Dual BIOS Motherboards

[Matthew Sullivan]

Hi All

I don't recall anyone saying anything about this in the recent future, but
I stumbled over the Specs to the new GigaByte motherboard (GA-BX2000)
earlier... - It offers 'Dual BIOS'

Follows is an extract of the related article on Gigabyte site:

Enter the concept of DualBIOS. The intent of DualBIOS, as you briefly
noted, is to provide a second BIOS in case of the infrequent occurrence of
a "BIOS failure". Without going into a long explanation, if the primary
BIOS "fails", the secondary BIOS takes over for the primary and you keep
running. Effectively the secondary BIOS acts as a "hot spare". If the BIOS
is actually good (electronically), but its data has become corrupted,
DualBIOS provides a copy of good BIOS code (stored in the secondary BIOS)
to allow you to use the included, automatically-activated utility to flash
the primary BIOS "back to health." If the BIOS is really dead
(electronically dead), then the secondary BIOS just keeps on going on its
own. Rather a simple, but innovative, approach to this problem.

The full story at: http://www.giga-byte.com/gigabyte-web/whydual.htm

Probably on of the most useful boards on teh market I would say, though we
would have to check that the first BIOS could be programed without altering
the second BIOS code....

Yours

Matthew Sullivan

[James Oakley]

Matthew Sullivan wrote:
> The full story at: http://www.giga-byte.com/gigabyte-web/whydual.htm
>
> Probably on of the most useful boards on teh market I would say, though we
> would have to check that the first BIOS could be programed without altering
> the second BIOS code....

I saw this board a while back and was quite interested (in fact, I was
thinking of picking one up). The two main reasons for having the dual
flash is to protect from 1) flash virii, and 2) interrupted/corrupted
flash updating. Therefore, It should be quite neat for our purposes and
it doesn't matter what code you put on it.

The OB ISA card could easily mirror this functionality. You could, say,
put code to start with the other BIOS in the boot-block and restore the
original BIOS. Oh yeah, I haven't had time to prototype the card yet,
but I'm moving to a place with more room for doing work this weekend, so
I may actually get it working. Anybody have comments on the circuit?
I've never designed an ISA card before so I don't know if I fell into
any traps (yes, I know there are no caps and I know there should be
some, I'll play with that on my prototype card).

I don't know why I didn't think of it before, but, maybe it would be a
good idea to implement this functionality into OB itself. Like
automatically detecting a bad flash and reading from the other one. This
requires a slight hardware change to the card to allow writing to the
backup flash, but that's no biggie.

Besides, we will have to implement that functionality in order to
support that board at all. What if this dual-flash thing catches on?

James Oakley
[e-mail]
-
To Unsubscribe: send mail to [e-mail]
with "unsubscribe openbios" in the body of the message

[Matthew Sullivan]

Hi,

James Oakley wrote:
> Matthew Sullivan wrote:
> > The full story at: http://www.giga-byte.com/gigabyte-web/whydual.htm
> >
> > Probably on of the most useful boards on teh market I would say, though we
> > would have to check that the first BIOS could be programed without altering
> > the second BIOS code....
>
> I saw this board a while back and was quite interested (in fact, I was
> thinking of picking one up). The two main reasons for having the dual
> flash is to protect from 1) flash virii, and 2) interrupted/corrupted
> flash updating. Therefore, It should be quite neat for our purposes and
> it doesn't matter what code you put on it.

Ah, but you missed the point that I was/am worried about.  Can _we_ write our
code to BIOS1 and not to BIOS2 and if our code is good in BIOS1 will it get
copied to BIOS2 automatically on successful boot? How does BIOS2 deturmine if
BIOS1 has corrupt code - will it detect our code as corrupt etc etc etc.... I
think we need to talk to gigabyte about the board, and functionality on a
technical level.

I am currently looking at building a new machine and that M/B maybe a good choice
if the DualBIOS is usable...

----[snip - out of time for a few hours]----

more later...

Yours

Mat

[James Oakley]

Matthew Sullivan wrote:
> Ah, but you missed the point that I was/am worried about.  Can _we_ write our
> code to BIOS1 and not to BIOS2 and if our code is good in BIOS1 will it get
> copied to BIOS2 automatically on successful boot? How does BIOS2 deturmine if
> BIOS1 has corrupt code - will it detect our code as corrupt etc etc etc.... I
> think we need to talk to gigabyte about the board, and functionality on a
> technical level.

Yeah, I'm not sure exactly how it works, we'll have to ask Gigabyte
about that. However, if I had to guess, I'd say that the functionality
is present in the boot block. To determine if a flashing went awry, it's
probably a simple checksum. For the copying between chips, I believe the
BIOS program does it interactively, IIRC.

I hope Gigabyte doesn't mind giving us that information. Remember that
they are the only ones with dual-flash and they probably don't want the
world to see the details...

Hmmm, maybe we can design a simple plug-in dual flash device (without
the extra features of the current design) to provide this functionality
on any board, meant for public consumption. We could sell them to fund
OpenBIOS development. They might interest people who have x86 Win
servers and the like, for protection from flash virii and bad flashing.
We could make it tiny enough to fit right into current sockets. We could
even make it the same size as a DIP flash using quad-flatpack flash
ROMs. It might be hard to automatically switch between them if we only
use the pins in the flash socket but we could place a switch on there
and have a 'bad' BIOS tell the user to manually switch to the other one.
It's only a tiny circuit and the boards would be quite small.

James Oakley
[e-mail]
-
To Unsubscribe: send mail to [e-mail]
with "unsubscribe openbios" in the body of the message