Re: Open a discussion about bug 253324
-------------------
> Hi, Sven Luther wrote:
>
> > This is bullshit. I don't think it is a great idea to modularize
the
> > kernel, and thus lose all the early debugging. We are going to
hurt us if
> > we go that way, especially as the boot process is now more fragile
as it
> > used to be, thanks to the initrd thingy,
>
> The workaround for this problem is to modularize everything that
could
> possibly break or hang the kernel. Load it after vesafb-or-whatever.
> Problem solved.
I'd like to know how a modularized PCI subsystem can be loaded _AFTER_
a PCI dependent FB driver.
What then if the PCI buswalk kills the kernel?
I'm not even dealing with AGP here :P
This is a gross example, still it proves that kind of idea can't be
implemented.
HTH,
Thibaut VARENE
PA/Linux ESIEE Team
http://www.pateam.org/
Reply to: