Differences and Usage of AF_INET and PF_INET in Socket Programming

Historical Context of AF_INET and PF_INET

In socket programming, AF_INET and PF_INET are often confused terms. According to Beej's Guide to Network Programming, it was historically envisioned that an address family (AF) might support multiple protocol families (PF), but this never materialized. Thus, in theory, AF_INET should be used in struct sockaddr_in, while PF_INET is for the socket() function call. However, in practice, they are interchangeable because they are defined as the same value in most systems.

Code Implementation and Equivalence

The Linux kernel source code explicitly shows the equivalence of PF_INET and AF_INET. For instance, in the include/linux/socket.h file of the Linux kernel, the definition is: #define PF_INET AF_INET. This confirms that protocol and address families are identical in implementation, removing any ambiguity in usage. Therefore, developers can safely use AF_INET in both socket() and bind(), as practiced by W. Richard Stevens in his seminal works.

Practical Application Examples

Setting the IP address correctly is a crucial step in socket programming. Below is a complete example using AF_INET to create a socket and bind an IP address:

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>

int main() {
    int sockfd;
    struct sockaddr_in serv_addr;

    // Create a socket using AF_INET as the address family
    sockfd = socket(AF_INET, SOCK_STREAM, 0);
    if (sockfd < 0) {
        perror("socket creation failed");
        exit(EXIT_FAILURE);
    }

    // Set up the server address structure
    memset(&serv_addr, 0, sizeof(serv_addr));
    serv_addr.sin_family = AF_INET; // Use AF_INET for address family
    serv_addr.sin_port = htons(8080); // Set the port number

    // Set the IP address using inet_pton to convert string to network byte order
    if (inet_pton(AF_INET, "192.168.1.100", &serv_addr.sin_addr) <= 0) {
        perror("invalid address");
        close(sockfd);
        exit(EXIT_FAILURE);
    }

    // Bind the socket to the address
    if (bind(sockfd, (struct sockaddr *)&serv_addr, sizeof(serv_addr)) < 0) {
        perror("bind failed");
        close(sockfd);
        exit(EXIT_FAILURE);
    }

    printf("Socket bound successfully to 192.168.1.100:8080\n");
    close(sockfd);
    return 0;
}

In this example, the serv_addr.sin_addr field is set using the inet_pton function, which converts a dotted-decimal string (e.g., "192.168.1.100") to binary in network byte order. This approach avoids the complexity of manually handling sin_addr and ensures cross-platform compatibility.

Summary and Best Practices

In summary, AF_INET and PF_INET are essentially the same in socket programming and can be used interchangeably. It is recommended to follow W. Richard Stevens' practice by consistently using AF_INET to simplify code. For IP address configuration, prefer standard functions like inet_pton to avoid manual byte order issues. By understanding these core concepts, developers can write more robust and maintainable network applications.