Greetings,
I agree that non volatile is preferred. In the ms7308e, it seems that the CMOS is
initialized from the regular flash BIOS whenever BIOS settings are stored. The only
catch to another location is that the sis900 driver would need a a small patch to work
properly.
Newer 630 (such as 630et) look like they use yet another method that isn't dependant
on the CMOS.
G'day,
sjames
Quoting "Eric W. Biederman" <[EMAIL PROTECTED]>:
> Eric Seppanen <[EMAIL PROTECTED]> writes:
>
> > On Mon, Dec 17, 2001 at 12:57:08PM -0700, Ronald G Minnich wrote:
> > > this keeps coming up.
> > >
> > > so:
> > > typedef unsigned char uc;
> > >
> > > struct cmos {
> > > uc desired_image; // the FLASH/DoC image we want to boot
> > > uc last_booted_image; // The last one to successfully boot
> > > uc unused[7];
> > > uc ethernet_mac[6]; // Ethernet mac address
> > > uc unused2[242];
> > > };
> >
> > Why would you want ethernet address stored there? The ethernet chip
> I
> > use requires a serial EEPROM to be attached to it for the storage of
> > address plus other design-specific details. Don't other chips work
> that
> > way?
>
> It appears that this is one of the cost reductions in the SiS chipset.
> I would prefer that this is stored in nonvaltile flash, than in the
> CMOS though.... I'm aiming at allowing parameters to be stored in
> other locations than just CMOS so from a table perspective I don't see
> a problem.
>
> Having a good description of this would be useful though.
>
> Eric
>
----------------------------steven james, director of research, linux labs
LinuxBIOS Cluster Solutions 230 peachtree st nw ste 2705
High-Speed Colocation, Hosting, atlanta.ga.us 30303
Linux Hardware, Development & Support http://www.linuxlabs.com
* Visit us at SuperComputing 2001, Booth 539 * office/fax 404.577.7747/3
--------------------------------------------------------------------------
