Hi, I increased MEMP_N_PBUF to 32, and it is running well. (TCP_SND_QUEUELEN seems to be fixed.)
Thanx a lot! Regards, Daniel Gesendet: Montag, 2. August 2010 13:32 Von: [email protected] > Hello, > > Just try to increase your TCP_SND_QUEUELEN and MEMP_N_PBUF to 32. > It's a bit strange with the debug messages, because you are closing the > connection with this package: (There is a FIN flag.) > > +-------------------------------+ > | 4006 | 9999 | (src port, dest port) > +-------------------------------+ > | 00f589592a | (seq no) > +-------------------------------+ > | 00000019cb | (ack no) > +-------------------------------+ > | 5 | |010001| 65455 | (hdrlen, flags (FIN ACK ), win) > +-------------------------------+ > | 0x53e9 | 0 | (chksum, urgp) > +-------------------------------+ > > > Maybe you check your application why it is closing the connection. > > > Best regards, > Matthias > > > On Mon, 2 Aug 2010 12:36:44 +0200, Daniel Klingler wrote: >> Hi, >> >> I don't know how I can update in Xilinx to LWIP 1.32. >> >> - MEM_SIZE 524288 >> - TCP_SND_QUEUELEN 16 * TCP_SND_BUF/TCP_MSS >> - MEMP_N_PBUF 16 >> >> These are the debug information: >> The packet with the length of 1518 Bytes doesn't go out. With wireshark I >> don't see it. After a timeout the connection will be closed. >> >> Regards, >> Daniel >> >> --- >> State: ESTABLISHED >> tcp_receive: ACK for 19bd, unacked->seqno 19b7:19bd >> tcp_receive: removing 19b7:19bd from pcb->unacked >> pbuf_free(0x8802ffc0) >> pbuf_free: deallocating 0x8802ffc0 >> tcp_write(pcb=0x880b1800, data=0x8d2785d0, len=14, apiflags=1) >> tcp_enqueue(pcb=0x880b1800, arg=0x8d2785d0, len=14, flags=0, apiflags=1) >> pbuf_alloc(length=14) >> pbuf_alloc(length=14) == 0x8802ffc0 >> pbuf_header: old 0x88030040 new 0x8803002c (20) >> tcp_enqueue: queueing 19bd:19cb (0x0) >> tcp_output_segment: 19bd:19cb >> pbuf_header: old 0x8803002c new 0x88030018 (20) >> pbuf_header: old 0x88030018 new 0x8803000a (14) >> tcp_recved: recveived 6 bytes, wnd 8192 (0). >> pbuf_free(0x88276080) >> pbuf_free: deallocating 0x88276080 >> tcp_output: nothing to send (0x0) >> State: ESTABLISHED >> tcp_fasttmr: delayed ACK >> pbuf_alloc(length=20) >> pbuf_alloc(length=20) == 0x880300c0 >> tcp_output: sending ACK for f589592a >> pbuf_header: old 0x88030140 new 0x8803012c (20) >> pbuf_header: old 0x8803012c new 0x8803011e (14) >> pbuf_free(0x880300c0) >> pbuf_free: deallocating 0x880300c0 >> pbuf_alloc(length=1518) >> pbuf_alloc: allocated pbuf 0x88276080 >> pbuf_alloc(length=1518) == 0x88276080 >> pbuf_free(0x88276080) >> pbuf_free: deallocating 0x88276080 >> tcp_slowtmr: processing active pcb >> tcp_slowtmr: polling application >> tcp_output: nothing to send (0x0) >> pbuf_alloc(length=1518) >> pbuf_alloc: allocated pbuf 0x88276080 >> pbuf_alloc(length=1518) == 0x88276080 >> pbuf_header: old 0x882760c0 new 0x882760ce (-14) >> pbuf_header: old 0x882760ce new 0x882760e2 (-20) >> pbuf_free(0x88276080) >> pbuf_free: deallocating 0x88276080 >> tcp_slowtmr: processing active pcb >> tcp_slowtmr: polling application >> tcp_output: nothing to send (0x0) >> tcp_slowtmr: processing active pcb >> tcp_slowtmr: polling application >> tcp_output: nothing to send (0x0) >> pbuf_alloc(length=1518) >> pbuf_alloc: allocated pbuf 0x88276080 >> pbuf_alloc(length=1518) == 0x88276080 >> pbuf_header: old 0x882760c0 new 0x882760ce (-14) >> TCP header: >> +-------------------------------+ >> | 4006 | 9999 | (src port, dest port) >> +-------------------------------+ >> | 00f589592a | (seq no) >> +-------------------------------+ >> | 00000019cb | (ack no) >> +-------------------------------+ >> | 5 | |010000| 65455 | (hdrlen, flags (ACK ), win) >> +-------------------------------+ >> | 0x53ea | 0 | (chksum, urgp) >> +-------------------------------+ >> pbuf_header: old 0x882760ce new 0x882760e2 (-20) >> pbuf_header: old 0x882760e2 new 0x882760f6 (-20) >> +-+-+-+-+-+-+-+-+-+-+-+-+-+- tcp_input: flags ACK >> -+-+-+-+-+-+-+-+-+-+-+-+-+-+ >> State: ESTABLISHED >> tcp_receive: ACK for 19cb, unacked->seqno 19bd:19cb >> tcp_receive: removing 19bd:19cb from pcb->unacked >> pbuf_free(0x8802ffc0) >> pbuf_free: deallocating 0x8802ffc0 >> tcp_output: nothing to send (0x0) >> pbuf_free(0x88276080) >> pbuf_free: deallocating 0x88276080 >> State: ESTABLISHED >> pbuf_alloc(length=1518) >> pbuf_alloc: allocated pbuf 0x88276080 >> pbuf_alloc(length=1518) == 0x88276080 >> pbuf_header: old 0x882760c0 new 0x882760ce (-14) >> TCP header: >> +-------------------------------+ >> | 4006 | 9999 | (src port, dest port) >> +-------------------------------+ >> | 00f589592a | (seq no) >> +-------------------------------+ >> | 00000019cb | (ack no) >> +-------------------------------+ >> | 5 | |010001| 65455 | (hdrlen, flags (FIN ACK ), win) >> +-------------------------------+ >> | 0x53e9 | 0 | (chksum, urgp) >> +-------------------------------+ >> pbuf_header: old 0x882760ce new 0x882760e2 (-20) >> pbuf_header: old 0x882760e2 new 0x882760f6 (-20) >> +-+-+-+-+-+-+-+-+-+-+-+-+-+- tcp_input: flags FIN ACK >> -+-+-+-+-+-+-+-+-+-+-+-+-+-+ >> State: ESTABLISHED >> tcp_receive: received FIN. >> tcp_close: closing in State: CLOSE_WAIT >> --- >> >> >> >> >> Gesendet: Montag, 2. August 2010 10:26 Von: [email protected] >>> Hi, >>> >>> first of all please update to LWIP 1.32. I can only recommend that. >>> Second what's your MEM_SIZE, TCP_SND_QUEUELEN, MEMP_NUM_PBUF? >>> Do you have access to the Debug messages, that could really help. >>> I think it could be maybe a memory problem. >>> >>> >>> >>> On Mon, 2 Aug 2010 10:19:06 +0200, Daniel Klingler wrote: >>>> Hi, >>>> >>>> I am using LWIP 1.3 with Xilinx (SDK). >>>> >>>> My settings: >>>> - TCP_SND_BUF 64240 >>>> - TCP_MSS 1460 >>>> >>>> My problem: >>>> - If I am do tcp_write with 18720 Bytes the transfer is well going. >>>> - If I am writing with 19968 Bytes the transfer with some PCs is well >>>> going, but with other PCs the return of tcp_write is -1. >>>> - If I am writing with 21504 Bytes or more then with all connected PCs I >>>> will return -1 of tcp_write. >>>> >>>> Does anybody know this issue or has any idea to solve this? >>>> >>>> Thanks for your replies. >>>> Regards, >>>> Daniel >>>> >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> lwip-users mailing list >>>> [email protected] >>>> http://lists.nongnu.org/mailman/listinfo/lwip-users >>> _______________________________________________ >>> lwip-users mailing list >>> [email protected] >>> http://lists.nongnu.org/mailman/listinfo/lwip-users >> >> _______________________________________________ >> lwip-users mailing list >> [email protected] >> http://lists.nongnu.org/mailman/listinfo/lwip-users > > _______________________________________________ > lwip-users mailing list > [email protected] > http://lists.nongnu.org/mailman/listinfo/lwip-users _______________________________________________ lwip-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/lwip-users
