Today I finally updated my lwip to 1.4.1 and now it works. Release 1.4.1 contained many changes in tcp code and it seems that one of them fixed my problem.
On Fri, Mar 1, 2013 at 5:53 PM, Maciej Gajdzica <maciej.gajdz...@762687.com>wrote: > I made some futher research and observed that in local network that > fragmented page worked only on one computer. On every other one the page > was not loaded. It was caused by "Content-length" field in http > header.After I deleted that line, page was loading on every PC on LAN, but > still not on 3G. > > My next finding was that whole http packed with configuration.html page > has about 2950 bytes. When I made that page smaller (about 2200 bytes) - it > was loading without any problem on 3G. That means it is not exactly a > problem with fragmentation. > > My 3G router which forwards port 80 to STM has MTU=1480 in 3G settings, > maximum allowed is 1492. Lwip has default MTU=1500. I tried to set lower > MTU and TCP_MSS on my application but it still doesnt work. > > > On Mon, Feb 25, 2013 at 3:23 PM, Maciej Gajdzica < > maciej.gajdz...@762687.com> wrote: > >> Today I finally captured some debug output. Active debug options: >> #define SOCKETS_DEBUG LWIP_DBG_ON >> #define IP_REASS_DEBUG LWIP_DBG_ON >> #define MEM_DEBUG LWIP_DBG_ON >> #define MEMP_DEBUG LWIP_DBG_ON >> #define TCP_DEBUG LWIP_DBG_ON >> #define TCP_INPUT_DEBUG LWIP_DBG_ON >> #define TCP_FR_DEBUG LWIP_DBG_ON >> #define TCP_RTO_DEBUG LWIP_DBG_ON >> #define TCP_CWND_DEBUG LWIP_DBG_ON >> #define TCP_WND_DEBUG LWIP_DBG_ON >> #define TCP_OUTPUT_DEBUG LWIP_DBG_ON >> #define TCP_RST_DEBUG LWIP_DBG_ON >> #define TCP_QLEN_DEBUG LWIP_DBG_ON >> #define TCPIP_DEBUG LWIP_DBG_ON >> >> I dont really know how to interpret these data. Maybe someone more >> familiar with tcp debug could help me? >> >> Maybe that issue is fixed in 1.4.1, as there was some changes in tcp code >> according to 1.4.1 changelog. I will check it out. >> >> >> On Fri, Feb 22, 2013 at 8:42 PM, Simon Goldschmidt <goldsi...@gmx.de>wrote: >> >>> "ha...@gawnet.ch" wrote: >>> > Perhaps it could be, that there is still a Problem with Bug #24212. >>> > It has been said, that the Bug has been fixed. I dont really trust in >>> this. >>> >>> A statement like this belongs into the bug tracker or at least in a >>> suspected bug report on lwip-devel, not into a side-note on lwip-users. >>> >>> Unfortunately, I don't have much time for lwIP, lately. If that bug >>> would be reopened, it would make sure things don't get forgotten. >>> >>> Simon >>> >>> _______________________________________________ >>> lwip-users mailing list >>> lwip-users@nongnu.org >>> https://lists.nongnu.org/mailman/listinfo/lwip-users >>> >> >> >> >> -- >> 762687 INNOVATION CODE <http://www.762687.com> >> Tel: +48 783 367 232 >> Kom: +48 519 136 410 >> E-mail: maciej.gajdz...@762687.com >> > > > > -- > 762687 INNOVATION CODE <http://www.762687.com> > Tel: +48 783 367 232 > Kom: +48 519 136 410 > E-mail: maciej.gajdz...@762687.com > -- 762687 INNOVATION CODE <http://www.762687.com> Tel: +48 783 367 232 Kom: +48 519 136 410 E-mail: maciej.gajdz...@762687.com
_______________________________________________ lwip-users mailing list lwip-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/lwip-users
