1@page page_component_sal Socket Abstraction Layer: SAL 2 3# SAL Introduction 4 5In order to adapt to more network protocol stack types and avoid the system's dependence on a single network protocol stack, the RT-Thread system provides a SAL (Socket Abstraction Layer) components that implement different network protocol stacks or network implementations. The abstraction of the interface provides a set of standard BSD Socket APIs to the upper layer, so that developers only need to care about and use the network interface provided by the network application layer, without concern for the underlying specific network protocol stack type and implementation, which greatly improves the system's compatibility makes it easy for developers to complete protocol stack adaptation and network-related development. Main features of the SAL component are as follows: 6 7- Abstract and unified multiple network protocol stack interfaces;; 8- Provide Socket-level TLS encrypted transport feature; 9- Support standard BSD Socket APIs; 10- Unified FD management for easy operation of network functions using read/write poll/select; 11 12## SAL Network Framework 13 14The SAL network framework of RT-Thread is mainly shown in the following structure: 15 16 17 18The top layer is the network's application layer, provides a set of standard BSD Socket APIs, such as `socket`, `connect` and other functions, for most of the system network development applications. 19 20The second part is the virtual file system layer. In the RT-Thread system, the DFS file system program can implement different file system operations using standard interface functions.The network socket interface also supports the file system structure. The network socket descriptor created when using the network socket interface is uniformly managed by the file system. Therefore, the network socket descriptor can also use the standard file operation interface.The interfaces provided for the upper application layer are: `read`, `write`, `close`, `poll`/`select`, and so on. 21 22The third part is the socket abstraction layer, through which RT-thread system can adapt to different network protocol stacks in the lower layer and provide unified network programming interface in the upper layer to facilitate access of different protocol stacks. The socket abstraction layer provides interfaces for the upper application layer, such as `accept`, `connect`, `send`, `recv`, etc. 23 24The fourth part is the protocol stack layer, which includes several commonly used TCP/IP protocol stacks, such as lwIP, a light TCP/IP protocol stack commonly used in embedded development, and AT Socket network function implementation developed by rt-thread. These protocol stacks or network functions directly contact the hardware to complete the transformation of data from the network layer to the transmission layer. 25 26The network application layer of RT-thread provides interfaces mainly based on the standard BSD Socket API, which ensures that programs can be written on PC, debugged, and then ported to the RT-thread operating system. 27 28## Working Principles 29 30The working principle of SAL component is mainly divided into the following three parts: 31 32- Unified abstract functions of multi-protocol stack access and interface functions; 33- SAL TLS encryption transmission function; 34 35### Multi-Protocol Stack Access and Unified Abstract Function Of Interface Function 36 37For different protocol stacks or network function implementations, the names of network interfaces may be different. Take the `connect` connection function as an example. The interface name in the lwIP protocol stack is `lwip_connect`, and the interface name in the AT Socket network implementation is `at_connect`. The SAL component provides abstraction and unification of the interface of different protocol stacks or networks. When the socket is created, the component **judges the protocol stack or network function used by judging the incoming protocol domain type**, and completes the RT-Thread. 38 39Currently, the protocol stack or network implementation types supported by the SAL component are: **lwIP protocol stack**, **AT Socket protocol stack**, **WIZnet hardware TCP/IP protocol stack**. 40 41```c 42int socket(int domain, int type, int protocol); 43``` 44 45The above is the definition of the socket creation function in the standard BSD Socket API. The `domain` indicates that the protocol domain is also called the protocol domain. It is used to determine which protocol stack or network implementation to use. The domain type used by the AT Socket protocol stack is **AF_AT**, lwIP protocol stack uses protocol domain type **AF_INET**, etc., WIZnet protocol stack uses protocol domain type **AF_WIZ**. 46 47For different software packages, the protocol domain type passed to the socket may be fixed and will not change depending on how the SAL component is accessed. **In order to dynamically adapt access to different protocol stacks or network implementations**, the SAL component provides two protocol domain type matching methods for each protocol stack or network implementation: **Primary protocol domain type and secondary protocol domain type**. When socket is created, it first determines whether the incoming protocol domain type has the supported primary protocol type. If it is, it uses the corresponding protocol stack or network implementation, if not, determine whether the subprotocol cluster type supports. The current system support protocol domain types are as follows: 48 491. lwIP Protocol stack: family = AF_INET、sec_family = AF_INET 50 512. AT Socket Protocol stack: family = AF_AT、sec_family = AF_INET 52 533. WIZnet hardware TCP/IP Protocol stack: family = AF_WIZ、sec_family = AF_INET 54 55 56 57The main function of the SAL component is to unify the underlying BSD Socket API interface. The following takes the `connect` function call flow as an example to illustrate the SAL component function call method: 58 59- `connect`: The abstract BSD Socket API provided by the SAL component for unified FD management; 60- `sal_connect`: The `connect` implementation function in the SAL component that is used to call the `operation` function registered by the underlying stack.。 61- `lwip_connect`: The layer `connect` connection function provided by the underlying protocol stack is registered in the SAL component when the NIC is initialized, and the final call operation function。 62 63```c 64/* SAL component provides BSD Socket APIs for the application layer */ 65int connect(int s, const struct sockaddr *name, socklen_t namelen) 66{ 67 /* Get the SAL socket descriptor */ 68 int socket = dfs_net_getsocket(s); 69 70 /* Execute the sal_connect function with a SAL socket descriptor*/ 71 return sal_connect(socket, name, namelen); 72} 73 74/* SAL component abstract function interface implementation */ 75int sal_connect(int socket, const struct sockaddr *name, socklen_t namelen) 76{ 77 struct sal_socket *sock; 78 struct sal_proto_family *pf; 79 int ret; 80 81 /* Check if the SAL socket structure is normal */ 82 SAL_SOCKET_OBJ_GET(sock, socket); 83 84 /* Check if the current socket network connection status is normal. */ 85 SAL_NETDEV_IS_COMMONICABLE(sock->netdev); 86 /* Check if the underlying operation function corresponding to the current socket is normal. */ 87 SAL_NETDEV_SOCKETOPS_VALID(sock->netdev, pf, connect); 88 89 /* The connect operation function that performs the underlying registration */ 90 ret = pf->skt_ops->connect((int) sock->user_data, name, namelen); 91#ifdef SAL_USING_TLS 92 if (ret >= 0 && SAL_SOCKOPS_PROTO_TLS_VALID(sock, connect)) 93 { 94 if (proto_tls->ops->connect(sock->user_data_tls) < 0) 95 { 96 return -1; 97 } 98 return ret; 99 } 100#endif 101 return ret; 102} 103 104/* The underlying connect function of the lwIP protocol stack function */ 105int lwip_connect(int socket, const struct sockaddr *name, socklen_t namelen) 106{ 107 ... 108} 109``` 110 111### SAL TLS Encrypted Transmission Function 112 113**1. SAL TLS Feature** 114 115In the transmission of protocol data such as TCP and UDP, since the data packet is plaintext, it is likely to be intercepted and parsed by others, which has a great impact on the secure transmission of information. In order to solve such problems, users generally need to add the SSL/TLS protocol between the application layer and the transport layer. 116 117TLS (Transport Layer Security) is a protocol based on the transport layer TCP protocol. Its predecessor is SSL (Secure Socket Layer). Its main function is to encrypt the application layer message asymmetrically and then transmit it by TCP protocol, so as to achieve secure data encryption interaction. 118 119Currently used TLS methods: **MbedTLS, OpenSSL, s2n**, etc., but for different encryption methods, you need to use their specified encryption interface and process for encryption, the migration of some application layer protocols is more complicated. Therefore, the SAL TLS function is generated, the main function is to **provide TLS-encrypted transmission characteristics at the Socket level, abstract multiple TLS processing methods, and provide a unified interface for completing TLS data interaction**. 120 121**2. How to use the SAL TLS feature** 122 123The process is as follows: 124 125- Configure to enable any network protocol stack support (such as lwIP protocol stack); 126 127- Configure to enable the MbedTLS package (currently only supports MbedTLS type encryption); 128 129- Configure to enable SAL_TLS support (as shown in the configuration options section below); 130 131After the configuration is complete, as long as the `protocol` type passed in the socket creation uses **PROTOCOL_TLS** or **PROTOCOL_DTLS **, the standard BSD Socket API interface can be used to complete the establishment of the TLS connection and the data transmission and reception. The sample code is as follows: 132 133```c 134#include <stdio.h> 135#include <string.h> 136 137#include <rtthread.h> 138#include <sys/socket.h> 139#include <netdb.h> 140 141/* RT-Thread offical website,supporting TLS function */ 142#define SAL_TLS_HOST "www.rt-thread.org" 143#define SAL_TLS_PORT 443 144#define SAL_TLS_BUFSZ 1024 145 146static const char *send_data = "GET /download/rt-thread.txt HTTP/1.1\r\n" 147 "Host: www.rt-thread.org\r\n" 148 "User-Agent: rtthread/4.0.1 rtt\r\n\r\n"; 149 150void sal_tls_test(void) 151{ 152 int ret, i; 153 char *recv_data; 154 struct hostent *host; 155 int sock = -1, bytes_received; 156 struct sockaddr_in server_addr; 157 158 /* Get the host address through the function entry parameter url (if it is a domain name, it will do domain name resolution) */ 159 host = gethostbyname(SAL_TLS_HOST); 160 161 recv_data = rt_calloc(1, SAL_TLS_BUFSZ); 162 if (recv_data == RT_NULL) 163 { 164 rt_kprintf("No memory\n"); 165 return; 166 } 167 168 /* Create a socket of type SOCKET_STREAM, TCP protocol, TLS type */ 169 if ((sock = socket(AF_INET, SOCK_STREAM, PROTOCOL_TLS)) < 0) 170 { 171 rt_kprintf("Socket error\n"); 172 goto __exit; 173 } 174 175 /* Initialize the server address */ 176 server_addr.sin_family = AF_INET; 177 server_addr.sin_port = htons(SAL_TLS_PORT); 178 server_addr.sin_addr = *((struct in_addr *)host->h_addr); 179 rt_memset(&(server_addr.sin_zero), 0, sizeof(server_addr.sin_zero)); 180 181 if (connect(sock, (struct sockaddr *)&server_addr, sizeof(struct sockaddr)) < 0) 182 { 183 rt_kprintf("Connect fail!\n"); 184 goto __exit; 185 } 186 187 /* Send data to the socket connection */ 188 ret = send(sock, send_data, strlen(send_data), 0); 189 if (ret <= 0) 190 { 191 rt_kprintf("send error,close the socket.\n"); 192 goto __exit; 193 } 194 195 /* Receive and print the response data, using encrypted data transmission */ 196 bytes_received = recv(sock, recv_data, SAL_TLS_BUFSZ - 1, 0); 197 if (bytes_received <= 0) 198 { 199 rt_kprintf("received error,close the socket.\n"); 200 goto __exit; 201 } 202 203 rt_kprintf("recv data:\n"); 204 for (i = 0; i < bytes_received; i++) 205 { 206 rt_kprintf("%c", recv_data[i]); 207 } 208 209__exit: 210 if (recv_data) 211 rt_free(recv_data); 212 213 if (sock >= 0) 214 closesocket(sock); 215} 216 217#ifdef FINSH_USING_MSH 218#include <finsh.h> 219MSH_CMD_EXPORT(sal_tls_test, SAL TLS function test); 220#endif /* FINSH_USING_MSH */ 221``` 222 223## Configuration Options 224 225When we use the SAL component we need to define the following macro definition in rtconfig.h: 226 227| **Macro definition** | **Description** | 228|--------------------------------|--------------------------------| 229| RT_USING_SAL | Enable the SAL function | 230| SAL_USING_LWIP | Enable lwIP stack support | 231| SAL_USING_AT | Enable AT Socket protocol stack support | 232| SAL_USING_TLS | Enable SAL TLS feature support | 233| SAL_USING_POSIX | Enable POSIX file system related function support, such as read, write, select/poll, etc. | 234 235Currently, the SAL abstraction layer supports the lwIP protocol stack, the AT Socket protocol stack, and the WIZnet hardware TCP/IP protocol stack. To enable SAL in the system, at least one protocol stack support is required. 236 237The above configuration options can be added directly to the `rtconfig.h` file or added by the component package management tool Env configuration option. The specific configuration path in the Env tool is as follows: 238 239```c 240RT-Thread Components ---> 241 Network ---> 242 Socket abstraction layer ---> 243 [*] Enable socket abstraction layer 244 protocol stack implement ---> 245 [ ] Support lwIP stack 246 [ ] Support AT Commands stack 247 [ ] Support MbedTLS protocol 248 [*] Enable BSD socket operated by file system API 249``` 250 251After the configuration is complete, you can use the `scons` command to regenerate the function and complete the addition of the SAL component. 252 253# Initialization 254 255配置开启 SAL 选项之后,需要在启动时对它进行初始化,开启 SAL 功能,如果程序中已经使用了组件自动初始化,则不再需要额外进行单独的初始化,否则需要在初始化任务中调用如下函数: 256 257```c 258int sal_init(void); 259``` 260 261The initialization function is mainly for initializing the SAL component, supporting the component to repeatedly initialize the judgment, and completing the initialization of the resource such as the mutex used in the component. There is no new thread created in the SAL component, which means that the SAL component resource is very small. Currently, the **SAL component resource is occupied by ROM 2.8K and RAM 0.6K**. 262 263 264# BSD Socket API Introduction 265 266The SAL component abstracts the standard BSD Socket API interface. The following is an introduction to common network interfaces: 267 268## Create a Socket (socket) 269 270``` c 271int socket(int domain, int type, int protocol); 272``` 273 274| **Parameter** | **Description** | 275|--------|-------------------------------------| 276| domain | protocol domain type | 277| type | protocol type | 278| protocol | Transport layer protocol for actual use | 279| **back** | -- | 280| >=0 | Success, an integer representing the socket descriptor will be returned | 281| -1 | Fail | 282 283This function is used to assign a socket descriptor and the resources it USES based on the specified address family, data type, and protocol. 284 285**domain ( protocol domain type ):** 286 287- AF_INET: IPv4 288- AF_INET6: IPv6 289 290**type ( protocol type ):** 291 292- SOCK_STREAM:Stream socket 293- SOCK_DGRAM: Datagram socket 294- SOCK_RAW: Raw socket 295 296## Bind Socket (bind) 297 298```c 299int bind(int s, const struct sockaddr *name, socklen_t namelen); 300``` 301 302| **P**arameter | **Description** | 303|---------|---------------------------------------------| 304| s | socket descriptor | 305| name | a pointer to the sockaddr structure representing the address to bind to | 306| namelen | length of sockaddr structure | 307| **back** | -- | 308| 0 | success | 309| -1 | fail | 310 311This function is used to bind the port number and IP address to the specified socket。 312 313SAL components depend on the `netdev` components, when using ` bind () ` function, can get IP address information through the netdev nic name, is used to create Socket bound to the specified object of network interface card. The following example completes the process of binding the IP address of the network interface card and connecting to the server through the name of the incoming network interface card: 314 315```c 316#include <rtthread.h> 317#include <arpa/inet.h> 318#include <netdev.h> 319 320#define SERVER_HOST "192.168.1.123" 321#define SERVER_PORT 1234 322 323static int bing_test(int argc, char **argv) 324{ 325 struct sockaddr_in client_addr; 326 struct sockaddr_in server_addr; 327 struct netdev *netdev = RT_NULL; 328 int sockfd = -1; 329 330 if (argc != 2) 331 { 332 rt_kprintf("bind_test [netdev_name] --bind network interface device by name.\n"); 333 return -RT_ERROR; 334 } 335 336 /* Get the netdev network interface card object by name */ 337 netdev = netdev_get_by_name(argv[1]); 338 if (netdev == RT_NULL) 339 { 340 rt_kprintf("get network interface device(%s) failed.\n", argv[1]); 341 return -RT_ERROR; 342 } 343 344 if ((sockfd = socket(AF_INET, SOCK_STREAM, 0)) < 0) 345 { 346 rt_kprintf("Socket create failed.\n"); 347 return -RT_ERROR; 348 } 349 350 /* Initializes the client address to bind to */ 351 client_addr.sin_family = AF_INET; 352 client_addr.sin_port = htons(8080); 353 /* Gets the IP address information in the network interface card object */ 354 client_addr.sin_addr.s_addr = netdev->ip_addr.addr; 355 rt_memset(&(client_addr.sin_zero), 0, sizeof(client_addr.sin_zero)); 356 357 if (bind(sockfd, (struct sockaddr *)&client_addr, sizeof(struct sockaddr)) < 0) 358 { 359 rt_kprintf("socket bind failed.\n"); 360 closesocket(sockfd); 361 return -RT_ERROR; 362 } 363 rt_kprintf("socket bind network interface device(%s) success!\n", netdev->name); 364 365 /* Initializes the server address for the pre-connection */ 366 server_addr.sin_family = AF_INET; 367 server_addr.sin_port = htons(SERVER_PORT); 368 server_addr.sin_addr.s_addr = inet_addr(SERVER_HOST); 369 rt_memset(&(server_addr.sin_zero), 0, sizeof(server_addr.sin_zero)); 370 371 /* Connect to the server */ 372 if (connect(sockfd, (struct sockaddr *)&server_addr, sizeof(struct sockaddr)) < 0) 373 { 374 rt_kprintf("socket connect failed!\n"); 375 closesocket(sockfd); 376 return -RT_ERROR; 377 } 378 else 379 { 380 rt_kprintf("socket connect success!\n"); 381 } 382 383 /* Close the connection */ 384 closesocket(sockfd); 385 return RT_EOK; 386} 387 388#ifdef FINSH_USING_MSH 389#include <finsh.h> 390MSH_CMD_EXPORT(bing_test, bind network interface device test); 391#endif /* FINSH_USING_MSH */ 392``` 393 394## Listen Socket (listen) 395 396```c 397int listen(int s, int backlog); 398``` 399 400| **Parameter** | **Description** | 401|-------|-------------------------------| 402| s | socket descriptor | 403| backlog | represents the maximum number of connections that can wait at a time | 404| **back** | -- | 405| 0 | success | 406| -1 | fail | 407 408This function is used by the TCP server to listen for a specified socket connection。 409 410## Accept Connection (accept) 411 412```c 413int accept(int s, struct sockaddr *addr, socklen_t *addrlen); 414``` 415 416| **Parameter** | **Descrption** | 417|-------|-------------------------------| 418| s | socket descriptor | 419| addr | represents the maximum number of connections that can wait at a time | 420| addrlen | length of client device address structure | 421| **back** | -- | 422| 0 | success, return the newly created socket descriptor | 423| -1 | fail | 424 425When the application listens for connections from other hosts, the connection is initialized with the `accept()` function, and `accept()` creates a new socket for each connection and removes the connection from the listen queue. 426 427## Establish Connection (connect) 428 429```c 430int connect(int s, const struct sockaddr *name, socklen_t namelen); 431``` 432 433| **Parameter** | **Description** | 434|-------|-------------------------------| 435| s | socket descriptor | 436| name | server address information | 437| namelen | server address structure length | 438| **back** | -- | 439| 0 | successful, return the newly created socket descriptor | 440| -1 | fail | 441 442This function is used to establish a connection to the specified socket. 443 444## Send TCP Data (send) 445 446```c 447int send(int s, const void *dataptr, size_t size, int flags); 448``` 449 450| **Parameter** | **Description** | 451|-------|---------------------------| 452| s | socket descriptor | 453| dataptr | sent data pointer | 454| size | length of sent data | 455| flags | flags, usually 0 | 456| **back** | -- | 457| >0 | success, the length of the sent data is returned | 458| <=0 | Fail | 459 460This function is commonly used to send data over a TCP connection。 461 462## Receive TCP Data (recv) 463 464```c 465int recv(int s, void *mem, size_t len, int flags); 466``` 467 468| **Parameter** | **Description** | 469|-----|---------------------------| 470| s | socket descriptor | 471| mem | received data pointer | 472| len | length of received data | 473| flags | flags, usually 0 | 474| **back** | -- | 475| >0 | success, the length of the received data will be returned | 476| =0 | the destination address has been transmitted and the connection is closed | 477| <0 | fail | 478 479This function is used to receive data over a TCP connection。 480 481## Send UDP Data (sendto) 482 483```c 484int sendto(int s, const void *dataptr, size_t size, int flags, const struct sockaddr *to, socklen_t tolen); 485``` 486 487| **Parameter** | **Description** | 488|-------|---------------------------| 489| s | socket descriptor | 490| dataptr | sent data pointer | 491| size | length of sent data | 492| flags | flags, usually 0 | 493| to | target address structure pointer | 494| tolen | length of the target address structure | 495| **return** | -- | 496| >0 | success, the length of the received data will be returned | 497| <=0 | fail | 498 499This function is used for UDP connections to send data。 500 501## Receive UDP Data (recvfrom) 502 503```c 504int recvfrom(int s, void *mem, size_t len, int flags, struct sockaddr *from, socklen_t *fromlen); 505``` 506 507| **Parameter** | **Description** | 508|-------|---------------------------| 509| s | socket descriptor | 510| mem | received data pointer | 511| len | length of data received | 512| flags | flags, usually 0 | 513| from | received address structure pointer | 514| fromlen | length of received address structure | 515| **back** | -- | 516| >0 | success, return the length of the received data | 517| =0 | the receiving address has been transferred and the connection is closed | 518| <0 | fail | 519 520This function is used to receive data on a UDP connection。 521 522## Close Socket (closesocket) 523 524```c 525int closesocket(int s); 526``` 527 528| **Parameter** | **Description** | 529|----|-------------| 530| s | Socket descriptor | 531| **back** | -- | 532| 0 | success | 533| -1 | fail | 534 535This function is used to close the connection and release the resource.。 536 537## Shutdown The Socket By Setting(shutdown) 538 539```c 540int shutdown(int s, int how); 541``` 542 543| **Parameter** | **Descrption** | 544|----|-----------------| 545| s | socket descriptor | 546| how | socket control | 547| **back** | -- | 548| 0 | fail | 549| -1 | success | 550 551This function provides more permissions to control the closing process of the socket.。 552 553**How ( socket control ):** 554 555- 0: Stop receiving current data and reject future data reception; 556- 1: Stop sending data and discard unsent data; 557- 2: Stop receiving and sending data。 558 559## Set Socket Options(setsockopt) 560 561```c 562int setsockopt(int s, int level, int optname, const void *optval, socklen_t optlen); 563``` 564 565| **Parameter** | **Description** | 566|-------|-----------------------| 567| s | socket descriptor | 568| level | protocol stack configuration options | 569| optname | option name to be set | 570| optval | set the buffer address of the option value | 571| optlen | set the buffer length of the option value | 572| **back** | -- | 573| =0 | success | 574| <0 | fail | 575 576This function is used to set the socket mode and modify the socket configuration options.。 577 578**level ( protocol stack configuration options ):** 579 580- SOL_SOCKET:Socket layer 581- IPPROTO_TCP:TCP layer 582- IPPROTO_IP:IP layer 583 584**optname ( Option name to be set ):** 585 586- SO_KEEPALIVE:Set keep alive options 587- SO_RCVTIMEO:Set socket data reception timeout 588- SO_SNDTIMEO:Set socket data sending timeout 589 590## Get Socket Options(getsockopt) 591 592```c 593int getsockopt(int s, int level, int optname, void *optval, socklen_t *optlen); 594``` 595 596| **Parameter** | **Description** | 597|-------|---------------------------| 598| s | socket descriptor | 599| level | protocol stack configuration options | 600| optname | option name to be set | 601| optval | get the buffer address of the option value | 602| optlen | get the buffer length address of the option value | 603| **back** | -- | 604| =0 | success | 605| <0 | fail | 606 607This function is used to get the socket configuration options。 608 609## Get Remote Address Information (getpeername) 610 611```c 612int getpeername(int s, struct sockaddr *name, socklen_t *namelen); 613``` 614 615| **Parameter** | **Description** | 616|-------|-------------------------| 617| s | socket descriptor | 618| name | received address structure pointer | 619| namelen | length of received address structure | 620| **back** | -- | 621| =0 | success | 622| <0 | fail | 623 624This function is used to get the remote address information associated with the socket。 625 626## Get Local Address Information (getsockname) 627 628```c 629int getsockname(int s, struct sockaddr *name, socklen_t *namelen); 630``` 631 632| **Parameter** | **Description** | 633|-------|-------------------------| 634| s | Socket descriptor | 635| name | received address structure pointer | 636| namelen | length of received address structure | 637| **back** | -- | 638| =0 | success | 639| <0 | fail | 640 641This function is used to get local socket address information。 642 643## Configure Socket Parameters (ioctlsocket)) 644 645```c 646int ioctlsocket(int s, long cmd, void *arg); 647``` 648 649| **Parameter** | **Description** | 650|-----|-----------------| 651| s | socket descriptor | 652| cmd | socket operation command | 653| arg | operation command's parameters | 654| **back** | -- | 655| =0 | success | 656| <0 | fail | 657 658This function sets the socket control mode。 659 660**The CMD supports following commands** 661 662- FIONBIO: Turns on or off the socket's non-blocking mode. Arg parameter 1 is open non-blocking and 0 is closed non-blocking. 663 664# Network Protocol Stack Access 665 666Access to the network protocol stack or network function implementation is mainly to initialize and register the protocol cluster structure, and add it to the protocol cluster list in SAL component. The protocol cluster structure is defined as follows: 667 668```c 669/* network interface socket opreations */ 670struct sal_socket_ops 671{ 672 int (*socket) (int domain, int type, int protocol); 673 int (*closesocket)(int s); 674 int (*bind) (int s, const struct sockaddr *name, socklen_t namelen); 675 int (*listen) (int s, int backlog); 676 int (*connect) (int s, const struct sockaddr *name, socklen_t namelen); 677 int (*accept) (int s, struct sockaddr *addr, socklen_t *addrlen); 678 int (*sendto) (int s, const void *data, size_t size, int flags, const struct sockaddr *to, socklen_t tolen); 679 int (*recvfrom) (int s, void *mem, size_t len, int flags, struct sockaddr *from, socklen_t *fromlen); 680 int (*getsockopt) (int s, int level, int optname, void *optval, socklen_t *optlen); 681 int (*setsockopt) (int s, int level, int optname, const void *optval, socklen_t optlen); 682 int (*shutdown) (int s, int how); 683 int (*getpeername)(int s, struct sockaddr *name, socklen_t *namelen); 684 int (*getsockname)(int s, struct sockaddr *name, socklen_t *namelen); 685 int (*ioctlsocket)(int s, long cmd, void *arg); 686#ifdef SAL_USING_POSIX 687 int (*poll) (struct dfs_file *file, struct rt_pollreq *req); 688#endif 689}; 690 691/* sal network database name resolving */ 692struct sal_netdb_ops 693{ 694 struct hostent* (*gethostbyname) (const char *name); 695 int (*gethostbyname_r)(const char *name, struct hostent *ret, char *buf, size_t buflen, struct hostent **result, int *h_errnop); 696 int (*getaddrinfo) (const char *nodename, const char *servname, const struct addrinfo *hints, struct addrinfo **res); 697 void (*freeaddrinfo) (struct addrinfo *ai); 698}; 699 700/* Protocol domain structure definition */ 701struct sal_proto_family 702{ 703 int family; /* primary protocol families type */ 704 int sec_family; /* secondary protocol families type */ 705 const struct sal_socket_ops *skt_ops; /* socket opreations */ 706 const struct sal_netdb_ops *netdb_ops; /* network database opreations */ 707}; 708 709``` 710 711- `family`: Each protocol stack supports the main protocol cluster types, such as AF_INET for lwIP, AF_AT Socket, and AF_WIZ for WIZnet。 712- `sec_family`:The type of sub-protocol domain supported by each protocol stack, used to support a single protocol stack or network implementation, that matches other types of protocol cluster types in the package。 713- `skt_ops`: Define socket related functions, such as connect, send, recv, etc., each protocol cluster has a different set of implementation。 714- `netdb_ops`:Define non-socket-related execution functions, such as `gethostbyname`, `getaddrinfo`, `freeaddrinfo`, etc. Each protocol cluster has a different set of implementations。 715 716The following is the access registration process implemented by AT Socket network. Developers can refer to other protocol stacks or network implementations for access: 717 718```c 719#include <rtthread.h> 720#include <netdb.h> 721#include <sal_low_lvl.h> /* SAL component structure holds the header file */ 722#include <at_socket.h> /* AT Socket related header file */ 723#include <af_inet.h> 724 725#include <netdev.h> /*network interface card function related header file */ 726 727#ifdef SAL_USING_POSIX 728#include <dfs_poll.h> /* Poll function related header file */ 729#endif 730 731#ifdef SAL_USING_AT 732 733/* A custom poll execution function that handles the events received in the poll */ 734static int at_poll(struct dfs_file *file, struct rt_pollreq *req) 735{ 736 int mask = 0; 737 struct at_socket *sock; 738 struct socket *sal_sock; 739 740 sal_sock = sal_get_socket((int) file->data); 741 if(!sal_sock) 742 { 743 return -1; 744 } 745 746 sock = at_get_socket((int)sal_sock->user_data); 747 if (sock != NULL) 748 { 749 rt_base_t level; 750 751 rt_poll_add(&sock->wait_head, req); 752 753 level = rt_hw_interrupt_disable(); 754 if (sock->rcvevent) 755 { 756 mask |= POLLIN; 757 } 758 if (sock->sendevent) 759 { 760 mask |= POLLOUT; 761 } 762 if (sock->errevent) 763 { 764 mask |= POLLERR; 765 } 766 rt_hw_interrupt_enable(level); 767 } 768 769 return mask; 770} 771#endif 772 773/* Define and assign socket execution functions, and the SAL component calls the underlying function of the registration when it executes relevant functions */ 774static const struct proto_ops at_inet_stream_ops = 775{ 776 at_socket, 777 at_closesocket, 778 at_bind, 779 NULL, 780 at_connect, 781 NULL, 782 at_sendto, 783 at_recvfrom, 784 at_getsockopt, 785 at_setsockopt, 786 at_shutdown, 787 NULL, 788 NULL, 789 NULL, 790 791#ifdef SAL_USING_POSIX 792 at_poll, 793#else 794 NULL, 795#endif /* SAL_USING_POSIX */ 796}; 797 798static const struct sal_netdb_ops at_netdb_ops = 799{ 800 at_gethostbyname, 801 NULL, 802 at_getaddrinfo, 803 at_freeaddrinfo, 804}; 805 806/* define and assign AT Socket protocol domain structure */ 807static const struct sal_proto_family at_inet_family = 808{ 809 AF_AT, 810 AF_INET, 811 &at_socket_ops, 812 &at_netdb_ops, 813}; 814 815/* Used to set the protocol domain information in the network interface card device */ 816int sal_at_netdev_set_pf_info(struct netdev *netdev) 817{ 818 RT_ASSERT(netdev); 819 820 netdev->sal_user_data = (void *) &at_inet_family; 821 return 0; 822} 823 824#endif /* SAL_USING_AT */ 825``` 826