powerpc: Partialy restore core of mpc8xx
CS Systemes d'Information (CSSI) manufactures 8xx boards for
critical communication systems. Those boards have been
running U-Boot since 2010 and will have to be maintained
until at least 2027.
commit 5b8e76c35ec312a3f73126bd1a2d2c0965b98a9f
("powerpc, 8xx: remove support for 8xx") orphaned those boards
by removing support for the mpc8xx CPU.
This commit partially restores support for the 8xx, with the
following limitations:
- Restores support for MPC866 and MPC885 only
- Does not restore IDE, PCMCIA, I2C, USB
- Does not restore examples
- Does not restore POST
- Does not restore Ethernet on SCC
- Does not restore console on SCC
- Does not restore bedbug and kgdb support
As the 866 and 885 do not support the following features,
they are not restored either:
- VIDEO / LCD
- RTC clock
The CPM uCODE patch is not restored either, because:
- 866 and 885 already have support for I2C and SPI relocation
without a uCODE patch
- relocation of SMC, I2C or SPI is only needed for using SCCs
for Ethernet or QMC
The dynamic setup/calculation of clocks is removed, we
expect the target being use with the clock and PLPRCR register
defined in the configuration.
All the clock settings for 8xx prior to 866 is removed as
well as we now only support 866 and 885.
This code is mature and addresses mature boards. Therefore
all code enclosed in '#if 0/#endif' and '#if XX_DEBUG/#endif'
is unneeded.
The following files are not restored by this patch:
- arch/powerpc/cpu/mpc8xx/bedbug_860.c
- arch/powerpc/cpu/mpc8xx/fec.h
- arch/powerpc/cpu/mpc8xx/kgdb.S
- arch/powerpc/cpu/mpc8xx/plprcr_write.S
- arch/powerpc/cpu/mpc8xx/scc.c
- arch/powerpc/cpu/mpc8xx/upatch.c
- arch/powerpc/cpu/mpc8xx/video.c
- arch/powerpc/include/asm/status_led.h
- arch/powerpc/lib/ide.c
- arch/powerpc/lib/ide.h
- doc/README.MPC866
- drivers/pcmcia/mpc8xx_pcmcia.c
- drivers/rtc/mpc8xx.c
- drivers/usb/gadget/mpc8xx_udc.c
- drivers/video/mpc8xx_lcd.c
- examples/standalone/test_burst.c
- examples/standalone/test_burst.h
- examples/standalone/test_burst_lib.S
- examples/standalone/timer.c
- include/mpc823_lcd.h
- include/usb/mpc8xx_udc.h
- post/cpu/mpc8xx/Makefile
- post/cpu/mpc8xx/cache.c
- post/cpu/mpc8xx/cache_8xx.S
- post/cpu/mpc8xx/ether.c
- post/cpu/mpc8xx/spr.c
- post/cpu/mpc8xx/uart.c
- post/cpu/mpc8xx/usb.c
- post/cpu/mpc8xx/watchdog.c
Some of the restored files are not located in a proper location.
In order to keep traceability of the changes, they will be
moved to their correct location and moved to Kconfig in a
followup patch.
This patch also declares CSSI as point of contact for the update
of the 8xx platform, as those boards are the only ones still
being maintained on the 8xx area. A later patch will add
those boards to the tree.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
diff --git a/arch/powerpc/cpu/mpc8xx/fec.c b/arch/powerpc/cpu/mpc8xx/fec.c
new file mode 100644
index 0000000..7aa526d
--- /dev/null
+++ b/arch/powerpc/cpu/mpc8xx/fec.c
@@ -0,0 +1,847 @@
+/*
+ * (C) Copyright 2000
+ * Wolfgang Denk, DENX Software Engineering, wd@denx.de.
+ *
+ * SPDX-License-Identifier: GPL-2.0+
+ */
+
+#include <common.h>
+#include <command.h>
+#include <commproc.h>
+#include <malloc.h>
+#include <net.h>
+
+#include <phy.h>
+
+DECLARE_GLOBAL_DATA_PTR;
+
+#if defined(CONFIG_CMD_NET) && \
+ (defined(FEC_ENET) || defined(CONFIG_ETHER_ON_FEC1) || defined(CONFIG_ETHER_ON_FEC2))
+
+/* compatibility test, if only FEC_ENET defined assume ETHER on FEC1 */
+#if defined(FEC_ENET) && !defined(CONFIG_ETHER_ON_FEC1) && !defined(CONFIG_ETHER_ON_FEC2)
+#define CONFIG_ETHER_ON_FEC1 1
+#endif
+
+/* define WANT_MII when MII support is required */
+#if defined(CONFIG_SYS_DISCOVER_PHY) || defined(CONFIG_FEC1_PHY) || defined(CONFIG_FEC2_PHY)
+#define WANT_MII
+#else
+#undef WANT_MII
+#endif
+
+#if defined(WANT_MII)
+#include <miiphy.h>
+
+#if !(defined(CONFIG_MII) || defined(CONFIG_CMD_MII))
+#error "CONFIG_MII has to be defined!"
+#endif
+
+#endif
+
+#if defined(CONFIG_RMII) && !defined(WANT_MII)
+#error RMII support is unusable without a working PHY.
+#endif
+
+#ifdef CONFIG_SYS_DISCOVER_PHY
+static int mii_discover_phy(struct eth_device *dev);
+#endif
+
+int fec8xx_miiphy_read(struct mii_dev *bus, int addr, int devad, int reg);
+int fec8xx_miiphy_write(struct mii_dev *bus, int addr, int devad, int reg,
+ u16 value);
+
+static struct ether_fcc_info_s
+{
+ int ether_index;
+ int fecp_offset;
+ int phy_addr;
+ int actual_phy_addr;
+ int initialized;
+}
+ ether_fcc_info[] = {
+#if defined(CONFIG_ETHER_ON_FEC1)
+ {
+ 0,
+ offsetof(immap_t, im_cpm.cp_fec1),
+#if defined(CONFIG_FEC1_PHY)
+ CONFIG_FEC1_PHY,
+#else
+ -1, /* discover */
+#endif
+ -1,
+ 0,
+
+ },
+#endif
+#if defined(CONFIG_ETHER_ON_FEC2)
+ {
+ 1,
+ offsetof(immap_t, im_cpm.cp_fec2),
+#if defined(CONFIG_FEC2_PHY)
+ CONFIG_FEC2_PHY,
+#else
+ -1,
+#endif
+ -1,
+ 0,
+ },
+#endif
+};
+
+/* Ethernet Transmit and Receive Buffers */
+#define DBUF_LENGTH 1520
+
+#define TX_BUF_CNT 2
+
+#define TOUT_LOOP 100
+
+#define PKT_MAXBUF_SIZE 1518
+#define PKT_MINBUF_SIZE 64
+#define PKT_MAXBLR_SIZE 1520
+
+#ifdef __GNUC__
+static char txbuf[DBUF_LENGTH] __attribute__ ((aligned(8)));
+#else
+#error txbuf must be aligned.
+#endif
+
+static uint rxIdx; /* index of the current RX buffer */
+static uint txIdx; /* index of the current TX buffer */
+
+/*
+ * FEC Ethernet Tx and Rx buffer descriptors allocated at the
+ * immr->udata_bd address on Dual-Port RAM
+ * Provide for Double Buffering
+ */
+
+typedef volatile struct CommonBufferDescriptor {
+ cbd_t rxbd[PKTBUFSRX]; /* Rx BD */
+ cbd_t txbd[TX_BUF_CNT]; /* Tx BD */
+} RTXBD;
+
+static RTXBD *rtx = NULL;
+
+static int fec_send(struct eth_device *dev, void *packet, int length);
+static int fec_recv(struct eth_device* dev);
+static int fec_init(struct eth_device* dev, bd_t * bd);
+static void fec_halt(struct eth_device* dev);
+#if defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
+static void __mii_init(void);
+#endif
+
+int fec_initialize(bd_t *bis)
+{
+ struct eth_device* dev;
+ struct ether_fcc_info_s *efis;
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(ether_fcc_info); i++) {
+
+ dev = malloc(sizeof(*dev));
+ if (dev == NULL)
+ hang();
+
+ memset(dev, 0, sizeof(*dev));
+
+ /* for FEC1 make sure that the name of the interface is the same
+ as the old one for compatibility reasons */
+ if (i == 0) {
+ strcpy(dev->name, "FEC");
+ } else {
+ sprintf (dev->name, "FEC%d",
+ ether_fcc_info[i].ether_index + 1);
+ }
+
+ efis = ðer_fcc_info[i];
+
+ /*
+ * reset actual phy addr
+ */
+ efis->actual_phy_addr = -1;
+
+ dev->priv = efis;
+ dev->init = fec_init;
+ dev->halt = fec_halt;
+ dev->send = fec_send;
+ dev->recv = fec_recv;
+
+ eth_register(dev);
+
+#if defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
+ int retval;
+ struct mii_dev *mdiodev = mdio_alloc();
+ if (!mdiodev)
+ return -ENOMEM;
+ strncpy(mdiodev->name, dev->name, MDIO_NAME_LEN);
+ mdiodev->read = fec8xx_miiphy_read;
+ mdiodev->write = fec8xx_miiphy_write;
+
+ retval = mdio_register(mdiodev);
+ if (retval < 0)
+ return retval;
+#endif
+ }
+ return 1;
+}
+
+static int fec_send(struct eth_device *dev, void *packet, int length)
+{
+ int j, rc;
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile fec_t *fecp = (volatile fec_t *)(CONFIG_SYS_IMMR + efis->fecp_offset);
+
+ /* section 16.9.23.3
+ * Wait for ready
+ */
+ j = 0;
+ while ((rtx->txbd[txIdx].cbd_sc & BD_ENET_TX_READY) && (j<TOUT_LOOP)) {
+ udelay(1);
+ j++;
+ }
+ if (j>=TOUT_LOOP) {
+ printf("TX not ready\n");
+ }
+
+ rtx->txbd[txIdx].cbd_bufaddr = (uint)packet;
+ rtx->txbd[txIdx].cbd_datlen = length;
+ rtx->txbd[txIdx].cbd_sc |= BD_ENET_TX_READY | BD_ENET_TX_LAST;
+ __asm__ ("eieio");
+
+ /* Activate transmit Buffer Descriptor polling */
+ fecp->fec_x_des_active = 0x01000000; /* Descriptor polling active */
+
+ j = 0;
+ while ((rtx->txbd[txIdx].cbd_sc & BD_ENET_TX_READY) && (j<TOUT_LOOP)) {
+ udelay(1);
+ j++;
+ }
+ if (j>=TOUT_LOOP) {
+ printf("TX timeout\n");
+ }
+ /* return only status bits */;
+ rc = (rtx->txbd[txIdx].cbd_sc & BD_ENET_TX_STATS);
+
+ txIdx = (txIdx + 1) % TX_BUF_CNT;
+
+ return rc;
+}
+
+static int fec_recv (struct eth_device *dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile fec_t *fecp =
+ (volatile fec_t *) (CONFIG_SYS_IMMR + efis->fecp_offset);
+ int length;
+
+ for (;;) {
+ /* section 16.9.23.2 */
+ if (rtx->rxbd[rxIdx].cbd_sc & BD_ENET_RX_EMPTY) {
+ length = -1;
+ break; /* nothing received - leave for() loop */
+ }
+
+ length = rtx->rxbd[rxIdx].cbd_datlen;
+
+ if (rtx->rxbd[rxIdx].cbd_sc & 0x003f) {
+ } else {
+ uchar *rx = net_rx_packets[rxIdx];
+
+ length -= 4;
+
+#if defined(CONFIG_CMD_CDP)
+ if ((rx[0] & 1) != 0 &&
+ memcmp((uchar *)rx, net_bcast_ethaddr, 6) != 0 &&
+ !is_cdp_packet((uchar *)rx))
+ rx = NULL;
+#endif
+ /*
+ * Pass the packet up to the protocol layers.
+ */
+ if (rx != NULL)
+ net_process_received_packet(rx, length);
+ }
+
+ /* Give the buffer back to the FEC. */
+ rtx->rxbd[rxIdx].cbd_datlen = 0;
+
+ /* wrap around buffer index when necessary */
+ if ((rxIdx + 1) >= PKTBUFSRX) {
+ rtx->rxbd[PKTBUFSRX - 1].cbd_sc =
+ (BD_ENET_RX_WRAP | BD_ENET_RX_EMPTY);
+ rxIdx = 0;
+ } else {
+ rtx->rxbd[rxIdx].cbd_sc = BD_ENET_RX_EMPTY;
+ rxIdx++;
+ }
+
+ __asm__ ("eieio");
+
+ /* Try to fill Buffer Descriptors */
+ fecp->fec_r_des_active = 0x01000000; /* Descriptor polling active */
+ }
+
+ return length;
+}
+
+/**************************************************************
+ *
+ * FEC Ethernet Initialization Routine
+ *
+ *************************************************************/
+
+#define FEC_ECNTRL_PINMUX 0x00000004
+#define FEC_ECNTRL_ETHER_EN 0x00000002
+#define FEC_ECNTRL_RESET 0x00000001
+
+#define FEC_RCNTRL_BC_REJ 0x00000010
+#define FEC_RCNTRL_PROM 0x00000008
+#define FEC_RCNTRL_MII_MODE 0x00000004
+#define FEC_RCNTRL_DRT 0x00000002
+#define FEC_RCNTRL_LOOP 0x00000001
+
+#define FEC_TCNTRL_FDEN 0x00000004
+#define FEC_TCNTRL_HBC 0x00000002
+#define FEC_TCNTRL_GTS 0x00000001
+
+#define FEC_RESET_DELAY 50
+
+#if defined(CONFIG_RMII)
+
+static inline void fec_10Mbps(struct eth_device *dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ int fecidx = efis->ether_index;
+ uint mask = (fecidx == 0) ? 0x0000010 : 0x0000008;
+
+ if ((unsigned int)fecidx >= 2)
+ hang();
+
+ ((volatile immap_t *)CONFIG_SYS_IMMR)->im_cpm.cp_cptr |= mask;
+}
+
+static inline void fec_100Mbps(struct eth_device *dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ int fecidx = efis->ether_index;
+ uint mask = (fecidx == 0) ? 0x0000010 : 0x0000008;
+
+ if ((unsigned int)fecidx >= 2)
+ hang();
+
+ ((volatile immap_t *)CONFIG_SYS_IMMR)->im_cpm.cp_cptr &= ~mask;
+}
+
+#endif
+
+static inline void fec_full_duplex(struct eth_device *dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile fec_t *fecp = (volatile fec_t *)(CONFIG_SYS_IMMR + efis->fecp_offset);
+
+ fecp->fec_r_cntrl &= ~FEC_RCNTRL_DRT;
+ fecp->fec_x_cntrl |= FEC_TCNTRL_FDEN; /* FD enable */
+}
+
+static inline void fec_half_duplex(struct eth_device *dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile fec_t *fecp = (volatile fec_t *)(CONFIG_SYS_IMMR + efis->fecp_offset);
+
+ fecp->fec_r_cntrl |= FEC_RCNTRL_DRT;
+ fecp->fec_x_cntrl &= ~FEC_TCNTRL_FDEN; /* FD disable */
+}
+
+static void fec_pin_init(int fecidx)
+{
+ bd_t *bd = gd->bd;
+ volatile immap_t *immr = (immap_t *) CONFIG_SYS_IMMR;
+
+ /*
+ * Set MII speed to 2.5 MHz or slightly below.
+ *
+ * According to the MPC860T (Rev. D) Fast ethernet controller user
+ * manual (6.2.14),
+ * the MII management interface clock must be less than or equal
+ * to 2.5 MHz.
+ * This MDC frequency is equal to system clock / (2 * MII_SPEED).
+ * Then MII_SPEED = system_clock / 2 * 2,5 MHz.
+ *
+ * All MII configuration is done via FEC1 registers:
+ */
+ immr->im_cpm.cp_fec1.fec_mii_speed = ((bd->bi_intfreq + 4999999) / 5000000) << 1;
+
+#if defined(CONFIG_MPC885_FAMILY) && defined(WANT_MII)
+ /* use MDC for MII */
+ immr->im_ioport.iop_pdpar |= 0x0080;
+ immr->im_ioport.iop_pddir &= ~0x0080;
+#endif
+
+ if (fecidx == 0) {
+#if defined(CONFIG_ETHER_ON_FEC1)
+
+#if defined(CONFIG_MPC885_FAMILY) /* MPC87x/88x have got 2 FECs and different pinout */
+
+#if !defined(CONFIG_RMII)
+
+ immr->im_ioport.iop_papar |= 0xf830;
+ immr->im_ioport.iop_padir |= 0x0830;
+ immr->im_ioport.iop_padir &= ~0xf000;
+
+ immr->im_cpm.cp_pbpar |= 0x00001001;
+ immr->im_cpm.cp_pbdir &= ~0x00001001;
+
+ immr->im_ioport.iop_pcpar |= 0x000c;
+ immr->im_ioport.iop_pcdir &= ~0x000c;
+
+ immr->im_cpm.cp_pepar |= 0x00000003;
+ immr->im_cpm.cp_pedir |= 0x00000003;
+ immr->im_cpm.cp_peso &= ~0x00000003;
+
+ immr->im_cpm.cp_cptr &= ~0x00000100;
+
+#else
+
+#if !defined(CONFIG_FEC1_PHY_NORXERR)
+ immr->im_ioport.iop_papar |= 0x1000;
+ immr->im_ioport.iop_padir &= ~0x1000;
+#endif
+ immr->im_ioport.iop_papar |= 0xe810;
+ immr->im_ioport.iop_padir |= 0x0810;
+ immr->im_ioport.iop_padir &= ~0xe000;
+
+ immr->im_cpm.cp_pbpar |= 0x00000001;
+ immr->im_cpm.cp_pbdir &= ~0x00000001;
+
+ immr->im_cpm.cp_cptr |= 0x00000100;
+ immr->im_cpm.cp_cptr &= ~0x00000050;
+
+#endif /* !CONFIG_RMII */
+
+#else
+ /*
+ * Configure all of port D for MII.
+ */
+ immr->im_ioport.iop_pdpar = 0x1fff;
+
+ immr->im_ioport.iop_pddir = 0x1fff; /* Rev. D and later */
+#endif
+
+#endif /* CONFIG_ETHER_ON_FEC1 */
+ } else if (fecidx == 1) {
+
+#if defined(CONFIG_ETHER_ON_FEC2)
+
+#if defined(CONFIG_MPC885_FAMILY) /* MPC87x/88x have got 2 FECs and different pinout */
+
+#if !defined(CONFIG_RMII)
+ immr->im_cpm.cp_pepar |= 0x0003fffc;
+ immr->im_cpm.cp_pedir |= 0x0003fffc;
+ immr->im_cpm.cp_peso &= ~0x000087fc;
+ immr->im_cpm.cp_peso |= 0x00037800;
+
+ immr->im_cpm.cp_cptr &= ~0x00000080;
+#else
+
+#if !defined(CONFIG_FEC2_PHY_NORXERR)
+ immr->im_cpm.cp_pepar |= 0x00000010;
+ immr->im_cpm.cp_pedir |= 0x00000010;
+ immr->im_cpm.cp_peso &= ~0x00000010;
+#endif
+ immr->im_cpm.cp_pepar |= 0x00039620;
+ immr->im_cpm.cp_pedir |= 0x00039620;
+ immr->im_cpm.cp_peso |= 0x00031000;
+ immr->im_cpm.cp_peso &= ~0x00008620;
+
+ immr->im_cpm.cp_cptr |= 0x00000080;
+ immr->im_cpm.cp_cptr &= ~0x00000028;
+#endif /* CONFIG_RMII */
+
+#endif /* CONFIG_MPC885_FAMILY */
+
+#endif /* CONFIG_ETHER_ON_FEC2 */
+
+ }
+}
+
+static int fec_reset(volatile fec_t *fecp)
+{
+ int i;
+
+ /* Whack a reset.
+ * A delay is required between a reset of the FEC block and
+ * initialization of other FEC registers because the reset takes
+ * some time to complete. If you don't delay, subsequent writes
+ * to FEC registers might get killed by the reset routine which is
+ * still in progress.
+ */
+
+ fecp->fec_ecntrl = FEC_ECNTRL_PINMUX | FEC_ECNTRL_RESET;
+ for (i = 0;
+ (fecp->fec_ecntrl & FEC_ECNTRL_RESET) && (i < FEC_RESET_DELAY);
+ ++i) {
+ udelay (1);
+ }
+ if (i == FEC_RESET_DELAY)
+ return -1;
+
+ return 0;
+}
+
+static int fec_init (struct eth_device *dev, bd_t * bd)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile immap_t *immr = (immap_t *) CONFIG_SYS_IMMR;
+ volatile fec_t *fecp =
+ (volatile fec_t *) (CONFIG_SYS_IMMR + efis->fecp_offset);
+ int i;
+
+#if defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
+ /* the MII interface is connected to FEC1
+ * so for the miiphy_xxx function to work we must
+ * call mii_init since fec_halt messes the thing up
+ */
+ if (efis->ether_index != 0)
+ __mii_init();
+#endif
+
+ if (fec_reset(fecp) < 0)
+ printf ("FEC_RESET_DELAY timeout\n");
+
+ /* We use strictly polling mode only
+ */
+ fecp->fec_imask = 0;
+
+ /* Clear any pending interrupt
+ */
+ fecp->fec_ievent = 0xffc0;
+
+ /* No need to set the IVEC register */
+
+ /* Set station address
+ */
+#define ea dev->enetaddr
+ fecp->fec_addr_low = (ea[0] << 24) | (ea[1] << 16) | (ea[2] << 8) | (ea[3]);
+ fecp->fec_addr_high = (ea[4] << 8) | (ea[5]);
+#undef ea
+
+#if defined(CONFIG_CMD_CDP)
+ /*
+ * Turn on multicast address hash table
+ */
+ fecp->fec_hash_table_high = 0xffffffff;
+ fecp->fec_hash_table_low = 0xffffffff;
+#else
+ /* Clear multicast address hash table
+ */
+ fecp->fec_hash_table_high = 0;
+ fecp->fec_hash_table_low = 0;
+#endif
+
+ /* Set maximum receive buffer size.
+ */
+ fecp->fec_r_buff_size = PKT_MAXBLR_SIZE;
+
+ /* Set maximum frame length
+ */
+ fecp->fec_r_hash = PKT_MAXBUF_SIZE;
+
+ /*
+ * Setup Buffers and Buffer Desriptors
+ */
+ rxIdx = 0;
+ txIdx = 0;
+
+ if (!rtx)
+ rtx = (RTXBD *)(immr->im_cpm.cp_dpmem + CPM_FEC_BASE);
+ /*
+ * Setup Receiver Buffer Descriptors (13.14.24.18)
+ * Settings:
+ * Empty, Wrap
+ */
+ for (i = 0; i < PKTBUFSRX; i++) {
+ rtx->rxbd[i].cbd_sc = BD_ENET_RX_EMPTY;
+ rtx->rxbd[i].cbd_datlen = 0; /* Reset */
+ rtx->rxbd[i].cbd_bufaddr = (uint) net_rx_packets[i];
+ }
+ rtx->rxbd[PKTBUFSRX - 1].cbd_sc |= BD_ENET_RX_WRAP;
+
+ /*
+ * Setup Ethernet Transmitter Buffer Descriptors (13.14.24.19)
+ * Settings:
+ * Last, Tx CRC
+ */
+ for (i = 0; i < TX_BUF_CNT; i++) {
+ rtx->txbd[i].cbd_sc = BD_ENET_TX_LAST | BD_ENET_TX_TC;
+ rtx->txbd[i].cbd_datlen = 0; /* Reset */
+ rtx->txbd[i].cbd_bufaddr = (uint) (&txbuf[0]);
+ }
+ rtx->txbd[TX_BUF_CNT - 1].cbd_sc |= BD_ENET_TX_WRAP;
+
+ /* Set receive and transmit descriptor base
+ */
+ fecp->fec_r_des_start = (unsigned int) (&rtx->rxbd[0]);
+ fecp->fec_x_des_start = (unsigned int) (&rtx->txbd[0]);
+
+ /* Enable MII mode
+ */
+ /* Half duplex mode */
+ fecp->fec_r_cntrl = FEC_RCNTRL_MII_MODE | FEC_RCNTRL_DRT;
+ fecp->fec_x_cntrl = 0;
+
+ /* Enable big endian and don't care about SDMA FC.
+ */
+ fecp->fec_fun_code = 0x78000000;
+
+ /*
+ * Setup the pin configuration of the FEC
+ */
+ fec_pin_init (efis->ether_index);
+
+ rxIdx = 0;
+ txIdx = 0;
+
+ /*
+ * Now enable the transmit and receive processing
+ */
+ fecp->fec_ecntrl = FEC_ECNTRL_PINMUX | FEC_ECNTRL_ETHER_EN;
+
+ if (efis->phy_addr == -1) {
+#ifdef CONFIG_SYS_DISCOVER_PHY
+ /*
+ * wait for the PHY to wake up after reset
+ */
+ efis->actual_phy_addr = mii_discover_phy (dev);
+
+ if (efis->actual_phy_addr == -1) {
+ printf ("Unable to discover phy!\n");
+ return -1;
+ }
+#else
+ efis->actual_phy_addr = -1;
+#endif
+ } else {
+ efis->actual_phy_addr = efis->phy_addr;
+ }
+
+#if defined(CONFIG_MII) && defined(CONFIG_RMII)
+ /*
+ * adapt the RMII speed to the speed of the phy
+ */
+ if (miiphy_speed (dev->name, efis->actual_phy_addr) == _100BASET) {
+ fec_100Mbps (dev);
+ } else {
+ fec_10Mbps (dev);
+ }
+#endif
+
+#if defined(CONFIG_MII)
+ /*
+ * adapt to the half/full speed settings
+ */
+ if (miiphy_duplex (dev->name, efis->actual_phy_addr) == FULL) {
+ fec_full_duplex (dev);
+ } else {
+ fec_half_duplex (dev);
+ }
+#endif
+
+ /* And last, try to fill Rx Buffer Descriptors */
+ fecp->fec_r_des_active = 0x01000000; /* Descriptor polling active */
+
+ efis->initialized = 1;
+
+ return 0;
+}
+
+
+static void fec_halt(struct eth_device* dev)
+{
+ struct ether_fcc_info_s *efis = dev->priv;
+ volatile fec_t *fecp = (volatile fec_t *)(CONFIG_SYS_IMMR + efis->fecp_offset);
+ int i;
+
+ /* avoid halt if initialized; mii gets stuck otherwise */
+ if (!efis->initialized)
+ return;
+
+ /* Whack a reset.
+ * A delay is required between a reset of the FEC block and
+ * initialization of other FEC registers because the reset takes
+ * some time to complete. If you don't delay, subsequent writes
+ * to FEC registers might get killed by the reset routine which is
+ * still in progress.
+ */
+
+ fecp->fec_ecntrl = FEC_ECNTRL_PINMUX | FEC_ECNTRL_RESET;
+ for (i = 0;
+ (fecp->fec_ecntrl & FEC_ECNTRL_RESET) && (i < FEC_RESET_DELAY);
+ ++i) {
+ udelay (1);
+ }
+ if (i == FEC_RESET_DELAY) {
+ printf ("FEC_RESET_DELAY timeout\n");
+ return;
+ }
+
+ efis->initialized = 0;
+}
+
+#if defined(CONFIG_SYS_DISCOVER_PHY) || defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
+
+/* Make MII read/write commands for the FEC.
+*/
+
+#define mk_mii_read(ADDR, REG) (0x60020000 | ((ADDR << 23) | \
+ (REG & 0x1f) << 18))
+
+#define mk_mii_write(ADDR, REG, VAL) (0x50020000 | ((ADDR << 23) | \
+ (REG & 0x1f) << 18) | \
+ (VAL & 0xffff))
+
+/* 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 */
+
+/* send command to phy using mii, wait for result */
+static uint
+mii_send(uint mii_cmd)
+{
+ uint mii_reply;
+ volatile fec_t *ep;
+ int cnt;
+
+ ep = &(((immap_t *)CONFIG_SYS_IMMR)->im_cpm.cp_fec);
+
+ ep->fec_mii_data = mii_cmd; /* command to phy */
+
+ /* wait for mii complete */
+ cnt = 0;
+ while (!(ep->fec_ievent & FEC_ENET_MII)) {
+ if (++cnt > 1000) {
+ printf("mii_send STUCK!\n");
+ break;
+ }
+ }
+ mii_reply = ep->fec_mii_data; /* result from phy */
+ ep->fec_ievent = FEC_ENET_MII; /* clear MII complete */
+ return (mii_reply & 0xffff); /* data read from phy */
+}
+#endif
+
+#if defined(CONFIG_SYS_DISCOVER_PHY)
+static int mii_discover_phy(struct eth_device *dev)
+{
+#define MAX_PHY_PASSES 11
+ uint phyno;
+ int pass;
+ uint phytype;
+ int phyaddr;
+
+ phyaddr = -1; /* didn't find a PHY yet */
+ for (pass = 1; pass <= MAX_PHY_PASSES && phyaddr < 0; ++pass) {
+ if (pass > 1) {
+ /* PHY may need more time to recover from reset.
+ * The LXT970 needs 50ms typical, no maximum is
+ * specified, so wait 10ms before try again.
+ * With 11 passes this gives it 100ms to wake up.
+ */
+ udelay(10000); /* wait 10ms */
+ }
+ for (phyno = 0; phyno < 32 && phyaddr < 0; ++phyno) {
+ phytype = mii_send(mk_mii_read(phyno, MII_PHYSID2));
+ if (phytype != 0xffff) {
+ phyaddr = phyno;
+ phytype |= mii_send(mk_mii_read(phyno,
+ MII_PHYSID1)) << 16;
+ }
+ }
+ }
+ if (phyaddr < 0) {
+ printf("No PHY device found.\n");
+ }
+ return phyaddr;
+}
+#endif /* CONFIG_SYS_DISCOVER_PHY */
+
+#if (defined(CONFIG_MII) || defined(CONFIG_CMD_MII)) && !defined(CONFIG_BITBANGMII)
+
+/****************************************************************************
+ * mii_init -- Initialize the MII via FEC 1 for MII command without ethernet
+ * This function is a subset of eth_init
+ ****************************************************************************
+ */
+static void __mii_init(void)
+{
+ volatile immap_t *immr = (immap_t *) CONFIG_SYS_IMMR;
+ volatile fec_t *fecp = &(immr->im_cpm.cp_fec);
+
+ if (fec_reset(fecp) < 0)
+ printf ("FEC_RESET_DELAY timeout\n");
+
+ /* We use strictly polling mode only
+ */
+ fecp->fec_imask = 0;
+
+ /* Clear any pending interrupt
+ */
+ fecp->fec_ievent = 0xffc0;
+
+ /* Now enable the transmit and receive processing
+ */
+ fecp->fec_ecntrl = FEC_ECNTRL_PINMUX | FEC_ECNTRL_ETHER_EN;
+}
+
+void mii_init (void)
+{
+ int i;
+
+ __mii_init();
+
+ /* Setup the pin configuration of the FEC(s)
+ */
+ for (i = 0; i < ARRAY_SIZE(ether_fcc_info); i++)
+ fec_pin_init(ether_fcc_info[i].ether_index);
+}
+
+/*****************************************************************************
+ * Read and write a MII PHY register, routines used by MII Utilities
+ *
+ * FIXME: These routines are expected to return 0 on success, but mii_send
+ * does _not_ return an error code. Maybe 0xFFFF means error, i.e.
+ * no PHY connected...
+ * For now always return 0.
+ * FIXME: These routines only work after calling eth_init() at least once!
+ * Otherwise they hang in mii_send() !!! Sorry!
+ *****************************************************************************/
+
+int fec8xx_miiphy_read(struct mii_dev *bus, int addr, int devad, int reg)
+{
+ unsigned short value = 0;
+ short rdreg; /* register working value */
+
+ rdreg = mii_send(mk_mii_read(addr, reg));
+
+ value = rdreg;
+ return value;
+}
+
+int fec8xx_miiphy_write(struct mii_dev *bus, int addr, int devad, int reg,
+ u16 value)
+{
+ (void)mii_send(mk_mii_write(addr, reg, value));
+
+ return 0;
+}
+#endif
+
+#endif