[lwip-users] netconn_write() causes stack to lock-up

[Josh Rothstein]

I am running lwIP 1.2.0 on an Atmel AVR32 UC3A.

 

I am trying to simply dump some data using the Netconn API to HyperTerm running on my PC.  After a few transmissions the entire stack locks up.  I had a similar problem a while back on a different project while attempting to do this with the Socket API.  It turns out that I had to call lwip_recv() after each lwip_send() or else the responses from HyperTerm would overload the stack.  I have done the following below, but I still get a lockup after a few transmits. Any ideas?

 

Here is a reduced version of my task that causes the lockup

 

portTASK_FUNCTION( vDspLowBandTask, pvParameters )

{

 

dsp32_t downsample[DOWNSAMPLE_SIZE];

 

      struct netbuf *pxRxBuffer;

      struct netconn *pxNewConnection;

      struct ip_addr pxRemoteAddr;

 

      pxNewConnection = netconn_new(NETCONN_TCP);

      pxRemoteAddr.addr = inet_addr("192.168.0.10");

      netconn_connect(pxNewConnection, &pxRemoteAddr, 23);

 

      for ( ;; )

      {

            netconn_write(pxNewConnection, downsample, (sizeof(dsp32_t) * DOWNSAMPLE_SIZE), NETCONN_NOCOPY);

            pxRxBuffer = netconn_recv(pxNewConnection);

            netbuf_delete(pxRxBuffer);

            }

}

           

Josh Rothstein

Embedded Systems Engineer

DCX-CHOL Enterprises, Inc.

12831 South Figueroa Street

Los Angeles, CA 90061-1157

310.516.1692x134