OEM looking at LinuxBIOS

Eric W. Biederman ebiederman at lnxi.com
Thu Mar 11 10:37:00 CET 2004

"Hendricks David W." <dwh at lanl.gov> writes:

> There is already 82803 and 82801ca code in the freebios1 tree, might as 
> well put it to good use :) I think Eric Beiderman was even able to fix 
> SPD, at least on the MCH4 (Or something).
> On Tue, 9 Mar 2004, Anthony Winner wrote:
> > Hello all,
> > 
> > I work for a OEM and I'm playing with the idea of using linuxBIOS, I have
> > all of the technical info for the machine short of the actual source code to
> > the Phoenix BIOS we use now as it was done by a contractor before I joined
> > the company.
> > 
> > Are there any pointers for a 'from scratch', implementation?
> > 
> > Should I start with the old code or start fresh from the new freebios2 tree?
> >
> > For reference, the machine is a Intel PIII with a 82830 Northbridge, a
> > 82801ca Southbridge, TI pci1450 card bus controller, and a pair of SMC SIO
> > chips....

I don't think we have good support for th 82830 at this time so the freebios2
tree would be the one to work with.  You get romcc so you can code the
memory initialization in C.  If you know what you are doing a huge chunk of
assembly is tolerable.  But doing things in C prevents a lot of stupid typos.

I just looked the code in the freebios1 tree for the 82830 does not even
attempt to do SPD based memory initialization so it will not work with
arbitrary memory.  So definitely freebios2 using some of the freebios1
code as reference. 

I don't think the freebios1 tree can code with a pair of superio chips either.

The freebios2 tree comes with multiple devices of the same type trivially.


More information about the coreboot mailing list