2010年9月14日 星期二

Using socket RAW to send an ICMP protocal

this is a simple program that i found on the internet showing how to send ICMP using RAW socket.
example code:
/*
 *    pinger.c 
 *    This is a ping imitation program 
 *    It will send an ICMP ECHO packet to the server of 
 *    your choice and listen for an ICMP REPLY packet
 *    Have fun!
 */
/*
 *    pinger.c 
 *    This is a ping imitation program 
 *    It will send an ICMP ECHO packet to the server of 
 *    your choice and listen for an ICMP REPLY packet
 *    Have fun!
 */
#include <stdio.h>
#include <stdlib.h>

#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <netdb.h>
#include <linux/ip.h>

#include <linux/icmp.h>
#include <string.h>
#include <unistd.h>


char dst_addr[15];

char src_addr[15];

unsigned short in_cksum(unsigned short *, int);

void parse_argvs(char**, char*, char* );

void usage();
char* getip();

int main(int argc, char* argv[])

{
    struct iphdr* ip;
    struct iphdr* ip_reply;

    struct icmphdr* icmp;
    struct sockaddr_in connection;

    char* packet;
    char* buffer;
    int sockfd;

    int optval;
    int addrlen;
    
    if (getuid() != 0)

    {
 fprintf(stderr, "%s: root privelidges needed\n", *(argv + 0));

 exit(EXIT_FAILURE);
    }

    parse_argvs(argv, dst_addr, src_addr);

    printf("Source address: %s\n", src_addr);
    printf("Destination address: %s\n", dst_addr);

    
    /*
     * allocate all necessary memory
    */
    ip = malloc(sizeof(struct iphdr));

    ip_reply = malloc(sizeof(struct iphdr));
    icmp = malloc(sizeof(struct icmphdr));

    packet = malloc(sizeof(struct iphdr) + sizeof(struct icmphdr));

    buffer = malloc(sizeof(struct iphdr) + sizeof(struct icmphdr));

    /****************************************************************/
    
    ip = (struct iphdr*) packet;

    icmp = (struct icmphdr*) (packet + sizeof(struct iphdr));

    
    /*  
     * here the ip packet is set up except checksum
     */
    ip->ihl   = 5;
    ip->version   = 4;

    ip->tos   = 0;
    ip->tot_len   = sizeof(struct iphdr) + sizeof(struct icmphdr);

    ip->id   = htons(random());
    ip->ttl   = 255;

    ip->protocol  = IPPROTO_ICMP;
    ip->saddr   = inet_addr(src_addr);

    ip->daddr   = inet_addr(dst_addr);

    
    if ((sockfd = socket(AF_INET, SOCK_RAW, IPPROTO_ICMP)) == -1)

    {
 perror("socket");
 exit(EXIT_FAILURE);

    }
    
    /* 
     * IP_HDRINCL must be set on the socket so that
     * the kernel does not attempt to automatically add
     * a default ip header to the packet
     */
    
    setsockopt(sockfd, IPPROTO_IP, IP_HDRINCL, &optval, sizeof(int));

    
    /*
     * here the icmp packet is created
     * also the ip checksum is generated
     */
    icmp->type   = ICMP_ECHO;
    icmp->code   = 0;

    icmp->un.echo.id  = 0;
    icmp->un.echo.sequence = 0;

    icmp->checksum   = 0;
    icmp-> checksum  = in_cksum((unsigned short *)icmp, sizeof(struct icmphdr));

    
    ip->check   = in_cksum((unsigned short *)ip, sizeof(struct iphdr));

    
    connection.sin_family = AF_INET;
    connection.sin_addr.s_addr = inet_addr(dst_addr);

    
    /*
     * now the packet is sent
     */
    
    sendto(sockfd, packet, ip->tot_len, 0, (struct sockaddr *)&connection, sizeof(struct sockaddr));

    printf("Sent %d byte packet to %s\n", sizeof(packet), dst_addr);
    
    /*
     * now we listen for responses
     */

    addrlen = sizeof(connection);
    if (recvfrom(sockfd, buffer, sizeof(struct iphdr) + sizeof(struct icmphdr), 0, (struct sockaddr *)&connection, &addrlen) == -1)

    {
 perror("recv");
    }
    else
    {

 printf("Received %d byte reply from %s:\n", sizeof(buffer), dst_addr);
        ip_reply = (struct iphdr*) buffer;

 printf("ID: %d\n", ntohs(ip_reply->id));
 printf("TTL: %d\n", ip_reply->ttl);

    }
    close(sockfd);
    return 0;
}

void parse_argvs(char** argv, char* dst, char* src)

{
    int i;
    if(!(*(argv + 1))) 
    {

 /* there are no options on the command line */
 usage();
 exit(EXIT_FAILURE); 
    }

    if (*(argv + 1) && (!(*(argv + 2)))) 
    {

 /* 
  *   only one argument provided
  *   assume it is the destination server
  *   source address is local host
  */
 strncpy(dst, *(argv + 1), 15);

 strncpy(src, getip(), 15);
 return;

    }
    else if ((*(argv + 1) && (*(argv + 2))))

    {
 /* 
  *    both the destination and source address are defined
  *    for now only implemented is a source address and 
  *    destination address
  */
 strncpy(dst, *(argv + 1), 15);

 i = 2;
 while(*(argv + i + 1))

 {
     if (strncmp(*(argv + i), "-s", 2) == 0)

     {
  strncpy(src, *(argv + i + 1), 15);

  break;
     }
     i++;
 }

    }
}

void usage()
{
    fprintf(stderr, "\nUsage: pinger [destination] <-s [source]>\n");

    fprintf(stderr, "Destination must be provided\n");
    fprintf(stderr, "Source is optional\n\n");

}

char* getip()
{
    char buffer[256];

    struct hostent* h;
    
    gethostname(buffer, 256);

    h = gethostbyname(buffer);
    
    return inet_ntoa(*(struct in_addr *)h->h_addr);

    
}
/*
 * in_cksum --
 * Checksum routine for Internet Protocol
 * family headers (C Version)
 */
unsigned short in_cksum(unsigned short *addr, int len)

{
    register int sum = 0;
    u_short answer = 0;

    register u_short *w = addr;
    register int nleft = len;

    /*
     * Our algorithm is simple, using a 32 bit accumulator (sum), we add
     * sequential 16 bit words to it, and at the end, fold back all the
     * carry bits from the top 16 bits into the lower 16 bits.
     */
    while (nleft > 1)
    {
   sum += *w++;

   nleft -= 2;
    }
    /* mop up an odd byte, if necessary */
    if (nleft == 1)

    {
   *(u_char *) (&answer) = *(u_char *) w;

   sum += answer;
    }
    /* add back carry outs from top 16 bits to low 16 bits */
    sum = (sum >> 16) + (sum & 0xffff);  /* add hi 16 to low 16 */

    sum += (sum >> 16);    /* add carry */
    answer = ~sum;    /* truncate to 16 bits */

    return (answer);
}



2010年5月3日 星期一

Stopping TCP retransmit

The TCP keep-alive is a great way to detect TCP connection lost; if a connection list lost, "select()" will show readable data, yet the data length returned by "recv()" is less or equal to zero. Yet, if you've already sent a data during the connection lost, the TCP retransmit is triggered, stopping the  mechanism of TCP keep-alive, the bad news is that the retransmit time can go on to 15-20 mins, and keep-alive will not be invoked during this time period.
In other words, you will not notice the connection lost unless you full the Tx buffer of the net device driver(the default size is 4096 bytes).
If you have little or no data to send, detecting the connection loss will take forever, and become a pain in the ass.
To solve this issue,you can reduce the tcp retransmit times by doing:
    system("sysctl -w net.ipv4.tcp_retries1=2");
    system("sysctl -w net.ipv4.tcp_retries2=3");

yet the effect is system wide.... not a very elegant solution.

P.S.
It was interesting to find out that linux socket always tries to send every packet, even if "netif_queue_stopped(netdev *)" (used for flow control on the Tx side of the driver) or "netif_carrier_off(netdev *)"(used when carrier is down ex. unplugged cable) is called by the net device driver, Linux queues the unsent packets and sends it the instance when carrier or queue is available. This mechanism guarantees the message you sent is delivered to the destination if possible.