* Jeff Garzik <[EMAIL PROTECTED]> [020305 21:08]:
> > need an OS to see whether your second graphics adapter works? Why handle
> > graphics adapters so much different than scsi host adapters.
>
> Because all graphics adapters are trying to decode the same VGA region.
This can be handled by hiding the other devices during setup, like done
in XFree86 afaik.
> But, as I keep saying, this is easily Just Another Option(tm) in the
> freebios config. You -must- implement init-only-one-VGA-adapter if the
> BIOS ever expects to be compliant with existing BIOS standards(sic).
Is it even suggested by a standard that only the first device has to be
initialized? Thought this is just a matter of hardware vendors not doing
it the right way due to compatibility reasons or whatever.
> Graphics adapters are always a special case. Yes, it's annoying :)
But this is on a hardware level. Firmware's task can and should be to
hide these misimplementations in a sane and (i.e. for the OS) reliable
way.
> If we are talking about freebios a.k.a. firmware console image, then
> sure, it can and -should- perform graphics adapter initialization,
> including support of the needed minimal set of BIOS INT calls.
>
> Here is my desired view of booting:
>
> linuxbios -> freebios -> Linux kernel
>
> linuxbios merely initializes CPU and DRAM. freebios is the more
> generic, portable piece that initializes PCI, graphics, and IDE. And
> loads the Linux kernel.
Exactly. This is the point. Certain PCI device initialization has
nothing to do with whether your CPU is 32/64bit, big or little endian,
your northbridge copes with this or that kind of ram modules. The
clearer we get this seperated, the easier it it to support a large
amount of existing hardware.
Best regards,
Stefan Reinauer
--
Ok hex 4666 dup negate do i 4000 dup 2* negate do " *" 0 dup 2dup 1e 0 do
2swap * e >>a 2* 3 pick + -rot - j + dup dup * e >>a rot dup dup * e >>a
rot swap 2dup + 10000 > if 3drop 3drop " " 0 dup 2dup leave then loop
2drop 2drop type 266 +loop cr 5de +loop
