bktr problems

[Danny Pansters]

On Sunday 11 June 2006 13:47, you wrote:
> --- Danny Pansters <danny at ricin.com> wrote:
> > /usr/src/sys/dev/bktr/bktr_card.c has some interesting comments about
> > (some?)
> >
> > bt878 based cards that have an EEPROM. In short: they've been broken by
> > the time MT2032 tuner support was added, and they're still broken. I've
> > had similar problems with an (old) WinCast TV card; it stopped working
> > somewhere around FreeBSD-5.3 or so, also because tuner detection fails
> > and it can't be bullied into using the correct one with sysctl. I tried
> > to raise the issue several times but never got a response.
>
> When u use a version of bktr_card.[ch] and bktr_tuner.[ch], that is old
> enough, the tuner can be set/detected again?
>
> Could u send us a diff between the current version and the version, that
> detects/behaves nicely?

I'd have to install 5-something and incremently test it. What I can say is 
that I remember that for a while it worked with simply putting the (wrong) 
product/manufactorer ID that was reported in card.c and treating it as 
hauppauge, but later the reported ID became 0xffff and that went nowhere. It 
could probably have been done better and perhaps work with something 
analogous to what Robert Krten wrote.

I also have a miro card and a saa based avermedia one, so I pretty much tossed 
the wintv aside.

I have an amd box here that I need to set up for testing, I'll look which 
install images I still have a CD for and slap in the hauppauge and see if I 
can recreate that situation a little. If I can point to some useful 
timeframes where things broke I'll let you know, if not, well I won't :)

IIRC it was between 5.2 -> 5.3 (hack as described above needed)
and 5.3 -> 5.4/6.0 (b0rken since)

Cheers,

Dan