nuttx/drivers/net/enc28j60.c

2496 lines
75 KiB
C
Raw Normal View History

/****************************************************************************
* drivers/net/enc28j60.c
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership. The
* ASF licenses this file to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance with the
* License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
* License for the specific language governing permissions and limitations
* under the License.
*
****************************************************************************/
/* References:
* - ENC28J60 Data Sheet, Stand-Alone Ethernet Controller with SPI Interface,
* DS39662C, 2008 Microchip Technology Inc.
*/
/****************************************************************************
* Included Files
****************************************************************************/
#include <nuttx/config.h>
#if defined(CONFIG_NET) && defined(CONFIG_ENC28J60)
#include <stdint.h>
#include <stdbool.h>
#include <stdint.h>
#include <time.h>
#include <string.h>
#include <assert.h>
#include <debug.h>
#include <errno.h>
#include <arpa/inet.h>
#include <nuttx/irq.h>
#include <nuttx/arch.h>
#include <nuttx/wdog.h>
#include <nuttx/spi/spi.h>
#include <nuttx/wqueue.h>
#include <nuttx/clock.h>
#include <nuttx/net/enc28j60.h>
2014-07-05 00:38:51 +02:00
#include <nuttx/net/net.h>
#include <nuttx/net/ip.h>
#include <nuttx/net/netdev.h>
#ifdef CONFIG_NET_PKT
# include <nuttx/net/pkt.h>
#endif
#include "enc28j60.h"
/****************************************************************************
* Pre-processor Definitions
****************************************************************************/
/* Configuration ************************************************************/
/* ENC28J60 Configuration Settings:
*
* CONFIG_ENC28J60 - Enabled ENC28J60 support
* CONFIG_ENC28J60_SPIMODE - Controls the SPI mode
* CONFIG_ENC28J60_FREQUENCY - Define to use a different bus frequency
* CONFIG_ENC28J60_NINTERFACES - Specifies the number of physical ENC28J60
* devices that will be supported.
* CONFIG_ENC28J60_HALFDUPPLEX - Default is full duplex
*/
/* The ENC28J60 spec says that it supports SPI mode 0,0 only: "The
* implementation used on this device supports SPI mode 0,0 only. In
* addition, the SPI port requires that SCK be at Idle in a low state;
* selectable clock polarity is not supported." However, sometimes you
* need to tinker with these things.
*/
#ifndef CONFIG_ENC28J60_SPIMODE
# define CONFIG_ENC28J60_SPIMODE SPIDEV_MODE0
#endif
/* CONFIG_ENC28J60_NINTERFACES determines the number of physical interfaces
* that will be supported.
*/
#ifndef CONFIG_ENC28J60_NINTERFACES
# define CONFIG_ENC28J60_NINTERFACES 1
#endif
This commit attempts remove some long standard confusion in naming and some actual problems that result from the naming confusion. The basic problem is the standard MTU does not include the size of the Ethernet header. For clarity, I changed the naming of most things called MTU to PKTSIZE. For example, CONFIG_NET_ETH_MTU is now CONFIG_NET_ETH_PKTSIZE. This makes the user interface a little hostile. People thing of an MTU of 1500 bytes, but the corresponding packet is really 1514 bytes (including the 14 byte Ethernet header). A more friendly solution would configure the MTU (as before), but then derive the packet buffer size by adding the MAC header length. Instead, we define the packet buffer size then derive the MTU. The MTU is not common currency in networking. On the wire, the only real issue is the MSS which is derived from MTU by subtracting the IP header and TCP header sizes (for the case of TCP). Now it is derived for the PKTSIZE by subtracting the IP header, the TCP header, and the MAC header sizes. So we should be all good and without the recurring 14 byte error in MTU's and MSS's. Squashed commit of the following: Trivial update to fix some spacing issues. net/: Rename several macros containing _MTU to _PKTSIZE. net/: Rename CONFIG_NET_SLIP_MTU to CONFIG_NET_SLIP_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_6LOWPAN_MTU to CONFIG_NET_6LOWPAN_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_ETH_MTU to CONFIG_NET_ETH_PKTSIZE. This is not the MTU which does not include the size of the link layer header. This is the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename the file d_mtu in the network driver structure to d_pktsize. That value saved there is not the MTU. The packetsize is the memory large enough to hold the maximum packet PLUS the size of the link layer header. The MTU does not include the link layer header.
2018-07-04 22:10:40 +02:00
/* CONFIG_NET_ETH_PKTSIZE must always be defined */
This commit attempts remove some long standard confusion in naming and some actual problems that result from the naming confusion. The basic problem is the standard MTU does not include the size of the Ethernet header. For clarity, I changed the naming of most things called MTU to PKTSIZE. For example, CONFIG_NET_ETH_MTU is now CONFIG_NET_ETH_PKTSIZE. This makes the user interface a little hostile. People thing of an MTU of 1500 bytes, but the corresponding packet is really 1514 bytes (including the 14 byte Ethernet header). A more friendly solution would configure the MTU (as before), but then derive the packet buffer size by adding the MAC header length. Instead, we define the packet buffer size then derive the MTU. The MTU is not common currency in networking. On the wire, the only real issue is the MSS which is derived from MTU by subtracting the IP header and TCP header sizes (for the case of TCP). Now it is derived for the PKTSIZE by subtracting the IP header, the TCP header, and the MAC header sizes. So we should be all good and without the recurring 14 byte error in MTU's and MSS's. Squashed commit of the following: Trivial update to fix some spacing issues. net/: Rename several macros containing _MTU to _PKTSIZE. net/: Rename CONFIG_NET_SLIP_MTU to CONFIG_NET_SLIP_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_6LOWPAN_MTU to CONFIG_NET_6LOWPAN_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_ETH_MTU to CONFIG_NET_ETH_PKTSIZE. This is not the MTU which does not include the size of the link layer header. This is the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename the file d_mtu in the network driver structure to d_pktsize. That value saved there is not the MTU. The packetsize is the memory large enough to hold the maximum packet PLUS the size of the link layer header. The MTU does not include the link layer header.
2018-07-04 22:10:40 +02:00
#if !defined(CONFIG_NET_ETH_PKTSIZE) && (CONFIG_NET_ETH_PKTSIZE <= MAX_FRAMELEN)
# error "CONFIG_NET_ETH_PKTSIZE is not valid for the ENC28J60"
#endif
/* We need to have the work queue to handle SPI interrupts */
#if !defined(CONFIG_SCHED_WORKQUEUE)
# error "Worker thread support is required (CONFIG_SCHED_WORKQUEUE)"
#endif
/* The low priority work queue is preferred. If it is not enabled, LPWORK
* will be the same as HPWORK.
*
* NOTE: However, the network should NEVER run on the high priority work
* queue! That queue is intended only to service short back end interrupt
* processing that never suspends. Suspending the high priority work queue
* may bring the system to its knees!
*/
#define ENCWORK LPWORK
/* CONFIG_ENC28J60_DUMPPACKET will dump the contents of each packet. */
#ifdef CONFIG_ENC28J60_DUMPPACKET
# define enc_dumppacket(m,a,n) lib_dumpbuffer(m,a,n)
#else
# define enc_dumppacket(m,a,n)
#endif
/* Low-level register debug */
#if !defined(CONFIG_DEBUG_FEATURES) || !defined(CONFIG_DEBUG_NET)
# undef CONFIG_ENC28J60_REGDEBUG
#endif
/* Timing *******************************************************************/
/* TX timeout = 1 minute */
#define ENC_TXTIMEOUT (60*CLK_TCK)
/* Poll timeout */
#define ENC_POLLTIMEOUT MSEC2TICK(50)
/* Packet Memory ************************************************************/
/* Packet memory layout */
This commit attempts remove some long standard confusion in naming and some actual problems that result from the naming confusion. The basic problem is the standard MTU does not include the size of the Ethernet header. For clarity, I changed the naming of most things called MTU to PKTSIZE. For example, CONFIG_NET_ETH_MTU is now CONFIG_NET_ETH_PKTSIZE. This makes the user interface a little hostile. People thing of an MTU of 1500 bytes, but the corresponding packet is really 1514 bytes (including the 14 byte Ethernet header). A more friendly solution would configure the MTU (as before), but then derive the packet buffer size by adding the MAC header length. Instead, we define the packet buffer size then derive the MTU. The MTU is not common currency in networking. On the wire, the only real issue is the MSS which is derived from MTU by subtracting the IP header and TCP header sizes (for the case of TCP). Now it is derived for the PKTSIZE by subtracting the IP header, the TCP header, and the MAC header sizes. So we should be all good and without the recurring 14 byte error in MTU's and MSS's. Squashed commit of the following: Trivial update to fix some spacing issues. net/: Rename several macros containing _MTU to _PKTSIZE. net/: Rename CONFIG_NET_SLIP_MTU to CONFIG_NET_SLIP_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_6LOWPAN_MTU to CONFIG_NET_6LOWPAN_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_ETH_MTU to CONFIG_NET_ETH_PKTSIZE. This is not the MTU which does not include the size of the link layer header. This is the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename the file d_mtu in the network driver structure to d_pktsize. That value saved there is not the MTU. The packetsize is the memory large enough to hold the maximum packet PLUS the size of the link layer header. The MTU does not include the link layer header.
2018-07-04 22:10:40 +02:00
#define ALIGNED_BUFSIZE ((CONFIG_NET_ETH_PKTSIZE + 255) & ~255)
2013-06-23 01:01:44 +02:00
/* Work around Errata #5 (spurious reset of ERXWRPT to 0) by placing the RX
* FIFO at the beginning of packet memory.
*/
#define ERRATA5 1
#if ERRATA5
# define PKTMEM_RX_START 0x0000 /* RX buffer must be at addr 0 for errata 5 */
# define PKTMEM_RX_END (PKTMEM_END-ALIGNED_BUFSIZE) /* RX buffer length is total SRAM minus TX buffer */
# define PKTMEM_TX_START (PKTMEM_RX_END+1) /* Start TX buffer after */
# define PKTMEM_TX_ENDP1 (PKTMEM_TX_START+ALIGNED_BUFSIZE) /* Allow TX buffer for one frame */
2013-06-22 16:55:16 +02:00
#else
2013-06-23 01:01:44 +02:00
# define PKTMEM_TX_START 0x0000 /* Start TX buffer at 0 */
# define PKTMEM_TX_ENDP1 ALIGNED_BUFSIZE /* Allow TX buffer for one frame */
# define PKTMEM_RX_START PKTMEM_TX_ENDP1 /* Followed by RX buffer */
# define PKTMEM_RX_END PKTMEM_END /* RX buffer goes to the end of SRAM */
2013-06-22 16:55:16 +02:00
#endif
/* Misc. Helper Macros ******************************************************/
#define enc_rdgreg(priv,ctrlreg) \
enc_rdgreg2(priv, ENC_RCR | GETADDR(ctrlreg))
#define enc_wrgreg(priv,ctrlreg,wrdata) \
enc_wrgreg2(priv, ENC_WCR | GETADDR(ctrlreg), wrdata)
#define enc_bfcgreg(priv,ctrlreg,clrbits) \
enc_wrgreg2(priv, ENC_BFC | GETADDR(ctrlreg), clrbits)
#define enc_bfsgreg(priv,ctrlreg,setbits) \
enc_wrgreg2(priv, ENC_BFS | GETADDR(ctrlreg), setbits)
/* Packet buffer size */
#define PKTBUF_SIZE (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE)
/* This is a helper pointer for accessing the contents of Ethernet header */
#define BUF ((FAR struct eth_hdr_s *)priv->dev.d_buf)
/* Debug ********************************************************************/
#ifdef CONFIG_ENC28J60_REGDEBUG
# define enc_wrdump(a,v) \
syslog(LOG_DEBUG, "ENC28J60: %02x<-%02x\n", a, v);
# define enc_rddump(a,v) \
syslog(LOG_DEBUG, "ENC28J60: %02x->%02x\n", a, v);
# define enc_cmddump(c) \
syslog(LOG_DEBUG, "ENC28J60: CMD: %02x\n", c);
# define enc_bmdump(c,b,s) \
syslog(LOG_DEBUG, "ENC28J60: CMD: %02x buffer: %p length: %d\n", c, b, s);
#else
# define enc_wrdump(a,v)
# define enc_rddump(a,v)
# define enc_cmddump(c)
# define enc_bmdump(c,b,s)
#endif
/****************************************************************************
* Private Types
****************************************************************************/
/* The state of the interface */
enum enc_state_e
{
ENCSTATE_UNINIT = 0, /* The interface is in an uninitialized state */
ENCSTATE_DOWN, /* The interface is down */
ENCSTATE_UP /* The interface is up */
};
/* The enc_driver_s encapsulates all state information for a single hardware
* interface
*/
struct enc_driver_s
{
/* Device control */
uint8_t ifstate; /* Interface state: See ENCSTATE_* */
uint8_t bank; /* Currently selected bank */
uint16_t nextpkt; /* Next packet address */
FAR const struct enc_lower_s *lower; /* Low-level MCU-specific support */
/* Timing */
struct wdog_s txtimeout; /* TX timeout timer */
/* If we don't own the SPI bus, then we cannot do SPI accesses from the
* interrupt handler.
*/
struct work_s irqwork; /* Interrupt continuation work queue support */
struct work_s towork; /* Tx timeout work queue support */
struct work_s pollwork; /* Poll timeout work queue support */
/* This is the contained SPI driver instance */
FAR struct spi_dev_s *spi;
/* This holds the information visible to the NuttX network */
struct net_driver_s dev; /* Interface understood by the network */
};
/****************************************************************************
* Private Data
****************************************************************************/
/* A single packet buffer is used */
static uint16_t g_pktbuf[CONFIG_ENC28J60_NINTERFACES][(PKTBUF_SIZE + 1) / 2];
/* Driver status structure */
static struct enc_driver_s g_enc28j60[CONFIG_ENC28J60_NINTERFACES];
/****************************************************************************
* Private Function Prototypes
****************************************************************************/
/* Low-level SPI helpers */
static inline void enc_configspi(FAR struct spi_dev_s *spi);
static void enc_lock(FAR struct enc_driver_s *priv);
static inline void enc_unlock(FAR struct enc_driver_s *priv);
/* SPI control register access */
static uint8_t enc_rdgreg2(FAR struct enc_driver_s *priv, uint8_t cmd);
static void enc_wrgreg2(FAR struct enc_driver_s *priv, uint8_t cmd,
uint8_t wrdata);
static inline void enc_src(FAR struct enc_driver_s *priv);
static void enc_setbank(FAR struct enc_driver_s *priv, uint8_t bank);
static uint8_t enc_rdbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg);
static void enc_wrbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg,
uint8_t wrdata);
static int enc_waitbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg,
uint8_t bits, uint8_t value);
#if 0 /* Sometimes useful */
static void enc_rxdump(FAR struct enc_driver_s *priv);
static void enc_txdump(FAR struct enc_driver_s *priv);
#endif
/* SPI buffer transfers */
static void enc_rdbuffer(FAR struct enc_driver_s *priv, FAR uint8_t *buffer,
size_t buflen);
static inline void enc_wrbuffer(FAR struct enc_driver_s *priv,
FAR const uint8_t *buffer, size_t buflen);
/* PHY register access */
static uint16_t enc_rdphy(FAR struct enc_driver_s *priv, uint8_t phyaddr);
static void enc_wrphy(FAR struct enc_driver_s *priv, uint8_t phyaddr,
uint16_t phydata);
/* Common TX logic */
static int enc_transmit(FAR struct enc_driver_s *priv);
static int enc_txpoll(struct net_driver_s *dev);
/* Interrupt handling */
static void enc_linkstatus(FAR struct enc_driver_s *priv);
static void enc_txif(FAR struct enc_driver_s *priv);
static void enc_txerif(FAR struct enc_driver_s *priv);
static void enc_txerif(FAR struct enc_driver_s *priv);
static void enc_rxerif(FAR struct enc_driver_s *priv);
static void enc_rxdispatch(FAR struct enc_driver_s *priv);
static void enc_pktif(FAR struct enc_driver_s *priv);
static void enc_irqworker(FAR void *arg);
static int enc_interrupt(int irq, FAR void *context, FAR void *arg);
/* Watchdog timer expirations */
static void enc_toworker(FAR void *arg);
static void enc_txtimeout(wdparm_t arg);
/* NuttX callback functions */
2014-06-28 00:48:12 +02:00
static int enc_ifup(struct net_driver_s *dev);
static int enc_ifdown(struct net_driver_s *dev);
static int enc_txavail(struct net_driver_s *dev);
#ifdef CONFIG_NET_MCASTGROUP
2014-06-28 00:48:12 +02:00
static int enc_addmac(struct net_driver_s *dev, FAR const uint8_t *mac);
static int enc_rmmac(struct net_driver_s *dev, FAR const uint8_t *mac);
#endif
/* Initialization */
static void enc_pwrsave(FAR struct enc_driver_s *priv);
static void enc_pwrfull(FAR struct enc_driver_s *priv);
static void enc_setmacaddr(FAR struct enc_driver_s *priv);
static int enc_reset(FAR struct enc_driver_s *priv);
/****************************************************************************
* Private Functions
****************************************************************************/
/****************************************************************************
* Name: enc_configspi
*
* Description:
* Configure the SPI for use with the ENC28J60
*
* Input Parameters:
* spi - Reference to the SPI driver structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static inline void enc_configspi(FAR struct spi_dev_s *spi)
{
/* Configure SPI for the ENC28J60. */
SPI_SETMODE(spi, CONFIG_ENC28J60_SPIMODE);
SPI_SETBITS(spi, 8);
SPI_HWFEATURES(spi, 0);
SPI_SETFREQUENCY(spi, CONFIG_ENC28J60_FREQUENCY);
}
/****************************************************************************
* Name: enc_lock
*
* Description:
* Select the SPI, locking and re-configuring if necessary
*
* Input Parameters:
* spi - Reference to the SPI driver structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_lock(FAR struct enc_driver_s *priv)
{
/* Lock the SPI bus in case there are multiple devices competing for the
* SPI bus.
*/
SPI_LOCK(priv->spi, true);
/* Now make sure that the SPI bus is configured for the ENC28J60 (it
* might have gotten configured for a different device while unlocked)
*/
SPI_SETMODE(priv->spi, CONFIG_ENC28J60_SPIMODE);
SPI_SETBITS(priv->spi, 8);
SPI_HWFEATURES(priv->spi, 0);
SPI_SETFREQUENCY(priv->spi, CONFIG_ENC28J60_FREQUENCY);
}
/****************************************************************************
* Name: enc_unlock
*
* Description:
* De-select the SPI
*
* Input Parameters:
* spi - Reference to the SPI driver structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static inline void enc_unlock(FAR struct enc_driver_s *priv)
{
/* Relinquish the lock on the bus. */
SPI_LOCK(priv->spi, false);
}
/****************************************************************************
* Name: enc_rdgreg2
*
* Description:
* Read a global register (EIE, EIR, ESTAT, ECON2, or ECON1). The cmd
* include the CMD 'OR'd with the global address register.
*
* Input Parameters:
* priv - Reference to the driver state structure
* cmd - The full command to received (cmd | address)
*
* Returned Value:
* The value read from the register
*
* Assumptions:
*
****************************************************************************/
static uint8_t enc_rdgreg2(FAR struct enc_driver_s *priv, uint8_t cmd)
{
uint8_t rddata;
DEBUGASSERT(priv && priv->spi);
/* Select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the read command and collect the data. The sequence requires
* 16-clocks: 8 to clock out the cmd + 8 to clock in the data.
*/
SPI_SEND(priv->spi, cmd); /* Clock out the command */
rddata = SPI_SEND(priv->spi, 0); /* Clock in the data */
/* De-select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_rddump(cmd, rddata);
return rddata;
}
/****************************************************************************
* Name: enc_wrgreg2
*
* Description:
* Write to a global register (EIE, EIR, ESTAT, ECON2, or ECON1). The cmd
* include the CMD 'OR'd with the global address register.
*
* Input Parameters:
* priv - Reference to the driver state structure
* cmd - The full command to received (cmd | address)
* wrdata - The data to send
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_wrgreg2(FAR struct enc_driver_s *priv, uint8_t cmd,
uint8_t wrdata)
{
DEBUGASSERT(priv && priv->spi);
/* Select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the write command and data. The sequence requires 16-clocks:
* 8 to clock out the cmd + 8 to clock out the data.
*/
SPI_SEND(priv->spi, cmd); /* Clock out the command */
SPI_SEND(priv->spi, wrdata); /* Clock out the data */
/* De-select ENC28J60 chip. */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_wrdump(cmd, wrdata);
}
/****************************************************************************
* Name: enc_src
*
* Description:
* Send the single byte system reset command (SRC).
*
* "The System Reset Command (SRC) allows the host controller to issue a
* System Soft Reset command. Unlike other SPI commands, the SRC is
* only a single byte command and does not operate on any register. The
* command is started by pulling the CS pin low. The SRC opcode is the
* sent, followed by a 5-bit Soft Reset command constant of 1Fh. The
* SRC operation is terminated by raising the CS pin."
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static inline void enc_src(FAR struct enc_driver_s *priv)
{
DEBUGASSERT(priv && priv->spi);
/* Select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the system reset command. */
SPI_SEND(priv->spi, ENC_SRC);
/* Check CLKRDY bit to see when the reset is complete. There is an errata
* that says the CLKRDY may be invalid. We'll wait a couple of msec to
* workaround this condition.
*
* Also, "After a System Reset, all PHY registers should not be read or
* written to until at least 50 us have passed since the Reset has ended.
* All registers will revert to their Reset default values. The dual
* port buffer memory will maintain state throughout the System Reset."
*/
up_mdelay(2);
#if 0
while ((enc_rdgreg(priv, ENC_ESTAT) & ESTAT_CLKRDY) != 0);
#endif
/* De-select ENC28J60 chip. */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_cmddump(ENC_SRC);
}
/****************************************************************************
* Name: enc_setbank
*
* Description:
* Set the bank for these next control register access.
*
* Assumption:
* The caller has exclusive access to the SPI bus
*
* Input Parameters:
* priv - Reference to the driver state structure
* bank - The bank to select (0-3)
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_setbank(FAR struct enc_driver_s *priv, uint8_t bank)
{
/* Check if the bank setting has changed */
if (bank != priv->bank)
{
/* Select bank 0 (just so that all of the bits are cleared) */
enc_bfcgreg(priv, ENC_ECON1, ECON1_BSEL_MASK);
/* Then OR in bits to get the correct bank */
if (bank != 0)
{
enc_bfsgreg(priv, ENC_ECON1, (bank << ECON1_BSEL_SHIFT));
}
/* Then remember the bank setting */
priv->bank = bank;
}
}
/****************************************************************************
* Name: enc_rdbreg
*
* Description:
* Read from a banked control register using the RCR command.
*
* Input Parameters:
* priv - Reference to the driver state structure
* ctrlreg - Bit encoded address of banked register to read
*
* Returned Value:
* The byte read from the banked register
*
* Assumptions:
*
****************************************************************************/
static uint8_t enc_rdbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg)
{
uint8_t rddata;
DEBUGASSERT(priv && priv->spi);
/* Set the bank */
enc_setbank(priv, GETBANK(ctrlreg));
/* Re-select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the RCR command and collect the data. How we collect the data
* depends on if this is a PHY/CAN or not. The normal sequence requires
* 16-clocks: 8 to clock out the cmd and 8 to clock in the data.
*/
SPI_SEND(priv->spi, ENC_RCR | GETADDR(ctrlreg)); /* Clock out the command */
if (ISPHYMAC(ctrlreg))
{
/* The PHY/MAC sequence requires 24-clocks: 8 to clock out the cmd,
* 8 dummy bits, and 8 to clock in the PHY/MAC data.
*/
SPI_SEND(priv->spi, 0); /* Clock in the dummy byte */
}
rddata = SPI_SEND(priv->spi, 0); /* Clock in the data */
/* De-select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_rddump(ENC_RCR | GETADDR(ctrlreg), rddata);
return rddata;
}
/****************************************************************************
* Name: enc_wrbreg
*
* Description:
* Write to a banked control register using the WCR command. Unlike
* reading, this same SPI sequence works for normal, MAC, and PHY
* registers.
*
* Input Parameters:
* priv - Reference to the driver state structure
* ctrlreg - Bit encoded address of banked register to write
* wrdata - The data to send
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_wrbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg,
uint8_t wrdata)
{
DEBUGASSERT(priv && priv->spi);
/* Set the bank */
enc_setbank(priv, GETBANK(ctrlreg));
/* Re-select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the WCR command and data. The sequence requires 16-clocks:
* 8 to clock out the cmd + 8 to clock out the data.
*/
SPI_SEND(priv->spi, ENC_WCR | GETADDR(ctrlreg)); /* Clock out the command */
SPI_SEND(priv->spi, wrdata); /* Clock out the data */
/* De-select ENC28J60 chip. */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_wrdump(ENC_WCR | GETADDR(ctrlreg), wrdata);
}
/****************************************************************************
* Name: enc_waitbreg
*
* Description:
* Wait until banked register bit(s) take a specific value (or a timeout
* occurs).
*
* Input Parameters:
* priv - Reference to the driver state structure
* ctrlreg - Bit encoded address of banked register to check
* bits - The bits to check (a mask)
* value - The value of the bits to return (value under mask)
*
* Returned Value:
* OK on success, negated errno on failure
*
* Assumptions:
*
****************************************************************************/
static int enc_waitbreg(FAR struct enc_driver_s *priv, uint8_t ctrlreg,
uint8_t bits, uint8_t value)
{
clock_t start = clock_systime_ticks();
clock_t elapsed;
uint8_t rddata;
/* Loop until the exit condition is met */
do
{
/* Read the byte from the requested banked register */
rddata = enc_rdbreg(priv, ctrlreg);
elapsed = clock_systime_ticks() - start;
}
while ((rddata & bits) != value && elapsed < ENC_POLLTIMEOUT);
return (rddata & bits) == value ? OK : -ETIMEDOUT;
}
/****************************************************************************
* Name: enc_txdump enc_rxdump
*
* Description:
* Dump registers associated with receiving or sending packets.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
#if 0 /* Sometimes useful */
static void enc_rxdump(FAR struct enc_driver_s *priv)
{
syslog(LOG_DEBUG, "Rx Registers:\n");
syslog(LOG_DEBUG, " EIE: %02x EIR: %02x\n",
enc_rdgreg(priv, ENC_EIE), enc_rdgreg(priv, ENC_EIR));
syslog(LOG_DEBUG, " ESTAT: %02x ECON1: %02x ECON2: %02x\n",
enc_rdgreg(priv, ENC_ESTAT), enc_rdgreg(priv, ENC_ECON1),
enc_rdgreg(priv, ENC_ECON2));
syslog(LOG_DEBUG, " ERXST: %02x %02x\n",
enc_rdbreg(priv, ENC_ERXSTH), enc_rdbreg(priv, ENC_ERXSTL));
syslog(LOG_DEBUG, " ERXND: %02x %02x\n",
enc_rdbreg(priv, ENC_ERXNDH), enc_rdbreg(priv, ENC_ERXNDL));
syslog(LOG_DEBUG, " ERXRDPT: %02x %02x\n",
enc_rdbreg(priv, ENC_ERXRDPTH), enc_rdbreg(priv, ENC_ERXRDPTL));
syslog(LOG_DEBUG, " ERXFCON: %02x EPKTCNT: %02x\n",
enc_rdbreg(priv, ENC_ERXFCON), enc_rdbreg(priv, ENC_EPKTCNT));
syslog(LOG_DEBUG, " MACON1: %02x MACON3: %02x\n",
enc_rdbreg(priv, ENC_MACON1), enc_rdbreg(priv, ENC_MACON3));
syslog(LOG_DEBUG, " MAMXFL: %02x %02x\n",
enc_rdbreg(priv, ENC_MAMXFLH), enc_rdbreg(priv, ENC_MAMXFLL));
syslog(LOG_DEBUG, " MAADR: %02x:%02x:%02x:%02x:%02x:%02x\n",
enc_rdbreg(priv, ENC_MAADR1), enc_rdbreg(priv, ENC_MAADR2),
enc_rdbreg(priv, ENC_MAADR3), enc_rdbreg(priv, ENC_MAADR4),
enc_rdbreg(priv, ENC_MAADR5), enc_rdbreg(priv, ENC_MAADR6));
}
#endif
#if 0 /* Sometimes useful */
static void enc_txdump(FAR struct enc_driver_s *priv)
{
syslog(LOG_DEBUG, "Tx Registers:\n");
syslog(LOG_DEBUG, " EIE: %02x EIR: %02x\n",
enc_rdgreg(priv, ENC_EIE), enc_rdgreg(priv, ENC_EIR));
syslog(LOG_DEBUG, " ESTAT: %02x ECON1: %02x\n",
enc_rdgreg(priv, ENC_ESTAT), enc_rdgreg(priv, ENC_ECON1));
syslog(LOG_DEBUG, " ETXST: %02x %02x\n",
enc_rdbreg(priv, ENC_ETXSTH), enc_rdbreg(priv, ENC_ETXSTL));
syslog(LOG_DEBUG, " ETXND: %02x %02x\n",
enc_rdbreg(priv, ENC_ETXNDH), enc_rdbreg(priv, ENC_ETXNDL));
syslog(LOG_DEBUG, " MACON1: %02x MACON3: %02x MACON4: %02x\n",
enc_rdbreg(priv, ENC_MACON1), enc_rdbreg(priv, ENC_MACON3),
enc_rdbreg(priv, ENC_MACON4));
syslog(LOG_DEBUG, " MACON1: %02x MACON3: %02x MACON4: %02x\n",
enc_rdbreg(priv, ENC_MACON1), enc_rdbreg(priv, ENC_MACON3),
enc_rdbreg(priv, ENC_MACON4));
syslog(LOG_DEBUG, " MABBIPG: %02x MAIPG %02x %02x\n",
enc_rdbreg(priv, ENC_MABBIPG), enc_rdbreg(priv, ENC_MAIPGH),
enc_rdbreg(priv, ENC_MAIPGL));
syslog(LOG_DEBUG, " MACLCON1: %02x MACLCON2: %02x\n",
enc_rdbreg(priv, ENC_MACLCON1), enc_rdbreg(priv, ENC_MACLCON2));
syslog(LOG_DEBUG, " MAMXFL: %02x %02x\n",
enc_rdbreg(priv, ENC_MAMXFLH), enc_rdbreg(priv, ENC_MAMXFLL));
}
#endif
/****************************************************************************
* Name: enc_rdbuffer
*
* Description:
* Read a buffer of data.
*
* Input Parameters:
* priv - Reference to the driver state structure
* buffer - A pointer to the buffer to read into
* buflen - The number of bytes to read
*
* Returned Value:
* None
*
* Assumptions:
* Read pointer is set to the correct address
*
****************************************************************************/
static void enc_rdbuffer(FAR struct enc_driver_s *priv, FAR uint8_t *buffer,
size_t buflen)
{
DEBUGASSERT(priv && priv->spi);
/* Select ENC28J60 chip */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the read buffer memory command (ignoring the response) */
SPI_SEND(priv->spi, ENC_RBM);
/* Then read the buffer data */
SPI_RECVBLOCK(priv->spi, buffer, buflen);
/* De-select ENC28J60 chip. */
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_bmdump(ENC_WBM, buffer, buflen);
}
/****************************************************************************
* Name: enc_wrbuffer
*
* Description:
* Write a buffer of data.
*
* Input Parameters:
* priv - Reference to the driver state structure
* buffer - A pointer to the buffer to write from
* buflen - The number of bytes to write
*
* Returned Value:
* None
*
* Assumptions:
* Read pointer is set to the correct address
*
****************************************************************************/
static inline void enc_wrbuffer(FAR struct enc_driver_s *priv,
FAR const uint8_t *buffer, size_t buflen)
{
DEBUGASSERT(priv && priv->spi);
/* Select ENC28J60 chip
*
* "The WBM command is started by lowering the CS pin. ..."
*/
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), true);
/* Send the write buffer memory command (ignoring the response)
*
* "...The [3-bit]WBM opcode should then be sent to the ENC28J60,
* followed by the 5-bit constant, 1Ah."
*/
SPI_SEND(priv->spi, ENC_WBM);
/* "...the ENC28J60 requires a single per packet control byte to
* precede the packet for transmission."
*
* POVERRIDE: Per Packet Override bit (Not set):
* 1 = The values of PCRCEN, PPADEN and PHUGEEN will override the
* configuration defined by MACON3.
* 0 = The values in MACON3 will be used to determine how the packet
* will be transmitted
* PCRCEN: Per Packet CRC Enable bit (Set, but won't be used because
* POVERRIDE is zero).
* PPADEN: Per Packet Padding Enable bit (Set, but won't be used because
* POVERRIDE is zero).
* PHUGEEN: Per Packet Huge Frame Enable bit (Set, but won't be used
* because POVERRIDE is zero).
*/
SPI_SEND(priv->spi,
(PKTCTRL_PCRCEN | PKTCTRL_PPADEN | PKTCTRL_PHUGEEN));
/* Then send the buffer
*
* "... After the WBM command and constant are sent, the data to
* be stored in the memory pointed to by EWRPT should be shifted
* out MSb first to the ENC28J60. After 8 data bits are received,
* the Write Pointer will automatically increment if AUTOINC is
* set. The host controller can continue to provide clocks on the
* SCK pin and send data on the SI pin, without raising CS, to
* keep writing to the memory. In this manner, with AUTOINC
* enabled, it is possible to continuously write sequential bytes
* to the buffer memory without any extra SPI command
* overhead.
*/
SPI_SNDBLOCK(priv->spi, buffer, buflen);
/* De-select ENC28J60 chip
*
* "The WBM command is terminated by bringing up the CS pin. ..."
*/
SPI_SELECT(priv->spi, SPIDEV_ETHERNET(0), false);
enc_bmdump(ENC_WBM, buffer, buflen + 1);
}
/****************************************************************************
* Name: enc_rdphy
*
* Description:
* Read 16-bits of PHY data.
*
* Input Parameters:
* priv - Reference to the driver state structure
* phyaddr - The PHY register address
*
* Returned Value:
* 16-bit value read from the PHY
*
* Assumptions:
*
****************************************************************************/
static uint16_t enc_rdphy(FAR struct enc_driver_s *priv, uint8_t phyaddr)
{
uint16_t data = 0;
/* "To read from a PHY register:
*
* 1. Write the address of the PHY register to read from into MIREGADR
* register.
*/
enc_wrbreg(priv, ENC_MIREGADR, phyaddr);
/* 2. Set the MICMD.MIIRD bit. The read operation begins and the
* MISTAT.BUSY bit is set.
*/
enc_wrbreg(priv, ENC_MICMD, MICMD_MIIRD);
/* 3. Wait 10.24 us. Poll the MISTAT.BUSY bit to be certain that the
* operation is complete. While busy, the host controller should not
* start any MIISCAN operations or write to the MIWRH register.
*
* When the MAC has obtained the register contents, the BUSY bit will
* clear itself.
*/
up_udelay(12);
if (enc_waitbreg(priv, ENC_MISTAT, MISTAT_BUSY, 0x00) == OK)
{
/* 4. Clear the MICMD.MIIRD bit. */
enc_wrbreg(priv, ENC_MICMD, 0x00);
/* 5. Read the desired data from the MIRDL and MIRDH registers. The
* order that these bytes are accessed is unimportant."
*/
data = (uint16_t)enc_rdbreg(priv, ENC_MIRDL);
data |= (uint16_t)enc_rdbreg(priv, ENC_MIRDH) << 8;
}
return data;
}
/****************************************************************************
* Name: enc_wrphy
*
* Description:
* write 16-bits of PHY data.
*
* Input Parameters:
* priv - Reference to the driver state structure
* phyaddr - The PHY register address
* phydata - 16-bit data to write to the PHY
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_wrphy(FAR struct enc_driver_s *priv, uint8_t phyaddr,
uint16_t phydata)
{
/* "To write to a PHY register:
*
* 1. Write the address of the PHY register to write to into the
* MIREGADR register.
*/
enc_wrbreg(priv, ENC_MIREGADR, phyaddr);
/* 2. Write the lower 8 bits of data to write into the MIWRL register. */
enc_wrbreg(priv, ENC_MIWRL, phydata);
/* 3. Write the upper 8 bits of data to write into MIWRH register.
* Writing to this register automatically begins MIIM transaction,
* so it must be written to after MIWRL. The MISTAT.BUSY bit becomes
* set.
*/
enc_wrbreg(priv, ENC_MIWRH, phydata >> 8);
/* The PHY register will be written after the MIIM operation completes,
* which takes 10.24 us. When the write operation has completed, BUSY
* bit will clear itself.
*
* The host controller should not start any MIISCAN or MIIRD operations
* while busy."
*/
up_udelay(12);
enc_waitbreg(priv, ENC_MISTAT, MISTAT_BUSY, 0x00);
}
/****************************************************************************
* Name: enc_transmit
*
* Description:
* Start hardware transmission. Called either from:
*
* - pkif interrupt when an application responds to the receipt of data
* by trying to send something, or
* - From watchdog based polling.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* OK on success; a negated errno on failure
*
* Assumptions:
*
****************************************************************************/
static int enc_transmit(FAR struct enc_driver_s *priv)
{
uint16_t txend;
/* Increment statistics */
ninfo("Sending packet, pktlen: %d\n", priv->dev.d_len);
NETDEV_TXPACKETS(&priv->dev);
/* Verify that the hardware is ready to send another packet. The driver
* starts a transmission process by setting ECON1.TXRTS. When the packet is
* finished transmitting or is aborted due to an error/cancellation, the
* ECON1.TXRTS bit will be cleared.
*
* NOTE: If we got here, then we have committed to sending a packet.
* higher level logic must have assured that (1) there is no transmission
* in progress, and that (2) TX-related interrupts are disabled.
*/
DEBUGASSERT((enc_rdgreg(priv, ENC_ECON1) & ECON1_TXRTS) == 0);
/* Send the packet: address=priv->dev.d_buf, length=priv->dev.d_len */
enc_dumppacket("Transmit Packet", priv->dev.d_buf, priv->dev.d_len);
/* Set transmit buffer start (is this necessary?). */
enc_wrbreg(priv, ENC_ETXSTL, PKTMEM_TX_START & 0xff);
enc_wrbreg(priv, ENC_ETXSTH, PKTMEM_TX_START >> 8);
/* Reset the write pointer to start of transmit buffer */
enc_wrbreg(priv, ENC_EWRPTL, PKTMEM_TX_START & 0xff);
enc_wrbreg(priv, ENC_EWRPTH, PKTMEM_TX_START >> 8);
/* Set the TX End pointer based on the size of the packet to send. Note
* that the offset accounts for the control byte at the beginning the
* buffer plus the size of the packet data.
*/
txend = PKTMEM_TX_START + priv->dev.d_len;
enc_wrbreg(priv, ENC_ETXNDL, txend & 0xff);
enc_wrbreg(priv, ENC_ETXNDH, txend >> 8);
/* Send the WBM command and copy the packet itself into the transmit
* buffer at the position of the EWRPT register.
*/
enc_wrbuffer(priv, priv->dev.d_buf, priv->dev.d_len);
/* Set TXRTS to send the packet in the transmit buffer */
enc_bfsgreg(priv, ENC_ECON1, ECON1_TXRTS);
/* Setup the TX timeout watchdog (perhaps restarting the timer). Note:
* Is there a race condition. Could the TXIF interrupt occur before
* the timer is started?
*/
wd_start(&priv->txtimeout, ENC_TXTIMEOUT,
enc_txtimeout, (wdparm_t)priv);
return OK;
}
/****************************************************************************
* Name: enc_txpoll
*
* Description:
* The transmitter is available, check if the network has any outgoing
* packets ready to send. This is a callback from devif_poll().
* devif_poll() may be called:
*
* 1. When the preceding TX packet send is complete,
* 2. When the preceding TX packet send timesout and the interface is
* reset
* 3. During normal TX polling
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
*
* Returned Value:
* OK on success; a negated errno on failure
*
* Assumptions:
* Interrupts are enabled but the caller holds the network lock.
*
****************************************************************************/
static int enc_txpoll(struct net_driver_s *dev)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
/* Send the packet */
enc_transmit(priv);
/* Stop the poll now because we can queue only one packet */
return -EBUSY;
}
/****************************************************************************
* Name: enc_linkstatus
*
* Description:
* The current link status can be obtained from the PHSTAT1.LLSTAT or
* PHSTAT2.LSTAT.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_linkstatus(FAR struct enc_driver_s *priv)
{
#if 0
uint16_t regval = enc_rdphy(priv, ENC_PHSTAT2);
priv->duplex = ((regval & PHSTAT2_DPXSTAT) != 0);
priv->carrier = ((regval & PHSTAT2_LSTAT) != 0);
#endif
}
/****************************************************************************
* Name: enc_txif
*
* Description:
* An TXIF interrupt was received indicating that the last TX packet(s) is
* done
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
* Interrupts are enabled but the caller holds the network lock.
*
****************************************************************************/
static void enc_txif(FAR struct enc_driver_s *priv)
{
/* Update statistics */
NETDEV_TXDONE(&priv->dev);
/* Clear the request to send bit */
enc_bfcgreg(priv, ENC_ECON1, ECON1_TXRTS);
/* If no further xmits are pending, then cancel the TX timeout */
wd_cancel(&priv->txtimeout);
/* Then poll the network for new XMIT data */
devif_poll(&priv->dev, enc_txpoll);
}
/****************************************************************************
* Name: enc_txerif
*
* Description:
* An TXERIF interrupt was received indicating that TX abort has occurred.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_txerif(FAR struct enc_driver_s *priv)
{
/* Update statistics */
NETDEV_TXERRORS(&priv->dev);
/* Reset TX */
enc_bfsgreg(priv, ENC_ECON1, ECON1_TXRST);
enc_bfcgreg(priv, ENC_ECON1, ECON1_TXRST | ECON1_TXRTS);
/* Here we really should re-transmit (I fact, if we want half duplex to
* work right, then it is necessary to do this!):
*
* 1. Read the TSV:
* - Read ETXNDL to get the end pointer
* - Read 7 bytes from that pointer + 1 using ENC_RMB.
* 2. Determine if we need to retransmit. Check the LATE COLLISION bit, if
* set, then we need to transmit.
* 3. Retranmit by resetting ECON1_TXRTS.
*/
#ifdef CONFIG_ENC28J60_HALFDUPLEX
# error "Missing logic for half duplex"
#endif
}
/****************************************************************************
* Name: enc_rxerif
*
* Description:
* An RXERIF interrupt was received indicating that the last TX packet(s)
* is done
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_rxerif(FAR struct enc_driver_s *priv)
{
/* REVISIT: Update statistics */
}
/****************************************************************************
* Name: enc_rxdispatch
*
* Description:
* Give the newly received packet to the network.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
* Interrupts are enabled but the caller holds the network lock.
*
****************************************************************************/
static void enc_rxdispatch(FAR struct enc_driver_s *priv)
{
#ifdef CONFIG_NET_PKT
/* When packet sockets are enabled, feed the frame into the tap */
pkt_input(&priv->dev);
#endif
/* We only accept IP packets of the configured type and ARP packets */
#ifdef CONFIG_NET_IPv4
2014-07-04 23:40:49 +02:00
if (BUF->type == HTONS(ETHTYPE_IP))
{
ninfo("IPv4 frame\n");
NETDEV_RXIPV4(&priv->dev);
/* Receive an IPv4 packet from the network device */
ipv4_input(&priv->dev);
/* If the above function invocation resulted in data that should be
* sent out on the network, d_len field will set to a value > 0.
*/
if (priv->dev.d_len > 0)
{
/* And send the packet */
enc_transmit(priv);
}
}
else
#endif
#ifdef CONFIG_NET_IPv6
if (BUF->type == HTONS(ETHTYPE_IP6))
{
2020-03-03 16:11:57 +01:00
ninfo("IPv6 frame\n");
NETDEV_RXIPV6(&priv->dev);
/* Give the IPv6 packet to the network layer */
ipv6_input(&priv->dev);
/* If the above function invocation resulted in data that should be
* sent out on the network, d_len field will set to a value > 0.
*/
if (priv->dev.d_len > 0)
{
/* And send the packet */
enc_transmit(priv);
}
}
else
#endif
#ifdef CONFIG_NET_ARP
if (BUF->type == HTONS(ETHTYPE_ARP))
{
ninfo("ARP packet received (%02x)\n", BUF->type);
NETDEV_RXARP(&priv->dev);
arp_input(&priv->dev);
/* If the above function invocation resulted in data that should be
* sent out on the network, d_len field will set to a value > 0.
*/
if (priv->dev.d_len > 0)
{
enc_transmit(priv);
}
}
else
#endif
{
nwarn("WARNING: Unsupported packet type dropped (%02x)\n",
HTONS(BUF->type));
NETDEV_RXDROPPED(&priv->dev);
}
}
/****************************************************************************
* Name: enc_pktif
*
* Description:
* An interrupt was received indicating the availability of a new RX packet
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
* Interrupts are enabled but the caller holds the network lock.
*
****************************************************************************/
static void enc_pktif(FAR struct enc_driver_s *priv)
{
uint8_t rsv[6];
uint16_t pktlen;
uint16_t rxstat;
/* Set the read pointer to the start of the received packet (ERDPT) */
DEBUGASSERT(priv->nextpkt <= PKTMEM_RX_END);
enc_wrbreg(priv, ENC_ERDPTL, (priv->nextpkt));
enc_wrbreg(priv, ENC_ERDPTH, (priv->nextpkt) >> 8);
/* Read the next packet pointer and the 4 byte read status vector (RSV)
* at the beginning of the received packet. (ERDPT should auto-increment
* and wrap to the beginning of the read buffer as necessary)
*/
enc_rdbuffer(priv, rsv, 6);
/* Decode the new next packet pointer, and the RSV. The
* RSV is encoded as:
*
* Bits 0-15: Indicates length of the received frame. This includes the
* destination address, source address, type/length, data,
* padding and CRC fields. This field is stored in little-
* endian format.
* Bits 16-31: Bit encoded RX status.
*/
priv->nextpkt = (uint16_t)rsv[1] << 8 | (uint16_t)rsv[0];
pktlen = (uint16_t)rsv[3] << 8 | (uint16_t)rsv[2];
rxstat = (uint16_t)rsv[5] << 8 | (uint16_t)rsv[4];
ninfo("Receiving packet, nextpkt: %04x pktlen: %d rxstat: %04x\n",
priv->nextpkt, pktlen, rxstat);
/* Check if the packet was received OK */
if ((rxstat & RXSTAT_OK) == 0)
{
nerr("ERROR: RXSTAT: %04x\n", rxstat);
NETDEV_RXERRORS(&priv->dev);
}
/* Check for a usable packet length (4 added for the CRC) */
else if (pktlen > (CONFIG_NET_ETH_PKTSIZE + 4) ||
pktlen <= (ETH_HDRLEN + 4))
{
nerr("ERROR: Bad packet size dropped (%d)\n", pktlen);
NETDEV_RXERRORS(&priv->dev);
}
/* Otherwise, read and process the packet */
else
{
/* Save the packet length (without the 4 byte CRC) in priv->dev.d_len */
priv->dev.d_len = pktlen - 4;
/* Copy the data data from the receive buffer to priv->dev.d_buf.
2019-09-29 20:52:20 +02:00
* ERDPT should be correctly positioned from the last call to
* end_rdbuffer (above).
*/
enc_rdbuffer(priv, priv->dev.d_buf, priv->dev.d_len);
enc_dumppacket("Received Packet", priv->dev.d_buf, priv->dev.d_len);
/* Dispatch the packet to the network */
enc_rxdispatch(priv);
}
/* Update statistics */
NETDEV_RXPACKETS(&priv->dev);
/* Move the RX read pointer to the start of the next received packet.
* This frees the memory we just read.
*/
enc_wrbreg(priv, ENC_ERXRDPTL, (priv->nextpkt));
enc_wrbreg(priv, ENC_ERXRDPTH, (priv->nextpkt) >> 8);
2015-10-04 23:04:00 +02:00
/* Decrement the packet counter indicate we are done with this packet */
enc_bfsgreg(priv, ENC_ECON2, ECON2_PKTDEC);
}
/****************************************************************************
* Name: enc_irqworker
*
* Description:
* Perform interrupt handling logic outside of the interrupt handler (on
* the work queue thread).
*
* Input Parameters:
* arg - The reference to the driver structure (case to void*)
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_irqworker(FAR void *arg)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)arg;
uint8_t eir;
DEBUGASSERT(priv);
/* Get exclusive access to both the network and the SPI bus. */
net_lock();
enc_lock(priv);
/* Disable further interrupts by clearing the global interrupt enable bit.
* "After an interrupt occurs, the host controller should clear the global
* enable bit for the interrupt pin before servicing the interrupt.
* Clearing the enable bit will cause the interrupt pin to return to the
* non-asserted state (high). Doing so will prevent the host controller
* from missing a falling edge should another interrupt occur while the
* immediate interrupt is being serviced."
*/
enc_bfcgreg(priv, ENC_EIE, EIE_INTIE);
/* Loop until all interrupts have been processed (EIR==0). Note that
* there is no infinite loop check... if there are always pending
* interrupts, we are just broken.
*/
while ((eir = enc_rdgreg(priv, ENC_EIR) & EIR_ALLINTS) != 0)
{
/* Handle interrupts according to interrupt register register bit
* settings.
*/
ninfo("EIR: %02x\n", eir);
/* DMAIF: The DMA interrupt indicates that the DMA module has completed
* its memory copy or checksum calculation. Additionally, this
* interrupt will be caused if the host controller cancels a DMA
* operation by manually clearing the DMAST bit. Once set, DMAIF can
* only be cleared by the host controller or by a Reset condition.
*/
if ((eir & EIR_DMAIF) != 0) /* DMA interrupt */
{
/* Not used by this driver. Just clear the interrupt request. */
enc_bfcgreg(priv, ENC_EIR, EIR_DMAIF);
}
/* LINKIF: The LINKIF indicates that the link status has changed.
* The actual current link status can be obtained from the
* PHSTAT1.LLSTAT or PHSTAT2.LSTAT. Unlike other interrupt sources, the
* link status change interrupt is created in the integrated PHY
* module.
*
* To receive it, the host controller must set the PHIE.PLNKIE and
* PGEIE bits. After setting the two PHY interrupt enable bits, the
* LINKIF bit will then shadow the contents of the PHIR.PGIF bit.
*
* Once LINKIF is set, it can only be cleared by the host controller or
* by a Reset. The LINKIF bit is read-only. Performing an MII read on
* the PHIR register will clear the LINKIF, PGIF and PLNKIF bits
* automatically and allow for future link status change interrupts.
*/
if ((eir & EIR_LINKIF) != 0) /* Link change interrupt */
{
enc_linkstatus(priv); /* Get current link status */
enc_rdphy(priv, ENC_PHIR); /* Clear the LINKIF interrupt */
}
/* TXIF: The Transmit Interrupt Flag (TXIF) is used to indicate that
* the requested packet transmission has ended. Upon transmission
* completion, abort or transmission cancellation by the host
* controller, the EIR.TXIF flag will be set to 1.
*
* Once TXIF is set, it can only be cleared by the host controller
* or by a Reset condition. Once processed, the host controller should
* use the BFC command to clear the EIR.TXIF bit.
*/
if ((eir & EIR_TXIF) != 0) /* Transmit interrupt */
{
enc_txif(priv); /* Handle TX completion */
enc_bfcgreg(priv, ENC_EIR, EIR_TXIF); /* Clear the TXIF interrupt */
}
/* TXERIF: The Transmit Error Interrupt Flag (TXERIF) is used to
* indicate that a transmit abort has occurred. An abort can occur
* because of any of the following:
*
* 1. Excessive collisions occurred as defined by the Retransmission
* Maximum (RETMAX) bits in the MACLCON1 register.
* 2. A late collision occurred as defined by the Collision Window
* (COLWIN) bits in the MACLCON2 register.
* 3. A collision after transmitting 64 bytes occurred (ESTAT.LATECOL
* set).
* 4. The transmission was unable to gain an opportunity to transmit
* the packet because the medium was constantly occupied for too
* long. The deferral limit (2.4287 ms) was reached and the
* MACON4.DEFER bit was clear.
* 5. An attempt to transmit a packet larger than the maximum frame
* length defined by the MAMXFL registers was made without setting
* the MACON3.HFRMEN bit or per packet POVERRIDE and PHUGEEN bits.
*
* Upon any of these conditions, the EIR.TXERIF flag is set to 1. Once
* set, it can only be cleared by the host controller or by a Reset
* condition.
*
* After a transmit abort, the TXRTS bit will be cleared, the
* ESTAT.TXABRT bit will be set and the transmit status vector will be
* written at ETXND + 1. The MAC will not automatically attempt to
* retransmit the packet. The host controller may wish to read the
* transmit status vector and LATECOL bit to determine the cause of
* the abort. After determining the problem and solution, the host
* controller should clear the LATECOL (if set) and TXABRT bits so
* that future aborts can be detected accurately.
*
* In Full-Duplex mode, condition 5 is the only one that should cause
* this interrupt. Collisions and other problems related to sharing
* the network are not possible on full-duplex networks. The conditions
* which cause the transmit error interrupt meet the requirements of
* the transmit interrupt. As a result, when this interrupt occurs,
* TXIF will also be simultaneously set.
*/
if ((eir & EIR_TXERIF) != 0) /* Transmit Error Interrupts */
{
enc_txerif(priv); /* Handle the TX error */
enc_bfcgreg(priv, ENC_EIR, EIR_TXERIF); /* Clear the TXERIF interrupt */
}
/* PKTIF The Receive Packet Pending Interrupt Flag (PKTIF) is used to
* indicate the presence of one or more data packets in the receive
* buffer and to provide a notification means for the arrival of new
* packets. When the receive buffer has at least one packet in it,
* EIR.PKTIF will be set. In other words, this interrupt flag will be
* set anytime the Ethernet Packet Count register (EPKTCNT) is
* non-zero.
*
* The PKTIF bit can only be cleared by the host controller or by a
* Reset condition. In order to clear PKTIF, the EPKTCNT register must
* be decremented to 0. If the last data packet in the receive buffer
* is processed, EPKTCNT will become zero and the PKTIF bit will
* automatically be cleared.
*/
#if 0
/* Ignore PKTIF because is unreliable. Use EPKTCNT instead */
if ((eir & EIR_PKTIF) != 0)
#endif
{
uint8_t pktcnt = enc_rdbreg(priv, ENC_EPKTCNT);
if (pktcnt > 0)
{
ninfo("EPKTCNT: %02x\n", pktcnt);
/* Handle packet receipt */
enc_pktif(priv);
}
}
/* RXERIF: The Receive Error Interrupt Flag (RXERIF) is used to
* indicate a receive buffer overflow condition. Alternately, this
* interrupt may indicate that too many packets are in the receive
* buffer and more cannot be stored without overflowing the EPKTCNT
* register. When a packet is being received and the receive buffer
* runs completely out of space, or EPKTCNT is 255 and cannot be
* incremented, the packet being received will be aborted (permanently
* lost) and the EIR.RXERIF bit will be set to 1.
*
* Once set, RXERIF can only be cleared by the host controller or by a
* Reset condition. Normally, upon the receive error condition, the
* host controller would process any packets pending from the receive
* buffer and then make additional room for future packets by
* advancing the ERXRDPT registers (low byte first) and decrementing
* the EPKTCNT register.
*
* Once processed, the host controller should use the BFC command to
* clear the EIR.RXERIF bit.
*/
if ((eir & EIR_RXERIF) != 0) /* Receive Error Interrupts */
{
enc_rxerif(priv); /* Handle the RX error */
enc_bfcgreg(priv, ENC_EIR, EIR_RXERIF); /* Clear the RXERIF interrupt */
}
}
/* Enable GPIO interrupts */
priv->lower->enable(priv->lower);
/* Enable Ethernet interrupts */
enc_bfsgreg(priv, ENC_EIE, EIE_INTIE);
/* Release lock on the SPI bus and the network */
enc_unlock(priv);
net_unlock();
}
/****************************************************************************
* Name: enc_interrupt
*
* Description:
* Hardware interrupt handler
*
* Input Parameters:
* irq - Number of the IRQ that generated the interrupt
* context - Interrupt register state save info (architecture-specific)
*
* Returned Value:
* OK on success
*
* Assumptions:
*
****************************************************************************/
static int enc_interrupt(int irq, FAR void *context, FAR void *arg)
{
FAR struct enc_driver_s *priv;
DEBUGASSERT(arg != NULL);
priv = (FAR struct enc_driver_s *)arg;
/* In complex environments, we cannot do SPI transfers from the interrupt
* handler because semaphores are probably used to lock the SPI bus. In
* this case, we will defer processing to the worker thread. This is also
* much kinder in the use of system resources and is, therefore, probably
* a good thing to do in any event.
*/
DEBUGASSERT(work_available(&priv->irqwork));
/* Notice that further GPIO interrupts are disabled until the work is
* actually performed. This is to prevent overrun of the worker thread.
* Interrupts are re-enabled in enc_irqworker() when the work is completed.
*/
priv->lower->disable(priv->lower);
return work_queue(ENCWORK, &priv->irqwork, enc_irqworker,
(FAR void *)priv, 0);
}
/****************************************************************************
* Name: enc_toworker
*
* Description:
* Our TX watchdog timed out. This is the worker thread continuation of
* the watchdog timer interrupt. Reset the hardware and start again.
*
* Input Parameters:
* arg - The reference to the driver structure (case to void*)
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_toworker(FAR void *arg)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)arg;
int ret;
nerr("ERROR: Tx timeout\n");
DEBUGASSERT(priv);
/* Get exclusive access to the network */
net_lock();
/* Increment statistics and dump debug info */
NETDEV_TXTIMEOUTS(&priv->dev);
/* Then reset the hardware: Take the interface down, then bring it
* back up
*/
ret = enc_ifdown(&priv->dev);
DEBUGASSERT(ret == OK);
ret = enc_ifup(&priv->dev);
DEBUGASSERT(ret == OK);
UNUSED(ret);
/* Then poll the network for new XMIT data */
devif_poll(&priv->dev, enc_txpoll);
/* Release lock on the network */
net_unlock();
}
/****************************************************************************
* Name: enc_txtimeout
*
* Description:
* Our TX watchdog timed out. Called from the timer interrupt handler.
* The last TX never completed. Perform work on the worker thread.
*
* Input Parameters:
* arg - The argument
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_txtimeout(wdparm_t arg)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)arg;
int ret;
/* In complex environments, we cannot do SPI transfers from the timeout
* handler because semaphores are probably used to lock the SPI bus. In
* this case, we will defer processing to the worker thread. This is also
* much kinder in the use of system resources and is, therefore, probably
* a good thing to do in any event.
*/
DEBUGASSERT(priv && work_available(&priv->towork));
/* Notice that Tx timeout watchdog is not active so further Tx timeouts
* can occur until we restart the Tx timeout watchdog.
*/
ret = work_queue(ENCWORK, &priv->towork, enc_toworker, priv, 0);
DEBUGASSERT(ret == OK);
UNUSED(ret);
}
/****************************************************************************
* Name: enc_ifup
*
* Description:
* NuttX Callback: Bring up the Ethernet interface when an IP address is
* provided
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
2014-06-28 00:48:12 +02:00
static int enc_ifup(struct net_driver_s *dev)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
int ret;
ninfo("Bringing up: %u.%u.%u.%u\n",
ip4_addr1(dev->d_ipaddr), ip4_addr2(dev->d_ipaddr),
ip4_addr3(dev->d_ipaddr), ip4_addr4(dev->d_ipaddr));
/* Lock the SPI bus so that we have exclusive access */
enc_lock(priv);
/* Initialize Ethernet interface, set the MAC address, and make sure that
* the ENC28J80 is not in power save mode.
*/
ret = enc_reset(priv);
if (ret == OK)
{
enc_setmacaddr(priv);
enc_pwrfull(priv);
/* Enable interrupts at the ENC28J60. Interrupts are still disabled
* at the interrupt controller.
*/
enc_wrphy(priv, ENC_PHIE, PHIE_PGEIE | PHIE_PLNKIE);
enc_bfcgreg(priv, ENC_EIR, EIR_ALLINTS);
enc_wrgreg(priv, ENC_EIE, EIE_INTIE | EIE_PKTIE | EIE_LINKIE |
EIE_TXIE | EIE_TXERIE | EIE_RXERIE);
/* Enable the receiver */
enc_bfsgreg(priv, ENC_ECON1, ECON1_RXEN);
/* Mark the interface up and enable the Ethernet interrupt at the
* controller
*/
priv->ifstate = ENCSTATE_UP;
priv->lower->enable(priv->lower);
}
/* Un-lock the SPI bus */
enc_unlock(priv);
return ret;
}
/****************************************************************************
* Name: enc_ifdown
*
* Description:
* NuttX Callback: Stop the interface.
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
2014-06-28 00:48:12 +02:00
static int enc_ifdown(struct net_driver_s *dev)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
irqstate_t flags;
int ret;
ninfo("Taking down: %u.%u.%u.%u\n",
ip4_addr1(dev->d_ipaddr), ip4_addr2(dev->d_ipaddr),
ip4_addr3(dev->d_ipaddr), ip4_addr4(dev->d_ipaddr));
/* Lock the SPI bus so that we have exclusive access */
enc_lock(priv);
/* Disable the Ethernet interrupt */
flags = enter_critical_section();
priv->lower->disable(priv->lower);
/* Cancel the TX timeout timers */
wd_cancel(&priv->txtimeout);
/* Reset the device and leave in the power save state */
ret = enc_reset(priv);
enc_pwrsave(priv);
priv->ifstate = ENCSTATE_DOWN;
leave_critical_section(flags);
/* Un-lock the SPI bus */
enc_unlock(priv);
return ret;
}
/****************************************************************************
* Name: enc_txavail
*
* Description:
* Driver callback invoked when new TX data is available. This is a
* stimulus perform an out-of-cycle poll and, thereby, reduce the TX
* latency.
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
*
* Returned Value:
* None
*
* Assumptions:
* Called in normal user mode
*
****************************************************************************/
2014-06-28 00:48:12 +02:00
static int enc_txavail(struct net_driver_s *dev)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
irqstate_t flags;
/* Lock the SPI bus so that we have exclusive access */
enc_lock(priv);
/* Ignore the notification if the interface is not yet up */
flags = enter_critical_section();
if (priv->ifstate == ENCSTATE_UP)
{
/* Check if the hardware is ready to send another packet. The driver
* starts a transmission process by setting ECON1.TXRTS. When the
* packet is finished transmitting or is aborted due to an error/
* cancellation, the ECON1.TXRTS bit will be cleared.
*/
if ((enc_rdgreg(priv, ENC_ECON1) & ECON1_TXRTS) == 0)
{
/* The interface is up and TX is idle;
* poll the network for new XMIT data
*/
devif_poll(&priv->dev, enc_txpoll);
}
}
/* Un-lock the SPI bus */
leave_critical_section(flags);
enc_unlock(priv);
return OK;
}
/****************************************************************************
* Name: enc_addmac
*
* Description:
* NuttX Callback: Add the specified MAC address to the hardware multicast
* address filtering
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
* mac - The MAC address to be added
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
#ifdef CONFIG_NET_MCASTGROUP
2014-06-28 00:48:12 +02:00
static int enc_addmac(struct net_driver_s *dev, FAR const uint8_t *mac)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
/* Lock the SPI bus so that we have exclusive access */
enc_lock(priv);
/* Add the MAC address to the hardware multicast routing table */
/* #warning "Multicast MAC support not implemented" */
/* Un-lock the SPI bus */
enc_unlock(priv);
return -ENOSYS;
}
#endif
/****************************************************************************
* Name: enc_rmmac
*
* Description:
* NuttX Callback: Remove the specified MAC address from the hardware
* multicast address filtering
*
* Input Parameters:
* dev - Reference to the NuttX driver state structure
* mac - The MAC address to be removed
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
#ifdef CONFIG_NET_MCASTGROUP
2014-06-28 00:48:12 +02:00
static int enc_rmmac(struct net_driver_s *dev, FAR const uint8_t *mac)
{
FAR struct enc_driver_s *priv = (FAR struct enc_driver_s *)dev->d_private;
/* Lock the SPI bus so that we have exclusive access */
enc_lock(priv);
/* Add the MAC address to the hardware multicast routing table */
/* #warning "Multicast MAC support not implemented" */
/* Un-lock the SPI bus */
enc_unlock(priv);
return -ENOSYS;
}
#endif
/****************************************************************************
* Name: enc_pwrsave
*
* Description:
* The ENC28J60 may be commanded to power-down via the SPI interface.
* When powered down, it will no longer be able to transmit and receive
* any packets. To maximize power savings:
*
* 1. Turn off packet reception by clearing ECON1.RXEN.
* 2. Wait for any in-progress packets to finish being received by
* polling ESTAT.RXBUSY. This bit should be clear before proceeding.
* 3. Wait for any current transmissions to end by confirming ECON1.TXRTS
* is clear.
* 4. Set ECON2.VRPS (if not already set).
* 5. Enter Sleep by setting ECON2.PWRSV. All MAC, MII and PHY registers
* become inaccessible as a result. Setting PWRSV also clears
* ESTAT.CLKRDY automatically.
*
* In Sleep mode, all registers and buffer memory will maintain their
* states. The ETH registers and buffer memory will still be accessible
* by the host controller. Additionally, the clock driver will continue
* to operate. The CLKOUT function will be unaffected.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_pwrsave(FAR struct enc_driver_s *priv)
{
ninfo("Set PWRSV\n");
/* 1. Turn off packet reception by clearing ECON1.RXEN. */
enc_bfcgreg(priv, ENC_ECON1, ECON1_RXEN);
/* 2. Wait for any in-progress packets to finish being received by
* polling ESTAT.RXBUSY. This bit should be clear before proceeding.
*/
if (enc_waitbreg(priv, ENC_ESTAT, ESTAT_RXBUSY, 0) == OK)
{
/* 3. Wait for any current transmissions to end by confirming
* ECON1.TXRTS is clear.
*/
enc_waitbreg(priv, ENC_ECON1, ECON1_TXRTS, 0);
/* 4. Set ECON2.VRPS (if not already set).
* (Set in enc_reset()
*
* 5. Enter Sleep by setting ECON2.PWRSV.
*/
enc_bfsgreg(priv, ENC_ECON2, ECON2_PWRSV);
}
}
/****************************************************************************
* Name: enc_pwrfull
*
* Description:
* When normal operation is desired, the host controller must perform
* a slightly modified procedure:
*
* 1. Wake-up by clearing ECON2.PWRSV.
* 2. Wait at least 300 us for the PHY to stabilize. To accomplish the
* delay, the host controller may poll ESTAT.CLKRDY and wait for it
* to become set.
* 3. Restore receive capability by setting ECON1.RXEN.
*
* After leaving Sleep mode, there is a delay of many milliseconds
* before a new link is established (assuming an appropriate link
* partner is present). The host controller may wish to wait until
* the link is established before attempting to transmit any packets.
* The link status can be determined by polling the PHSTAT2.LSTAT bit.
* Alternatively, the link change interrupt may be used if it is
* enabled.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_pwrfull(FAR struct enc_driver_s *priv)
{
ninfo("Clear PWRSV\n");
/* 1. Wake-up by clearing ECON2.PWRSV. */
enc_bfcgreg(priv, ENC_ECON2, ECON2_PWRSV);
/* 2. Wait at least 300 us for the PHY to stabilize. To accomplish the
* delay, the host controller may poll ESTAT.CLKRDY and wait for it to
* become set.
*/
enc_waitbreg(priv, ENC_ESTAT, ESTAT_CLKRDY, ESTAT_CLKRDY);
/* 3. Restore receive capability by setting ECON1.RXEN.
*
* The caller will do this when it is read to receive packets
*/
}
/****************************************************************************
* Name: enc_setmacaddr
*
* Description:
* Set the MAC address to the configured value. This is done after ifup
* or after a TX timeout. Note that this means that the interface must
* be down before configuring the MAC addr.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static void enc_setmacaddr(FAR struct enc_driver_s *priv)
{
/* Program the hardware with it's MAC address (for filtering).
* MAADR1 MAC Address Byte 1 (MAADR<47:40>), OUI Byte 1
* MAADR2 MAC Address Byte 2 (MAADR<39:32>), OUI Byte 2
* MAADR3 MAC Address Byte 3 (MAADR<31:24>), OUI Byte 3
* MAADR4 MAC Address Byte 4 (MAADR<23:16>)
* MAADR5 MAC Address Byte 5 (MAADR<15:8>)
* MAADR6 MAC Address Byte 6 (MAADR<7:0>)
*/
enc_wrbreg(priv, ENC_MAADR1, priv->dev.d_mac.ether.ether_addr_octet[0]);
enc_wrbreg(priv, ENC_MAADR2, priv->dev.d_mac.ether.ether_addr_octet[1]);
enc_wrbreg(priv, ENC_MAADR3, priv->dev.d_mac.ether.ether_addr_octet[2]);
enc_wrbreg(priv, ENC_MAADR4, priv->dev.d_mac.ether.ether_addr_octet[3]);
enc_wrbreg(priv, ENC_MAADR5, priv->dev.d_mac.ether.ether_addr_octet[4]);
enc_wrbreg(priv, ENC_MAADR6, priv->dev.d_mac.ether.ether_addr_octet[5]);
}
/****************************************************************************
* Name: enc_reset
*
* Description:
* Stop, reset, re-initialize, and restart the ENC28J60. This is done
* initially, on ifup, and after a TX timeout.
*
* Input Parameters:
* priv - Reference to the driver state structure
*
* Returned Value:
* None
*
* Assumptions:
*
****************************************************************************/
static int enc_reset(FAR struct enc_driver_s *priv)
{
uint8_t regval;
nwarn("WARNING: Reset\n");
/* Configure SPI for the ENC28J60 */
enc_configspi(priv->spi);
/* Reset the ENC28J60 */
enc_src(priv);
/* Initialize ECON1: Clear ECON1 */
enc_wrgreg(priv, ENC_ECON1, 0x00);
/* Initialize ECON2: Enable address auto increment and voltage
* regulator powersave.
*/
enc_wrgreg(priv, ENC_ECON2, ECON2_AUTOINC | ECON2_VRPS);
/* Initialize receive buffer.
* First, set the receive buffer start address.
*/
priv->nextpkt = PKTMEM_RX_START;
enc_wrbreg(priv, ENC_ERXSTL, PKTMEM_RX_START & 0xff);
enc_wrbreg(priv, ENC_ERXSTH, PKTMEM_RX_START >> 8);
/* Set the receive data pointer */
enc_wrbreg(priv, ENC_ERXRDPTL, PKTMEM_RX_START & 0xff);
enc_wrbreg(priv, ENC_ERXRDPTH, PKTMEM_RX_START >> 8);
/* Set the receive buffer end. */
enc_wrbreg(priv, ENC_ERXNDL, PKTMEM_RX_END & 0xff);
enc_wrbreg(priv, ENC_ERXNDH, PKTMEM_RX_END >> 8);
/* Set transmit buffer start. */
enc_wrbreg(priv, ENC_ETXSTL, PKTMEM_TX_START & 0xff);
enc_wrbreg(priv, ENC_ETXSTH, PKTMEM_TX_START >> 8);
/* Check if we are actually communicating with the ENC28J60. If its
* 0x00 or 0xff, then we are probably not communicating correctly
* via SPI.
*/
regval = enc_rdbreg(priv, ENC_EREVID);
if (regval == 0x00 || regval == 0xff)
{
nerr("ERROR: Bad Rev ID: %02x\n", regval);
return -ENODEV;
}
ninfo("Rev ID: %02x\n", regval);
/* Set filter mode: unicast OR broadcast AND crc valid */
enc_wrbreg(priv, ENC_ERXFCON, ERXFCON_UCEN | ERXFCON_CRCEN |
ERXFCON_BCEN);
/* Enable MAC receive */
enc_wrbreg(priv, ENC_MACON1, MACON1_MARXEN | MACON1_TXPAUS |
MACON1_RXPAUS);
/* Enable automatic padding and CRC operations */
#ifdef CONFIG_ENC28J60_HALFDUPLEX
enc_wrbreg(priv, ENC_MACON3, MACON3_PADCFG0 | MACON3_TXCRCEN |
MACON3_FRMLNEN);
enc_wrbreg(priv, ENC_MACON4, MACON4_DEFER); /* Defer transmission enable */
/* Set Non-Back-to-Back Inter-Packet Gap */
enc_wrbreg(priv, ENC_MAIPGL, 0x12);
enc_wrbreg(priv, ENC_MAIPGH, 0x0c);
/* Set Back-to-Back Inter-Packet Gap */
enc_wrbreg(priv, ENC_MABBIPG, 0x12);
#else
/* Set filter mode: unicast OR broadcast AND crc valid AND Full Duplex */
enc_wrbreg(priv, ENC_MACON3,
MACON3_PADCFG0 | MACON3_TXCRCEN | MACON3_FRMLNEN |
MACON3_FULDPX);
/* Set Non-Back-to-Back Inter-Packet Gap */
enc_wrbreg(priv, ENC_MAIPGL, 0x12);
/* Set Back-to-Back Inter-Packet Gap */
enc_wrbreg(priv, ENC_MABBIPG, 0x15);
#endif
/* Set the maximum packet size which the controller will accept */
This commit attempts remove some long standard confusion in naming and some actual problems that result from the naming confusion. The basic problem is the standard MTU does not include the size of the Ethernet header. For clarity, I changed the naming of most things called MTU to PKTSIZE. For example, CONFIG_NET_ETH_MTU is now CONFIG_NET_ETH_PKTSIZE. This makes the user interface a little hostile. People thing of an MTU of 1500 bytes, but the corresponding packet is really 1514 bytes (including the 14 byte Ethernet header). A more friendly solution would configure the MTU (as before), but then derive the packet buffer size by adding the MAC header length. Instead, we define the packet buffer size then derive the MTU. The MTU is not common currency in networking. On the wire, the only real issue is the MSS which is derived from MTU by subtracting the IP header and TCP header sizes (for the case of TCP). Now it is derived for the PKTSIZE by subtracting the IP header, the TCP header, and the MAC header sizes. So we should be all good and without the recurring 14 byte error in MTU's and MSS's. Squashed commit of the following: Trivial update to fix some spacing issues. net/: Rename several macros containing _MTU to _PKTSIZE. net/: Rename CONFIG_NET_SLIP_MTU to CONFIG_NET_SLIP_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_6LOWPAN_MTU to CONFIG_NET_6LOWPAN_PKTSIZE and similarly for CONFIG_NET_TUN_MTU. These are not the MTU which does not include the size of the link layer header. These are the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename CONFIG_NET_ETH_MTU to CONFIG_NET_ETH_PKTSIZE. This is not the MTU which does not include the size of the link layer header. This is the full size of the packet buffer memory (minus any GUARD bytes). net/: Rename the file d_mtu in the network driver structure to d_pktsize. That value saved there is not the MTU. The packetsize is the memory large enough to hold the maximum packet PLUS the size of the link layer header. The MTU does not include the link layer header.
2018-07-04 22:10:40 +02:00
enc_wrbreg(priv, ENC_MAMXFLL, CONFIG_NET_ETH_PKTSIZE & 0xff);
enc_wrbreg(priv, ENC_MAMXFLH, CONFIG_NET_ETH_PKTSIZE >> 8);
/* Configure LEDs (No, just use the defaults for now) */
/* Setup up PHCON1 & 2 */
#ifdef CONFIG_ENC28J60_HALFDUPLEX
enc_wrphy(priv, ENC_PHCON1, 0x00);
enc_wrphy(priv, ENC_PHCON2, PHCON2_HDLDIS);
#else
enc_wrphy(priv, ENC_PHCON1, PHCON1_PDPXMD);
enc_wrphy(priv, ENC_PHCON2, 0x00);
#endif
return OK;
}
/****************************************************************************
* Public Functions
****************************************************************************/
/****************************************************************************
* Name: enc_initialize
*
* Description:
* Initialize the Ethernet driver. The ENC28J60 device is assumed to be
* in the post-reset state upon entry to this function.
*
* Input Parameters:
* spi - A reference to the platform's SPI driver for the ENC28J60
* lower - The MCU-specific interrupt used to control low-level MCU
* functions (i.e., ENC28J60 GPIO interrupts).
* devno - If more than one ENC28J60 is supported, then this is the
* zero based number that identifies the ENC28J60;
*
* Returned Value:
* OK on success; Negated errno on failure.
*
* Assumptions:
*
****************************************************************************/
int enc_initialize(FAR struct spi_dev_s *spi,
FAR const struct enc_lower_s *lower, unsigned int devno)
{
FAR struct enc_driver_s *priv;
DEBUGASSERT(devno < CONFIG_ENC28J60_NINTERFACES);
priv = &g_enc28j60[devno];
/* Initialize the driver structure */
memset(g_enc28j60, 0,
CONFIG_ENC28J60_NINTERFACES * sizeof(struct enc_driver_s));
priv->dev.d_buf = (FAR uint8_t *)g_pktbuf[devno]; /* Single packet buffer */
priv->dev.d_ifup = enc_ifup; /* I/F down callback */
priv->dev.d_ifdown = enc_ifdown; /* I/F up (new IP address) callback */
priv->dev.d_txavail = enc_txavail; /* New TX data callback */
#ifdef CONFIG_NET_MCASTGROUP
priv->dev.d_addmac = enc_addmac; /* Add multicast MAC address */
priv->dev.d_rmmac = enc_rmmac; /* Remove multicast MAC address */
#endif
priv->dev.d_private = priv; /* Used to recover private state from dev */
priv->spi = spi; /* Save the SPI instance */
priv->lower = lower; /* Save the low-level MCU interface */
/* The interface should be in the down state. However, this function is
* called too early in initialization to perform the ENC28J60 reset in
* enc_ifdown. We are depending upon the fact that the application level
* logic will call enc_ifdown later to reset the ENC28J60. NOTE: The MAC
* address will not be set up until enc_ifup() is called. That gives the
* app time to set the MAC address before bringing the interface up.
*/
priv->ifstate = ENCSTATE_UNINIT;
/* Attach the interrupt to the driver (but don't enable it yet) */
if (lower->attach(lower, enc_interrupt, priv) < 0)
{
/* We could not attach the ISR to the interrupt */
return -EAGAIN;
}
/* Register the device with the OS so that socket IOCTLs can be performed */
return netdev_register(&priv->dev, NET_LL_ETHERNET);
}
#endif /* CONFIG_NET && CONFIG_ENC28J60_NET */