Universal Single Method USB Flash Drive Boot and Run
Mike Gonta
http://board.flatassembler.net/topic.php?t=12488
Mike Gonta
look and see - many look but few see
Rod Pemberton
news:iikck4$7e7$1...@speranza.aioe.org...
"1. The USB flash drive was first released commercially in 2000, so any PC
before then can't be expected to boot and run from USB flash drive."
Is that true? ...
Well, you meant just with the stock BIOS, yes? You weren't referring to
after market USB cards, yes? I have an add-in USB card for an older PC.
I've not yet checked to see if it can boot a USB stick. I hope it can...
"2. All three major BIOS (Phoenix acquired Award in 1998 but operates as
separate divisions) with a date after mid 2004 can be expected to universal
single method USB flash drive boot and run (green squares)."
How do you know this is supported after 2004? Is it part of a spec.?
Who do Compaq and HP use for their BIOS? Both are major computer
vendors, yes?
"... (green squares)"
Mike? What green squares? Neither Mozilla Firefox nor Opera (current
versions) show any green squares... OK, IE6 shows green squares... Hmm, it
seems that the two "modern" browsers are clipping the image. Maybe .png is
an issue? Or, maybe the size of the .png is an issue?
"The boot sector 0AA55h signature at 510h"
Is this always required? For floppies it wasn't, per the IBM tech. ref.
man. Until recently, I never found a PC that require it for floppies.
However, this new PC requires it on floppies and USB floppy images also.
It's the first I'm aware of. I've not been able to locate if 0AA55h on
floppies and floppy USB images is a industry-wide BIOS change, or just a
BIOS bug for this machine. I'd like to know. If you know the spec., could
you mention it?
Rod Pemberton
Mike Gonta
> "Mike Gonta" wrote in message
> news:iikck4$7e7$1...@speranza.aioe.org...
>> Time line, description and simple test code:
Note: the code can be assembled in FASM or NASM.
>> http://board.flatassembler.net/topic.php?t=12488
> "1. The USB flash drive was first released commercially in 2000, so any PC
> before then can't be expected to boot and run from USB flash drive."
>
> Is that true? ...
>
> Well, you meant just with the stock BIOS, yes?
Hi Rod,
Yes, this is a generalization, since one cannot be expected to test all
PC's past present and future. I've currently checked 3 PC's with Award BIOS
08/08/01, 09/24/01 and 05/28/02 (month/day/year located at 0FFFF5h) which
have no setup for USB booting.
> You weren't referring to after market USB cards, yes? I have an add-in
> USB card for an older PC. I've not yet checked to see if it can boot a
> USB stick. I hope it can...
Without an option ROM it would depend on the BIOS version.
> "2. All three major BIOS (Phoenix acquired Award in 1998 but operates as
> separate divisions) with a date after mid 2004 can be expected to
> universal
> single method USB flash drive boot and run (green squares)."
>
> How do you know this is supported after 2004?
Technically there is no problem (other and beginner issues) USB booting
with the MBR method on a current PC. http://reboot.pro
I'm generalizing that success prior to 2004 with the "big 3" would not
change. I've currently tested 10 PC's with BIOS dates between then and now.
> Who do Compaq and HP use for their BIOS? Both are major computer
> vendors, yes?
Today, HP owns Compaq and like Dell it's going to one of the "big 3"
depending the year. The original Compaq wrote their own BIOS, which
is famous for the 7C00h:0 boot sector segment:offset and a lawsuit
with IBM.
http://wimsbios.com has information on the current versions of 223
manufactures.
An examination of the Debug (d)ump of F000:0 will reveal the BIOS maker.
1. make a debug script and name it dump.txt:
d F000:0 FFFF
q
1
Note: the "1" on the third line is ignored and can be left out, the "q" on
the second line needs to be followed by a carriage return. This will q(uit)
debug.
2. make a dump batch file and name it dump.cmd:
debug < dump.txt > f000.txt
3. run dump.cmd and examine the f000.txt file for the BIOS maker
> "... (green squares)"
>
> Mike? What green squares? Neither Mozilla Firefox nor Opera (current
> versions) show any green squares... OK, IE6 shows green squares... Hmm,
> it
> seems that the two "modern" browsers are clipping the image. Maybe .png
> is
> an issue? Or, maybe the size of the .png is an issue?
I've now reduced the image size.
> "The boot sector 0AA55h signature at 510h"
>
> Is this always required? For floppies it wasn't, per the IBM tech. ref.
> man. Until recently, I never found a PC that require it for floppies.
> However, this new PC requires it on floppies and USB floppy images also.
> It's the first I'm aware of. I've not been able to locate if 0AA55h on
> floppies and floppy USB images is a industry-wide BIOS change, or just a
> BIOS bug for this machine. I'd like to know. If you know the spec.,
> could you mention it?
This was historically true for the floppy diskette. The signature was
checked for fixed disks and option ROMs only.
Some BIOS (maker and version) interpret the USB booting as HDD and require
the signature.
Maxim S. Shatskih
> is famous for the 7C00h:0 boot sector segment:offset and a lawsuit
I had Compaq Presario produced about 1998 - Pentium MMX with VX chipset.
It surely had boot entry point of 07c0:0000
--
Maxim S. Shatskih
Windows DDK MVP
ma...@storagecraft.com
http://www.storagecraft.com
Mike Gonta
> depending the year. The original Compaq wrote their own BIOS, which
> is famous for the 7C00h:0 boot sector segment:offset and a lawsuit
> I had Compaq Presario produced about 1998 - Pentium MMX with VX chipset.
> It surely had boot entry point of 07c0:0000
Hi Maxim,
Thanks for correcting my typo.
s_dub...@yahoo.com
wrote:
> "Mike Gonta" <mikego...@REMOVEgmailTHIS.com> wrote in message
>
> news:iikck4$7e7$1...@speranza.aioe.org...
>
> > Time line, description and simple test code:
>
> >http://board.flatassembler.net/topic.php?t=12488
>
> "1. The USB flash drive was first released commercially in 2000, so any PC
> before then can't be expected to boot and run from USB flash drive."
>
> Is that true? ...
>
> Well, you meant just with the stock BIOS, yes? You weren't referring to
> after market USB cards, yes? I have an add-in USB card for an older PC.
> I've not yet checked to see if it can boot a USB stick. I hope it can...
>
to speeds supported. I don't really know, I'm stating a supposition.
Your motherboard RomBios may only support usb 1.0, yet the add-in usb
option rom may support usb 2.0 and booting from usb.
> "2. All three major BIOS (Phoenix acquired Award in 1998 but operates as
> separate divisions) with a date after mid 2004 can be expected to universal
> single method USB flash drive boot and run (green squares)."
>
> How do you know this is supported after 2004? Is it part of a spec.?
>
> Who do Compaq and HP use for their BIOS? Both are major computer
> vendors, yes?
>
> "... (green squares)"
>
> Mike? What green squares? Neither Mozilla Firefox nor Opera (current
> versions) show any green squares... OK, IE6 shows green squares... Hmm, it
> seems that the two "modern" browsers are clipping the image. Maybe .png is
> an issue? Or, maybe the size of the .png is an issue?
>
> "The boot sector 0AA55h signature at 510h"
>
> Is this always required? For floppies it wasn't, per the IBM tech. ref.
> man. Until recently, I never found a PC that require it for floppies.
In my case with my XT clone, the 0AA55h signature _is_ required for
floppies _only_ with its seagate ST-01 scsi hd controller installed.
It turns out that the option rom firmware for the ST-01 uses int 40h
and revectors int 13h thru int 40h, and enforces the signature for
floppies, 5 1/4 floppies at that. Much later I obtained a Future
Domain controller which didn't enforce that requirement. - the point
is there's more to it than the motherboard RomBios.
Steve
Rod Pemberton
> "Rod Pemberton" wrote in message news:iillte$8uj$1...@speranza.aioe.org...
> > "Mike Gonta" wrote in message
> > news:iikck4$7e7$1...@speranza.aioe.org...
> >> Time line, description and simple test code:
>
> Note: the code can be assembled in FASM or NASM.
>
> I've currently tested 10 PC's with BIOS dates between then and now.
>
Uh, why does your test work - to confirm USB booting? It seems to just use
extended read... ? Is that your "gauge" for yes or no on USB booting? Will
the test confirm if used on an actual floppy? Or, only if used from USB
device?
> > "The boot sector 0AA55h signature at 510h"
> >
> > Is this always required? For floppies it wasn't, per the IBM tech. ref.
> > man. Until recently, I never found a PC that require it for floppies.
> > However, this new PC requires it on floppies and USB floppy images also.
> > It's the first I'm aware of. I've not been able to locate if 0AA55h on
> > floppies and floppy USB images is a industry-wide BIOS change, or just a
> > BIOS bug for this machine. I'd like to know. If you know the spec.,
> > could you mention it?
>
> This was historically true for the floppy diskette. The signature was
> checked for fixed disks and option ROMs only.
> Some BIOS (maker and version) interpret the USB booting as HDD and require
> the signature.
>
That's interesting... So, they detect a floppy USB image as harddisk and
therefore needs a signature, but does not also need a partition? Can they
boot from USB with large harddisk images *without* partitions? I know my PC
detects floppy images and harddisk images on USB stick, and harddisk images
on USB harddrive. They need the signature, but not the partition.
Strange... There's got to be spec. out there somewhere on this.
Rod Pemberton
Rod Pemberton
> "Rod Pemberton" wrote in message news:iillte$8uj$1...@speranza.aioe.org...
> > "Mike Gonta" wrote in message
> > news:iikck4$7e7$1...@speranza.aioe.org...
>
> http://reboot.pro
>
Thanks, one of the comments in the bazillion in their forum (ahem...) leads
to a website for a program called RMPrepUSB.
It seems the RMPrepUSB.pdf has a section that explains USB booting. But, it
doesn't say where or how they located the information. I.e., no mention of
an industry specification.
http://sites.google.com/site/rmprepusb/documents/rmprepusb-beta-versions/RMPrepUSB.pdf
They are using slightly different terminology...
For those 223 BIOS manufacturers, do they list critical BIOS info? Like
extended Int13h? USB booting? 7c0h "bug"? etc.
What I predominantly see is BIOS IDs and links to manufacturers pages, and
some trivial FAQs.
Rod Pemberton
not seeing it...
Mike Gonta
> "Mike Gonta" wrote in message
> news:iim2rt$73c$1...@speranza.aioe.org...
>> "Rod Pemberton" wrote in message news:iillte$8uj$1...@speranza.aioe.org...
>> > "Mike Gonta" wrote in message
>> > news:iikck4$7e7$1...@speranza.aioe.org...
>> >> Time line, description and simple test code:
>>
>> Note: the code can be assembled in FASM or NASM.
>>
> ...
>
>> I've currently tested 10 PC's with BIOS dates between then and now.
>>
>
> Uh, why does your test work - to confirm USB booting? It seems to just
> use
> extended read... ? Is that your "gauge" for yes or no on USB booting?
Three things:
1. An MBR table with a minimal single partition entry (the actual partition
volume need not exist - the sector at LBA 63 should be zeroed out for
the test).
2. Use the int 13h extended functions.
3. The boot sector 0AA55h signature at 510h.
> Will the test confirm if used on an actual floppy? Or, only if used from
> USB device?
Only USB flash drive.
>> > "The boot sector 0AA55h signature at 510h"
>> >
>> > Is this always required? For floppies it wasn't, per the IBM tech.
>> > ref.
>> > man. Until recently, I never found a PC that require it for floppies.
>> > However, this new PC requires it on floppies and USB floppy images
>> > also.
>> > It's the first I'm aware of. I've not been able to locate if 0AA55h on
>> > floppies and floppy USB images is a industry-wide BIOS change, or just
>> > a
>> > BIOS bug for this machine. I'd like to know. If you know the spec.,
>> > could you mention it?
>>
>> This was historically true for the floppy diskette. The signature was
>> checked for fixed disks and option ROMs only.
>> Some BIOS (maker and version) interpret the USB booting as HDD and
>> require
>> the signature.
>>
>
> That's interesting... So, they detect a floppy USB image as harddisk and
> therefore needs a signature, but does not also need a partition?
My testing indicates that the BIOS does not look past the boot sector (LBA
0).
Other than a couple of sanity checks and for certain BIOS maker/version the
signature requirement - everything boots.
http://board.flatassembler.net/topic.php?t=12389
This means that the BIOS loads the boot sector to absolute address 7C00h
and jumps to it, the success or failure of that code depends on that code.
> Can they boot from USB with large harddisk images *without* partitions?
> I know my PC detects floppy images and harddisk images on USB stick,
> and harddisk images > on USB harddrive. They need the signature,
> but not the partition. Strange... There's got to be spec. out there
> somewhere on this.
The regular USB flash drive reports itself as a mass storage device,
a USB floppy drive, USB hard drive, USB CD drive, etc., reports itself as
that physical device.