|
| From: | Josh Rothstein |
| Subject: | RE: [lwip-users] Unable to allocate TCP_SEG |
| Date: | Wed, 16 Jan 2008 17:20:48 -0800 |
|
One more behavioral thing I noticed: If I do lwip_send() and watch the debug
output I will see the TCP header information showing the stack received an ack for
that packet. If I do a couple more lwip_send(), I will eventually
stop seeing the TCP header information and the tcp_enqueue() will begin to fill
with TCP_SEGs. I if then close() the socket all the pending receives will
then come through the debug output. How can I stop this bottleneck? Any suggestions? From: address@hidden
[mailto:address@hidden On Behalf Of Josh Rothstein I have done some more debugging and it
looks like tcp_enqueue cannot allocate memory for another TCP_SEG because acks
are not getting received by tcp_receive(), who would in turn would free the
pool for MEMP_TCP_SEG. Further examination shows that the problem
lies within tcpip.c. For some reason tcpip_input() fails when trying to
memp_malloc(MEMP_TCPIP_MSG). This occurs before tcp_enqueue() fills the MEMP_TCP_SEG
pool. tcpip_thread() does call memp_free(MEMP_TCPIP_MSG, msg) at the end of the
loop, so I am at a loss as to why the MEMP_TCPIP_MSG pool fills up. Any suggestions? From:
address@hidden [mailto:address@hidden
On Behalf Of Josh Rothstein In order to learn how to properly use LwIP in conjunction
with FreeRTOS, I am trying to write a simple Telnet client on an Atmel AVR32
Demo Board. Disregarding the reset button, there are three tactile
buttons on the board. I have enabled the first button to open a connection to
the development PC, the second button to transmit a character string, and the
third button to close the connection. I have included the code for my Telnet
task at the bottom of the email; formatting is a little funny due to pasting
into MS Outlook. For some reason after, a certain amount of transmissions
(about 9 or 10) my tcp/ip stack is unable to allocate memory for the next
transmission and the entire stack locks up. The error occurs in the LwIP
Raw API implementation in lwip/core/tcp_out.c
within the function tcp_enqueue(). I have pasted the area below where the
stack is unable to allocated memory. err_t tcp_enqueue(struct tcp_pcb *pcb, void *arg, u16_t len, u8_t
flags, u8_t copy, u8_t
*optdata, u8_t optlen) { …… …… …… ……
/* Allocate
memory for tcp_seg, and fill in fields. */
seg = memp_malloc(MEMP_TCP_SEG);
if (seg == NULL) {
LWIP_DEBUGF(TCP_OUTPUT_DEBUG | 2, ("tcp_enqueue: could not allocate memory for
tcp_seg\n"));
goto memerr;
} …… …… …… ……
return ERR_OK; } memp_malloc() allocates memory from a predefined memory
pool. It seems that the TCP/IP stack is not properly freeing memory from the
pool after successful transmissions. I am able to tweak the lwipopts.h
file to allow for more successful transmissions but eventually the stack always
uses up the pool. Oddly, I have noticed that if I quickly and repeatedly press
my transmit button, I can send endless amounts of data without any stack
lockup. If I transmit with approximately 1 or more seconds between transmissions
the memory pool runs dry after 9 or 10 presses. I have attached a log of the debug output from lwip as well
as a Wireshark log from the Host. The Host (windows PC) is 192.168.0.10
and the lwip stack sits on 192.168.0.2. portTASK_FUNCTION(
vBasicTELNETClient, pvParameters ) {
t_ButtonStatus xData; struct sockaddr_in xSockAddr;
portSHORT sErrorEverOccurred = pdFALSE;
portLONG l_socketFD = 0;
Bool bConnectionOpen = FALSE; pxButtonQueueParameters
= ( xBlockingQueueParameters * ) pvParameters;
b_pushb1_init(); /* configures GPIO and interrupts */
b_pushb2_init(); /* configures GPIO and interrupts */
b_pushb3_init(); /* configures GPIO and interrupts */ memset(&xSockAddr,
0, sizeof(xSockAddr));
xSockAddr.sin_family = AF_INET;
xSockAddr.sin_len = sizeof(xSockAddr);
xSockAddr.sin_port = htons(23);
xSockAddr.sin_addr.s_addr = inet_addr("192.168.0.10");
for( ;; )
{ if( xQueueReceive( pxButtonQueueParameters->xQueue, &xData,
pxButtonQueueParameters->xBlockTime ) == pdPASS )
{
/* We have
successfully received a message, so increment the
variable used to check we are still running and process queue. */
if(
sErrorEverOccurred == pdFALSE )
{
if (
xData.xStates.uc_pushb1 == PUSHB_EVENT_PRESS && !bConnectionOpen )
{
l_socketFD = lwip_socket(PF_INET, SOCK_STREAM,
0);
if
(lwip_connect(l_socketFD, (struct sockaddr*) &xSockAddr, sizeof(xSockAddr)) >= 0)
bConnectionOpen = TRUE;
}
if (
xData.xStates.uc_pushb2 == PUSHB_EVENT_PRESS && bConnectionOpen)
{
if (
lwip_send(l_socketFD, (void *) MSG_HELLO, 20, 0) < 0 )
lwip_close(l_socketFD);
}
if (
xData.xStates.uc_pushb3 == PUSHB_EVENT_PRESS && bConnectionOpen)
{
lwip_shutdown(l_socketFD, 2);
lwip_close(l_socketFD);
bConnectionOpen = FALSE;
}
}
}
} } Any suggestions
would be greatly appreciated. Thanks, Josh |
| [Prev in Thread] | Current Thread | [Next in Thread] |