Re: [microblaze-uclinux] tcp/ip

[Johan Bernspång <johan.bernspang@xxxxxx>]

Hi,

I'm not sure about how to solve that, I didn't use the webserver example 
very much. I just made sure it worked and carried on with other 
applications...

/Johan

Matthias Heise wrote:
> Hi Johan,
>
> thanks for you answer, yeah, thats what I found out yesterday too :- ) I 
> now delete the "buf" buffer in the "smsc_rcv_poll" routine if the 
> RCV_INT flag isn't active. Now all of the four pings go through with 1-2 
> ms and no extra sent frames. Now there is still the http thing : a 
> "crippled" frame is sent there too after the correct http frame 
> tranmission. I'm still searching where that comes from but obviously the 
> GET command is taken there twice and I don't know yet which buffer or 
> flag to delete at which point.
> Matthias
>
> Johan Bernspång schrieb:
>
> > Hi Matthias,
> >
> > Do you use interrupts or do you poll the chip? If you use interrupts 
> > your system might respond to the same packet generating interrupt over 
> > and over again if you haven't acked the interrupt correctly or the 
> > receive buffer isn't reset properly. If you're polling you might poll 
> > the same ping req over and over again if you're not clearing the 
> > packet after read correctly.
> >
> > At least that is my guess.
> >
> > /Johan
> >
> > Matthias Heise wrote:
> >
> > > Hi again,
> > > as the situation doesn't improve I again try to get help from you 
> > > regarding my network issue. I'm still at the web_server example, I've 
> > > got a little functionality (one http page, and an average of two 
> > > successful pings of four) . So I went back to the ping and checked 
> > > the communication with ethereal. See the attached ethereal "ping" 
> > > file : after the first request there are 40 replies sent and after 
> > > that one request and one reply frame then again request and 40 
> > > replies and the fourth sequence has again one request and one reply. 
> > > The strange thing is that only sequence 1 and 3 (with the 40 replies) 
> > > seem to be successful (with 14 and 15ms) the short ones result in a 
> > > timeout.
> > > Please, even if you don't know yourself, could you try to give me a 
> > > contact to someone who maybe has a clue ?
> > > Thanks very much,

--
-----------------------------------------------
Johan Bernspång, johan.bernspang@xxxxxx
Research engineer

Phone: +46(0)13 378386

Swedish Defense Research Agency - FOI
Division of Command & Control Systems
Department of Electronic Warfare Systems

www.foi.se
-----------------------------------------------
___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/