Ronald G Minnich wrote:
On Thu, 4 May 2000, Jeff Garzik wrote:
Ronald G Minnich wrote: What are the symptoms of the failure? Does the Linux kernel print out any keyboard-related messages? Does booting even get far?
The kernel is up with a login prompt, displaying very nicely on the Matrox (cool logo in the upper left, BTW). When it tries to do anything with the keyboard, however, nothing happens.
Turns out that you have to do a bit of work with the PIIX4E power management registers to get things like keyboard going. I'm still digging.
docs: SiS not on the web yet. They will be happy to send docs, however, those guys are really cool. For Intel, I will try to remember the links, but at the least you need the PIIX4E manual.
I've got the PIIX4* manuals, so all I would need is the southbridge/northbridge databooks. Those are the ones which IMHO may be more important to getting keyboard going than the PIIX4 manual...
Comments from the PIIX4 manual:
* The PM registers relating to keyboard are probably red herrings -- it appears you are only signalling that keyboard events should generate PM event interrupts, which isn't what you are looking for * The Fast A20 bit (bit 1 of port 92 register) is logically combined with the input from the keyboard controller. (x-bus interface, a20gate signal) maybe that is significant, maybe not. * have you enabled X-BUS accesses (function 0, register XBCS 0x4E-4F) * have you enabled the keyboard interrupt? * have you cleared the kbd interrupts (RIRQ, 0x60) * is the USB keyboard interface "stealing" the keyboard port for itself? (register C0-C1) * port 92 (P92) register mentions keyboard * possibly important bits: keyboard bit on DEVRESA register (function 3, bit 28, register offset 0x5C-5F) and DEVRESB register (function 3, bit 30, register offset 0x60-63)