hate to be clogging up the list but memcpy works!! It looks like the VT8622/8623/8635 all have the same way of dealing with this shadow copy stuff. After losing a couple of hours sleep, I've hit another snag here (partial log):
==================================== setting pci slot setting vt8235 slot Assigning IRQ 5 to 0:11.1 Readback = 5 Assigning IRQ 12 to 0:11.5 Readback = 12 INSTALL REAL-MODE IDT DO THE VGA BIOS found VGA: vid=1106, did=3122 Setting up VGA, VGABIOS_START: 0xfffc0000 write_protect_vgabios Looks like VGA_ROM is good: 0x55 0xaa VGA copy looks good bus/devfn = 0x100 biosint: # 0x15, eax 0x5f00 ebx 0x100 ecx 0x100 edx 0x12 biosint: ebp 0x138dc esp 0xff2 edi 0xf934 esi 0x141f8 biosint: ip 0x637f cs 0xc000 flags 0x46 calling handleint21() done with handleint21() returning from biosint()...ret=-1 biosint: # 0x1a, eax 0xb108 ebx 0x0 ecx 0x0 edx 0x3d5 biosint: ebp 0x138dc esp 0xfcc edi 0xf6 esi 0x155eb biosint: ip 0x40da cs 0xc000 flags 0x46 calling pcibios() 0xb108: bus 0 devfn 0x0 reg 0xf6 val 0x3 done with pcibios() returning from biosint()...ret=0 =================================================
I've tried to put some debugging outputs but the system just grinds to a halt after the second call to biosint(). It's now my belief that directly manipulating the registers on the Apollo CLE266 northbridge might yield better results for the epia M. VIA doesn't seem inclined to give out the datasheets or source code, as the posts on viaarena indicate. I wonder if looking at the linux driver patches for kernel 2.4 and greater might offer some clues. Any help would be greatly appreciated!