u-boot/drivers/net/ks8851_mll.c
Joe Hershberger 1fd92db83d net: cosmetic: Fix var naming net <-> eth drivers
Update the naming convention used in the network stack functions and
variables that Ethernet drivers use to interact with it.

This cleans up the temporary hacks that were added to this interface
along with the DM support.

This patch has a few remaining checkpatch.pl failures that would be out
of the scope of this patch to fix (drivers that are in gross violation
of checkpatch.pl).

Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
2015-04-18 11:11:33 -06:00

633 lines
15 KiB
C

/*
* Micrel KS8851_MLL 16bit Network driver
* Copyright (c) 2011 Roberto Cerati <roberto.cerati@bticino.it>
*
* SPDX-License-Identifier: GPL-2.0+
*/
#include <asm/io.h>
#include <common.h>
#include <command.h>
#include <malloc.h>
#include <net.h>
#include <miiphy.h>
#include "ks8851_mll.h"
#define DRIVERNAME "ks8851_mll"
#define MAX_RECV_FRAMES 32
#define MAX_BUF_SIZE 2048
#define TX_BUF_SIZE 2000
#define RX_BUF_SIZE 2000
static const struct chip_id chip_ids[] = {
{CIDER_ID, "KSZ8851"},
{0, NULL},
};
/*
* union ks_tx_hdr - tx header data
* @txb: The header as bytes
* @txw: The header as 16bit, little-endian words
*
* A dual representation of the tx header data to allow
* access to individual bytes, and to allow 16bit accesses
* with 16bit alignment.
*/
union ks_tx_hdr {
u8 txb[4];
__le16 txw[2];
};
/*
* struct ks_net - KS8851 driver private data
* @net_device : The network device we're bound to
* @txh : temporaly buffer to save status/length.
* @frame_head_info : frame header information for multi-pkt rx.
* @statelock : Lock on this structure for tx list.
* @msg_enable : The message flags controlling driver output (see ethtool).
* @frame_cnt : number of frames received.
* @bus_width : i/o bus width.
* @irq : irq number assigned to this device.
* @rc_rxqcr : Cached copy of KS_RXQCR.
* @rc_txcr : Cached copy of KS_TXCR.
* @rc_ier : Cached copy of KS_IER.
* @sharedbus : Multipex(addr and data bus) mode indicator.
* @cmd_reg_cache : command register cached.
* @cmd_reg_cache_int : command register cached. Used in the irq handler.
* @promiscuous : promiscuous mode indicator.
* @all_mcast : mutlicast indicator.
* @mcast_lst_size : size of multicast list.
* @mcast_lst : multicast list.
* @mcast_bits : multicast enabed.
* @mac_addr : MAC address assigned to this device.
* @fid : frame id.
* @extra_byte : number of extra byte prepended rx pkt.
* @enabled : indicator this device works.
*/
/* Receive multiplex framer header info */
struct type_frame_head {
u16 sts; /* Frame status */
u16 len; /* Byte count */
} fr_h_i[MAX_RECV_FRAMES];
struct ks_net {
struct net_device *netdev;
union ks_tx_hdr txh;
struct type_frame_head *frame_head_info;
u32 msg_enable;
u32 frame_cnt;
int bus_width;
int irq;
u16 rc_rxqcr;
u16 rc_txcr;
u16 rc_ier;
u16 sharedbus;
u16 cmd_reg_cache;
u16 cmd_reg_cache_int;
u16 promiscuous;
u16 all_mcast;
u16 mcast_lst_size;
u8 mcast_lst[MAX_MCAST_LST][MAC_ADDR_LEN];
u8 mcast_bits[HW_MCAST_SIZE];
u8 mac_addr[6];
u8 fid;
u8 extra_byte;
u8 enabled;
} ks_str, *ks;
#define BE3 0x8000 /* Byte Enable 3 */
#define BE2 0x4000 /* Byte Enable 2 */
#define BE1 0x2000 /* Byte Enable 1 */
#define BE0 0x1000 /* Byte Enable 0 */
static u8 ks_rdreg8(struct eth_device *dev, u16 offset)
{
u8 shift_bit = offset & 0x03;
u8 shift_data = (offset & 1) << 3;
writew(offset | (BE0 << shift_bit), dev->iobase + 2);
return (u8)(readw(dev->iobase) >> shift_data);
}
static u16 ks_rdreg16(struct eth_device *dev, u16 offset)
{
writew(offset | ((BE1 | BE0) << (offset & 0x02)), dev->iobase + 2);
return readw(dev->iobase);
}
static void ks_wrreg8(struct eth_device *dev, u16 offset, u8 val)
{
u8 shift_bit = (offset & 0x03);
u16 value_write = (u16)(val << ((offset & 1) << 3));
writew(offset | (BE0 << shift_bit), dev->iobase + 2);
writew(value_write, dev->iobase);
}
static void ks_wrreg16(struct eth_device *dev, u16 offset, u16 val)
{
writew(offset | ((BE1 | BE0) << (offset & 0x02)), dev->iobase + 2);
writew(val, dev->iobase);
}
/*
* ks_inblk - read a block of data from QMU. This is called after sudo DMA mode
* enabled.
* @ks: The chip state
* @wptr: buffer address to save data
* @len: length in byte to read
*/
static inline void ks_inblk(struct eth_device *dev, u16 *wptr, u32 len)
{
len >>= 1;
while (len--)
*wptr++ = readw(dev->iobase);
}
/*
* ks_outblk - write data to QMU. This is called after sudo DMA mode enabled.
* @ks: The chip information
* @wptr: buffer address
* @len: length in byte to write
*/
static inline void ks_outblk(struct eth_device *dev, u16 *wptr, u32 len)
{
len >>= 1;
while (len--)
writew(*wptr++, dev->iobase);
}
static void ks_enable_int(struct eth_device *dev)
{
ks_wrreg16(dev, KS_IER, ks->rc_ier);
}
static void ks_set_powermode(struct eth_device *dev, unsigned pwrmode)
{
unsigned pmecr;
ks_rdreg16(dev, KS_GRR);
pmecr = ks_rdreg16(dev, KS_PMECR);
pmecr &= ~PMECR_PM_MASK;
pmecr |= pwrmode;
ks_wrreg16(dev, KS_PMECR, pmecr);
}
/*
* ks_read_config - read chip configuration of bus width.
* @ks: The chip information
*/
static void ks_read_config(struct eth_device *dev)
{
u16 reg_data = 0;
/* Regardless of bus width, 8 bit read should always work. */
reg_data = ks_rdreg8(dev, KS_CCR) & 0x00FF;
reg_data |= ks_rdreg8(dev, KS_CCR + 1) << 8;
/* addr/data bus are multiplexed */
ks->sharedbus = (reg_data & CCR_SHARED) == CCR_SHARED;
/*
* There are garbage data when reading data from QMU,
* depending on bus-width.
*/
if (reg_data & CCR_8BIT) {
ks->bus_width = ENUM_BUS_8BIT;
ks->extra_byte = 1;
} else if (reg_data & CCR_16BIT) {
ks->bus_width = ENUM_BUS_16BIT;
ks->extra_byte = 2;
} else {
ks->bus_width = ENUM_BUS_32BIT;
ks->extra_byte = 4;
}
}
/*
* ks_soft_reset - issue one of the soft reset to the device
* @ks: The device state.
* @op: The bit(s) to set in the GRR
*
* Issue the relevant soft-reset command to the device's GRR register
* specified by @op.
*
* Note, the delays are in there as a caution to ensure that the reset
* has time to take effect and then complete. Since the datasheet does
* not currently specify the exact sequence, we have chosen something
* that seems to work with our device.
*/
static void ks_soft_reset(struct eth_device *dev, unsigned op)
{
/* Disable interrupt first */
ks_wrreg16(dev, KS_IER, 0x0000);
ks_wrreg16(dev, KS_GRR, op);
mdelay(10); /* wait a short time to effect reset */
ks_wrreg16(dev, KS_GRR, 0);
mdelay(1); /* wait for condition to clear */
}
void ks_enable_qmu(struct eth_device *dev)
{
u16 w;
w = ks_rdreg16(dev, KS_TXCR);
/* Enables QMU Transmit (TXCR). */
ks_wrreg16(dev, KS_TXCR, w | TXCR_TXE);
/* Enable RX Frame Count Threshold and Auto-Dequeue RXQ Frame */
w = ks_rdreg16(dev, KS_RXQCR);
ks_wrreg16(dev, KS_RXQCR, w | RXQCR_RXFCTE);
/* Enables QMU Receive (RXCR1). */
w = ks_rdreg16(dev, KS_RXCR1);
ks_wrreg16(dev, KS_RXCR1, w | RXCR1_RXE);
}
static void ks_disable_qmu(struct eth_device *dev)
{
u16 w;
w = ks_rdreg16(dev, KS_TXCR);
/* Disables QMU Transmit (TXCR). */
w &= ~TXCR_TXE;
ks_wrreg16(dev, KS_TXCR, w);
/* Disables QMU Receive (RXCR1). */
w = ks_rdreg16(dev, KS_RXCR1);
w &= ~RXCR1_RXE;
ks_wrreg16(dev, KS_RXCR1, w);
}
static inline void ks_read_qmu(struct eth_device *dev, u16 *buf, u32 len)
{
u32 r = ks->extra_byte & 0x1;
u32 w = ks->extra_byte - r;
/* 1. set sudo DMA mode */
ks_wrreg16(dev, KS_RXFDPR, RXFDPR_RXFPAI);
ks_wrreg8(dev, KS_RXQCR, (ks->rc_rxqcr | RXQCR_SDA) & 0xff);
/*
* 2. read prepend data
*
* read 4 + extra bytes and discard them.
* extra bytes for dummy, 2 for status, 2 for len
*/
if (r)
ks_rdreg8(dev, 0);
ks_inblk(dev, buf, w + 2 + 2);
/* 3. read pkt data */
ks_inblk(dev, buf, ALIGN(len, 4));
/* 4. reset sudo DMA Mode */
ks_wrreg8(dev, KS_RXQCR, (ks->rc_rxqcr & ~RXQCR_SDA) & 0xff);
}
static void ks_rcv(struct eth_device *dev, uchar **pv_data)
{
struct type_frame_head *frame_hdr = ks->frame_head_info;
int i;
ks->frame_cnt = ks_rdreg16(dev, KS_RXFCTR) >> 8;
/* read all header information */
for (i = 0; i < ks->frame_cnt; i++) {
/* Checking Received packet status */
frame_hdr->sts = ks_rdreg16(dev, KS_RXFHSR);
/* Get packet len from hardware */
frame_hdr->len = ks_rdreg16(dev, KS_RXFHBCR);
frame_hdr++;
}
frame_hdr = ks->frame_head_info;
while (ks->frame_cnt--) {
if ((frame_hdr->sts & RXFSHR_RXFV) &&
(frame_hdr->len < RX_BUF_SIZE) &&
frame_hdr->len) {
/* read data block including CRC 4 bytes */
ks_read_qmu(dev, (u16 *)(*pv_data), frame_hdr->len);
/* net_rx_packets buffer size is ok (*pv_data) */
net_process_received_packet(*pv_data, frame_hdr->len);
pv_data++;
} else {
ks_wrreg16(dev, KS_RXQCR, (ks->rc_rxqcr | RXQCR_RRXEF));
printf(DRIVERNAME ": bad packet\n");
}
frame_hdr++;
}
}
/*
* ks_read_selftest - read the selftest memory info.
* @ks: The device state
*
* Read and check the TX/RX memory selftest information.
*/
static int ks_read_selftest(struct eth_device *dev)
{
u16 both_done = MBIR_TXMBF | MBIR_RXMBF;
u16 mbir;
int ret = 0;
mbir = ks_rdreg16(dev, KS_MBIR);
if ((mbir & both_done) != both_done) {
printf(DRIVERNAME ": Memory selftest not finished\n");
return 0;
}
if (mbir & MBIR_TXMBFA) {
printf(DRIVERNAME ": TX memory selftest fails\n");
ret |= 1;
}
if (mbir & MBIR_RXMBFA) {
printf(DRIVERNAME ": RX memory selftest fails\n");
ret |= 2;
}
debug(DRIVERNAME ": the selftest passes\n");
return ret;
}
static void ks_setup(struct eth_device *dev)
{
u16 w;
/* Setup Transmit Frame Data Pointer Auto-Increment (TXFDPR) */
ks_wrreg16(dev, KS_TXFDPR, TXFDPR_TXFPAI);
/* Setup Receive Frame Data Pointer Auto-Increment */
ks_wrreg16(dev, KS_RXFDPR, RXFDPR_RXFPAI);
/* Setup Receive Frame Threshold - 1 frame (RXFCTFC) */
ks_wrreg16(dev, KS_RXFCTR, 1 & RXFCTR_THRESHOLD_MASK);
/* Setup RxQ Command Control (RXQCR) */
ks->rc_rxqcr = RXQCR_CMD_CNTL;
ks_wrreg16(dev, KS_RXQCR, ks->rc_rxqcr);
/*
* set the force mode to half duplex, default is full duplex
* because if the auto-negotiation fails, most switch uses
* half-duplex.
*/
w = ks_rdreg16(dev, KS_P1MBCR);
w &= ~P1MBCR_FORCE_FDX;
ks_wrreg16(dev, KS_P1MBCR, w);
w = TXCR_TXFCE | TXCR_TXPE | TXCR_TXCRC | TXCR_TCGIP;
ks_wrreg16(dev, KS_TXCR, w);
w = RXCR1_RXFCE | RXCR1_RXBE | RXCR1_RXUE | RXCR1_RXME | RXCR1_RXIPFCC;
/* Normal mode */
w |= RXCR1_RXPAFMA;
ks_wrreg16(dev, KS_RXCR1, w);
}
static void ks_setup_int(struct eth_device *dev)
{
ks->rc_ier = 0x00;
/* Clear the interrupts status of the hardware. */
ks_wrreg16(dev, KS_ISR, 0xffff);
/* Enables the interrupts of the hardware. */
ks->rc_ier = (IRQ_LCI | IRQ_TXI | IRQ_RXI);
}
static int ks8851_mll_detect_chip(struct eth_device *dev)
{
unsigned short val, i;
ks_read_config(dev);
val = ks_rdreg16(dev, KS_CIDER);
if (val == 0xffff) {
/* Special case -- no chip present */
printf(DRIVERNAME ": is chip mounted ?\n");
return -1;
} else if ((val & 0xfff0) != CIDER_ID) {
printf(DRIVERNAME ": Invalid chip id 0x%04x\n", val);
return -1;
}
debug("Read back KS8851 id 0x%x\n", val);
/* only one entry in the table */
val &= 0xfff0;
for (i = 0; chip_ids[i].id != 0; i++) {
if (chip_ids[i].id == val)
break;
}
if (!chip_ids[i].id) {
printf(DRIVERNAME ": Unknown chip ID %04x\n", val);
return -1;
}
dev->priv = (void *)&chip_ids[i];
return 0;
}
static void ks8851_mll_reset(struct eth_device *dev)
{
/* wake up powermode to normal mode */
ks_set_powermode(dev, PMECR_PM_NORMAL);
mdelay(1); /* wait for normal mode to take effect */
/* Disable interrupt and reset */
ks_soft_reset(dev, GRR_GSR);
/* turn off the IRQs and ack any outstanding */
ks_wrreg16(dev, KS_IER, 0x0000);
ks_wrreg16(dev, KS_ISR, 0xffff);
/* shutdown RX/TX QMU */
ks_disable_qmu(dev);
}
static void ks8851_mll_phy_configure(struct eth_device *dev)
{
u16 data;
ks_setup(dev);
ks_setup_int(dev);
/* Probing the phy */
data = ks_rdreg16(dev, KS_OBCR);
ks_wrreg16(dev, KS_OBCR, data | OBCR_ODS_16MA);
debug(DRIVERNAME ": phy initialized\n");
}
static void ks8851_mll_enable(struct eth_device *dev)
{
ks_wrreg16(dev, KS_ISR, 0xffff);
ks_enable_int(dev);
ks_enable_qmu(dev);
}
static int ks8851_mll_init(struct eth_device *dev, bd_t *bd)
{
struct chip_id *id = dev->priv;
debug(DRIVERNAME ": detected %s controller\n", id->name);
if (ks_read_selftest(dev)) {
printf(DRIVERNAME ": Selftest failed\n");
return -1;
}
ks8851_mll_reset(dev);
/* Configure the PHY, initialize the link state */
ks8851_mll_phy_configure(dev);
/* static allocation of private informations */
ks->frame_head_info = fr_h_i;
/* Turn on Tx + Rx */
ks8851_mll_enable(dev);
return 0;
}
static void ks_write_qmu(struct eth_device *dev, u8 *pdata, u16 len)
{
/* start header at txb[0] to align txw entries */
ks->txh.txw[0] = 0;
ks->txh.txw[1] = cpu_to_le16(len);
/* 1. set sudo-DMA mode */
ks_wrreg16(dev, KS_TXFDPR, TXFDPR_TXFPAI);
ks_wrreg8(dev, KS_RXQCR, (ks->rc_rxqcr | RXQCR_SDA) & 0xff);
/* 2. write status/lenth info */
ks_outblk(dev, ks->txh.txw, 4);
/* 3. write pkt data */
ks_outblk(dev, (u16 *)pdata, ALIGN(len, 4));
/* 4. reset sudo-DMA mode */
ks_wrreg8(dev, KS_RXQCR, (ks->rc_rxqcr & ~RXQCR_SDA) & 0xff);
/* 5. Enqueue Tx(move the pkt from TX buffer into TXQ) */
ks_wrreg16(dev, KS_TXQCR, TXQCR_METFE);
/* 6. wait until TXQCR_METFE is auto-cleared */
do { } while (ks_rdreg16(dev, KS_TXQCR) & TXQCR_METFE);
}
static int ks8851_mll_send(struct eth_device *dev, void *packet, int length)
{
u8 *data = (u8 *)packet;
u16 tmplen = (u16)length;
u16 retv;
/*
* Extra space are required:
* 4 byte for alignment, 4 for status/length, 4 for CRC
*/
retv = ks_rdreg16(dev, KS_TXMIR) & 0x1fff;
if (retv >= tmplen + 12) {
ks_write_qmu(dev, data, tmplen);
return 0;
} else {
printf(DRIVERNAME ": failed to send packet: No buffer\n");
return -1;
}
}
static void ks8851_mll_halt(struct eth_device *dev)
{
ks8851_mll_reset(dev);
}
/*
* Maximum receive ring size; that is, the number of packets
* we can buffer before overflow happens. Basically, this just
* needs to be enough to prevent a packet being discarded while
* we are processing the previous one.
*/
static int ks8851_mll_recv(struct eth_device *dev)
{
u16 status;
status = ks_rdreg16(dev, KS_ISR);
ks_wrreg16(dev, KS_ISR, status);
if ((status & IRQ_RXI))
ks_rcv(dev, (uchar **)net_rx_packets);
if ((status & IRQ_LDI)) {
u16 pmecr = ks_rdreg16(dev, KS_PMECR);
pmecr &= ~PMECR_WKEVT_MASK;
ks_wrreg16(dev, KS_PMECR, pmecr | PMECR_WKEVT_LINK);
}
return 0;
}
static int ks8851_mll_write_hwaddr(struct eth_device *dev)
{
u16 addrl, addrm, addrh;
addrh = (dev->enetaddr[0] << 8) | dev->enetaddr[1];
addrm = (dev->enetaddr[2] << 8) | dev->enetaddr[3];
addrl = (dev->enetaddr[4] << 8) | dev->enetaddr[5];
ks_wrreg16(dev, KS_MARH, addrh);
ks_wrreg16(dev, KS_MARM, addrm);
ks_wrreg16(dev, KS_MARL, addrl);
return 0;
}
int ks8851_mll_initialize(u8 dev_num, int base_addr)
{
struct eth_device *dev;
dev = malloc(sizeof(*dev));
if (!dev) {
printf("Error: Failed to allocate memory\n");
return -1;
}
memset(dev, 0, sizeof(*dev));
dev->iobase = base_addr;
ks = &ks_str;
/* Try to detect chip. Will fail if not present. */
if (ks8851_mll_detect_chip(dev)) {
free(dev);
return -1;
}
dev->init = ks8851_mll_init;
dev->halt = ks8851_mll_halt;
dev->send = ks8851_mll_send;
dev->recv = ks8851_mll_recv;
dev->write_hwaddr = ks8851_mll_write_hwaddr;
sprintf(dev->name, "%s-%hu", DRIVERNAME, dev_num);
eth_register(dev);
return 0;
}