On Mon, Feb 04, 2002 at 07:25:17PM +0000, Ben Dooks wrote: > On Mon, Feb 04, 2002 at 07:16:49PM +0000, Russell King - ARM Linux wrote: > > On Mon, Feb 04, 2002 at 07:03:48PM +0000, Ben Dooks wrote: > > > I think we have a PCI mapping problem, as registers seem to be comming > > > back as 0xff... > > > > Are those the config registers, or the io/mmio registers? > > io/mmio > > > The card will play dumb until you wake it up with a specific sequence > > of writes (see the first 3 writes in cyber2000fb.c > > cyberpro_pci_enable_mmio). > > ok, will check
solved the problem now... seems the PC BIOS wasn't leaving the card in an state to work with MMIO... will try and clean off my debugging and submit the patch to get this working. > > > > The solution? I have no idea - I don't have any clues to suggest why > > > > these chips behave like this, nor what tweaks may be needed to make > > > > them behave. However, there are rumours around that indicate that > > > > IGS did admit that the PCI parity logic inside the CyberPro chips > > > > were broken. > > > > > > Your bios32.c is disabling parity errors (iirr, correctly) > > > > You also need to disable SERR as well. > > yes, it is already doing that. -- Ben Nothing is so absurd that it cannot have been said by a philosopher. _______________________________________________ http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm http://www.arm.linux.org.uk/armlinux/mailinglists.php Please visit the above addresses for information on this list.
