summaryrefslogtreecommitdiff
path: root/drivers/net
diff options
context:
space:
mode:
authorJason Jin <Jason.jin@freescale.com>2012-10-12 16:59:47 +0800
committerJason Jin <Jason.jin@freescale.com>2012-10-12 16:59:47 +0800
commitebbd3debaaef3344e89d3a0b2fc1338c4be3ee86 (patch)
tree780cca7aeca872dea59814dcdf9b1c8396a87536 /drivers/net
parent0fb1497f28e7bc9df1b6c8267da7d58fa32e2a65 (diff)
ENGR00216078-2: switch: add L2 switch driver for VybridRel3
Port the L2 switch driver from CF platform for Vybridi plaform. This driver usually should work together with the RSTP demo applications provided by fsl. Signed-off-by: Alison Wang <b18965@freescale.com> Signed-off-by: Jason Jin <Jason.jin@freescale.com>
Diffstat (limited to 'drivers/net')
-rw-r--r--drivers/net/Kconfig9
-rw-r--r--drivers/net/Makefile1
-rw-r--r--drivers/net/fsl_l2_switch.c4266
3 files changed, 4276 insertions, 0 deletions
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 0f4ff7a68a05..de2905a67e76 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -1958,6 +1958,15 @@ config FEC1
Say Y here if you want to use the second built-in 10/100 Fast
ethernet controller on Vybrid and some ColdFire processors.
+config FSL_L2_SWITCH
+ bool "Ethernet switch controller(Freescale ColdFire and Vybrid platform)"
+ depends on !FEC && SOC_MVFA5
+ help
+ Say Y here if you want to use the built-in ethernet switch
+ controller on some ColdFire processors.
+ The Integrated Ethernet switch engine is compatible with
+ 10/100 MAC-NET core.
+
config FEC_1588
bool "Enable FEC 1588 timestamping"
depends on FEC
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index f7e26baa26a3..6d1f6624bada 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -124,6 +124,7 @@ obj-$(CONFIG_PCMCIA_PCNET) += 8390.o
obj-$(CONFIG_HP100) += hp100.o
obj-$(CONFIG_SMC9194) += smc9194.o
obj-$(CONFIG_FEC) += fec.o
+obj-$(CONFIG_FSL_L2_SWITCH) += fsl_l2_switch.o
obj-$(CONFIG_FEC_MPC52xx) += fec_mpc52xx.o
ifeq ($(CONFIG_FEC_MPC52xx_MDIO),y)
obj-$(CONFIG_FEC_MPC52xx) += fec_mpc52xx_phy.o
diff --git a/drivers/net/fsl_l2_switch.c b/drivers/net/fsl_l2_switch.c
new file mode 100644
index 000000000000..0a6b9f56dd96
--- /dev/null
+++ b/drivers/net/fsl_l2_switch.c
@@ -0,0 +1,4266 @@
+/*
+ * L2 switch Controller (Ethernet switch) driver
+ * for Freescale M5441x and Vybrid.
+ *
+ * Copyright 2010-2012 Freescale Semiconductor, Inc.
+ * Alison Wang (b18965@freescale.com)
+ * Jason Jin (Jason.jin@freescale.com)
+ *
+ * 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/kernel.h>
+#include <linux/string.h>
+#include <linux/ptrace.h>
+#include <linux/errno.h>
+#include <linux/ioport.h>
+#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/pci.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+#include <linux/bitops.h>
+#include <linux/platform_device.h>
+#include <linux/fsl_devices.h>
+#include <linux/phy.h>
+#include <linux/kthread.h>
+#include <linux/syscalls.h>
+#include <linux/uaccess.h>
+#include <linux/io.h>
+#include <linux/signal.h>
+#include <linux/clk.h>
+
+#include <asm/irq.h>
+#include <asm/pgtable.h>
+#include <asm/cacheflush.h>
+#include <mach/hardware.h>
+#include <mach/fsl_l2_switch.h>
+
+#define SWITCH_MAX_PORTS 1
+#define CONFIG_FEC_SHARED_PHY
+
+static unsigned char macaddr[ETH_ALEN];
+module_param_array(macaddr, byte, NULL, 0);
+MODULE_PARM_DESC(macaddr, "FEC Ethernet MAC address");
+
+/* Interrupt events/masks */
+#define FEC_ENET_HBERR ((uint)0x80000000) /* Heartbeat error */
+#define FEC_ENET_BABR ((uint)0x40000000) /* Babbling receiver */
+#define FEC_ENET_BABT ((uint)0x20000000) /* Babbling transmitter */
+#define FEC_ENET_GRA ((uint)0x10000000) /* Graceful stop complete */
+#define FEC_ENET_TXF ((uint)0x08000000) /* Full frame transmitted */
+#define FEC_ENET_TXB ((uint)0x04000000) /* A buffer was transmitted */
+#define FEC_ENET_RXF ((uint)0x02000000) /* Full frame received */
+#define FEC_ENET_RXB ((uint)0x01000000) /* A buffer was received */
+#define FEC_ENET_MII ((uint)0x00800000) /* MII interrupt */
+#define FEC_ENET_EBERR ((uint)0x00400000) /* SDMA bus error */
+
+static int switch_enet_open(struct net_device *dev);
+static int switch_enet_start_xmit(struct sk_buff *skb, struct net_device *dev);
+static irqreturn_t switch_enet_interrupt(int irq, void *dev_id);
+static void switch_enet_tx(struct net_device *dev);
+static void switch_enet_rx(struct net_device *dev);
+static int switch_enet_close(struct net_device *dev);
+static void set_multicast_list(struct net_device *dev);
+static void switch_restart(struct net_device *dev, int duplex);
+static void switch_stop(struct net_device *dev);
+static void switch_get_mac_address(struct net_device *dev);
+
+#define NMII 20
+
+/* Make MII read/write commands for the FEC */
+#define mk_mii_read(REG) (0x60020000 | ((REG & 0x1f) << 18))
+#define mk_mii_write(REG, VAL) (0x50020000 | ((REG & 0x1f) << 18) | \
+ (VAL & 0xffff))
+/* MII MMFR bits definition */
+#define ESW_MMFR_ST (1 << 30)
+#define ESW_MMFR_OP_READ (2 << 28)
+#define ESW_MMFR_OP_WRITE (1 << 28)
+#define ESW_MMFR_PA(v) ((v & 0x1f) << 23)
+#define ESW_MMFR_RA(v) ((v & 0x1f) << 18)
+#define ESW_MMFR_TA (2 << 16)
+#define ESW_MMFR_DATA(v) (v & 0xffff)
+
+#define ESW_MII_TIMEOUT 30 /* ms */
+
+/* Transmitter timeout.*/
+#define TX_TIMEOUT (2*HZ)
+
+/*last read entry from learning interface*/
+eswPortInfo g_info;
+/* switch ports status */
+struct port_status ports_link_status;
+
+/* the user space pid, used to send the link change to user space */
+long user_pid = 1;
+
+/* ----------------------------------------------------------------*/
+/*
+ * Calculate Galois Field Arithmetic CRC for Polynom x^8+x^2+x+1.
+ * It omits the final shift in of 8 zeroes a "normal" CRC would do
+ * (getting the remainder).
+ *
+ * Examples (hexadecimal values):<br>
+ * 10-11-12-13-14-15 => CRC=0xc2
+ * 10-11-cc-dd-ee-00 => CRC=0xe6
+ *
+ * param: pmacaddress
+ * A 6-byte array with the MAC address.
+ * The first byte is the first byte transmitted
+ * return The 8-bit CRC in bits 7:0
+ */
+int crc8_calc(unsigned char *pmacaddress)
+{
+ /* byte index */
+ int byt;
+ /* bit index */
+ int bit;
+ int inval;
+ int crc;
+ /* preset */
+ crc = 0x12;
+ for (byt = 0; byt < 6; byt++) {
+ inval = (((int)pmacaddress[byt]) & 0xff);
+ /*
+ * shift bit 0 to bit 8 so all our bits
+ * travel through bit 8
+ * (simplifies below calc)
+ */
+ inval <<= 8;
+
+ for (bit = 0; bit < 8; bit++) {
+ /* next input bit comes into d7 after shift */
+ crc |= inval & 0x100;
+ if (crc & 0x01)
+ /* before shift */
+ crc ^= 0x1c0;
+
+ crc >>= 1;
+ inval >>= 1;
+ }
+
+ }
+ /* upper bits are clean as we shifted in zeroes! */
+ return crc;
+}
+
+void read_atable(struct switch_enet_private *fep,
+ int index, unsigned long *read_lo, unsigned long *read_hi)
+{
+ unsigned long atable_base = (unsigned long)fep->macbase;
+
+ *read_lo = readl(atable_base + (index << 3));
+ *read_hi = readl(atable_base + (index << 3) + 4);
+}
+
+void write_atable(struct switch_enet_private *fep,
+ int index, unsigned long write_lo, unsigned long write_hi)
+{
+ unsigned long atable_base = (unsigned long)fep->macbase;
+
+ writel(write_lo, atable_base + (index << 3));
+ writel(write_hi, atable_base + (index<<3) + 4);
+}
+
+/* Check if the Port Info FIFO has data available
+ * for reading. 1 valid, 0 invalid*/
+int esw_portinfofifo_status(struct switch_enet_private *fep)
+{
+ return readl(fep->membase + FEC_ESW_LSR);
+}
+
+/* Initialize the Port Info FIFO. */
+void esw_portinfofifo_initialize(struct switch_enet_private *fep)
+{
+ unsigned long tmp;
+
+ /* disable all learn */
+ tmp = readl(fep->membase + FEC_ESW_IMR);
+ tmp &= (~FSL_ESW_IMR_LRN);
+ writel(tmp, fep->membase + FEC_ESW_IMR);
+
+ /* remove all entries from FIFO */
+ while (esw_portinfofifo_status(fep)) {
+ /* read one data word */
+ tmp = readl(fep->membase + FEC_ESW_LREC0);
+ tmp = readl(fep->membase + FEC_ESW_LREC1);
+ }
+
+}
+
+/* Read one element from the HW receive FIFO (Queue)
+ * if available and return it.
+ * return ms_HwPortInfo or null if no data is available
+ */
+eswPortInfo *esw_portinfofifo_read(struct switch_enet_private *fep)
+{
+ unsigned long tmp;
+
+ /* check learning record valid */
+ if (!(readl(fep->membase + FEC_ESW_LSR)))
+ return NULL;
+
+ /*read word from FIFO*/
+ g_info.maclo = readl(fep->membase + FEC_ESW_LREC0);
+
+ /*but verify that we actually did so
+ * (0=no data available)*/
+ if (g_info.maclo == 0)
+ return NULL;
+
+ /* read 2nd word from FIFO */
+ tmp = readl(fep->membase + FEC_ESW_LREC1);
+ g_info.machi = tmp & 0xffff;
+ g_info.hash = (tmp >> 16) & 0xff;
+ g_info.port = (tmp >> 24) & 0xf;
+
+ return &g_info;
+}
+
+/* Clear complete MAC Look Up Table */
+void esw_clear_atable(struct switch_enet_private *fep)
+{
+ int index;
+ for (index = 0; index < 2048; index++)
+ write_atable(fep, index, 0, 0);
+}
+
+void esw_dump_atable(struct switch_enet_private *fep)
+{
+ int index;
+ unsigned long read_lo, read_hi;
+ for (index = 0; index < 2048; index++)
+ read_atable(fep, index, &read_lo, &read_hi);
+}
+
+/*
+ * pdates MAC address lookup table with a static entry
+ * Searches if the MAC address is already there in the block and replaces
+ * the older entry with new one. If MAC address is not there then puts a
+ * new entry in the first empty slot available in the block
+ *
+ * mac_addr Pointer to the array containing MAC address to
+ * be put as static entry
+ * port Port bitmask numbers to be added in static entry,
+ * valid values are 1-7
+ * priority Priority for the static entry in table
+ *
+ * return 0 for a successful update else -1 when no slot available
+ */
+int esw_update_atable_static(unsigned char *mac_addr,
+ unsigned int port, unsigned int priority,
+ struct switch_enet_private *fep)
+{
+ unsigned long block_index, entry, index_end;
+ unsigned long read_lo, read_hi;
+ unsigned long write_lo, write_hi;
+
+ write_lo = (unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) |
+ (mac_addr[1] << 8) |
+ mac_addr[0]);
+ write_hi = (unsigned long)(0 |
+ (port << AT_SENTRY_PORTMASK_shift) |
+ (priority << AT_SENTRY_PRIO_shift) |
+ (AT_ENTRY_TYPE_STATIC << AT_ENTRY_TYPE_shift) |
+ (AT_ENTRY_RECORD_VALID << AT_ENTRY_VALID_shift) |
+ (mac_addr[5] << 8) | (mac_addr[4]));
+
+ block_index = GET_BLOCK_PTR(crc8_calc(mac_addr));
+ index_end = block_index + ATABLE_ENTRY_PER_SLOT;
+ /* Now search all the entries in the selected block */
+ for (entry = block_index; entry < index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ /*
+ * MAC address matched, so update the
+ * existing entry
+ * even if its a dynamic one
+ */
+ if ((read_lo == write_lo) && ((read_hi & 0x0000ffff) ==
+ (write_hi & 0x0000ffff))) {
+ write_atable(fep, entry, write_lo, write_hi);
+ return 0;
+ } else if (!(read_hi & (1 << 16))) {
+ /*
+ * Fill this empty slot (valid bit zero),
+ * assuming no holes in the block
+ */
+ write_atable(fep, entry, write_lo, write_hi);
+ fep->atCurrEntries++;
+ return 0;
+ }
+ }
+
+ /* No space available for this static entry */
+ return -1;
+}
+
+/* lookup entry in given Address Table slot and
+ * insert (learn) it if it is not found.
+ * return 0 if entry was found and updated.
+ * 1 if entry was not found and has been inserted (learned).
+ */
+int esw_update_atable_dynamic(unsigned char *mac_addr, unsigned int port,
+ unsigned int currTime, struct switch_enet_private *fep)
+{
+ unsigned long block_index, entry, index_end;
+ unsigned long read_lo, read_hi;
+ unsigned long write_lo, write_hi;
+ unsigned long tmp;
+ int time, timeold, indexold;
+
+ /* prepare update port and timestamp */
+ write_hi = (mac_addr[5] << 8) | (mac_addr[4]);
+ write_lo = (unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) |
+ (mac_addr[1] << 8) |
+ mac_addr[0]);
+ tmp = AT_ENTRY_RECORD_VALID << AT_ENTRY_VALID_shift;
+ tmp |= AT_ENTRY_TYPE_DYNAMIC << AT_ENTRY_TYPE_shift;
+ tmp |= currTime << AT_DENTRY_TIME_shift;
+ tmp |= port << AT_DENTRY_PORT_shift;
+ tmp |= write_hi;
+
+ /*
+ * linear search through all slot
+ * entries and update if found
+ */
+ block_index = GET_BLOCK_PTR(crc8_calc(mac_addr));
+ index_end = block_index + ATABLE_ENTRY_PER_SLOT;
+ /* Now search all the entries in the selected block */
+ for (entry = block_index; entry < index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+
+ if ((read_lo == write_lo) &&
+ ((read_hi & 0x0000ffff) ==
+ (write_hi & 0x0000ffff))) {
+ /* found correct address,
+ * update timestamp. */
+ write_atable(fep, entry, write_lo, tmp);
+ return 0;
+ } else if (!(read_hi & (1 << 16))) {
+ /* slot is empty, then use it
+ * for new entry
+ * Note: There are no holes,
+ * therefore cannot be any
+ * more that need to be compared.
+ */
+ write_atable(fep, entry, write_lo, tmp);
+ /* statistics (we do it between writing
+ * .hi an .lo due to
+ * hardware limitation...
+ */
+ fep->atCurrEntries++;
+ /* newly inserted */
+ return 1;
+ }
+ }
+
+ /*
+ * no more entry available in blockk ...
+ * overwrite oldest
+ */
+ timeold = 0;
+ indexold = 0;
+ for (entry = block_index; entry < index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ time = AT_EXTRACT_TIMESTAMP(read_hi);
+ time = TIMEDELTA(currTime, time);
+ if (time > timeold) {
+ /* is it older ?*/
+ timeold = time;
+ indexold = entry;
+ }
+ }
+
+ write_atable(fep, indexold, write_lo, tmp);
+ /* Statistics (do it inbetween
+ * writing to .lo and .hi*/
+ fep->atBlockOverflows++;
+ /* newly inserted */
+ return 1;
+}
+
+int esw_update_atable_dynamic1(unsigned long write_lo, unsigned long write_hi,
+ int block_index, unsigned int port, unsigned int currTime,
+ struct switch_enet_private *fep)
+{
+ unsigned long entry, index_end;
+ unsigned long read_lo, read_hi;
+ unsigned long tmp;
+ int time, timeold, indexold;
+
+ /* prepare update port and timestamp */
+ tmp = AT_ENTRY_RECORD_VALID << AT_ENTRY_VALID_shift;
+ tmp |= AT_ENTRY_TYPE_DYNAMIC << AT_ENTRY_TYPE_shift;
+ tmp |= currTime << AT_DENTRY_TIME_shift;
+ tmp |= port << AT_DENTRY_PORT_shift;
+ tmp |= write_hi;
+
+ /*
+ * linear search through all slot
+ * entries and update if found
+ */
+ index_end = block_index + ATABLE_ENTRY_PER_SLOT;
+ /* Now search all the entries in the selected block */
+ for (entry = block_index; entry < index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ if ((read_lo == write_lo) &&
+ ((read_hi & 0x0000ffff) ==
+ (write_hi & 0x0000ffff))) {
+ /* found correct address,
+ * update timestamp. */
+ write_atable(fep, entry, write_lo, tmp);
+ return 0;
+ } else if (!(read_hi & (1 << 16))) {
+ /* slot is empty, then use it
+ * for new entry
+ * Note: There are no holes,
+ * therefore cannot be any
+ * more that need to be compared.
+ */
+ write_atable(fep, entry, write_lo, tmp);
+ /* statistics (we do it between writing
+ * .hi an .lo due to
+ * hardware limitation...
+ */
+ fep->atCurrEntries++;
+ /* newly inserted */
+ return 1;
+ }
+ }
+
+ /*
+ * no more entry available in block ...
+ * overwrite oldest
+ */
+ timeold = 0;
+ indexold = 0;
+ for (entry = block_index; entry < index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ time = AT_EXTRACT_TIMESTAMP(read_hi);
+ time = TIMEDELTA(currTime, time);
+ if (time > timeold) {
+ /* is it older ?*/
+ timeold = time;
+ indexold = entry;
+ }
+ }
+
+ write_atable(fep, indexold, write_lo, tmp);
+ /* Statistics (do it inbetween
+ * writing to .lo and .hi*/
+ fep->atBlockOverflows++;
+ /* newly inserted */
+ return 1;
+}
+
+/*
+ * Delete one dynamic entry within the given block
+ * of 64-bit entries.
+ * return number of valid entries in the block after deletion.
+ */
+int esw_del_atable_dynamic(struct switch_enet_private *fep,
+ int blockidx, int entryidx)
+{
+ unsigned long index_start, index_end;
+ int i;
+ unsigned long read_lo, read_hi;
+
+ /* the entry to delete */
+ index_start = blockidx + entryidx;
+ /* one after last */
+ index_end = blockidx + ATABLE_ENTRY_PER_SLOT;
+ /* Statistics */
+ fep->atCurrEntries--;
+
+ if (entryidx == (ATABLE_ENTRY_PER_SLOT - 1)) {
+ /* if it is the very last entry,
+ * just delete it without further efford*/
+ write_atable(fep, index_start, 0, 0);
+ /*number of entries left*/
+ i = ATABLE_ENTRY_PER_SLOT - 1;
+ return i;
+ } else {
+ /*not the last in the block, then
+ * shift all that follow the one
+ * that is deleted to avoid "holes".
+ */
+ for (i = index_start; i < (index_end - 1); i++) {
+ read_atable(fep, i + 1, &read_lo, &read_hi);
+ /* move it down */
+ write_atable(fep, i, read_lo, read_hi);
+ if (!(read_hi & (1 << 16))) {
+ /* stop if we just copied the last */
+ return i - blockidx;
+ }
+ }
+
+ /*moved all entries up to the last.
+ * then set invalid flag in the last*/
+ write_atable(fep, index_end - 1, 0, 0);
+ /* number of valid entries left */
+ return i - blockidx;
+ }
+}
+
+void esw_atable_dynamicms_del_entries_for_port(
+ struct switch_enet_private *fep, int port_index)
+{
+ unsigned long read_lo, read_hi;
+ unsigned int port_idx;
+ int i;
+
+ for (i = 0; i < ESW_ATABLE_MEM_NUM_ENTRIES; i++) {
+ read_atable(fep, i, &read_lo, &read_hi);
+ if (read_hi & (1 << 16)) {
+ port_idx = AT_EXTRACT_PORT(read_hi);
+
+ if (port_idx == port_index)
+ write_atable(fep, i, 0, 0);
+ }
+ }
+}
+
+void esw_atable_dynamicms_del_entries_for_other_port(
+ struct switch_enet_private *fep,
+ int port_index)
+{
+ unsigned long read_lo, read_hi;
+ unsigned int port_idx;
+ int i;
+
+ for (i = 0; i < ESW_ATABLE_MEM_NUM_ENTRIES; i++) {
+ read_atable(fep, i, &read_lo, &read_hi);
+ if (read_hi & (1 << 16)) {
+ port_idx = AT_EXTRACT_PORT(read_hi);
+
+ if (port_idx != port_index)
+ write_atable(fep, i, 0, 0);
+ }
+ }
+}
+
+/*
+ * Scan one complete block (Slot) for outdated entries and delete them.
+ * blockidx index of block of entries that should be analyzed.
+ * return number of deleted entries, 0 if nothing was modified.
+ */
+int esw_atable_dynamicms_check_block_age(
+ struct switch_enet_private *fep, int blockidx) {
+
+ int i, tm, tdelta;
+ int deleted = 0, entries = 0;
+ unsigned long read_lo, read_hi;
+ /* Scan all entries from last down to
+ * have faster deletion speed if necessary*/
+ for (i = (blockidx + ATABLE_ENTRY_PER_SLOT - 1);
+ i >= blockidx; i--) {
+ read_atable(fep, i, &read_lo, &read_hi);
+
+ if (read_hi & (1 << 16)) {
+ /* the entry is valide*/
+ tm = AT_EXTRACT_TIMESTAMP(read_hi);
+ tdelta = TIMEDELTA(fep->currTime, tm);
+ if (tdelta > fep->ageMax) {
+ esw_del_atable_dynamic(fep,
+ blockidx, i-blockidx);
+ deleted++;
+ } else {
+ /* statistics */
+ entries++;
+ }
+ }
+ }
+
+ /*update statistics*/
+ if (fep->atMaxEntriesPerBlock < entries)
+ fep->atMaxEntriesPerBlock = entries;
+
+ return deleted;
+}
+
+/* scan the complete address table and find the most current entry.
+ * The time of the most current entry then is used as current time
+ * for the context structure.
+ * In addition the atCurrEntries value is updated as well.
+ * return time that has been set in the context.
+ */
+int esw_atable_dynamicms_find_set_latesttime(
+ struct switch_enet_private *fep) {
+
+ int tm_min, tm_max, tm;
+ int delta, cur, i;
+ unsigned long read_lo, read_hi;
+
+ tm_min = (1 << AT_DENTRY_TIMESTAMP_WIDTH) - 1;
+ tm_max = 0;
+ cur = 0;
+
+ for (i = 0; i < ESW_ATABLE_MEM_NUM_ENTRIES; i++) {
+ read_atable(fep, i, &read_lo, &read_hi);
+ if (read_hi & (1 << 16)) {
+ /*the entry is valid*/
+ tm = AT_EXTRACT_TIMESTAMP(read_hi);
+ if (tm > tm_max)
+ tm_max = tm;
+ if (tm < tm_min)
+ tm_min = tm;
+ cur++;
+ }
+ }
+
+ delta = TIMEDELTA(tm_max, tm_min);
+ if (delta < fep->ageMax) {
+ /*Difference must be in range*/
+ fep->currTime = tm_max;
+ } else {
+ fep->currTime = tm_min;
+ }
+
+ fep->atCurrEntries = cur;
+ return fep->currTime;
+}
+
+int esw_atable_dynamicms_get_port(
+ struct switch_enet_private *fep,
+ unsigned long write_lo,
+ unsigned long write_hi,
+ int block_index)
+{
+ int i, index_end;
+ unsigned long read_lo, read_hi, port;
+
+ index_end = block_index + ATABLE_ENTRY_PER_SLOT;
+ /* Now search all the entries in the selected block */
+ for (i = block_index; i < index_end; i++) {
+ read_atable(fep, i, &read_lo, &read_hi);
+
+ if ((read_lo == write_lo) &&
+ ((read_hi & 0x0000ffff) ==
+ (write_hi & 0x0000ffff))) {
+ /* found correct address,*/
+ if (read_hi & (1 << 16)) {
+ /*extract the port index from the valid entry*/
+ port = AT_EXTRACT_PORT(read_hi);
+ return port;
+ }
+ }
+ }
+
+ return -1;
+}
+
+/* Get the port index from the source MAC address
+ * of the received frame
+ * @return port index
+ */
+int esw_atable_dynamicms_get_portindex_from_mac(
+ struct switch_enet_private *fep,
+ unsigned char *mac_addr,
+ unsigned long write_lo,
+ unsigned long write_hi)
+{
+ int blockIdx;
+ int rc;
+ /*compute the block index*/
+ blockIdx = GET_BLOCK_PTR(crc8_calc(mac_addr));
+ /* Get the ingress port index of the received BPDU */
+ rc = esw_atable_dynamicms_get_port(fep,
+ write_lo, write_hi, blockIdx);
+
+ return rc;
+}
+
+/* dynamicms MAC address table learn and migration*/
+int esw_atable_dynamicms_learn_migration(
+ struct switch_enet_private *fep,
+ int currTime)
+{
+ eswPortInfo *pESWPortInfo;
+ int index;
+ int inserted = 0;
+
+ pESWPortInfo = esw_portinfofifo_read(fep);
+ /* Anything to learn */
+ if (pESWPortInfo != 0) {
+ /*get block index from lookup table*/
+ index = GET_BLOCK_PTR(pESWPortInfo->hash);
+ inserted = esw_update_atable_dynamic1(
+ pESWPortInfo->maclo,
+ pESWPortInfo->machi, index,
+ pESWPortInfo->port, currTime, fep);
+ }
+
+ return 0;
+}
+/* -----------------------------------------------------------------*/
+/*
+ * esw_forced_forward
+ * The frame is forwared to the forced destination ports.
+ * It only replace the MAC lookup function,
+ * all other filtering(eg.VLAN verification) act as normal
+ */
+int esw_forced_forward(struct switch_enet_private *fep,
+ int port1, int port2, int enable)
+{
+ unsigned long tmp = 0;
+
+ /* Enable Forced forwarding for port num */
+ if ((port1 == 1) && (port2 == 1))
+ tmp |= FSL_ESW_P0FFEN_FD(3);
+ else if (port1 == 1)
+ /*Enable Forced forwarding for port 1 only*/
+ tmp |= FSL_ESW_P0FFEN_FD(1);
+ else if (port2 == 1)
+ /*Enable Forced forwarding for port 2 only*/
+ tmp |= FSL_ESW_P0FFEN_FD(2);
+ else {
+ printk(KERN_ERR "%s:do not support "
+ "the forced forward mode"
+ "port1 %x port2 %x\n",
+ __func__, port1, port2);
+ return -1;
+ }
+
+ if (enable == 1)
+ tmp |= FSL_ESW_P0FFEN_FEN;
+ else if (enable == 0)
+ tmp &= ~FSL_ESW_P0FFEN_FEN;
+ else {
+ printk(KERN_ERR "%s: the enable %x is error\n",
+ __func__, enable);
+ return -2;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_P0FFEN);
+ return 0;
+}
+
+void esw_get_forced_forward(
+ struct switch_enet_private *fep,
+ unsigned long *ulForceForward)
+{
+ *ulForceForward = readl(fep->membase + FEC_ESW_P0FFEN);
+}
+
+void esw_get_port_enable(
+ struct switch_enet_private *fep,
+ unsigned long *ulPortEnable)
+{
+ *ulPortEnable = readl(fep->membase + FEC_ESW_PER);
+}
+
+/*
+ * enable or disable port n tx or rx
+ * tx_en 0 disable port n tx
+ * tx_en 1 enable port n tx
+ * rx_en 0 disbale port n rx
+ * rx_en 1 enable port n rx
+ */
+int esw_port_enable_config(struct switch_enet_private *fep,
+ int port, int tx_en, int rx_en)
+{
+ unsigned long tmp = 0;
+
+ tmp = readl(fep->membase + FEC_ESW_PER);
+ if (tx_en == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_PER_TE0;
+ else if (port == 1)
+ tmp |= FSL_ESW_PER_TE1;
+ else if (port == 2)
+ tmp |= FSL_ESW_PER_TE2;
+ else {
+ printk(KERN_ERR "%s:do not support the"
+ " port %x tx enable\n",
+ __func__, port);
+ return -1;
+ }
+ } else if (tx_en == 0) {
+ if (port == 0)
+ tmp &= (~FSL_ESW_PER_TE0);
+ else if (port == 1)
+ tmp &= (~FSL_ESW_PER_TE1);
+ else if (port == 2)
+ tmp &= (~FSL_ESW_PER_TE2);
+ else {
+ printk(KERN_ERR "%s:do not support "
+ "the port %x tx disable\n",
+ __func__, port);
+ return -2;
+ }
+ } else {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " tx op value %x\n",
+ __func__, port, tx_en);
+ return -3;
+ }
+
+ if (rx_en == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_PER_RE0;
+ else if (port == 1)
+ tmp |= FSL_ESW_PER_RE1;
+ else if (port == 2)
+ tmp |= FSL_ESW_PER_RE2;
+ else {
+ printk(KERN_ERR "%s:do not support the "
+ "port %x rx enable\n",
+ __func__, port);
+ return -4;
+ }
+ } else if (rx_en == 0) {
+ if (port == 0)
+ tmp &= (~FSL_ESW_PER_RE0);
+ else if (port == 1)
+ tmp &= (~FSL_ESW_PER_RE1);
+ else if (port == 2)
+ tmp &= (~FSL_ESW_PER_RE2);
+ else {
+ printk(KERN_ERR "%s:do not support the "
+ "port %x rx disable\n",
+ __func__, port);
+ return -5;
+ }
+ } else {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " rx op value %x\n",
+ __func__, port, tx_en);
+ return -6;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_PER);
+ return 0;
+}
+
+
+void esw_get_port_broadcast(struct switch_enet_private *fep,
+ unsigned long *ulPortBroadcast)
+{
+ *ulPortBroadcast = readl(fep->membase + FEC_ESW_DBCR);
+}
+
+int esw_port_broadcast_config(struct switch_enet_private *fep,
+ int port, int enable)
+{
+ unsigned long tmp = 0;
+
+ if ((port > 2) || (port < 0)) {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " default broadcast\n",
+ __func__, port);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_DBCR);
+ if (enable == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_DBCR_P0;
+ else if (port == 1)
+ tmp |= FSL_ESW_DBCR_P1;
+ else if (port == 2)
+ tmp |= FSL_ESW_DBCR_P2;
+ } else if (enable == 0) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_DBCR_P0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_DBCR_P1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_DBCR_P2;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_DBCR);
+ return 0;
+}
+
+
+void esw_get_port_multicast(struct switch_enet_private *fep,
+ unsigned long *ulPortMulticast)
+{
+ *ulPortMulticast = readl(fep->membase + FEC_ESW_DMCR);
+}
+
+int esw_port_multicast_config(struct switch_enet_private *fep,
+ int port, int enable)
+{
+ unsigned long tmp = 0;
+
+ if ((port > 2) || (port < 0)) {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " default broadcast\n",
+ __func__, port);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_DMCR);
+ if (enable == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_DMCR_P0;
+ else if (port == 1)
+ tmp |= FSL_ESW_DMCR_P1;
+ else if (port == 2)
+ tmp |= FSL_ESW_DMCR_P2;
+ } else if (enable == 0) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_DMCR_P0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_DMCR_P1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_DMCR_P2;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_DMCR);
+ return 0;
+}
+
+
+void esw_get_port_blocking(struct switch_enet_private *fep,
+ unsigned long *ulPortBlocking)
+{
+ *ulPortBlocking = readl(fep->membase + FEC_ESW_BKLR) & 0x0000000f;
+}
+
+int esw_port_blocking_config(struct switch_enet_private *fep,
+ int port, int enable)
+{
+ unsigned long tmp = 0;
+
+ if ((port > 2) || (port < 0)) {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " default broadcast\n",
+ __func__, port);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_BKLR);
+ if (enable == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_BKLR_BE0;
+ else if (port == 1)
+ tmp |= FSL_ESW_BKLR_BE1;
+ else if (port == 2)
+ tmp |= FSL_ESW_BKLR_BE2;
+ } else if (enable == 0) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_BKLR_BE0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_BKLR_BE1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_BKLR_BE2;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_BKLR);
+ return 0;
+}
+
+
+void esw_get_port_learning(struct switch_enet_private *fep,
+ unsigned long *ulPortLearning)
+{
+ *ulPortLearning =
+ (readl(fep->membase + FEC_ESW_BKLR) & 0x000f0000) >> 16;
+}
+
+int esw_port_learning_config(struct switch_enet_private *fep,
+ int port, int disable)
+{
+ unsigned long tmp = 0;
+
+ if ((port > 2) || (port < 0)) {
+ printk(KERN_ERR "%s:do not support the port %x"
+ " default broadcast\n",
+ __func__, port);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_BKLR);
+ if (disable == 0) {
+ fep->learning_irqhandle_enable = 0;
+ if (port == 0)
+ tmp |= FSL_ESW_BKLR_LD0;
+ else if (port == 1)
+ tmp |= FSL_ESW_BKLR_LD1;
+ else if (port == 2)
+ tmp |= FSL_ESW_BKLR_LD2;
+ } else if (disable == 1) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_BKLR_LD0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_BKLR_LD1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_BKLR_LD2;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_BKLR);
+ return 0;
+}
+
+/*********************************************************************/
+void esw_mac_lookup_table_range(struct switch_enet_private *fep)
+{
+ int index;
+ unsigned long read_lo, read_hi;
+ /* Pointer to switch address look up memory*/
+ for (index = 0; index < 2048; index++)
+ write_atable(fep, index, index, (~index));
+
+ /* Pointer to switch address look up memory*/
+ for (index = 0; index < 2048; index++) {
+ read_atable(fep, index, &read_lo, &read_hi);
+ if (read_lo != index) {
+ printk(KERN_ERR "%s:Mismatch at low %d\n",
+ __func__, index);
+ return;
+ }
+
+ if (read_hi != (~index)) {
+ printk(KERN_ERR "%s:Mismatch at high %d\n",
+ __func__, index);
+ return;
+ }
+ }
+}
+
+/*
+ * Checks IP Snoop options of handling the snooped frame.
+ * mode 0 : The snooped frame is forward only to management port
+ * mode 1 : The snooped frame is copy to management port and
+ * normal forwarding is checked.
+ * mode 2 : The snooped frame is discarded.
+ * mode 3 : Disable the ip snoop function
+ * ip_header_protocol : the IP header protocol field
+ */
+int esw_ip_snoop_config(struct switch_enet_private *fep,
+ int mode, unsigned long ip_header_protocol)
+{
+ unsigned long tmp = 0, protocol_type = 0;
+ int num = 0;
+
+ /* Config IP Snooping */
+ if (mode == 0) {
+ /* Enable IP Snooping */
+ tmp = FSL_ESW_IPSNP_EN;
+ tmp |= FSL_ESW_IPSNP_MODE(0);/*For Forward*/
+ } else if (mode == 1) {
+ /* Enable IP Snooping */
+ tmp = FSL_ESW_IPSNP_EN;
+ /*For Forward and copy_to_mangmnt_port*/
+ tmp |= FSL_ESW_IPSNP_MODE(1);
+ } else if (mode == 2) {
+ /* Enable IP Snooping */
+ tmp = FSL_ESW_IPSNP_EN;
+ tmp |= FSL_ESW_IPSNP_MODE(2);/*discard*/
+ } else if (mode == 3) {
+ /* disable IP Snooping */
+ tmp = FSL_ESW_IPSNP_EN;
+ tmp &= ~FSL_ESW_IPSNP_EN;
+ } else {
+ printk(KERN_ERR "%s: the mode %x "
+ "we do not support\n", __func__, mode);
+ return -1;
+ }
+
+ protocol_type = ip_header_protocol;
+ for (num = 0; num < 8; num++) {
+ unsigned long reg = readl(fep->membase + FEC_ESW_IPSNP(num));
+ if (protocol_type ==
+ AT_EXTRACT_IP_PROTOCOL(reg)) {
+ writel(tmp | FSL_ESW_IPSNP_PROTOCOL(protocol_type),
+ fep->membase + FEC_ESW_IPSNP(num));
+ break;
+ } else if (!reg) {
+ writel(tmp | FSL_ESW_IPSNP_PROTOCOL(protocol_type),
+ fep->membase + FEC_ESW_IPSNP(num));
+ break;
+ }
+ }
+ if (num == 8) {
+ printk(KERN_INFO "IP snooping table is full\n");
+ return 0;
+ }
+
+ return 0;
+}
+
+void esw_get_ip_snoop_config(struct switch_enet_private *fep,
+ unsigned long *ulpESW_IPSNP)
+{
+ int i;
+
+ for (i = 0; i < 8; i++)
+ *(ulpESW_IPSNP + i) = readl(fep->membase + FEC_ESW_IPSNP(i));
+}
+/*
+ * Checks TCP/UDP Port Snoop options of handling the snooped frame.
+ * mode 0 : The snooped frame is forward only to management port
+ * mode 1 : The snooped frame is copy to management port and
+ * normal forwarding is checked.
+ * mode 2 : The snooped frame is discarded.
+ * mode 3 : Disable the TCP/UDP port snoop function
+ * compare_port : port number in the TCP/UDP header
+ * compare_num 1: TCP/UDP source port number is compared
+ * compare_num 2: TCP/UDP destination port number is compared
+ * compare_num 3: TCP/UDP source and destination port number is compared
+ */
+int esw_tcpudp_port_snoop_config(struct switch_enet_private *fep,
+ int mode, int compare_port, int compare_num)
+{
+ unsigned long tmp;
+ int num;
+
+ /* Enable TCP/UDP port Snooping */
+ tmp = FSL_ESW_PSNP_EN;
+ if (mode == 0)
+ tmp |= FSL_ESW_PSNP_MODE(0);/*For Forward*/
+ else if (mode == 1)/*For Forward and copy_to_mangmnt_port*/
+ tmp |= FSL_ESW_PSNP_MODE(1);
+ else if (mode == 2)
+ tmp |= FSL_ESW_PSNP_MODE(2);/*discard*/
+ else if (mode == 3) /*disable the port function*/
+ tmp &= (~FSL_ESW_PSNP_EN);
+ else {
+ printk(KERN_ERR "%s: the mode %x we do not support\n",
+ __func__, mode);
+ return -1;
+ }
+
+ if (compare_num == 1)
+ tmp |= FSL_ESW_PSNP_CS;
+ else if (compare_num == 2)
+ tmp |= FSL_ESW_PSNP_CD;
+ else if (compare_num == 3)
+ tmp |= FSL_ESW_PSNP_CD | FSL_ESW_PSNP_CS;
+ else {
+ printk(KERN_ERR "%s: the compare port address %x"
+ " we do not support\n",
+ __func__, compare_num);
+ return -1;
+ }
+
+ for (num = 0; num < 8; num++) {
+ u32 reg = readl(fep->membase + FEC_ESW_PSNP(num));
+ if (compare_port ==
+ AT_EXTRACT_TCP_UDP_PORT(reg)) {
+ writel(tmp | FSL_ESW_PSNP_PORT_COMPARE(compare_port),
+ fep->membase + FEC_ESW_PSNP(num));
+ break;
+ } else if (!reg) {
+ writel(tmp | FSL_ESW_PSNP_PORT_COMPARE(compare_port),
+ fep->membase + FEC_ESW_PSNP(num));
+ break;
+ }
+ }
+ if (num == 8) {
+ printk(KERN_INFO "TCP/UDP port snooping table is full\n");
+ return 0;
+ }
+
+ return 0;
+}
+
+void esw_get_tcpudp_port_snoop_config(
+ struct switch_enet_private *fep,
+ unsigned long *ulpESW_PSNP)
+{
+ int i;
+
+ for (i = 0; i < 8; i++)
+ *(ulpESW_PSNP + i) = readl(fep->membase + FEC_ESW_PSNP(i));
+}
+
+/*mirror*/
+void esw_get_port_mirroring(struct switch_enet_private *fep)
+{
+ unsigned long tmp, tmp1, tmp2;
+
+ tmp = readl(fep->membase + FEC_ESW_MCR);
+
+ printk(KERN_INFO "Mirror Port: %ld Egress Port Match:%s "
+ "Ingress Port Match:%s\n", tmp & 0xf,
+ tmp & 1 ? "Y" : "N",
+ tmp & 1 ? "Y" : "N");
+
+ if ((tmp >> 6) & 1)
+ printk(KERN_INFO "Egress Port to be mirrored: Port %d\n",
+ readl(fep->membase + FEC_ESW_EGMAP) >> 1);
+ if ((tmp >> 5) & 1)
+ printk(KERN_INFO "Ingress Port to be mirrored: Port %d\n",
+ readl(fep->membase + FEC_ESW_INGMAP) >> 1);
+
+ printk(KERN_INFO "Egress Des Address Match:%s "
+ "Egress Src Address Match:%s\n",
+ (tmp >> 10) & 1 ? "Y" : "N",
+ (tmp >> 9) & 1 ? "Y" : "N");
+ printk(KERN_INFO "Ingress Des Address Match:%s "
+ "Ingress Src Address Match:%s\n",
+ (tmp >> 8) & 1 ? "Y" : "N",
+ (tmp >> 7) & 1 ? "Y" : "N");
+
+ if ((tmp >> 10) & 1) {
+ tmp1 = readl(fep->membase + FEC_ESW_ENGDAL);
+ tmp2 = readl(fep->membase + FEC_ESW_ENGDAH);
+ printk(KERN_INFO "Egress Des Address to be mirrored: "
+ "%02lx-%02lx-%02lx-%02lx-%02lx-%02lx\n",
+ tmp1 & 0xff,
+ (tmp1 >> 8) & 0xff,
+ (tmp1 >> 16) & 0xff,
+ (tmp1 >> 24) & 0xff,
+ tmp2 & 0xff,
+ (tmp2 >> 8) & 0xff);
+ }
+ if ((tmp >> 9) & 1) {
+ tmp1 = readl(fep->membase + FEC_ESW_ENGSAL);
+ tmp2 = readl(fep->membase + FEC_ESW_ENGSAH);
+ printk("Egress Src Address to be mirrored: "
+ "%02lx-%02lx-%02lx-%02lx-%02lx-%02lx\n",
+ tmp1 & 0xff,
+ (tmp1 >> 8) & 0xff,
+ (tmp1 >> 16) & 0xff,
+ (tmp1 >> 24) & 0xff,
+ tmp2 & 0xff,
+ (tmp2 >> 8) & 0xff);
+ }
+ if ((tmp >> 8) & 1) {
+ tmp1 = readl(fep->membase + FEC_ESW_INGDAL);
+ tmp2 = readl(fep->membase + FEC_ESW_INGDAH);
+ printk("Ingress Des Address to be mirrored: "
+ "%02lx-%02lx-%02lx-%02lx-%02lx-%02lx\n",
+ tmp1 & 0xff,
+ (tmp1 >> 8) & 0xff,
+ (tmp1 >> 16) & 0xff,
+ (tmp1 >> 24) & 0xff,
+ tmp2 & 0xff,
+ (tmp2 >> 8) & 0xff);
+ }
+ if ((tmp >> 7) & 1) {
+ tmp1 = readl(fep->membase + FEC_ESW_INGSAL);
+ tmp2 = readl(fep->membase + FEC_ESW_INGSAH);
+ printk("Ingress Src Address to be mirrored: "
+ "%02lx-%02lx-%02lx-%02lx-%02lx-%02lx\n",
+ tmp1 & 0xff,
+ (tmp1 >> 8) & 0xff,
+ (tmp1 >> 16) & 0xff,
+ (tmp1 >> 24) & 0xff,
+ tmp2 & 0xff,
+ (tmp2 >> 8) & 0xff);
+ }
+}
+
+int esw_port_mirroring_config_port_match(struct switch_enet_private *fep,
+ int mirror_port, int port_match_en, int port)
+{
+ unsigned long tmp = 0;
+
+ tmp = readl(fep->membase + FEC_ESW_MCR);
+ if (mirror_port != (tmp & 0xf))
+ tmp = 0;
+
+ switch (port_match_en) {
+ case MIRROR_EGRESS_PORT_MATCH:
+ tmp |= FSL_ESW_MCR_EGMAP;
+ if (port == 0)
+ writel(FSL_ESW_EGMAP_EG0, fep->membase + FEC_ESW_EGMAP);
+ else if (port == 1)
+ writel(FSL_ESW_EGMAP_EG1, fep->membase + FEC_ESW_EGMAP);
+ else if (port == 2)
+ writel(FSL_ESW_EGMAP_EG2, fep->membase + FEC_ESW_EGMAP);
+ break;
+ case MIRROR_INGRESS_PORT_MATCH:
+ tmp |= FSL_ESW_MCR_INGMAP;
+ if (port == 0)
+ writel(FSL_ESW_INGMAP_ING0,
+ fep->membase + FEC_ESW_INGMAP);
+ else if (port == 1)
+ writel(FSL_ESW_INGMAP_ING1,
+ fep->membase + FEC_ESW_INGMAP);
+ else if (port == 2)
+ writel(FSL_ESW_INGMAP_ING2,
+ fep->membase + FEC_ESW_INGMAP);
+ break;
+ default:
+ tmp = 0;
+ break;
+ }
+
+ tmp = tmp & 0x07e0;
+ if (port_match_en)
+ tmp |= FSL_ESW_MCR_MEN | FSL_ESW_MCR_PORT(mirror_port);
+
+ writel(tmp, fep->membase + FEC_ESW_MCR);
+ return 0;
+}
+
+int esw_port_mirroring_config(struct switch_enet_private *fep,
+ int mirror_port, int port, int mirror_enable,
+ unsigned char *src_mac, unsigned char *des_mac,
+ int egress_en, int ingress_en,
+ int egress_mac_src_en, int egress_mac_des_en,
+ int ingress_mac_src_en, int ingress_mac_des_en)
+{
+ unsigned long tmp;
+
+ /*mirroring config*/
+ tmp = 0;
+ if (egress_en == 1) {
+ tmp |= FSL_ESW_MCR_EGMAP;
+ if (port == 0)
+ writel(FSL_ESW_EGMAP_EG0, fep->membase + FEC_ESW_EGMAP);
+ else if (port == 1)
+ writel(FSL_ESW_EGMAP_EG1, fep->membase + FEC_ESW_EGMAP);
+ else if (port == 2)
+ writel(FSL_ESW_EGMAP_EG2, fep->membase + FEC_ESW_EGMAP);
+ else {
+ printk(KERN_ERR "%s: the port %x we do not support\n",
+ __func__, port);
+ return -1;
+ }
+ } else if (egress_en == 0) {
+ tmp &= (~FSL_ESW_MCR_EGMAP);
+ } else {
+ printk(KERN_ERR "%s: egress_en %x we do not support\n",
+ __func__, egress_en);
+ return -1;
+ }
+
+ if (ingress_en == 1) {
+ tmp |= FSL_ESW_MCR_INGMAP;
+ if (port == 0)
+ writel(FSL_ESW_INGMAP_ING0,
+ fep->membase + FEC_ESW_INGMAP);
+ else if (port == 1)
+ writel(FSL_ESW_INGMAP_ING1,
+ fep->membase + FEC_ESW_INGMAP);
+ else if (port == 2)
+ writel(FSL_ESW_INGMAP_ING2,
+ fep->membase + FEC_ESW_INGMAP);
+ else {
+ printk(KERN_ERR "%s: the port %x we do not support\n",
+ __func__, port);
+ return -1;
+ }
+ } else if (ingress_en == 0) {
+ tmp &= ~FSL_ESW_MCR_INGMAP;
+ } else{
+ printk(KERN_ERR "%s: ingress_en %x we do not support\n",
+ __func__, ingress_en);
+ return -1;
+ }
+
+ if (egress_mac_src_en == 1) {
+ tmp |= FSL_ESW_MCR_EGSA;
+ writel((src_mac[5] << 8) | (src_mac[4]),
+ fep->membase + FEC_ESW_ENGSAH);
+ writel((unsigned long)((src_mac[3] << 24) | (src_mac[2] << 16) |
+ (src_mac[1] << 8) | src_mac[0]),
+ fep->membase + FEC_ESW_ENGSAL);
+ } else if (egress_mac_src_en == 0) {
+ tmp &= ~FSL_ESW_MCR_EGSA;
+ } else {
+ printk(KERN_ERR "%s: egress_mac_src_en %x we do not support\n",
+ __func__, egress_mac_src_en);
+ return -1;
+ }
+
+ if (egress_mac_des_en == 1) {
+ tmp |= FSL_ESW_MCR_EGDA;
+ writel((des_mac[5] << 8) | (des_mac[4]),
+ fep->membase + FEC_ESW_ENGDAH);
+ writel((unsigned long)((des_mac[3] << 24) | (des_mac[2] << 16) |
+ (des_mac[1] << 8) | des_mac[0]),
+ fep->membase + FEC_ESW_ENGDAL);
+ } else if (egress_mac_des_en == 0) {
+ tmp &= ~FSL_ESW_MCR_EGDA;
+ } else {
+ printk(KERN_ERR "%s: egress_mac_des_en %x we do not support\n",
+ __func__, egress_mac_des_en);
+ return -1;
+ }
+
+ if (ingress_mac_src_en == 1) {
+ tmp |= FSL_ESW_MCR_INGSA;
+ writel((src_mac[5] << 8) | (src_mac[4]),
+ fep->membase + FEC_ESW_INGSAH);
+ writel((unsigned long)((src_mac[3] << 24) | (src_mac[2] << 16) |
+ (src_mac[1] << 8) | src_mac[0]),
+ fep->membase + FEC_ESW_INGSAL);
+ } else if (ingress_mac_src_en == 0) {
+ tmp &= ~FSL_ESW_MCR_INGSA;
+ } else {
+ printk(KERN_ERR "%s: ingress_mac_src_en %x "
+ "we do not support\n",
+ __func__, ingress_mac_src_en);
+ return -1;
+ }
+
+ if (ingress_mac_des_en == 1) {
+ tmp |= FSL_ESW_MCR_INGDA;
+ writel((des_mac[5] << 8) | (des_mac[4]),
+ fep->membase + FEC_ESW_INGDAH);
+ writel((unsigned long)((des_mac[3] << 24) | (des_mac[2] << 16) |
+ (des_mac[1] << 8) | des_mac[0]),
+ fep->membase + FEC_ESW_INGDAL);
+ } else if (ingress_mac_des_en == 0) {
+ tmp &= ~FSL_ESW_MCR_INGDA;
+ } else {
+ printk(KERN_ERR "%s: ingress_mac_des_en %x we do not support\n",
+ __func__, ingress_mac_des_en);
+ return -1;
+ }
+
+ if (mirror_enable == 1)
+ tmp |= FSL_ESW_MCR_MEN | FSL_ESW_MCR_PORT(mirror_port);
+ else if (mirror_enable == 0)
+ tmp &= ~FSL_ESW_MCR_MEN;
+ else
+ printk(KERN_ERR "%s: the mirror enable %x is error\n",
+ __func__, mirror_enable);
+
+
+ writel(tmp, fep->membase + FEC_ESW_MCR);
+ return 0;
+}
+
+int esw_port_mirroring_config_addr_match(struct switch_enet_private *fep,
+ int mirror_port, int addr_match_enable, unsigned char *mac_addr)
+{
+ unsigned long tmp = 0;
+
+ tmp = readl(fep->membase + FEC_ESW_MCR);
+ if (mirror_port != (tmp & 0xf))
+ tmp = 0;
+
+ switch (addr_match_enable) {
+ case MIRROR_EGRESS_SOURCE_MATCH:
+ tmp |= FSL_ESW_MCR_EGSA;
+ writel((mac_addr[5] << 8) | (mac_addr[4]),
+ fep->membase + FEC_ESW_ENGSAH);
+ writel((unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) | (mac_addr[1] << 8) | mac_addr[0]),
+ fep->membase + FEC_ESW_ENGSAL);
+ break;
+ case MIRROR_INGRESS_SOURCE_MATCH:
+ tmp |= FSL_ESW_MCR_INGSA;
+ writel((mac_addr[5] << 8) | (mac_addr[4]),
+ fep->membase + FEC_ESW_INGSAH);
+ writel((unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) | (mac_addr[1] << 8) | mac_addr[0]),
+ fep->membase + FEC_ESW_INGSAL);
+ break;
+ case MIRROR_EGRESS_DESTINATION_MATCH:
+ tmp |= FSL_ESW_MCR_EGDA;
+ writel((mac_addr[5] << 8) | (mac_addr[4]),
+ fep->membase + FEC_ESW_ENGDAH);
+ writel((unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) | (mac_addr[1] << 8) | mac_addr[0]),
+ fep->membase + FEC_ESW_ENGDAL);
+ break;
+ case MIRROR_INGRESS_DESTINATION_MATCH:
+ tmp |= FSL_ESW_MCR_INGDA;
+ writel((mac_addr[5] << 8) | (mac_addr[4]),
+ fep->membase + FEC_ESW_INGDAH);
+ writel((unsigned long)((mac_addr[3] << 24) |
+ (mac_addr[2] << 16) | (mac_addr[1] << 8) | mac_addr[0]),
+ fep->membase + FEC_ESW_INGDAL);
+ break;
+ default:
+ tmp = 0;
+ break;
+ }
+
+ tmp = tmp & 0x07e0;
+ if (addr_match_enable)
+ tmp |= FSL_ESW_MCR_MEN | FSL_ESW_MCR_PORT(mirror_port);
+
+ writel(tmp, fep->membase + FEC_ESW_MCR);
+ return 0;
+}
+
+void esw_get_vlan_verification(struct switch_enet_private *fep,
+ unsigned long *ulValue)
+{
+ *ulValue = readl(fep->membase + FEC_ESW_VLANV);
+}
+
+int esw_set_vlan_verification(struct switch_enet_private *fep, int port,
+ int vlan_domain_verify_en, int vlan_discard_unknown_en)
+{
+ u32 tmp;
+
+ if ((port < 0) || (port > 2)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+ tmp = readl(fep->membase + FEC_ESW_VLANV);
+ if (vlan_domain_verify_en == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_VLANV_VV0;
+ else if (port == 1)
+ tmp |= FSL_ESW_VLANV_VV1;
+ else if (port == 2)
+ tmp |= FSL_ESW_VLANV_VV2;
+ } else if (vlan_domain_verify_en == 0) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_VLANV_VV0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_VLANV_VV1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_VLANV_VV2;
+ } else {
+ printk(KERN_INFO "%s: donot support "
+ "vlan_domain_verify %x\n",
+ __func__, vlan_domain_verify_en);
+ return -2;
+ }
+
+ if (vlan_discard_unknown_en == 1) {
+ if (port == 0)
+ tmp |= FSL_ESW_VLANV_DU0;
+ else if (port == 1)
+ tmp |= FSL_ESW_VLANV_DU1;
+ else if (port == 2)
+ tmp |= FSL_ESW_VLANV_DU2;
+ } else if (vlan_discard_unknown_en == 0) {
+ if (port == 0)
+ tmp &= ~FSL_ESW_VLANV_DU0;
+ else if (port == 1)
+ tmp &= ~FSL_ESW_VLANV_DU1;
+ else if (port == 2)
+ tmp &= ~FSL_ESW_VLANV_DU2;
+ } else {
+ printk(KERN_INFO "%s: donot support "
+ "vlan_discard_unknown %x\n",
+ __func__, vlan_discard_unknown_en);
+ return -3;
+ }
+
+ writel(tmp, fep->membase + FEC_ESW_VLANV);
+ return 0;
+}
+
+void esw_get_vlan_resolution_table(struct switch_enet_private *fep,
+ struct eswVlanTableItem *tableaddr)
+{
+ int vnum = 0;
+ int i;
+ u32 tmp;
+
+ for (i = 0; i < 32; i++) {
+ tmp = readl(fep->membase + FEC_ESW_VRES(i));
+ if (tmp) {
+ tableaddr->table[i].port_vlanid =
+ tmp >> 3;
+ tableaddr->table[i].vlan_domain_port =
+ tmp & 7;
+ vnum++;
+ }
+ }
+ tableaddr->valid_num = vnum;
+}
+
+int esw_set_vlan_id(struct switch_enet_private *fep, unsigned long configData)
+{
+ int i;
+ u32 tmp;
+
+ for (i = 0; i < 32; i++) {
+ tmp = readl(fep->membase + FEC_ESW_VRES(i));
+ if (tmp == 0) {
+ writel(FSL_ESW_VRES_VLANID(configData),
+ fep->membase + FEC_ESW_VRES(i));
+ return 0;
+ } else if (((tmp >> 3) & 0xfff) == configData) {
+ printk(KERN_INFO "The VLAN already exists\n");
+ return 0;
+ }
+ }
+
+ printk(KERN_INFO "The VLAN can't create, because VLAN table is full\n");
+ return 0;
+}
+
+int esw_set_vlan_id_cleared(struct switch_enet_private *fep,
+ unsigned long configData)
+{
+ int i;
+ u32 tmp;
+
+ for (i = 0; i < 32; i++) {
+ tmp = readl(fep->membase + FEC_ESW_VRES(i));
+ if (((tmp >> 3) & 0xfff) == configData) {
+ writel(0, fep->membase + FEC_ESW_VRES(i));
+ break;
+ }
+ }
+ return 0;
+}
+
+int esw_set_port_in_vlan_id(struct switch_enet_private *fep,
+ eswIoctlVlanResoultionTable configData)
+{
+ int i;
+ int lastnum = 0;
+ u32 tmp;
+
+ for (i = 0; i < 32; i++) {
+ tmp = readl(fep->membase + FEC_ESW_VRES(i));
+ if (tmp == 0) {
+ lastnum = i;
+ break;
+ } else if (((tmp >> 3) & 0xfff) ==
+ configData.port_vlanid) {
+ /* update the port members of this vlan */
+ tmp |= 1 << configData.vlan_domain_port;
+ writel(tmp, fep->membase + FEC_ESW_VRES(i));
+ return 0;
+ }
+ }
+ /* creat a new vlan in vlan table */
+ writel(FSL_ESW_VRES_VLANID(configData.port_vlanid) |
+ (1 << configData.vlan_domain_port),
+ fep->membase + FEC_ESW_VRES(lastnum));
+ return 0;
+}
+
+int esw_set_vlan_resolution_table(struct switch_enet_private *fep,
+ unsigned short port_vlanid, int vlan_domain_num,
+ int vlan_domain_port)
+{
+ if ((vlan_domain_num < 0)
+ || (vlan_domain_num > 31)) {
+ printk(KERN_ERR "%s: do not support the "
+ "vlan_domain_num %d\n",
+ __func__, vlan_domain_num);
+ return -1;
+ }
+
+ if ((vlan_domain_port < 0)
+ || (vlan_domain_port > 7)) {
+ printk(KERN_ERR "%s: do not support the "
+ "vlan_domain_port %d\n",
+ __func__, vlan_domain_port);
+ return -2;
+ }
+
+ writel(
+ FSL_ESW_VRES_VLANID(port_vlanid)
+ | vlan_domain_port,
+ fep->membase + FEC_ESW_VRES(vlan_domain_num));
+
+ return 0;
+}
+
+void esw_get_vlan_input_config(struct switch_enet_private *fep,
+ eswIoctlVlanInputStatus *pVlanInputConfig)
+{
+ int i;
+
+ for (i = 0; i < 3; i++)
+ pVlanInputConfig->ESW_PID[i] =
+ readl(fep->membase + FEC_ESW_PID(i));
+
+ pVlanInputConfig->ESW_VLANV = readl(fep->membase + FEC_ESW_VLANV);
+ pVlanInputConfig->ESW_VIMSEL = readl(fep->membase + FEC_ESW_VIMSEL);
+ pVlanInputConfig->ESW_VIMEN = readl(fep->membase + FEC_ESW_VIMEN);
+
+ for (i = 0; i < 32; i++)
+ pVlanInputConfig->ESW_VRES[i] =
+ readl(fep->membase + FEC_ESW_VRES(i));
+}
+
+
+int esw_vlan_input_process(struct switch_enet_private *fep,
+ int port, int mode, unsigned short port_vlanid)
+{
+ u32 tmp;
+
+ if ((mode < 0) || (mode > 5)) {
+ printk(KERN_ERR "%s: do not support the"
+ " VLAN input processing mode %d\n",
+ __func__, mode);
+ return -1;
+ }
+
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, mode);
+ return -2;
+ }
+
+ writel(FSL_ESW_PID_VLANID(port_vlanid),
+ fep->membase + FEC_ESW_PID(port));
+
+ if (port == 0) {
+ tmp = readl(fep->membase + FEC_ESW_VIMEN);
+ if (mode == 4)
+ tmp &= ~FSL_ESW_VIMEN_EN0;
+ else
+ tmp |= FSL_ESW_VIMEN_EN0;
+ writel(tmp, fep->membase + FEC_ESW_VIMEN);
+
+ tmp = readl(fep->membase + FEC_ESW_VIMSEL);
+ tmp &= ~FSL_ESW_VIMSEL_IM0(3);
+ tmp |= FSL_ESW_VIMSEL_IM0(mode);
+ writel(tmp, fep->membase + FEC_ESW_VIMSEL);
+ } else if (port == 1) {
+ tmp = readl(fep->membase + FEC_ESW_VIMEN);
+ if (mode == 4)
+ tmp &= ~FSL_ESW_VIMEN_EN1;
+ else
+ tmp |= FSL_ESW_VIMEN_EN1;
+ writel(tmp, fep->membase + FEC_ESW_VIMEN);
+
+ tmp = readl(fep->membase + FEC_ESW_VIMSEL);
+ tmp &= ~FSL_ESW_VIMSEL_IM1(3);
+ tmp |= FSL_ESW_VIMSEL_IM1(mode);
+ writel(tmp, fep->membase + FEC_ESW_VIMSEL);
+ } else if (port == 2) {
+ tmp = readl(fep->membase + FEC_ESW_VIMEN);
+ if (mode == 4)
+ tmp &= ~FSL_ESW_VIMEN_EN2;
+ else
+ tmp |= FSL_ESW_VIMEN_EN2;
+ writel(tmp, fep->membase + FEC_ESW_VIMEN);
+
+ tmp = readl(fep->membase + FEC_ESW_VIMSEL);
+ tmp &= ~FSL_ESW_VIMSEL_IM2(3);
+ tmp |= FSL_ESW_VIMSEL_IM2(mode);
+ writel(tmp, fep->membase + FEC_ESW_VIMSEL);
+ } else {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -2;
+ }
+
+ return 0;
+}
+
+void esw_get_vlan_output_config(struct switch_enet_private *fep,
+ unsigned long *ulVlanOutputConfig)
+{
+ *ulVlanOutputConfig = readl(fep->membase + FEC_ESW_VOMSEL);
+}
+
+int esw_vlan_output_process(struct switch_enet_private *fep,
+ int port, int mode)
+{
+ u32 tmp;
+
+ if ((port < 0) || (port > 2)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, mode);
+ return -1;
+ }
+ tmp = readl(fep->membase + FEC_ESW_VOMSEL);
+ if (port == 0) {
+ tmp &= ~FSL_ESW_VOMSEL_OM0(3);
+ tmp |= FSL_ESW_VOMSEL_OM0(mode);
+ } else if (port == 1) {
+ tmp &= ~FSL_ESW_VOMSEL_OM1(3);
+ tmp |= FSL_ESW_VOMSEL_OM1(mode);
+ } else if (port == 2) {
+ tmp &= ~FSL_ESW_VOMSEL_OM2(3);
+ tmp |= FSL_ESW_VOMSEL_OM2(mode);
+ } else {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+ writel(tmp, fep->membase + FEC_ESW_VOMSEL);
+ return 0;
+}
+
+/*
+ * frame calssify and priority resolution
+ * vlan priority lookup
+ */
+int esw_framecalssify_vlan_priority_lookup(struct switch_enet_private *fep,
+ int port, int func_enable, int vlan_pri_table_num,
+ int vlan_pri_table_value)
+{
+ u32 tmp;
+
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+
+ if (func_enable == 0) {
+ tmp = readl(fep->membase + FEC_ESW_PRES(port));
+ tmp &= ~FSL_ESW_PRES_VLAN;
+ writel(tmp, fep->membase + FEC_ESW_PRES(port));
+ printk(KERN_ERR "%s: disable port %d VLAN priority "
+ "lookup function\n", __func__, port);
+ return 0;
+ }
+
+ if ((vlan_pri_table_num < 0) || (vlan_pri_table_num > 7)) {
+ printk(KERN_ERR "%s: do not support the priority %d\n",
+ __func__, vlan_pri_table_num);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_PVRES(port));
+ tmp |= ((vlan_pri_table_value & 0x3)
+ << (vlan_pri_table_num*3));
+ writel(tmp, fep->membase + FEC_ESW_PVRES(port));
+
+ /* enable port VLAN priority lookup function*/
+ tmp = readl(fep->membase + FEC_ESW_PRES(port));
+ tmp |= FSL_ESW_PRES_VLAN;
+ writel(tmp, fep->membase + FEC_ESW_PRES(port));
+ return 0;
+}
+
+int esw_framecalssify_ip_priority_lookup(struct switch_enet_private *fep,
+ int port, int func_enable, int ipv4_en, int ip_priority_num,
+ int ip_priority_value)
+{
+ unsigned long tmp = 0, tmp_prio = 0;
+
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+
+ if (func_enable == 0) {
+ tmp = readl(fep->membase + FEC_ESW_PRES(port));
+ tmp &= ~FSL_ESW_PRES_IP;
+ writel(tmp, fep->membase + FEC_ESW_PRES(port));
+ printk(KERN_ERR "%s: disable port %d ip priority "
+ "lookup function\n", __func__, port);
+ return 0;
+ }
+
+ /* IPV4 priority 64 entry table lookup*/
+ /* IPv4 head 6 bit TOS field*/
+ if (ipv4_en == 1) {
+ if ((ip_priority_num < 0) || (ip_priority_num > 63)) {
+ printk(KERN_ERR "%s: do not support the table entry %d\n",
+ __func__, ip_priority_num);
+ return -2;
+ }
+ } else { /* IPV6 priority 256 entry table lookup*/
+ /* IPv6 head 8 bit COS field*/
+ if ((ip_priority_num < 0) || (ip_priority_num > 255)) {
+ printk(KERN_ERR "%s: do not support the table entry %d\n",
+ __func__, ip_priority_num);
+ return -3;
+ }
+ }
+
+ /* IP priority table lookup : address*/
+ tmp = FSL_ESW_IPRES_ADDRESS(ip_priority_num);
+ /* IP priority table lookup : ipv4sel*/
+ if (ipv4_en == 1)
+ tmp = tmp | FSL_ESW_IPRES_IPV4SEL;
+ /* IP priority table lookup : priority*/
+ if (port == 0)
+ tmp |= FSL_ESW_IPRES_PRI0(ip_priority_value);
+ else if (port == 1)
+ tmp |= FSL_ESW_IPRES_PRI1(ip_priority_value);
+ else if (port == 2)
+ tmp |= FSL_ESW_IPRES_PRI2(ip_priority_value);
+
+ /* configure*/
+ writel(FSL_ESW_IPRES_READ | FSL_ESW_IPRES_ADDRESS(ip_priority_num),
+ fep->membase + FEC_ESW_IPRES);
+ tmp_prio = readl(fep->membase + FEC_ESW_IPRES);
+
+ writel(tmp | tmp_prio, fep->membase + FEC_ESW_IPRES);
+
+ writel(FSL_ESW_IPRES_READ | FSL_ESW_IPRES_ADDRESS(ip_priority_num),
+ fep->membase + FEC_ESW_IPRES);
+ tmp_prio = readl(fep->membase + FEC_ESW_IPRES);
+
+ /* enable port IP priority lookup function*/
+ tmp = readl(fep->membase + FEC_ESW_PRES(port));
+ tmp |= FSL_ESW_PRES_IP;
+ writel(tmp, fep->membase + FEC_ESW_PRES(port));
+ return 0;
+}
+
+int esw_framecalssify_mac_priority_lookup(
+ struct switch_enet_private *fep, int port)
+{
+ u32 tmp;
+
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+
+ tmp = readl(fep->membase + FEC_ESW_PRES(port));
+ tmp |= FSL_ESW_PRES_MAC;
+ writel(tmp, fep->membase + FEC_ESW_PRES(port));
+
+ return 0;
+}
+
+int esw_frame_calssify_priority_init(struct switch_enet_private *fep,
+ int port, unsigned char priority_value)
+{
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+ /*disable all priority lookup function*/
+ writel(0, fep->membase + FEC_ESW_PRES(port));
+ writel(FSL_ESW_PRES_DFLT_PRI(priority_value & 0x7),
+ fep->membase + FEC_ESW_PRES(port));
+
+ return 0;
+}
+
+int esw_get_statistics_status(struct switch_enet_private *fep,
+ esw_statistics_status *pStatistics)
+{
+ pStatistics->ESW_DISCN = readl(fep->membase + FEC_ESW_DISCN);
+ pStatistics->ESW_DISCB = readl(fep->membase + FEC_ESW_DISCB);
+ pStatistics->ESW_NDISCN = readl(fep->membase + FEC_ESW_NDISCN);
+ pStatistics->ESW_NDISCB = readl(fep->membase + FEC_ESW_NDISCB);
+ return 0;
+}
+
+int esw_get_port_statistics_status(struct switch_enet_private *fep,
+ int port, esw_port_statistics_status *pPortStatistics)
+{
+ if ((port < 0) || (port > 3)) {
+ printk(KERN_ERR "%s: do not support the port %d\n",
+ __func__, port);
+ return -1;
+ }
+
+ pPortStatistics->FSL_ESW_POQC =
+ readl(fep->membase + FEC_ESW_POQC(port));
+ pPortStatistics->FSL_ESW_PMVID =
+ readl(fep->membase + FEC_ESW_PMVID(port));
+ pPortStatistics->FSL_ESW_PMVTAG =
+ readl(fep->membase + FEC_ESW_PMVTAG(port));
+ pPortStatistics->FSL_ESW_PBL =
+ readl(fep->membase + FEC_ESW_PBL(port));
+ return 0;
+}
+
+int esw_get_output_queue_status(struct switch_enet_private *fep,
+ esw_output_queue_status *pOutputQueue)
+{
+ pOutputQueue->ESW_MMSR = readl(fep->membase + FEC_ESW_MMSR);
+ pOutputQueue->ESW_LMT = readl(fep->membase + FEC_ESW_LMT);
+ pOutputQueue->ESW_LFC = readl(fep->membase + FEC_ESW_LFC);
+ pOutputQueue->ESW_IOSR = readl(fep->membase + FEC_ESW_IOSR);
+ pOutputQueue->ESW_PCSR = readl(fep->membase + FEC_ESW_PCSR);
+ pOutputQueue->ESW_QWT = readl(fep->membase + FEC_ESW_QWT);
+ pOutputQueue->ESW_P0BCT = readl(fep->membase + FEC_ESW_P0BCT);
+ return 0;
+}
+
+/* set output queue memory status and configure*/
+int esw_set_output_queue_memory(struct switch_enet_private *fep,
+ int fun_num, esw_output_queue_status *pOutputQueue)
+{
+ if (fun_num == 1) {
+ /* memory manager status*/
+ writel(pOutputQueue->ESW_MMSR, fep->membase + FEC_ESW_MMSR);
+ } else if (fun_num == 2) {
+ /*low memory threshold*/
+ writel(pOutputQueue->ESW_LMT, fep->membase + FEC_ESW_LMT);
+ } else if (fun_num == 3) {
+ /*lowest number of free cells*/
+ writel(pOutputQueue->ESW_LFC, fep->membase + FEC_ESW_LFC);
+ } else if (fun_num == 4) {
+ /*queue weights*/
+ writel(pOutputQueue->ESW_QWT, fep->membase + FEC_ESW_QWT);
+ } else if (fun_num == 5) {
+ /*port 0 backpressure congenstion thresled*/
+ writel(pOutputQueue->ESW_P0BCT, fep->membase + FEC_ESW_P0BCT);
+ } else {
+ printk(KERN_ERR "%s: do not support the cmd %x\n",
+ __func__, fun_num);
+ return -1;
+ }
+ return 0;
+}
+
+int esw_get_irq_status(struct switch_enet_private *fep,
+ eswIoctlIrqStatus *pIrqStatus)
+{
+ pIrqStatus->isr = readl(fep->membase + FEC_ESW_ISR);
+ pIrqStatus->imr = readl(fep->membase + FEC_ESW_IMR);
+ pIrqStatus->rx_buf_pointer = readl(fep->membase + FEC_ESW_RDSR);
+ pIrqStatus->tx_buf_pointer = readl(fep->membase + FEC_ESW_TDSR);
+ pIrqStatus->rx_max_size = readl(fep->membase + FEC_ESW_MRBR);
+ pIrqStatus->rx_buf_active = readl(fep->membase + FEC_ESW_RDAR);
+ pIrqStatus->tx_buf_active = readl(fep->membase + FEC_ESW_TDAR);
+ return 0;
+}
+
+int esw_set_irq_mask(struct switch_enet_private *fep,
+ unsigned long mask, int enable)
+{
+ u32 tmp = readl(fep->membase + FEC_ESW_IMR);
+
+ if (enable == 1)
+ tmp |= mask;
+ else if (enable == 1)
+ tmp &= (~mask);
+ else {
+ printk(KERN_INFO "%s: enable %lx is error value\n",
+ __func__, mask);
+ return -1;
+ }
+ writel(tmp, fep->membase + FEC_ESW_IMR);
+ return 0;
+}
+
+void esw_clear_irq_event(struct switch_enet_private *fep,
+ unsigned long mask)
+{
+ u32 tmp = readl(fep->membase + FEC_ESW_ISR);
+ tmp |= mask;
+ writel(tmp, fep->membase + FEC_ESW_ISR);
+}
+
+void esw_get_switch_mode(struct switch_enet_private *fep,
+ unsigned long *ulModeConfig)
+{
+ *ulModeConfig = readl(fep->membase + FEC_ESW_MODE);
+}
+
+void esw_switch_mode_configure(struct switch_enet_private *fep,
+ unsigned long configure)
+{
+ u32 tmp = readl(fep->membase + FEC_ESW_MODE);
+ tmp |= configure;
+ writel(tmp, fep->membase + FEC_ESW_MODE);
+}
+
+void esw_get_bridge_port(struct switch_enet_private *fep,
+ unsigned long *ulBMPConfig)
+{
+ *ulBMPConfig = readl(fep->membase + FEC_ESW_BMPC);
+}
+
+void esw_bridge_port_configure(struct switch_enet_private *fep,
+ unsigned long configure)
+{
+ writel(configure, fep->membase + FEC_ESW_BMPC);
+}
+
+int esw_get_port_all_status(struct switch_enet_private *fep,
+ unsigned char portnum, struct port_all_status *port_alstatus)
+{
+ unsigned long PortBlocking;
+ unsigned long PortLearning;
+ unsigned long VlanVerify;
+ unsigned long DiscardUnknown;
+ unsigned long MultiReso;
+ unsigned long BroadReso;
+ unsigned long FTransmit;
+ unsigned long FReceive;
+
+ PortBlocking = readl(fep->membase + FEC_ESW_BKLR) & 0x0000000f;
+ PortLearning = (readl(fep->membase + FEC_ESW_BKLR) & 0x000f0000) >> 16;
+ VlanVerify = readl(fep->membase + FEC_ESW_VLANV) & 0x0000000f;
+ DiscardUnknown = (readl(fep->membase + FEC_ESW_VLANV) & 0x000f0000)
+ >> 16;
+ MultiReso = readl(fep->membase + FEC_ESW_DMCR) & 0x0000000f;
+ BroadReso = readl(fep->membase + FEC_ESW_DBCR) & 0x0000000f;
+ FTransmit = readl(fep->membase + FEC_ESW_PER) & 0x0000000f;
+ FReceive = (readl(fep->membase + FEC_ESW_PER) & 0x000f0000) >> 16;
+
+ switch (portnum) {
+ case 0:
+ port_alstatus->link_status = 1;
+ port_alstatus->block_status = PortBlocking & 1;
+ port_alstatus->learn_status = PortLearning & 1;
+ port_alstatus->vlan_verify = VlanVerify & 1;
+ port_alstatus->discard_unknown = DiscardUnknown & 1;
+ port_alstatus->multi_reso = MultiReso & 1;
+ port_alstatus->broad_reso = BroadReso & 1;
+ port_alstatus->ftransmit = FTransmit & 1;
+ port_alstatus->freceive = FReceive & 1;
+ break;
+ case 1:
+ port_alstatus->link_status =
+ ports_link_status.port1_link_status;
+ port_alstatus->block_status = (PortBlocking >> 1) & 1;
+ port_alstatus->learn_status = (PortLearning >> 1) & 1;
+ port_alstatus->vlan_verify = (VlanVerify >> 1) & 1;
+ port_alstatus->discard_unknown = (DiscardUnknown >> 1) & 1;
+ port_alstatus->multi_reso = (MultiReso >> 1) & 1;
+ port_alstatus->broad_reso = (BroadReso >> 1) & 1;
+ port_alstatus->ftransmit = (FTransmit >> 1) & 1;
+ port_alstatus->freceive = (FReceive >> 1) & 1;
+ break;
+ case 2:
+ port_alstatus->link_status =
+ ports_link_status.port2_link_status;
+ port_alstatus->block_status = (PortBlocking >> 2) & 1;
+ port_alstatus->learn_status = (PortLearning >> 2) & 1;
+ port_alstatus->vlan_verify = (VlanVerify >> 2) & 1;
+ port_alstatus->discard_unknown = (DiscardUnknown >> 2) & 1;
+ port_alstatus->multi_reso = (MultiReso >> 2) & 1;
+ port_alstatus->broad_reso = (BroadReso >> 2) & 1;
+ port_alstatus->ftransmit = (FTransmit >> 2) & 1;
+ port_alstatus->freceive = (FReceive >> 2) & 1;
+ break;
+ default:
+ printk(KERN_ERR "%s:do not support the port %d",
+ __func__, portnum);
+ break;
+ }
+ return 0;
+}
+
+int esw_atable_get_entry_port_number(struct switch_enet_private *fep,
+ unsigned char *mac_addr, unsigned char *port)
+{
+ int block_index, block_index_end, entry;
+ unsigned long read_lo, read_hi;
+ unsigned long mac_addr_lo, mac_addr_hi;
+
+ mac_addr_lo = (unsigned long)((mac_addr[3]<<24) | (mac_addr[2]<<16) |
+ (mac_addr[1]<<8) | mac_addr[0]);
+ mac_addr_hi = (unsigned long)((mac_addr[5]<<8) | (mac_addr[4]));
+
+ block_index = GET_BLOCK_PTR(crc8_calc(mac_addr));
+ block_index_end = block_index + ATABLE_ENTRY_PER_SLOT;
+
+ /* now search all the entries in the selected block */
+ for (entry = block_index; entry < block_index_end; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ if ((read_lo == mac_addr_lo) &&
+ ((read_hi & 0x0000ffff) ==
+ (mac_addr_hi & 0x0000ffff))) {
+ /* found the correct address */
+ if ((read_hi & (1 << 16)) && (!(read_hi & (1 << 17))))
+ *port = AT_EXTRACT_PORT(read_hi);
+ break;
+ } else
+ *port = -1;
+ }
+
+ return 0;
+}
+
+int esw_get_mac_address_lookup_table(struct switch_enet_private *fep,
+ unsigned long *tableaddr, unsigned long *dnum, unsigned long *snum)
+{
+ unsigned long read_lo, read_hi;
+ unsigned long entry;
+ unsigned long dennum = 0;
+ unsigned long sennum = 0;
+
+ for (entry = 0; entry < ESW_ATABLE_MEM_NUM_ENTRIES; entry++) {
+ read_atable(fep, entry, &read_lo, &read_hi);
+ if ((read_hi & (1 << 17)) && (read_hi & (1 << 16))) {
+ /* static entry */
+ *(tableaddr + (2047 - sennum) * 11) = entry;
+ *(tableaddr + (2047 - sennum) * 11 + 2) =
+ read_lo & 0x000000ff;
+ *(tableaddr + (2047 - sennum) * 11 + 3) =
+ (read_lo & 0x0000ff00) >> 8;
+ *(tableaddr + (2047 - sennum) * 11 + 4) =
+ (read_lo & 0x00ff0000) >> 16;
+ *(tableaddr + (2047 - sennum) * 11 + 5) =
+ (read_lo & 0xff000000) >> 24;
+ *(tableaddr + (2047 - sennum) * 11 + 6) =
+ read_hi & 0x000000ff;
+ *(tableaddr + (2047 - sennum) * 11 + 7) =
+ (read_hi & 0x0000ff00) >> 8;
+ *(tableaddr + (2047 - sennum) * 11 + 8) =
+ AT_EXTRACT_PORTMASK(read_hi);
+ *(tableaddr + (2047 - sennum) * 11 + 9) =
+ AT_EXTRACT_PRIO(read_hi);
+ sennum++;
+ } else if ((read_hi & (1 << 16)) && (!(read_hi & (1 << 17)))) {
+ /* dynamic entry */
+ *(tableaddr + dennum * 11) = entry;
+ *(tableaddr + dennum * 11 + 2) = read_lo & 0xff;
+ *(tableaddr + dennum * 11 + 3) =
+ (read_lo & 0x0000ff00) >> 8;
+ *(tableaddr + dennum * 11 + 4) =
+ (read_lo & 0x00ff0000) >> 16;
+ *(tableaddr + dennum * 11 + 5) =
+ (read_lo & 0xff000000) >> 24;
+ *(tableaddr + dennum * 11 + 6) = read_hi & 0xff;
+ *(tableaddr + dennum * 11 + 7) =
+ (read_hi & 0x0000ff00) >> 8;
+ *(tableaddr + dennum * 11 + 8) =
+ AT_EXTRACT_PORT(read_hi);
+ *(tableaddr + dennum * 11 + 9) =
+ AT_EXTRACT_TIMESTAMP(read_hi);
+ dennum++;
+ }
+ }
+
+ *dnum = dennum;
+ *snum = sennum;
+ return 0;
+}
+
+/* The timer should create an interrupt every 4 seconds*/
+static void l2switch_aging_timer(unsigned long data)
+{
+ struct switch_enet_private *fep;
+
+ fep = (struct switch_enet_private *)data;
+
+ if (fep) {
+ TIMEINCREMENT(fep->currTime);
+ fep->timeChanged++;
+ }
+
+ mod_timer(&fep->timer_aging, jiffies + LEARNING_AGING_TIMER);
+}
+
+static int switch_enet_learning(void *arg)
+{
+ struct switch_enet_private *fep = arg;
+
+ while (!kthread_should_stop()) {
+
+ set_current_state(TASK_INTERRUPTIBLE);
+
+ /* check learning record valid */
+ if (readl(fep->membase + FEC_ESW_LSR))
+ esw_atable_dynamicms_learn_migration(fep,
+ fep->currTime);
+ else
+ schedule_timeout(HZ/100);
+ }
+
+ return 0;
+}
+
+static int switch_enet_ioctl(struct net_device *dev,
+ struct ifreq *ifr, int cmd)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+ int ret = 0;
+
+ switch (cmd) {
+ case ESW_SET_PORTENABLE_CONF:
+ {
+ eswIoctlPortEnableConfig configData;
+ ret = copy_from_user(&configData,
+ ifr->ifr_data,
+ sizeof(eswIoctlPortEnableConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_enable_config(fep,
+ configData.port,
+ configData.tx_enable,
+ configData.rx_enable);
+ }
+ break;
+ case ESW_SET_BROADCAST_CONF:
+ {
+ eswIoctlPortConfig configData;
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_broadcast_config(fep,
+ configData.port, configData.enable);
+ }
+ break;
+
+ case ESW_SET_MULTICAST_CONF:
+ {
+ eswIoctlPortConfig configData;
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_multicast_config(fep,
+ configData.port, configData.enable);
+ }
+ break;
+
+ case ESW_SET_BLOCKING_CONF:
+ {
+ eswIoctlPortConfig configData;
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortConfig));
+
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_blocking_config(fep,
+ configData.port, configData.enable);
+ }
+ break;
+
+ case ESW_SET_LEARNING_CONF:
+ {
+ eswIoctlPortConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_learning_config(fep,
+ configData.port, configData.enable);
+ }
+ break;
+
+ case ESW_SET_PORT_ENTRY_EMPTY:
+ {
+ unsigned long portnum;
+
+ ret = copy_from_user(&portnum,
+ ifr->ifr_data, sizeof(portnum));
+ if (ret)
+ return -EFAULT;
+ esw_atable_dynamicms_del_entries_for_port(fep, portnum);
+ }
+ break;
+
+ case ESW_SET_OTHER_PORT_ENTRY_EMPTY:
+ {
+ unsigned long portnum;
+
+ ret = copy_from_user(&portnum,
+ ifr->ifr_data, sizeof(portnum));
+ if (ret)
+ return -EFAULT;
+
+ esw_atable_dynamicms_del_entries_for_other_port(fep, portnum);
+ }
+ break;
+
+ case ESW_SET_IP_SNOOP_CONF:
+ {
+ eswIoctlIpsnoopConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlIpsnoopConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_ip_snoop_config(fep, configData.mode,
+ configData.ip_header_protocol);
+ }
+ break;
+
+ case ESW_SET_PORT_SNOOP_CONF:
+ {
+ eswIoctlPortsnoopConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortsnoopConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_tcpudp_port_snoop_config(fep, configData.mode,
+ configData.compare_port,
+ configData.compare_num);
+ }
+ break;
+
+ case ESW_SET_PORT_MIRROR_CONF_PORT_MATCH:
+ {
+ struct eswIoctlMirrorCfgPortMatch configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(configData));
+ if (ret)
+ return -EFAULT;
+ ret = esw_port_mirroring_config_port_match(fep,
+ configData.mirror_port, configData.port_match_en,
+ configData.port);
+ }
+ break;
+
+ case ESW_SET_PORT_MIRROR_CONF:
+ {
+ eswIoctlPortMirrorConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPortMirrorConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_mirroring_config(fep,
+ configData.mirror_port, configData.port,
+ configData.mirror_enable,
+ configData.src_mac, configData.des_mac,
+ configData.egress_en, configData.ingress_en,
+ configData.egress_mac_src_en,
+ configData.egress_mac_des_en,
+ configData.ingress_mac_src_en,
+ configData.ingress_mac_des_en);
+ }
+ break;
+
+ case ESW_SET_PORT_MIRROR_CONF_ADDR_MATCH:
+ {
+ struct eswIoctlMirrorCfgAddrMatch configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(configData));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_port_mirroring_config_addr_match(fep,
+ configData.mirror_port, configData.addr_match_en,
+ configData.mac_addr);
+ }
+ break;
+
+ case ESW_SET_PIRORITY_VLAN:
+ {
+ eswIoctlPriorityVlanConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPriorityVlanConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_framecalssify_vlan_priority_lookup(fep,
+ configData.port, configData.func_enable,
+ configData.vlan_pri_table_num,
+ configData.vlan_pri_table_value);
+ }
+ break;
+
+ case ESW_SET_PIRORITY_IP:
+ {
+ eswIoctlPriorityIPConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPriorityIPConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_framecalssify_ip_priority_lookup(fep,
+ configData.port, configData.func_enable,
+ configData.ipv4_en, configData.ip_priority_num,
+ configData.ip_priority_value);
+ }
+ break;
+
+ case ESW_SET_PIRORITY_MAC:
+ {
+ eswIoctlPriorityMacConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPriorityMacConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_framecalssify_mac_priority_lookup(fep,
+ configData.port);
+ }
+ break;
+
+ case ESW_SET_PIRORITY_DEFAULT:
+ {
+ eswIoctlPriorityDefaultConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlPriorityDefaultConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_frame_calssify_priority_init(fep,
+ configData.port, configData.priority_value);
+ }
+ break;
+
+ case ESW_SET_P0_FORCED_FORWARD:
+ {
+ eswIoctlP0ForcedForwardConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlP0ForcedForwardConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_forced_forward(fep, configData.port1,
+ configData.port2, configData.enable);
+ }
+ break;
+
+ case ESW_SET_BRIDGE_CONFIG:
+ {
+ unsigned long configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+
+ esw_bridge_port_configure(fep, configData);
+ }
+ break;
+
+ case ESW_SET_SWITCH_MODE:
+ {
+ unsigned long configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+
+ esw_switch_mode_configure(fep, configData);
+ }
+ break;
+
+ case ESW_SET_OUTPUT_QUEUE_MEMORY:
+ {
+ eswIoctlOutputQueue configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlOutputQueue));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_output_queue_memory(fep,
+ configData.fun_num, &configData.sOutputQueue);
+ }
+ break;
+
+ case ESW_SET_VLAN_OUTPUT_PROCESS:
+ {
+ eswIoctlVlanOutputConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlVlanOutputConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_vlan_output_process(fep,
+ configData.port, configData.mode);
+ }
+ break;
+
+ case ESW_SET_VLAN_INPUT_PROCESS:
+ {
+ eswIoctlVlanInputConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data,
+ sizeof(eswIoctlVlanInputConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_vlan_input_process(fep, configData.port,
+ configData.mode, configData.port_vlanid);
+ }
+ break;
+
+ case ESW_SET_VLAN_DOMAIN_VERIFICATION:
+ {
+ eswIoctlVlanVerificationConfig configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data,
+ sizeof(eswIoctlVlanVerificationConfig));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_vlan_verification(
+ fep, configData.port,
+ configData.vlan_domain_verify_en,
+ configData.vlan_discard_unknown_en);
+ }
+ break;
+
+ case ESW_SET_VLAN_RESOLUTION_TABLE:
+ {
+ eswIoctlVlanResoultionTable configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data,
+ sizeof(eswIoctlVlanResoultionTable));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_vlan_resolution_table(
+ fep, configData.port_vlanid,
+ configData.vlan_domain_num,
+ configData.vlan_domain_port);
+
+ }
+ break;
+
+ case ESW_SET_VLAN_ID:
+ {
+ unsigned long configData;
+ ret = copy_from_user(&configData, ifr->ifr_data,
+ sizeof(configData));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_vlan_id(fep, configData);
+ }
+ break;
+
+ case ESW_SET_VLAN_ID_CLEARED:
+ {
+ unsigned long configData;
+ ret = copy_from_user(&configData, ifr->ifr_data,
+ sizeof(configData));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_vlan_id_cleared(fep, configData);
+ }
+ break;
+
+ case ESW_SET_PORT_IN_VLAN_ID:
+ {
+ eswIoctlVlanResoultionTable configData;
+
+ ret = copy_from_user(&configData, ifr->ifr_data,
+ sizeof(configData));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_set_port_in_vlan_id(fep, configData);
+ }
+ break;
+
+ case ESW_UPDATE_STATIC_MACTABLE:
+ {
+ eswIoctlUpdateStaticMACtable configData;
+
+ ret = copy_from_user(&configData,
+ ifr->ifr_data, sizeof(eswIoctlUpdateStaticMACtable));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_update_atable_static(configData.mac_addr,
+ configData.port, configData.priority, fep);
+ }
+ break;
+
+ case ESW_CLEAR_ALL_MACTABLE:
+ {
+ esw_clear_atable(fep);
+ }
+ break;
+
+ /*get*/
+ case ESW_GET_STATISTICS_STATUS:
+ {
+ esw_statistics_status Statistics;
+ esw_port_statistics_status PortSta;
+ int i;
+
+ ret = esw_get_statistics_status(fep, &Statistics);
+ if (ret != 0) {
+ printk(KERN_ERR "%s: cmd %x fail\n", __func__, cmd);
+ return -1;
+ }
+ printk(KERN_INFO "DISCN : %10ld DISCB : %10ld\n",
+ Statistics.ESW_DISCN, Statistics.ESW_DISCB);
+ printk(KERN_INFO "NDISCN: %10ld NDISCB: %10ld\n",
+ Statistics.ESW_NDISCN, Statistics.ESW_NDISCB);
+
+ for (i = 0; i < 3; i++) {
+ ret = esw_get_port_statistics_status(fep, i,
+ &PortSta);
+ if (ret != 0) {
+ printk(KERN_ERR "%s: cmd %x fail\n",
+ __func__, cmd);
+ return -1;
+ }
+ printk(KERN_INFO "port %d: POQC : %ld\n",
+ i, PortSta.FSL_ESW_POQC);
+ printk(KERN_INFO " PMVID : %ld\n",
+ PortSta.FSL_ESW_PMVID);
+ printk(KERN_INFO " PMVTAG: %ld\n",
+ PortSta.FSL_ESW_PMVTAG);
+ printk(KERN_INFO " PBL : %ld\n",
+ PortSta.FSL_ESW_PBL);
+ }
+ }
+ break;
+
+ case ESW_GET_LEARNING_CONF:
+ {
+ unsigned long PortLearning;
+
+ esw_get_port_learning(fep, &PortLearning);
+ ret = copy_to_user(ifr->ifr_data, &PortLearning,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_BLOCKING_CONF:
+ {
+ unsigned long PortBlocking;
+
+ esw_get_port_blocking(fep, &PortBlocking);
+ ret = copy_to_user(ifr->ifr_data, &PortBlocking,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_MULTICAST_CONF:
+ {
+ unsigned long PortMulticast;
+
+ esw_get_port_multicast(fep, &PortMulticast);
+ ret = copy_to_user(ifr->ifr_data, &PortMulticast,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_BROADCAST_CONF:
+ {
+ unsigned long PortBroadcast;
+
+ esw_get_port_broadcast(fep, &PortBroadcast);
+ ret = copy_to_user(ifr->ifr_data, &PortBroadcast,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_PORTENABLE_CONF:
+ {
+ unsigned long PortEnable;
+
+ esw_get_port_enable(fep, &PortEnable);
+ ret = copy_to_user(ifr->ifr_data, &PortEnable,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_IP_SNOOP_CONF:
+ {
+ unsigned long ESW_IPSNP[8];
+ int i;
+
+ esw_get_ip_snoop_config(fep, (unsigned long *)ESW_IPSNP);
+ printk(KERN_INFO "IP Protocol Mode Type\n");
+ for (i = 0; i < 8; i++) {
+ if (ESW_IPSNP[i] != 0)
+ printk(KERN_INFO "%3ld "
+ "%1ld %s\n",
+ (ESW_IPSNP[i] >> 8) & 0xff,
+ (ESW_IPSNP[i] >> 1) & 3,
+ ESW_IPSNP[i] & 1 ? "Active" :
+ "Inactive");
+ }
+ }
+ break;
+
+ case ESW_GET_PORT_SNOOP_CONF:
+ {
+ unsigned long ESW_PSNP[8];
+ int i;
+
+ esw_get_tcpudp_port_snoop_config(fep,
+ (unsigned long *)ESW_PSNP);
+ printk(KERN_INFO "TCP/UDP Port SrcCompare DesCompare "
+ "Mode Type\n");
+ for (i = 0; i < 8; i++) {
+ if (ESW_PSNP[i] != 0)
+ printk(KERN_INFO "%5ld %s "
+ "%s %1ld %s\n",
+ (ESW_PSNP[i] >> 16) & 0xffff,
+ (ESW_PSNP[i] >> 4) & 1 ? "Y" : "N",
+ (ESW_PSNP[i] >> 3) & 1 ? "Y" : "N",
+ (ESW_PSNP[i] >> 1) & 3,
+ ESW_PSNP[i] & 1 ? "Active" :
+ "Inactive");
+ }
+ }
+ break;
+
+ case ESW_GET_PORT_MIRROR_CONF:
+ esw_get_port_mirroring(fep);
+ break;
+
+ case ESW_GET_P0_FORCED_FORWARD:
+ {
+ unsigned long ForceForward;
+
+ esw_get_forced_forward(fep, &ForceForward);
+ ret = copy_to_user(ifr->ifr_data, &ForceForward,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_SWITCH_MODE:
+ {
+ unsigned long Config;
+
+ esw_get_switch_mode(fep, &Config);
+ ret = copy_to_user(ifr->ifr_data, &Config,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_BRIDGE_CONFIG:
+ {
+ unsigned long Config;
+
+ esw_get_bridge_port(fep, &Config);
+ ret = copy_to_user(ifr->ifr_data, &Config,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+ case ESW_GET_OUTPUT_QUEUE_STATUS:
+ {
+ esw_output_queue_status Config;
+ esw_get_output_queue_status(fep,
+ &Config);
+ ret = copy_to_user(ifr->ifr_data, &Config,
+ sizeof(esw_output_queue_status));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_VLAN_OUTPUT_PROCESS:
+ {
+ unsigned long Config;
+ int tmp;
+ int i;
+
+ esw_get_vlan_output_config(fep, &Config);
+
+ for (i = 0; i < 3; i++) {
+ tmp = (Config >> (i << 1)) & 3;
+
+ if (tmp != 0)
+ printk(KERN_INFO "port %d: vlan output "
+ "manipulation enable (mode %d)\n",
+ i, tmp);
+ else
+ printk(KERN_INFO "port %d: vlan output "
+ "manipulation disable\n", i);
+ }
+ }
+ break;
+
+ case ESW_GET_VLAN_INPUT_PROCESS:
+ {
+ eswIoctlVlanInputStatus Config;
+ int i;
+
+ esw_get_vlan_input_config(fep, &Config);
+
+ for (i = 0; i < 3; i++) {
+ if (((Config.ESW_VIMEN >> i) & 1) == 0)
+ printk(KERN_INFO "port %d: vlan input "
+ "manipulation disable\n", i);
+ else
+ printk("port %d: vlan input manipulation enable"
+ " (mode %ld, vlan id %ld)\n", i,
+ (((Config.ESW_VIMSEL >> (i << 1)) & 3)
+ + 1), Config.ESW_PID[i]);
+ }
+ }
+ break;
+
+ case ESW_GET_VLAN_RESOLUTION_TABLE:
+ {
+ struct eswVlanTableItem vtableitem;
+ unsigned char tmp0, tmp1, tmp2;
+ int i;
+
+ esw_get_vlan_resolution_table(fep, &vtableitem);
+
+ printk(KERN_INFO "VLAN Name VLAN Id Ports\n");
+ for (i = 0; i < vtableitem.valid_num; i++) {
+ tmp0 = vtableitem.table[i].vlan_domain_port & 1;
+ tmp1 = (vtableitem.table[i].vlan_domain_port >> 1) & 1;
+ tmp2 = (vtableitem.table[i].vlan_domain_port >> 2) & 1;
+ printk(KERN_INFO "%2d %4d %s%s%s\n",
+ i, vtableitem.table[i].port_vlanid,
+ tmp0 ? "0 " : "", tmp1 ? "1 " : "",
+ tmp2 ? "2" : "");
+ }
+ }
+ break;
+
+ case ESW_GET_VLAN_DOMAIN_VERIFICATION:
+ {
+ unsigned long Config;
+
+ esw_get_vlan_verification(fep, &Config);
+ ret = copy_to_user(ifr->ifr_data, &Config,
+ sizeof(unsigned long));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_ENTRY_PORT_NUMBER:
+ {
+ unsigned char mac_addr[6];
+ unsigned char portnum;
+
+ ret = copy_from_user(mac_addr,
+ ifr->ifr_data, sizeof(mac_addr));
+ if (ret)
+ return -EFAULT;
+
+ ret = esw_atable_get_entry_port_number(fep, mac_addr,
+ &portnum);
+
+ ret = copy_to_user(ifr->ifr_data, &portnum,
+ sizeof(unsigned char));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_LOOKUP_TABLE:
+ {
+ unsigned long *ConfigData;
+ unsigned long dennum, sennum;
+ int i;
+ int tmp;
+
+ ConfigData = kmalloc(sizeof(struct eswAddrTableEntryExample) *
+ ESW_ATABLE_MEM_NUM_ENTRIES, GFP_KERNEL);
+ ret = esw_get_mac_address_lookup_table(fep, ConfigData,
+ &dennum, &sennum);
+ printk(KERN_INFO "Dynamic entries number: %ld\n", dennum);
+ printk(KERN_INFO "Static entries number: %ld\n", sennum);
+ printk(KERN_INFO "Type MAC address Port Timestamp\n");
+ for (i = 0; i < dennum; i++) {
+ printk(KERN_INFO "dynamic "
+ "%02lx-%02lx-%02lx-%02lx-%02lx-%02lx "
+ "%01lx %4ld\n", *(ConfigData + i * 11 + 2),
+ *(ConfigData + i * 11 + 3),
+ *(ConfigData + i * 11 + 4),
+ *(ConfigData + i * 11 + 5),
+ *(ConfigData + i * 11 + 6),
+ *(ConfigData + i * 11 + 7),
+ *(ConfigData + i * 11 + 8),
+ *(ConfigData + i * 11 + 9));
+ }
+
+ if (sennum != 0)
+ printk(KERN_INFO "Type MAC address"
+ " Port Priority\n");
+
+ for (i = 0; i < sennum; i++) {
+ printk(KERN_INFO "static %02lx-%02lx-%02lx-%02lx"
+ "-%02lx-%02lx ",
+ *(ConfigData + (2047 - i) * 11 + 2),
+ *(ConfigData + (2047 - i) * 11 + 3),
+ *(ConfigData + (2047 - i) * 11 + 4),
+ *(ConfigData + (2047 - i) * 11 + 5),
+ *(ConfigData + (2047 - i) * 11 + 6),
+ *(ConfigData + (2047 - i) * 11 + 7));
+
+ tmp = *(ConfigData + (2047 - i) * 11 + 8);
+ if ((tmp == 0) || (tmp == 2) || (tmp == 4))
+ printk("%01x ", tmp >> 1);
+ else if (tmp == 3)
+ printk("0,1 ");
+ else if (tmp == 5)
+ printk("0,2 ");
+ else if (tmp == 6)
+ printk("1,2 ");
+
+ printk("%4ld\n", *(ConfigData + (2047 - i) * 11 + 9));
+ }
+ kfree(ConfigData);
+ }
+ break;
+
+ case ESW_GET_PORT_STATUS:
+ {
+ unsigned long PortBlocking;
+
+ esw_get_port_blocking(fep, &PortBlocking);
+
+ ports_link_status.port0_block_status = PortBlocking & 1;
+ ports_link_status.port1_block_status = (PortBlocking >> 1) & 1;
+ ports_link_status.port2_block_status = PortBlocking >> 2;
+
+ ret = copy_to_user(ifr->ifr_data, &ports_link_status,
+ sizeof(ports_link_status));
+ if (ret)
+ return -EFAULT;
+ }
+ break;
+
+ case ESW_GET_PORT_ALL_STATUS:
+ {
+ unsigned char portnum;
+ struct port_all_status port_astatus;
+
+ ret = copy_from_user(&portnum,
+ ifr->ifr_data, sizeof(portnum));
+ if (ret)
+ return -EFAULT;
+
+ esw_get_port_all_status(fep, portnum, &port_astatus);
+ printk(KERN_INFO "Port %d status:\n", portnum);
+ printk(KERN_INFO "Link:%-4s Blocking:%1s "
+ "Learning:%1s\n",
+ port_astatus.link_status ? "Up" : "Down",
+ port_astatus.block_status ? "Y" : "N",
+ port_astatus.learn_status ? "N" : "Y");
+ printk(KERN_INFO "VLAN Verify:%1s Discard Unknown:%1s "
+ "Multicast Res:%1s\n",
+ port_astatus.vlan_verify ? "Y" : "N",
+ port_astatus.discard_unknown ? "Y" : "N",
+ port_astatus.multi_reso ? "Y" : "N");
+ printk(KERN_INFO "Broadcast Res:%1s Transmit:%-7s "
+ "Receive:%7s\n",
+ port_astatus.broad_reso ? "Y" : "N",
+ port_astatus.ftransmit ? "Enable" : "Disable",
+ port_astatus.freceive ? "Enable" : "Disable");
+
+ }
+ break;
+
+ case ESW_GET_USER_PID:
+ {
+ long get_pid = 0;
+ ret = copy_from_user(&get_pid,
+ ifr->ifr_data, sizeof(get_pid));
+
+ if (ret)
+ return -EFAULT;
+ user_pid = get_pid;
+ }
+ break;
+
+ default:
+ return -EOPNOTSUPP;
+ }
+
+ return ret;
+}
+
+static netdev_tx_t switch_enet_start_xmit(struct sk_buff *skb,
+ struct net_device *dev)
+{
+ struct switch_enet_private *fep;
+ cbd_t *bdp;
+ unsigned short status;
+ unsigned long flags;
+ void *bufaddr;
+
+ fep = netdev_priv(dev);
+
+ spin_lock_irqsave(&fep->hw_lock, flags);
+ /* Fill in a Tx ring entry */
+ bdp = fep->cur_tx;
+
+ status = bdp->cbd_sc;
+
+ /* Clear all of the status flags */
+ status &= ~BD_ENET_TX_STATS;
+
+ /* Set buffer length and buffer pointer.
+ */
+ bufaddr = skb->data;
+ bdp->cbd_datlen = skb->len;
+
+ /*
+ * On some FEC implementations data must be aligned on
+ * 4-byte boundaries. Use bounce buffers to copy data
+ * and get it aligned. Ugh.
+ */
+ if (((unsigned long)bufaddr) & 0xf) {
+ unsigned int index;
+ index = bdp - fep->tx_bd_base;
+ memcpy(fep->tx_bounce[index],
+ (void *)skb->data, bdp->cbd_datlen);
+ bufaddr = fep->tx_bounce[index];
+ }
+
+ /* Save skb pointer. */
+ fep->tx_skbuff[fep->skb_cur] = skb;
+
+ dev->stats.tx_bytes += skb->len;
+ fep->skb_cur = (fep->skb_cur+1) & TX_RING_MOD_MASK;
+
+ /* Push the data cache so the CPM does not get stale memory
+ * data.
+ */
+ bdp->cbd_bufaddr = dma_map_single(&fep->pdev->dev, bufaddr,
+ SWITCH_ENET_TX_FRSIZE, DMA_TO_DEVICE);
+
+ /* Send it on its way. Tell FEC it's ready, interrupt when done,
+ * it's the last BD of the frame, and to put the CRC on the end.
+ */
+ status |= (BD_ENET_TX_READY | BD_ENET_TX_INTR
+ | BD_ENET_TX_LAST | BD_ENET_TX_TC);
+ bdp->cbd_sc = status;
+
+ dev->trans_start = jiffies;
+
+ /* Trigger transmission start */
+ writel(FSL_ESW_TDAR_X_DES_ACTIVE, fep->membase + FEC_ESW_TDAR);
+
+ /* If this was the last BD in the ring,
+ * start at the beginning again.*/
+ if (status & BD_ENET_TX_WRAP)
+ bdp = fep->tx_bd_base;
+ else
+ bdp++;
+
+ if (bdp == fep->dirty_tx) {
+ fep->tx_full = 1;
+ netif_stop_queue(dev);
+ printk(KERN_ERR "%s: net stop\n", __func__);
+ }
+
+ fep->cur_tx = (cbd_t *)bdp;
+
+ spin_unlock_irqrestore(&fep->hw_lock, flags);
+
+ return NETDEV_TX_OK;
+}
+
+static void switch_timeout(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+
+ printk(KERN_ERR "%s: transmit timed out.\n", dev->name);
+ dev->stats.tx_errors++;
+ switch_restart(dev, fep->full_duplex);
+ netif_wake_queue(dev);
+}
+
+/* The interrupt handler.
+ * This is called from the MPC core interrupt.
+ */
+static irqreturn_t switch_enet_interrupt(int irq, void *dev_id)
+{
+ struct net_device *dev = dev_id;
+ struct switch_enet_private *fep = netdev_priv(dev);
+ uint int_events;
+ irqreturn_t ret = IRQ_NONE;
+
+ /* Get the interrupt events that caused us to be here.
+ */
+ do {
+ int_events = readl(fep->membase + FEC_ESW_ISR);
+ writel(int_events, fep->membase + FEC_ESW_ISR);
+
+ /* Handle receive event in its own function. */
+
+ if (int_events & FSL_ESW_ISR_RXF) {
+ ret = IRQ_HANDLED;
+ switch_enet_rx(dev);
+ }
+
+ if (int_events & FSL_ESW_ISR_TXF) {
+ ret = IRQ_HANDLED;
+ switch_enet_tx(dev);
+ }
+
+ } while (int_events);
+
+ return ret;
+}
+
+static void switch_enet_tx(struct net_device *dev)
+{
+ struct switch_enet_private *fep;
+ struct bufdesc *bdp;
+ unsigned short status;
+ struct sk_buff *skb;
+ unsigned long flags;
+
+ fep = netdev_priv(dev);
+ spin_lock_irqsave(&fep->hw_lock, flags);
+ bdp = fep->dirty_tx;
+
+ while (((status = bdp->cbd_sc) & BD_ENET_TX_READY) == 0) {
+ if (bdp == fep->cur_tx && fep->tx_full == 0)
+ break;
+
+ if (bdp->cbd_bufaddr)
+ dma_unmap_single(&fep->pdev->dev, bdp->cbd_bufaddr,
+ SWITCH_ENET_TX_FRSIZE, DMA_TO_DEVICE);
+ bdp->cbd_bufaddr = 0;
+
+ skb = fep->tx_skbuff[fep->skb_dirty];
+ /* Check for errors. */
+ if (status & (BD_ENET_TX_HB | BD_ENET_TX_LC |
+ BD_ENET_TX_RL | BD_ENET_TX_UN |
+ BD_ENET_TX_CSL)) {
+ dev->stats.tx_errors++;
+ if (status & BD_ENET_TX_HB) /* No heartbeat */
+ dev->stats.tx_heartbeat_errors++;
+ if (status & BD_ENET_TX_LC) /* Late collision */
+ dev->stats.tx_window_errors++;
+ if (status & BD_ENET_TX_RL) /* Retrans limit */
+ dev->stats.tx_aborted_errors++;
+ if (status & BD_ENET_TX_UN) /* Underrun */
+ dev->stats.tx_fifo_errors++;
+ if (status & BD_ENET_TX_CSL) /* Carrier lost */
+ dev->stats.tx_carrier_errors++;
+ } else {
+ dev->stats.tx_packets++;
+ }
+
+ /* Deferred means some collisions occurred during transmit,
+ * but we eventually sent the packet OK.
+ */
+ if (status & BD_ENET_TX_DEF)
+ dev->stats.collisions++;
+
+ /* Free the sk buffer associated with this last transmit.
+ */
+ dev_kfree_skb_any(skb);
+ fep->tx_skbuff[fep->skb_dirty] = NULL;
+ fep->skb_dirty = (fep->skb_dirty + 1) & TX_RING_MOD_MASK;
+
+ /* Update pointer to next buffer descriptor to be transmitted.
+ */
+ if (status & BD_ENET_TX_WRAP)
+ bdp = fep->tx_bd_base;
+ else
+ bdp++;
+
+ /* Since we have freed up a buffer, the ring is no longer
+ * full.
+ */
+ if (fep->tx_full) {
+ fep->tx_full = 0;
+ printk(KERN_ERR "%s: tx full is zero\n", __func__);
+ if (netif_queue_stopped(dev))
+ netif_wake_queue(dev);
+ }
+ }
+ fep->dirty_tx = (cbd_t *)bdp;
+ spin_unlock_irqrestore(&fep->hw_lock, flags);
+}
+
+
+/* During a receive, the cur_rx points to the current incoming buffer.
+ * When we update through the ring, if the next incoming buffer has
+ * not been given to the system, we just set the empty indicator,
+ * effectively tossing the packet.
+ */
+static void switch_enet_rx(struct net_device *dev)
+{
+ struct switch_enet_private *fep;
+ cbd_t *bdp;
+ unsigned short status;
+ struct sk_buff *skb;
+ ushort pkt_len;
+ __u8 *data;
+ unsigned long flags;
+
+ fep = netdev_priv(dev);
+
+ spin_lock_irqsave(&fep->hw_lock, flags);
+ /* First, grab all of the stats for the incoming packet.
+ * These get messed up if we get called due to a busy condition.
+ */
+ bdp = fep->cur_rx;
+
+ while (!((status = bdp->cbd_sc) & BD_ENET_RX_EMPTY)) {
+
+ /* Since we have allocated space to hold a complete frame,
+ * the last indicator should be set.
+ * */
+ if ((status & BD_ENET_RX_LAST) == 0)
+ printk(KERN_ERR "SWITCH ENET: rcv is not +last\n");
+
+ if (!fep->opened)
+ goto rx_processing_done;
+
+ /* Check for errors. */
+ if (status & (BD_ENET_RX_LG | BD_ENET_RX_SH | BD_ENET_RX_NO |
+ BD_ENET_RX_CR | BD_ENET_RX_OV)) {
+ dev->stats.rx_errors++;
+ if (status & (BD_ENET_RX_LG | BD_ENET_RX_SH)) {
+ /* Frame too long or too short. */
+ dev->stats.rx_length_errors++;
+ }
+ if (status & BD_ENET_RX_NO) /* Frame alignment */
+ dev->stats.rx_frame_errors++;
+ if (status & BD_ENET_RX_CR) /* CRC Error */
+ dev->stats.rx_crc_errors++;
+ if (status & BD_ENET_RX_OV) /* FIFO overrun */
+ dev->stats.rx_fifo_errors++;
+ }
+ /* Report late collisions as a frame error.
+ * On this error, the BD is closed, but we don't know what we
+ * have in the buffer. So, just drop this frame on the floor.
+ * */
+ if (status & BD_ENET_RX_CL) {
+ dev->stats.rx_errors++;
+ dev->stats.rx_frame_errors++;
+ goto rx_processing_done;
+ }
+ /* Process the incoming frame */
+ dev->stats.rx_packets++;
+ pkt_len = bdp->cbd_datlen;
+ dev->stats.rx_bytes += pkt_len;
+ data = (__u8 *)__va(bdp->cbd_bufaddr);
+
+ if (bdp->cbd_bufaddr)
+ dma_unmap_single(&fep->pdev->dev, bdp->cbd_bufaddr,
+ SWITCH_ENET_TX_FRSIZE, DMA_FROM_DEVICE);
+
+ /* This does 16 byte alignment, exactly what we need.
+ * The packet length includes FCS, but we don't want to
+ * include that when passing upstream as it messes up
+ * bridging applications.
+ * */
+ skb = dev_alloc_skb(pkt_len - 4 + NET_IP_ALIGN);
+
+ if (skb == NULL)
+ dev->stats.rx_dropped++;
+
+ if (unlikely(!skb)) {
+ printk(KERN_ERR "%s: Memory squeeze, dropping packet.\n",
+ dev->name);
+ dev->stats.rx_dropped++;
+ } else {
+ skb_reserve(skb, NET_IP_ALIGN);
+ skb_put(skb, pkt_len - 4); /* Make room */
+ skb_copy_to_linear_data(skb, data, pkt_len - 4);
+ skb->protocol = eth_type_trans(skb, dev);
+ netif_rx(skb);
+ }
+
+ bdp->cbd_bufaddr = dma_map_single(&fep->pdev->dev, data,
+ SWITCH_ENET_TX_FRSIZE, DMA_FROM_DEVICE);
+rx_processing_done:
+
+ /* Clear the status flags for this buffer */
+ status &= ~BD_ENET_RX_STATS;
+
+ /* Mark the buffer empty */
+ status |= BD_ENET_RX_EMPTY;
+ bdp->cbd_sc = status;
+
+ /* Update BD pointer to next entry */
+ if (status & BD_ENET_RX_WRAP)
+ bdp = fep->rx_bd_base;
+ else
+ bdp++;
+
+ /* Doing this here will keep the FEC running while we process
+ * incoming frames. On a heavily loaded network, we should be
+ * able to keep up at the expense of system resources.
+ * */
+ writel(FSL_ESW_RDAR_R_DES_ACTIVE, fep->membase + FEC_ESW_RDAR);
+ }
+ fep->cur_rx = (cbd_t *)bdp;
+
+ spin_unlock_irqrestore(&fep->hw_lock, flags);
+}
+
+static int fec_mdio_transfer(struct mii_bus *bus, int phy_id,
+ int reg, int regval)
+{
+ unsigned long flags;
+ struct switch_enet_private *fep = bus->priv;
+ int retval = 0;
+ int tries = 100;
+
+ spin_lock_irqsave(&fep->mii_lock, flags);
+
+ fep->mii_timeout = 0;
+ init_completion(&fep->mdio_done);
+
+ regval |= phy_id << 23;
+ writel(regval, fep->enetbase + FSL_FEC_MMFR0);
+
+ /* wait for it to finish, this takes about 23 us on lite5200b */
+ while (!(readl(fep->enetbase + FSL_FEC_EIR0) & FEC_ENET_MII) && --tries)
+ udelay(5);
+ if (!tries) {
+ printk(KERN_ERR "%s timeout\n", __func__);
+ return -ETIMEDOUT;
+ }
+
+ writel(FEC_ENET_MII, fep->enetbase + FSL_FEC_EIR0);
+ retval = (readl(fep->enetbase + FSL_FEC_MMFR0) & 0xffff);
+ spin_unlock_irqrestore(&fep->mii_lock, flags);
+
+ return retval;
+}
+
+static int fec_enet_mdio_read(struct mii_bus *bus,
+ int phy_id, int reg)
+{
+ int ret;
+ ret = fec_mdio_transfer(bus, phy_id, reg,
+ mk_mii_read(reg));
+ return ret;
+}
+
+static int fec_enet_mdio_write(struct mii_bus *bus,
+ int phy_id, int reg, u16 data)
+{
+ return fec_mdio_transfer(bus, phy_id, reg,
+ mk_mii_write(reg, data));
+}
+
+static void switch_adjust_link1(struct net_device *dev)
+{
+ struct switch_enet_private *priv = netdev_priv(dev);
+ struct phy_device *phydev1 = priv->phydev[0];
+ int new_state = 0;
+
+ if (phydev1->link != PHY_DOWN) {
+ if (phydev1->duplex != priv->phy1_duplex) {
+ new_state = 1;
+ priv->phy1_duplex = phydev1->duplex;
+ }
+
+ if (phydev1->speed != priv->phy1_speed) {
+ new_state = 1;
+ priv->phy1_speed = phydev1->speed;
+ }
+
+ if (priv->phy1_old_link == PHY_DOWN) {
+ new_state = 1;
+ priv->phy1_old_link = phydev1->link;
+ }
+ } else if (priv->phy1_old_link) {
+ new_state = 1;
+ priv->phy1_old_link = PHY_DOWN;
+ priv->phy1_speed = 0;
+ priv->phy1_duplex = -1;
+ }
+
+ if (new_state) {
+ ports_link_status.port1_link_status = phydev1->link;
+ if (phydev1->link == PHY_DOWN)
+ esw_atable_dynamicms_del_entries_for_port(priv, 1);
+
+ /*Send the new status to user space*/
+ if (user_pid != 1)
+ sys_tkill(user_pid, SIGUSR1);
+ phy_print_status(phydev1);
+ }
+}
+
+static void switch_adjust_link2(struct net_device *dev)
+{
+ struct switch_enet_private *priv = netdev_priv(dev);
+ struct phy_device *phydev2 = priv->phydev[1];
+ int new_state = 0;
+
+ if (phydev2->link != PHY_DOWN) {
+ if (phydev2->duplex != priv->phy2_duplex) {
+ new_state = 1;
+ priv->phy2_duplex = phydev2->duplex;
+ }
+
+ if (phydev2->speed != priv->phy2_speed) {
+ new_state = 1;
+ priv->phy2_speed = phydev2->speed;
+ }
+
+ if (priv->phy2_old_link == PHY_DOWN) {
+ new_state = 1;
+ priv->phy2_old_link = phydev2->link;
+ }
+ } else if (priv->phy2_old_link) {
+ new_state = 1;
+ priv->phy2_old_link = PHY_DOWN;
+ priv->phy2_speed = 0;
+ priv->phy2_duplex = -1;
+ }
+
+ if (new_state) {
+ ports_link_status.port2_link_status = phydev2->link;
+ if (phydev2->link == PHY_DOWN)
+ esw_atable_dynamicms_del_entries_for_port(priv, 2);
+
+ /*Send the new status to user space*/
+ if (user_pid != 1)
+ sys_tkill(user_pid, SIGUSR1);
+ phy_print_status(phydev2);
+ }
+}
+
+static int switch_init_phy(struct net_device *dev)
+{
+ struct switch_enet_private *priv = netdev_priv(dev);
+ struct phy_device *phydev[SWITCH_EPORT_NUMBER] = {NULL, NULL};
+ int i, j = 0;
+
+ /* search for connect PHY device */
+ for (i = 0; i < PHY_MAX_ADDR; i++) {
+ struct phy_device *const tmp_phydev =
+ priv->mdio_bus->phy_map[i];
+
+ if (!tmp_phydev)
+ continue;
+
+ phydev[j++] = tmp_phydev;
+ if (j >= SWITCH_EPORT_NUMBER)
+ break;
+ }
+
+ /* now we are supposed to have a proper phydev, to attach to... */
+ if ((!phydev[0]) && (!phydev[1])) {
+ printk(KERN_INFO "%s: Don't found any phy device at all\n",
+ dev->name);
+ return -ENODEV;
+ }
+
+ priv->phy1_link = PHY_DOWN;
+ priv->phy1_old_link = PHY_DOWN;
+ priv->phy1_speed = 0;
+ priv->phy1_duplex = -1;
+
+ priv->phy2_link = PHY_DOWN;
+ priv->phy2_old_link = PHY_DOWN;
+ priv->phy2_speed = 0;
+ priv->phy2_duplex = -1;
+
+ phydev[0] = phy_connect(dev, dev_name(&phydev[0]->dev),
+ &switch_adjust_link1, 0, PHY_INTERFACE_MODE_RMII);
+ if (IS_ERR(phydev[0])) {
+ printk(KERN_ERR " %s phy_connect failed\n", __func__);
+ return PTR_ERR(phydev[0]);
+ }
+
+ phydev[1] = phy_connect(dev, dev_name(&phydev[1]->dev),
+ &switch_adjust_link2, 0, PHY_INTERFACE_MODE_RMII);
+ if (IS_ERR(phydev[1])) {
+ printk(KERN_ERR " %s phy_connect failed\n", __func__);
+ return PTR_ERR(phydev[1]);
+ }
+
+ for (i = 0; i < SWITCH_EPORT_NUMBER; i++) {
+ phydev[i]->supported &= PHY_BASIC_FEATURES;
+ phydev[i]->advertising = phydev[i]->supported;
+ priv->phydev[i] = phydev[i];
+ printk(KERN_INFO "attached phy %i to driver %s "
+ "(mii_bus:phy_addr=%s, irq=%d)\n",
+ phydev[i]->addr, phydev[i]->drv->name,
+ dev_name(&priv->phydev[i]->dev),
+ priv->phydev[i]->irq);
+ }
+
+ return 0;
+}
+
+static void switch_enet_free_buffers(struct net_device *ndev)
+{
+ struct switch_enet_private *fep = netdev_priv(ndev);
+ int i;
+ struct sk_buff *skb;
+ cbd_t *bdp;
+
+ bdp = fep->rx_bd_base;
+ for (i = 0; i < RX_RING_SIZE; i++) {
+ skb = fep->rx_skbuff[i];
+
+ if (bdp->cbd_bufaddr)
+ dma_unmap_single(&fep->pdev->dev, bdp->cbd_bufaddr,
+ SWITCH_ENET_RX_FRSIZE, DMA_FROM_DEVICE);
+ if (skb)
+ dev_kfree_skb(skb);
+ bdp++;
+ }
+
+ bdp = fep->tx_bd_base;
+ for (i = 0; i < TX_RING_SIZE; i++)
+ kfree(fep->tx_bounce[i]);
+}
+
+static int switch_alloc_buffers(struct net_device *ndev)
+{
+ struct switch_enet_private *fep = netdev_priv(ndev);
+ int i;
+ struct sk_buff *skb;
+ cbd_t *bdp;
+
+ bdp = fep->rx_bd_base;
+ for (i = 0; i < RX_RING_SIZE; i++) {
+ skb = dev_alloc_skb(SWITCH_ENET_RX_FRSIZE);
+ if (!skb) {
+ switch_enet_free_buffers(ndev);
+ return -ENOMEM;
+ }
+ fep->rx_skbuff[i] = skb;
+
+ bdp->cbd_bufaddr = dma_map_single(&fep->pdev->dev, skb->data,
+ SWITCH_ENET_RX_FRSIZE, DMA_FROM_DEVICE);
+ bdp->cbd_sc = BD_ENET_RX_EMPTY;
+
+ bdp++;
+ }
+
+ /* Set the last buffer to wrap. */
+ bdp--;
+ bdp->cbd_sc |= BD_SC_WRAP;
+
+ bdp = fep->tx_bd_base;
+ for (i = 0; i < TX_RING_SIZE; i++) {
+ fep->tx_bounce[i] = kmalloc(SWITCH_ENET_TX_FRSIZE, GFP_KERNEL);
+
+ bdp->cbd_sc = 0;
+ bdp->cbd_bufaddr = 0;
+ bdp++;
+ }
+
+ /* Set the last buffer to wrap. */
+ bdp--;
+ bdp->cbd_sc |= BD_SC_WRAP;
+
+ return 0;
+}
+
+static int switch_enet_open(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+ int i;
+
+ fep->phy1_link = 0;
+ fep->phy2_link = 0;
+
+ switch_init_phy(dev);
+ for (i = 0; i < SWITCH_EPORT_NUMBER; i++) {
+ phy_write(fep->phydev[i], MII_BMCR, BMCR_RESET);
+ udelay(10);
+ phy_start(fep->phydev[i]);
+ }
+
+ fep->phy1_old_link = 0;
+ fep->phy2_old_link = 0;
+ fep->phy1_link = 1;
+ fep->phy2_link = 1;
+
+ /* no phy, go full duplex, it's most likely a hub chip */
+ switch_restart(dev, 1);
+
+ /* if the fec open firstly, we need to do nothing*/
+ /* otherwise, we need to restart the FEC*/
+ if (fep->sequence_done == 0)
+ switch_restart(dev, 1);
+ else
+ fep->sequence_done = 0;
+
+ fep->currTime = 0;
+ fep->learning_irqhandle_enable = 0;
+
+ writel(0x70007, fep->membase + FEC_ESW_PER);
+ writel(FSL_ESW_DBCR_P0 | FSL_ESW_DBCR_P1 | FSL_ESW_DBCR_P2,
+ fep->membase + FEC_ESW_DBCR);
+ writel(FSL_ESW_DMCR_P0 | FSL_ESW_DMCR_P1 | FSL_ESW_DMCR_P2,
+ fep->membase + FEC_ESW_DMCR);
+
+ writel(0,fep->membase + FEC_ESW_BKLR);
+
+ netif_start_queue(dev);
+
+ /* And last, enable the receive processing.*/
+ writel(FSL_ESW_RDAR_R_DES_ACTIVE, fep->membase + FEC_ESW_RDAR);
+
+ fep->opened = 1;
+
+ return 0;
+}
+
+static int switch_enet_close(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+ int i;
+
+ /* Don't know what to do yet.*/
+ fep->opened = 0;
+ netif_stop_queue(dev);
+ switch_stop(dev);
+
+ for (i = 0; i < SWITCH_EPORT_NUMBER; i++) {
+ phy_disconnect(fep->phydev[i]);
+ phy_stop(fep->phydev[i]);
+ phy_write(fep->phydev[i], MII_BMCR, BMCR_PDOWN);
+ }
+
+ return 0;
+}
+
+#define HASH_BITS 6 /* #bits in hash */
+#define CRC32_POLY 0xEDB88320
+
+static void set_multicast_list(struct net_device *dev)
+{
+ struct switch_enet_private *fep;
+ unsigned int i, bit, data, crc;
+ struct netdev_hw_addr *ha;
+
+ fep = netdev_priv(dev);
+
+ if (dev->flags & IFF_PROMISC) {
+ printk(KERN_INFO "%s IFF_PROMISC\n", __func__);
+ } else {
+ if (dev->flags & IFF_ALLMULTI)
+ /* Catch all multicast addresses, so set the
+ * filter to all 1's.
+ */
+ printk(KERN_INFO "%s IFF_ALLMULTI\n", __func__);
+ else {
+ netdev_for_each_mc_addr(ha, dev) {
+ if (!(ha->addr[0] & 1))
+ continue;
+
+ /* calculate crc32 value of mac address
+ */
+ crc = 0xffffffff;
+
+ for (i = 0; i < dev->addr_len; i++) {
+ data = ha->addr[i];
+ for (bit = 0; bit < 8; bit++,
+ data >>= 1) {
+ crc = (crc >> 1) ^
+ (((crc ^ data) & 1) ?
+ CRC32_POLY : 0);
+ }
+ }
+
+ }
+ }
+ }
+}
+
+/* Set a MAC change in hardware.*/
+static void switch_get_mac_address(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+ struct switch_platform_data *pdata = fep->pdev->dev.platform_data;
+ unsigned char *iap, tmpaddr[ETH_ALEN];
+
+ iap = macaddr;
+
+ if (!is_valid_ether_addr(iap))
+ if (pdata)
+ memcpy(iap, pdata->mac, ETH_ALEN);
+
+ if (!is_valid_ether_addr(iap)) {
+ *((unsigned long *) &tmpaddr[0]) =
+ be32_to_cpu(readl(fep->enetbase + FSL_FEC_PALR0));
+ *((unsigned short *) &tmpaddr[4]) =
+ be16_to_cpu(readl(fep->enetbase + FSL_FEC_PAUR0) >> 16);
+ iap = &tmpaddr[0];
+ }
+
+ memcpy(dev->dev_addr, iap, ETH_ALEN);
+
+ /* Adjust MAC if using macaddr */
+ if (iap == macaddr)
+ dev->dev_addr[ETH_ALEN-1] =
+ macaddr[ETH_ALEN-1] + fep->pdev->id;
+}
+
+static void switch_hw_init(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+
+ /* Initialize MAC 0/1 */
+ writel(FSL_FEC_RCR_MAX_FL(1522) | FSL_FEC_RCR_RMII_MODE | FSL_FEC_RCR_PROM
+ | FSL_FEC_RCR_MII_MODE | FSL_FEC_RCR_MII_MODE,
+ fep->enetbase + FSL_FEC_RCR0);
+ writel(FSL_FEC_RCR_MAX_FL(1522) | FSL_FEC_RCR_RMII_MODE | FSL_FEC_RCR_PROM
+ | FSL_FEC_RCR_MII_MODE | FSL_FEC_RCR_MII_MODE,
+ fep->enetbase + FSL_FEC_RCR1);
+
+ writel(FSL_FEC_TCR_FDEN, fep->enetbase + FSL_FEC_TCR0);
+ writel(FSL_FEC_TCR_FDEN, fep->enetbase + FSL_FEC_TCR1);
+
+ writel(0x1a, fep->enetbase + FSL_FEC_MSCR0);
+
+ /* Set the station address for the ENET Adapter */
+ writel(dev->dev_addr[3] |
+ dev->dev_addr[2] << 8 |
+ dev->dev_addr[1] << 16 |
+ dev->dev_addr[0] << 24, fep->enetbase + FSL_FEC_PALR0);
+ writel((dev->dev_addr[5] << 16) |
+ (dev->dev_addr[4] << 24),
+ fep->enetbase + FSL_FEC_PAUR0);
+ writel(dev->dev_addr[3] |
+ dev->dev_addr[2] << 8 |
+ dev->dev_addr[1] << 16 |
+ dev->dev_addr[0] << 24, fep->enetbase + FSL_FEC_PALR1);
+ writel((dev->dev_addr[5] << 16) |
+ (dev->dev_addr[4] << 24),
+ fep->enetbase + FSL_FEC_PAUR1);
+
+ writel(FEC_ENET_TXF | FEC_ENET_RXF, fep->enetbase + FSL_FEC_EIMR0);
+ writel(FEC_ENET_TXF | FEC_ENET_RXF, fep->enetbase + FSL_FEC_EIMR1);
+
+ writel(FSL_FEC_ECR_ETHER_EN | (0x1 << 8), fep->enetbase + FSL_FEC_ECR0);
+ writel(FSL_FEC_ECR_ETHER_EN | (0x1 << 8), fep->enetbase + FSL_FEC_ECR1);
+ udelay(20);
+}
+
+static const struct net_device_ops switch_netdev_ops = {
+ .ndo_open = switch_enet_open,
+ .ndo_stop = switch_enet_close,
+ .ndo_start_xmit = switch_enet_start_xmit,
+ .ndo_set_multicast_list = set_multicast_list,
+ .ndo_do_ioctl = switch_enet_ioctl,
+ .ndo_tx_timeout = switch_timeout,
+};
+
+/* Initialize the FEC Ethernet.
+ */
+ /*
+ * XXX: We need to clean up on failure exits here.
+ */
+static int switch_enet_init(struct platform_device *pdev)
+{
+ struct net_device *dev = platform_get_drvdata(pdev);
+ struct switch_enet_private *fep = netdev_priv(dev);
+ cbd_t *cbd_base;
+ int ret = 0;
+
+ /* Allocate memory for buffer descriptors. */
+ cbd_base = dma_alloc_coherent(NULL, PAGE_SIZE, &fep->bd_dma,
+ GFP_KERNEL);
+ if (!cbd_base) {
+ printk(KERN_ERR "FEC: allocate descriptor memory failed?\n");
+ return -ENOMEM;
+ }
+
+ spin_lock_init(&fep->hw_lock);
+ spin_lock_init(&fep->mii_lock);
+
+ writel(FSL_ESW_MODE_SW_RST, fep->membase + FEC_ESW_MODE);
+ udelay(10);
+ writel(FSL_ESW_MODE_STATRST, fep->membase + FEC_ESW_MODE);
+ writel(FSL_ESW_MODE_SW_EN, fep->membase + FEC_ESW_MODE);
+
+ /* Enable transmit/receive on all ports */
+ writel(0xffffffff, fep->membase + FEC_ESW_PER);
+
+ /* Management port configuration,
+ * make port 0 as management port */
+ writel(0, fep->membase + FEC_ESW_BMPC);
+
+ /* Clear any outstanding interrupt.*/
+ writel(0xffffffff, fep->membase + FEC_ESW_ISR);
+ writel(0, fep->membase + FEC_ESW_IMR);
+ udelay(100);
+
+ switch_get_mac_address(dev);
+
+ /* Set receive and transmit descriptor base.
+ */
+ fep->rx_bd_base = cbd_base;
+ fep->tx_bd_base = cbd_base + RX_RING_SIZE;
+
+ dev->base_addr = (unsigned long)fep->membase;
+
+ /* The FEC Ethernet specific entries in the device structure. */
+ dev->watchdog_timeo = TX_TIMEOUT;
+ dev->netdev_ops = &switch_netdev_ops;
+
+ fep->skb_cur = fep->skb_dirty = 0;
+
+ ret = switch_alloc_buffers(dev);
+ if (ret)
+ return ret;
+
+ /* Set receive and transmit descriptor base.*/
+ writel(fep->bd_dma, fep->membase + FEC_ESW_RDSR);
+ writel((unsigned long)fep->bd_dma +
+ sizeof(struct bufdesc) * RX_RING_SIZE,
+ fep->membase + FEC_ESW_TDSR);
+
+ /*set mii*/
+ switch_hw_init(dev);
+
+ /* Clear any outstanding interrupt.*/
+ writel(0xffffffff, fep->membase + FEC_ESW_ISR);
+ writel(FSL_ESW_IMR_RXF | FSL_ESW_IMR_TXF, fep->membase + FEC_ESW_IMR);
+ esw_clear_atable(fep);
+
+ /* Queue up command to detect the PHY and initialize the
+ * remainder of the interface.
+ */
+#ifndef CONFIG_FEC_SHARED_PHY
+ fep->phy_addr = 0;
+#else
+ fep->phy_addr = fep->index;
+#endif
+
+ fep->sequence_done = 1;
+
+ return ret;
+}
+
+/* This function is called to start or restart the FEC during a link
+ * change. This only happens when switching between half and full
+ * duplex.
+ */
+static void switch_restart(struct net_device *dev, int duplex)
+{
+ struct switch_enet_private *fep;
+ int i;
+
+ fep = netdev_priv(dev);
+
+ /* Whack a reset. We should wait for this.*/
+ writel(1, fep->enetbase + FSL_FEC_ECR0);
+ writel(1, fep->enetbase + FSL_FEC_ECR1);
+ udelay(10);
+
+ writel(FSL_ESW_MODE_SW_RST, fep->membase + FEC_ESW_MODE);
+ udelay(10);
+ writel(FSL_ESW_MODE_STATRST, fep->membase + FEC_ESW_MODE);
+ writel(FSL_ESW_MODE_SW_EN, fep->membase + FEC_ESW_MODE);
+
+ /* Enable transmit/receive on all ports */
+ writel(0xffffffff, fep->membase + FEC_ESW_PER);
+
+ /* Management port configuration,
+ * make port 0 as management port */
+ writel(0, fep->membase + FEC_ESW_BMPC);
+
+ /* Clear any outstanding interrupt.*/
+ writel(0xffffffff, fep->membase + FEC_ESW_ISR);
+
+ switch_hw_init(dev);
+
+ /* Set station address.*/
+ switch_get_mac_address(dev);
+
+ writel(0, fep->membase + FEC_ESW_IMR);
+ udelay(10);
+
+ /* Set maximum receive buffer size.
+ */
+ writel(PKT_MAXBLR_SIZE, fep->membase + FEC_ESW_MRBR);
+
+ /* Set receive and transmit descriptor base.
+ */
+ writel(fep->bd_dma, fep->membase + FEC_ESW_RDSR);
+ writel((unsigned long)fep->bd_dma +
+ sizeof(struct bufdesc) * RX_RING_SIZE,
+ fep->membase + FEC_ESW_TDSR);
+
+ fep->dirty_tx = fep->cur_tx = fep->tx_bd_base;
+ fep->cur_rx = fep->rx_bd_base;
+
+ /* Reset SKB transmit buffers.
+ */
+ fep->skb_cur = fep->skb_dirty = 0;
+ for (i = 0; i <= TX_RING_MOD_MASK; i++) {
+ if (fep->tx_skbuff[i] != NULL) {
+ dev_kfree_skb_any(fep->tx_skbuff[i]);
+ fep->tx_skbuff[i] = NULL;
+ }
+ }
+
+ /*hardware has set in hw_init*/
+ fep->full_duplex = duplex;
+
+ /* Clear any outstanding interrupt.*/
+ writel(0xffffffff, fep->membase + FEC_ESW_ISR);
+ writel(FSL_ESW_IMR_RXF | FSL_ESW_IMR_TXF, fep->membase + FEC_ESW_IMR);
+}
+
+static void switch_stop(struct net_device *dev)
+{
+ struct switch_enet_private *fep = netdev_priv(dev);
+
+ /* We cannot expect a graceful transmit
+ * stop without link */
+ if (fep->phy1_link)
+ udelay(10);
+ if (fep->phy2_link)
+ udelay(10);
+
+ /* Whack a reset. We should wait for this */
+ udelay(10);
+}
+
+static int fec_mdio_register(struct net_device *dev)
+{
+ int i, err = 0;
+ struct switch_enet_private *fep = netdev_priv(dev);
+
+ fep->mdio_bus = mdiobus_alloc();
+ if (!fep->mdio_bus) {
+ printk(KERN_ERR "ethernet switch mdiobus_alloc fail\n");
+ return -ENOMEM;
+ }
+
+ fep->mdio_bus->name = "fsl l2 switch MII Bus";
+
+ snprintf(fep->mdio_bus->id, MII_BUS_ID_SIZE, "%x", fep->pdev->id);
+
+ fep->mdio_bus->read = &fec_enet_mdio_read;
+ fep->mdio_bus->write = &fec_enet_mdio_write;
+ fep->mdio_bus->priv = fep;
+
+ fep->mdio_bus->irq = kmalloc(sizeof(int) * PHY_MAX_ADDR, GFP_KERNEL);
+ if (!fep->mdio_bus->irq) {
+ err = -ENOMEM;
+ return err;
+ }
+
+ for (i = 0; i < PHY_MAX_ADDR; i++)
+ fep->mdio_bus->irq[i] = PHY_POLL;
+
+ err = mdiobus_register(fep->mdio_bus);
+ if (err) {
+ mdiobus_free(fep->mdio_bus);
+ printk(KERN_ERR "%s: ethernet mdiobus_register fail\n",
+ dev->name);
+ return -EIO;
+ }
+
+ printk(KERN_INFO "%s mdiobus(%s) register ok.\n",
+ fep->mdio_bus->name, fep->mdio_bus->id);
+ return err;
+}
+
+static int __devinit eth_switch_probe(struct platform_device *pdev)
+{
+ struct net_device *ndev;
+ int err;
+ struct switch_enet_private *fep;
+ struct task_struct *task;
+ int i, irq, ret = 0;
+ struct resource *r;
+
+ printk(KERN_INFO "Ethernet Switch Version 1.0\n");
+
+ r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (!r) {
+ printk(KERN_ERR "get resource error!\n");
+ return -ENXIO;
+ }
+ r = request_mem_region(r->start, resource_size(r), pdev->name);
+ if (!r)
+ return -EBUSY;
+
+ ndev = alloc_etherdev(sizeof(struct switch_enet_private));
+ if (!ndev) {
+ printk(KERN_ERR "%s: ethernet switch alloc_etherdev fail\n",
+ ndev->name);
+ return -ENOMEM;
+ }
+ SET_NETDEV_DEV(ndev, &pdev->dev);
+
+ fep = netdev_priv(ndev);
+
+ memset(fep, 0, sizeof(*fep));
+
+ fep->membase = ioremap(r->start, resource_size(r));
+ if (!fep->membase)
+ return -ENOMEM;
+
+ r = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+ if (!r)
+ return -ENXIO;
+
+ r = request_mem_region(r->start, resource_size(r), pdev->name);
+ if (!r)
+ return -EBUSY;
+ fep->macbase = ioremap(r->start, resource_size(r));
+
+ r = platform_get_resource(pdev, IORESOURCE_MEM, 2);
+ if (!r)
+ return -ENXIO;
+
+ r = request_mem_region(r->start, resource_size(r), pdev->name);
+ if (!r)
+ return -EBUSY;
+ fep->enetbase = ioremap(r->start, resource_size(r));
+ if (!fep->enetbase)
+ return -ENOMEM;
+
+ fep->pdev = pdev;
+ platform_set_drvdata(pdev, ndev);
+
+ printk(KERN_INFO "%s: ethernet switch init\n", __func__);
+
+ /* This device has up to three irqs on some platforms */
+ for (i = 0; i < 3; i++) {
+ irq = platform_get_irq(pdev, i);
+ if (i && irq < 0)
+ break;
+ ret = request_irq(irq, switch_enet_interrupt, IRQF_DISABLED,
+ pdev->name, ndev);
+ if (ret) {
+ while (--i >= 0) {
+ irq = platform_get_irq(pdev, i);
+ free_irq(irq, ndev);
+ }
+ return -ENOMEM;
+ }
+ }
+
+ fep->clk = clk_get(&pdev->dev, "switch_clk");
+
+ if (IS_ERR(fep->clk)) {
+ ret = PTR_ERR(fep->clk);
+ goto failed_clk;
+ }
+ clk_enable(fep->clk);
+
+ err = switch_enet_init(pdev);
+ if (err) {
+ free_netdev(ndev);
+ platform_set_drvdata(pdev, NULL);
+ return -EIO;
+ }
+
+ err = fec_mdio_register(ndev);
+ if (err) {
+ printk(KERN_ERR "%s: L2 switch fec_mdio_register error!\n",
+ ndev->name);
+ free_netdev(ndev);
+ platform_set_drvdata(pdev, NULL);
+ return -ENOMEM;
+ }
+
+ /* setup timer for Learning Aging function */
+ init_timer(&fep->timer_aging);
+ fep->timer_aging.function = l2switch_aging_timer;
+ fep->timer_aging.data = (unsigned long) fep;
+ fep->timer_aging.expires = jiffies + LEARNING_AGING_TIMER;
+ add_timer(&fep->timer_aging);
+
+ /* register network device*/
+ if (register_netdev(ndev) != 0) {
+ /* XXX: missing cleanup here */
+ free_netdev(ndev);
+ platform_set_drvdata(pdev, NULL);
+ printk(KERN_ERR "%s: L2 switch register_netdev fail\n",
+ ndev->name);
+ return -EIO;
+ }
+
+ task = kthread_run(switch_enet_learning, fep,
+ "fsl_l2switch_learning");
+ if (IS_ERR(task)) {
+ err = PTR_ERR(task);
+ return err;
+ }
+
+ printk(KERN_INFO "%s: ethernet switch %pM\n",
+ ndev->name, ndev->dev_addr);
+ return 0;
+failed_clk:
+ iounmap(fep->membase);
+ return ret;
+}
+
+static int __devexit eth_switch_remove(struct platform_device *pdev)
+{
+ int i;
+ struct net_device *dev;
+ struct switch_enet_private *fep;
+ struct switch_platform_private *chip;
+
+ chip = platform_get_drvdata(pdev);
+ if (chip) {
+ for (i = 0; i < chip->num_slots; i++) {
+ fep = chip->fep_host[i];
+ dev = fep->netdev;
+ fep->sequence_done = 1;
+ unregister_netdev(dev);
+ free_netdev(dev);
+
+ del_timer_sync(&fep->timer_aging);
+ }
+
+ platform_set_drvdata(pdev, NULL);
+ kfree(chip);
+
+ } else
+ printk(KERN_ERR "%s: can not get the "
+ "switch_platform_private %x\n", __func__,
+ (unsigned int)chip);
+
+ return 0;
+}
+
+static struct platform_driver eth_switch_driver = {
+ .probe = eth_switch_probe,
+ .remove = __devexit_p(eth_switch_remove),
+ .driver = {
+ .name = "switch",
+ .owner = THIS_MODULE,
+ },
+};
+
+static int fec_mac_addr_setup(char *mac_addr)
+{
+ char *ptr, *p = mac_addr;
+ unsigned long tmp;
+ int i = 0, ret = 0;
+
+ while (p && (*p) && i < 6) {
+ ptr = strchr(p, ':');
+ if (ptr)
+ *ptr++ = '\0';
+
+ if (strlen(p)) {
+ ret = strict_strtoul(p, 16, &tmp);
+ if (ret < 0 || tmp > 0xff)
+ break;
+ macaddr[i++] = tmp;
+ }
+ p = ptr;
+ }
+
+ return 0;
+}
+
+__setup("fec_mac=", fec_mac_addr_setup);
+
+static int __init fsl_l2_switch_init(void)
+{
+ return platform_driver_register(&eth_switch_driver);
+}
+
+static void __exit fsl_l2_switch_exit(void)
+{
+ platform_driver_unregister(&eth_switch_driver);
+}
+
+module_init(fsl_l2_switch_init);
+module_exit(fsl_l2_switch_exit);
+MODULE_LICENSE("GPL");