On Thu, Nov 26, 2015 at 11:01:21AM +0000, Umair Ali wrote:
> Hello Xenomai team,
Hi, we received your message three times. One time would have been
enough.
>
> I am working on sending the raw Ethernet packets through RTnet
> driver (rt_8139t). I am using the the example i.e. raw-ethernet.c
> as my starting point for sending the raw Ethernet packets from
> Rtnet source file.
So, you are not using the advices I gave you in answer to your first
question.
> I have managed to compile the using skin=Posix
> and also able run it with rteth0 interface. The code file is
> attached with this email along with makefile. I have configured
> the rteth0 interface with ip and netmask with promisc mode.
> Morover over i have added the route by using the command
> "./rtroute solicit 10.0.0.11 dev rteth0". I am able to ping also
> as below:
You do not need to add a network route to send raw packets. The
network routes are only used to route IP packets.
>
> root@cpc:/usr/xenomai/sbin# ./rtping 10.0.0.11
> Real-time PING 10.0.0.11 56(84) bytes of data.
> ioctl: No route to host
> root@cpc:/usr/xenomai/sbin# ./rtroute solicit 10.0.0.11 dev rteth0
> root@cpc:/usr/xenomai/sbin# ./rtping 10.0.0.11
> Real-time PING 10.0.0.11 56(84) bytes of data.
> 64 bytes from 10.0.0.11: icmp_seq=1 time=408.6 us
> 64 bytes from 10.0.0.11: icmp_seq=2 time=370.1 us
> 64 bytes from 10.0.0.11: icmp_seq=3 time=387.2 us
> 64 bytes from 10.0.0.11: icmp_seq=4 time=413.4 us
> ^C
> --- 10.0.0.11 rtping statistics ---
> 4 packets transmitted, 4 received, 0% packet loss
> worst case rtt = 413.4 us"
So, RTnet works on your machine. This is a really high value for an
rtnet ping, though.
>
> The other PC is non real time windows PC.
That explains the long ping time.
> And I can see the ping traffic on the windows PC. But when i run
> the code after compilation I can not see the traffic on the
> network. The output of the code is
So, obviously, the code is wrong. I am not really interested in
knowing why, so I have implemented a program doing what you are
trying to do, and tested that it works with RTnet. Why your example
fail is left as an exercise for you if you are interested.
#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <string.h>
#include <pthread.h>
#include <sched.h>
#include <net/if.h>
#include <sys/ioctl.h>
#include <sys/socket.h>
#include <netpacket/packet.h>
#include <net/ethernet.h>
#include <arpa/inet.h>
#define check_unix(expr) \
({ \
int _rc = (expr); \
if (_rc < 0) { \
fprintf(stderr, "%s:%d: %s failed: %d (%m)\n", \
__FILE__, __LINE__, #expr, errno); \
exit(EXIT_FAILURE); \
} \
_rc; \
})
#define check_pthread(expr) \
({ \
int _rc = (expr); \
if (_rc != 0) { \
fprintf(stderr, "%s:%d: %s failed: %d (%s)\n", \
__FILE__, __LINE__, #expr, _rc, strerror(_rc));
\
exit(EXIT_FAILURE); \
} \
_rc; \
})
int main(int argc, const char *argv[])
{
struct sched_param sparm;
struct sockaddr_ll peer;
struct ethhdr *header;
struct ifreq ifr;
char buffer[2048];
int sock, err;
if (argc != 3)
exit(EXIT_FAILURE);
sparm.sched_priority = 1;
check_pthread(pthread_setschedparam(pthread_self(), SCHED_FIFO,
&sparm));
sock = check_unix(socket(PF_PACKET, SOCK_RAW, 0));
check_unix(snprintf(ifr.ifr_name, sizeof(ifr.ifr_name), "%s", argv[1]));
check_unix(ioctl(sock, SIOCGIFHWADDR, &ifr));
header = (struct ethhdr *)buffer;
memset(&header->h_dest, 0xff, sizeof(header->h_dest));
memcpy(&header->h_source, ifr.ifr_hwaddr.sa_data,
sizeof(header->h_source));
header->h_proto = htons(ETH_P_802_EX1);
check_unix(snprintf(ifr.ifr_name, sizeof(ifr.ifr_name), "%s", argv[1]));
check_unix(ioctl(sock, SIOCGIFINDEX, &ifr));
memset(&peer, '\0', sizeof(peer));
peer.sll_family = AF_PACKET;
peer.sll_halen = 6;
peer.sll_ifindex = ifr.ifr_ifindex;
err = check_unix(
snprintf(buffer + sizeof(*header),
sizeof(buffer) - sizeof(*header), "%s", argv[2]));
for (;;) {
struct timespec ts;
check_unix(
sendto(sock, buffer, err + sizeof(*header), 0,
(struct sockaddr *)&peer, sizeof(peer)));
ts.tv_sec = 1;
ts.tv_nsec = 0;
check_unix(nanosleep(&ts, NULL));
}
}
--
Gilles.
https://click-hack.org
_______________________________________________
Xenomai mailing list
[email protected]
http://xenomai.org/mailman/listinfo/xenomai
