Hi, > > [...] > > I need to divide between Bus 0 and Bus 1, since the SRM firmware > > gets screwed if it finds multiple VGA-Bioses on Bus 0. It seems it > > doesn't try to x86-emulate Boards in Bus 1. > > Can that be an explanation? If a card is not initialised by its BIOS, > it may be in an undocumented state with which Xserver will not work > correctly.
Maybe. On the other hand, running the G200's in an x86 system as multihead *requires* turning of their BIOSes - I did that, by telling the systems Bios to initialize AGP-VGA first. I can't tell, what kind of init the cards really need. Also, for exactly *that* scenario, XFree uses it's libint10.a, to initialize cards that have previously not been initialized by POST procedures. I assume, this fixes Alpha conflicts as well. But it sure is a pointer, that I will follow. > I guess the x86 machine runs all the VGA BIOSes, doesn't it? Hard to tell ( see above ). If I run the x86 with one of the G200's alone, it produces no grafics output ( yes, I switched BIOS back from AGP to PCI VGA ). The card's Bios appears to be turned off. There is a ( DOS based ) Matrox utillity, that is supposed to display several characteristics of an installed Board. Running this with the AGP Voodoo as primary display and the G200 as secondary, it gives *nothing* on the G200 board - I therefore assume, the card has not been inited by BIOS routines. > > There is an x86 emulator which is reported to initialise Matrox cards and > was apparently developed for use in multihead configurations. Check out > ftp://ftp.scitechsoft.com/devel/x86emu/ . Thanx for the pointer. I'll check that anyway. However, one question still remains. Has anyone gotten XFree 4.0 running multihead+xinerama on Alpha's ? Sincerely, Thomas Weyergraf -- Thomas Weyergraf [EMAIL PROTECTED] My Favorite IA64 Opcode-guess ( see arch/ia64/lib/memset.S ) "br.ret.spnt.few" - got back from getting beer, did not spend a lot.
