On Thu, 2004-09-09 at 11:58, Richard Smith wrote:
Li-Ta Lo wrote:
Did they do anything on the "core" x86emu? Last time I checked, XF86 4.4.0 has almost the same x86emu as testbios.
Something is differnt even with 4.3. I can run my ATI M1 bios under testbios and it dosen't work but X with InitPrimary enabled works.
Interesting. Do you have any idea why ?
Not really. There's lots of variables. I do have machines that testbios does work with my M1 bios. But these machines boot a COTS bios rather than LB.
For one thing I think the emu in X handles the reads and writes to the timer rather than letting the bios get to the hardware. The stock M1 bios won't run under vgabios at all. The delay functions go into loops that only exit by chance. The X emu will run that code fine.
I have a modified version of the M1 bios that works around the delay issue by effectively removing them so it runs under testbios but it still dosen't bring the card to life. Its possible that my "fix" is part of the problem since timeing can be critical but I wouldn't expect the emulator to running too fast and I don't think too slow would be a problem.
So you still think it is the timer ? Did you try to access the timer HW in testbios ?
Ollie