/* $Id: e100lpslavenet.c,v 1.5 2002/04/22 11:47:24 johana Exp $
*
* e100lpslavenet.c: A network driver for the ETRAX 100LX slave controller.
*
* Copyright (c) 1998-2001 Axis Communications AB.
*
* The outline of this driver comes from skeleton.c.
*
* $Log: e100lpslavenet.c,v $
* Revision 1.5 2002/04/22 11:47:24 johana
* Fix according to 2.4.19-pre7. time_after/time_before and
* missing end of comment.
* The patch has a typo for ethernet.c in e100_clear_network_leds(),
* that is fixed here.
*
* Revision 1.4 2001/06/21 16:55:26 olof
* Minimized par port setup time to gain bandwidth
*
* Revision 1.3 2001/06/21 15:49:02 olof
* Removed setting of default MAC address
*
* Revision 1.2 2001/06/11 15:39:52 olof
* Clean up and sync with ethernet.c rev 1.16. Increased reset time of slave.
*
* Revision 1.1 2001/06/06 08:56:26 olof
* Added support for slave Etrax defined by CONFIG_ETRAX_ETHERNET_LPSLAVE
*
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/delay.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/interrupt.h>
#include <linux/ptrace.h>
#include <linux/ioport.h>
#include <linux/in.h>
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/spinlock.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <asm/svinto.h> /* DMA and register descriptions */
#include <asm/io.h> /* LED_* I/O functions */
#include <asm/irq.h>
#include <asm/dma.h>
#include <asm/system.h>
#include <asm/bitops.h>
#include "e100lpslave.h"
/* #define ETHDEBUG */
#define D(x)
/*
* The name of the card. Is used for messages and in the requests for
* io regions, irqs and dma channels
*/
static const char* cardname = "Etrax 100LX ethernet slave controller";
/* A default ethernet address. Highlevel SW will set the real one later */
static struct sockaddr default_mac = {
0,
{ 0x00, 0x40, 0x8C, 0xCD, 0x00, 0x00 }
};
/* Information that need to be kept for each board. */
struct net_local {
struct net_device_stats stats;
/* Tx control lock. This protects the transmit buffer ring
* state along with the "tx full" state of the driver. This
* means all netif_queue flow control actions are protected
* by this lock as well.
*/
spinlock_t lock;
};
/* Dma descriptors etc. */
#define RX_BUF_SIZE 32768
#define ETHER_HEAD_LEN 14
#define PAR0_ECP_IRQ_NBR 4
#define RX_DESC_BUF_SIZE 256
#define NBR_OF_RX_DESC (RX_BUF_SIZE / \
RX_DESC_BUF_SIZE)
/* Size of slave etrax boot image */
#define ETRAX_PAR_BOOT_LENGTH 784
static etrax_dma_descr *myNextRxDesc; /* Points to the next descriptor to
to be processed */
static etrax_dma_descr *myLastRxDesc; /* The last processed descriptor */
static etrax_dma_descr *myPrevRxDesc; /* The descriptor right before myNextRxDesc */
static unsigned char RxBuf[RX_BUF_SIZE];
static etrax_dma_descr RxDescList[NBR_OF_RX_DESC] __attribute__ ((aligned(4)));
static etrax_dma_descr TxDescList[3] __attribute__ ((aligned(4)));
/* host command, data, bogus ECP command */
static struct sk_buff *tx_skb;
/* Index to functions, as function prototypes. */
static int etrax_ethernet_lpslave_init(struct net_device *dev);
static int e100_open(struct net_device *dev);
static int e100_set_mac_address(struct net_device *dev, void *addr);
static int e100_send_packet(struct sk_buff *skb, struct net_device *dev);
static void e100rx_interrupt(int irq, void *dev_id, struct pt_regs *regs);
static void e100tx_interrupt(int irq, void *dev_id, struct pt_regs *regs);
static void ecp_interrupt(int irq, void *dev_id, struct pt_regs *regs);
static void e100_rx(struct net_device *dev);
static int e100_close(struct net_device *dev);
static struct net_device_stats *e100_get_stats(struct net_device *dev);
static void set_multicast_list(struct net_device *dev);
static void e100_hardware_send_packet(unsigned long hostcmd, char *buf, int length);
static void update_rx_stats(struct net_device_stats *);
static void update_tx_stats(struct net_device_stats *);
static void e100_reset_tranceiver(void);
static void boot_slave(unsigned char *code);
#ifdef ETHDEBUG
static void dump_parport_status(void);
#endif
#define tx_done(dev) (*R_DMA_CH0_CMD == 0)
static unsigned long host_command;
extern unsigned char e100lpslaveprog;
/*
* This driver uses PAR0 to recevice data from slave ETRAX and PAR1 to boot
* and send data to slave ETRAX.
* Used ETRAX100 DMAchannels with corresponding IRQ:
* PAR0 RX : DMA3 - IRQ 19
* PAR1 TX : DMA4 - IRQ 20
* IRQ 4 is used to detect ECP commands from slave ETRAX
*
* NOTE! PAR0 and PAR1 shares DMA and IRQ numbers with SER2 and SER3
*/
/*
* Check for a network adaptor of this type, and return '0' if one exists.
* If dev->base_addr == 0, probe all likely locations.
* If dev->base_addr == 1, always return failure.
* If dev->base_addr == 2, allocate space for the device and return success
* (detachable devices only).
*/
static int __init
etrax_ethernet_lpslave_init(struct net_device *dev)
{
int i;
int anOffset = 0;
printk("Etrax/100 lpslave ethernet driver v0.3, (c) 1999 Axis Communications AB\n");
dev->base_addr = 2;
printk("%s initialized\n", dev->name);
/* make Linux aware of the new hardware */
if (!dev) {
printk(KERN_WARNING "%s: dev == NULL. Should this happen?\n",
cardname);
dev = init_etherdev(dev, sizeof(struct net_local));
if (!dev)
panic("init_etherdev failed\n");
}
/* setup generic handlers and stuff in the dev struct */
ether_setup(dev);
/* make room for the local structure containing stats etc */
dev->priv = kmalloc(sizeof(struct net_local), GFP_KERNEL);
if (dev->priv == NULL)
return -ENOMEM;
memset(dev->priv, 0, sizeof(struct net_local));
/* now setup our etrax specific stuff */
dev->irq = DMA3_RX_IRQ_NBR; /* we really use DMATX as well... */
dev->dma = PAR0_RX_DMA_NBR;
/* fill in our handlers so the network layer can talk to us in the future */
dev->open = e100_open;
dev->hard_start_xmit = e100_send_packet;
dev->stop = e100_close;
dev->get_stats = e100_get_stats;
dev->set_multicast_list = set_multicast_list;
dev->set_mac_address = e100_set_mac_address;
/* Initialise the list of Etrax DMA-descriptors */
/* Initialise receive descriptors */
for(i = 0; i < (NBR_OF_RX_DESC - 1); i++) {
RxDescList[i].ctrl = 0;
RxDescList[i].sw_len = RX_DESC_BUF_SIZE;
RxDescList[i].next = virt_to_phys(&RxDescList[i + 1]);
RxDescList[i].buf = virt_to_phys(RxBuf + anOffset);
RxDescList[i].status = 0;
RxDescList[i].hw_len = 0;
anOffset += RX_DESC_BUF_SIZE;
}
RxDescList[i].ctrl = d_eol;
RxDescList[i].sw_len = RX_DESC_BUF_SIZE;
RxDescList[i].next = virt_to_phys(&RxDescList[0]);
RxDescList[i].buf = virt_to_phys(RxBuf + anOffset);
RxDescList[i].status = 0;
RxDescList[i].hw_len = 0;
/* Initialise initial pointers */
myNextRxDesc = &RxDescList[0];
myLastRxDesc = &RxDescList[NBR_OF_RX_DESC - 1];
myPrevRxDesc = &RxDescList[NBR_OF_RX_DESC - 1];
/* setup some TX descriptor data */
TxDescList[0].sw_len = 4;
TxDescList[0].ctrl = 0;
TxDescList[0].buf = virt_to_phys(&host_command);
TxDescList[0].next = virt_to_phys(&TxDescList[1]);
return 0;
}
/* set MAC address of the interface. called from the core after a
* SIOCSIFADDR ioctl, and from the bootup above.
*/
static int
e100_set_mac_address(struct net_device *dev, void *p)
{
struct sockaddr *addr = p;
int i;
/* remember it */
memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
/* Write it to the hardware.
* Note the way the address is wrapped:
* *R_NETWORK_SA_0 = a0_0 | (a0_1 << 8) | (a0_2 << 16) | (a0_3 << 24);
* *R_NETWORK_SA_1 = a0_4 | (a0_5 << 8);
*/
tx_skb = 0;
e100_hardware_send_packet(HOST_CMD_SETMAC, dev->dev_addr, 6);
/* show it in the log as well */
printk("%s: changed MAC to ", dev->name);
for (i = 0; i < 5; i++)
printk("%02X:", dev->dev_addr[i]);
printk("%02X\n", dev->dev_addr[i]);
return 0;
}
/*
* Open/initialize the board. This is called (in the current kernel)
* sometime after booting when the 'ifconfig' program is run.
*
* This routine should set everything up anew at each open, even
* registers that "should" only need to be set once at boot, so that
* there is non-reboot way to recover if something goes wrong.
*/
static int
e100_open(struct net_device *dev)
{
unsigned long flags;
/* configure the PAR0 (RX) and PAR1 (TX) ports
*
* perror is nAckReverse, which must be 1 at the TX side,
* and 0 at the RX side
*
* select is XFlag, which must be 1 at both sides
*/
#ifdef ETHDEBUG
printk("Setting up PAR ports\n");
#endif
*R_PAR0_CONFIG =
/* We do not have an external buffer, don't care */
IO_STATE(R_PAR0_CONFIG, ioe, noninv) |
/* Not connected, don't care */
IO_STATE(R_PAR0_CONFIG, iseli, noninv) |
/* iautofd is not inverted, noninv */
IO_STATE(R_PAR0_CONFIG, iautofd, noninv) |
/* Not used in reverse direction, don't care */
IO_STATE(R_PAR0_CONFIG, istrb, noninv) |
/* Not connected, don't care */
IO_STATE(R_PAR0_CONFIG, iinit, noninv) |
/* perror is GND and reverse wants 0, noninv */
IO_STATE(R_PAR0_CONFIG, iperr, noninv) |
/* ack is not inverted, noninv */
IO_STATE(R_PAR0_CONFIG, iack, noninv) |
/* busy is not inverted, noninv */
IO_STATE(R_PAR0_CONFIG, ibusy, noninv) |
/* fault is not inverted, noninv */
IO_STATE(R_PAR0_CONFIG, ifault, noninv) |
/* select is Vcc and we want 1, noninv */
IO_STATE(R_PAR0_CONFIG, isel, noninv) |
/* We will run dma, enable */
IO_STATE(R_PAR0_CONFIG, dma, enable) |
/* No run length encoding, disable */
IO_STATE(R_PAR0_CONFIG, rle_in, disable) |
/* No run length encoding, disable */
IO_STATE(R_PAR0_CONFIG, rle_out, disable) |
/* Enable parallel port */
IO_STATE(R_PAR0_CONFIG, enable, on) |
/* Force mode regardless of pin status */
IO_STATE(R_PAR0_CONFIG, force, on) |
/* We want ECP forward mode since PAR0 is RX */
IO_STATE(R_PAR0_CONFIG, mode, ecp_rev);
*R_PAR1_CONFIG =
/* We do not have an external buffer, don't care */
IO_STATE(R_PAR1_CONFIG, ioe, noninv) |
/* Not connected, don't care */
IO_STATE(R_PAR1_CONFIG, iseli, noninv) |
/* HostAck must indicate data cycle, noninv */
IO_STATE(R_PAR1_CONFIG, iautofd, noninv) |
/* HostClk has no external inverter, noninv */
IO_STATE(R_PAR1_CONFIG, istrb, noninv) |
/* Not connected, don't care */
IO_STATE(R_PAR1_CONFIG, iinit, noninv) |
/* nAckReverse must be 1 in forward mode but is grounded, inv */
IO_STATE(R_PAR1_CONFIG, iperr, inv) |
/* PeriphClk must be 1 in forward mode, noninv */
IO_STATE(R_PAR1_CONFIG, iack, noninv) |
/* PeriphAck has no external inverter, noninv */
IO_STATE(R_PAR1_CONFIG, ibusy, noninv) |
/* nPerihpRequest has no external inverter, noniv */
IO_STATE(R_PAR1_CONFIG, ifault, noninv) |
/* Select is VCC and we want 1, noninv */
IO_STATE(R_PAR1_CONFIG, isel, noninv) |
/* No EPP mode, disable */
IO_STATE(R_PAR1_CONFIG, ext_mode, disable) |
/* We will run dma, enable */
IO_STATE(R_PAR1_CONFIG, dma, enable) |
/* No run length encoding, disable */
IO_STATE(R_PAR1_CONFIG, rle_in, disable) |
/* No run length encoding, disable */
IO_STATE(R_PAR1_CONFIG, rle_out, disable) |
/* Enable parallel port */
IO_STATE(R_PAR1_CONFIG, enable, on) |
/* Force mode regardless of pin status */
IO_STATE(R_PAR1_CONFIG, force, on) |
/* We want ECP forward mode since PAR1 is TX */
IO_STATE(R_PAR1_CONFIG, mode, ecp_fwd);
/* Setup time of value * 160 + 20 ns == 20 ns below */
*R_PAR1_DELAY = IO_FIELD(R_PAR1_DELAY, setup, 0);
*R_PAR1_CTRL = 0;
while ((((*R_PAR1_STATUS)&0xE000) >> 13) != 5); /* Wait for ECP_FWD mode */
#ifdef ETHDEBUG
dump_parport_status();
#endif
/* make sure ECP irq is acked when we enable it below */
(void)*R_PAR0_STATUS_DATA;
(void)*R_PAR1_STATUS_DATA;
/* Reset and wait for the DMA channels */
RESET_DMA(4); /* PAR1_TX_DMA_NBR */
RESET_DMA(3); /* PAR0_RX_DMA_NBR */
WAIT_DMA(4);
WAIT_DMA(3);
/* boot the slave Etrax, by sending code on PAR1.
* do this before we start up the IRQ handlers and stuff,
* beacuse we simply poll for completion in boot_slave.
*/
boot_slave(&e100lpslaveprog);
/* allocate the irq corresponding to the receiving DMA */
if (request_irq(DMA3_RX_IRQ_NBR, e100rx_interrupt, 0,
cardname, (void *)dev)) {
printk("Failed to allocate DMA3_RX_IRQ_NBR\n");
goto grace_exit;
}
/* allocate the irq corresponding to the transmitting DMA */
if (request_irq(DMA4_TX_IRQ_NBR, e100tx_interrupt, 0,
cardname, (void *)dev)) {
printk("Failed to allocate DMA4_TX_IRQ_NBR\n");
goto grace_exit;
}
/* allocate the irq used for detecting ECP commands on the RX port (PAR0) */
if (request_irq(PAR0_ECP_IRQ_NBR, ecp_interrupt, 0,
cardname, (void *)dev)) {
printk("Failed to allocate PAR0_ECP_IRQ_NBR\n");
grace_exit:
free_irq(PAR0_ECP_IRQ_NBR, (void *)dev);
free_irq(DMA4_TX_IRQ_NBR, (void *)dev);
free_irq(DMA3_RX_IRQ_NBR, (void *)dev);
return -EAGAIN;
}
#if 0
/* We are not allocating DMA since DMA4 is reserved for 'cascading'
* and will always fail with the current dma.c
*/
/*
* Always allocate the DMA channels after the IRQ,
* and clean up on failure.
*/
if(request_dma(PAR0_RX_DMA_NBR, cardname)) {
printk("Failed to allocate PAR0_RX_DMA_NBR\n");
goto grace_exit;
}
if(request_dma(PAR1_TX_DMA_NBR, cardname)) {
printk("Failed to allocate PAR1_TX_DMA_NBR\n");
grace_exit:
/* this will cause some 'trying to free free irq' but what the heck... */
free_dma(PAR1_TX_DMA_NBR);
free_dma(PAR0_RX_DMA_NBR);
free_irq(PAR0_ECP_IRQ_NBR, (void *)dev);
free_irq(DMA4_TX_IRQ_NBR, (void *)dev);
free_irq(DMA3_RX_IRQ_NBR, (void *)dev);
return -EAGAIN;
}
#endif
#ifdef ETHDEBUG
printk("Par port IRQ and DMA allocated\n");
#endif
save_flags(flags);
cli();
/* enable the irq's for PAR0/1 DMA */
*R_IRQ_MASK2_SET =
IO_STATE(R_IRQ_MASK2_SET, dma3_eop, set) |
IO_STATE(R_IRQ_MASK2_SET, dma4_descr, set);
*R_IRQ_MASK0_SET =
IO_STATE(R_IRQ_MASK0_SET, par0_ecp_cmd, set);
tx_skb = 0;
/* make sure the irqs are cleared */
*R_DMA_CH3_CLR_INTR = IO_STATE(R_DMA_CH3_CLR_INTR, clr_eop, do);
*R_DMA_CH4_CLR_INTR = IO_STATE(R_DMA_CH4_CLR_INTR, clr_descr, do);
/* Write the MAC address to the slave HW */
udelay(5000);
e100_hardware_send_packet(HOST_CMD_SETMAC, dev->dev_addr, 6);
/* make sure the rec and transmit error counters are cleared */
(void)*R_REC_COUNTERS; /* dummy read */
(void)*R_TR_COUNTERS; /* dummy read */
/* start the receiving DMA channel so we can receive packets from now on */
*R_DMA_CH3_FIRST = virt_to_phys(myNextRxDesc);
*R_DMA_CH3_CMD = IO_STATE(R_DMA_CH3_CMD, cmd, start);
restore_flags(flags);
/* We are now ready to accept transmit requeusts from
* the queueing layer of the networking.
*/
#ifdef ETHDEBUG
printk("Starting slave network transmit queue\n");
#endif
netif_start_queue(dev);
return 0;
}
static void
e100_reset_tranceiver(void)
{
/* To do: Reboot and setup slave Etrax */
}
/* Called by upper layers if they decide it took too long to complete
* sending a packet - we need to reset and stuff.
*/
static void
e100_tx_timeout(struct net_device *dev)
{
struct net_local *np = (struct net_local *)dev->priv;
printk(KERN_WARNING "%s: transmit timed out, %s?\n", dev->name,
tx_done(dev) ? "IRQ problem" : "network cable problem");
/* remember we got an error */
np->stats.tx_errors++;
/* reset the TX DMA in case it has hung on something */
RESET_DMA(4);
WAIT_DMA(4);
/* Reset the tranceiver. */
e100_reset_tranceiver();
/* and get rid of the packet that never got an interrupt */
dev_kfree_skb(tx_skb);
tx_skb = 0;
/* tell the upper layers we're ok again */
netif_wake_queue(dev);
}
/* This will only be invoked if the driver is _not_ in XOFF state.
* What this means is that we need not check it, and that this
* invariant will hold if we make sure that the netif_*_queue()
* calls are done at the proper times.
*/
static int
e100_send_packet(struct sk_buff *skb, struct net_device *dev)
{
struct net_local *np = (struct net_local *)dev->priv;
int length = ETH_ZLEN < skb->len ? skb->len : ETH_ZLEN;
unsigned char *buf = skb->data;
#ifdef ETHDEBUG
unsigned char *temp_data_ptr = buf;
int i;
printk("Sending a packet of length %d:\n", length);
/* dump the first bytes in the packet */
for(i = 0; i < 8; i++) {
printk("%d: %.2x %.2x %.2x %.2x %.2x %.2x %.2x %.2x\n", i * 8,
temp_data_ptr[0],temp_data_ptr[1],temp_data_ptr[2],
temp_data_ptr[3],temp_data_ptr[4],temp_data_ptr[5],
temp_data_ptr[6],temp_data_ptr[7]);
temp_data_ptr += 8;
}
#endif
spin_lock_irq(&np->lock); /* protect from tx_interrupt */
tx_skb = skb; /* remember it so we can free it in the tx irq handler later */
dev->trans_start = jiffies;
e100_hardware_send_packet(HOST_CMD_SENDPACK, buf, length);
/* this simple TX driver has only one send-descriptor so we're full
* directly. If this had a send-ring instead, we would only do this if
* the ring got full.
*/
netif_stop_queue(dev);
spin_unlock_irq(&np->lock);
return 0;
}
/*
* The typical workload of the driver:
* Handle the network interface interrupts.
*/
static void
e100rx_interrupt(int irq, void *dev_id, struct pt_regs * regs)
{
struct net_device *dev = (struct net_device *)dev_id;
unsigned long irqbits = *R_IRQ_MASK2_RD;
if(irqbits & IO_STATE(R_IRQ_MASK2_RD, dma3_eop, active)) {
/* acknowledge the eop interrupt */
*R_DMA_CH3_CLR_INTR = IO_STATE(R_DMA_CH3_CLR_INTR, clr_eop, do);
/* check if one or more complete packets were indeed received */
while(*R_DMA_CH3_FIRST != virt_to_phys(myNextRxDesc)) {
/* Take out the buffer and give it to the OS, then
* allocate a new buffer to put a packet in.
*/
e100_rx(dev);
((struct net_local *)dev->priv)->stats.rx_packets++;
/* restart/continue on the channel, for safety */
*R_DMA_CH3_CMD = IO_STATE(R_DMA_CH3_CMD, cmd, restart);
/* clear dma channel 3 eop/descr irq bits */
*R_DMA_CH3_CLR_INTR =
IO_STATE(R_DMA_CH3_CLR_INTR, clr_eop, do) |
IO_STATE(R_DMA_CH3_CLR_INTR, clr_descr, do);
/* now, we might have gotten another packet
so we have to loop back and check if so */
}
}
}
/* the transmit dma channel interrupt
*
* this is supposed to free the skbuff which was pending during transmission,
* and inform the kernel that we can send one more buffer
*/
static void
e100tx_interrupt(int irq, void *dev_id, struct pt_regs * regs)
{
struct net_device *dev = (struct net_device *)dev_id;
unsigned long irqbits = *R_IRQ_MASK2_RD;
struct net_local *np = (struct net_local *)dev->priv;
#ifdef ETHDEBUG
printk("We got tx interrupt\n");
#endif
/* check for a dma4_eop interrupt */
if(irqbits & IO_STATE(R_IRQ_MASK2_RD, dma4_descr, active)) {
/* This protects us from concurrent execution of
* our dev->hard_start_xmit function above.
*/
spin_lock(&np->lock);
/* acknowledge the eop interrupt */
*R_DMA_CH4_CLR_INTR = IO_STATE(R_DMA_CH4_CLR_INTR, clr_descr, do);
/* skip *R_DMA_CH4_FIRST == 0 test since we use d_wait... */
if(tx_skb) {
np->stats.tx_bytes += tx_skb->len;
np->stats.tx_packets++;
/* dma is ready with the transmission of the data in tx_skb, so now we can release the skb memory */
dev_kfree_skb_irq(tx_skb);
tx_skb = 0;
netif_wake_queue(dev);
} else {
printk(KERN_WARNING "%s: tx weird interrupt\n",
cardname);
}
spin_unlock(&np->lock);
}
}
static void
ecp_interrupt(int irq, void *dev_id, struct pt_regs * regs)
{
struct net_device *dev = (struct net_device *)dev_id;
struct net_local *lp = (struct net_local *)dev->priv;
unsigned long temp, irqbits = *R_IRQ_MASK0_RD;
/* check for ecp irq */
if(irqbits & IO_MASK(R_IRQ_MASK0_RD, par0_ecp_cmd)) {
/* acknowledge by reading the bit */
temp = *R_PAR0_STATUS_DATA;
/* force an EOP on the incoming channel, so we'll get an rx interrupt */
*R_SET_EOP = IO_STATE(R_SET_EOP, ch3_eop, set);
}
}
/* We have a good packet(s), get it/them out of the buffers. */
static void
e100_rx(struct net_device *dev)
{
struct sk_buff *skb;
int length=0;
int i;
struct net_local *np = (struct net_local *)dev->priv;
struct etrax_dma_descr *mySaveRxDesc = myNextRxDesc;
unsigned char *skb_data_ptr;
/* If the packet is broken down in many small packages then merge
* count how much space we will need to alloc with skb_alloc() for
* it to fit.
*/
while (!(myNextRxDesc->status & d_eop)) {
length += myNextRxDesc->sw_len; /* use sw_len for the first descs */
myNextRxDesc->status = 0;
myNextRxDesc = phys_to_virt(myNextRxDesc->next);
}
length += myNextRxDesc->hw_len; /* use hw_len for the last descr */
#ifdef ETHDEBUG
printk("Got a packet of length %d:\n", length);
/* dump the first bytes in the packet */
skb_data_ptr = (unsigned char *)phys_to_virt(mySaveRxDesc->buf);
for(i = 0; i < 8; i++) {
printk("%d: %.2x %.2x %.2x %.2x %.2x %.2x %.2x %.2x\n", i * 8,
skb_data_ptr[0],skb_data_ptr[1],skb_data_ptr[2],skb_data_ptr[3],
skb_data_ptr[4],skb_data_ptr[5],skb_data_ptr[6],skb_data_ptr[7]);
skb_data_ptr += 8;
}
#endif
skb = dev_alloc_skb(length - ETHER_HEAD_LEN);
if (!skb) {
np->stats.rx_errors++;
printk(KERN_NOTICE "%s: Memory squeeze, dropping packet.\n",
dev->name);
return;
}
skb_put(skb, length - ETHER_HEAD_LEN); /* allocate room for the packet body */
skb_data_ptr = skb_push(skb, ETHER_HEAD_LEN); /* allocate room for the header */
#ifdef ETHDEBUG
printk("head = 0x%x, data = 0x%x, tail = 0x%x, end = 0x%x\n",
skb->head, skb->data, skb->tail, skb->end);
printk("copying packet to 0x%x.\n", skb_data_ptr);
#endif
/* this loop can be made using max two memcpy's if optimized */
while(mySaveRxDesc != myNextRxDesc) {
memcpy(skb_data_ptr, phys_to_virt(mySaveRxDesc->buf),
mySaveRxDesc->sw_len);
skb_data_ptr += mySaveRxDesc->sw_len;
mySaveRxDesc = phys_to_virt(mySaveRxDesc->next);
}
memcpy(skb_data_ptr, phys_to_virt(mySaveRxDesc->buf),
mySaveRxDesc->hw_len);
skb->dev = dev;
skb->protocol = eth_type_trans(skb, dev);
/* Send the packet to the upper layers */
netif_rx(skb);
/* Prepare for next packet */
myNextRxDesc->status = 0;
myPrevRxDesc = myNextRxDesc;
myNextRxDesc = phys_to_virt(myNextRxDesc->next);
myPrevRxDesc->ctrl |= d_eol;
myLastRxDesc->ctrl &= ~d_eol;
myLastRxDesc = myPrevRxDesc;
return;
}
/* The inverse routine to net_open(). */
static int
e100_close(struct net_device *dev)
{
struct net_local *np = (struct net_local *)dev->priv;
printk("Closing %s.\n", dev->name);
netif_stop_queue(dev);
*R_IRQ_MASK0_CLR = IO_STATE(R_IRQ_MASK0_CLR, par0_ecp_cmd, clr);
*R_IRQ_MASK2_CLR =
IO_STATE(R_IRQ_MASK2_CLR, dma3_eop, clr) |
IO_STATE(R_IRQ_MASK2_CLR, dma4_descr, clr);
/* Stop the receiver and the transmitter */
RESET_DMA(3);
RESET_DMA(4);
/* Flush the Tx and disable Rx here. */
free_irq(DMA3_RX_IRQ_NBR, (void *)dev);
free_irq(DMA4_TX_IRQ_NBR, (void *)dev);
free_irq(PAR0_ECP_IRQ_NBR, (void *)dev);
free_dma(PAR1_TX_DMA_NBR);
free_dma(PAR0_RX_DMA_NBR);
/* Update the statistics here. */
update_rx_stats(&np->stats);
update_tx_stats(&np->stats);
return 0;
}
static void
update_rx_stats(struct net_device_stats *es)
{
unsigned long r = *R_REC_COUNTERS;
/* update stats relevant to reception errors */
es->rx_fifo_errors += r >> 24; /* fifo overrun */
es->rx_crc_errors += r & 0xff; /* crc error */
es->rx_frame_errors += (r >> 8) & 0xff; /* alignment error */
es->rx_length_errors += (r >> 16) & 0xff; /* oversized frames */
}
static void
update_tx_stats(struct net_device_stats *es)
{
unsigned long r = *R_TR_COUNTERS;
/* update stats relevant to transmission errors */
es->collisions += (r & 0xff) + ((r >> 8) & 0xff); /* single_col + multiple_col */
es->tx_errors += (r >> 24) & 0xff; /* deferred transmit frames */
}
/*
* Get the current statistics.
* This may be called with the card open or closed.
*/
static struct net_device_stats *
e100_get_stats(struct net_device *dev)
{
struct net_local *lp = (struct net_local *)dev->priv;
update_rx_stats(&lp->stats);
update_tx_stats(&lp->stats);
return &lp->stats;
}
/*
* Set or clear the multicast filter for this adaptor.
* num_addrs == -1 Promiscuous mode, receive all packets
* num_addrs == 0 Normal mode, clear multicast list
* num_addrs > 0 Multicast mode, receive normal and MC packets,
* and do best-effort filtering.
*/
static void
set_multicast_list(struct net_device *dev)
{
/* To do */
}
void
e100_hardware_send_packet(unsigned long hostcmd, char *buf, int length)
{
static char bogus_ecp[] = { 42, 42 };
int i;
#ifdef ETHDEBUG
printk("e100 send pack, buf 0x%x len %d\n", buf, length);
#endif
host_command = hostcmd;
/* Configure the tx dma descriptor. Desc 0 is already configured.*/
TxDescList[1].sw_len = length;
/* bug workaround - etrax100 needs d_wait on the descriptor _before_
* a descriptor containing an ECP command
*/
TxDescList[1].ctrl = d_wait;
TxDescList[1].buf = virt_to_phys(buf);
TxDescList[1].next = virt_to_phys(&TxDescList[2]);
/* append the ecp dummy descriptor - its only purpose is to
* make the receiver generate an irq due to the ecp command
* so the receiver knows where packets end
*/
TxDescList[2].sw_len = 1;
TxDescList[2].ctrl = d_ecp | d_eol | d_int;
TxDescList[2].buf = virt_to_phys(bogus_ecp);
/* setup the dma channel and start it */
*R_DMA_CH4_FIRST = virt_to_phys(TxDescList);
*R_DMA_CH4_CMD = IO_STATE(R_DMA_CH4_CMD, cmd, start);
#ifdef ETHDEBUG
printk("done\n");
#endif
}
/* send a chunk of code to the slave chip to boot it. */
static void
boot_slave(unsigned char *code)
{
int i;
#ifdef ETHDEBUG
printk(" booting slave ETRAX...\n");
#endif
*R_PORT_PB_DATA = 0x7F; /* Reset slave */
udelay(15); /* Time enough to reset WAN tranciever */
*R_PORT_PB_DATA = 0xFF; /* Reset slave */
/* configure the tx dma data descriptor */
TxDescList[1].sw_len = ETRAX_PAR_BOOT_LENGTH;
TxDescList[1].ctrl = d_eol | d_int;
TxDescList[1].buf = virt_to_phys(code);
TxDescList[1].next = 0;
/* setup the dma channel and start it */
*R_DMA_CH4_FIRST = virt_to_phys(&TxDescList[1]);
*R_DMA_CH4_CMD = IO_STATE(R_DMA_CH4_CMD, cmd, start);
/* wait for completion */
while(!(*R_IRQ_READ2 & IO_MASK(R_IRQ_READ2, dma4_descr)));
/* ack the irq */
*R_DMA_CH4_CLR_INTR = IO_STATE(R_DMA_CH4_CLR_INTR, clr_descr, do);
#if 0
/* manual transfer of boot code - requires dma turned off */
for (i=0; i<ETRAX_PAR_BOOT_LENGTH; i++)
{
printk(" sending byte: %u value: %x\n",i,code[i]);
while (((*R_PAR1_STATUS)&0x02) == 0); /* Wait while tr_rdy is busy*/
*R_PAR1_CTRL_DATA = code[i];
}
#endif
#ifdef ETHDEBUG
printk(" done\n");
#endif
}
#ifdef ETHDEBUG
/* debug code to check the current status of PAR1 */
static void
dump_parport_status(void)
{
unsigned long temp;
printk("Parport1 status:\n");
temp = (*R_PAR1_STATUS)&0xE000;
temp = temp >> 13;
printk("Reg mode: %u (ecp_fwd(5), ecp_rev(6))\n", temp);
temp = (*R_PAR1_STATUS)&0x1000;
temp = temp >> 12;
printk("Reg perr: %u (ecp_rev(0))\n", temp);
temp = (*R_PAR1_STATUS)&0x0800;
temp = temp >> 11;
printk("Reg ack: %u (inactive (1), active (0))\n", temp);
temp = (*R_PAR1_STATUS)&0x0400;
temp = temp >> 10;
printk("Reg busy: %u (inactive (0), active (1))\n", temp);
temp = (*R_PAR1_STATUS)&0x0200;
temp = temp >> 9;
printk("Reg fault: %u (inactive (1), active (0))\n", temp);
temp = (*R_PAR1_STATUS)&0x0100;
temp = temp >> 8;
printk("Reg sel: %u (inactive (0), active (1), xflag(1))\n", temp);
temp = (*R_PAR1_STATUS)&0x02;
temp = temp >> 1;
printk("Reg tr_rdy: %u (busy (0), ready (1))\n", temp);
}
#endif /* ETHDEBUG */
static struct net_device dev_etrax_slave_ethernet;
static int
etrax_init_module(void)
{
struct net_device *d = &dev_etrax_slave_ethernet;
d->init = etrax_ethernet_lpslave_init;
if(register_netdev(d) == 0)
return 0;
else
return -ENODEV;
}
module_init(etrax_init_module);