[netdrvr] add driver "meth", for SGI O2 MACE fast eth

drivers/net/Kconfig

@@ -511,6 +511,10 @@ config SGI_IOC3_ETH
 	  the Ethernet-HOWTO, available from
 	  <http://www.tldp.org/docs.html#howto>.
 
+config SGI_O2MACE_ETH
+	tristate "SGI O2 MACE Fast Ethernet support"
+	depends on NET_ETHERNET && SGI_IP32=y
+
 config STNIC
 	tristate "National DP83902AV  support"
 	depends on NET_ETHERNET && SUPERH

drivers/net/Makefile

@@ -117,6 +117,7 @@ obj-$(CONFIG_SUN3_82586) += sun3_82586.o
 obj-$(CONFIG_SUN3LANCE) += sun3lance.o
 obj-$(CONFIG_DEFXX) += defxx.o
 obj-$(CONFIG_SGISEEQ) += sgiseeq.o
+obj-$(CONFIG_SGI_O2MACE_ETH) += meth.o
 obj-$(CONFIG_AT1700) += at1700.o
 obj-$(CONFIG_FMV18X) += fmv18x.o
 obj-$(CONFIG_EL1) += 3c501.o

drivers/net/meth.c

+/*
+ * meth.c -- O2 Builtin 10/100 Ethernet driver
+ *
+ * Copyright (C) 2001 Ilya Volynets
+ *
+ *	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.
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+
+#include <linux/sched.h>
+#include <linux/kernel.h> /* printk() */
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/errno.h>  /* error codes */
+#include <linux/types.h>  /* size_t */
+#include <linux/interrupt.h> /* mark_bh */
+#include <linux/pci.h>
+
+#include <linux/in.h>
+#include <linux/netdevice.h>   /* struct device, and other headers */
+#include <linux/etherdevice.h> /* eth_type_trans */
+#include <linux/ip.h>          /* struct iphdr */
+#include <linux/tcp.h>         /* struct tcphdr */
+#include <linux/skbuff.h>
+#include <linux/mii.h> /*MII definitions */
+
+#include <asm/ip32/crime.h>
+#include <asm/ip32/mace.h>
+#include <asm/ip32/ip32_ints.h>
+
+#include "meth.h"
+
+#include <linux/in6.h>
+#include <asm/checksum.h>
+
+#ifndef MFE_DEBUG
+#define MFE_DEBUG 0
+#endif
+
+#if MFE_DEBUG>=1
+#define DPRINTK(str,args...) printk (KERN_DEBUG "meth(%ld): %s: " str, jiffies, __FUNCTION__ , ## args)
+#define MFE_RX_DEBUG 2
+#else
+#define DPRINTK(str,args...)
+#define MFE_RX_DEBUG 0
+#endif
+
+
+static const char *version="meth.c: Ilya Volynets (ilya@theIlya.com)";
+static const char *meth_str="SGI O2 Fast Ethernet";
+MODULE_AUTHOR("Ilya Volynets");
+MODULE_DESCRIPTION("SGI O2 Builtin Fast Ethernet driver");
+
+/* This is a load-time options */
+/*static int eth = 0;
+MODULE_PARM(eth, "i");*/
+
+#define HAVE_TX_TIMEOUT
+/* The maximum time waited (in jiffies) before assuming a Tx failed. (400ms) */
+#define TX_TIMEOUT (400*HZ/1000)
+
+#ifdef HAVE_TX_TIMEOUT
+static int timeout = TX_TIMEOUT;
+MODULE_PARM(timeout, "i");
+#endif
+
+int meth_eth;
+
+/*
+ * This structure is private to each device. It is used to pass
+ * packets in and out, so there is place for a packet
+ */
+
+typedef struct meth_private {
+    struct net_device_stats stats;
+	volatile struct meth_regs *regs;
+	u64 mode; /* in-memory copy of MAC control register */
+	int  phy_addr; /* address of phy, used by mdio_* functions, initialized in mdio_probe*/
+	tx_packet *tx_ring;
+	dma_addr_t tx_ring_dma;
+	int free_space;
+	struct sk_buff *tx_skbs[TX_RING_ENTRIES];
+	dma_addr_t      tx_skb_dmas[TX_RING_ENTRIES];
+	int tx_read,tx_write;
+	int tx_count;
+
+	rx_packet *rx_ring[RX_RING_ENTRIES];
+	dma_addr_t rx_ring_dmas[RX_RING_ENTRIES];
+	int rx_write;
+
+    spinlock_t meth_lock;
+} meth_private;
+
+extern struct net_device meth_devs[];
+void meth_tx_timeout (struct net_device *dev);
+void meth_interrupt(int irq, void *dev_id, struct pt_regs *pregs);
+        
+/* global, initialized in ip32-setup.c */
+char o2meth_eaddr[8]={0,0,0,0,0,0,0,0};
+
+static inline void load_eaddr(struct net_device *dev,
+			      volatile struct meth_regs *regs)
+{
+	int i;
+	DPRINTK("Loading MAC Address: %02x:%02x:%02x:%02x:%02x:%02x\n",
+		(int)o2meth_eaddr[0]&0xFF,(int)o2meth_eaddr[1]&0xFF,(int)o2meth_eaddr[2]&0xFF,
+		(int)o2meth_eaddr[3]&0xFF,(int)o2meth_eaddr[4]&0xFF,(int)o2meth_eaddr[5]&0xFF);
+	//memcpy(dev->dev_addr,o2meth_eaddr+2,6);
+	for (i=0; i<6; i++)
+		dev->dev_addr[i]=o2meth_eaddr[i];
+	regs->mac_addr= //dev->dev_addr[0]|(dev->dev_addr[1]<<8)|
+					//dev->dev_addr[2]<<16|(dev->dev_addr[3]<<24)|
+					//dev->dev_addr[4]<<32|(dev->dev_addr[5]<<40);
+	(*(u64*)o2meth_eaddr)>>16;
+	DPRINTK("MAC, finally is %0lx\n",regs->mac_addr);
+}
+
+/*
+ *Waits for BUSY status of mdio bus to clear
+ */
+#define WAIT_FOR_PHY(___regs, ___rval)			\
+	while((___rval=___regs->phy_data)&MDIO_BUSY){	\
+		udelay(25);								\
+	}
+/*read phy register, return value read */
+static int mdio_read(meth_private *priv,int phyreg)
+{
+	volatile meth_regs* regs=priv->regs;
+	volatile u32 rval;
+	WAIT_FOR_PHY(regs,rval)
+	regs->phy_registers=(priv->phy_addr<<5)|(phyreg&0x1f);
+	udelay(25);
+	regs->phy_trans_go=1;
+	udelay(25);
+	WAIT_FOR_PHY(regs,rval)
+	return rval&MDIO_DATA_MASK;
+}
+
+/*write phy register */
+static void mdio_write(meth_private* priv,int pfyreg,int val)
+{
+	volatile meth_regs* regs=priv->regs;
+	int rval;
+///	DPRINTK("Trying to write value %i to reguster %i\n",val, pfyreg);
+	spin_lock_irq(&priv->meth_lock);
+	WAIT_FOR_PHY(regs,rval)
+	regs->phy_registers=(priv->phy_addr<<5)|(pfyreg&0x1f);
+	regs->phy_data=val;
+	udelay(25);
+	WAIT_FOR_PHY(regs,rval)
+	spin_unlock_irq(&priv->meth_lock);
+}
+
+/* Modify phy register using given mask and value */
+static void mdio_update(meth_private* priv,int phyreg, int val, int mask)
+{
+	int rval;
+	DPRINTK("RMW value %i to PHY register %i with mask %i\n",val,phyreg,mask);
+	rval=mdio_read(priv,phyreg);
+	rval=(rval&~mask)|(val&mask);
+	mdio_write(priv,phyreg,rval);
+}
+
+/* handle errata data on MDIO bus */
+//static void mdio_errata(meth_private *priv)
+//{
+	/* Hmmm... what the hell is phyerrata? does it come from sys init parameters in IRIX */
+//}
+static int mdio_probe(meth_private *priv)
+{
+	int i, p2, p3;
+	DPRINTK("Detecting PHY kind\n");
+	/* check if phy is detected already */
+	if(priv->phy_addr>=0&&priv->phy_addr<32)
+		return 0;
+	spin_lock_irq(&priv->meth_lock);
+	for (i=0;i<32;++i){
+		priv->phy_addr=(char)i;
+		p2=mdio_read(priv,2);
+#ifdef MFE_DEBUG
+		p3=mdio_read(priv,3);
+		switch ((p2<<12)|(p3>>4)){
+			case PHY_QS6612X:
+				DPRINTK("PHY is QS6612X\n");
+				break;
+			case PHY_ICS1889:
+				DPRINTK("PHY is ICS1889\n");
+				break;
+			case PHY_ICS1890:
+				DPRINTK("PHY is ICS1890\n");
+				break;
+			case PHY_DP83840:
+				DPRINTK("PHY is DP83840\n");
+				break;
+		}
+#endif
+		if(p2!=0xffff&&p2!=0x0000){
+			DPRINTK("PHY code: %x\n",(p2<<12)|(p3>>4));
+			break;
+		}
+	}
+	spin_unlock_irq(&priv->meth_lock);
+	if(priv->phy_addr<32) {
+		return 0;
+	}
+	DPRINTK("Oopsie! PHY is not known!\n");
+	priv->phy_addr=-1;
+	return -ENODEV;
+}
+
+static void meth_check_link(struct net_device *dev)
+{
+	struct meth_private *priv = (struct meth_private *) dev->priv;
+	int mii_partner = mdio_read(priv, 5);
+	int mii_advertising = mdio_read(priv, 4);
+	int negotiated = mii_advertising & mii_partner;
+	int duplex, speed;
+
+	if (mii_partner == 0xffff)
+		return;
+
+	duplex = ((negotiated & 0x0100) || (negotiated & 0x01C0) == 0x0040)?METH_PHY_FDX:0;
+	speed = (negotiated & 0x0380)?METH_100MBIT:0;
+
+	if ((priv->mode & METH_PHY_FDX) ^ duplex)
+	{
+		DPRINTK("Setting %s-duplex\n", duplex ? "full" : "half");
+		if (duplex)
+			priv->mode |= METH_PHY_FDX;
+		else
+			priv->mode &= ~METH_PHY_FDX;
+		priv->regs->mac_ctrl = priv->mode;
+	}
+
+	if ((priv->mode & METH_100MBIT) ^ speed)
+	{
+		DPRINTK("Setting %dMbs mode\n", speed ? 100 : 10);
+		if (duplex)
+			priv->mode |= METH_100MBIT;
+		else
+			priv->mode &= ~METH_100MBIT;
+		priv->regs->mac_ctrl = priv->mode;
+	}
+}
+
+
+static int meth_init_tx_ring(meth_private *priv)
+{
+	/* Init TX ring */
+	DPRINTK("Initializing TX ring\n");
+	priv->tx_ring = (tx_packet *) pci_alloc_consistent(NULL,TX_RING_BUFFER_SIZE,&(priv->tx_ring_dma));
+	if(!priv->tx_ring)
+		return -ENOMEM;
+	memset(priv->tx_ring, 0, TX_RING_BUFFER_SIZE);
+	priv->tx_count = priv->tx_read = priv->tx_write = 0;
+	priv->regs->tx_ring_base = priv->tx_ring_dma;
+	priv->free_space = TX_RING_ENTRIES;
+	/* Now init skb save area */
+	memset(priv->tx_skbs,0,sizeof(priv->tx_skbs));
+	memset(priv->tx_skb_dmas,0,sizeof(priv->tx_skb_dmas));
+	DPRINTK("Done with TX ring init\n");
+	return 0;
+}
+
+static int meth_init_rx_ring(meth_private *priv)
+{
+	int i;
+	DPRINTK("Initializing RX ring\n");
+	for(i=0;i<RX_RING_ENTRIES;i++){
+		DPRINTK("\t1:\t%i\t",i);
+		/*if(!(priv->rx_ring[i]=get_free_page(GFP_KERNEL)))
+			return -ENOMEM;
+		DPRINTK("\t2:\t%i\n",i);*/
+		priv->rx_ring[i]=(rx_packet*)pci_alloc_consistent(NULL,METH_RX_BUFF_SIZE,&(priv->rx_ring_dmas[i]));
+		/* I'll need to re-sync it after each RX */
+		DPRINTK("\t%p\n",priv->rx_ring[i]);
+		priv->regs->rx_fifo=priv->rx_ring_dmas[i];
+	}
+	priv->rx_write = 0;
+	DPRINTK("Done with RX ring\n");
+	return 0;
+}
+static void meth_free_tx_ring(meth_private *priv)
+{
+	int i;
+
+	/* Remove any pending skb */
+	for (i = 0; i < TX_RING_ENTRIES; i++) {
+	  if (priv->tx_skbs[i])
+		dev_kfree_skb(priv->tx_skbs[i]);
+		priv->tx_skbs[i] = NULL;
+	}
+	pci_free_consistent(NULL,
+			    TX_RING_BUFFER_SIZE,
+			    priv->tx_ring,
+			    priv->tx_ring_dma);
+}
+static void meth_free_rx_ring(meth_private *priv)
+{
+	int i;
+
+	for(i=0;i<RX_RING_ENTRIES;i++)
+		pci_free_consistent(NULL,
+				    METH_RX_BUFF_SIZE,
+				    priv->rx_ring[i],
+				    priv->rx_ring_dmas[i]);
+}
+
+int meth_reset(struct net_device *dev)
+{
+	struct meth_private *priv = (struct meth_private *) dev->priv;
+
+	/* Reset card */
+	priv->regs->mac_ctrl = SGI_MAC_RESET;
+	priv->regs->mac_ctrl = 0;
+	udelay(25);
+	DPRINTK("MAC control after reset: %016lx\n", priv->regs->mac_ctrl);
+
+	/* Load ethernet address */
+	load_eaddr(dev, priv->regs);
+	/* Should load some "errata", but later */
+	
+	/* Check for device */
+	if(mdio_probe(priv) < 0) {
+		DPRINTK("Unable to find PHY\n");
+		return -ENODEV;
+	}
+
+	/* Initial mode -- 10|Half-duplex|Accept normal packets */
+	priv->mode=METH_ACCEPT_MCAST|METH_DEFAULT_IPG;
+	if(dev->flags | IFF_PROMISC)
+		priv->mode |= METH_PROMISC;
+	priv->regs->mac_ctrl = priv->mode;
+
+	/* Autonegociate speed and duplex mode */
+	meth_check_link(dev);
+
+	/* Now set dma control, but don't enable DMA, yet */
+	priv->regs->dma_ctrl= (4 << METH_RX_OFFSET_SHIFT) |
+		              (RX_RING_ENTRIES << METH_RX_DEPTH_SHIFT);
+
+	return(0);
+}
+
+/*============End Helper Routines=====================*/
+
+/*
+ * Open and close
+ */
+
+int meth_open(struct net_device *dev)
+{
+	meth_private *priv=dev->priv;
+	volatile meth_regs *regs=priv->regs;
+
+	MOD_INC_USE_COUNT;
+
+	/* Start DMA */
+	regs->dma_ctrl|=
+	        METH_DMA_TX_EN|/*METH_DMA_TX_INT_EN|*/
+		METH_DMA_RX_EN|METH_DMA_RX_INT_EN;
+
+	if(request_irq(dev->irq,meth_interrupt,SA_SHIRQ,meth_str,dev)){
+		printk(KERN_ERR "%s: Can't get irq %d\n", dev->name, dev->irq);
+		return -EAGAIN;
+	}
+	netif_start_queue(dev);
+	DPRINTK("Opened... DMA control=0x%08lx\n", regs->dma_ctrl);
+	return 0;
+}
+
+int meth_release(struct net_device *dev)
+{
+    netif_stop_queue(dev); /* can't transmit any more */
+	/* shut down dma */
+	((meth_private*)(dev->priv))->regs->dma_ctrl&=
+		~(METH_DMA_TX_EN|METH_DMA_TX_INT_EN|
+		METH_DMA_RX_EN|METH_DMA_RX_INT_EN);
+	free_irq(dev->irq, dev);
+    MOD_DEC_USE_COUNT;
+    return 0;
+}
+
+/*
+ * Configuration changes (passed on by ifconfig)
+ */
+int meth_config(struct net_device *dev, struct ifmap *map)
+{
+    if (dev->flags & IFF_UP) /* can't act on a running interface */
+        return -EBUSY;
+
+    /* Don't allow changing the I/O address */
+    if (map->base_addr != dev->base_addr) {
+        printk(KERN_WARNING "meth: Can't change I/O address\n");
+        return -EOPNOTSUPP;
+    }
+
+    /* Allow changing the IRQ */
+    if (map->irq != dev->irq) {
+        printk(KERN_WARNING "meth: Can't change IRQ\n");
+        return -EOPNOTSUPP;
+    }
+	DPRINTK("Configured\n");
+
+    /* ignore other fields */
+    return 0;
+}
+
+/*
+ * Receive a packet: retrieve, encapsulate and pass over to upper levels
+ */
+void meth_rx(struct net_device* dev)
+{
+    struct sk_buff *skb;
+    struct meth_private *priv = (struct meth_private *) dev->priv;
+	rx_packet *rxb;
+	DPRINTK("RX...\n");
+	// TEMP	while((rxb=priv->rx_ring[priv->rx_write])->status.raw&0x8000000000000000){
+	while((rxb=priv->rx_ring[priv->rx_write])->status.raw&0x8000000000000000){
+	        int len=rxb->status.parsed.rx_len - 4; /* omit CRC */
+		DPRINTK("(%i)\n",priv->rx_write);
+		/* length sanity check */
+		if(len < 60 || len > 1518) {
+		  printk(KERN_DEBUG "%s: bogus packet size: %d, status=%#2x.\n",
+			 dev->name, priv->rx_write, rxb->status.raw);
+		  priv->stats.rx_errors++;
+		  priv->stats.rx_length_errors++;
+		}
+		if(!(rxb->status.raw&METH_RX_STATUS_ERRORS)){
+			skb=alloc_skb(len+2,GFP_ATOMIC);/* Should be atomic -- we are in interrupt */
+			if(!skb){
+				/* Ouch! No memory! Drop packet on the floor */
+				DPRINTK("!!!>>>Ouch! Not enough Memory for RX buffer!\n");
+				priv->stats.rx_dropped++;
+			} else {
+				skb_reserve(skb, 2); /* align IP on 16B boundary */  
+    			memcpy(skb_put(skb, len), rxb->buf, len);
+			    /* Write metadata, and then pass to the receive level */
+			    skb->dev = dev;
+			    skb->protocol = eth_type_trans(skb, dev);
+				//skb->ip_summed = CHECKSUM_UNNECESSARY; /* don't check it */
+			   
+				DPRINTK("passing packet\n");
+				DPRINTK("len = %d rxb->status = %x\n",
+					len, rxb->status.raw);
+				netif_rx(skb);
+				dev->last_rx = jiffies;
+				priv->stats.rx_packets++;
+				priv->stats.rx_bytes+=len;
+				DPRINTK("There we go... Whew...\n");
+			}
+		}
+		priv->regs->rx_fifo=priv->rx_ring_dmas[priv->rx_write];
+		rxb->status.raw=0;		
+		priv->rx_write=(priv->rx_write+1)&(RX_RING_ENTRIES-1);
+	}
+}
+
+static int meth_tx_full(struct net_device *dev)
+{
+	struct meth_private *priv = (struct meth_private *) dev->priv;
+
+	return(priv->tx_count >= TX_RING_ENTRIES-1);
+}
+
+void meth_tx_cleanup(struct net_device* dev, int rptr)
+{
+	meth_private *priv=dev->priv;
+	tx_packet* status;
+	struct sk_buff *skb;
+
+	spin_lock(&priv->meth_lock);
+
+	/* Stop DMA */
+	priv->regs->dma_ctrl &= ~(METH_DMA_TX_INT_EN);
+
+	while(priv->tx_read != rptr){
+		skb = priv->tx_skbs[priv->tx_read];
+		status = &priv->tx_ring[priv->tx_read];
+		if(!status->header.res.sent)
+			break;
+		if(status->header.raw & METH_TX_STATUS_DONE) {
+			priv->stats.tx_packets++;
+			priv->stats.tx_bytes += skb->len;
+		}
+		dev_kfree_skb_irq(skb);
+		priv->tx_skbs[priv->tx_read] = NULL;
+		status->header.raw = 0;
+		priv->tx_read = (priv->tx_read+1)&(TX_RING_ENTRIES-1);
+		priv->tx_count --;
+	}
+
+	/* wake up queue if it was stopped */
+	if (netif_queue_stopped(dev) && ! meth_tx_full(dev)) {
+		netif_wake_queue(dev);
+	}
+
+	spin_unlock(priv->meth_lock);
+}
+
+/*
+ * The typical interrupt entry point
+ */
+void meth_interrupt(int irq, void *dev_id, struct pt_regs *pregs)
+{
+	struct meth_private *priv;
+	union {
+		u32	reg; /*Whole status register */
+		struct {
+			u32			:	2,
+				rx_seq	:	5,
+				tx_read	:	9,
+				
+				rx_read	:	8,
+				int_mask:	8;
+		} parsed;
+	} status;
+	/*
+	 * As usual, check the "device" pointer for shared handlers.
+	 * Then assign "struct device *dev"
+	 */
+	struct net_device *dev = (struct net_device *)dev_id;
+	/* ... and check with hw if it's really ours */
+
+	if (!dev /*paranoid*/ ) return;
+
+	/* Lock the device */
+	priv = (struct meth_private *) dev->priv;
+
+	status.reg = priv->regs->int_flags;
+    
+	DPRINTK("Interrupt, status %08x...\n",status.reg);
+	if (status.parsed.int_mask & METH_INT_RX_THRESHOLD) {
+		/* send it to meth_rx for handling */
+		meth_rx(dev);
+	}
+
+	if (status.parsed.int_mask & (METH_INT_TX_EMPTY|METH_INT_TX_PKT)) {
+		/* a transmission is over: free the skb */
+		meth_tx_cleanup(dev, status.parsed.tx_read);
+	}
+	/* check for errors too... */
+	if (status.parsed.int_mask & (METH_INT_TX_LINK_FAIL))
+		printk(KERN_WARNING "meth: link failure\n");
+	if (status.parsed.int_mask & (METH_INT_MEM_ERROR))
+		printk(KERN_WARNING "meth: memory error\n");
+	if (status.parsed.int_mask & (METH_INT_TX_ABORT))
+		printk(KERN_WARNING "meth: aborted\n");
+	DPRINTK("Interrupt handling done...\n");
+	
+	priv->regs->int_flags=status.reg&0xff; /* clear interrupts */
+}
+
+/*
+ * Transmits packets that fit into TX descriptor (are <=120B)
+ */
+static void meth_tx_short_prepare(meth_private* priv, struct sk_buff* skb)
+{
+	tx_packet *desc=&priv->tx_ring[priv->tx_write];
+	int len = (skb->len<ETH_ZLEN)?ETH_ZLEN:skb->len;
+
+	DPRINTK("preparing short packet\n");
+	/* maybe I should set whole thing to 0 first... */
+	memcpy(desc->data.dt+(120-len),skb->data,skb->len);
+	if(skb->len < len)
+		memset(desc->data.dt+120-len+skb->len,0,len-skb->len);
+	desc->header.raw=METH_TX_CMD_INT_EN|(len-1)|((128-len)<<16);
+	DPRINTK("desc=%016lx\n",desc->header.raw);
+}
+#define TX_CATBUF1 BIT(25)
+static void meth_tx_1page_prepare(meth_private* priv, struct sk_buff* skb)
+{
+	tx_packet *desc=&priv->tx_ring[priv->tx_write];
+	void *buffer_data = (void *)(((u64)skb->data + 7ULL) & (~7ULL));
+	int unaligned_len = (int)((u64)buffer_data - (u64)skb->data);
+	int buffer_len = skb->len - unaligned_len;
+	dma_addr_t catbuf;
+
+	DPRINTK("preparing 1 page...\n");
+	DPRINTK("length=%d data=%p\n", skb->len, skb->data);
+	DPRINTK("unaligned_len=%d\n", unaligned_len);
+	DPRINTK("buffer_data=%p buffer_len=%d\n",
+	       buffer_data,
+	       buffer_len);
+
+	desc->header.raw=METH_TX_CMD_INT_EN|TX_CATBUF1|(skb->len-1);
+
+	/* unaligned part */
+	if(unaligned_len){
+		memcpy(desc->data.dt+(120-unaligned_len),
+		       skb->data, unaligned_len);
+		desc->header.raw |= (128-unaligned_len) << 16;
+	}
+
+	/* first page */
+	catbuf = pci_map_single(NULL,
+				buffer_data,
+				buffer_len,
+				PCI_DMA_TODEVICE);
+	DPRINTK("catbuf=%x\n", catbuf);
+	desc->data.cat_buf[0].form.start_addr = catbuf >> 3;
+	desc->data.cat_buf[0].form.len = buffer_len-1;
+	DPRINTK("desc=%016lx\n",desc->header.raw);
+	DPRINTK("cat_buf[0].raw=%016lx\n",desc->data.cat_buf[0].raw);
+}
+#define TX_CATBUF2 BIT(26)
+static void meth_tx_2page_prepare(meth_private* priv, struct sk_buff* skb)
+{
+	tx_packet *desc=&priv->tx_ring[priv->tx_write];
+	void *buffer1_data = (void *)(((u64)skb->data + 7ULL) & (~7ULL));
+	void *buffer2_data = (void *)PAGE_ALIGN((u64)skb->data);
+	int unaligned_len = (int)((u64)buffer1_data - (u64)skb->data);
+	int buffer1_len = (int)((u64)buffer2_data - (u64)buffer1_data);
+	int buffer2_len = skb->len - buffer1_len - unaligned_len;
+	dma_addr_t catbuf1, catbuf2;
+
+	DPRINTK("preparing 2 pages... \n");
+	DPRINTK("length=%d data=%p\n", skb->len, skb->data);
+	DPRINTK("unaligned_len=%d\n", unaligned_len);
+	DPRINTK("buffer1_data=%p buffer1_len=%d\n",
+	       buffer1_data,
+	       buffer1_len);
+	DPRINTK("buffer2_data=%p buffer2_len=%d\n",
+	       buffer2_data,
+	       buffer2_len);
+
+	desc->header.raw=METH_TX_CMD_INT_EN|TX_CATBUF1|TX_CATBUF2|(skb->len-1);
+	/* unaligned part */
+	if(unaligned_len){
+		memcpy(desc->data.dt+(120-unaligned_len),
+		       skb->data, unaligned_len);
+		desc->header.raw |= (128-unaligned_len) << 16;
+	}
+
+	/* first page */
+	catbuf1 = pci_map_single(NULL,
+				 buffer1_data,
+				 buffer1_len,
+				 PCI_DMA_TODEVICE);
+	DPRINTK("catbuf1=%x\n", catbuf1);
+	desc->data.cat_buf[0].form.start_addr = catbuf1 >> 3;
+	desc->data.cat_buf[0].form.len = buffer1_len-1;
+	/* second page */
+	catbuf2 = pci_map_single(NULL,
+				 buffer2_data,
+				 buffer2_len,
+				 PCI_DMA_TODEVICE);
+	DPRINTK("catbuf2=%x\n", catbuf2);
+	desc->data.cat_buf[1].form.start_addr = catbuf2 >> 3;
+	desc->data.cat_buf[1].form.len = buffer2_len-1;
+	DPRINTK("desc=%016lx\n",desc->header.raw);
+	DPRINTK("cat_buf[0].raw=%016lx\n",desc->data.cat_buf[0].raw);
+	DPRINTK("cat_buf[1].raw=%016lx\n",desc->data.cat_buf[1].raw);
+}
+
+
+void meth_add_to_tx_ring(meth_private *priv, struct sk_buff* skb)
+{
+	DPRINTK("Transmitting data...\n");
+	if(skb->len <= 120) {
+		/* Whole packet fits into descriptor */
+		meth_tx_short_prepare(priv,skb);
+	} else if(PAGE_ALIGN((u64)skb->data) !=
+		  PAGE_ALIGN((u64)skb->data+skb->len-1)) {
+		/* Packet crosses page boundary */
+		meth_tx_2page_prepare(priv,skb);
+	} else {
+		/* Packet is in one page */
+		meth_tx_1page_prepare(priv,skb);
+	}
+
+	/* Remember the skb, so we can free it at interrupt time */
+	priv->tx_skbs[priv->tx_write] = skb;
+	priv->tx_write = (priv->tx_write+1) & (TX_RING_ENTRIES-1);
+	priv->regs->tx_info.wptr = priv->tx_write;
+	priv->tx_count ++;
+	/* Enable DMA transfer */
+	priv->regs->dma_ctrl |= METH_DMA_TX_INT_EN;
+}
+
+/*
+ * Transmit a packet (called by the kernel)
+ */
+int meth_tx(struct sk_buff *skb, struct net_device *dev)
+{
+	struct meth_private *priv = (struct meth_private *) dev->priv;
+
+	spin_lock_irq(&priv->meth_lock);
+
+	meth_add_to_tx_ring(priv, skb);
+	dev->trans_start = jiffies; /* save the timestamp */
+
+	/* If TX ring is full, tell the upper layer to stop sending packets */
+	if (meth_tx_full(dev)) {
+	        DPRINTK("TX full: stopping\n");
+		netif_stop_queue(dev);
+	}
+
+	spin_unlock_irq(&priv->meth_lock);
+
+	return 0;
+}
+
+/*
+ * Deal with a transmit timeout.
+ */
+
+void meth_tx_timeout (struct net_device *dev)
+{
+	struct meth_private *priv = (struct meth_private *) dev->priv;
+	
+	printk(KERN_WARNING "%s: transmit timed out\n", dev->name);
+
+	/* Protect against concurrent rx interrupts */
+	spin_lock_irq(&priv->meth_lock);
+
+	/* Try to reset the adaptor. */
+	meth_reset(dev);
+
+	priv->stats.tx_errors++;
+
+	/* Clear all rings */
+	meth_free_tx_ring(priv);
+	meth_free_rx_ring(priv);
+	meth_init_tx_ring(priv);
+	meth_init_rx_ring(priv);
+
+	/* Restart dma */
+	priv->regs->dma_ctrl|=METH_DMA_TX_EN|METH_DMA_RX_EN|METH_DMA_RX_INT_EN;
+
+	/* Enable interrupt */
+	spin_unlock_irq(&priv->meth_lock);
+
+	dev->trans_start = jiffies;
+	netif_wake_queue(dev);
+
+	return;
+}
+
+/*
+ * Ioctl commands 
+ */
+int meth_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
+{
+ 
+    DPRINTK("ioctl\n");
+    return 0;
+}
+
+/*
+ * Return statistics to the caller
+ */
+struct net_device_stats *meth_stats(struct net_device *dev)
+{
+    struct meth_private *priv = (struct meth_private *) dev->priv;
+    return &priv->stats;
+}
+
+/*
+ * The init function (sometimes called probe).
+ * It is invoked by register_netdev()
+ */
+int meth_init(struct net_device *dev)
+{
+	meth_private *priv;
+	int ret;
+	/* 
+	 * Then, assign other fields in dev, using ether_setup() and some
+	 * hand assignments
+	 */
+	ether_setup(dev); /* assign some of the fields */
+
+	dev->open            = meth_open;
+	dev->stop            = meth_release;
+	dev->set_config      = meth_config;
+	dev->hard_start_xmit = meth_tx;
+	dev->do_ioctl        = meth_ioctl;
+	dev->get_stats       = meth_stats;
+#ifdef HAVE_TX_TIMEOUT
+	dev->tx_timeout      = meth_tx_timeout;
+	dev->watchdog_timeo  = timeout;
+#endif
+	dev->irq		 = MACE_ETHERNET_IRQ;
+	SET_MODULE_OWNER(dev);
+
+	/*
+	 * Then, allocate the priv field. This encloses the statistics
+	 * and a few private fields.
+	 */
+	priv = kmalloc(sizeof(struct meth_private), GFP_KERNEL);
+	if (priv == NULL)
+		return -ENOMEM;
+	dev->priv=priv;
+	memset(priv, 0, sizeof(struct meth_private));
+	spin_lock_init(&((struct meth_private *) dev->priv)->meth_lock);
+	/*
+	 * Make the usual checks: check_region(), probe irq, ...  -ENODEV
+	 * should be returned if no device found.  No resource should be
+	 * grabbed: this is done on open(). 
+	 */
+	priv->regs=(meth_regs*)SGI_MFE;
+	dev->base_addr=SGI_MFE;
+	priv->phy_addr = -1; /* No phy is known yet... */
+
+	/* Initialize the hardware */
+	if((ret=meth_reset(dev)) < 0)
+	        return ret;
+
+	/* Allocate the ring buffers */
+	if((ret=meth_init_tx_ring(priv))<0||(ret=meth_init_rx_ring(priv))<0){
+		meth_free_tx_ring(priv);
+		meth_free_rx_ring(priv);
+		return ret;
+	}
+
+	printk("SGI O2 Fast Ethernet rev. %ld\n", priv->regs->mac_ctrl >> 29);
+
+    return 0;
+}
+
+/*
+ * The devices
+ */
+
+struct net_device meth_devs[1] = {
+    { init: meth_init, }  /* init, nothing more */
+};
+
+/*
+ * Finally, the module stuff
+ */
+
+int meth_init_module(void)
+{
+	int result, device_present = 0;
+
+	strcpy(meth_devs[0].name, "eth%d");
+
+	if ( (result = register_netdev(meth_devs)) )
+		printk("meth: error %i registering device \"%s\"\n",
+		       result, meth_devs->name);
+	else device_present++;
+#ifndef METH_DEBUG
+	EXPORT_NO_SYMBOLS;
+#endif
+	
+	return device_present ? 0 : -ENODEV;
+}
+
+void meth_cleanup(void)
+{
+    kfree(meth_devs->priv);
+    unregister_netdev(meth_devs);
+    return;
+}
+
+module_init(meth_init_module);
+module_exit(meth_cleanup);

drivers/net/meth.h

+
+/*
+ * snull.h -- definitions for the network module
+ *
+ * Copyright (C) 2001 Alessandro Rubini and Jonathan Corbet
+ * Copyright (C) 2001 O'Reilly & Associates
+ *
+ * The source code in this file can be freely used, adapted,
+ * and redistributed in source or binary form, so long as an
+ * acknowledgment appears in derived source files.  The citation
+ * should list that the code comes from the book "Linux Device
+ * Drivers" by Alessandro Rubini and Jonathan Corbet, published
+ * by O'Reilly & Associates.   No warranty is attached;
+ * we cannot take responsibility for errors or fitness for use.
+ */
+
+/* version dependencies have been confined to a separate file */
+
+#define SGI_MFE		(MACE_BASE+MACE_ENET)
+/*		(0xBF280000)*/
+
+/* Tunable parameters */
+#define TX_RING_ENTRIES 64	/* 64-512?*/
+
+#define RX_RING_ENTRIES 16 /* Do not change */
+/* Internal constants */
+#define TX_RING_BUFFER_SIZE	(TX_RING_ENTRIES*sizeof(tx_packet))
+#define RX_BUFFER_SIZE 1546 /* ethenet packet size */
+#define METH_RX_BUFF_SIZE 4096
+#define RX_BUFFER_OFFSET (sizeof(rx_status_vector)+2) /* staus vector + 2 bytes of padding */
+#define RX_BUCKET_SIZE 256
+
+
+
+/* For more detailed explanations of what each field menas,
+   see Nick's great comments to #defines below (or docs, if
+   you are lucky enough toget hold of them :)*/
+
+/* tx status vector is written over tx command header upon
+   dma completion. */
+
+typedef struct tx_status_vector {
+	u64		sent:1; /* always set to 1...*/
+	u64		pad0:34;/* always set to 0 */
+	u64		flags:9;			/*I'm too lazy to specify each one separately at the moment*/
+	u64		col_retry_cnt:4;	/*collision retry count*/
+	u64		len:16;				/*Transmit length in bytes*/
+} tx_status_vector;
+
+/*
+ * Each packet is 128 bytes long.
+ * It consists of header, 0-3 concatination
+ * buffer pointers and up to 120 data bytes.
+ */
+typedef struct tx_packet_hdr {
+	u64		pad1:36; /*should be filled with 0 */
+	u64		cat_ptr3_valid:1,	/*Concatination pointer valid flags*/
+			cat_ptr2_valid:1,
+			cat_ptr1_valid:1;
+	u64		tx_int_flag:1;		/*Generate TX intrrupt when packet has been sent*/
+	u64		term_dma_flag:1;	/*Terminate transmit DMA on transmit abort conditions*/
+	u64		data_offset:7;		/*Starting byte offset in ring data block*/
+	u64		data_len:16;		/*Length of valid data in bytes-1*/
+} tx_packet_hdr;
+typedef union tx_cat_ptr {
+	struct {
+		u64		pad2:16; /* should be 0 */
+		u64		len:16;				/*length of buffer data - 1*/
+		u64		start_addr:29;		/*Physical starting address*/
+		u64		pad1:3; /* should be zero */
+	} form;
+	u64 raw;
+} tx_cat_ptr;
+
+typedef struct tx_packet {
+	union {
+		tx_packet_hdr header;
+		tx_status_vector res;
+		u64 raw;
+	}header;
+	union {
+		tx_cat_ptr cat_buf[3];
+		char dt[120];
+	} data;
+} tx_packet;
+
+typedef union rx_status_vector {
+	struct {
+		u64		pad1:1;/*fill it with ones*/
+		u64		pad2:15;/*fill with 0*/
+		u64		ip_chk_sum:16;
+		u64		seq_num:5;
+		u64		mac_addr_match:1;
+		u64		mcast_addr_match:1;
+		u64		carrier_event_seen:1;
+		u64		bad_packet:1;
+		u64		long_event_seen:1;
+		u64		invalid_preamble:1;
+		u64		broadcast:1;
+		u64		multicast:1;
+		u64		crc_error:1;
+		u64		huh:1;/*???*/
+		u64		rx_code_violation:1;
+		u64		rx_len:16;
+	} parsed;
+	u64 raw;
+} rx_status_vector;
+
+typedef struct rx_packet {
+	rx_status_vector status;
+        u64 pad[3]; /* For whatever reason, there needs to be 4 double-word offset */
+        u16 pad2;
+	char buf[METH_RX_BUFF_SIZE-sizeof(rx_status_vector)-3*sizeof(u64)-sizeof(u16)];/* data */
+} rx_packet;
+
+typedef struct meth_regs {
+	u64		mac_ctrl;		/*0x00,rw,31:0*/
+	u64		int_flags;		/*0x08,rw,30:0*/
+	u64		dma_ctrl;		/*0x10,rw,15:0*/
+	u64		timer;			/*0x18,rw,5:0*/
+	u64		int_tx;			/*0x20,wo,0:0*/
+	u64		int_rx;			/*0x28,wo,9:4*/
+	struct {
+		u32 tx_info_pad;
+		u32 rptr:16,wptr:16;
+	}		tx_info;		/*0x30,rw,31:0*/
+	u64		tx_info_al;		/*0x38,rw,31:0*/
+	struct {
+		u32	rx_buff_pad1;
+		u32	rx_buff_pad2:8,
+			wptr:8,
+			rptr:8,
+			depth:8;
+	}		rx_buff;		/*0x40,ro,23:0*/
+	u64		rx_buff_al1;	/*0x48,ro,23:0*/
+	u64		rx_buff_al2;	/*0x50,ro,23:0*/
+	u64		int_update;		/*0x58,wo,31:0*/
+	u32		phy_data_pad;
+	u32		phy_data;		/*0x60,rw,16:0*/
+	u32		phy_reg_pad;
+	u32		phy_registers;	/*0x68,rw,9:0*/
+	u64		phy_trans_go;	/*0x70,wo,0:0*/
+	u64		backoff_seed;	/*0x78,wo,10:0*/
+	u64		imq_reserved[4];/*0x80,ro,64:0(x4)*/
+	/*===================================*/
+	u64		mac_addr;		/*0xA0,rw,47:0, I think it's MAC address, but I'm not sure*/
+	u64		mcast_addr;		/*0xA8,rw,47:0, This seems like secondary MAC address*/
+	u64		mcast_filter;	/*0xB0,rw,63:0*/
+	u64		tx_ring_base;	/*0xB8,rw,31:13*/
+	/* Following are read-only debugging info register */
+	u64		tx_pkt1_hdr;	/*0xC0,ro,63:0*/
+	u64		tx_pkt1_ptr[3];	/*0xC8,ro,63:0(x3)*/
+	u64		tx_pkt2_hdr;	/*0xE0,ro,63:0*/
+	u64		tx_pkt2_ptr[3];	/*0xE8,ro,63:0(x3)*/
+	/*===================================*/
+	u32		rx_pad;
+	u32		rx_fifo;
+	u64		reserved[31];
+}meth_regs;
+
+	/* Bits in METH_MAC */
+
+#define SGI_MAC_RESET		BIT(0)	/* 0: MAC110 active in run mode, 1: Global reset signal to MAC110 core is active */
+#define METH_PHY_FDX		BIT(1) /* 0: Disable full duplex, 1: Enable full duplex */
+#define METH_PHY_LOOP	BIT(2) /* 0: Normal operation, follows 10/100mbit and M10T/MII select, 1: loops internal MII bus */
+				       /*    selects ignored */
+#define METH_100MBIT		BIT(3) /* 0: 10meg mode, 1: 100meg mode */
+#define METH_PHY_MII		BIT(4) /* 0: MII selected, 1: SIA selected */
+				       /*   Note: when loopback is set this bit becomes collision control.  Setting this bit will */
+				       /*         cause a collision to be reported. */
+
+				       /* Bits 5 and 6 are used to determine the the Destination address filter mode */
+#define METH_ACCEPT_MY 0			/* 00: Accept PHY address only */
+#define METH_ACCEPT_MCAST 0x20	/* 01: Accept physical, broadcast, and multicast filter matches only */
+#define METH_ACCEPT_AMCAST 0x40	/* 10: Accept physical, broadcast, and all multicast packets */
+#define METH_PROMISC 0x60		/* 11: Promiscious mode */
+
+#define METH_PHY_LINK_FAIL	BIT(7) /* 0: Link failure detection disabled, 1: Hardware scans for link failure in PHY */
+
+#define METH_MAC_IPG	0x1ffff00
+
+#define METH_DEFAULT_IPG ((17<<15) | (11<<22) | (21<<8))
+						/* 0x172e5c00 */ /* 23, 23, 23 */ /*0x54A9500 *//*21,21,21*/
+				       /* Bits 8 through 14 are used to determine Inter-Packet Gap between "Back to Back" packets */
+				       /* The gap depends on the clock speed of the link, 80ns per increment for 100baseT, 800ns  */
+				       /* per increment for 10BaseT */
+
+				       /* Bits 15 through 21 are used to determine IPGR1 */
+
+				       /* Bits 22 through 28 are used to determine IPGR2 */
+
+#define METH_REV_SHIFT 29       /* Bits 29 through 31 are used to determine the revision */
+				       /* 000: Inital revision */
+				       /* 001: First revision, Improved TX concatenation */
+
+
+/* DMA control bits */
+#define METH_RX_OFFSET_SHIFT 12 /* Bits 12:14 of DMA control register indicate starting offset of packet data for RX operation */
+#define METH_RX_DEPTH_SHIFT 4 /* Bits 8:4 define RX fifo depth -- when # of RX fifo entries != depth, interrupt is generted */
+
+#define METH_DMA_TX_EN BIT(1) /* enable TX DMA */
+#define METH_DMA_TX_INT_EN BIT(0) /* enable TX Buffer Empty interrupt */
+#define METH_DMA_RX_EN BIT(15) /* Enable RX */
+#define METH_DMA_RX_INT_EN BIT(9) /* Enable interrupt on RX packet */
+
+
+/* RX status bits */
+
+#define METH_RX_ST_RCV_CODE_VIOLATION BIT(16)
+#define METH_RX_ST_DRBL_NBL BIT(17)
+#define METH_RX_ST_CRC_ERR BIT(18)
+#define METH_RX_ST_MCAST_PKT BIT(19)
+#define METH_RX_ST_BCAST_PKT BIT(20)
+#define METH_RX_ST_INV_PREAMBLE_CTX BIT(21)
+#define METH_RX_ST_LONG_EVT_SEEN BIT(22)
+#define METH_RX_ST_BAD_PACKET BIT(23)
+#define METH_RX_ST_CARRIER_EVT_SEEN BIT(24)
+#define METH_RX_ST_MCAST_FILTER_MATCH BIT(25)
+#define METH_RX_ST_PHYS_ADDR_MATCH BIT(26)
+
+#define METH_RX_STATUS_ERRORS \
+	( \
+	METH_RX_ST_RCV_CODE_VIOLATION| \
+	METH_RX_ST_CRC_ERR| \
+	METH_RX_ST_INV_PREAMBLE_CTX| \
+	METH_RX_ST_LONG_EVT_SEEN| \
+	METH_RX_ST_BAD_PACKET| \
+	METH_RX_ST_CARRIER_EVT_SEEN \
+	)
+	/* Bits in METH_INT */
+	/* Write _1_ to corresponding bit to clear */
+#define METH_INT_TX_EMPTY	BIT(0)	/* 0: No interrupt pending, 1: The TX ring buffer is empty */
+#define METH_INT_TX_PKT		BIT(1)	/* 0: No interrupt pending */
+					      	/* 1: A TX message had the INT request bit set, the packet has been sent. */
+#define METH_INT_TX_LINK_FAIL	BIT(2)	/* 0: No interrupt pending, 1: PHY has reported a link failure */
+#define METH_INT_MEM_ERROR	BIT(3)	/* 0: No interrupt pending */
+						/* 1: A memory error occurred durring DMA, DMA stopped, Fatal */
+#define METH_INT_TX_ABORT		BIT(4)	/* 0: No interrupt pending, 1: The TX aborted operation, DMA stopped, FATAL */
+#define METH_INT_RX_THRESHOLD	BIT(5)	/* 0: No interrupt pending, 1: Selected receive threshold condition Valid */
+#define METH_INT_RX_UNDERFLOW	BIT(6)	/* 0: No interrupt pending, 1: FIFO was empty, packet could not be queued */
+#define METH_INT_RX_OVERFLOW		BIT(7)	/* 0: No interrupt pending, 1: DMA FIFO Overflow, DMA stopped, FATAL */
+
+#define METH_INT_RX_RPTR_MASK 0x0001F00		/* Bits 8 through 12 alias of RX read-pointer */
+
+						/* Bits 13 through 15 are always 0. */
+
+#define METH_INT_TX_RPTR_MASK 0x1FF0000	        /* Bits 16 through 24 alias of TX read-pointer */
+
+#define METH_INT_SEQ_MASK    0x2E000000	        /* Bits 25 through 29 are the starting seq number for the message at the */
+						/* top of the queue */
+
+#define METH_ERRORS ( \
+	METH_INT_RX_OVERFLOW|	\
+	METH_INT_RX_UNDERFLOW|	\
+	METH_INT_MEM_ERROR|			\
+	METH_INT_TX_ABORT)
+
+#define METH_INT_MCAST_HASH		BIT(30) /* If RX DMA is enabled the hash select logic output is latched here */
+
+/* TX status bits */
+#define METH_TX_STATUS_DONE BIT(23) /* Packet was transmitted successfully */
+
+/* Tx command header bits */
+#define METH_TX_CMD_INT_EN BIT(24) /* Generate TX interrupt when packet is sent */
+
+/* Phy MDIO interface busy flag */
+#define MDIO_BUSY    BIT(16)
+#define MDIO_DATA_MASK 0xFFFF
+/* PHY defines */
+#define PHY_QS6612X    0x0181441    /* Quality TX */
+#define PHY_ICS1889    0x0015F41    /* ICS FX */
+#define PHY_ICS1890    0x0015F42    /* ICS TX */
+#define PHY_DP83840    0x20005C0    /* National TX */