Re: [lwip-users] Remote tcp connection with TAPIF interface

[K.J. Mansley]

On Fri, 2004-08-27 at 16:09, address@hidden
wrote:
> > The problem is that when the client (A) tries to connect to (B)
> > 192.168.0.2 address, port 19 (server application) it returns an error with
> > value: -3.
> > It seems that in the "etharp_output(...)" function in the client side, all
> > the items in the "arp_table" are set to ETHARP_STATE_EMPTY state. 

> > Does somebody has encounter this type of problem ?

This would appear to have the same symptoms as the problems reported by
"xcb" earlier in the week.

> > Is our system environment configuration correct ?
> > Do we need to add the server interface in the client ("add_netif of
> > "192.168.0.2" in the "test_tcp_client") ?
> > How the ARP query has been made in the code in order for the "arp_table"
> > to be filled ?

I think it should automatically queue packets while it sends out an ARP
query to find the mac address of the server.  Could you get a packet
trace using tcpdump or ethereal in order to check if this is actually
happening, and if the request is being responded to with an ARP reply?

There have been a couple of minor errors reported in the ARP code
recently, so this could be another.  I'm not familiar enough with that
bit of lwIP to comment though.  Hopefully someone with more knowledge
will be able to!

Thanks

Kieran