[LinuxBIOS] Gigabyte's DualBios and recovery from a botched Linuxbios flash?

I was preparing to flash an ECS NForce4-A939 motherboard with Linuxbios
using advice I received earlier this week on this mailing list.  Since the
bios chip is a surface mount package, I have no easy way of restoring the
original manufacturer bios image should I flash an invalid rom image to the
motherboard.

To minimize my risk, I planned to flash a Gigabyte motherboard with a
similar chipset configuration, just to prove my image would theoreticaly
work on the ECS board.  This motherboard has something called Dualbios,
which essentially allows you to boot from a secondary bios image should
something go wrong.  However, I noticed that to activate this feature and
access the backup BIOS image, you had to enter the CMOS setup for the
motherboard, whereas I initally assumed the backup mechanism was jumper
activated.

So my questions are:

1. Does anyone know if the Gigabyte Dualbios feature will still work once
I flash the primary bios image with Linuxbios.  If so, it would allow me to
recover if my first image is bad.  I don't see how it would work, since I
assume Linuxbios will overwrite any recovery logic, rendering the backup
bios image inaccessible.

2. Other than using a fallback image in my Linuxbios image, or using this
Dualbios feature, are there any other economical ways to recover from a bad
Linuxbios flash, when dealing with a non-socketed bios chip?

I just have this sinking feeling that I'll flash my MB, and end up with a
$75 brick, hence these questions...

Thanks,
Brad Brown

--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .

[LinuxBIOS] Gigabyte's DualBios and recovery from a botched Linuxbios flash?

[David Hendricks]

MSI does the same thing with their dual BIOS approach--I found
out the hard way from purchasing a K8N-Neo3. My advice is to avoid
these gimmicky 0v3rc!0ck3r BIOSes and get a BIOS Savior RD-1 or
something if you can get your hands on a socketed mainboard.

Or, I think someone on this list (Bari Ari?) can modify your board to
use socketed flash parts as well.

--David

[Richard Smith]

On 12/27/05, David Hendricks  wrote:
> MSI does the same thing with their dual BIOS approach--I found
> out the hard way from purchasing a K8N-Neo3. My advice is to avoid
> these gimmicky 0v3rc!0ck3r BIOSes and get a BIOS Savior RD-1 or
> something if you can get your hands on a socketed mainboard.

If you can find out what the address is that the recovery image jumps
to after boot then it seems possible to add Linuxbios there without
nuking your recovery code.

You would have to modify the flash programer to not erase the sections
where your recovery stuff was and add some sort of offset ability as
well.
> Or, I think someone on this list (Bari Ari?) can modify your board to
> use socketed flash parts as well.

Is the flash a PLCC or TSOP?

[LinuxBIOS] Gigabyte's DualBios and recovery from a botched Linuxbios flash?

[Bari Ari]

David Hendricks wrote:
> Or, I think someone on this list (Bari Ari?) can modify your board to
> use socketed flash parts as well.

If there is enough room around the tsop flash device on the mainboard, 
the flash device may be replaced by a ZIF tsop socket. Such as one by:

http://www.emulation.com/catalog/off-the-shelf_solutions/sockets/tsop/

http://www.emulation.com/pdf/skt518.pdf

The tsop zif sockets are larger than the tsop package so you need the 
extra space to solder them in place.

-Bari

[LinuxBIOS] Gigabyte's DualBios and recovery from a botched Linuxbios flash?

[David Hendricks]

Oops--I forgot that I'm not running my old crappy mainboard anymore and
my current one *does* have a socketed PLCC.

Richard does have a good idea, though. Yhlu added a start and end
exclusion option to flash rom to specify which portions of the ROM to
erase and reprogram. I have not tried it yet as this patch was made
specifically for the pm49fl004 (The standard ROM on a bunch of Tyan
boards). I also seem to recall that this required IO chipset support
for legacy programming mode, but I could be wrong on that point. Still,
perhaps this would help out Mr. Brown if it could be adapted to
whatever his board uses.

PS -- Sorry for the dupe, Richard!

On Wed, 28 Dec 2005 00:06:18 -0600
Richard Smith  wrote:
> On 12/27/05, David Hendricks  wrote:
> > MSI does the same thing with their dual BIOS approach--I found
> > out the hard way from purchasing a K8N-Neo3. My advice is to avoid
> > these gimmicky 0v3rc!0ck3r BIOSes and get a BIOS Savior RD-1 or
> > something if you can get your hands on a socketed mainboard.
>
> If you can find out what the address is that the recovery image jumps
> to after boot then it seems possible to add Linuxbios there without
> nuking your recovery code.
>
> You would have to modify the flash programer to not erase the sections
> where your recovery stuff was and add some sort of offset ability as
> well.
> > Or, I think someone on this list (Bari Ari?) can modify your board
> > to use socketed flash parts as well.
>
> Is the flash a PLCC or TSOP?