Re: X -configure seems to choose wrong driver for CT65555 video hardware

beasley wrote:
>
[Lots of test reports]

Thanks for all this data. I think I now have an idea where the problem
is. I'll try to send you a patch to test later today.

> There are certainly many differences between the two files above
> but one that I noticed - that might be important - is that the
> probe results are different. The 4.6-beta log contains this line:
> 
> (--) PCI:*(0@0:4:0) Chips and Technologies 65555 rev 198, Mem @
> 0xfd000000/0
> 
> While the 4.4-release log contains this line:
> 
> (--) PCI:*(0:4:0) Chips and Technologies F65555 HiQVPro rev 198, Mem @
> 0xfd000000/24
> 
> The chips man page mentions three architectures: "Basic", "Wingine",
> and "HiQV" - with the ct65555 chipset falling into this last one.
> 
> Is it possible that the 4.6-beta probe reported information is wrong?
> ... and if so could that cause the chips driver to fail?

This is only cased by the fact that xserver 1.5 now uses the system
pcidevs data to display the name of the PCI chips found by X, instead of
having a duplicated database in X. The OpenBSD one is more terse to save
 room in the kernel.

> 
>> I guess the chips driver fails to load for some reason, I'd like to see
>> if it can be fixed for 4.6.
>>
> 
> Thanks very much for your help on this problem.
> 
> --
> Mike Beasley
> 


-- 
Matthieu Herrb