Re: comedi_config of NI DAQCard fails

> Hi all,
>
> I have tried to get a NI DAQCard 6024E to work on my Debian Sarge system.
> Linux kernels I tried were either 2.4.29 or 2.6.10 - both patched with
> adeos.

I have now got the card to work on a different notebook. It "just worked"
with comedilib configured like this:
./configure --sysconfdir=/etc --localstatedir=/var

This seemed to be a general PCMCIA problem instead of a special one with
comedi or the card.
To sum it up I could _not_ get the DAQCard to work on two notebooks with a
Texas Instruments PCMCIA bridge. One with a Ricoh bridge worked.

At the non-working notebooks cardmgr complained when probing the DAQCard's
memory regions. I should have mentioned _this_ in my lengthy email - but
overlooked it.

Suddenly some of my observations of this make sense:

> This is the output of the cardmgr:
> Jul 22 20:03:22 db4774 cardmgr[1074]: watching 2 sockets
> Jul 22 20:03:22 db4774 cardmgr[1075]: starting, version is 3.2.5
> Jul 22 20:03:22 db4774 cardmgr[1075]: initializing socket 0
> Jul 22 20:03:22 db4774 cardmgr[1075]: socket 0: National Instruments
> DAQCard-6024E
> Jul 22 20:03:22 db4774 cardmgr[1075]:   product info: "National
> Instruments", "DAQCard-6024E"
> Jul 22 20:03:22 db4774 cardmgr[1075]:   manfid: 0x010b, 0x075e
-------------------------------------------------------------------------
> Jul 22 20:03:22 db4774 cardmgr[1075]: executing: 'modprobe ni_mio_cs'
-------------------------------------------------------------------------
>
> Curiously cardmgr does not seem to execute '/etc/pcmcia/comedi start
> ni_mio_cs'.
Maybe because it's called by hotplug and not by PCMCIA system?

> This is the problem I couldn't get around - the driver module _is_ loaded
> but comedi_config says that it's not.
I have learnt that comedi drivers don't seem to access card hardware until
comedi_configure is invoked - is that correct?

   Andreas Leuner

Received on 2005-08-19Z09:31:16