Hi, I am trying to use lwip 7.1
with ucos-ii on a mc9s12 microcontroller.
When I call netif_add(&netif,
&ipaddr, &netmask, &gw, NULL, cs8900if_init, ip_input) that
machine crashes shortly after, however control is returned to the task after
the function is called, so I know its completing the netif_add function
call.
By increasing the tasks stack
size, I can prolong the crash by a few seconds, however I think I am a victim
of stack corruption, not sure.
When I call netif_add() there are
some things I am unsure of.
What do I do for ipaddr netmask
etc if I plan to use DHCP?
Do I still
set:
IP4_ADDR(&gw, 192,168,0,1);
IP4_ADDR(&ipaddr, 192,168,0,102);
IP4_ADDR(&netmask, 255,255,255,0);
For the function call, even if
they will be ignored?
As for debugging, if I comment out
the netif_add() the machine runs fine, however, if I call it, it runs and
returns, but crashes the machine at some point later in time (a few seconds
later depending on the stack size for the task). IF I comment out the body of
the netif_add() it does the same thing.
This is why I think the function
call itself is causing the stack to become corrupted! (from the parameters
being passed).
Here is what I have done to
initialize everything so far:
sys_init();
mem_init();
memp_init();
pbuf_init();
netif_init();
ip_init();
udp_init();
tcp_init();
I have also setup separate tasks
for various timers:
dhcp_fine_tmr(); every 0.5
seconds
dhcp_coarse_tmr(); every
second
etharp_tmr(); every 10
seconds
tcp_tmr(); every
250ms
I have registered the following
callbacks:
tcp_accept(listen_pcb,
accept_connection);
tcp_recv(listen_pcb, recv);
tcp_sent(connect_pcb,
sent);
thanks for the
help,
--eric
shufro