从 sockaddr_in* 到 sockaddr* 的类型转换是否违反了“严格的别名规则”？

Question

Is the following type cast from sockaddr_in * to sockaddr * a violation of "strict aliasing rule"?

Example code snippet from "Beej's Guide to Network programming" (version 2.3.23). The typecast is happening at the last line.

...
#include <arpa/inet.h>
#define MYPORT 3490
main()
{
int sockfd;
struct sockaddr_in my_addr;
sockfd = socket(PF_INET, SOCK_STREAM, 0); 
my_addr.sin_family = AF_INET; 
my_addr.sin_port = htons(MYPORT); 
my_addr.sin_addr.s_addr = inet_addr("10.12.110.57");
memset(&(my_addr.sin_zero), ’\0’, 8); 
bind(sockfd, (struct sockaddr *)&my_addr, sizeof(struct sockaddr));

Here, for convenience, I'm including those struct definitions:

struct sockaddr {
    unsigned short    sa_family;    // address family, AF_xxx
    char              sa_data[14];  // 14 bytes of protocol address
};


// IPv4 AF_INET sockets:

struct sockaddr_in {
    short            sin_family;   // e.g. AF_INET, AF_INET6
    unsigned short   sin_port;     // e.g. htons(3490)
    struct in_addr   sin_addr;     // see struct in_addr, below
    char             sin_zero[8];  // zero this if you want to
};

struct in_addr {
    unsigned long s_addr;          // load with inet_pton()
};

Answer 1

A lot of the UNIX networking foundations are built on these sorts of creative "abuses" of structures. This can make using these functions and structures in non C code quite difficult as many languages, like C++, forbid these sorts of arbitrary recasting operations by default.

In C++ you will need to deal with the fact that, yes, technically these are not valid casts, but the UNIX networking specification has been around for decades and is a known commodity. You can make reasonable assumptions about what is and isn't a valid conversion regardless of what the compiler insists.

As always you will need to tread carefully, but especially when doing these sorts of casts.