AP_Networking: added ArduPilot HAL port of lwip

This commit is contained in:
Andrew Tridgell 2023-12-26 06:31:16 +11:00
parent b28912c351
commit 22938e99ea
8 changed files with 1290 additions and 45 deletions

View File

@ -6,9 +6,16 @@
#include "AP_Networking_ChibiOS.h"
#include <GCS_MAVLink/GCS.h>
#include <lwipthread.h>
#include <lwip/udp.h>
#include <lwip/ip_addr.h>
#include <lwip/tcpip.h>
#include <lwip/netifapi.h>
#if LWIP_DHCP
#include <lwip/dhcp.h>
#endif
#include <hal.h>
#include "../../modules/ChibiOS/os/various/evtimer.h"
#include <AP_HAL_ChibiOS/hwdef/common/stm32_util.h>
extern const AP_HAL::HAL& hal;
@ -18,23 +25,34 @@ extern const AP_HAL::HAL& hal;
/*
these are referenced as globals inside lwip
*/
*/
stm32_eth_rx_descriptor_t *__eth_rd;
stm32_eth_tx_descriptor_t *__eth_td;
uint32_t *__eth_rb[STM32_MAC_RECEIVE_BUFFERS];
uint32_t *__eth_tb[STM32_MAC_TRANSMIT_BUFFERS];
#define LWIP_SEND_TIMEOUT_MS 50
#define LWIP_NETIF_MTU 1500
#define LWIP_LINK_POLL_INTERVAL TIME_S2I(5)
#define PERIODIC_TIMER_ID 1
#define FRAME_RECEIVED_ID 2
#if CH_CFG_ST_RESOLUTION != 32
#error "ethernet requires 32 bit timer"
#endif
/*
allocate buffers for LWIP
*/
*/
bool AP_Networking_ChibiOS::allocate_buffers()
{
#define AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE ((((STM32_MAC_BUFFERS_SIZE - 1) | 3) + 1) / 4) // typically == 381
#define AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE ((((STM32_MAC_BUFFERS_SIZE - 1) | 3) + 1) / 4) // typically == 381
// check total size of buffers
const uint32_t total_size = sizeof(stm32_eth_rx_descriptor_t)*STM32_MAC_RECEIVE_BUFFERS +
sizeof(stm32_eth_tx_descriptor_t)*STM32_MAC_TRANSMIT_BUFFERS +
sizeof(uint32_t)*STM32_MAC_RECEIVE_BUFFERS*AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE +
sizeof(uint32_t)*STM32_MAC_TRANSMIT_BUFFERS*AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE; // typically == 9240
sizeof(stm32_eth_tx_descriptor_t)*STM32_MAC_TRANSMIT_BUFFERS +
sizeof(uint32_t)*STM32_MAC_RECEIVE_BUFFERS*AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE +
sizeof(uint32_t)*STM32_MAC_TRANSMIT_BUFFERS*AP_NETWORKING_EXTERN_MAC_BUFFER_SIZE; // typically == 9240
// ensure that we allocate 32-bit aligned memory, and mark it non-cacheable
uint32_t size = 1;
@ -84,7 +102,7 @@ bool AP_Networking_ChibiOS::allocate_buffers()
/*
initialise ChibiOS network backend using LWIP
*/
*/
bool AP_Networking_ChibiOS::init()
{
#ifdef HAL_GPIO_ETH_ENABLE
@ -102,21 +120,6 @@ bool AP_Networking_ChibiOS::init()
return false;
}
lwip_options = new lwipthread_opts;
if (frontend.get_dhcp_enabled()) {
lwip_options->addrMode = NET_ADDRESS_DHCP;
} else {
lwip_options->addrMode = NET_ADDRESS_STATIC;
lwip_options->address = htonl(frontend.get_ip_param());
lwip_options->netmask = htonl(frontend.get_netmask_param());
lwip_options->gateway = htonl(frontend.get_gateway_param());
}
frontend.param.macaddr.get_address(macaddr);
lwip_options->macaddress = macaddr;
lwipInit(lwip_options);
#if LWIP_IGMP
if (ETH != nullptr) {
// enbale "permit multicast" so we can receive multicast packets
@ -124,17 +127,252 @@ bool AP_Networking_ChibiOS::init()
}
#endif
thisif = new netif;
if (thisif == nullptr) {
return false;
}
if (!hal.scheduler->thread_create(FUNCTOR_BIND_MEMBER(&AP_Networking_ChibiOS::thread, void),
"network",
2048, AP_HAL::Scheduler::PRIORITY_NET, 0)) {
return false;
}
return true;
}
void AP_Networking_ChibiOS::link_up_cb(void *p)
{
auto *driver = (AP_Networking_ChibiOS *)p;
#if LWIP_DHCP
if (driver->frontend.get_dhcp_enabled()) {
dhcp_start(driver->thisif);
}
#endif
}
void AP_Networking_ChibiOS::link_down_cb(void *p)
{
auto *driver = (AP_Networking_ChibiOS *)p;
#if LWIP_DHCP
if (driver->frontend.get_dhcp_enabled()) {
dhcp_stop(driver->thisif);
}
#endif
}
/*
* This function does the actual transmission of the packet. The packet is
* contained in the pbuf that is passed to the function. This pbuf
* might be chained.
*
* @param netif the lwip network interface structure for this ethernetif
* @param p the MAC packet to send (e.g. IP packet including MAC addresses and type)
* @return ERR_OK if the packet could be sent
* an err_t value if the packet couldn't be sent
*
* @note Returning ERR_MEM here if a DMA queue of your MAC is full can lead to
* strange results. You might consider waiting for space in the DMA queue
* to become available since the stack doesn't retry to send a packet
* dropped because of memory failure (except for the TCP timers).
*/
int8_t AP_Networking_ChibiOS::low_level_output(struct netif *netif, struct pbuf *p)
{
struct pbuf *q;
MACTransmitDescriptor td;
(void)netif;
if (macWaitTransmitDescriptor(&ETHD1, &td, TIME_MS2I(LWIP_SEND_TIMEOUT_MS)) != MSG_OK) {
return ERR_TIMEOUT;
}
#if ETH_PAD_SIZE
pbuf_header(p, -ETH_PAD_SIZE); /* drop the padding word */
#endif
/* Iterates through the pbuf chain. */
for(q = p; q != NULL; q = q->next) {
macWriteTransmitDescriptor(&td, (uint8_t *)q->payload, (size_t)q->len);
}
macReleaseTransmitDescriptorX(&td);
#if ETH_PAD_SIZE
pbuf_header(p, ETH_PAD_SIZE); /* reclaim the padding word */
#endif
return ERR_OK;
}
/*
* Receives a frame.
* Allocates a pbuf and transfers the bytes of the incoming
* packet from the interface into the pbuf.
*
* @param netif the lwip network interface structure for this ethernetif
* @return a pbuf filled with the received packet (including MAC header)
* NULL on memory error
*/
bool AP_Networking_ChibiOS::low_level_input(struct netif *netif, struct pbuf **pbuf)
{
MACReceiveDescriptor rd;
struct pbuf *q;
u16_t len;
(void)netif;
if (macWaitReceiveDescriptor(&ETHD1, &rd, TIME_IMMEDIATE) != MSG_OK) {
return false;
}
len = (u16_t)rd.size;
#if ETH_PAD_SIZE
len += ETH_PAD_SIZE; /* allow room for Ethernet padding */
#endif
/* We allocate a pbuf chain of pbufs from the pool. */
*pbuf = pbuf_alloc(PBUF_RAW, len, PBUF_POOL);
if (*pbuf != nullptr) {
#if ETH_PAD_SIZE
pbuf_header(*pbuf, -ETH_PAD_SIZE); /* drop the padding word */
#endif
/* Iterates through the pbuf chain. */
for(q = *pbuf; q != NULL; q = q->next) {
macReadReceiveDescriptor(&rd, (uint8_t *)q->payload, (size_t)q->len);
}
macReleaseReceiveDescriptorX(&rd);
#if ETH_PAD_SIZE
pbuf_header(*pbuf, ETH_PAD_SIZE); /* reclaim the padding word */
#endif
} else {
macReleaseReceiveDescriptorX(&rd); // Drop packet
}
return true;
}
int8_t AP_Networking_ChibiOS::ethernetif_init(struct netif *netif)
{
netif->state = NULL;
netif->name[0] = 'm';
netif->name[1] = 's';
netif->output = etharp_output;
netif->linkoutput = low_level_output;
/* set MAC hardware address length */
netif->hwaddr_len = ETHARP_HWADDR_LEN;
/* maximum transfer unit */
netif->mtu = LWIP_NETIF_MTU;
/* device capabilities */
netif->flags = NETIF_FLAG_BROADCAST | NETIF_FLAG_ETHARP;
#if LWIP_IGMP
// also enable multicast
netif->flags |= NETIF_FLAG_IGMP;
#endif
return ERR_OK;
}
/*
networking thread
*/
void AP_Networking_ChibiOS::thread()
{
while (!hal.scheduler->is_system_initialized()) {
hal.scheduler->delay_microseconds(1000);
}
/* start tcpip thread */
tcpip_init(NULL, NULL);
frontend.param.macaddr.get_address(thisif->hwaddr);
struct {
ip4_addr_t ip, gateway, netmask;
} addr {};
if (!frontend.get_dhcp_enabled()) {
addr.ip.addr = htonl(frontend.get_ip_param());
addr.gateway.addr = htonl(frontend.get_gateway_param());
addr.netmask.addr = htonl(frontend.get_netmask_param());
}
const MACConfig mac_config = {thisif->hwaddr};
macStart(&ETHD1, &mac_config);
/* Add interface. */
auto result = netifapi_netif_add(thisif, &addr.ip, &addr.netmask, &addr.gateway, NULL, ethernetif_init, tcpip_input);
if (result != ERR_OK) {
AP_HAL::panic("Failed to initialise netif");
}
netifapi_netif_set_default(thisif);
netifapi_netif_set_up(thisif);
/* Setup event sources.*/
event_timer_t evt;
event_listener_t el0, el1;
evtObjectInit(&evt, LWIP_LINK_POLL_INTERVAL);
evtStart(&evt);
chEvtRegisterMask(&evt.et_es, &el0, PERIODIC_TIMER_ID);
chEvtRegisterMaskWithFlags(macGetEventSource(&ETHD1), &el1,
FRAME_RECEIVED_ID, MAC_FLAGS_RX);
chEvtAddEvents(PERIODIC_TIMER_ID | FRAME_RECEIVED_ID);
while (true) {
eventmask_t mask = chEvtWaitAny(ALL_EVENTS);
if (mask & PERIODIC_TIMER_ID) {
bool current_link_status = macPollLinkStatus(&ETHD1);
if (current_link_status != netif_is_link_up(thisif)) {
if (current_link_status) {
tcpip_callback_with_block((tcpip_callback_fn) netif_set_link_up, thisif, 0);
tcpip_callback_with_block(link_up_cb, this, 0);
}
else {
tcpip_callback_with_block((tcpip_callback_fn) netif_set_link_down, thisif, 0);
tcpip_callback_with_block(link_down_cb, this, 0);
}
}
}
if (mask & FRAME_RECEIVED_ID) {
struct pbuf *p;
while (low_level_input(thisif, &p)) {
if (p != NULL) {
struct eth_hdr *ethhdr = (struct eth_hdr *)p->payload;
switch (htons(ethhdr->type)) {
/* IP or ARP packet? */
case ETHTYPE_IP:
case ETHTYPE_ARP:
/* full packet send to tcpip_thread to process */
if (thisif->input(p, thisif) == ERR_OK) {
break;
}
/* Falls through */
default:
pbuf_free(p);
}
}
}
}
}
}
/*
update called at 10Hz
*/
*/
void AP_Networking_ChibiOS::update()
{
const uint32_t ip = ntohl(lwipGetIp());
const uint32_t nm = ntohl(lwipGetNetmask());
const uint32_t gw = ntohl(lwipGetGateway());
const uint32_t ip = ntohl(thisif->ip_addr.addr);
const uint32_t nm = ntohl(thisif->netmask.addr);
const uint32_t gw = ntohl(thisif->gw.addr);
if (ip != activeSettings.ip ||
nm != activeSettings.nm ||

View File

@ -18,10 +18,17 @@ public:
private:
bool allocate_buffers(void);
void thread(void);
static void link_up_cb(void*);
static void link_down_cb(void*);
static int8_t ethernetif_init(struct netif *netif);
static int8_t low_level_output(struct netif *netif, struct pbuf *p);
static bool low_level_input(struct netif *netif, struct pbuf **pbuf);
private:
struct lwipthread_opts *lwip_options;
uint8_t macaddr[6];
struct netif *thisif;
};
#endif // AP_NETWORKING_BACKEND_CHIBIOS

View File

@ -12,20 +12,11 @@
#include <netif/ppp/pppapi.h>
#include <netif/ppp/pppos.h>
#include <lwip/tcpip.h>
#include <stdio.h>
extern const AP_HAL::HAL& hal;
#if CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
/*
uint32_t timestamp in smallest available units
*/
uint32_t sys_jiffies(void)
{
return AP_HAL::micros();
}
#endif
#if LWIP_TCPIP_CORE_LOCKING
#define LWIP_TCPIP_LOCK() sys_lock_tcpip_core()
#define LWIP_TCPIP_UNLOCK() sys_unlock_tcpip_core()
@ -42,7 +33,7 @@ uint32_t AP_Networking_PPP::ppp_output_cb(ppp_pcb *pcb, const void *data, uint32
auto &driver = *(AP_Networking_PPP *)ctx;
LWIP_UNUSED_ARG(pcb);
uint32_t remaining = len;
uint8_t *ptr = const_cast<uint8_t*>((const uint8_t *)data);
const uint8_t *ptr = (const uint8_t *)data;
while (remaining > 0) {
auto n = driver.uart->write(ptr, remaining);
if (n > 0) {
@ -132,12 +123,12 @@ void AP_Networking_PPP::ppp_status_callback(struct ppp_pcb_s *pcb, int code, voi
*/
bool AP_Networking_PPP::init()
{
uart = AP::serialmanager().find_serial(AP_SerialManager::SerialProtocol_PPP, 0);
auto &sm = AP::serialmanager();
uart = sm.find_serial(AP_SerialManager::SerialProtocol_PPP, 0);
if (uart == nullptr) {
return false;
}
uart->set_unbuffered_writes(true);
pppif = new netif;
if (pppif == nullptr) {
return false;
@ -163,7 +154,7 @@ bool AP_Networking_PPP::init()
GCS_SEND_TEXT(MAV_SEVERITY_INFO, "PPP: started");
hal.scheduler->thread_create(FUNCTOR_BIND_MEMBER(&AP_Networking_PPP::ppp_loop, void),
"ppp",
2048, AP_HAL::Scheduler::PRIORITY_UART, 0);
2048, AP_HAL::Scheduler::PRIORITY_NET, 0);
return true;
}
@ -186,6 +177,7 @@ void AP_Networking_PPP::ppp_loop(void)
// ensure this thread owns the uart
uart->begin(0);
uart->set_unbuffered_writes(true);
while (true) {
uint8_t buf[1024];
@ -194,9 +186,8 @@ void AP_Networking_PPP::ppp_loop(void)
LWIP_TCPIP_LOCK();
pppos_input(ppp, buf, n);
LWIP_TCPIP_UNLOCK();
hal.scheduler->delay_microseconds(100);
} else {
hal.scheduler->delay_microseconds(1000);
hal.scheduler->delay_microseconds(200);
}
}
}

View File

@ -0,0 +1,365 @@
/*
* Copyright (c) 2001-2003 Swedish Institute of Computer Science.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without modification,
* are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT
* SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
* OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
* IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY
* OF SUCH DAMAGE.
*
* This file is part of the lwIP TCP/IP stack.
*
* Author: Adam Dunkels <adam@sics.se>
*
*/
#pragma once
#ifdef __cplusplus
extern "C"
{
#endif
#ifdef LWIP_OPTTEST_FILE
#include "lwipopts_test.h"
#else /* LWIP_OPTTEST_FILE */
#define LWIP_IPV4 1
#define LWIP_IPV6 0
#define MEM_LIBC_MALLOC 1
#define MEMP_MEM_MALLOC 1
#define LWIP_NETCONN_SEM_PER_THREAD 0
#define NO_SYS 0
#define LWIP_SOCKET (NO_SYS==0)
#define LWIP_NETCONN (NO_SYS==0)
#define LWIP_NETIF_API (NO_SYS==0)
#define LWIP_IGMP LWIP_IPV4
#define LWIP_ICMP LWIP_IPV4
#define LWIP_SNMP LWIP_UDP
#define MIB2_STATS LWIP_SNMP
#ifdef LWIP_HAVE_MBEDTLS
#define LWIP_SNMP_V3 (LWIP_SNMP)
#endif
#define LWIP_DNS LWIP_UDP
#define LWIP_MDNS_RESPONDER LWIP_UDP
#define LWIP_NUM_NETIF_CLIENT_DATA (LWIP_MDNS_RESPONDER)
#define LWIP_HAVE_LOOPIF 1
#define LWIP_NETIF_LOOPBACK 1
#define LWIP_LOOPBACK_MAX_PBUFS 10
#define TCP_LISTEN_BACKLOG 1
#define LWIP_COMPAT_SOCKETS 0
#define LWIP_SO_RCVTIMEO 1
#define LWIP_SO_RCVBUF 1
#define LWIP_TCPIP_CORE_LOCKING 1
#define LWIP_NETIF_LINK_CALLBACK 1
#define LWIP_NETIF_STATUS_CALLBACK 1
#define LWIP_NETIF_EXT_STATUS_CALLBACK 1
#define USE_PPP 1
#define LWIP_TIMEVAL_PRIVATE 0
#define LWIP_FD_SET_PRIVATE 0
#define TCP_WND 12000
#define TCP_SND_BUF 12000
// #define LWIP_DEBUG
#ifdef LWIP_DEBUG
#define LWIP_DBG_MIN_LEVEL 0
#define PPP_DEBUG LWIP_DBG_ON
#define MEM_DEBUG LWIP_DBG_ON
#define MEMP_DEBUG LWIP_DBG_ON
#define PBUF_DEBUG LWIP_DBG_OFF
#define API_LIB_DEBUG LWIP_DBG_OFF
#define API_MSG_DEBUG LWIP_DBG_OFF
#define TCPIP_DEBUG LWIP_DBG_ON
#define NETIF_DEBUG LWIP_DBG_ON
#define SOCKETS_DEBUG LWIP_DBG_OFF
#define DNS_DEBUG LWIP_DBG_OFF
#define AUTOIP_DEBUG LWIP_DBG_OFF
#define DHCP_DEBUG LWIP_DBG_OFF
#define IP_DEBUG LWIP_DBG_OFF
#define IP_REASS_DEBUG LWIP_DBG_OFF
#define ICMP_DEBUG LWIP_DBG_OFF
#define IGMP_DEBUG LWIP_DBG_OFF
#define UDP_DEBUG LWIP_DBG_OFF
#define TCP_DEBUG LWIP_DBG_ON
#define TCP_INPUT_DEBUG LWIP_DBG_ON
#define TCP_OUTPUT_DEBUG LWIP_DBG_ON
#define TCP_RTO_DEBUG LWIP_DBG_OFF
#define TCP_CWND_DEBUG LWIP_DBG_OFF
#define TCP_WND_DEBUG LWIP_DBG_OFF
#define TCP_FR_DEBUG LWIP_DBG_OFF
#define TCP_QLEN_DEBUG LWIP_DBG_OFF
#define TCP_RST_DEBUG LWIP_DBG_OFF
#endif
#define LWIP_DBG_TYPES_ON (LWIP_DBG_ON|LWIP_DBG_TRACE|LWIP_DBG_STATE|LWIP_DBG_FRESH|LWIP_DBG_HALT)
/* ---------- Memory options ---------- */
/* MEM_ALIGNMENT: should be set to the alignment of the CPU for which
lwIP is compiled. 4 byte alignment -> define MEM_ALIGNMENT to 4, 2
byte alignment -> define MEM_ALIGNMENT to 2. */
/* MSVC port: intel processors don't need 4-byte alignment,
but are faster that way! */
#define MEM_ALIGNMENT 4U
/* MEM_SIZE: the size of the heap memory. If the application will send
a lot of data that needs to be copied, this should be set high. */
#define MEM_SIZE 10240
/* MEMP_NUM_PBUF: the number of memp struct pbufs. If the application
sends a lot of data out of ROM (or other static memory), this
should be set high. */
#define MEMP_NUM_PBUF 100
/* MEMP_NUM_RAW_PCB: the number of UDP protocol control blocks. One
per active RAW "connection". */
#define MEMP_NUM_RAW_PCB 3
/* MEMP_NUM_UDP_PCB: the number of UDP protocol control blocks. One
per active UDP "connection". */
#define MEMP_NUM_UDP_PCB 8
/* MEMP_NUM_TCP_PCB: the number of simultaneously active TCP
connections. */
#define MEMP_NUM_TCP_PCB 5
/* MEMP_NUM_TCP_PCB_LISTEN: the number of listening TCP
connections. */
#define MEMP_NUM_TCP_PCB_LISTEN 8
/* MEMP_NUM_TCP_SEG: the number of simultaneously queued TCP
segments. */
#define MEMP_NUM_TCP_SEG 16
/* MEMP_NUM_SYS_TIMEOUT: the number of simultaneously active
timeouts. */
#define MEMP_NUM_SYS_TIMEOUT 17
/* The following four are used only with the sequential API and can be
set to 0 if the application only will use the raw API. */
/* MEMP_NUM_NETBUF: the number of struct netbufs. */
#define MEMP_NUM_NETBUF 100
/* MEMP_NUM_NETCONN: the number of struct netconns. */
#define MEMP_NUM_NETCONN 64
/* MEMP_NUM_TCPIP_MSG_*: the number of struct tcpip_msg, which is used
for sequential API communication and incoming packets. Used in
src/api/tcpip.c. */
#define MEMP_NUM_TCPIP_MSG_API 16
#define MEMP_NUM_TCPIP_MSG_INPKT 16
/* ---------- Pbuf options ---------- */
/* PBUF_POOL_SIZE: the number of buffers in the pbuf pool. */
#define PBUF_POOL_SIZE 120
/* PBUF_POOL_BUFSIZE: the size of each pbuf in the pbuf pool. */
#define PBUF_POOL_BUFSIZE 256
/** SYS_LIGHTWEIGHT_PROT
* define SYS_LIGHTWEIGHT_PROT in lwipopts.h if you want inter-task protection
* for certain critical regions during buffer allocation, deallocation and memory
* allocation and deallocation.
*/
#define SYS_LIGHTWEIGHT_PROT (NO_SYS==0)
/* ---------- TCP options ---------- */
#define LWIP_TCP 1
#define TCP_TTL 255
#define LWIP_ALTCP (LWIP_TCP)
#ifdef LWIP_HAVE_MBEDTLS
#define LWIP_ALTCP_TLS (LWIP_TCP)
#define LWIP_ALTCP_TLS_MBEDTLS (LWIP_TCP)
#endif
/* Controls if TCP should queue segments that arrive out of
order. Define to 0 if your device is low on memory. */
#define TCP_QUEUE_OOSEQ 1
/* TCP Maximum segment size. */
#define TCP_MSS 1024
/* TCP sender buffer space (bytes). */
#ifndef TCP_SND_BUF
#define TCP_SND_BUF 2048
#endif
/* TCP sender buffer space (pbufs). This must be at least = 2 *
TCP_SND_BUF/TCP_MSS for things to work. */
#define TCP_SND_QUEUELEN (4 * TCP_SND_BUF/TCP_MSS)
/* TCP writable space (bytes). This must be less than or equal
to TCP_SND_BUF. It is the amount of space which must be
available in the tcp snd_buf for select to return writable */
#define TCP_SNDLOWAT (TCP_SND_BUF/2)
/* TCP receive window. */
#ifndef TCP_WND
#define TCP_WND (20 * 1024)
#endif
/* Maximum number of retransmissions of data segments. */
#define TCP_MAXRTX 12
/* Maximum number of retransmissions of SYN segments. */
#define TCP_SYNMAXRTX 4
/* ---------- ARP options ---------- */
#define LWIP_ARP 1
#define ARP_TABLE_SIZE 10
#define ARP_QUEUEING 1
/* ---------- IP options ---------- */
/* Define IP_FORWARD to 1 if you wish to have the ability to forward
IP packets across network interfaces. If you are going to run lwIP
on a device with only one network interface, define this to 0. */
#define IP_FORWARD 1
/* IP reassembly and segmentation.These are orthogonal even
* if they both deal with IP fragments */
#define IP_REASSEMBLY 1
#define IP_REASS_MAX_PBUFS (10 * ((1500 + PBUF_POOL_BUFSIZE - 1) / PBUF_POOL_BUFSIZE))
#define MEMP_NUM_REASSDATA IP_REASS_MAX_PBUFS
#define IP_FRAG 1
#define IPV6_FRAG_COPYHEADER 1
/* ---------- ICMP options ---------- */
#define ICMP_TTL 255
/* ---------- DHCP options ---------- */
/* Define LWIP_DHCP to 1 if you want DHCP configuration of
interfaces. */
#define LWIP_DHCP 1
/* 1 if you want to do an ARP check on the offered address
(recommended). */
#define DHCP_DOES_ARP_CHECK (LWIP_DHCP)
/* ---------- AUTOIP options ------- */
#define LWIP_AUTOIP (LWIP_DHCP)
#define LWIP_DHCP_AUTOIP_COOP (LWIP_DHCP && LWIP_AUTOIP)
/* ---------- UDP options ---------- */
#define LWIP_UDP 1
#define LWIP_UDPLITE LWIP_UDP
#define UDP_TTL 255
/* ---------- RAW options ---------- */
#define LWIP_RAW 1
/* ---------- Statistics options ---------- */
#define LWIP_STATS 1
#define LWIP_STATS_DISPLAY 1
#if LWIP_STATS
#define LINK_STATS 1
#define IP_STATS 1
#define ICMP_STATS 1
#define IGMP_STATS 1
#define IPFRAG_STATS 1
#define UDP_STATS 1
#define TCP_STATS 1
#define MEM_STATS 1
#define MEMP_STATS 1
#define PBUF_STATS 1
#define SYS_STATS 1
#endif /* LWIP_STATS */
/* ---------- NETBIOS options ---------- */
#define LWIP_NETBIOS_RESPOND_NAME_QUERY 1
/* ---------- PPP options ---------- */
#define PPP_SUPPORT 1 /* Set > 0 for PPP */
#if PPP_SUPPORT
#define NUM_PPP 1 /* Max PPP sessions. */
/* Select modules to enable. Ideally these would be set in the makefile but
* we're limited by the command line length so you need to modify the settings
* in this file.
*/
#define PPPOE_SUPPORT 0
#define PPPOS_SUPPORT 1
#define PAP_SUPPORT 0 /* Set > 0 for PAP. */
#define CHAP_SUPPORT 0 /* Set > 0 for CHAP. */
#define MSCHAP_SUPPORT 0 /* Set > 0 for MSCHAP */
#define CBCP_SUPPORT 0 /* Set > 0 for CBCP (NOT FUNCTIONAL!) */
#define CCP_SUPPORT 0 /* Set > 0 for CCP */
#define VJ_SUPPORT 1 /* Set > 0 for VJ header compression. */
#define MD5_SUPPORT 0 /* Set > 0 for MD5 (see also CHAP) */
#endif /* PPP_SUPPORT */
#endif /* LWIP_OPTTEST_FILE */
/* The following defines must be done even in OPTTEST mode: */
#if !defined(NO_SYS) || !NO_SYS /* default is 0 */
void sys_check_core_locking(void);
#define LWIP_ASSERT_CORE_LOCKED() sys_check_core_locking()
#endif
#ifndef LWIP_PLATFORM_ASSERT
/* Define LWIP_PLATFORM_ASSERT to something to catch missing stdio.h includes */
void ap_networking_platform_assert(const char *msg, int line, const char *file);
#define LWIP_PLATFORM_ASSERT(x) ap_networking_platform_assert(x, __LINE__, __FILE__)
#endif
/*
map LWIP debugging onto ap_networking_printf to allow for easier
redirection to a file or dedicated serial port
*/
#ifdef __cplusplus
extern "C" {
#endif
int ap_networking_printf(const char *fmt, ...);
#ifdef __cplusplus
}
#endif
#define LWIP_PLATFORM_DIAG(x) do {ap_networking_printf x; } while(0)
// #define AP_NETWORKING_LWIP_DEBUG_FILE "lwip.log"
#ifdef __cplusplus
}
#endif

View File

@ -0,0 +1,538 @@
/*
port of lwip to ArduPilot AP_HAL
*/
#include <AP_HAL/AP_HAL.h>
#include <AP_HAL/Semaphores.h>
#include <AP_Math/AP_Math.h>
#include <string.h>
#include <sys/time.h>
#include <sys/types.h>
#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <pthread.h>
#include <errno.h>
#include <lwipopts.h>
extern "C" {
#include "lwip/debug.h"
#include "lwip/def.h"
#include "lwip/sys.h"
#include "lwip/opt.h"
#include "lwip/stats.h"
#include "lwip/tcpip.h"
}
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
#include <semaphore.h>
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
#include "hal.h"
#include "../../../libraries/AP_HAL_ChibiOS/hwdef/common/stm32_util.h"
#endif
extern const AP_HAL::HAL &hal;
unsigned int
lwip_port_rand(void)
{
return (u32_t)rand();
}
static HAL_Semaphore lwprot_mutex;
static HAL_Semaphore tcpip_mutex;
struct sys_mbox_msg {
struct sys_mbox_msg *next;
void *msg;
};
#define SYS_MBOX_SIZE 128
struct sys_mbox {
int first, last;
void *msgs[SYS_MBOX_SIZE];
struct sys_sem *not_empty;
struct sys_sem *not_full;
struct sys_sem *mutex;
int wait_send;
};
struct sys_sem {
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
sem_t sem;
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
semaphore_t sem;
#else
#error "Need sys_sem implementation"
#endif
};
static struct sys_sem *sys_sem_new_internal(u8_t count);
static void sys_sem_free_internal(struct sys_sem *sem);
/* Threads */
struct thread_wrapper_data {
lwip_thread_fn function;
void *arg;
};
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
static void *
thread_wrapper(void *arg)
{
auto *thread_data = (struct thread_wrapper_data *)arg;
thread_data->function(thread_data->arg);
return NULL;
}
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
static void
thread_wrapper(void *arg)
{
auto *thread_data = (struct thread_wrapper_data *)arg;
thread_data->function(thread_data->arg);
}
#endif
sys_thread_t
sys_thread_new(const char *name, lwip_thread_fn function, void *arg, int stacksize, int prio)
{
sys_thread_t ret = nullptr;
struct thread_wrapper_data *thread_data;
thread_data = new thread_wrapper_data;
thread_data->arg = arg;
thread_data->function = function;
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
pthread_t t;
if (pthread_create(&t, NULL, thread_wrapper, thread_data) == 0) {
pthread_setname_np(t, name);
ret = (void*)t;
}
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
ret = thread_create_alloc(THD_WORKING_AREA_SIZE(stacksize+1024),
name,
prio+60, // need to use HAL thread call
thread_wrapper,
thread_data);
#endif
if (ret == nullptr) {
AP_HAL::panic("Failed to create thread %s", name);
}
return ret;
}
void sys_lock_tcpip_core(void)
{
if (hal.scheduler != nullptr) {
tcpip_mutex.take_blocking();
}
}
void sys_unlock_tcpip_core(void)
{
if (hal.scheduler != nullptr) {
tcpip_mutex.give();
}
}
void sys_mark_tcpip_thread(void)
{
}
void sys_check_core_locking(void)
{
/* Embedded systems should check we are NOT in an interrupt
* context here */
}
/*-----------------------------------------------------------------------------------*/
/* Mailbox */
err_t
sys_mbox_new(struct sys_mbox **mb, int size)
{
struct sys_mbox *mbox;
LWIP_UNUSED_ARG(size);
mbox = new sys_mbox;
if (mbox == NULL) {
return ERR_MEM;
}
mbox->first = mbox->last = 0;
mbox->not_empty = sys_sem_new_internal(0);
mbox->not_full = sys_sem_new_internal(0);
mbox->mutex = sys_sem_new_internal(1);
mbox->wait_send = 0;
*mb = mbox;
return ERR_OK;
}
void
sys_mbox_free(struct sys_mbox **mb)
{
if ((mb != NULL) && (*mb != SYS_MBOX_NULL)) {
struct sys_mbox *mbox = *mb;
sys_arch_sem_wait(&mbox->mutex, 0);
sys_sem_free_internal(mbox->not_empty);
sys_sem_free_internal(mbox->not_full);
sys_sem_free_internal(mbox->mutex);
mbox->not_empty = mbox->not_full = mbox->mutex = NULL;
/* LWIP_DEBUGF("sys_mbox_free: mbox 0x%lx\n", mbox); */
delete mbox;
}
}
err_t
sys_mbox_trypost(struct sys_mbox **mb, void *msg)
{
u8_t first;
struct sys_mbox *mbox;
LWIP_ASSERT("invalid mbox", (mb != NULL) && (*mb != NULL));
mbox = *mb;
sys_arch_sem_wait(&mbox->mutex, 0);
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_trypost: mbox %p msg %p\n",
(void *)mbox, (void *)msg));
if ((mbox->last + 1) >= (mbox->first + SYS_MBOX_SIZE)) {
sys_sem_signal(&mbox->mutex);
return ERR_MEM;
}
mbox->msgs[mbox->last % SYS_MBOX_SIZE] = msg;
if (mbox->last == mbox->first) {
first = 1;
} else {
first = 0;
}
mbox->last++;
if (first) {
sys_sem_signal(&mbox->not_empty);
}
sys_sem_signal(&mbox->mutex);
return ERR_OK;
}
void
sys_mbox_post(struct sys_mbox **mb, void *msg)
{
u8_t first;
struct sys_mbox *mbox;
LWIP_ASSERT("invalid mbox", (mb != NULL) && (*mb != NULL));
mbox = *mb;
sys_arch_sem_wait(&mbox->mutex, 0);
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_post: mbox %p msg %p\n", (void *)mbox, (void *)msg));
while ((mbox->last + 1) >= (mbox->first + SYS_MBOX_SIZE)) {
mbox->wait_send++;
sys_sem_signal(&mbox->mutex);
sys_arch_sem_wait(&mbox->not_full, 0);
sys_arch_sem_wait(&mbox->mutex, 0);
mbox->wait_send--;
}
mbox->msgs[mbox->last % SYS_MBOX_SIZE] = msg;
if (mbox->last == mbox->first) {
first = 1;
} else {
first = 0;
}
mbox->last++;
if (first) {
sys_sem_signal(&mbox->not_empty);
}
sys_sem_signal(&mbox->mutex);
}
u32_t
sys_arch_mbox_tryfetch(struct sys_mbox **mb, void **msg)
{
struct sys_mbox *mbox = *mb;
sys_arch_sem_wait(&mbox->mutex, 0);
if (mbox->first == mbox->last) {
sys_sem_signal(&mbox->mutex);
return SYS_MBOX_EMPTY;
}
if (msg != NULL) {
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_tryfetch: mbox %p msg %p\n", (void *)mbox, *msg));
*msg = mbox->msgs[mbox->first % SYS_MBOX_SIZE];
} else {
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_tryfetch: mbox %p, null msg\n", (void *)mbox));
}
mbox->first++;
if (mbox->wait_send) {
sys_sem_signal(&mbox->not_full);
}
sys_sem_signal(&mbox->mutex);
return 0;
}
u32_t
sys_arch_mbox_fetch(struct sys_mbox **mb, void **msg, u32_t timeout)
{
struct sys_mbox *mbox;
LWIP_ASSERT("invalid mbox", (mb != NULL) && (*mb != NULL));
mbox = *mb;
/* The mutex lock is quick so we don't bother with the timeout
stuff here. */
sys_arch_sem_wait(&mbox->mutex, 0);
while (mbox->first == mbox->last) {
sys_sem_signal(&mbox->mutex);
/* We block while waiting for a mail to arrive in the mailbox. We
must be prepared to timeout. */
if (timeout != 0) {
u32_t time_needed = sys_arch_sem_wait(&mbox->not_empty, timeout);
if (time_needed == SYS_ARCH_TIMEOUT) {
return SYS_ARCH_TIMEOUT;
}
} else {
sys_arch_sem_wait(&mbox->not_empty, 0);
}
sys_arch_sem_wait(&mbox->mutex, 0);
}
if (msg != NULL) {
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_fetch: mbox %p msg %p\n", (void *)mbox, *msg));
*msg = mbox->msgs[mbox->first % SYS_MBOX_SIZE];
} else {
LWIP_DEBUGF(SYS_DEBUG, ("sys_mbox_fetch: mbox %p, null msg\n", (void *)mbox));
}
mbox->first++;
if (mbox->wait_send) {
sys_sem_signal(&mbox->not_full);
}
sys_sem_signal(&mbox->mutex);
return 0;
}
/*-----------------------------------------------------------------------------------*/
/* Semaphore */
static struct sys_sem *
sys_sem_new_internal(u8_t count)
{
auto *ret = new sys_sem;
if (ret != nullptr) {
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
sem_init(&ret->sem, 0, count);
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
chSemObjectInit(&ret->sem, (cnt_t)count);
#endif
}
return ret;
}
err_t
sys_sem_new(struct sys_sem **sem, u8_t count)
{
*sem = sys_sem_new_internal(count);
if (*sem == NULL) {
return ERR_MEM;
}
return ERR_OK;
}
u32_t
sys_arch_sem_wait(struct sys_sem **s, u32_t timeout_ms)
{
struct sys_sem *sem = *s;
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
if (timeout_ms == 0) {
sem_wait(&sem->sem);
return 0;
}
struct timespec ts;
if (clock_gettime(CLOCK_REALTIME, &ts) != 0) {
return SYS_ARCH_TIMEOUT;
}
ts.tv_sec += timeout_ms/1000UL;
ts.tv_nsec += (timeout_ms % 1000U) * 1000000UL;
if (ts.tv_nsec >= 1000000000L) {
ts.tv_sec++;
ts.tv_nsec -= 1000000000L;
}
auto ret = sem_timedwait(&sem->sem, &ts);
if (ret != 0) {
return SYS_ARCH_TIMEOUT;
}
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
chSysLock();
sysinterval_t tmo = timeout_ms > 0 ? MIN(TIME_MAX_INTERVAL, TIME_MS2I((time_msecs_t)timeout_ms)) : TIME_INFINITE;
if (chSemWaitTimeoutS(&sem->sem, tmo) != MSG_OK) {
chSysUnlock();
return SYS_ARCH_TIMEOUT;
}
chSysUnlock();
#endif
return 0;
}
void
sys_sem_signal(struct sys_sem **s)
{
struct sys_sem *sem = *s;
#if CONFIG_HAL_BOARD == HAL_BOARD_SITL
sem_post(&sem->sem);
#elif CONFIG_HAL_BOARD == HAL_BOARD_CHIBIOS
chSemSignal(&sem->sem);
#endif
}
static void
sys_sem_free_internal(struct sys_sem *sem)
{
delete sem;
}
void
sys_sem_free(struct sys_sem **sem)
{
if ((sem != NULL) && (*sem != SYS_SEM_NULL)) {
sys_sem_free_internal(*sem);
}
}
/*-----------------------------------------------------------------------------------*/
/* Mutex */
/** Create a new mutex
* @param mutex pointer to the mutex to create
* @return a new mutex */
err_t
sys_mutex_new(struct sys_mutex **mutex)
{
auto *sem = new HAL_Semaphore;
if (sem == nullptr) {
return ERR_MEM;
}
*mutex = (struct sys_mutex *)sem;
return ERR_OK;
}
/** Lock a mutex
* @param mutex the mutex to lock */
void
sys_mutex_lock(struct sys_mutex **mutex)
{
auto *sem = (HAL_Semaphore *)*mutex;
if (hal.scheduler != nullptr) {
sem->take_blocking();
}
}
/** Unlock a mutex
* @param mutex the mutex to unlock */
void
sys_mutex_unlock(struct sys_mutex **mutex)
{
auto *sem = (HAL_Semaphore *)*mutex;
if (hal.scheduler != nullptr) {
sem->give();
}
}
/** Delete a mutex
* @param mutex the mutex to delete */
void
sys_mutex_free(struct sys_mutex **mutex)
{
auto *sem = (HAL_Semaphore *)*mutex;
delete sem;
}
/*-----------------------------------------------------------------------------------*/
/* Time */
u32_t
sys_now(void)
{
return AP_HAL::millis();
}
u32_t
sys_jiffies(void)
{
return AP_HAL::micros();
}
/*-----------------------------------------------------------------------------------*/
/* Init */
void
sys_init(void)
{
}
/*-----------------------------------------------------------------------------------*/
/* Critical section */
/** sys_prot_t sys_arch_protect(void)
This optional function does a "fast" critical region protection and returns
the previous protection level. This function is only called during very short
critical regions. An embedded system which supports ISR-based drivers might
want to implement this function by disabling interrupts. Task-based systems
might want to implement this by using a mutex or disabling tasking. This
function should support recursive calls from the same task or interrupt. In
other words, sys_arch_protect() could be called while already protected. In
that case the return value indicates that it is already protected.
sys_arch_protect() is only required if your port is supporting an operating
system.
*/
sys_prot_t
sys_arch_protect(void)
{
if (hal.scheduler != nullptr) {
lwprot_mutex.take_blocking();
}
return 0;
}
/** void sys_arch_unprotect(sys_prot_t pval)
This optional function does a "fast" set of critical region protection to the
value specified by pval. See the documentation for sys_arch_protect() for
more information. This function is only required if your port is supporting
an operating system.
*/
void
sys_arch_unprotect(sys_prot_t pval)
{
LWIP_UNUSED_ARG(pval);
if (hal.scheduler != nullptr) {
lwprot_mutex.give();
}
}

View File

@ -0,0 +1,21 @@
#pragma once
#include <stdint.h>
#include <sys/select.h>
#ifdef __cplusplus
extern "C"
{
#endif
#define LWIP_ERRNO_STDINCLUDE 1
extern unsigned int lwip_port_rand(void);
#define LWIP_RAND() (lwip_port_rand())
typedef uint32_t sys_prot_t;
#ifdef __cplusplus
}
#endif

View File

@ -0,0 +1,49 @@
#pragma once
#ifdef __cplusplus
extern "C"
{
#endif
#define SYS_MBOX_NULL NULL
#define SYS_SEM_NULL NULL
/*typedef u32_t sys_prot_t;*/
struct sys_sem;
typedef struct sys_sem * sys_sem_t;
#define sys_sem_valid(sem) (((sem) != NULL) && (*(sem) != NULL))
#define sys_sem_valid_val(sem) ((sem) != NULL)
#define sys_sem_set_invalid(sem) do { if((sem) != NULL) { *(sem) = NULL; }}while(0)
#define sys_sem_set_invalid_val(sem) do { (sem) = NULL; }while(0)
struct sys_mutex;
typedef struct sys_mutex * sys_mutex_t;
#define sys_mutex_valid(mutex) sys_sem_valid(mutex)
#define sys_mutex_set_invalid(mutex) sys_sem_set_invalid(mutex)
struct sys_mbox;
typedef struct sys_mbox * sys_mbox_t;
#define sys_mbox_valid(mbox) sys_sem_valid(mbox)
#define sys_mbox_valid_val(mbox) sys_sem_valid_val(mbox)
#define sys_mbox_set_invalid(mbox) sys_sem_set_invalid(mbox)
#define sys_mbox_set_invalid_val(mbox) sys_sem_set_invalid_val(mbox)
typedef void *sys_thread_t;
void sys_mark_tcpip_thread(void);
#define LWIP_MARK_TCPIP_THREAD() sys_mark_tcpip_thread()
#if LWIP_TCPIP_CORE_LOCKING
void sys_lock_tcpip_core(void);
#define LOCK_TCPIP_CORE() sys_lock_tcpip_core()
void sys_unlock_tcpip_core(void);
#define UNLOCK_TCPIP_CORE() sys_unlock_tcpip_core()
#endif
struct timeval;
#ifdef __cplusplus
}
#endif

View File

@ -0,0 +1,36 @@
#!/usr/bin/env python3
import pathlib
def configure(cfg):
extra_src = [
'modules/ChibiOS/ext/lwip/src/core/*c',
'modules/ChibiOS/ext/lwip/src/core/ipv4/*c',
'modules/ChibiOS/ext/lwip/src/api/*c',
'modules/ChibiOS/ext/lwip/src/netif/*c',
'modules/ChibiOS/ext/lwip/src/netif/ppp/*c',
]
extra_src_inc = [
'modules/ChibiOS/ext/lwip/src/include',
]
if cfg.env.BOARD_CLASS == "SITL":
extra_src.extend(['libraries/AP_Networking/lwip_hal/arch/*cpp'])
extra_src_inc.extend(['libraries/AP_Networking/config',
'libraries/AP_Networking/lwip_hal/include'])
if cfg.env.BOARD_CLASS == "ChibiOS":
extra_src.extend(['libraries/AP_Networking/lwip_hal/arch/*cpp'])
extra_src_inc.extend(['libraries/AP_Networking/config',
'libraries/AP_Networking/lwip_hal/include'])
cfg.env.AP_LIB_EXTRA_SOURCES['AP_Networking'] = []
for pattern in extra_src:
s = cfg.srcnode.ant_glob(pattern, dir=False, src=True)
for x in s:
cfg.env.AP_LIB_EXTRA_SOURCES['AP_Networking'].append(str(x))
for inc in extra_src_inc:
cfg.env.INCLUDES += [str(cfg.srcnode.make_node(inc))]