> > Anyone who knows what was the good reason for initializing the console _before_
> > the PCI code initialization is called? This causes problems for systems where
> > the console is on a PCI device (e.g. a TGA console, for which a workaround is
> > used).
> >
> > If there's no good reason for this, I'd like to invert the order.
>
> Is there any chance that PCI scan will hang the computer etc.? If so,
> than it is because this way you can _see_ where it died.
>
> IMO console should be enabled as soon as possible so we do not get
> 'freeze after Uncompressing linux...' error reports.
I agree. I had an AMD 486/100 w/ an AMI Bios that didn't fully support PCI
(it had unadvertised PNP support) and its structure to point to the PCI
BIOS32 extensions was there and filled, but it pointed to invalid memory.
This machine froze many times. Because it spit out all that handy info, I
was able to easily (this was the first time I was installing linux)
recompile the kernel on a friend's box just disabling PCI -- if it hadn't
displayed the error messages, I probably wouldn't be using Linux today.
Greg Alexander - also <gralexan@indiana.edu> - http://sietch.home.ml.org/
----
"What is the best manner in which to seek the nonexistant?" asked a
young man of a wise man.
"All methods are equally wrotten except those which succeed, those
are the ones to be feared," responded the wise man.
-- Rare saying