[LinuxBIOS] Intel 440bx

Al Boldi a1426z at gawab.com
Tue Feb 27 12:07:14 CET 2007


Corey Osgood wrote:
> Al Boldi wrote:
> >>> What I can't understand, why does v1 work whereas v2 doesn't?
> >>
> >> Probably because the code currently in the svn looks for the vt8601,
> >> and attempts to program the wrong registers :-p
> >
> > I still don't get it.  What's the big difference between v1 and v2?
>
> No one ported the 440bx from v1 up to v2. Instead, someone copied the
> work that was done on via vt8601 into the i440bx folder probably with
> the intention of porting that, but never got to it. And the way v1 and
> v2 work are very different, not to mention the coding styles. In v2, we
> try to avoid assembly code as much as possible, keeping almost strictly
> to c. v1 otoh has lots of assembly. not to mention that the file
> structures are very very different,

Ok, I think I get it now, thanks for the clarification.

> just compare and see for yourself.

I probably will, but I need v1 to compile first as a starting point.

> >>> Can you post a working v1 tarball link as well?
> >>
> >> http://www.openbios.org/viewvc/trunk/LinuxBIOSv1.tar.gz?view=tar
> >
> > I downloaded this once, but it did not compile with the current
> > distributions.  Has this been updated now?
>
> Use an older compiler, newer gcc versions don't seem to like it. I think
> I used gcc-3.3 on ubuntu.

I have gcc-3.2, which compiles v2 ok, but doesn't compile v1.  Does v1 need 
something special?

> OTOH, compiling v1 is of very little value: we
> know it works, and we have the code, that's all that really matters.

Ok, but when porting code, you should always make sure that the code you are 
porting from works on the specific hw you want to port to, otherwise you may 
wind up wasting a lot of time debugging minute variations due to hw.


Thanks!

--
Al





More information about the coreboot mailing list