Bios Mods -The Best BIOS Update and Modification Source

Full Version: Assistance with the A25 series (Toshiba)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
So...something as old as these is probly dead to most of y'all...but this is what
has brought me to your humble(?) group....namely...

A host of A25 laptops...all of these route to the same bios on Toshiba's
support site...sure, v 1.40 sounds flashy, but here I am with this poor ol'
2.66/512/400 chip sitting in the machine...and a flashy 2.66/1m/533 chip
right next to me. Or perhaps a 3.06/1m/800. Both of the other chips, aside
from having double the L2, are hyperthreaders. I at one point thought that
might be part of the limitation, but lo' and behold, a 2.4/512/800 HT chip
boots just fine. So obviously, this is more a matter of updating microcode.

Now, I know that we are supposed to have a donor bios, and the original it possible to just use the linux microcode database for adding to
the rom? The list as such shows well more than just my socket 478 cpu's
I seek to add support for, but what the [censored], right?

Donor link: (hopefully)

Original Bios link:

Now, from here, I want to point's not that I would mind doing it myself...
by hand if need be with a hex editor, or some such...
but I lack the knowledge of what tools are even going to touch the files,
let alone allow me to alter them, and see if I brick my machine or not lol

So...if someone can try this, I'd appreciate it...if you can direct me to at least
what editor will open this (Is it an ACPI bios like the Tecras? I'm finding
nothing for good reference material on this...)

For lack of anything better to do in the meantime, I'm going to try a pheonix
editor and see if anything happens, heh.

Updates, suggestions, and advice are welcome, always!
Update on progress thus far:

After digging, and searching, and attempting every tool i can think of, I've now tried a variety of resources.

On one hand, I now have a handy-dandy list of CPU microcodes straight from intel...ignoring that for my processor and chipset i have like 6 codes listed up, I'll worry about revision later.

The biggest problem I seem to be having is in the nature of the bios itself. Being a basic Toshiba written bios, labelled 'ACPI bios' (which has been utterly irrelevant to my efforts!), it is for a 512k chip...seems most of the tools available are for 1mb and 2mb bios's and chips...(now, I could just be wrong and not have been using a proper tool version, so, again, help? Suggestions? Advice?)

I've taken to looking at the bios via hex-editor, and have now used the bios-backup toolkit to get a different look at it (which makes it look VASTLY different in the hex-editor, which leads me to believe the original bios file may be encrypted, or just written in a different language? maybe if I open it in a C+ program language editor?) and i'm able to find the identifier for the CPU code header...but therein the similarities end, for looking at any of the information AROUND the header doesn't look remotely like anything in the microcode strings, leaving me with (surprise surprise) no forward progress.

I'm beginning to wonder if my actual microcode strings are stored on some completely different chip that is motherboard specific and all it's actually doing is referring to said flashed, this would explain why the A25-S207 is using the exact same bios as the A25-S3072, yet will not support a processor higher than the P4 2.66/512/400 (or the P4 2.4/512/800 HT that is presently in it), as opposed to the P4 3.06/512/400 that is in the 3072 model. Same motherboard, same BIOS...similar limitations.

This may come down to my limitations are all going to be whatever works on the chipset microcode ID. My actual goal being to potentially increase to Gallatin support, as that would be freakin' sweet for my 1000's of tabs I have loaded at any given time, plus the occasional game of Diablo via IPX network (kickin' it old school with my roomies), and perhaps letting it crank away at some video conversion in the background (It really is quite remarkable what you can get away with with a strong enough P4).

Doing some identification with AIDA64 between this laptop, and an ABIT IC-7G running a 3.2 Gallatin returns this set of references:

00000F24h Pentium 4 2.4/512/800 HT

00000F25h Pentium 4 3.2/2m/800 HT

If you look at the CPUs physically, the capacitors on the bottom of the chip double noticeably from the initial 2.66, to the 2.4HT, to the Gallatin 3.2HT, yet the microcode revisions for the first two run under the 00000F24h string.

So far as I can tell, the only actually important information is the 00000F24...I cannot figure out which revision/platform 'h' refers to, if any. There are 3 separate files of 00000F24 and 4 for 00000F25, and I won't rule out the idea that some of the later F25 turns into socket 775 support before shifting to F26...

Suggestions? Thoughts? Anyone? Tongue
Ok, well, more digging...and some updated info on what I've found and done so far...

((I would have posted this HOURS ago, except the bleedin' server was down...anyhew...))

So, the initial bios is encrypted (thus, I'm going to be looking for a different flash utility when this is done, as I do not trust Toshiba's programmers from this time period to be intelligent enough for their flash utility to not run a decryption algorithm on a non-encrypted bios image...), which is half of what makes the bloody .com files so ridiculous to figure out. However, as previously stated, I ingeniously copied the active bios file with the Bios Backup Toolkit...thus, unencrypted.

So scanning through the bios file in a hex editor, looking for the F24h headers...and not finding them. Anywhere. Yet I keep seeing a couple groups of code that have very -similar- headers...start loading up other series into the hex editor, and 'lo and behold! It's got 3 F27h, and 2 different F29h microcodes. This confuses me, as first, none of these laptops (maybe some of the A30 series?) use those processors, and secondly, a good portion of the processors running those microcodes are socket this seems a very odd choice for the Toshiba programmers.

Think I'm gonna skip the F24h codes and just load all 3 F25h microcodes I have to see if I can get Gallatin support...which should support anything 'lower' as well...and from there work backwards if I need to to gain extra support...I'm going to try to figure out how to limit myself to 3 codes so I don't have to move anything else in the hex editor to make space, as I'm not sure I CAN put a 4th in without overwriting some of the other data...not that there isn't plenty of blank space to work with and move things around within, but I don't want to get to tweaking too terribly much, as I now can't seem to find any form of emergency flasher...Phoenix Crisis Recovery Disk seems to be rather hard to find right now (about 3/4 of the download links I could find were broken, and a USB floppy drive is not a present option for me)...though I have -a- copy or two of it downloaded at present, so hoping that at least one of what I have found will do USB stick support, or at least CD support.

Oh, and btw, for those of you who are having problems finding cross referencing information beyond the identifiers as given by Aida64, which will tell you what chipset the processor IN a machine at the moment needs for CPU microcode, and doesn't tell you a dang thing about what the processor your system DOES NOT support will need, I finally tracked down the reference sheets in Intel's 'Ark' site, and once you select a processor, and load up 'documents and downloads' links for it, you can get a link for 'Specification updates' which will load up a PDF with, amazingly, a table of information on the CPU's in a given family with their CPUID to work I just need to figure out what the 'platform' number means...

This is fun :-P Almost there!