patch-2.1.72 linux/drivers/net/dmascc.c
Next file: linux/drivers/net/e2100.c
Previous file: linux/drivers/net/Space.c
Back to the patch index
Back to the overall index
- Lines: 1261
- Date:
Tue Dec 9 09:49:58 1997
- Orig file:
v2.1.71/linux/drivers/net/dmascc.c
- Orig date:
Wed Dec 31 16:00:00 1969
diff -u --recursive --new-file v2.1.71/linux/drivers/net/dmascc.c linux/drivers/net/dmascc.c
@@ -0,0 +1,1260 @@
+/*
+ * $Id: dmascc.c,v 1.2 1997/12/02 16:49:49 oe1kib Exp $
+ *
+ * Driver for high-speed SCC boards (those with DMA support)
+ * Copyright (C) 1997 Klaus Kudielka
+ *
+ * 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., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+
+#include <linux/module.h>
+#include <linux/dmascc.h>
+#include <linux/errno.h>
+#include <linux/if_arp.h>
+#include <linux/in.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/ioport.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/netdevice.h>
+#include <linux/sockios.h>
+#include <linux/tqueue.h>
+#include <linux/version.h>
+#include <asm/atomic.h>
+#include <asm/bitops.h>
+#include <asm/dma.h>
+#include <asm/io.h>
+#include <asm/irq.h>
+#include <asm/segment.h>
+#include <asm/uaccess.h>
+#include <net/ax25.h>
+#include <stdio.h>
+#include "z8530.h"
+
+
+/* Number of buffers per channel */
+
+#define NUM_TX_BUF 2 /* NUM_TX_BUF >= 1 (2 recommended) */
+#define NUM_RX_BUF 2 /* NUM_RX_BUF >= 1 (2 recommended) */
+#define BUF_SIZE 2016
+
+
+/* Cards supported */
+
+#define HW_PI { "Ottawa PI", 0x300, 0x20, 0x10, 8, \
+ 0, 8, 1843200, 3686400 }
+#define HW_PI2 { "Ottawa PI2", 0x300, 0x20, 0x10, 8, \
+ 0, 8, 3686400, 7372800 }
+#define HW_TWIN { "Gracilis PackeTwin", 0x200, 0x10, 0x10, 32, \
+ 0, 4, 6144000, 6144000 }
+
+#define HARDWARE { HW_PI, HW_PI2, HW_TWIN }
+
+#define TYPE_PI 0
+#define TYPE_PI2 1
+#define TYPE_TWIN 2
+#define NUM_TYPES 3
+
+#define MAX_NUM_DEVS 32
+
+
+/* SCC chips supported */
+
+#define Z8530 0
+#define Z85C30 1
+#define Z85230 2
+
+#define CHIPNAMES { "Z8530", "Z85C30", "Z85230" }
+
+
+/* I/O registers */
+
+/* 8530 registers relative to card base */
+#define SCCB_CMD 0x00
+#define SCCB_DATA 0x01
+#define SCCA_CMD 0x02
+#define SCCA_DATA 0x03
+
+/* 8254 registers relative to card base */
+#define TMR_CNT0 0x00
+#define TMR_CNT1 0x01
+#define TMR_CNT2 0x02
+#define TMR_CTRL 0x03
+
+/* Additional PI/PI2 registers relative to card base */
+#define PI_DREQ_MASK 0x04
+
+/* Additional PackeTwin registers relative to card base */
+#define TWIN_INT_REG 0x08
+#define TWIN_CLR_TMR1 0x09
+#define TWIN_CLR_TMR2 0x0a
+#define TWIN_SPARE_1 0x0b
+#define TWIN_DMA_CFG 0x08
+#define TWIN_SERIAL_CFG 0x09
+#define TWIN_DMA_CLR_FF 0x0a
+#define TWIN_SPARE_2 0x0b
+
+
+/* PackeTwin I/O register values */
+
+/* INT_REG */
+#define TWIN_SCC_MSK 0x01
+#define TWIN_TMR1_MSK 0x02
+#define TWIN_TMR2_MSK 0x04
+#define TWIN_INT_MSK 0x07
+
+/* SERIAL_CFG */
+#define TWIN_DTRA_ON 0x01
+#define TWIN_DTRB_ON 0x02
+#define TWIN_EXTCLKA 0x04
+#define TWIN_EXTCLKB 0x08
+#define TWIN_LOOPA_ON 0x10
+#define TWIN_LOOPB_ON 0x20
+#define TWIN_EI 0x80
+
+/* DMA_CFG */
+#define TWIN_DMA_HDX_T1 0x08
+#define TWIN_DMA_HDX_R1 0x0a
+#define TWIN_DMA_HDX_T3 0x14
+#define TWIN_DMA_HDX_R3 0x16
+#define TWIN_DMA_FDX_T3R1 0x1b
+#define TWIN_DMA_FDX_T1R3 0x1d
+
+
+/* Status values */
+
+/* tx_state */
+#define TX_IDLE 0
+#define TX_OFF 1
+#define TX_TXDELAY 2
+#define TX_ACTIVE 3
+#define TX_SQDELAY 4
+
+
+/* Data types */
+
+struct scc_hardware {
+ char *name;
+ int io_region;
+ int io_delta;
+ int io_size;
+ int num_devs;
+ int scc_offset;
+ int tmr_offset;
+ int tmr_hz;
+ int pclk_hz;
+};
+
+struct scc_priv {
+ char name[10];
+ struct enet_statistics stats;
+ struct scc_info *info;
+ int channel;
+ int cmd, data, tmr;
+ struct scc_param param;
+ char rx_buf[NUM_RX_BUF][BUF_SIZE];
+ int rx_len[NUM_RX_BUF];
+ int rx_ptr;
+ struct tq_struct rx_task;
+ int rx_head, rx_tail, rx_count;
+ int rx_over;
+ char tx_buf[NUM_TX_BUF][BUF_SIZE];
+ int tx_len[NUM_TX_BUF];
+ int tx_ptr;
+ int tx_head, tx_tail, tx_count;
+ int tx_sem, tx_state;
+ unsigned long tx_start;
+ int status;
+};
+
+struct scc_info {
+ int type;
+ int chip;
+ int open;
+ int scc_base;
+ int tmr_base;
+ int twin_serial_cfg;
+ struct device dev[2];
+ struct scc_priv priv[2];
+ struct scc_info *next;
+};
+
+
+/* Function declarations */
+
+int dmascc_init(void) __init;
+static int setup_adapter(int io, int h, int n) __init;
+
+static inline void write_scc(int ctl, int reg, int val);
+static inline int read_scc(int ctl, int reg);
+static int scc_open(struct device *dev);
+static int scc_close(struct device *dev);
+static int scc_ioctl(struct device *dev, struct ifreq *ifr, int cmd);
+static int scc_send_packet(struct sk_buff *skb, struct device *dev);
+static struct enet_statistics *scc_get_stats(struct device *dev);
+static int scc_set_mac_address(struct device *dev, void *sa);
+static void scc_isr(int irq, void *dev_id, struct pt_regs * regs);
+static inline void z8530_isr(struct scc_info *info);
+static void rx_isr(struct device *dev);
+static void special_condition(struct device *dev, int rc);
+static void rx_bh(void *arg);
+static void tx_isr(struct device *dev);
+static void es_isr(struct device *dev);
+static void tm_isr(struct device *dev);
+static inline void delay(struct device *dev, int t);
+static inline unsigned char random(void);
+
+
+/* Initialization variables */
+
+static int io[MAX_NUM_DEVS] __initdata = { 0, };
+/* Beware! hw[] is also used in cleanup_module(). If __initdata also applies
+ to modules, we may not declare hw[] as __initdata */
+static struct scc_hardware hw[NUM_TYPES] __initdata = HARDWARE;
+static char ax25_broadcast[7] __initdata =
+ { 'Q'<<1, 'S'<<1, 'T'<<1, ' '<<1, ' '<<1, ' '<<1, '0'<<1 };
+static char ax25_test[7] __initdata =
+ { 'L'<<1, 'I'<<1, 'N'<<1, 'U'<<1, 'X'<<1, ' '<<1, '1'<<1 };
+
+
+/* Global variables */
+
+static struct scc_info *first = NULL;
+static unsigned long rand;
+
+
+
+/* Module functions */
+
+#ifdef MODULE
+
+
+MODULE_AUTHOR("Klaus Kudielka <oe1kib@oe1xtu.ampr.org>");
+MODULE_DESCRIPTION("Driver for high-speed SCC boards");
+MODULE_PARM(io, "1-" __MODULE_STRING(MAX_NUM_DEVS) "i");
+
+
+int init_module(void)
+{
+ return dmascc_init();
+}
+
+
+void cleanup_module(void)
+{
+ int i;
+ struct scc_info *info;
+
+ while (first) {
+ info = first;
+
+ /* Unregister devices */
+ for (i = 0; i < 2; i++) {
+ if (info->dev[i].name)
+ unregister_netdev(&info->dev[i]);
+ }
+
+ /* Reset board */
+ if (info->type == TYPE_TWIN)
+ outb_p(0, info->dev[0].base_addr + TWIN_SERIAL_CFG);
+ write_scc(info->priv[0].cmd, R9, FHWRES);
+ release_region(info->dev[0].base_addr,
+ hw[info->type].io_size);
+
+ /* Free memory */
+ first = info->next;
+ kfree_s(info, sizeof(struct scc_info));
+ }
+}
+
+
+#else
+
+
+__initfunc(void dmascc_setup(char *str, int *ints))
+{
+ int i;
+
+ for (i = 0; i < MAX_NUM_DEVS && i < ints[0]; i++)
+ io[i] = ints[i+1];
+}
+
+
+#endif
+
+
+/* Initialization functions */
+
+__initfunc(int dmascc_init(void))
+{
+ int h, i, j, n, base[MAX_NUM_DEVS], tcmd, t0, t1, status;
+ unsigned long time, start[MAX_NUM_DEVS], stop[MAX_NUM_DEVS];
+
+ /* Initialize random number generator */
+ rand = jiffies;
+
+ /* Cards found = 0 */
+ n = 0;
+
+ /* Run autodetection for each card type */
+ for (h = 0; h < NUM_TYPES; h++) {
+
+ if (io[0]) {
+ /* User-specified I/O address regions */
+ for (i = 0; i < hw[h].num_devs; i++) base[i] = 0;
+ for (i = 0; i < MAX_NUM_DEVS && io[i]; i++) {
+ j = (io[i] - hw[h].io_region) / hw[h].io_delta;
+ if (j >= 0 &&
+ j < hw[h].num_devs &&
+ hw[h].io_region + j * hw[h].io_delta == io[i])
+ base[j] = io[i];
+ }
+ } else {
+ /* Default I/O address regions */
+ for (i = 0; i < hw[h].num_devs; i++)
+ base[i] = hw[h].io_region + i * hw[h].io_delta;
+ }
+
+ /* Check valid I/O address regions */
+ for (i = 0; i < hw[h].num_devs; i++)
+ if (base[i] && check_region(base[i], hw[h].io_size))
+ base[i] = 0;
+
+ /* Start timers */
+ for (i = 0; i < hw[h].num_devs; i++)
+ if (base[i]) {
+ tcmd = base[i] + hw[h].tmr_offset + TMR_CTRL;
+ t0 = base[i] + hw[h].tmr_offset + TMR_CNT0;
+ t1 = base[i] + hw[h].tmr_offset + TMR_CNT1;
+ /* Timer 0: LSB+MSB, Mode 3, TMR_0_HZ */
+ outb_p(0x36, tcmd);
+ outb_p((hw[h].tmr_hz/TMR_0_HZ) & 0xFF, t0);
+ outb_p((hw[h].tmr_hz/TMR_0_HZ) >> 8, t0);
+ /* Timer 1: LSB+MSB, Mode 0, HZ/10 */
+ outb_p(0x70, tcmd);
+ outb_p((TMR_0_HZ/HZ*10) & 0xFF, t1);
+ outb_p((TMR_0_HZ/HZ*10) >> 8, t1);
+ /* Timer 2: LSB+MSB, Mode 0 */
+ outb_p(0xb0, tcmd);
+ }
+
+ /* Initialize start values in case we miss the null count bit */
+ time = jiffies;
+ for (i = 0; i < hw[h].num_devs; i++) start[i] = time;
+
+ /* Timing loop */
+ while (jiffies - time < 12) {
+ for (i = 0; i < hw[h].num_devs; i++)
+ if (base[i]) {
+ /* Read back Timer 1: Status */
+ outb_p(0xE4, base[i] + hw[h].tmr_offset + TMR_CTRL);
+ status = inb_p(base[i] + hw[h].tmr_offset + TMR_CNT1);
+ if ((status & 0x3F) != 0x30) base[i] = 0;
+ if (status & 0x40) start[i] = jiffies;
+ if (~status & 0x80) stop[i] = jiffies;
+ }
+ }
+
+ /* Evaluate measurements */
+ for (i = 0; i < hw[h].num_devs; i++)
+ if (base[i]) {
+ time = stop[i] - start[i];
+ if (time < 9 || time > 11)
+ /* The time expired doesn't match */
+ base[i] = 0;
+ else {
+ /* Ok, we have found an adapter */
+ if (setup_adapter(base[i], h, n) == 0)
+ n++;
+ }
+ }
+
+ } /* NUM_TYPES */
+
+ /* If any adapter was successfully initialized, return ok */
+ if (n) return 0;
+
+ /* If no adapter found, return error */
+ printk("dmascc: no adapters found\n");
+ return -EIO;
+}
+
+
+__initfunc(int setup_adapter(int io, int h, int n))
+{
+ int i, irq, chip;
+ struct scc_info *info;
+ struct device *dev;
+ struct scc_priv *priv;
+ unsigned long time;
+ unsigned int irqs;
+ int tmr = io + hw[h].tmr_offset;
+ int scc = io + hw[h].scc_offset;
+ int cmd = scc + SCCA_CMD;
+ char *chipnames[] = CHIPNAMES;
+
+ /* Reset 8530 */
+ write_scc(cmd, R9, FHWRES | MIE | NV);
+
+ /* Determine type of chip */
+ write_scc(cmd, R15, 1);
+ if (!read_scc(cmd, R15)) {
+ /* WR7' not present. This is an ordinary Z8530 SCC. */
+ chip = Z8530;
+ } else {
+ /* Put one character in TX FIFO */
+ write_scc(cmd, R8, 0);
+ if (read_scc(cmd, R0) & Tx_BUF_EMP) {
+ /* TX FIFO not full. This is a Z85230 ESCC with a 4-byte FIFO. */
+ chip = Z85230;
+ } else {
+ /* TX FIFO full. This is a Z85C30 SCC with a 1-byte FIFO. */
+ chip = Z85C30;
+ }
+ }
+ write_scc(cmd, R15, 0);
+
+ /* Start IRQ auto-detection */
+ sti();
+ irqs = probe_irq_on();
+
+ /* Enable interrupts */
+ switch (h) {
+ case TYPE_PI:
+ case TYPE_PI2:
+ outb_p(0, io + PI_DREQ_MASK);
+ write_scc(cmd, R15, CTSIE);
+ write_scc(cmd, R0, RES_EXT_INT);
+ write_scc(cmd, R1, EXT_INT_ENAB);
+ break;
+ case TYPE_TWIN:
+ outb_p(0, io + TWIN_DMA_CFG);
+ inb_p(io + TWIN_CLR_TMR1);
+ inb_p(io + TWIN_CLR_TMR2);
+ outb_p(TWIN_EI, io + TWIN_SERIAL_CFG);
+ break;
+ }
+
+ /* Start timer */
+ outb_p(1, tmr + TMR_CNT1);
+ outb_p(0, tmr + TMR_CNT1);
+ /* Wait and detect IRQ */
+ time = jiffies; while (jiffies - time < 2 + HZ / TMR_0_HZ);
+ irq = probe_irq_off(irqs);
+
+ /* Clear pending interrupt, disable interrupts */
+ switch (h) {
+ case TYPE_PI:
+ case TYPE_PI2:
+ write_scc(cmd, R1, 0);
+ write_scc(cmd, R15, 0);
+ write_scc(cmd, R0, RES_EXT_INT);
+ break;
+ case TYPE_TWIN:
+ inb_p(io + TWIN_CLR_TMR1);
+ outb_p(0, io + TWIN_SERIAL_CFG);
+ break;
+ }
+
+ if (irq <= 0) {
+ printk("dmascc: could not find irq of %s at %#3x (irq=%d)\n",
+ hw[h].name, io, irq);
+ return -1;
+ }
+
+ /* Allocate memory */
+ info = kmalloc(sizeof(struct scc_info), GFP_KERNEL | GFP_DMA);
+ if (!info) {
+ printk("dmascc: could not allocate memory for %s at %#3x\n",
+ hw[h].name, io);
+ return -1;
+ }
+
+ /* Set up data structures */
+ memset(info, 0, sizeof(struct scc_info));
+ info->type = h;
+ info->chip = chip;
+ info->scc_base = io + hw[h].scc_offset;
+ info->tmr_base = io + hw[h].tmr_offset;
+ info->twin_serial_cfg = 0;
+ for (i = 0; i < 2; i++) {
+ dev = &info->dev[i];
+ priv = &info->priv[i];
+ sprintf(priv->name, "dmascc%i", 2*n+i);
+ priv->info = info;
+ priv->channel = i;
+ priv->cmd = info->scc_base + (i ? SCCB_CMD : SCCA_CMD);
+ priv->data = info->scc_base + (i ? SCCB_DATA : SCCA_DATA);
+ priv->tmr = info->tmr_base + (i ? TMR_CNT2 : TMR_CNT1);
+ priv->param.pclk_hz = hw[h].pclk_hz;
+ priv->param.brg_tc = -1;
+ priv->param.clocks = TCTRxCP | RCRTxCP;
+ priv->param.txdelay = TMR_0_HZ * 10 / 1000;
+ priv->param.txtime = HZ * 3;
+ priv->param.sqdelay = TMR_0_HZ * 1 / 1000;
+ priv->param.slottime = TMR_0_HZ * 10 / 1000;
+ priv->param.waittime = TMR_0_HZ * 100 / 1000;
+ priv->param.persist = 32;
+ priv->rx_task.routine = rx_bh;
+ priv->rx_task.data = dev;
+ dev->priv = priv;
+ dev->name = priv->name;
+ dev->base_addr = io;
+ dev->irq = irq;
+ dev->open = scc_open;
+ dev->stop = scc_close;
+ dev->do_ioctl = scc_ioctl;
+ dev->hard_start_xmit = scc_send_packet;
+ dev->get_stats = scc_get_stats;
+ dev->hard_header = ax25_encapsulate;
+ dev->rebuild_header = ax25_rebuild_header;
+ dev->set_mac_address = scc_set_mac_address;
+ dev->type = ARPHRD_AX25;
+ dev->hard_header_len = 73;
+ dev->mtu = 1500;
+ dev->addr_len = 7;
+ dev->tx_queue_len = 64;
+ memcpy(dev->broadcast, ax25_broadcast, 7);
+ memcpy(dev->dev_addr, ax25_test, 7);
+ dev->flags = 0;
+ dev_init_buffers(dev);
+ if (register_netdev(dev)) {
+ printk("dmascc: could not register %s\n", dev->name);
+ dev->name = NULL;
+ }
+ }
+
+ request_region(io, hw[h].io_size, "dmascc");
+
+ info->next = first;
+ first = info;
+ printk("dmascc: found %s (%s) at %#3x, irq %d\n", hw[h].name,
+ chipnames[chip], io, irq);
+ return 0;
+}
+
+
+/* Driver functions */
+
+static inline void write_scc(int ctl, int reg, int val)
+{
+ outb_p(reg, ctl);
+ outb_p(val, ctl);
+}
+
+
+static inline int read_scc(int ctl, int reg)
+{
+ outb_p(reg, ctl);
+ return inb_p(ctl);
+}
+
+
+static int scc_open(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int io = dev->base_addr;
+ int cmd = priv->cmd;
+
+ /* Request IRQ if not already used by other channel */
+ if (!info->open) {
+ if (request_irq(dev->irq, scc_isr, SA_INTERRUPT, "dmascc", info))
+ return -EAGAIN;
+ }
+
+ /* Request DMA if required */
+ if (dev->dma && request_dma(dev->dma, "dmascc")) {
+ if (!info->open) free_irq(dev->irq, info);
+ return -EAGAIN;
+ }
+
+ /* Initialize local variables */
+ dev->tbusy = 0;
+ priv->rx_ptr = 0;
+ priv->rx_over = 0;
+ priv->rx_head = priv->rx_tail = priv->rx_count = 0;
+ priv->tx_state = TX_IDLE;
+ priv->tx_head = priv->tx_tail = priv->tx_count = 0;
+ priv->tx_ptr = 0;
+ priv->tx_sem = 0;
+
+ /* Reset channel */
+ write_scc(cmd, R9, (priv->channel ? CHRB : CHRA) | MIE | NV);
+ /* X1 clock, SDLC mode */
+ write_scc(cmd, R4, SDLC | X1CLK);
+ /* DMA */
+ write_scc(cmd, R1, EXT_INT_ENAB | WT_FN_RDYFN);
+ /* 8 bit RX char, RX disable */
+ write_scc(cmd, R3, Rx8);
+ /* 8 bit TX char, TX disable */
+ write_scc(cmd, R5, Tx8);
+ /* SDLC address field */
+ write_scc(cmd, R6, 0);
+ /* SDLC flag */
+ write_scc(cmd, R7, FLAG);
+ switch (info->chip) {
+ case Z85C30:
+ /* Select WR7' */
+ write_scc(cmd, R15, 1);
+ /* Auto EOM reset */
+ write_scc(cmd, R7, 0x02);
+ write_scc(cmd, R15, 0);
+ break;
+ case Z85230:
+ /* Select WR7' */
+ write_scc(cmd, R15, 1);
+ /* RX FIFO half full (interrupt only), Auto EOM reset,
+ TX FIFO empty (DMA only) */
+ write_scc(cmd, R7, dev->dma ? 0x22 : 0x0a);
+ write_scc(cmd, R15, 0);
+ break;
+ }
+ /* Preset CRC, NRZ(I) encoding */
+ write_scc(cmd, R10, CRCPS | (priv->param.nrzi ? NRZI : NRZ));
+
+ /* Configure baud rate generator */
+ if (priv->param.brg_tc >= 0) {
+ /* Program BR generator */
+ write_scc(cmd, R12, priv->param.brg_tc & 0xFF);
+ write_scc(cmd, R13, (priv->param.brg_tc>>8) & 0xFF);
+ /* BRG source = SYS CLK; enable BRG; DTR REQ function (required by
+ PackeTwin, not connected on the PI2); set DPLL source to BRG */
+ write_scc(cmd, R14, SSBR | DTRREQ | BRSRC | BRENABL);
+ /* Enable DPLL */
+ write_scc(cmd, R14, SEARCH | DTRREQ | BRSRC | BRENABL);
+ } else {
+ /* Disable BR generator */
+ write_scc(cmd, R14, DTRREQ | BRSRC);
+ }
+
+ /* Configure clocks */
+ if (info->type == TYPE_TWIN) {
+ /* Disable external TX clock receiver */
+ outb_p((info->twin_serial_cfg &=
+ ~(priv->channel ? TWIN_EXTCLKB : TWIN_EXTCLKA)),
+ io + TWIN_SERIAL_CFG);
+ }
+ write_scc(cmd, R11, priv->param.clocks);
+ if ((info->type == TYPE_TWIN) && !(priv->param.clocks & TRxCOI)) {
+ /* Enable external TX clock receiver */
+ outb_p((info->twin_serial_cfg |=
+ (priv->channel ? TWIN_EXTCLKB : TWIN_EXTCLKA)),
+ io + TWIN_SERIAL_CFG);
+ }
+
+ /* Configure PackeTwin */
+ if (info->type == TYPE_TWIN) {
+ /* Assert DTR, enable interrupts */
+ outb_p((info->twin_serial_cfg |= TWIN_EI |
+ (priv->channel ? TWIN_DTRB_ON : TWIN_DTRA_ON)),
+ io + TWIN_SERIAL_CFG);
+ }
+
+ /* Read current status */
+ priv->status = read_scc(cmd, R0);
+ /* Enable SYNC, DCD, and CTS interrupts */
+ write_scc(cmd, R15, DCDIE | CTSIE | SYNCIE);
+
+ /* Configure PI2 DMA */
+ if (info->type <= TYPE_PI2) outb_p(1, io + PI_DREQ_MASK);
+
+ dev->start = 1;
+ info->open++;
+ MOD_INC_USE_COUNT;
+
+ return 0;
+}
+
+
+static int scc_close(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int io = dev->base_addr;
+ int cmd = priv->cmd;
+
+ dev->start = 0;
+ info->open--;
+ MOD_DEC_USE_COUNT;
+
+ if (info->type == TYPE_TWIN)
+ /* Drop DTR */
+ outb_p((info->twin_serial_cfg &=
+ (priv->channel ? ~TWIN_DTRB_ON : ~TWIN_DTRA_ON)),
+ io + TWIN_SERIAL_CFG);
+
+ /* Reset channel, free DMA */
+ write_scc(cmd, R9, (priv->channel ? CHRB : CHRA) | MIE | NV);
+ if (dev->dma) {
+ if (info->type == TYPE_TWIN) outb_p(0, io + TWIN_DMA_CFG);
+ free_dma(dev->dma);
+ }
+
+ if (!info->open) {
+ if (info->type <= TYPE_PI2) outb_p(0, io + PI_DREQ_MASK);
+ free_irq(dev->irq, info);
+ }
+ return 0;
+}
+
+
+static int scc_ioctl(struct device *dev, struct ifreq *ifr, int cmd)
+{
+ int rc;
+ struct scc_priv *priv = dev->priv;
+
+ switch (cmd) {
+ case SIOCGSCCPARAM:
+ rc = verify_area(VERIFY_WRITE, ifr->ifr_data, sizeof(struct scc_param));
+ if (rc) return rc;
+ copy_to_user(ifr->ifr_data, &priv->param, sizeof(struct scc_param));
+ return 0;
+ case SIOCSSCCPARAM:
+ if (!suser()) return -EPERM;
+ rc = verify_area(VERIFY_READ, ifr->ifr_data, sizeof(struct scc_param));
+ if (rc) return rc;
+ if (dev->start) return -EAGAIN;
+ copy_from_user(&priv->param, ifr->ifr_data, sizeof(struct scc_param));
+ dev->dma = priv->param.dma;
+ return 0;
+ default:
+ return -EINVAL;
+ }
+}
+
+
+static int scc_send_packet(struct sk_buff *skb, struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int cmd = priv->cmd;
+ unsigned long flags;
+ int i;
+
+ /* Block a timer-based transmit from overlapping */
+ if (test_and_set_bit(0, (void *) &priv->tx_sem) != 0) {
+ atomic_inc((void *) &priv->stats.tx_dropped);
+ dev_kfree_skb(skb, FREE_WRITE);
+ return 0;
+ }
+
+ /* Return with an error if we cannot accept more data */
+ if (dev->tbusy) {
+ priv->tx_sem = 0;
+ return -1;
+ }
+
+ /* Transfer data to DMA buffer */
+ i = priv->tx_head;
+ memcpy(priv->tx_buf[i], skb->data+1, skb->len-1);
+ priv->tx_len[i] = skb->len-1;
+
+ save_flags(flags);
+ cli();
+
+ /* Set the busy flag if we just filled up the last buffer */
+ priv->tx_head = (i + 1) % NUM_TX_BUF;
+ priv->tx_count++;
+ if (priv->tx_count == NUM_TX_BUF) dev->tbusy = 1;
+
+ /* Set new TX state */
+ if (priv->tx_state == TX_IDLE) {
+ /* Assert RTS, start timer */
+ priv->tx_state = TX_TXDELAY;
+ if (info->type <= TYPE_PI2) outb_p(0, dev->base_addr + PI_DREQ_MASK);
+ write_scc(cmd, R5, TxCRC_ENAB | RTS | TxENAB | Tx8);
+ if (info->type <= TYPE_PI2) outb_p(1, dev->base_addr + PI_DREQ_MASK);
+ priv->tx_start = jiffies;
+ delay(dev, priv->param.txdelay);
+ }
+
+ restore_flags(flags);
+
+ dev_kfree_skb(skb, FREE_WRITE);
+
+ priv->tx_sem = 0;
+ return 0;
+}
+
+
+static struct enet_statistics *scc_get_stats(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+
+ return &priv->stats;
+}
+
+
+static int scc_set_mac_address(struct device *dev, void *sa)
+{
+ memcpy(dev->dev_addr, ((struct sockaddr *)sa)->sa_data, dev->addr_len);
+ return 0;
+}
+
+
+static void scc_isr(int irq, void *dev_id, struct pt_regs * regs)
+{
+ struct scc_info *info = dev_id;
+ int is, io = info->dev[0].base_addr;
+
+ /* We're a fast IRQ handler and are called with interrupts disabled */
+
+ /* IRQ sharing doesn't make sense due to ISA's edge-triggered
+ interrupts, hence it is safe to return if we have found and
+ processed a single device. */
+
+ /* Interrupt processing: We loop until we know that the IRQ line is
+ low. If another positive edge occurs afterwards during the ISR,
+ another interrupt will be triggered by the interrupt controller
+ as soon as the IRQ level is enabled again (see asm/irq.h). */
+
+ switch (info->type) {
+ case TYPE_PI:
+ case TYPE_PI2:
+ outb_p(0, io + PI_DREQ_MASK);
+ z8530_isr(info);
+ outb_p(1, io + PI_DREQ_MASK);
+ return;
+ case TYPE_TWIN:
+ while ((is = ~inb_p(io + TWIN_INT_REG)) &
+ TWIN_INT_MSK) {
+ if (is & TWIN_SCC_MSK) {
+ z8530_isr(info);
+ } else if (is & TWIN_TMR1_MSK) {
+ inb_p(io + TWIN_CLR_TMR1);
+ tm_isr(&info->dev[0]);
+ } else {
+ inb_p(io + TWIN_CLR_TMR2);
+ tm_isr(&info->dev[1]);
+ }
+ }
+ /* No interrupts pending from the PackeTwin */
+ return;
+ }
+}
+
+
+static inline void z8530_isr(struct scc_info *info)
+{
+ int is, a_cmd;
+
+ a_cmd = info->scc_base + SCCA_CMD;
+
+ while ((is = read_scc(a_cmd, R3))) {
+ if (is & CHARxIP) {
+ rx_isr(&info->dev[0]);
+ } else if (is & CHATxIP) {
+ tx_isr(&info->dev[0]);
+ } else if (is & CHAEXT) {
+ es_isr(&info->dev[0]);
+ } else if (is & CHBRxIP) {
+ rx_isr(&info->dev[1]);
+ } else if (is & CHBTxIP) {
+ tx_isr(&info->dev[1]);
+ } else {
+ es_isr(&info->dev[1]);
+ }
+ }
+ /* Ok, no interrupts pending from this 8530. The INT line should
+ be inactive now. */
+}
+
+
+static void rx_isr(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ int cmd = priv->cmd;
+
+ if (dev->dma) {
+ /* Check special condition and perform error reset. See 2.4.7.5. */
+ special_condition(dev, read_scc(cmd, R1));
+ write_scc(cmd, R0, ERR_RES);
+ } else {
+ /* Check special condition for each character. Error reset not necessary.
+ Same algorithm for SCC and ESCC. See 2.4.7.1 and 2.4.7.4. */
+ int rc;
+ while (read_scc(cmd, R0) & Rx_CH_AV) {
+ rc = read_scc(cmd, R1);
+ if (priv->rx_ptr < BUF_SIZE)
+ priv->rx_buf[priv->rx_head][priv->rx_ptr++] = read_scc(cmd, R8);
+ else {
+ priv->rx_over = 2;
+ read_scc(cmd, R8);
+ }
+ special_condition(dev, rc);
+ }
+ }
+}
+
+
+static void special_condition(struct device *dev, int rc)
+{
+ struct scc_priv *priv = dev->priv;
+ int cb, cmd = priv->cmd;
+
+ /* See Figure 2-15. Only overrun and EOF need to be checked. */
+
+ if (rc & Rx_OVR) {
+ /* Receiver overrun */
+ priv->rx_over = 1;
+ if (!dev->dma) write_scc(cmd, R0, ERR_RES);
+ } else if (rc & END_FR) {
+ /* End of frame. Get byte count */
+ if (dev->dma) {
+ disable_dma(dev->dma);
+ clear_dma_ff(dev->dma);
+ cb = BUF_SIZE - get_dma_residue(dev->dma) - 2;
+ } else {
+ cb = priv->rx_ptr - 2;
+ }
+ if (priv->rx_over) {
+ /* We had an overrun */
+ priv->stats.rx_errors++;
+ if (priv->rx_over == 2) priv->stats.rx_length_errors++;
+ else priv->stats.rx_fifo_errors++;
+ priv->rx_over = 0;
+ } else if (rc & CRC_ERR) {
+ /* Count invalid CRC only if packet length >= minimum */
+ if (cb >= 8) {
+ priv->stats.rx_errors++;
+ priv->stats.rx_crc_errors++;
+ }
+ } else {
+ if (cb >= 8) {
+ /* Put good frame in FIFO */
+ priv->rx_len[priv->rx_head] = cb;
+ priv->rx_head = (priv->rx_head + 1) % NUM_RX_BUF;
+ priv->rx_count++;
+ if (priv->rx_count == NUM_RX_BUF) {
+ /* Disable receiver if FIFO full */
+ write_scc(cmd, R3, Rx8);
+ priv->stats.rx_errors++;
+ priv->stats.rx_over_errors++;
+ }
+ /* Mark bottom half handler */
+ queue_task(&priv->rx_task, &tq_immediate);
+ mark_bh(IMMEDIATE_BH);
+ }
+ }
+ /* Get ready for new frame */
+ if (dev->dma) {
+ set_dma_addr(dev->dma, (int) priv->rx_buf[priv->rx_head]);
+ set_dma_count(dev->dma, BUF_SIZE);
+ enable_dma(dev->dma);
+ } else {
+ priv->rx_ptr = 0;
+ }
+ }
+}
+
+
+static void rx_bh(void *arg)
+{
+ struct device *dev = arg;
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int cmd = priv->cmd;
+ int i = priv->rx_tail;
+ int cb;
+ unsigned long flags;
+ struct sk_buff *skb;
+ unsigned char *data;
+
+ save_flags(flags);
+ cli();
+
+ while (priv->rx_count) {
+ restore_flags(flags);
+ cb = priv->rx_len[i];
+ /* Allocate buffer */
+ skb = dev_alloc_skb(cb+1);
+ if (skb == NULL) {
+ /* Drop packet */
+ priv->stats.rx_dropped++;
+ } else {
+ /* Fill buffer */
+ data = skb_put(skb, cb+1);
+ data[0] = 0;
+ memcpy(&data[1], priv->rx_buf[i], cb);
+ skb->dev = dev;
+ skb->protocol = ntohs(ETH_P_AX25);
+ skb->mac.raw = skb->data;
+ netif_rx(skb);
+ priv->stats.rx_packets++;
+ }
+ save_flags(flags);
+ cli();
+ /* Enable receiver if RX buffers have been unavailable */
+ if ((priv->rx_count == NUM_RX_BUF) && (priv->status & DCD)) {
+ if (info->type <= TYPE_PI2) outb_p(0, dev->base_addr + PI_DREQ_MASK);
+ write_scc(cmd, R3, RxENABLE | Rx8 | RxCRC_ENAB);
+ if (info->type <= TYPE_PI2) outb_p(1, dev->base_addr + PI_DREQ_MASK);
+ }
+ /* Move tail */
+ priv->rx_tail = i = (i + 1) % NUM_RX_BUF;
+ priv->rx_count--;
+ }
+
+ restore_flags(flags);
+}
+
+
+static void tx_isr(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ int cmd = priv->cmd;
+ int i = priv->tx_tail, p = priv->tx_ptr;
+
+ /* Suspend TX interrupts if we don't want to send anything.
+ See Figure 2-22. */
+ if (p == priv->tx_len[i]) {
+ write_scc(cmd, R0, RES_Tx_P);
+ return;
+ }
+
+ /* Write characters */
+ while ((read_scc(cmd, R0) & Tx_BUF_EMP) && p < priv->tx_len[i]) {
+ write_scc(cmd, R8, priv->tx_buf[i][p++]);
+ }
+ priv->tx_ptr = p;
+
+}
+
+
+static void es_isr(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int i, cmd = priv->cmd;
+ int st, dst, res;
+
+ /* Read status and reset interrupt bit */
+ st = read_scc(cmd, R0);
+ write_scc(cmd, R0, RES_EXT_INT);
+ dst = priv->status ^ st;
+ priv->status = st;
+
+ /* Since the EOM latch is reset automatically, we assume that
+ it has been zero if and only if we are in the TX_ACTIVE state.
+ Otherwise we follow 2.4.9.6. */
+
+ /* Transmit underrun */
+ if ((priv->tx_state == TX_ACTIVE) && (st & TxEOM)) {
+ /* Get remaining bytes */
+ i = priv->tx_tail;
+ if (dev->dma) {
+ disable_dma(dev->dma);
+ clear_dma_ff(dev->dma);
+ res = get_dma_residue(dev->dma);
+ } else {
+ res = priv->tx_len[i] - priv->tx_ptr;
+ if (res) write_scc(cmd, R0, RES_Tx_P);
+ priv->tx_ptr = 0;
+ }
+ /* Remove frame from FIFO */
+ priv->tx_tail = (i + 1) % NUM_TX_BUF;
+ priv->tx_count--;
+ dev->tbusy = 0;
+ /* Check if another frame is available and we are allowed to transmit */
+ if (priv->tx_count && (jiffies - priv->tx_start) < priv->param.txtime) {
+ if (dev->dma) {
+ set_dma_addr(dev->dma, (int) priv->tx_buf[priv->tx_tail]);
+ set_dma_count(dev->dma, priv->tx_len[priv->tx_tail]);
+ enable_dma(dev->dma);
+ } else {
+ /* If we have an ESCC, we are allowed to write data bytes
+ immediately. Otherwise we have to wait for the next
+ TX interrupt. See Figure 2-22. */
+ if (info->chip == Z85230) {
+ tx_isr(dev);
+ }
+ }
+ } else {
+ /* No frame available. Disable interrupts. */
+ priv->tx_state = TX_SQDELAY;
+ delay(dev, priv->param.sqdelay);
+ write_scc(cmd, R15, DCDIE | CTSIE | SYNCIE);
+ write_scc(cmd, R1, EXT_INT_ENAB | WT_FN_RDYFN);
+ }
+ /* Update packet statistics */
+ if (res) {
+ priv->stats.tx_errors++;
+ priv->stats.tx_fifo_errors++;
+ } else {
+ priv->stats.tx_packets++;
+ }
+ /* Inform upper layers */
+ mark_bh(NET_BH);
+ }
+
+ /* DCD transition */
+ if ((priv->tx_state < TX_TXDELAY) && (dst & DCD)) {
+ /* Transmitter state change */
+ priv->tx_state = TX_OFF;
+ /* Enable or disable receiver */
+ if (st & DCD) {
+ if (dev->dma) {
+ /* Program DMA controller */
+ disable_dma(dev->dma);
+ clear_dma_ff(dev->dma);
+ set_dma_mode(dev->dma, DMA_MODE_READ);
+ set_dma_addr(dev->dma, (int) priv->rx_buf[priv->rx_head]);
+ set_dma_count(dev->dma, BUF_SIZE);
+ enable_dma(dev->dma);
+ /* Configure PackeTwin DMA */
+ if (info->type == TYPE_TWIN) {
+ outb_p((dev->dma == 1) ? TWIN_DMA_HDX_R1 : TWIN_DMA_HDX_R3,
+ dev->base_addr + TWIN_DMA_CFG);
+ }
+ /* Sp. cond. intr. only, ext int enable */
+ write_scc(cmd, R1, EXT_INT_ENAB | INT_ERR_Rx |
+ WT_RDY_RT | WT_FN_RDYFN | WT_RDY_ENAB);
+ } else {
+ /* Intr. on all Rx characters and Sp. cond., ext int enable */
+ write_scc(cmd, R1, EXT_INT_ENAB | INT_ALL_Rx | WT_RDY_RT |
+ WT_FN_RDYFN);
+ }
+ if (priv->rx_count < NUM_RX_BUF) {
+ /* Enable receiver */
+ write_scc(cmd, R3, RxENABLE | Rx8 | RxCRC_ENAB);
+ }
+ } else {
+ /* Disable DMA */
+ if (dev->dma) disable_dma(dev->dma);
+ /* Disable receiver */
+ write_scc(cmd, R3, Rx8);
+ /* DMA disable, RX int disable, Ext int enable */
+ write_scc(cmd, R1, EXT_INT_ENAB | WT_RDY_RT | WT_FN_RDYFN);
+ /* Transmitter state change */
+ if (random() > priv->param.persist)
+ delay(dev, priv->param.slottime);
+ else {
+ if (priv->tx_count) {
+ priv->tx_state = TX_TXDELAY;
+ write_scc(cmd, R5, TxCRC_ENAB | RTS | TxENAB | Tx8);
+ priv->tx_start = jiffies;
+ delay(dev, priv->param.txdelay);
+ } else {
+ priv->tx_state = TX_IDLE;
+ }
+ }
+ }
+ }
+
+ /* CTS transition */
+ if ((info->type <= TYPE_PI2) && (dst & CTS) && (~st & CTS)) {
+ /* Timer has expired */
+ tm_isr(dev);
+ }
+
+ /* /SYNC/HUNT transition */
+ if ((dst & SYNC_HUNT) && (~st & SYNC_HUNT)) {
+ /* Reset current frame and clear RX FIFO */
+ while (read_scc(cmd, R0) & Rx_CH_AV) read_scc(cmd, R8);
+ priv->rx_over = 0;
+ if (dev->dma) {
+ disable_dma(dev->dma);
+ clear_dma_ff(dev->dma);
+ set_dma_addr(dev->dma, (int) priv->rx_buf[priv->rx_head]);
+ set_dma_count(dev->dma, BUF_SIZE);
+ enable_dma(dev->dma);
+ } else {
+ priv->rx_ptr = 0;
+ }
+ }
+}
+
+
+static void tm_isr(struct device *dev)
+{
+ struct scc_priv *priv = dev->priv;
+ struct scc_info *info = priv->info;
+ int cmd = priv->cmd;
+
+ switch (priv->tx_state) {
+ case TX_OFF:
+ if (~priv->status & DCD) {
+ if (random() > priv->param.persist) delay(dev, priv->param.slottime);
+ else {
+ if (priv->tx_count) {
+ priv->tx_state = TX_TXDELAY;
+ write_scc(cmd, R5, TxCRC_ENAB | RTS | TxENAB | Tx8);
+ priv->tx_start = jiffies;
+ delay(dev, priv->param.txdelay);
+ } else {
+ priv->tx_state = TX_IDLE;
+ }
+ }
+ }
+ break;
+ case TX_TXDELAY:
+ priv->tx_state = TX_ACTIVE;
+ if (dev->dma) {
+ /* Program DMA controller */
+ disable_dma(dev->dma);
+ clear_dma_ff(dev->dma);
+ set_dma_mode(dev->dma, DMA_MODE_WRITE);
+ set_dma_addr(dev->dma, (int) priv->tx_buf[priv->tx_tail]);
+ set_dma_count(dev->dma, priv->tx_len[priv->tx_tail]);
+ enable_dma(dev->dma);
+ /* Configure PackeTwin DMA */
+ if (info->type == TYPE_TWIN) {
+ outb_p((dev->dma == 1) ? TWIN_DMA_HDX_T1 : TWIN_DMA_HDX_T3,
+ dev->base_addr + TWIN_DMA_CFG);
+ }
+ /* Enable interrupts and DMA. On the PackeTwin, the DTR//REQ pin
+ is used for TX DMA requests, but we enable the WAIT/DMA request
+ pin, anyway */
+ write_scc(cmd, R15, TxUIE | DCDIE | CTSIE | SYNCIE);
+ write_scc(cmd, R1, EXT_INT_ENAB | WT_FN_RDYFN | WT_RDY_ENAB);
+ } else {
+ write_scc(cmd, R15, TxUIE | DCDIE | CTSIE | SYNCIE);
+ write_scc(cmd, R1, EXT_INT_ENAB | WT_FN_RDYFN | TxINT_ENAB);
+ tx_isr(dev);
+ }
+ if (info->chip == Z8530) write_scc(cmd, R0, RES_EOM_L);
+ break;
+ case TX_SQDELAY:
+ /* Disable transmitter */
+ write_scc(cmd, R5, TxCRC_ENAB | Tx8);
+ /* Transmitter state change: Switch to TX_OFF and wait at least
+ 1 slottime. */
+ priv->tx_state = TX_OFF;
+ if (~priv->status & DCD) delay(dev, priv->param.waittime);
+ }
+}
+
+
+static inline void delay(struct device *dev, int t)
+{
+ struct scc_priv *priv = dev->priv;
+ int tmr = priv->tmr;
+
+ outb_p(t & 0xFF, tmr);
+ outb_p((t >> 8) & 0xFF, tmr);
+}
+
+
+static inline unsigned char random(void)
+{
+ /* See "Numerical Recipes in C", second edition, p. 284 */
+ rand = rand * 1664525L + 1013904223L;
+ return (unsigned char) (rand >> 24);
+}
+
+
FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen, slshen@lbl.gov