Re: [lwip-users] SOCK_RAW?

[Ed Sutter]

Thanks for responding.
Just a little bit more information...
To make sure that I didn't just have a bug in the code, I
tried changing the socket() call from...
   socket(AF_INET,SOCK_RAW,IP_PROTO_ICMP);
to...
   socket(AF_INET,SOCK_DGRAM,0);

Then I just reran the code with etherpeek running and a packet
was transmitted.  It was junk obviously but it was transmitted.
In the SOCK_RAW case, the lwip_sendto() call always hangs and
I traced it to the point shown in the below stack trace dump.
 
trace PingThread
   0x000317a0: OS_Sched()
   0x00036358: OSQPend()
   0x0006a4c0: sys_arch_mbox_fetch()
   0x0004db08: sys_mbox_fetch()
   0x00062914: netconn_connect()
   0x00066e18: lwip_sendto()
   0x00071e58: ping()
   0x00071f40: taskPing()

Note that this is a uCOS-ii port that I've been using for quite
some time now.  Just recently upgraded to 0.7.0 (and pulled in the
ARP patch), and everything else I have (that doesn't use SOCK_RAW)
appears to be working well.
Ed

"Zschocke, Florian" wrote:
> 
> Ed Sutter wrote on Wednesday, February 04, 2004 7:24 PM:
> 
> > Does anyone know the status of the ability to use SOCK_RAW
> > in lwip?
> 
> The general status is "possible" in the sense "can be done", but I don't
> know if the current raw sockets implementation in the CVS has a bug. We
> haven't tested it since we use a different raw sockets implementation for
> lwIP. But AFAIK the one in the CVS had been reported to work before.
> 
> Florian
> 
> _______________________________________________
> lwip-users mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/lwip-users