[linux-dvb] Re: [rfc/patch] pll handling and cx22702 update

[Gerd Knorr <kraxel@bytesex.org>]

On Thu, Dec 09, 2004 at 03:09:52PM +0000, Andrew Wilson wrote:
> Hi Gerd
> 
> Does this change preserve functionality?

Yes.  At least I don't *intentionally* break things ;)

> I'm looking at a bug in the LR6650 card (cx88/cx22702) where it's not 
> tuning to certain muxes - from earlier messages in this forum other 
> people have had the same problem.

Hmm, I have that problem as well.  The very same muxes have problems
with completely different hardware through (old + new hauppauge nova-t),
so I think is's because of reflections here.

> The odd thing is that it's the strongest mux that has difficulties.
> I've got it working ok in windows, so it does look like a driver bug.

Never tried to use the card with windows, maybe I should and compare
the results ...

> I'm no hardware guy, and a lot of the info on the datasheet goes over my 
> head. Do you (or anyone else here) have time to look at this?

Well, it's hard to figure if you can't re-produce it locally.  I also
didn't wrote the cx22702 frontend initially, I only did a number of
cleanups.  The actual hardware handling is untouched.

Note that the driver allways configures the hardware to autodetect
because manually setting the parameters doesn't word (a comment in
the source says so).  So it isn't that unlikely that some bugs left
in there ...

> Would it be possible to reverse engineer the windows driver or snoop the 
> PCI bus to find out what values it is poking into these registers?

Don't know, never tried that.

  Gerd

-- 
#define printk(args...) fprintf(stderr, ## args)