RE: Re: Re: comedi + two ni6070E

From: David <damachon_at_eresmas.net> · Date: Thu, 16 Oct 2003 15:06:59 GMT

Hello again

I have managed to compile the CVS version after installing some new
programs versions (aclocal-1.7, automake-1.7, autoconf2.54,
libtool1.4.2, bison1.8.75). Comedi looks to work fine. 

Besides, now we are using the mmap model from kernel space using
comedi_map() like you told us.

But our problems with adquisition continues appearing (exactly the
same). We are using comedi CVS (0.7.67.1) and comedilib CVS, two NI
6070E cards, RTAI 24.1.11 with kernel 2.4.20-rthal5. The adquisition
starts correctly and the real time task bring data from the comedi
buffer that belongs each card, but when we use high adquisition
frequencies (above 700KHz more or less) one of the buffers stop working
 and the result of "comedi_get_buffer_contents ()" is always 0. the
other card continue working correctly.

Any idea about what is happening?

thank you very much again.
-- Daniel Machón [damachon_at_eresmas.net] --

----- Mensaje Original -----
Remitente: David Schleef ds_at_schleef.org
Destinatario: danmachon_at_eresmas.net
Fecha: Lunes, Octubre 13, 2003 7:34pm
Asunto: Re: Re: comedi + two ni6070E

>On Mon, Oct 13, 2003 at 07:18:46AM +0000, David wrote:
>> I have never used CVS and I dont'n how to compile it because it has 
>not> scrpits to copile. ¿Any help please?
>
>Run installed.
>
>> Does exist any function in kcomedilib to have the same results as
>> cfc_read_array_from_buffer? If I shouldn't use that in my code I 
>would> like to change it.
>
>Using Comedi from kernel space requires you to use the mmap model
>of data transfer.  The code can be almost identical to user space,
>in the demo/mmap.c example.
>
>
>
>dave...
>
>
>_______________________________________________
>comedi mailing list
>comedi_at_comedi.org
>https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi
>