[lwip-users] Re: lwip_send bug with RAM data???

[Piero 74]

SORRY.... i found a bug in my driver....

packet had an odd len... when i copy pbuff data into eth buffer, i have to copy a even number of bytes (hw requests)... so, i do a check before send, and this check FAIL!!!!

i hope this is the last bug in driver.... 
i did a lot of work around lwip code for debugging.... now i know a lot of things about lwip code.... but i have yet a lot of work to do for my application!!! :O(

anyway, thanks to all lwip comunity... this mailing list are realy very usefull, and lwip seems a very professsional middleware!!!

bye.... see you... on the next problem! :O)

Piero

2008/3/7, Piero 74 <address@hidden>:

Hi all.

i'm seeing a strange problem.
I have lwip 130rc1 and i'm using socket.

recv function seems works well, but i have a strange behaviour of lwip_send function.

if i write:

send(newsockfd, &buffROM, len, 0);

where buffROM is a vector in ROM, data are sent.

if i write

send(newsockfd, &buffRAM, len, 0);

where buffRAM is a buffer in ram modified in run time (it's build in my application using data in ram) lwip seems doesn't send packet with these data. (i 'm using a protocol analyzer to capture packet in network... network is simple, only a pc, my board  connected through an hub)

Any idea???

please HELP me.... :O(
thanks,
Piero