19 replies to this topic

[noel]

Hello,
I make a bootable usb pen for winpe3. It works fine on many differents computers. But not on nec VL 280 (not so old). Some computers (hp 6005, a new one) doens't boot on USB pen even Setup Bios is ok (or seems ok). The V280 see the usb pen but mbr display "no system" (or like this, message is at office, tomorrow i'll see who display it).
Some bios speak about usb_zip or usb_hdd.

My questions :
Is exist a "good tutorial" which explain the "art" of inter-action BIOS / boot USB?

And what are différences between a usb pen of 16go and an extern disk usb (ide) of 80go when they are NTFS?
I think they have the same logical structure. Perhaps some thing change in the usb part. I compare with usbview.exe of DDK (W2K3) but i don't see any difference. And one or more differences exist really...for vl280 for sample.

But boot usb has many "mysteres".
Thanks

PS : I know that 80h is the number for int13h that identify the first disk, and 80h is required for an active partition, data and code of MBR/BR, boot NTFS, bootmgr, BCD, bootsect, diskpart, i display MBR/BR with diskprobe of ResourceKit of NT4 (i began with dos 3 and debug to explore hd)...

And i write like a french man. So, don't be worry and scratch this post if it's too difficult to understand me.

[steve6375]

Download RMPrepUSB v2 and click on Help and then OK. All will be explained (in Queens English)!

[was_jaclaz]

There is no full fledged "tutorial", AFAIK

But you may want to read the FAQ #10 on my old page:
http://home.graffiti...SB/USBfaqs.html

Then you might want to read these sometimes seemingly unrelated threads:
http://www.911cd.net...showtopic=21049
http://www.911cd.net...showtopic=21702
http://www.911cd.net...showtopic=23408
http://www.boot-land...?showtopic=9776

And these ones:
http://www.boot-land...?showtopic=7932
http://www.boot-land...?showtopic=7739

If you are still OK after reading those , you should have a fair enough knowledge of the problems that may be involved.
Basically:

• CHS geometry <-which is not the same on "any" BIOS and may be needed to be changed on particular machines
• BIOS code <- that may prove to be a non-resolvable problem, but for which there are workarounds
• MBR code <- actually rare/unlikely
• Bootsector CODE <- "wrong" for the MS FAT32 and NTFS filesystems, but patchable

Once you have this clear, you should be able to understand the approaches of the newish utilities, like RMPREPUSB and FBINST.

In the worst of the cases, you may need to take this kind of approach:
http://www.boot-land...?...ic=7138&hl=
(on-the-fly patching)
Don't even think to read the last linked to thread before having read and digested the other ones.



jaclaz

[vmbray]

I wonder if what I have been doing the last few days would work for you. It is pretty easy to try...

http://www.boot-land...showtopic=10503

If you're not using grub4dos don't be intimidated by it, once you have it going (and there are lots of great examples to help) it is the best thing right after booting on usb

[Wonko the Sane]

@vmbray

The problem at hand is NOT about which loader to use, it is about the partitioning/formatting of the stick.

Wonko

[vmbray]

Hello! I get it, did you read my thread? It is a combination of formatting and grub4dos that has proven to work great for me.

[Wonko the Sane]

Hello! I get it, did you read my thread?

Sure I have.

that has proven to work great for me.

Exactly.

Since I have been using the method you described (as well as most of the other ones) long before you even began thinking it could exist , I feel allowed to tell you that that is one of the numerous possible ways, and it is NOT - by any means - "universal", on the contrary, it is one of the most "basic" ones that will not work on a number of "particular" PC's, including the one the OP mentioned.

Maybe something you missed:
http://www.boot-land...?...c=10068&hl=



Wonko

[vmbray]

it is one of the most "basic" ones that will not work on a number of "particular" PC's, including the one the OP mentioned.

Okay, I was just trying to help and also see what boxes the method I am using will not work on. Is there a way to get the OPs box to boot from usb?

[Wonko the Sane]

As you can see from the links in the previous post, once set aside a thankfully very large base of "good" BIOSes, the rest can have "strange" behaviours.

The site on graffiti.net is no more, here is the copy:
http://jaclaz.alterv...SB/USBfaqs.html
To the list in FAQ #10 you can add:

• PC's that boot from "back" USB ports, but not from "front" ones
• PC's that only boot sticks up to 512 Mb in size
• PC's that only boot from "actual" floppy, meaning a device with a capacity of 1,474,560 bytes, 2/18/80

and the list is not exhaustive.

Over the years a number of utilities have been developed to try to mitigate this problem by making it more "probable" to "boot everywhere", but in some cases is simply not possible.

You already know the currently recommended utilities:
http://www.boot-land...?showtopic=9460

If you read the related threads, you will find a "history" of most common problmes and how the found workarounds work.

In such "problematic" PC's, and of course IF it works on them, and IF it is possible to install it, the best option is usually using PLoP:
http://www.plop.at/en/bootmanager.html
thus by-passing the "strange" BIOS, but again itwon't work "everywhere".



Wonko

[vmbray]

Like a premonition, I immediately after this conversation found that my laptop and my partners both will not boot the new stick using ntldr. How could I think that I had a way that solved all those issues? Well I didn't really but it does work on every box motherboard I have tried. So anyway I set up another stick with grub4dos mbr and it works on at least one of the laptops. So I'm back to two sticks, but I have a feeling that the combination of these two will be more universal than the last two I was using (one hp util, one make_usb_2 both with grub4dos mbr) but we'll see

Thanks for your comments, PLOP is interesting, will have to try it soon!

[Wonko the Sane]

Of the available utilities fbinst is in theory the most compatible with everything, but, on the other hand it is un-standard and thus some care must be used when using other utilities to manage/format partitions after fbinst did it's partitioning.

Wonko

[noel]

hello,
an other time, sorry for my poor english.

I trie about 30 PC (notebook, desktop), 5 years old and newest, a dozen of USB stick, a dozen of external disk UsB.

External Disk USB always boot in these PC, only a few of stick boot. Some PB boots with sticks less 4Go but don't boot with the stick usb of 16Go.

With plob on a stick usb, stick UsB boot very quickly (it's very good when Pc don't propose parameters for different vitesses : V1, V2 higth, V2 full). But not work for all of these PC. Plob on CD works fine but "not very friendly".

It's the "jungle".

Is USB V3 make something to normalize boot stick USB?

also an other question, but perhaps i must cut in two posts....

I read in a forum that stick usb use SCSI commands. And during format time, it's possible to reserve some space "unwritable". I found that when i was looking for a stick i can't format (XP says "write protected"). I used the program ( on a pc without hard disk because i was afraid by virus) :
http://www.apacer.co...ir_v2.9.1.1.zip
The link is ok to day but virus or not, i don't kown.
It works for me and after that i can format the stick under xp.

So, my question : is it possible to use this special space (write protected), mask and unmask data without a "proprietary" mecanisme of encryption protection (which use, i think, a driver loaded in the Os) ?
Could someone explain me more than these scsi command, these reserved space, and eventually how to use it?

Thank.

Ps : if my english is too difficult to understand, don't be worry and forget me.

[steve6375]

Try my program RMPrepUSB and read the Help pdf that comes with it.

I do not understand what you mean by write protected space.

USB sticks appear as Fixed Disks, these can be accessed via Windows APIs in various ways and some Windows SCSI APIs do work but get translated to USB hard disk commands.

As you have found, the way a USB drive is formatted and the way a BIOS treats these USB drives can cause different behaviours. Please read the RMPrepUSB PDF for an explanation and ask any Questions here.

[noel]

hello,
I know and use RMPrepUsb and i read the pdf more than one time. Thank for this program.

Well, i look in the "hard", in the network usb, in the commands on the frame usb, the scsi commands used to read/write/format the stick. Not the ones used by the API win32. But theses whitch are "under the hood".

The source of my informations :

- about the technique of "lockable storage"
http://www.usb.org/d...s_Agreement.zip
...Today there is no standard infrastructure in USB to control access to USB storage devices.
Loss and theft of these devices concerns both consumers and corporations. Consumers want to avoid accidentally sharing personal information. Corporations want to avoid accidentally sharing trade secrets and other confidential data.
This feature specification extends the USB Mass Storage class to allow hosts and devices to lock and unlock storage, without breaking legacy software behavior that exists today and without resorting to vendor-specific descriptors and requests....

- about scsi above usb (on the network usb)
http://www.usb.org/d...cs/uasp_1_0.zip
....UASP (USB Attached SCSI Protocol) will improve efficiency in [USB2]....

- for the commande "format" and the "defect list" :
http://www.usb.org/d...bmass-ufi10.pdf
it speaks about the command "format". It also introduce differents layers, the UFI... I think the floppy implementation is "conceptually" identical as hard disk implemenation in the hardware stick.

Too many texts en english for me.

The "defect list" technique can't show/hide datas in stick because i think it's need format and so erase datas.
The "lockable storage" technique is built for "that" but i don't know to use it.

My question is perhaps too generic : how to use this "lockable storage" technique?

Ps : many differences between API scsi (library depreciated?) but scsi commands send in the embeded frame on the network usb. And API windows like createfile or formatEx mask the real layers of protocols, like smb, socket or namedpipe.

Perhaps formatting with the win32api formatEx is not really the same thing that formatting with building a frame to put on the usb network.

Perhaps one day, it comes a filter driver whitch display the different layers on the network usb, like whireshark can do for ethernet.

[steve6375]

From the first spec. this looks very similar to the current spec used by most modern hard drives and already incorporated into some BIOSes.

Basically, the drive will not respond to any read or write commands until you send it a password/passphrase. This password unlocks the drive and allows read and write commands. Some BIOSes allow you to set a hard disk password. Before the system will boot, the BIOS sends the password to the hard drive to unlock it.

If anyone removes the hard disk, then they cannot access any data on the hard disk unless they know the password (which the BIOS may 'encrypt' before sending it to the hard disk.

By setting a user password in the BIOS, no-one can access the PC or the hard disk.

[Wonko the Sane]

noel, right now you are mixing THREE different things:

• USB stick booting/bootability
• USB stick Controller commands
• USB protocols

They are NOT the same and they are often not much connected to each other.
Up to post #11 and first half of post #12 we were talking of point 1) above, and the cause of failing or "strange" bootability is generally poor implementation in the BIOS of the specific machines.
Second half of post #12 is about point 2) above, the utility you used is a (reduced version) of the "Manufacturer Tool" a "low-level" interface to the controller inside the USB stick, with which you set (or reset) certain data or settings in the USB stick controller firmware.
Post #14 is about point 3) above and is mostly "fluff", theoretical discussions that have not ended in a real standard implemented in the actual hardware we are currently using.

It is not clear which part of your posts are actual statements and which part are actual questions.

I won't discuss anythng related to point 3) as it is completely pointless from a practical standpoint.

If you have questions about points 1) and 2), please ask them, and I'll try my best to try and answer them.



Wonko

[noel]

Sorry Wonko.

In post 14, i give the sources that i read. I give also some lines that they contain because my english is too poor to explain that i know about the context.
I read many text of www.uvb.org. And, until today, i can't imagine they are not implemented in stick.

I think it's difficult to explain "dysfuntion" about USB without speaking about the different layers. It's like SMB, tcp and ethernet csma/cd. Otherwise it's only a constat.

I have tow questions ( but before i must explain the context because my english...)
- on the same recent (2009) PC :
boot is OK for : a stick USB 4Go, external disk USB
boot is NOK for a stick 16Go (this stick boot on an other PC)
All are prepared with the same maner, under winpe3/ntfs/diskpart (with only one partition).
i notice that under WMI, stick are "removable" and external disk USB are "fixed".

Q1 : is it a way to identify why the usb 16G dosen't boot on this PC?
Q2 : is an information coming from the layer "scsi" that identify fixe/removable or coming from MBR/BR? If you think "Yes", can i modify it to test the result?

For the "protected data", i must read and find document before asking you later.
I hope i don't "disturb" anyone. If so, say to me, i undestand that my english in an handicap.

thank you very much, stev6375. If a bios can send password to a stick, a program under winpe or other can do. I'm interesting to know how.

[Wonko the Sane]

Forget the SCSI layer: it is NOT implemented in current sticks, and however has NOTHING to do with the problem you are having, believe me.

Most USB sticks - please read as ALL of them - are set (actually the controller in them) as "Removable".
Most USB HD's and the HD cases you can find - please read as ALL of them - are set (actually the controller in them) as "Fixed".

The "by design" behaviour of windows on a device that is set as "Removable" is that ONLY one (the active Partition) is accessible/given a drive letter.

The "by design" behaviour of various BIOSes is UNKNOWN, every manufacturer uses his own ways, it is perfectly possible that:

• a size limit is forced in the BIOS for "Removable"
• a partition ID (say 06, FAT16 CHS) is needed by BIOS
• ...
• ...

(please do review the links I already gave you)

There has been reports of some BIOS that "like" having 2 partitions on the "Removable" stick in order to boot from one, hence the 2 PTNS in RMPREPUSB.

There has NEVER been reports of a stick that has been "forced", by "flipping the bit" to become "Fixed" failing to boot, just like there has NEVER been a report of an external USB HD (which is already by default "Fixed") that failed to boot - of course provided that the motherboard BIOS accepts "USB HD" as bootable device.

The solution for sticks is then to "flip the bit".
To do so you need the "Manufacturer Tool" (and nothing else, and nothing else will work).

Now to your questions:

Q1 : is it a way to identify why the usb 16G dosen't boot on this PC?

There is NO sense in identifying why that particular stick does not boot on that PC, the reason is because the particular BIOS of that PC "wants" something in order to boot a USB stick (see any item of the lists already given), so the why is already clear: non-standard PC BIOS, the problem is identifying which particular requirements the BIOS is looking for, but the problem is making it boot.
If this is the question, there are only TWO possible solutions:

• try partitioning/formatting it with one of the ways that have been reported to be more compatible
• find the "Manufacturer Tool" for the specific stick and "flip the bit"

On a probability scale, using another partitioning/formatting scheme has definitely more than 50% probabilities of being successful, let's say 70÷80%, while flipping the bit is very, very near to 100%.

Q2 : is an information coming from the layer "scsi" that identify fixe/removable or coming from MBR/BR? If you think "Yes", can i modify it to test the result?

We don't think , we KNOW there is NO way if not through the "Manufacturer's Tool", as each single chip maker has his own particular way to change the Firmware of the device, which is where the "Removable" bit is, I repeat there is NO standard layer/language/protocol/whatever to communicate with the proprietary device if not the particular language/code the manufcturer of the chip has implemented.


And now, allow me one of my grumpy comments :

I don't understand WHY you come here, ask advice about a problem, you are told to try using RMPREPUSB, and then you come back to whine about the fact that NOT using it your stick won't boot.

A quick review of the "common sense advice" attached to Rules (point #f) may be of interest to you:
http://www.boot-land...?showtopic=9101

Google translate gives a good enough translation:
http://translate.goo...t...l=&ie=UTF-8

So, and again my contribution in this thread is only aimed at trying to help you with the problem you are having , and not necessarily connected to the "philosophy" of layer/protocols and whatever , you are advised to:

• EASY -> try using RMPREPUSB or any other tool that allows for a "more compatible" partitioning/formatting scheme, you will need to experiment with different settings.
• DIFFICULT -> Get Chipgenius: http://www.boot-land...?showtopic=4661
  
• identify the chip used in that particular USB stick
• find the corresponding "Manufacturer Tool" (if available) - if you cannot find the tool post the info Chipgenius gives and I'll try and help you finding it.
• flip the "Removable" bit to "Fixed"

Now up to you if you want to risk trusting my advice and possibly have that d@mn thing booting or if you prefer to go on with theories and hypothesys.




Wonko

[noel]

Hello,
I do not want to be impertinent, or disruptive and am not too stupid.
I have no specific problem with a particular material. Just love and understand.
The formula "But then why ..." let me think I could get more information that I can not find anywhere else (since usb.org documents are not taken into account by the industry).
Sorry to have bothered you.
The site http://www.boot-land...cons/icon12.gif is very informative and I will continue to consult to avoid disturbing more.
Greeting.

[Wonko the Sane]

@noel

no problem whatever , it is simply that I like more to solve an actual problem than discussing the how and why of the phylosophy that led to them, as I see it is more productive to first solve the problem, and only then find alternate ways to solve it should the same problem happen again.




Wonko