Hi Simon,
... in that case it must be some other bug!
Every time it crashed, I got ERR_USE. But not only did it crash this
thread, but killed all other LWIP functions (killing the HTTP server
thread too - making de-bug difficult). The RTOS seems to keep running,
as the GPIO toggle markers still appear every 4mS (I had markers in
other threads also, and they kept running too).
I'm beginning to wonder if there's some kind of "memory leak"
somewhere. I can't think where, as apart from a few "fixed" memory
buffers - which aren't being used at the moment - the only memory usage
is stacks allocated by the RTOS and buffers allocated by LWIP. I have a
status page on the HTTP server which shows the RTOS free stack for each
thread - none anywhere near a problem level. Is it possible to extract
similar information from LWIP? I see in lwipopts a switch for
statistics, but don't know how to access the results - I guess watching
that would indicate any growing problem.
I tried one of Noam's linked examples (included below) - but that showed
exactly the same problem, stopped sending after about 20 minutes (I
couldn't tell if it was ERR_USE again, but no reason to think it was any
different!).
void udp_test2(void *arg)
{
struct netconn *conn;
char msg[]="testing" ;
struct netbuf *buf;
char * data;
conn = netconn_new( NETCONN_UDP );
netconn_bind(conn, IP_ADDR_ANY, 1234 ); //local port
netconn_connect(conn, IP_ADDR_BROADCAST, 1235 );
for( ;; )
{
buf = netbuf_new();
data =netbuf_alloc(buf, sizeof(msg));
memcpy (data, msg, sizeof (msg));
netconn_send(conn, buf);
netbuf_delete(buf); // De-allocate packet buffer
osDelay(4); //some delay!
}
}
I altered the function name and the delay function, but otherwise it's a
direct copy.
Many thanks,
Mike.
On 28/07/2016 10:24, Simon Goldschmidt wrote:
Mike Fleetwood wrote:
Question 1: Why should send return with ERR_USE?
It shouldn't. ERR_USE is only returned by tcp_bind() and udp_bind() (and you
already bound your netconn outside the loop).
The only way it could be returned is by your netif driver's output function -
or due to some other bug...
Simon
_______________________________________________
lwip-users mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/lwip-users
--
FACE Systems Ltd
The Old Boat House
Cadgwith
Cornwall TR12 7JX
T:01326 291031
M:07831 401464
_______________________________________________
lwip-users mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/lwip-users
