Hauppage Conexent 878A video capture, follow up from -current

[bertrand kotewall]

On Wed, Oct 29, 2003 at 08:50:58PM +0100, Steve O'Hara-Smith wrote:
> 	The problem seems to be that the chip ID doesn't match any of
> those listed in the bktr code - 0x036c isn't in the list (the 109e bit
> means Brooktree). You could try hacking a bit and changing the #define
> for PCI_PRODUCT_BROOKTREE_BT878 in sys/dev/bktr/bktr_reg.h (if it
> works then something cleaner involving PCI_PRODUCT_BROOKTREE_BT878A
> getting added would seem to be in order).
> 
> 	I have no idea what the difference between an 878 and an 878A
> is, hopefully not enough to cause flying sparks.
> 
>

I went on the conexant website and they actually assured pin for pin 
compatability.  Feeling confident, I went into /usr/src/sys/dev/bktr/bktr_reg.h,
 changed the 878 entry to 0x036c, then that didn't fly.  It booted, then I did 
a pciconf -lv, and the line was "chip=0x016e109", so I was like...  change the 
entry in 878 to 0x016e...  crash on boot.  Tried it again on 879 and that 
didn't work either.  Crash on boot.  I haven't tried taking it out of the 
kernel and loading it as a module, but I suspect the results would be the same.
  So much for the chips touted pin for pin compatability with the older 
versions...  

If anyone would care to look at the -current archives, I remember Doug White 
responded to my request saying that it wasn't so straight forward as to add in 
the pci id.  I don't know what else to go from here.  conexant's website does 
post the references for the chip, I don't know if they are at all helpful.  
Many thanks though, I'm sure it will work itself out sooner or latter.

							sincerely,

							Bertrand Kotewall