RXIDLE	0x00000010
#define ISR_RXEARLY	0x00000008
#define ISR_RXERR	0x00000004
#define ISR_RXDESC	0x00000002
#define ISR_RXOK	0x00000001

#define TXCFG_CSI	0x80000000
#define TXCFG_HBI	0x40000000
#define TXCFG_MLB	0x20000000
#define TXCFG_ATP	0x10000000
#define TXCFG_ECRETRY	0x00800000
#define TXCFG_BRST_DIS	0x00080000
#define TXCFG_MXDMA1024	0x00000000
#define TXCFG_MXDMA512	0x00700000
#define TXCFG_MXDMA256	0x00600000
#define TXCFG_MXDMA128	0x00500000
#define TXCFG_MXDMA64	0x00400000
#define TXCFG_MXDMA32	0x00300000
#define TXCFG_MXDMA16	0x00200000
#define TXCFG_MXDMA8	0x00100000

#define CFG_LNKSTS	0x80000000
#define CFG_SPDSTS	0x60000000
#define CFG_SPDSTS1	0x40000000
#define CFG_SPDSTS0	0x20000000
#define CFG_DUPSTS	0x10000000
#define CFG_TBI_EN	0x01000000
#define CFG_MODE_1000	0x00400000
/* Ramit : Dont' ever use AUTO_1000, it never works and is buggy.
 * Read the Phy response and then configure the MAC accordingly */
#define CFG_AUTO_1000	0x00200000
#define CFG_PINT_CTL	0x001c0000
#define CFG_PINT_DUPSTS	0x00100000
#define CFG_PINT_LNKSTS	0x00080000
#define CFG_PINT_SPDSTS	0x00040000
#define CFG_TMRTEST	0x00020000
#define CFG_MRM_DIS	0x00010000
#define CFG_MWI_DIS	0x00008000
#define CFG_T64ADDR	0x00004000
#define CFG_PCI64_DET	0x00002000
#define CFG_DATA64_EN	0x00001000
#define CFG_M64ADDR	0x00000800
#define CFG_PHY_RST	0x00000400
#define CFG_PHY_DIS	0x00000200
#define CFG_EXTSTS_EN	0x00000100
#define CFG_REQALG	0x00000080
#define CFG_SB		0x00000040
#define CFG_POW		0x00000020
#define CFG_EXD		0x00000010
#define CFG_PESEL	0x00000008
#define CFG_BROM_DIS	0x00000004
#define CFG_EXT_125	0x00000002
#define CFG_BEM		0x00000001

#define EXTSTS_UDPPKT	0x00200000
#define EXTSTS_TCPPKT	0x00080000
#define EXTSTS_IPPKT	0x00020000
#define EXTSTS_VPKT	0x00010000
#define EXTSTS_VTG_MASK	0x0000ffff

#define SPDSTS_POLARITY	(CFG_SPDSTS1 | CFG_SPDSTS0 | CFG_DUPSTS | (lnksts ? CFG_LNKSTS : 0))

#define MIBC_MIBS	0x00000008
#define MIBC_ACLR	0x00000004
#define MIBC_FRZ	0x00000002
#define MIBC_WRN	0x00000001

#define PCR_PSEN	(1 << 31)
#define PCR_PS_MCAST	(1 << 30)
#define PCR_PS_DA	(1 << 29)
#define PCR_STHI_8	(3 << 23)
#define PCR_STLO_4	(1 << 23)
#define PCR_FFHI_8K	(3 << 21)
#define PCR_FFLO_4K	(1 << 21)
#define PCR_PAUSE_CNT	0xFFFE

#define RXCFG_AEP	0x80000000
#define RXCFG_ARP	0x40000000
#define RXCFG_STRIPCRC	0x20000000
#define RXCFG_RX_FD	0x10000000
#define RXCFG_ALP	0x08000000
#define RXCFG_AIRL	0x04000000
#define RXCFG_MXDMA512	0x00700000
#define RXCFG_DRTH	0x0000003e
#define RXCFG_DRTH0	0x00000002

#define RFCR_RFEN	0x80000000
#define RFCR_AAB	0x40000000
#define RFCR_AAM	0x20000000
#define RFCR_AAU	0x10000000
#define RFCR_APM	0x08000000
#define RFCR_APAT	0x07800000
#define RFCR_APAT3	0x04000000
#define RFCR_APAT2	0x02000000
#define RFCR_APAT1	0x01000000
#define RFCR_APAT0	0x00800000
#define RFCR_AARP	0x00400000
#define RFCR_MHEN	0x00200000
#define RFCR_UHEN	0x00100000
#define RFCR_ULM	0x00080000

#define VRCR_RUDPE	0x00000080
#define VRCR_RTCPE	0x00000040
#define VRCR_RIPE	0x00000020
#define VRCR_IPEN	0x00000010
#define VRCR_DUTF	0x00000008
#define VRCR_DVTF	0x00000004
#define VRCR_VTREN	0x00000002
#define VRCR_VTDEN	0x00000001

#define VTCR_PPCHK	0x00000008
#define VTCR_GCHK	0x00000004
#define VTCR_VPPTI	0x00000002
#define VTCR_VGTI	0x00000001

#define CR		0x00
#define CFG		0x04
#define MEAR		0x08
#define PTSCR		0x0c
#define	ISR		0x10
#define	IMR		0x14
#define	IER		0x18
#define	IHR		0x1c
#define TXDP		0x20
#define TXDP_HI		0x24
#define TXCFG		0x28
#define GPIOR		0x2c
#define RXDP		0x30
#define RXDP_HI		0x34
#define RXCFG		0x38
#define PQCR		0x3c
#define WCSR		0x40
#define PCR		0x44
#define RFCR		0x48
#define RFDR		0x4c

#define SRR		0x58

#define VRCR		0xbc
#define VTCR		0xc0
#define VDR		0xc4
#define CCSR		0xcc

#define TBICR		0xe0
#define TBISR		0xe4
#define TANAR		0xe8
#define TANLPAR		0xec
#define TANER		0xf0
#define TESR		0xf4

#define TBICR_MR_AN_ENABLE	0x00001000
#define TBICR_MR_RESTART_AN	0x00000200

#define TBISR_MR_LINK_STATUS	0x00000020
#define TBISR_MR_AN_COMPLETE	0x00000004

#define TANAR_PS2 		0x00000100
#define TANAR_PS1 		0x00000080
#define TANAR_HALF_DUP 		0x00000040
#define TANAR_FULL_DUP 		0x00000020

#define GPIOR_GP5_OE		0x00000200
#define GPIOR_GP4_OE		0x00000100
#define GPIOR_GP3_OE		0x00000080
#define GPIOR_GP2_OE		0x00000040
#define GPIOR_GP1_OE		0x00000020
#define GPIOR_GP3_OUT		0x00000004
#define GPIOR_GP1_OUT		0x00000001

#define LINK_AUTONEGOTIATE	0x01
#define LINK_DOWN		0x02
#define LINK_UP			0x04

#define HW_ADDR_LEN	sizeof(dma_addr_t)
#define desc_addr_set(desc, addr)				\
	do {							\
		((desc)[0] = cpu_to_le32(addr));		\
		if (HW_ADDR_LEN == 8)		 		\
			(desc)[1] = cpu_to_le32(((u64)addr) >> 32);	\
	} while(0)
#define desc_addr_get(desc)					\
	(le32_to_cpu((desc)[0]) | \
	(HW_ADDR_LEN == 8 ? ((dma_addr_t)le32_to_cpu((desc)[1]))<<32 : 0))

#define DESC_LINK		0
#define DESC_BUFPTR		(DESC_LINK + HW_ADDR_LEN/4)
#define DESC_CMDSTS		(DESC_BUFPTR + HW_ADDR_LEN/4)
#define DESC_EXTSTS		(DESC_CMDSTS + 4/4)

#define CMDSTS_OWN	0x80000000
#define CMDSTS_MORE	0x40000000
#define CMDSTS_INTR	0x20000000
#define CMDSTS_ERR	0x10000000
#define CMDSTS_OK	0x08000000
#define CMDSTS_RUNT	0x00200000
#define CMDSTS_LEN_MASK	0x0000ffff

#define CMDSTS_DEST_MASK	0x01800000
#define CMDSTS_DEST_SELF	0x00800000
#define CMDSTS_DEST_MULTI	0x01000000

#define DESC_SIZE	8		/* Should be cache line sized */

struct rx_info {
	spinlock_t	lock;
	int		up;
	long		idle;

	struct sk_buff	*skbs[NR_RX_DESC];

	__le32		*next_rx_desc;
	u16		next_rx, next_empty;

	__le32		*descs;
	dma_addr_t	phy_descs;
};


struct ns83820 {
	struct net_device_stats	stats;
	u8			__iomem *base;

	struct pci_dev		*pci_dev;
	struct net_device	*ndev;

#ifdef NS83820_VLAN_ACCEL_SUPPORT
	struct vlan_group	*vlgrp;
#endif

	struct rx_info		rx_info;
	struct tasklet_struct	rx_tasklet;

	unsigned		ihr;
	struct work_struct	tq_refill;

	/* protects everything below.  irqsave when using. */
	spinlock_t		misc_lock;

	u32			CFG_cache;

	u32			MEAR_cache;
	u32			IMR_cache;

	unsigned		linkstate;

	spinlock_t	tx_lock;

	u16		tx_done_idx;
	u16		tx_idx;
	volatile u16	tx_free_idx;	/* idx of free desc chain */
	u16		tx_intr_idx;

	atomic_t	nr_tx_skbs;
	struct sk_buff	*tx_skbs[NR_TX_DESC];

	char		pad[16] __attribute__((aligned(16)));
	__le32		*tx_descs;
	dma_addr_t	tx_phy_descs;

	struct timer_list	tx_watchdog;
};

static inline struct ns83820 *PRIV(struct net_device *dev)
{
	return netdev_priv(dev);
}

#define __kick_rx(dev)	writel(CR_RXE, dev->base + CR)

static inline void kick_rx(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	dprintk("kick_rx: maybe kicking\n");
	if (test_and_clear_bit(0, &dev->rx_info.idle)) {
		dprintk("actually kicking\n");
		writel(dev->rx_info.phy_descs +
			(4 * DESC_SIZE * dev->rx_info.next_rx),
		       dev->base + RXDP);
		if (dev->rx_info.next_rx == dev->rx_info.next_empty)
			printk(KERN_DEBUG "%s: uh-oh: next_rx == next_empty???\n",
				ndev->name);
		__kick_rx(dev);
	}
}

//free = (tx_done_idx + NR_TX_DESC-2 - free_idx) % NR_TX_DESC
#define start_tx_okay(dev)	\
	(((NR_TX_DESC-2 + dev->tx_done_idx - dev->tx_free_idx) % NR_TX_DESC) > MIN_TX_DESC_FREE)


#ifdef NS83820_VLAN_ACCEL_SUPPORT
static void ns83820_vlan_rx_register(struct net_device *ndev, struct vlan_group *grp)
{
	struct ns83820 *dev = PRIV(ndev);

	spin_lock_irq(&dev->misc_lock);
	spin_lock(&dev->tx_lock);

	dev->vlgrp = grp;

	spin_unlock(&dev->tx_lock);
	spin_unlock_irq(&dev->misc_lock);
}

static void ns83820_vlan_rx_kill_vid(struct net_device *ndev, unsigned short vid)
{
	struct ns83820 *dev = PRIV(ndev);

	spin_lock_irq(&dev->misc_lock);
	spin_lock(&dev->tx_lock);
	if (dev->vlgrp)
		dev->vlgrp->vlan_devices[vid] = NULL;
	spin_unlock(&dev->tx_lock);
	spin_unlock_irq(&dev->misc_lock);
}
#endif

/* Packet Receiver
 *
 * The hardware supports linked lists of receive descriptors for
 * which ownership is transfered back and forth by means of an
 * ownership bit.  While the hardware does support the use of a
 * ring for receive descriptors, we only make use of a chain in
 * an attempt to reduce bus traffic under heavy load scenarios.
 * This will also make bugs a bit more obvious.  The current code
 * only makes use of a single rx chain; I hope to implement
 * priority based rx for version 1.0.  Goal: even under overload
 * conditions, still route realtime traffic with as low jitter as
 * possible.
 */
static inline void build_rx_desc(struct ns83820 *dev, __le32 *desc, dma_addr_t link, dma_addr_t buf, u32 cmdsts, u32 extsts)
{
	desc_addr_set(desc + DESC_LINK, link);
	desc_addr_set(desc + DESC_BUFPTR, buf);
	desc[DESC_EXTSTS] = cpu_to_le32(extsts);
	mb();
	desc[DESC_CMDSTS] = cpu_to_le32(cmdsts);
}

#define nr_rx_empty(dev) ((NR_RX_DESC-2 + dev->rx_info.next_rx - dev->rx_info.next_empty) % NR_RX_DESC)
static inline int ns83820_add_rx_skb(struct ns83820 *dev, struct sk_buff *skb)
{
	unsigned next_empty;
	u32 cmdsts;
	__le32 *sg;
	dma_addr_t buf;

	next_empty = dev->rx_info.next_empty;

	/* don't overrun last rx marker */
	if (unlikely(nr_rx_empty(dev) <= 2)) {
		kfree_skb(skb);
		return 1;
	}

#if 0
	dprintk("next_empty[%d] nr_used[%d] next_rx[%d]\n",
		dev->rx_info.next_empty,
		dev->rx_info.nr_used,
		dev->rx_info.next_rx
		);
#endif

	sg = dev->rx_info.descs + (next_empty * DESC_SIZE);
	BUG_ON(NULL != dev->rx_info.skbs[next_empty]);
	dev->rx_info.skbs[next_empty] = skb;

	dev->rx_info.next_empty = (next_empty + 1) % NR_RX_DESC;
	cmdsts = REAL_RX_BUF_SIZE | CMDSTS_INTR;
	buf = pci_map_single(dev->pci_dev, skb->data,
			     REAL_RX_BUF_SIZE, PCI_DMA_FROMDEVICE);
	build_rx_desc(dev, sg, 0, buf, cmdsts, 0);
	/* update link of previous rx */
	if (likely(next_empty != dev->rx_info.next_rx))
		dev->rx_info.descs[((NR_RX_DESC + next_empty - 1) % NR_RX_DESC) * DESC_SIZE] = cpu_to_le32(dev->rx_info.phy_descs + (next_empty * DESC_SIZE * 4));

	return 0;
}

static inline int rx_refill(struct net_device *ndev, gfp_t gfp)
{
	struct ns83820 *dev = PRIV(ndev);
	unsigned i;
	unsigned long flags = 0;

	if (unlikely(nr_rx_empty(dev) <= 2))
		return 0;

	dprintk("rx_refill(%p)\n", ndev);
	if (gfp == GFP_ATOMIC)
		spin_lock_irqsave(&dev->rx_info.lock, flags);
	for (i=0; i<NR_RX_DESC; i++) {
		struct sk_buff *skb;
		long res;
		/* extra 16 bytes for alignment */
		skb = __dev_alloc_skb(REAL_RX_BUF_SIZE+16, gfp);
		if (unlikely(!skb))
			break;

		res = (long)skb->data & 0xf;
		res = 0x10 - res;
		res &= 0xf;
		skb_reserve(skb, res);

		skb->dev = ndev;
		if (gfp != GFP_ATOMIC)
			spin_lock_irqsave(&dev->rx_info.lock, flags);
		res = ns83820_add_rx_skb(dev, skb);
		if (gfp != GFP_ATOMIC)
			spin_unlock_irqrestore(&dev->rx_info.lock, flags);
		if (res) {
			i = 1;
			break;
		}
	}
	if (gfp == GFP_ATOMIC)
		spin_unlock_irqrestore(&dev->rx_info.lock, flags);

	return i ? 0 : -ENOMEM;
}

static void FASTCALL(rx_refill_atomic(struct net_device *ndev));
static void fastcall rx_refill_atomic(struct net_device *ndev)
{
	rx_refill(ndev, GFP_ATOMIC);
}

/* REFILL */
static inline void queue_refill(struct work_struct *work)
{
	struct ns83820 *dev = container_of(work, struct ns83820, tq_refill);
	struct net_device *ndev = dev->ndev;

	rx_refill(ndev, GFP_KERNEL);
	if (dev->rx_info.up)
		kick_rx(ndev);
}

static inline void clear_rx_desc(struct ns83820 *dev, unsigned i)
{
	build_rx_desc(dev, dev->rx_info.descs + (DESC_SIZE * i), 0, 0, CMDSTS_OWN, 0);
}

static void FASTCALL(phy_intr(struct net_device *ndev));
static void fastcall phy_intr(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	static const char *speeds[] = { "10", "100", "1000", "1000(?)", "1000F" };
	u32 cfg, new_cfg;
	u32 tbisr, tanar, tanlpar;
	int speed, fullduplex, newlinkstate;

	cfg = readl(dev->base + CFG) ^ SPDSTS_POLARITY;

	if (dev->CFG_cache & CFG_TBI_EN) {
		/* we have an optical transceiver */
		tbisr = readl(dev->base + TBISR);
		tanar = readl(dev->base + TANAR);
		tanlpar = readl(dev->base + TANLPAR);
		dprintk("phy_intr: tbisr=%08x, tanar=%08x, tanlpar=%08x\n",
			tbisr, tanar, tanlpar);

		if ( (fullduplex = (tanlpar & TANAR_FULL_DUP)
		      && (tanar & TANAR_FULL_DUP)) ) {

			/* both of us are full duplex */
			writel(readl(dev->base + TXCFG)
			       | TXCFG_CSI | TXCFG_HBI | TXCFG_ATP,
			       dev->base + TXCFG);
			writel(readl(dev->base + RXCFG) | RXCFG_RX_FD,
			       dev->base + RXCFG);
			/* Light up full duplex LED */
			writel(readl(dev->base + GPIOR) | GPIOR_GP1_OUT,
			       dev->base + GPIOR);

		} else if(((tanlpar & TANAR_HALF_DUP)
			   && (tanar & TANAR_HALF_DUP))
			|| ((tanlpar & TANAR_FULL_DUP)
			    && (tanar & TANAR_HALF_DUP))
			|| ((tanlpar & TANAR_HALF_DUP)
			    && (tanar & TANAR_FULL_DUP))) {

			/* one or both of us are half duplex */
			writel((readl(dev->base + TXCFG)
				& ~(TXCFG_CSI | TXCFG_HBI)) | TXCFG_ATP,
			       dev->base + TXCFG);
			writel(readl(dev->base + RXCFG) & ~RXCFG_RX_FD,
			       dev->base + RXCFG);
			/* Turn off full duplex LED */
			writel(readl(dev->base + GPIOR) & ~GPIOR_GP1_OUT,
			       dev->base + GPIOR);
		}

		speed = 4; /* 1000F */

	} else {
		/* we have a copper transceiver */
		new_cfg = dev->CFG_cache & ~(CFG_SB | CFG_MODE_1000 | CFG_SPDSTS);

		if (cfg & CFG_SPDSTS1)
			new_cfg |= CFG_MODE_1000;
		else
			new_cfg &= ~CFG_MODE_1000;

		speed = ((cfg / CFG_SPDSTS0) & 3);
		fullduplex = (cfg & CFG_DUPSTS);

		if (fullduplex) {
			new_cfg |= CFG_SB;
			writel(readl(dev->base + TXCFG)
					| TXCFG_CSI | TXCFG_HBI,
			       dev->base + TXCFG);
			writel(readl(dev->base + RXCFG) | RXCFG_RX_FD,
			       dev->base + RXCFG);
		} else {
			writel(readl(dev->base + TXCFG)
					& ~(TXCFG_CSI | TXCFG_HBI),
			       dev->base + TXCFG);
			writel(readl(dev->base + RXCFG) & ~(RXCFG_RX_FD),
			       dev->base + RXCFG);
		}

		if ((cfg & CFG_LNKSTS) &&
		    ((new_cfg ^ dev->CFG_cache) != 0)) {
			writel(new_cfg, dev->base + CFG);
			dev->CFG_cache = new_cfg;
		}

		dev->CFG_cache &= ~CFG_SPDSTS;
		dev->CFG_cache |= cfg & CFG_SPDSTS;
	}

	newlinkstate = (cfg & CFG_LNKSTS) ? LINK_UP : LINK_DOWN;

	if (newlinkstate & LINK_UP
	    && dev->linkstate != newlinkstate) {
		netif_start_queue(ndev);
		netif_wake_queue(ndev);
		printk(KERN_INFO "%s: link now %s mbps, %s duplex and up.\n",
			ndev->name,
			speeds[speed],
			fullduplex ? "full" : "half");
	} else if (newlinkstate & LINK_DOWN
		   && dev->linkstate != newlinkstate) {
		netif_stop_queue(ndev);
		printk(KERN_INFO "%s: link now down.\n", ndev->name);
	}

	dev->linkstate = newlinkstate;
}

static int ns83820_setup_rx(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	unsigned i;
	int ret;

	dprintk("ns83820_setup_rx(%p)\n", ndev);

	dev->rx_info.idle = 1;
	dev->rx_info.next_rx = 0;
	dev->rx_info.next_rx_desc = dev->rx_info.descs;
	dev->rx_info.next_empty = 0;

	for (i=0; i<NR_RX_DESC; i++)
		clear_rx_desc(dev, i);

	writel(0, dev->base + RXDP_HI);
	writel(dev->rx_info.phy_descs, dev->base + RXDP);

	ret = rx_refill(ndev, GFP_KERNEL);
	if (!ret) {
		dprintk("starting receiver\n");
		/* prevent the interrupt handler from stomping on us */
		spin_lock_irq(&dev->rx_info.lock);

		writel(0x0001, dev->base + CCSR);
		writel(0, dev->base + RFCR);
		writel(0x7fc00000, dev->base + RFCR);
		writel(0xffc00000, dev->base + RFCR);

		dev->rx_info.up = 1;

		phy_intr(ndev);

		/* Okay, let it rip */
		spin_lock_irq(&dev->misc_lock);
		dev->IMR_cache |= ISR_PHY;
		dev->IMR_cache |= ISR_RXRCMP;
		//dev->IMR_cache |= ISR_RXERR;
		//dev->IMR_cache |= ISR_RXOK;
		dev->IMR_cache |= ISR_RXORN;
		dev->IMR_cache |= ISR_RXSOVR;
		dev->IMR_cache |= ISR_RXDESC;
		dev->IMR_cache |= ISR_RXIDLE;
		dev->IMR_cache |= ISR_TXDESC;
		dev->IMR_cache |= ISR_TXIDLE;

		writel(dev->IMR_cache, dev->base + IMR);
		writel(1, dev->base + IER);
		spin_unlock(&dev->misc_lock);

		kick_rx(ndev);

		spin_unlock_irq(&dev->rx_info.lock);
	}
	return ret;
}

static void ns83820_cleanup_rx(struct ns83820 *dev)
{
	unsigned i;
	unsigned long flags;

	dprintk("ns83820_cleanup_rx(%p)\n", dev);

	/* disable receive interrupts */
	spin_lock_irqsave(&dev->misc_lock, flags);
	dev->IMR_cache &= ~(ISR_RXOK | ISR_RXDESC | ISR_RXERR | ISR_RXEARLY | ISR_RXIDLE);
	writel(dev->IMR_cache, dev->base + IMR);
	spin_unlock_irqrestore(&dev->misc_lock, flags);

	/* synchronize with the interrupt handler and kill it */
	dev->rx_info.up = 0;
	synchronize_irq(dev->pci_dev->irq);

	/* touch the pci bus... */
	readl(dev->base + IMR);

	/* assumes the transmitter is already disabled and reset */
	writel(0, dev->base + RXDP_HI);
	writel(0, dev->base + RXDP);

	for (i=0; i<NR_RX_DESC; i++) {
		struct sk_buff *skb = dev->rx_info.skbs[i];
		dev->rx_info.skbs[i] = NULL;
		clear_rx_desc(dev, i);
		if (skb)
			kfree_skb(skb);
	}
}

static void FASTCALL(ns83820_rx_kick(struct net_device *ndev));
static void fastcall ns83820_rx_kick(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	/*if (nr_rx_empty(dev) >= NR_RX_DESC/4)*/ {
		if (dev->rx_info.up) {
			rx_refill_atomic(ndev);
			kick_rx(ndev);
		}
	}

	if (dev->rx_info.up && nr_rx_empty(dev) > NR_RX_DESC*3/4)
		schedule_work(&dev->tq_refill);
	else
		kick_rx(ndev);
	if (dev->rx_info.idle)
		printk(KERN_DEBUG "%s: BAD\n", ndev->name);
}

/* rx_irq
 *
 */
static void FASTCALL(rx_irq(struct net_device *ndev));
static void fastcall rx_irq(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	struct rx_info *info = &dev->rx_info;
	unsigned next_rx;
	int rx_rc, len;
	u32 cmdsts;
	__le32 *desc;
	unsigned long flags;
	int nr = 0;

	dprintk("rx_irq(%p)\n", ndev);
	dprintk("rxdp: %08x, descs: %08lx next_rx[%d]: %p next_empty[%d]: %p\n",
		readl(dev->base + RXDP),
		(long)(dev->rx_info.phy_descs),
		(int)dev->rx_info.next_rx,
		(dev->rx_info.descs + (DESC_SIZE * dev->rx_info.next_rx)),
		(int)dev->rx_info.next_empty,
		(dev->rx_info.descs + (DESC_SIZE * dev->rx_info.next_empty))
		);

	spin_lock_irqsave(&info->lock, flags);
	if (!info->up)
		goto out;

	dprintk("walking descs\n");
	next_rx = info->next_rx;
	desc = info->next_rx_desc;
	while ((CMDSTS_OWN & (cmdsts = le32_to_cpu(desc[DESC_CMDSTS]))) &&
	       (cmdsts != CMDSTS_OWN)) {
		struct sk_buff *skb;
		u32 extsts = le32_to_cpu(desc[DESC_EXTSTS]);
		dma_addr_t bufptr = desc_addr_get(desc + DESC_BUFPTR);

		dprintk("cmdsts: %08x\n", cmdsts);
		dprintk("link: %08x\n", cpu_to_le32(desc[DESC_LINK]));
		dprintk("extsts: %08x\n", extsts);

		skb = info->skbs[next_rx];
		info->skbs[next_rx] = NULL;
		info->next_rx = (next_rx + 1) % NR_RX_DESC;

		mb();
		clear_rx_desc(dev, next_rx);

		pci_unmap_single(dev->pci_dev, bufptr,
				 RX_BUF_SIZE, PCI_DMA_FROMDEVICE);
		len = cmdsts & CMDSTS_LEN_MASK;
#ifdef NS83820_VLAN_ACCEL_SUPPORT
		/* NH: As was mentioned below, this chip is kinda
		 * brain dead about vlan tag stripping.  Frames
		 * that are 64 bytes with a vlan header appended
		 * like arp frames, or pings, are flagged as Runts
		 * when the tag is stripped and hardware.  This
		 * also means that the OK bit in the descriptor
		 * is cleared when the frame comes in so we have
		 * to do a specific length check here to make sure
		 * the frame would have been ok, had we not stripped
		 * the tag.
		 */
		if (likely((CMDSTS_OK & cmdsts) ||
			((cmdsts & CMDSTS_RUNT) && len >= 56))) {
#else
		if (likely(CMDSTS_OK & cmdsts)) {
#endif
			skb_put(skb, len);
			if (unlikely(!skb))
				goto netdev_mangle_me_harder_failed;
			if (cmdsts & CMDSTS_DEST_MULTI)
				dev->stats.multicast ++;
			dev->stats.rx_packets ++;
			dev->stats.rx_bytes += len;
			if ((extsts & 0x002a0000) && !(extsts & 0x00540000)) {
				skb->ip_summed = CHECKSUM_UNNECESSARY;
			} else {
				skb->ip_summed = CHECKSUM_NONE;
			}
			skb->protocol = eth_type_trans(skb, ndev);
#ifdef NS83820_VLAN_ACCEL_SUPPORT
			if(extsts & EXTSTS_VPKT) {
				unsigned short tag;
				tag = ntohs(extsts & EXTSTS_VTG_MASK);
				rx_rc = vlan_hwaccel_rx(skb,dev->vlgrp,tag);
			} else {
				rx_rc = netif_rx(skb);
			}
#else
			rx_rc = netif_rx(skb);
#endif
			if (NET_RX_DROP == rx_rc) {
netdev_mangle_me_harder_failed:
				dev->stats.rx_dropped ++;
			}
		} else {
			kfree_skb(skb);
		}

		nr++;
		next_rx = info->next_rx;
		desc = info->descs + (DESC_SIZE * next_rx);
	}
	info->next_rx = next_rx;
	info->next_rx_desc = info->descs + (DESC_SIZE * next_rx);

out:
	if (0 && !nr) {
		Dprintk("dazed: cmdsts_f: %08x\n", cmdsts);
	}

	spin_unlock_irqrestore(&info->lock, flags);
}

static void rx_action(unsigned long _dev)
{
	struct net_device *ndev = (void *)_dev;
	struct ns83820 *dev = PRIV(ndev);
	rx_irq(ndev);
	writel(ihr, dev->base + IHR);

	spin_lock_irq(&dev->misc_lock);
	dev->IMR_cache |= ISR_RXDESC;
	writel(dev->IMR_cache, dev->base + IMR);
	spin_unlock_irq(&dev->misc_lock);

	rx_irq(ndev);
	ns83820_rx_kick(ndev);
}

/* Packet Transmit code
 */
static inline void kick_tx(struct ns83820 *dev)
{
	dprintk("kick_tx(%p): tx_idx=%d free_idx=%d\n",
		dev, dev->tx_idx, dev->tx_free_idx);
	writel(CR_TXE, dev->base + CR);
}

/* No spinlock needed on the transmit irq path as the interrupt handler is
 * serialized.
 */
static void do_tx_done(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	u32 cmdsts, tx_done_idx;
	__le32 *desc;

	dprintk("do_tx_done(%p)\n", ndev);
	tx_done_idx = dev->tx_done_idx;
	desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);

	dprintk("tx_done_idx=%d free_idx=%d cmdsts=%08x\n",
		tx_done_idx, dev->tx_free_idx, le32_to_cpu(desc[DESC_CMDSTS]));
	while ((tx_done_idx != dev->tx_free_idx) &&
	       !(CMDSTS_OWN & (cmdsts = le32_to_cpu(desc[DESC_CMDSTS]))) ) {
		struct sk_buff *skb;
		unsigned len;
		dma_addr_t addr;

		if (cmdsts & CMDSTS_ERR)
			dev->stats.tx_errors ++;
		if (cmdsts & CMDSTS_OK)
			dev->stats.tx_packets ++;
		if (cmdsts & CMDSTS_OK)
			dev->stats.tx_bytes += cmdsts & 0xffff;

		dprintk("tx_done_idx=%d free_idx=%d cmdsts=%08x\n",
			tx_done_idx, dev->tx_free_idx, cmdsts);
		skb = dev->tx_skbs[tx_done_idx];
		dev->tx_skbs[tx_done_idx] = NULL;
		dprintk("done(%p)\n", skb);

		len = cmdsts & CMDSTS_LEN_MASK;
		addr = desc_addr_get(desc + DESC_BUFPTR);
		if (skb) {
			pci_unmap_single(dev->pci_dev,
					addr,
					len,
					PCI_DMA_TODEVICE);
			dev_kfree_skb_irq(skb);
			atomic_dec(&dev->nr_tx_skbs);
		} else
			pci_unmap_page(dev->pci_dev,
					addr,
					len,
					PCI_DMA_TODEVICE);

		tx_done_idx = (tx_done_idx + 1) % NR_TX_DESC;
		dev->tx_done_idx = tx_done_idx;
		desc[DESC_CMDSTS] = cpu_to_le32(0);
		mb();
		desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);
	}

	/* Allow network stack to resume queueing packets after we've
	 * finished transmitting at least 1/4 of the packets in the queue.
	 */
	if (netif_queue_stopped(ndev) && start_tx_okay(dev)) {
		dprintk("start_queue(%p)\n", ndev);
		netif_start_queue(ndev);
		netif_wake_queue(ndev);
	}
}

static void ns83820_cleanup_tx(struct ns83820 *dev)
{
	unsigned i;

	for (i=0; i<NR_TX_DESC; i++) {
		struct sk_buff *skb = dev->tx_skbs[i];
		dev->tx_skbs[i] = NULL;
		if (skb) {
			__le32 *desc = dev->tx_descs + (i * DESC_SIZE);
			pci_unmap_single(dev->pci_dev,
					desc_addr_get(desc + DESC_BUFPTR),
					le32_to_cpu(desc[DESC_CMDSTS]) & CMDSTS_LEN_MASK,
					PCI_DMA_TODEVICE);
			dev_kfree_skb_irq(skb);
			atomic_dec(&dev->nr_tx_skbs);
		}
	}

	memset(dev->tx_descs, 0, NR_TX_DESC * DESC_SIZE * 4);
}

/* transmit routine.  This code relies on the network layer serializing
 * its calls in, but will run happily in parallel with the interrupt
 * handler.  This code currently has provisions for fragmenting tx buffers
 * while trying to track down a bug in either the zero copy code or
 * the tx fifo (hence the MAX_FRAG_LEN).
 */
static int ns83820_hard_start_xmit(struct sk_buff *skb, struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	u32 free_idx, cmdsts, extsts;
	int nr_free, nr_frags;
	unsigned tx_done_idx, last_idx;
	dma_addr_t buf;
	unsigned len;
	skb_frag_t *frag;
	int stopped = 0;
	int do_intr = 0;
	volatile __le32 *first_desc;

	dprintk("ns83820_hard_start_xmit\n");

	nr_frags =  skb_shinfo(skb)->nr_frags;
again:
	if (unlikely(dev->CFG_cache & CFG_LNKSTS)) {
		netif_stop_queue(ndev);
		if (unlikely(dev->CFG_cache & CFG_LNKSTS))
			return 1;
		netif_start_queue(ndev);
	}

	last_idx = free_idx = dev->tx_free_idx;
	tx_done_idx = dev->tx_done_idx;
	nr_free = (tx_done_idx + NR_TX_DESC-2 - free_idx) % NR_TX_DESC;
	nr_free -= 1;
	if (nr_free <= nr_frags) {
		dprintk("stop_queue - not enough(%p)\n", ndev);
		netif_stop_queue(ndev);

		/* Check again: we may have raced with a tx done irq */
		if (dev->tx_done_idx != tx_done_idx) {
			dprintk("restart queue(%p)\n", ndev);
			netif_start_queue(ndev);
			goto again;
		}
		return 1;
	}

	if (free_idx == dev->tx_intr_idx) {
		do_intr = 1;
		dev->tx_intr_idx = (dev->tx_intr_idx + NR_TX_DESC/4) % NR_TX_DESC;
	}

	nr_free -= nr_frags;
	if (nr_free < MIN_TX_DESC_FREE) {
		dprintk("stop_queue - last entry(%p)\n", ndev);
		netif_stop_queue(ndev);
		stopped = 1;
	}

	frag = skb_shinfo(skb)->frags;
	if (!nr_frags)
		frag = NULL;
	extsts = 0;
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		extsts |= EXTSTS_IPPKT;
		if (IPPROTO_TCP == skb->nh.iph->protocol)
			extsts |= EXTSTS_TCPPKT;
		else if (IPPROTO_UDP == skb->nh.iph->protocol)
			extsts |= EXTSTS_UDPPKT;
	}

#ifdef NS83820_VLAN_ACCEL_SUPPORT
	if(vlan_tx_tag_present(skb)) {
		/* fetch the vlan tag info out of the
		 * ancilliary data if the vlan code
		 * is using hw vlan acceleration
		 */
		short tag = vlan_tx_tag_get(skb);
		extsts |= (EXTSTS_VPKT | htons(tag));
	}
#endif

	len = skb->len;
	if (nr_frags)
		len -= skb->data_len;
	buf = pci_map_single(dev->pci_dev, skb->data, len, PCI_DMA_TODEVICE);

	first_desc = dev->tx_descs + (free_idx * DESC_SIZE);

	for (;;) {
		volatile __le32 *desc = dev->tx_descs + (free_idx * DESC_SIZE);

		dprintk("frag[%3u]: %4u @ 0x%08Lx\n", free_idx, len,
			(unsigned long long)buf);
		last_idx = free_idx;
		free_idx = (free_idx + 1) % NR_TX_DESC;
		desc[DESC_LINK] = cpu_to_le32(dev->tx_phy_descs + (free_idx * DESC_SIZE * 4));
		desc_addr_set(desc + DESC_BUFPTR, buf);
		desc[DESC_EXTSTS] = cpu_to_le32(extsts);

		cmdsts = ((nr_frags) ? CMDSTS_MORE : do_intr ? CMDSTS_INTR : 0);
		cmdsts |= (desc == first_desc) ? 0 : CMDSTS_OWN;
		cmdsts |= len;
		desc[DESC_CMDSTS] = cpu_to_le32(cmdsts);

		if (!nr_frags)
			break;

		buf = pci_map_page(dev->pci_dev, frag->page,
				   frag->page_offset,
				   frag->size, PCI_DMA_TODEVICE);
		dprintk("frag: buf=%08Lx  page=%08lx offset=%08lx\n",
			(long long)buf, (long) page_to_pfn(frag->page),
			frag->page_offset);
		len = frag->size;
		frag++;
		nr_frags--;
	}
	dprintk("done pkt\n");

	spin_lock_irq(&dev->tx_lock);
	dev->tx_skbs[last_idx] = skb;
	first_desc[DESC_CMDSTS] |= cpu_to_le32(CMDSTS_OWN);
	dev->tx_free_idx = free_idx;
	atomic_inc(&dev->nr_tx_skbs);
	spin_unlock_irq(&dev->tx_lock);

	kick_tx(dev);

	/* Check again: we may have raced with a tx done irq */
	if (stopped && (dev->tx_done_idx != tx_done_idx) && start_tx_okay(dev))
		netif_start_queue(ndev);

	/* set the transmit start time to catch transmit timeouts */
	ndev->trans_start = jiffies;
	return 0;
}

static void ns83820_update_stats(struct ns83820 *dev)
{
	u8 __iomem *base = dev->base;

	/* the DP83820 will freeze counters, so we need to read all of them */
	dev->stats.rx_errors		+= readl(base + 0x60) & 0xffff;
	dev->stats.rx_crc_errors	+= readl(base + 0x64) & 0xffff;
	dev->stats.rx_missed_errors	+= readl(base + 0x68) & 0xffff;
	dev->stats.rx_frame_errors	+= readl(base + 0x6c) & 0xffff;
	/*dev->stats.rx_symbol_errors +=*/ readl(base + 0x70);
	dev->stats.rx_length_errors	+= readl(base + 0x74) & 0xffff;
	dev->stats.rx_length_errors	+= readl(base + 0x78) & 0xffff;
	/*dev->stats.rx_badopcode_errors += */ readl(base + 0x7c);
	/*dev->stats.rx_pause_count += */  readl(base + 0x80);
	/*dev->stats.tx_pause_count += */  readl(base + 0x84);
	dev->stats.tx_carrier_errors	+= readl(base + 0x88) & 0xff;
}

static struct net_device_stats *ns83820_get_stats(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);

	/* somewhat overkill */
	spin_lock_irq(&dev->misc_lock);
	ns83820_update_stats(dev);
	spin_unlock_irq(&dev->misc_lock);

	return &dev->stats;
}

static void ns83820_get_drvinfo(struct net_device *ndev, struct ethtool_drvinfo *info)
{
	struct ns83820 *dev = PRIV(ndev);
	strcpy(info->driver, "ns83820");
	strcpy(info->version, VERSION);
	strcpy(info->bus_info, pci_name(dev->pci_dev));
}

static u32 ns83820_get_link(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	u32 cfg = readl(dev->base + CFG) ^ SPDSTS_POLARITY;
	return cfg & CFG_LNKSTS ? 1 : 0;
}

static const struct ethtool_ops ops = {
	.get_drvinfo = ns83820_get_drvinfo,
	.get_link = ns83820_get_link
};

/* this function is called in irq context from the ISR */
static void ns83820_mib_isr(struct ns83820 *dev)
{
	unsigned long flags;
	spin_lock_irqsave(&dev->misc_lock, flags);
	ns83820_update_stats(dev);
	spin_unlock_irqrestore(&dev->misc_lock, flags);
}

static void ns83820_do_isr(struct net_device *ndev, u32 isr);
static irqreturn_t ns83820_irq(int foo, void *data)
{
	struct net_device *ndev = data;
	struct ns83820 *dev = PRIV(ndev);
	u32 isr;
	dprintk("ns83820_irq(%p)\n", ndev);

	dev->ihr = 0;

	isr = readl(dev->base + ISR);
	dprintk("irq: %08x\n", isr);
	ns83820_do_isr(ndev, isr);
	return IRQ_HANDLED;
}

static void ns83820_do_isr(struct net_device *ndev, u32 isr)
{
	struct ns83820 *dev = PRIV(ndev);
	unsigned long flags;

#ifdef DEBUG
	if (isr & ~(ISR_PHY | ISR_RXDESC | ISR_RXEARLY | ISR_RXOK | ISR_RXERR | ISR_TXIDLE | ISR_TXOK | ISR_TXDESC))
		Dprintk("odd isr? 0x%08x\n", isr);
#endif

	if (ISR_RXIDLE & isr) {
		dev->rx_info.idle = 1;
		Dprintk("oh dear, we are idle\n");
		ns83820_rx_kick(ndev);
	}

	if ((ISR_RXDESC | ISR_RXOK) & isr) {
		prefetch(dev->rx_info.next_rx_desc);

		spin_lock_irqsave(&dev->misc_lock, flags);
		dev->IMR_cache &= ~(ISR_RXDESC | ISR_RXOK);
		writel(dev->IMR_cache, dev->base + IMR);
		spin_unlock_irqrestore(&dev->misc_lock, flags);

		tasklet_schedule(&dev->rx_tasklet);
		//rx_irq(ndev);
		//writel(4, dev->base + IHR);
	}

	if ((ISR_RXIDLE | ISR_RXORN | ISR_RXDESC | ISR_RXOK | ISR_RXERR) & isr)
		ns83820_rx_kick(ndev);

	if (unlikely(ISR_RXSOVR & isr)) {
		//printk("overrun: rxsovr\n");
		dev->stats.rx_fifo_errors ++;
	}

	if (unlikely(ISR_RXORN & isr)) {
		//printk("overrun: rxorn\n");
		dev->stats.rx_fifo_errors ++;
	}

	if ((ISR_RXRCMP & isr) && dev->rx_info.up)
		writel(CR_RXE, dev->base + CR);

	if (ISR_TXIDLE & isr) {
		u32 txdp;
		txdp = readl(dev->base + TXDP);
		dprintk("txdp: %08x\n", txdp);
		txdp -= dev->tx_phy_descs;
		dev->tx_idx = txdp / (DESC_SIZE * 4);
		if (dev->tx_idx >= NR_TX_DESC) {
			printk(KERN_ALERT "%s: BUG -- txdp out of range\n", ndev->name);
			dev->tx_idx = 0;
		}
		/* The may have been a race between a pci originated read
		 * and the descriptor update from the cpu.  Just in case,
		 * kick the transmitter if the hardware thinks it is on a
		 * different descriptor than we are.
		 */
		if (dev->tx_idx != dev->tx_free_idx)
			kick_tx(dev);
	}

	/* Defer tx ring processing until more than a minimum amount of
	 * work has accumulated
	 */
	if ((ISR_TXDESC | ISR_TXIDLE | ISR_TXOK | ISR_TXERR) & isr) {
		spin_lock_irqsave(&dev->tx_lock, flags);
		do_tx_done(ndev);
		spin_unlock_irqrestore(&dev->tx_lock, flags);

		/* Disable TxOk if there are no outstanding tx packets.
		 */
		if ((dev->tx_done_idx == dev->tx_free_idx) &&
		    (dev->IMR_cache & ISR_TXOK)) {
			spin_lock_irqsave(&dev->misc_lock, flags);
			dev->IMR_cache &= ~ISR_TXOK;
			writel(dev->IMR_cache, dev->base + IMR);
			spin_unlock_irqrestore(&dev->misc_lock, flags);
		}
	}

	/* The TxIdle interrupt can come in before the transmit has
	 * completed.  Normally we reap packets off of the combination
	 * of TxDesc and TxIdle and leave TxOk disabled (since it
	 * occurs on every packet), but when no further irqs of this
	 * nature are expected, we must enable TxOk.
	 */
	if ((ISR_TXIDLE & isr) && (dev->tx_done_idx != dev->tx_free_idx)) {
		spin_lock_irqsave(&dev->misc_lock, flags);
		dev->IMR_cache |= ISR_TXOK;
		writel(dev->IMR_cache, dev->base + IMR);
		spin_unlock_irqrestore(&dev->misc_lock, flags);
	}

	/* MIB interrupt: one of the statistics counters is about to overflow */
	if (unlikely(ISR_MIB & isr))
		ns83820_mib_isr(dev);

	/* PHY: Link up/down/negotiation state change */
	if (unlikely(ISR_PHY & isr))
		phy_intr(ndev);

#if 0	/* Still working on the interrupt mitigation strategy */
	if (dev->ihr)
		writel(dev->ihr, dev->base + IHR);
#endif
}

static void ns83820_do_reset(struct ns83820 *dev, u32 which)
{
	Dprintk("resetting chip...\n");
	writel(which, dev->base + CR);
	do {
		schedule();
	} while (readl(dev->base + CR) & which);
	Dprintk("okay!\n");
}

static int ns83820_stop(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);

	/* FIXME: protect against interrupt handler? */
	del_timer_sync(&dev->tx_watchdog);

	/* disable interrupts */
	writel(0, dev->base + IMR);
	writel(0, dev->base + IER);
	readl(dev->base + IER);

	dev->rx_info.up = 0;
	synchronize_irq(dev->pci_dev->irq);

	ns83820_do_reset(dev, CR_RST);

	synchronize_irq(dev->pci_dev->irq);

	spin_lock_irq(&dev->misc_lock);
	dev->IMR_cache &= ~(ISR_TXURN | ISR_TXIDLE | ISR_TXERR | ISR_TXDESC | ISR_TXOK);
	spin_unlock_irq(&dev->misc_lock);

	ns83820_cleanup_rx(dev);
	ns83820_cleanup_tx(dev);

	return 0;
}

static void ns83820_tx_timeout(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
        u32 tx_done_idx;
	__le32 *desc;
	unsigned long flags;

	spin_lock_irqsave(&dev->tx_lock, flags);

	tx_done_idx = dev->tx_done_idx;
	desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);

	printk(KERN_INFO "%s: tx_timeout: tx_done_idx=%d free_idx=%d cmdsts=%08x\n",
		ndev->name,
		tx_done_idx, dev->tx_free_idx, le32_to_cpu(desc[DESC_CMDSTS]));

#if defined(DEBUG)
	{
		u32 isr;
		isr = readl(dev->base + ISR);
		printk("irq: %08x imr: %08x\n", isr, dev->IMR_cache);
		ns83820_do_isr(ndev, isr);
	}
#endif

	do_tx_done(ndev);

	tx_done_idx = dev->tx_done_idx;
	desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);

	printk(KERN_INFO "%s: after: tx_done_idx=%d free_idx=%d cmdsts=%08x\n",
		ndev->name,
		tx_done_idx, dev->tx_free_idx, le32_to_cpu(desc[DESC_CMDSTS]));

	spin_unlock_irqrestore(&dev->tx_lock, flags);
}

static void ns83820_tx_watch(unsigned long data)
{
	struct net_device *ndev = (void *)data;
	struct ns83820 *dev = PRIV(ndev);

#if defined(DEBUG)
	printk("ns83820_tx_watch: %u %u %d\n",
		dev->tx_done_idx, dev->tx_free_idx, atomic_read(&dev->nr_tx_skbs)
		);
#endif

	if (time_after(jiffies, ndev->trans_start + 1*HZ) &&
	    dev->tx_done_idx != dev->tx_free_idx) {
		printk(KERN_DEBUG "%s: ns83820_tx_watch: %u %u %d\n",
			ndev->name,
			dev->tx_done_idx, dev->tx_free_idx,
			atomic_read(&dev->nr_tx_skbs));
		ns83820_tx_timeout(ndev);
	}

	mod_timer(&dev->tx_watchdog, jiffies + 2*HZ);
}

static int ns83820_open(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	unsigned i;
	u32 desc;
	int ret;

	dprintk("ns83820_open\n");

	writel(0, dev->base + PQCR);

	ret = ns83820_setup_rx(ndev);
	if (ret)
		goto failed;

	memset(dev->tx_descs, 0, 4 * NR_TX_DESC * DESC_SIZE);
	for (i=0; i<NR_TX_DESC; i++) {
		dev->tx_descs[(i * DESC_SIZE) + DESC_LINK]
				= cpu_to_le32(
				  dev->tx_phy_descs
				  + ((i+1) % NR_TX_DESC) * DESC_SIZE * 4);
	}

	dev->tx_idx = 0;
	dev->tx_done_idx = 0;
	desc = dev->tx_phy_descs;
	writel(0, dev->base + TXDP_HI);
	writel(desc, dev->base + TXDP);

	init_timer(&dev->tx_watchdog);
	dev->tx_watchdog.data = (unsigned long)ndev;
	dev->tx_watchdog.function = ns83820_tx_watch;
	mod_timer(&dev->tx_watchdog, jiffies + 2*HZ);

	netif_start_queue(ndev);	/* FIXME: wait for phy to come up */

	return 0;

failed:
	ns83820_stop(ndev);
	return ret;
}

static void ns83820_getmac(struct ns83820 *dev, u8 *mac)
{
	unsigned i;
	for (i=0; i<3; i++) {
		u32 data;

		/* Read from the perfect match memory: this is loaded by
		 * the chip from the EEPROM via the EELOAD self test.
		 */
		writel(i*2, dev->base + RFCR);
		data = readl(dev->base + RFDR);

		*mac++ = data;
		*mac++ = data >> 8;
	}
}

static int ns83820_change_mtu(struct net_device *ndev, int new_mtu)
{
	if (new_mtu > RX_BUF_SIZE)
		return -EINVAL;
	ndev->mtu = new_mtu;
	return 0;
}

static void ns83820_set_multicast(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	u8 __iomem *rfcr = dev->base + RFCR;
	u32 and_mask = 0xffffffff;
	u32 or_mask = 0;
	u32 val;

	if (ndev->flags & IFF_PROMISC)
		or_mask |= RFCR_AAU | RFCR_AAM;
	else
		and_mask &= ~(RFCR_AAU | RFCR_AAM);

	if (ndev->flags & IFF_ALLMULTI)
		or_mask |= RFCR_AAM;
	else
		and_mask &= ~RFCR_AAM;

	spin_lock_irq(&dev->misc_lock);
	val = (readl(rfcr) & and_mask) | or_mask;
	/* Ramit : RFCR Write Fix doc says RFEN must be 0 modify other bits */
	writel(val & ~RFCR_RFEN, rfcr);
	writel(val, rfcr);
	spin_unlock_irq(&dev->misc_lock);
}

static void ns83820_run_bist(struct net_device *ndev, const char *name, u32 enable, u32 done, u32 fail)
{
	struct ns83820 *dev = PRIV(ndev);
	int timed_out = 0;
	unsigned long start;
	u32 status;
	int loops = 0;

	dprintk("%s: start %s\n", ndev->name, name);

	start = jiffies;

	writel(enable, dev->base + PTSCR);
	for (;;) {
		loops++;
		status = readl(dev->base + PTSCR);
		if (!(status & enable))
			break;
		if (status & done)
			break;
		if (status & fail)
			break;
		if (time_after_eq(jiffies, start + HZ)) {
			timed_out = 1;
			break;
		}
		schedule_timeout_uninterruptible(1);
	}

	if (status & fail)
		printk(KERN_INFO "%s: %s failed! (0x%08x & 0x%08x)\n",
			ndev->name, name, status, fail);
	else if (timed_out)
		printk(KERN_INFO "%s: run_bist %s timed out! (%08x)\n",
			ndev->name, name, status);

	dprintk("%s: done %s in %d loops\n", ndev->name, name, loops);
}

#ifdef PHY_CODE_IS_FINISHED
static void ns83820_mii_write_bit(struct ns83820 *dev, int bit)
{
	/* drive MDC low */
	dev->MEAR_cache &= ~MEAR_MDC;
	writel(dev->MEAR_cache, dev->base + MEAR);
	readl(dev->base + MEAR);

	/* enable output, set bit */
	dev->MEAR_cache |= MEAR_MDDIR;
	if (bit)
		dev->MEAR_cache |= MEAR_MDIO;
	else
		dev->MEAR_cache &= ~MEAR_MDIO;

	/* set the output bit */
	writel(dev->MEAR_cache, dev->base + MEAR);
	readl(dev->base + MEAR);

	/* Wait.  Max clock rate is 2.5MHz, this way we come in under 1MHz */
	udelay(1);

	/* drive MDC high causing the data bit to be latched */
	dev->MEAR_cache |= MEAR_MDC;
	writel(dev->MEAR_cache, dev->base + MEAR);
	readl(dev->base + MEAR);

	/* Wait again... */
	udelay(1);
}

static int ns83820_mii_read_bit(struct ns83820 *dev)
{
	int bit;

	/* drive MDC low, disable output */
	dev->MEAR_cache &= ~MEAR_MDC;
	dev->MEAR_cache &= ~MEAR_MDDIR;
	writel(dev->MEAR_cache, dev->base + MEAR);
	readl(dev->base + MEAR);

	/* Wait.  Max clock rate is 2.5MHz, this way we come in under 1MHz */
	udelay(1);

	/* drive MDC high causing the data bit to be latched */
	bit = (readl(dev->base + MEAR) & MEAR_MDIO) ? 1 : 0;
	dev->MEAR_cache |= MEAR_MDC;
	writel(dev->MEAR_cache, dev->base + MEAR);

	/* Wait again... */
	udelay(1);

	return bit;
}

static unsigned ns83820_mii_read_reg(struct ns83820 *dev, unsigned phy, unsigned reg)
{
	unsigned data = 0;
	int i;

	/* read some garbage so that we eventually sync up */
	for (i=0; i<64; i++)
		ns83820_mii_read_bit(dev);

	ns83820_mii_write_bit(dev, 0);	/* start */
	ns83820_mii_write_bit(dev, 1);
	ns83820_mii_write_bit(dev, 1);	/* opcode read */
	ns83820_mii_write_bit(dev, 0);

	/* write out the phy address: 5 bits, msb first */
	for (i=0; i<5; i++)
		ns83820_mii_write_bit(dev, phy & (0x10 >> i));

	/* write out the register address, 5 bits, msb first */
	for (i=0; i<5; i++)
		ns83820_mii_write_bit(dev, reg & (0x10 >> i));

	ns83820_mii_read_bit(dev);	/* turn around cycles */
	ns83820_mii_read_bit(dev);

	/* read in the register data, 16 bits msb first */
	for (i=0; i<16; i++) {
		data <<= 1;
		data |= ns83820_mii_read_bit(dev);
	}

	return data;
}

static unsigned ns83820_mii_write_reg(struct ns83820 *dev, unsigned phy, unsigned reg, unsigned data)
{
	int i;

	/* read some garbage so that we eventually sync up */
	for (i=0; i<64; i++)
		ns83820_mii_read_bit(dev);

	ns83820_mii_write_bit(dev, 0);	/* start */
	ns83820_mii_write_bit(dev, 1);
	ns83820_mii_write_bit(dev, 0);	/* opcode read */
	ns83820_mii_write_bit(dev, 1);

	/* write out the phy address: 5 bits, msb first */
	for (i=0; i<5; i++)
		ns83820_mii_write_bit(dev, phy & (0x10 >> i));

	/* write out the register address, 5 bits, msb first */
	for (i=0; i<5; i++)
		ns83820_mii_write_bit(dev, reg & (0x10 >> i));

	ns83820_mii_read_bit(dev);	/* turn around cycles */
	ns83820_mii_read_bit(dev);

	/* read in the register data, 16 bits msb first */
	for (i=0; i<16; i++)
		ns83820_mii_write_bit(dev, (data >> (15 - i)) & 1);

	return data;
}

static void ns83820_probe_phy(struct net_device *ndev)
{
	struct ns83820 *dev = PRIV(ndev);
	static int first;
	int i;
#define MII_PHYIDR1	0x02
#define MII_PHYIDR2	0x03

#if 0
	if (!first) {
		unsigned tmp;
		ns83820_mii_read_reg(dev, 1, 0x09);
		ns83820_mii_write_reg(dev, 1, 0x10, 0x0d3e);

		tmp = ns83820_mii_read_reg(dev, 1, 0x00);
		ns83820_mii_write_reg(dev, 1, 0x00, tmp | 0x8000);
		udelay(1300);
		ns83820_mii_read_reg(dev, 1, 0x09);
	}
#endif
	first = 1;

	for (i=1; i<2; i++) {
		int j;
		unsigned a, b;
		a = ns83820_mii_read_reg(dev, i, MII_PHYIDR1);
		b = ns83820_mii_read_reg(dev, i, MII_PHYIDR2);

		//printk("%s: phy %d: 0x%04x 0x%04x\n",
		//	ndev->name, i, a, b);

		for (j=0; j<0x16; j+=4) {
			dprintk("%s: [0x%02x] %04x %04x %04x %04x\n",
				ndev->name, j,
				ns83820_mii_read_reg(dev, i, 0 + j),
				ns83820_mii_read_reg(dev, i, 1 + j),
				ns83820_mii_read_reg(dev, i, 2 + j),
				ns83820_mii_read_reg(dev, i, 3 + j)
				);
		}
	}
	{
		unsigned a, b;
		/* read firmware version: memory addr is 0x8402 and 0x8403 */
		ns83820_mii_write_reg(dev, 1, 0x16, 0x000d);
		ns83820_mii_write_reg(dev, 1, 0x1e, 0x810e);
		a = ns83820_mii_read_reg(dev, 1, 0x1d);

		ns83820_mii_write_reg(dev, 1, 0x16, 0x000d);
		ns83820_mii_write_reg(dev, 1, 0x1e, 0x810e);
		b = ns83820_mii_read_reg(dev, 1, 0x1d);
		dprintk("version: 0x%04x 0x%04x\n", a, b);
	}
}
#endif

static int __devinit ns83820_init_one(struct pci_dev *pci_dev, const struct pci_device_id *id)
{
	struct net_device *ndev;
	struct ns83820 *dev;
	long addr;
	int err;
	int using_dac = 0;

	/* See if we can set the dma mask early on; failure is fatal. */
	if (sizeof(dma_addr_t) == 8 &&
	 	!pci_set_dma_mask(pci_dev, DMA_64BIT_MASK)) {
		using_dac = 1;
	} else if (!pci_set_dma_mask(pci_dev, DMA_32BIT_MASK)) {
		using_dac = 0;
	} else {
		dev_warn(&pci_dev->dev, "pci_set_dma_mask failed!\n");
		return -ENODEV;
	}

	ndev = alloc_etherdev(sizeof(struct ns83820));
	dev = PRIV(ndev);
	dev->ndev = ndev;
	err = -ENOMEM;
	if (!dev)
		goto out;

	spin_lock_init(&dev->rx_info.lock);
	spin_lock_init(&dev->tx_lock);
	spin_lock_init(&dev->misc_lock);
	dev->pci_dev = pci_dev;

	SET_MODULE_OWNER(ndev);
	SET_NETDEV_DEV(ndev, &pci_dev->dev);

	INIT_WORK(&dev->tq_refill, queue_refill);
	tasklet_init(&dev->rx_tasklet, rx_action, (unsigned long)ndev);

	err = pci_enable_device(pci_dev);
	if (err) {
		dev_info(&pci_dev->dev, "pci_enable_dev failed: %d\n", err);
		goto out_free;
	}

	pci_set_master(pci_dev);
	addr = pci_resource_start(pci_dev, 1);
	dev->base = ioremap_nocache(addr, PAGE_SIZE);
	dev->tx_descs = pci_alloc_consistent(pci_dev,
			4 * DESC_SIZE * NR_TX_DESC, &dev->tx_phy_descs);
	dev->rx_info.descs = pci_alloc_consistent(pci_dev,
			4 * DESC_SIZE * NR_RX_DESC, &dev->rx_info.phy_descs);
	err = -ENOMEM;
	if (!dev->base || !dev->tx_descs || !dev->rx_info.descs)
		goto out_disable;

	dprintk("%p: %08lx  %p: %08lx\n",
		dev->tx_descs, (long)dev->tx_phy_descs,
		dev->rx_info.descs, (long)dev->rx_info.phy_descs);

	/* disable interrupts */
	writel(0, dev->base + IMR);
	writel(0, dev->base + IER);
	readl(dev->base + IER);

	dev->IMR_cache = 0;

	err = request_irq(pci_dev->irq, ns83820_irq, IRQF_SHARED,
			  DRV_NAME, ndev);
	if (err) {
		dev_info(&pci_dev->dev, "unable to register irq %d, err %d\n",
			pci_dev->irq, err);
		goto out_disable;
	}

	/*
	 * FIXME: we are holding rtnl_lock() over obscenely long area only
	 * because some of the setup code uses dev->name.  It's Wrong(tm) -
	 * we should be using driver-specific names for all that stuff.
	 * For now that will do, but we really need to come back and kill
	 * most of the dev_alloc_name() users later.
	 */
	rtnl_lock();
	err = dev_alloc_name(ndev, ndev->name);
	if (err < 0) {
		dev_info(&pci_dev->dev, "unable to get netdev name: %d\n", err);
		goto out_free_irq;
	}

	printk("%s: ns83820.c: 0x22c: %08x, subsystem: %04x:%04x\n",
		ndev->name, le32_to_cpu(readl(dev->base + 0x22c)),
		pci_dev->subsystem_vendor, pci_dev->subsystem_device);

	ndev->open = ns83820_open;
	ndev->stop = ns83820_stop;
	ndev->hard_start_xmit = ns83820_hard_start_xmit;
	ndev->get_stats = ns83820_get_stats;
	ndev->change_mtu = ns83820_change_mtu;
	ndev->set_multicast_list = ns83820_set_multicast;
	SET_ETHTOOL_OPS(ndev, &ops);
	ndev->tx_timeout = ns83820_tx_timeout;
	ndev->watchdog_timeo = 5 * HZ;
	pci_set_drvdata(pci_dev, ndev);

	ns83820_do_reset(dev, CR_RST);

	/* Must reset the ram bist before running it */
	writel(PTSCR_RBIST_RST, dev->base + PTSCR);
	ns83820_run_bist(ndev, "sram bist",   PTSCR_RBIST_EN,
			 PTSCR_RBIST_DONE, PTSCR_RBIST_FAIL);
	ns83820_run_bist(ndev, "eeprom bist", PTSCR_EEBIST_EN, 0,
			 PTSCR_EEBIST_FAIL);
	ns83820_run_bist(ndev, "eeprom load", PTSCR_EELOAD_EN, 0, 0);

	/* I love config registers */
	dev->CFG_cache = readl(dev->base + CFG);

	if ((dev->CFG_cache & CFG_PCI64_DET)) {
		printk(KERN_INFO "%s: detected 64 bit PCI data bus.\n",
			ndev->name);
		/*dev->CFG_cache |= CFG_DATA64_EN;*/
		if (!(dev->CFG_cache & CFG_DATA64_EN))
			printk(KERN_INFO "%s: EEPROM did not enable 64 bit bus.  Disabled.\n",
				ndev->name);
	} else
		dev->CFG_cache &= ~(CFG_DATA64_EN);

	dev->CFG_cache &= (CFG_TBI_EN  | CFG_MRM_DIS   | CFG_MWI_DIS |
			   CFG_T64ADDR | CFG_DATA64_EN | CFG_EXT_125 |
			   CFG_M64ADDR);
	dev->CFG_cache |= CFG_PINT_DUPSTS | CFG_PINT_LNKSTS | CFG_PINT_SPDSTS |
			  CFG_EXTSTS_EN   | CFG_EXD         | CFG_PESEL;
	dev->CFG_cache |= CFG_REQALG;
	dev->CFG_cache |= CFG_POW;
	dev->CFG_cache |= CFG_TMRTEST;

	/* When compiled with 64 bit addressing, we must always enable
	 * the 64 bit descriptor format.
	 */
	if (sizeof(dma_addr_t) == 8)
		dev->CFG_cache |= CFG_M64ADDR;
	if (using_dac)
		dev->CFG_cache |= CFG_T64ADDR;

	/* Big endian mode does not seem to do what the docs suggest */
	dev->CFG_cache &= ~CFG_BEM;

	/* setup optical transceiver if we have one */
	if (dev->CFG_cache & CFG_TBI_EN) {
		printk(KERN_INFO "%s: enabling optical transceiver\n",
			ndev->name);
		writel(readl(dev->base + GPIOR) | 0x3e8, dev->base + GPIOR);

		/* setup auto negotiation feature advertisement */
		writel(readl(dev->base + TANAR)
		       | TANAR_HALF_DUP | TANAR_FULL_DUP,
		       dev->base + TANAR);

		/* start auto negotiation */
		writel(TBICR_MR_AN_ENABLE | TBICR_MR_RESTART_AN,
		       dev->base + TBICR);
		writel(TBICR_MR_AN_ENABLE, dev->base + TBICR);
		dev->linkstate = LINK_AUTONEGOTIATE;

		dev->CFG_cache |= CFG_MODE_1000;
	}

	writel(dev->CFG_cache, dev->base + CFG);
	dprintk("CFG: %08x\n", dev->CFG_cache);

	if (reset_phy) {
		printk(KERN_INFO "%s: resetting phy\n", ndev->name);
		writel(dev->CFG_cache | CFG_PHY_RST, dev->base + CFG);
		msleep(10);
		writel(dev->CFG_cache, dev->base + CFG);
	}

#if 0	/* Huh?  This sets the PCI latency register.  Should be done via
	 * the PCI layer.  FIXME.
	 */
	if (readl(dev->base + SRR))
		writel(readl(dev->base+0x20c) | 0xfe00, dev->base + 0x20c);
#endif

	/* Note!  The DMA burst size interacts with packet
	 * transmission, such that the largest packet that
	 * can be transmitted is 8192 - FLTH - burst size.
	 * If only the transmit fifo was larger...
	 */
	/* Ramit : 1024 DMA is not a good idea, it ends up banging
	 * some DELL and COMPAQ SMP systems */
	writel(TXCFG_CSI | TXCFG_HBI | TXCFG_ATP | TXCFG_MXDMA512
		| ((1600 / 32) * 0x100),
		dev->base + TXCFG);

	/* Flush the interrupt holdoff timer */
	writel(0x000, dev->base + IHR);
	writel(0x100, dev->base + IHR);
	writel(0x000, dev->base + IHR);

	/* Set Rx to full duplex, don't accept runt, errored, long or length
	 * range errored packets.  Use 512 byte DMA.
	 */
	/* Ramit : 1024 DMA is not a good idea, it ends up banging
	 * some DELL and COMPAQ SMP systems
	 * Turn on ALP, only we are accpeting Jumbo Packets */
	writel(RXCFG_AEP | RXCFG_ARP | RXCFG_AIRL | RXCFG_RX_FD
		| RXCFG_STRIPCRC
		//| RXCFG_ALP
		| (RXCFG_MXDMA512) | 0, dev->base + RXCFG);

	/* Disable priority queueing */
	writel(0, dev->base + PQCR);

	/* Enable IP checksum validation and detetion of VLAN headers.
	 * Note: do not set the reject options as at least the 0x102
	 * revision of the chip does not properly accept IP fragments
	 * at least for UDP.
	 */
	/* Ramit : Be sure to turn on RXCFG_ARP if VLAN's are enabled, since
	 * the MAC it calculates the packetsize AFTER stripping the VLAN
	 * header, and if a VLAN Tagged packet of 64 bytes is received (like
	 * a ping with a VLAN header) then the card, strips the 4 byte VLAN
	 * tag and then checks the packet size, so if RXCFG_ARP is not enabled,
	 * it discrards it!.  These guys......
	 * also turn on tag stripping if hardware acceleration is enabled
	 */
#ifdef NS83820_VLAN_ACCEL_SUPPORT
#define VRCR_INIT_VALUE (VRCR_IPEN|VRCR_VTDEN|VRCR_VTREN)
#else
#define VRCR_INIT_VALUE (VRCR_IPEN|VRCR_VTDEN)
#endif
	writel(VRCR_INIT_VALUE, dev->base + VRCR);

	/* Enable per-packet TCP/UDP/IP checksumming
	 * and per packet vlan tag insertion if
	 * vlan hardware acceleration is enabled
	 */
#ifdef NS83820_VLAN_ACCEL_SUPPORT
#define VTCR_INIT_VALUE (VTCR_PPCHK|VTCR_VPPTI)
#else
#define VTCR_INIT_VALUE VTCR_PPCHK
#endif
	writel(VTCR_INIT_VALUE, dev->base + VTCR);

	/* Ramit : Enable async and sync pause frames */
	/* writel(0, dev->base + PCR); */
	writel((PCR_PS_MCAST | PCR_PS_DA | PCR_PSEN | PCR_FFLO_4K |
		PCR_FFHI_8K | PCR_STLO_4 | PCR_STHI_8 | PCR_PAUSE_CNT),
		dev->base + PCR);

	/* Disable Wake On Lan */
	writel(0, dev->base + WCSR);

	ns83820_getmac(dev, ndev->dev_addr);

	/* Yes, we support dumb IP checksum on transmit */
	ndev->features |= NETIF_F_SG;
	ndev->features |= NETIF_F_IP_CSUM;

#ifdef NS83820_VLAN_ACCEL_SUPPORT
	/* We also support hardware vlan acceleration */
	ndev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX;
	ndev->vlan_rx_register = ns83820_vlan_rx_register;
	ndev->vlan_rx_kill_vid = ns83820_vlan_rx_kill_vid;
#endif

	if (using_dac) {
		printk(KERN_INFO "%s: using 64 bit addressing.\n",
			ndev->name);
		ndev->features |= NETIF_F_HIGHDMA;
	}

	printk(KERN_INFO "%s: ns83820 v" VERSION ": DP83820 v%u.%u: %02x:%02x:%02x:%02x:%02x:%02x io=0x%08lx irq=%d f=%s\n",
		ndev->name,
		(unsigned)readl(dev->base + SRR) >> 8,
		(unsigned)readl(dev->base + SRR) & 0xff,
		ndev->dev_addr[0], ndev->dev_addr[1],
		ndev->dev_addr[2], ndev->dev_addr[3],
		ndev->dev_addr[4], ndev->dev_addr[5],
		addr, pci_dev->irq,
		(ndev->features & NETIF_F_HIGHDMA) ? "h,sg" : "sg"
		);

#ifdef PHY_CODE_IS_FINISHED
	ns83820_probe_phy(ndev);
#endif

	err = register_netdevice(ndev);
	if (err) {
		printk(KERN_INFO "ns83820: unable to register netdev: %d\n", err);
		goto out_cleanup;
	}
	rtnl_unlock();

	return 0;

out_cleanup:
	writel(0, dev->base + IMR);	/* paranoia */
	writel(0, dev->base + IER);
	readl(dev->base + IER);
out_free_irq:
	rtnl_unlock();
	free_irq(pci_dev->irq, ndev);
out_disable:
	if (dev->base)
		iounmap(dev->base);
	pci_free_consistent(pci_dev, 4 * DESC_SIZE * NR_TX_DESC, dev->tx_descs, dev->tx_phy_descs);
	pci_free_consistent(pci_dev, 4 * DESC_SIZE * NR_RX_DESC, dev->rx_info.descs, dev->rx_info.phy_descs);
	pci_disable_device(pci_dev);
out_free:
	free_netdev(ndev);
	pci_set_drvdata(pci_dev, NULL);
out:
	return err;
}

static void __devexit ns83820_remove_one(struct pci_dev *pci_dev)
{
	struct net_device *ndev = pci_get_drvdata(pci_dev);
	struct ns83820 *dev = PRIV(ndev); /* ok even if NULL */

	if (!ndev)			/* paranoia */
		return;

	writel(0, dev->base + IMR);	/* paranoia */
	writel(0, dev->base + IER);
	readl(dev->base + IER);

	unregister_netdev(ndev);
	free_irq(dev->pci_dev->irq, ndev);
	iounmap(dev->base);
	pci_free_consistent(dev->pci_dev, 4 * DESC_SIZE * NR_TX_DESC,
			dev->tx_descs, dev->tx_phy_descs);
	pci_free_consistent(dev->pci_dev, 4 * DESC_SIZE * NR_RX_DESC,
			dev->rx_info.descs, dev->rx_info.phy_descs);
	pci_disable_device(dev->pci_dev);
	free_netdev(ndev);
	pci_set_drvdata(pci_dev, NULL);
}

static struct pci_device_id ns83820_pci_tbl[] = {
	{ 0x100b, 0x0022, PCI_ANY_ID, PCI_ANY_ID, 0, .driver_data = 0, },
	{ 0, },
};

static struct pci_driver driver = {
	.name		= "ns83820",
	.id_table	= ns83820_pci_tbl,
	.probe		= ns83820_init_one,
	.remove		= __devexit_p(ns83820_remove_one),
#if 0	/* FIXME: implement */
	.suspend	= ,
	.resume		= ,
#endif
};


static int __init ns83820_init(void)
{
	printk(KERN_INFO "ns83820.c: National Semiconductor DP83820 10/100/1000 driver.\n");
	return pci_register_driver(&driver);
}

static void __exit ns83820_exit(void)
{
	pci_unregister_driver(&driver);
}

MODULE_AUTHOR("Benjamin LaHaise <[e-mail]>");
MODULE_DESCRIPTION("National Semiconductor DP83820 10/100/1000 driver");
MODULE_LICENSE("GPL");

MODULE_DEVICE_TABLE(pci, ns83820_pci_tbl);

module_param(lnksts, int, 0);
MODULE_PARM_DESC(lnksts, "Polarity of LNKSTS bit");

module_param(ihr, int, 0);
MODULE_PARM_DESC(ihr, "Time in 100 us increments to delay interrupts (range 0-127)");

module_param(reset_phy, int, 0);
MODULE_PARM_DESC(reset_phy, "Set to 1 to reset the PHY on startup");

module_init(ns83820_init);
module_exit(ns83820_exit);