Re: Natl. Inst. pci-dio-96 woes

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 24 September 2002 05:50 pm, David Schleef wrote:
> On Tue, Sep 24, 2002 at 06:10:29PM -0400, Alan Chen wrote:
> > I am having a troublesome problem with comedi and/or the dio board .  It
> > seems that I must perform some akward operations for the board to work
> > correctly.
> >
> > So I am trying to set all the channels (0-23) in all the subdevices
> > (0-3) to either 1 or 0.  It turns out that the following graph is what I
> > get (source pasted below):
> >
> > comedi_dio_write(...,1) alone gets a random amount of 1's and 0's
> > ~ (although they seem clustered)
>
> This sounds like a driver bug.  I've suspected for a while that
> there might be problems in that area, but the code does exactly
> what the manual says, and there have been no complaints with
> other 8255 boards.  I don't have the hardware to experiment with.
>

As I recall my experience with the 8255 chip (or maybe it's just 8255 - 
compatible), when the input/output configuration of any pin is set, it can 
change the state of other output pins.  The only way to make it work right 
was to set up the input/output configuration of all the pins at the beginning 
and then leave them alone.  I also looked for bugs in the driver and found 
none.  I suppose it could force a write to all the output pins every time the 
configuration is changed, although that still would leave transient glitches.

- -- 
Frank

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE9kPaV5vihyNWuA4URAoBqAJ9pqZm3CjYzc3Ef6g5xQkJKSDjvJQCgwMeK
rTdomt3aMuvqtMy5pLVR+QQ=
=DJvs
-----END PGP SIGNATURE-----

Received on 2002-09-24Z22:34:43