Added ethernet driver for EP93XX SoCs
Signed-off-by: Matthias Kaehlcke <matthias(a)kaehlcke.net>
---
drivers/net/Makefile | 1 +
drivers/net/ep93xx.c | 677 ++++++++++++++++++++++++++++++++++++++++++++++++++
drivers/net/ep93xx.h | 153 ++++++++++++
include/common.h | 10 +-
4 files changed, 840 insertions(+), 1 deletions(-)
create mode 100644 drivers/net/ep93xx.c
create mode 100644 drivers/net/ep93xx.h
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 904727e..0694133 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -37,6 +37,7 @@ COBJS-$(CONFIG_DNET) += dnet.o
COBJS-$(CONFIG_E1000) += e1000.o
COBJS-$(CONFIG_EEPRO100) += eepro100.o
COBJS-$(CONFIG_ENC28J60) += enc28j60.o
+COBJS-$(CONFIG_EP93XX) += ep93xx.o
COBJS-$(CONFIG_FEC_MXC) += fec_mxc.o
COBJS-$(CONFIG_FSLDMAFEC) += fsl_mcdmafec.o mcfmii.o
COBJS-$(CONFIG_FTMAC100) += ftmac100.o
diff --git a/drivers/net/ep93xx.c b/drivers/net/ep93xx.c
new file mode 100644
index 0000000..1eb48ca
--- /dev/null
+++ b/drivers/net/ep93xx.c
@@ -0,0 +1,677 @@
+/*
+ * Cirrus Logic EP93xx ethernet MAC / MII driver.
+ *
+ * Copyright (C) 2009 Matthias Kaehlcke <matthias(a)kaehlcke.net>
+ *
+ * Copyright (C) 2004, 2005
+ * Cory T. Tusar, Videon Central, Inc., <ctusar(a)videon-central.com>
+ *
+ * Based on the original eth.[ch] Cirrus Logic EP93xx Rev D. Ethernet Driver,
+ * which is
+ *
+ * (C) Copyright 2002 2003
+ * Adam Bezanson, Network Audio Technologies, Inc.
+ * <bezanson(a)netaudiotech.com>
+ *
+ * See file CREDITS for list of people who contributed to this project.
+ *
+ * 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 <command.h>
+#include <common.h>
+#include <asm/arch/ep93xx.h>
+#include <asm/io.h>
+#include <malloc.h>
+#include <miiphy.h>
+#include <linux/types.h>
+#include "ep93xx.h"
+
+static int ep93xx_eth_send_packet(struct eth_device *dev,
+ volatile void * const packet, int const length);
+static int ep93xx_eth_rcv_packet(struct eth_device *dev);
+
+/* ----------------------------------------------------------------------------
+ * EP93xx ethernet MAC functionality
+ */
+#if defined(CONFIG_DRIVER_EP93XX_MAC)
+
+/* ep93xx_miiphy ops forward declarations */
+static int ep93xx_miiphy_read(char * const dev, unsigned char const addr,
+ unsigned char const reg, unsigned short * const value);
+static int ep93xx_miiphy_write(char * const dev, unsigned char const addr,
+ unsigned char const reg, unsigned short const value);
+
+/* Reserve memory for the MAC's private data */
+static struct ep93xx_mac ep93xx_mac = { 0 };
+
+#if defined(EP93XX_MAC_DEBUG)
+/**
+ * Dump ep93xx_mac values to the terminal.
+ */
+inline void dump_dev(void)
+{
+ int i;
+
+ printf("\ndump_dev()\n");
+ printf(" is_initialized %02X\n", ep93xx_mac.is_initialized);
+ printf(" rx_dq.base %08X\n", ep93xx_mac.rx_dq.base);
+ printf(" rx_dq.current %08X\n", ep93xx_mac.rx_dq.current);
+ printf(" rx_dq.end %08X\n", ep93xx_mac.rx_dq.end);
+ printf(" rx_sq.base %08X\n", ep93xx_mac.rx_sq.base);
+ printf(" rx_sq.current %08X\n", ep93xx_mac.rx_sq.current);
+ printf(" rx_sq.end %08X\n", ep93xx_mac.rx_sq.end);
+
+ for (i = 0; i < NUMRXDESC; i++)
+ printf(" rx_buffer[%2.d] %08X\n", i, NetRxPackets[i]);
+
+ printf(" tx_dq.base %08X\n", ep93xx_mac.tx_dq.base);
+ printf(" tx_dq.current %08X\n", ep93xx_mac.tx_dq.current);
+ printf(" tx_dq.end %08X\n", ep93xx_mac.tx_dq.end);
+ printf(" tx_sq.base %08X\n", ep93xx_mac.tx_sq.base);
+ printf(" tx_sq.current %08X\n", ep93xx_mac.tx_sq.current);
+ printf(" tx_sq.end %08X\n", ep93xx_mac.tx_sq.end);
+}
+#else
+#define dump_dev(x)
+#endif /* defined(EP93XX_MAC_DEBUG) */
+
+#if defined(EP93XX_MAC_DEBUG)
+/**
+ * Dump all RX descriptor queue entries to the terminal.
+ */
+inline void dump_rx_descriptor_queue(void)
+{
+ int i;
+
+ printf("\ndump_rx_descriptor_queue()\n");
+ printf(" descriptor address word1 word2\n");
+ for (i = 0; i < NUMRXDESC; i++) {
+ printf(" [ %08X ] %08X %08X\n",
+ (ep93xx_mac.rx_dq.base + i),
+ (ep93xx_mac.rx_dq.base + i)->word1,
+ (ep93xx_mac.rx_dq.base + i)->word2);
+ }
+}
+#else
+#define dump_rx_descriptor_queue()
+#endif /* defined(EP93XX_MAC_DEBUG) */
+
+#if defined(EP93XX_MAC_DEBUG)
+/**
+ * Dump all RX status queue entries to the terminal.
+ */
+inline void dump_rx_status_queue(void)
+{
+ int i;
+
+ printf("\ndump_rx_status_queue()\n");
+ printf(" descriptor address word1 word2\n");
+ for (i = 0; i < NUMRXDESC; i++) {
+ printf(" [ %08X ] %08X %08X\n",
+ (ep93xx_mac.rx_sq.base + i),
+ (ep93xx_mac.rx_sq.base + i)->word1,
+ (ep93xx_mac.rx_sq.base + i)->word2);
+ }
+}
+#else
+#define dump_rx_status_queue()
+#endif /* defined(EP93XX_MAC_DEBUG) */
+
+#if defined(EP93XX_MAC_DEBUG)
+/**
+ * Dump all TX descriptor queue entries to the terminal.
+ */
+inline void dump_tx_descriptor_queue(void)
+{
+ int i;
+
+ printf("\ndump_tx_descriptor_queue()\n");
+ printf(" descriptor address word1 word2\n");
+ for (i = 0; i < NUMTXDESC; i++) {
+ printf(" [ %08X ] %08X %08X\n",
+ (ep93xx_mac.tx_dq.base + i),
+ (ep93xx_mac.tx_dq.base + i)->word1,
+ (ep93xx_mac.tx_dq.base + i)->word2);
+ }
+}
+#else
+#define dump_tx_descriptor_queue()
+#endif /* defined(EP93XX_MAC_DEBUG) */
+
+#if defined(EP93XX_MAC_DEBUG)
+/**
+ * Dump all TX status queue entries to the terminal.
+ */
+inline void dump_tx_status_queue(void)
+{
+ int i;
+
+ printf("\ndump_tx_status_queue()\n");
+ printf(" descriptor address word1\n");
+ for (i = 0; i < NUMTXDESC; i++) {
+ printf(" [ %08X ] %08X\n",
+ (ep93xx_mac.rx_sq.base + i),
+ (ep93xx_mac.rx_sq.base + i)->word1);
+ }
+}
+#else
+#define dump_tx_status_queue()
+#endif /* defined(EP93XX_MAC_DEBUG) */
+
+/**
+ * Reset the EP93xx MAC by twiddling the soft reset bit and spinning until
+ * it's cleared.
+ */
+static void ep93xx_mac_reset(void)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+
+ debug("+ep93xx_mac_reset");
+
+ uint32_t value = readl(&mac->selfctl);
+ value |= SELFCTL_RESET;
+ writel(value, &mac->selfctl);
+
+ while (readl(&mac->selfctl) & SELFCTL_RESET)
+ ; /* noop */
+
+ debug("-ep93xx_mac_reset");
+}
+
+/* Eth device open */
+static int ep93xx_eth_open(struct eth_device *dev, bd_t *bd)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+ uchar enetaddr[6];
+ int i;
+
+ /* Reset the descriptor queues' current and end address values */
+ ep93xx_mac.tx_dq.current = ep93xx_mac.tx_dq.base;
+ ep93xx_mac.tx_dq.end = (ep93xx_mac.tx_dq.base + NUMTXDESC);
+
+ ep93xx_mac.tx_sq.current = ep93xx_mac.tx_sq.base;
+ ep93xx_mac.tx_sq.end = (ep93xx_mac.tx_sq.base + NUMTXDESC);
+
+ ep93xx_mac.rx_dq.current = ep93xx_mac.rx_dq.base;
+ ep93xx_mac.rx_dq.end = (ep93xx_mac.rx_dq.base + NUMRXDESC);
+
+ ep93xx_mac.rx_sq.current = ep93xx_mac.rx_sq.base;
+ ep93xx_mac.rx_sq.end = (ep93xx_mac.rx_sq.base + NUMRXDESC);
+
+ /*
+ * Set the transmit descriptor and status queues' base address,
+ * current address, and length registers. Set the maximum frame
+ * length and threshold. Enable the transmit descriptor processor.
+ */
+ writel((uint32_t)ep93xx_mac.tx_dq.base, &mac->txdq.badd);
+ writel((uint32_t)ep93xx_mac.tx_dq.base, &mac->txdq.curadd);
+ writel(sizeof(struct tx_descriptor) * NUMTXDESC, &mac->txdq.blen);
+
+ writel((uint32_t)ep93xx_mac.tx_sq.base, &mac->txstsq.badd);
+ writel((uint32_t)ep93xx_mac.tx_sq.base, &mac->txstsq.curadd);
+ writel(sizeof(struct tx_status) * NUMTXDESC, &mac->txstsq.blen);
+
+ writel(0x00040000, &mac->txdthrshld);
+ writel(0x00040000, &mac->txststhrshld);
+
+ writel((TXSTARTMAX << 0) | (PKTSIZE_ALIGN << 16), &mac->maxfrmlen);
+ writel(BMCTL_TXEN, &mac->bmctl);
+
+ /*
+ * Set the receive descriptor and status queues' base address,
+ * current address, and length registers. Enable the receive
+ * descriptor processor.
+ */
+ writel((uint32_t)ep93xx_mac.rx_dq.base, &mac->rxdq.badd);
+ writel((uint32_t)ep93xx_mac.rx_dq.base, &mac->rxdq.curadd);
+ writel(sizeof(struct rx_descriptor) * NUMRXDESC, &mac->rxdq.blen);
+
+ writel((uint32_t)ep93xx_mac.rx_sq.base, &mac->rxstsq.badd);
+ writel((uint32_t)ep93xx_mac.rx_sq.base, &mac->rxstsq.curadd);
+ writel(sizeof(struct rx_status) * NUMRXDESC, &mac->rxstsq.blen);
+
+ writel(0x00040000, &mac->rxdthrshld);
+
+ writel(BMCTL_RXEN, &mac->bmctl);
+
+ writel(0x00040000, &mac->rxststhrshld);
+
+ /* Wait until the receive descriptor processor is active */
+ while (!(readl(&mac->bmsts) & BMSTS_RXACT))
+ ; /* noop */
+
+ /*
+ * Initialize the RX descriptor queue. Clear the TX descriptor queue.
+ * Clear the RX and TX status queues. Enqueue the RX descriptor and
+ * status entries to the MAC.
+ */
+ for (i = 0; i < NUMRXDESC; i++) {
+ /* set buffer address */
+ (ep93xx_mac.rx_dq.base + i)->word1 = (uint32_t)NetRxPackets[i];
+
+ /* set buffer length, clear buffer index and NSOF */
+ (ep93xx_mac.rx_dq.base + i)->word2 = PKTSIZE_ALIGN;
+ }
+
+ memset(ep93xx_mac.tx_dq.base, 0,
+ (sizeof(struct tx_descriptor) * NUMTXDESC));
+ memset(ep93xx_mac.rx_sq.base, 0,
+ (sizeof(struct rx_status) * NUMRXDESC));
+ memset(ep93xx_mac.tx_sq.base, 0,
+ (sizeof(struct tx_status) * NUMTXDESC));
+
+ writel(NUMRXDESC, &mac->rxdqenq);
+ writel(NUMRXDESC, &mac->rxstsqenq);
+
+ /* Set the primary MAC address */
+ writel(AFP_IAPRIMARY, &mac->afp);
+ eth_getenv_enetaddr("ethaddr", enetaddr);
+ writel(enetaddr[0] | (enetaddr[1] << 8) |
+ (enetaddr[2] << 16) | (enetaddr[3] << 24),
+ &mac->indad);
+ writel(enetaddr[4] | (enetaddr[5] << 8), &mac->indad_upper);
+
+ /* Turn on RX and TX */
+ writel(RXCTL_IA0 | RXCTL_BA | RXCTL_SRXON |
+ RXCTL_RCRCA | RXCTL_MA, &mac->rxctl);
+ writel(TXCTL_STXON, &mac->txctl);
+
+ /* Dump data structures if we're debugging */
+ dump_dev();
+ dump_rx_descriptor_queue();
+ dump_rx_status_queue();
+ dump_tx_descriptor_queue();
+ dump_tx_status_queue();
+
+ return 1;
+}
+
+/**
+ * Halt EP93xx MAC transmit and receive by clearing the TxCTL and RxCTL
+ * registers.
+ */
+static void ep93xx_eth_close(struct eth_device *dev)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+
+ debug("+ep93xx_eth_close");
+
+ writel(0x00000000, &mac->rxctl);
+ writel(0x00000000, &mac->txctl);
+
+ debug("-ep93xx_eth_close");
+}
+
+#if defined(CONFIG_MII)
+int ep93xx_miiphy_initialize(bd_t * const bd)
+{
+ miiphy_register("ep93xx_eth0", ep93xx_miiphy_read, ep93xx_miiphy_write);
+ return 0;
+}
+#endif
+
+/**
+ * Initialize the EP93xx MAC. The MAC hardware is reset. Buffers are
+ * allocated, if necessary, for the TX and RX descriptor and status queues,
+ * as well as for received packets. The EP93XX MAC hardware is initialized.
+ * Transmit and receive operations are enabled.
+ */
+int ep93xx_eth_init(bd_t * const bd)
+{
+ int ret = ETH_STATUS_FAILURE;
+ struct eth_device *dev;
+
+ debug("+ep93xx_eth_init");
+
+ /* Parameter check */
+ BUG_ON(bd == NULL);
+
+ /* Reset the MAC */
+ ep93xx_mac_reset();
+
+ /*
+ * Allocate space for the queues and RX packet buffers if we're not
+ * already initialized
+ */
+ if (!ep93xx_mac.is_initialized) {
+ ep93xx_mac.tx_dq.base = calloc(NUMTXDESC,
+ sizeof(struct tx_descriptor));
+ if (ep93xx_mac.tx_dq.base == NULL) {
+ ERROR("calloc() failed");
+ goto eth_init_failed_0;
+ }
+
+ ep93xx_mac.tx_sq.base = calloc(NUMTXDESC,
+ sizeof(struct tx_status));
+ if (ep93xx_mac.tx_sq.base == NULL) {
+ ERROR("calloc() failed");
+ goto eth_init_failed_1;
+ }
+
+ ep93xx_mac.rx_dq.base = calloc(NUMRXDESC,
+ sizeof(struct rx_descriptor));
+ if (ep93xx_mac.rx_dq.base == NULL) {
+ ERROR("calloc() failed");
+ goto eth_init_failed_2;
+ }
+
+ ep93xx_mac.rx_sq.base = calloc(NUMRXDESC,
+ sizeof(struct rx_status));
+ if (ep93xx_mac.rx_sq.base == NULL) {
+ ERROR("calloc() failed");
+ goto eth_init_failed_3;
+ }
+
+ dev = malloc(sizeof *dev);
+ if (dev == NULL) {
+ ERROR("malloc() failed");
+ goto eth_init_failed_4;
+ }
+
+ memset(dev, 0, sizeof *dev);
+ sprintf(dev->name, "ep93xx MAC");
+ dev->iobase = 0;
+ dev->init = ep93xx_eth_open;
+ dev->halt = ep93xx_eth_close;
+ dev->send = ep93xx_eth_send_packet;
+ dev->recv = ep93xx_eth_rcv_packet;
+
+ eth_register(dev);
+
+ /*
+ * Set is_initialized flag so we don't go through allocation
+ * portion of init again.
+ */
+ ep93xx_mac.is_initialized = 1;
+ }
+
+ /* Done! */
+ ret = ETH_STATUS_SUCCESS;
+ goto eth_init_done;
+
+eth_init_failed_4:
+ free(ep93xx_mac.rx_sq.base);
+ /* Fall through */
+
+eth_init_failed_3:
+ free(ep93xx_mac.rx_dq.base);
+ /* Fall through */
+
+eth_init_failed_2:
+ free(ep93xx_mac.tx_sq.base);
+ /* Fall through */
+
+eth_init_failed_1:
+ free(ep93xx_mac.tx_dq.base);
+ /* Fall through */
+
+eth_init_failed_0:
+ /* Fall through */
+
+eth_init_done:
+ debug("-ep93xx_eth_init %d", ret);
+ return ret;
+}
+
+/**
+ * Copy a frame of data from the MAC into the protocol layer for further
+ * processing.
+ *
+ * TODO: Enhance this to deal with as many packets as are available at
+ * the MAC at one time? */
+static int ep93xx_eth_rcv_packet(struct eth_device *dev)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+ int ret = ETH_STATUS_FAILURE;
+
+ debug("+ep93xx_eth_rcv_packet");
+
+ if (RX_STATUS_RFP(ep93xx_mac.rx_sq.current)) {
+ if (RX_STATUS_RWE(ep93xx_mac.rx_sq.current)) {
+ /*
+ * We have a good frame. Extract the frame's length
+ * from the current rx_status_queue entry, and copy
+ * the frame's data into NetRxPackets[] of the
+ * protocol stack. We track the total number of
+ * bytes in the frame (nbytes_frame) which will be
+ * used when we pass the data off to the protocol
+ * layer via NetReceive().
+ */
+ NetReceive((uchar *)ep93xx_mac.rx_dq.current->word1,
+ RX_STATUS_FRAME_LEN(ep93xx_mac.rx_sq.current));
+ debug("reporting %d bytes...\n",
+ ep93xx_mac.rx_sq.current->frame_length);
+
+ ret = ETH_STATUS_SUCCESS;
+
+ } else {
+ /* Do we have an erroneous packet? */
+ ERROR("packet rx error, status %08X %08X",
+ ep93xx_mac.rx_sq.current->word1,
+ ep93xx_mac.rx_sq.current->word2);
+ dump_rx_descriptor_queue();
+ dump_rx_status_queue();
+ }
+
+ /*
+ * Clear the associated status queue entry, and
+ * increment our current pointers to the next RX
+ * descriptor and status queue entries (making sure
+ * we wrap properly).
+ */
+ memset((void *)ep93xx_mac.rx_sq.current, 0,
+ sizeof(struct rx_status));
+
+ ep93xx_mac.rx_sq.current++;
+ if (ep93xx_mac.rx_sq.current >= ep93xx_mac.rx_sq.end)
+ ep93xx_mac.rx_sq.current = ep93xx_mac.rx_sq.base;
+
+ ep93xx_mac.rx_dq.current++;
+ if (ep93xx_mac.rx_dq.current >= ep93xx_mac.rx_dq.end)
+ ep93xx_mac.rx_dq.current = ep93xx_mac.rx_dq.base;
+
+ /*
+ * Finally, return the RX descriptor and status entries
+ * back to the MAC engine, and loop again, checking for
+ * more descriptors to process.
+ */
+ writel(1, &mac->rxdqenq);
+ writel(1, &mac->rxstsqenq);
+ } else {
+ ret = ETH_STATUS_SUCCESS;
+ }
+
+ debug("-ep93xx_eth_rcv_packet %d", ret);
+ return ret;
+}
+
+/**
+ * Send a block of data via ethernet.
+ *
+ * TODO: Enhance this to deal with as much data as are available at one time? */
+static int ep93xx_eth_send_packet(struct eth_device *dev,
+ volatile void * const packet, int const length)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+ int ret = ETH_STATUS_FAILURE;
+
+ debug("+ep93xx_eth_send_packet");
+
+ /* Parameter check */
+ BUG_ON(packet == NULL);
+
+ /*
+ * Initialize the TX descriptor queue with the new packet's info.
+ * Clear the associated status queue entry. Enqueue the packet
+ * to the MAC for transmission.
+ */
+
+ /* set buffer address */
+ ep93xx_mac.tx_dq.current->word1 = (uint32_t)packet;
+
+ /* set buffer length and EOF bit */
+ ep93xx_mac.tx_dq.current->word2 = length | TX_DESC_EOF;
+
+ /* clear tx status */
+ ep93xx_mac.tx_sq.current->word1 = 0;
+
+ /* enqueue the TX descriptor */
+ writel(1, &mac->txdqenq);
+
+ /* wait for the frame to become processed */
+ while (!TX_STATUS_TXFP(ep93xx_mac.tx_sq.current))
+ ; /* noop */
+
+ if (!TX_STATUS_TXWE(ep93xx_mac.tx_sq.current)) {
+ ERROR("packet tx error, status %08X",
+ ep93xx_mac.tx_sq.current->word1);
+ dump_tx_descriptor_queue();
+ dump_tx_status_queue();
+
+ /* TODO: Add better error handling? */
+ goto eth_send_failed_0;
+ }
+
+ ret = ETH_STATUS_SUCCESS;
+ /* Fall through */
+
+eth_send_failed_0:
+ debug("-ep93xx_eth_send_packet %d", ret);
+ return ret;
+}
+#endif /* defined(CONFIG_DRIVER_EP93XX_MAC) */
+
+/* -----------------------------------------------------------------------------
+ * EP93xx ethernet MII functionality.
+ */
+#if defined(CONFIG_MII)
+
+/**
+ * Maximum MII address we support
+ */
+#define MII_ADDRESS_MAX (31)
+
+/**
+ * Maximum MII register address we support
+ */
+#define MII_REGISTER_MAX (31)
+
+/**
+ * Ethernet MII interface return values for public functions.
+ */
+enum mii_status {
+ MII_STATUS_SUCCESS = 0,
+ MII_STATUS_FAILURE = 1,
+};
+
+/**
+ * Read a 16-bit value from an MII register.
+ */
+static int ep93xx_miiphy_read(char * const dev, unsigned char const addr,
+ unsigned char const reg, unsigned short * const value)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+ int ret = MII_STATUS_FAILURE;
+ uint32_t self_ctl;
+
+ debug("+ep93xx_miiphy_read");
+
+ /* Parameter checks */
+ BUG_ON(dev == NULL);
+ BUG_ON(addr > MII_ADDRESS_MAX);
+ BUG_ON(reg > MII_REGISTER_MAX);
+ BUG_ON(value == NULL);
+
+ /*
+ * Save the current SelfCTL register value. Set MAC to suppress
+ * preamble bits. Wait for any previous MII command to complete
+ * before issuing the new command.
+ */
+ self_ctl = readl(&mac->selfctl);
+#if defined(CONFIG_MII_SUPPRESS_PREAMBLE)
+ writel(self_ctl & ~(1 << 8), &mac->selfctl);
+#endif /* defined(CONFIG_MII_SUPPRESS_PREAMBLE) */
+
+ while (readl(&mac->miists) & MIISTS_BUSY)
+ ; /* noop */
+
+ /*
+ * Issue the MII 'read' command. Wait for the command to complete.
+ * Read the MII data value.
+ */
+ writel(MIICMD_OPCODE_READ | ((uint32_t)addr << 5) | (uint32_t)reg,
+ &mac->miicmd);
+ while (readl(&mac->miists) & MIISTS_BUSY)
+ ; /* noop */
+
+ *value = (unsigned short)readl(&mac->miidata);
+
+ /* Restore the saved SelfCTL value and return. */
+ writel(self_ctl, &mac->selfctl);
+
+ ret = MII_STATUS_SUCCESS;
+ /* Fall through */
+
+ debug("-ep93xx_miiphy_read");
+ return ret;
+}
+
+/**
+ * Write a 16-bit value to an MII register.
+ */
+static int ep93xx_miiphy_write(char * const dev, unsigned char const addr,
+ unsigned char const reg, unsigned short const value)
+{
+ struct mac_regs *mac = (struct mac_regs *)MAC_BASE;
+ int ret = MII_STATUS_FAILURE;
+ uint32_t self_ctl;
+
+ debug("+ep93xx_miiphy_write");
+
+ /* Parameter checks */
+ BUG_ON(dev == NULL);
+ BUG_ON(addr > MII_ADDRESS_MAX);
+ BUG_ON(reg > MII_REGISTER_MAX);
+
+ /*
+ * Save the current SelfCTL register value. Set MAC to suppress
+ * preamble bits. Wait for any previous MII command to complete
+ * before issuing the new command.
+ */
+ self_ctl = readl(&mac->selfctl);
+#if defined(CONFIG_MII_SUPPRESS_PREAMBLE)
+ writel(self_ctl & ~(1 << 8), &mac->selfctl);
+#endif /* defined(CONFIG_MII_SUPPRESS_PREAMBLE) */
+
+ while (readl(&mac->miists) & MIISTS_BUSY)
+ ; /* noop */
+
+ /* Issue the MII 'write' command. Wait for the command to complete. */
+ writel((uint32_t)value, &mac->miidata);
+ writel(MIICMD_OPCODE_WRITE | ((uint32_t)addr << 5) | (uint32_t)reg,
+ &mac->miicmd);
+ while (readl(&mac->miists) & MIISTS_BUSY)
+ ; /* noop */
+
+ /* Restore the saved SelfCTL value and return. */
+ writel(self_ctl, &mac->selfctl);
+
+ ret = MII_STATUS_SUCCESS;
+ /* Fall through */
+
+ debug("-ep93xx_miiphy_write");
+ return ret;
+}
+#endif /* defined(CONFIG_MII) */
diff --git a/drivers/net/ep93xx.h b/drivers/net/ep93xx.h
new file mode 100644
index 0000000..79337f1
--- /dev/null
+++ b/drivers/net/ep93xx.h
@@ -0,0 +1,153 @@
+/*
+ * Copyright (C) 2009 Matthias Kaehlcke <matthias(a)kaehlcke.net>
+ *
+ * Copyright (C) 2004, 2005
+ * Cory T. Tusar, Videon Central, Inc., <ctusar(a)videon-central.com>
+ *
+ * See file CREDITS for list of people who contributed to this
+ * project.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ *
+ */
+
+#ifndef _ETH_H
+#define _ETH_H
+
+#include <net.h>
+
+/**
+ * #define this to dump device status and queue info during initialization and
+ * following errors.
+ */
+#undef EP93XX_MAC_DEBUG
+
+/**
+ * Number of descriptor and status entries in our RX queues.
+ * It must be power of 2 !
+ */
+#define NUMRXDESC PKTBUFSRX
+
+/**
+ * Number of descriptor and status entries in our TX queues.
+ */
+#define NUMTXDESC 1
+
+/**
+ * 944 = (1024 - 64) - 16, Fifo size - Minframesize - 16 (Chip FACT)
+ */
+#define TXSTARTMAX 944
+
+/**
+ * Ethernet MAC interface return values for public functions.
+ */
+enum eth_status {
+ ETH_STATUS_SUCCESS = 0,
+ ETH_STATUS_FAILURE = 1,
+};
+
+/**
+ * Receive descriptor queue entry
+ */
+struct rx_descriptor {
+ uint32_t word1;
+ uint32_t word2;
+} __attribute__((packed));
+
+/**
+ * Receive status queue entry
+ */
+struct rx_status {
+ uint32_t word1;
+ uint32_t word2;
+} __attribute__((packed));
+
+#define RX_STATUS_RWE(rx_status) ((rx_status->word1 >> 30) & 0x01)
+#define RX_STATUS_RFP(rx_status) ((rx_status->word1 >> 31) & 0x01)
+#define RX_STATUS_FRAME_LEN(rx_status) (rx_status->word2 & 0xFFFF)
+
+
+/**
+ * Transmit descriptor queue entry
+ */
+struct tx_descriptor {
+ uint32_t word1;
+ uint32_t word2;
+} __attribute__((packed));
+
+#define TX_DESC_EOF (1 << 31)
+
+/**
+ * Transmit status queue entry
+ */
+struct tx_status {
+ uint32_t word1;
+} __attribute__((packed));
+
+#define TX_STATUS_TXWE(tx_status) (((tx_status)->word1 >> 30) & 0x01)
+#define TX_STATUS_TXFP(tx_status) (((tx_status)->word1 >> 31) & 0x01)
+
+/**
+ * Transmit descriptor queue
+ */
+struct tx_descriptor_queue {
+ struct tx_descriptor *base;
+ struct tx_descriptor *current;
+ struct tx_descriptor *end;
+};
+
+/**
+ * Transmit status queue
+ */
+struct tx_status_queue {
+ struct tx_status *base;
+ volatile struct tx_status *current;
+ struct tx_status *end;
+};
+
+/**
+ * Receive descriptor queue
+ */
+struct rx_descriptor_queue {
+ struct rx_descriptor *base;
+ struct rx_descriptor *current;
+ struct rx_descriptor *end;
+};
+
+/**
+ * Receive status queue
+ */
+struct rx_status_queue {
+ struct rx_status *base;
+ volatile struct rx_status *current;
+ struct rx_status *end;
+};
+
+/**
+ * EP93xx MAC private data structure
+ */
+struct ep93xx_mac {
+ int is_initialized;
+
+ struct rx_descriptor_queue rx_dq;
+ struct rx_status_queue rx_sq;
+ void *rx_buffer[NUMRXDESC];
+
+ struct tx_descriptor_queue tx_dq;
+ struct tx_status_queue tx_sq;
+};
+
+#endif
diff --git a/include/common.h b/include/common.h
index 07897f6..6fecf39 100644
--- a/include/common.h
+++ b/include/common.h
@@ -123,6 +123,13 @@ typedef volatile unsigned char vu_char;
#define debugX(level,fmt,args...)
#endif /* DEBUG */
+#ifndef ERROR
+#define ERROR(fmt, args...) do { \
+ printf("ERROR: " fmt "\nat %s:%d/%s()\n", \
+ ##args, __FILE__, __LINE__, __func__); \
+} while (0)
+#endif
+
#ifndef BUG
#define BUG() do { \
printf("BUG: failure at %s:%d/%s()!\n", __FILE__, __LINE__, __FUNCTION__); \
@@ -500,7 +507,8 @@ ulong get_PCI_freq (void);
#endif
#if defined(CONFIG_S3C24X0) || \
defined(CONFIG_LH7A40X) || \
- defined(CONFIG_S3C6400)
+ defined(CONFIG_S3C6400) || \
+ defined(CONFIG_EP93XX)
ulong get_FCLK (void);
ulong get_HCLK (void);
ulong get_PCLK (void);
--
1.6.3.1
