Re: [lwip-users] Possible bug in src/api/tcpip.c

[Rishi Khan]

On May 2, 2008, at 5:14 AM, address@hidden wrote:

> Kieran Mansley wrote:
> > On Thu, 2008-05-01 at 14:16 -0400, Rishi Khan wrote:
> >
> > > P.S. I've written a unix driver that actually binds to eth0 (or  
> > > other  device) and operates there. This allows you to test inter- 
> > > computer  communication in unix (tap/tun seem to only work on the  
> > > machine  running the program and no other machine can access  
> > > this). Is this  something that is useful to the community?
> >
> > Yes, it would be useful, particularly if you can maintain it as lwIP
> > changes.
> I'm curious (since I didn't really use lwIP under linux so far):  
> how does your driver work? Is this a "real" driver running in the  
> kernel context or did you use libpcap or something like that?
>
> Simon

Here's the general idea (some code skipped for simplicity):
init()
{
  sockfd = socket(PF_PACKET,
                  SOCK_RAW,
                  htons(ETH_P_ALL));
...

  memset(&sll, 0, sizeof(sll));
  sll.sll_family   = PF_PACKET;
  sll.sll_ifindex  = ifindex;
  sll.sll_protocol = htons(ETH_P_ALL);

  bind(sockfd, (struct sockaddr *)&sll, sizeof(sll));

  ethernetif->sockfd = sockfd;
  /* set MAC hardware address */
  memcpy(ethernetif->ethaddr->addr, hwaddr, ETH_ALEN);
}

lowLeveloutput()
{
send(ethernetif->sockfd, buf, p->tot_len,0)
}

lowLevelinput()
{
len = recv(ethernetif->sockfd, buf, netif->mtu + ETH_HLEN,  
MSG_DONTWAIT);
<copy to pbuf>
}

ethif_input(),  ethif_thread(), ethif_init()  ... same as ethernet  
skeleton