Jump to content











Photo
- - - - -

About bootability of USB sticks


  • This topic is locked This topic is locked
60 replies to this topic

#1 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 10 April 2009 - 07:15 AM

SERVICE ADDITION by jaclaz:

This and the following posts up to #14 were splitted from this thread:
http://www.boot-land...?showtopic=7661






That is the main reason we have to "reformat" them with a special tool which actually adds the partition table and make flash drive look hard drive alike to BIOS so it can boot properly

Are you sure about your statement? (bold is mine)

In my opinion it is not really so, and then I disagree.
I assume that you refer to HP USB Disk Storage Format Tool when you say "a special tool".
So if you take a look here http://www.boot-land...?...ost&p=14607 you can discover the real reason (at least that we know until now) that make your UFD formatted by HP Format Tool quite bootable, but in my experience then it does not mean that the BIOS will see it as a Hard Disk: if your UFD (formatted by HP Format Tool) will boot, then it boots for the reason described in the above linked post and not because the BIOS sees it as a Hard Disk.

If you really want that your UFD is seen by the BIOS as a Hard Disk (reaching a real better bootability) then you have to partitioning your UFD with (at least) 2 partitions just as described here http://www.boot-land...?...ost&p=63162

This way (that's my linked tutorial) assures that the BIOS will really see your UFD as a Hard Disk reaching a high hardware compatibility UFD/BIOSes (please, read: a high bootability) and then your UFD will boot (reasonably) from the most machines out there, and so from more machines rather than when it is formatted by HP Format Tool. :)

#2 wimb

wimb

    Platinum Member

  • Developer
  • 2585 posts
  • Interests:Boot and Install from USB
  •  
    Netherlands

Posted 10 April 2009 - 07:25 AM

Actually the HP Format Tool Creates a single partition on the USB-stick and gives the stick a harddisk like configuration with MBR + partition table.

The BootSector of the created partition is at sector 63

In this way the USB-stick is seen by BIOS as harddisk.

Besides, as you indicated already, it might be that changing some bytes in the MBR as done by the HP Format Tool, might improve the bootability on various systems.

#3 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 10 April 2009 - 07:38 AM

Actually the HP Format Tool Creates a single partition on the USB-stick and gives the stick a harddisk like configuration with MBR + partition table.

The BootSector of the created partition is at sector 63

In this way the USB-stick is seen by BIOS as harddisk.

Besides, as you indicated already, it might be that changing some bytes in the MBR as done by the HP Format Tool, might improve the bootability on various systems.

I've just noticed you've edited your post adding something, so now I've entirely quoted your post in order to entirely reply. :)

Actually the HP Format Tool Creates a single partition on the USB-stick and gives the stick a harddisk like configuration with MBR + partition table.

May be it could be as you state, but then it does not mean that your UFD has achieved the highest grade of hardware compatibility (please, read: bootability).
As a matter of fact my UFDs on one of my machines (that refuses to boot ANYTHING is not as a Hard Disk) does NOT boot if formatted by HP Format Tool and instead it perfectly boots if it is partitioned in one of the two ways just as described in the tutorial in my signature.

It seems to me that the related knowledge still is in progress. :)

In this way the USB-stick is seen by BIOS as harddisk.

I'm not too sure about your statement: my experienced issues are different.
Just as above already replied, it seems to me that HP Format Tool make your UFD quite bootable for the reason explained by jaclaz here http://www.boot-land...?...ost&p=14607
but this does NOT mean that "the" BIOS (you say: "by BIOS"; then I ask: "which BIOS?") will see your UFD as a Hard Disk.

If you want that the most BIOSes out there see your UFD as a Hard Disk you have to perform the tutorial in my signature. :)

Besides, as you indicated already, it might be that changing some bytes in the MBR as done by the HP Format Tool, might improve the bootability on various systems.

Really I have not said that: I have said that HP Format Tool provides an average bootability and that if your UFD boots then it boots for the reason explained in the above linked post.

As already said, if you want to really increase your UFD bootability (that's the most BIOSes will see your UFD really as a Hard Disk) then you have to perform one of the two ways just described in the tutorial in my signature. :)

So - about the USB booting - the simple partitioning of your UFD seems to succeed EXACTLY as the flipping of removable-bit does and then the bootability of your UFD will result highly increased!



#4 wimb

wimb

    Platinum Member

  • Developer
  • 2585 posts
  • Interests:Boot and Install from USB
  •  
    Netherlands

Posted 10 April 2009 - 07:44 AM

It might quite well be that creating more than 1 partition improves the bootability.

BIOS will definitely define your USB-stick then as a Harddisk.

#5 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 10 April 2009 - 08:21 AM

It might quite well be that creating more than 1 partition improves the bootability.

BIOS will definitely define your USB-stick then as a Harddisk.

Just as I have already discovered and described in my tutorial here http://www.boot-land...?...ost&p=63162 :)

#6 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 10 April 2009 - 01:43 PM

@wimb
@online

If I may intervene in your little discussion, both of you have some "right" points, but also a few "misconceptions", briefly and in five points:
1) If first sector of the device contains a partition table, and the partition table contains a single valid entry, Windows will see it as a HD-like device. (please note that it is not necessary to have bootable MBR CODE in first sector, ONLY one valid partition entry and the magic bytes 55AA at the end of sector)
2) No need for a filter driver for a single partition.
3) Windows NT by default adopts a geometry of nx64x32
4) Windows 2K/XP/2003/Vista/2008/7 by default adopts a geometry of nx255x63
5) Latest Windows however use LBA access only, and, due to current size of sticks, it is rare to use anyway CHS filesystems, but some BIOS do use CHS, that is WHY the HP utility may produce an unbootable on some systems device.

Read this post:
http://www.boot-land...?...=3191&st=33

jaclaz

#7 Nour El2mar

Nour El2mar
  • Members
  • 7 posts
  •  
    Egypt

Posted 10 April 2009 - 07:04 PM

OMG jaclaz you speak right always :)
i hope you tell me the errors in book to modify them :)

#8 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 10 April 2009 - 09:21 PM

@jaclaz

Please, let me specify/explain/confront some my "conception" points just related to your posted ones.

1) If first sector of the device contains a partition table, and the partition table contains a single valid entry, Windows will see it as a HD-like device. (please note that it is not necessary to have bootable MBR CODE in first sector, ONLY one valid partition entry and the magic bytes 55AA at the end of sector)

Really, since the already mentioned procedure in order to increase my UFDs bootability is reliably working for me then I've not been more interested in HD-like feature/behaviour of my UFDs under Windows provided by HP Format Tool, because I've been more interested in my UFDs behaviour booting-BIOSes-related, and then in a better and wider compatibility about the best (that's also the simpler and the less tweaked possible) way for me in order to reach this goal.
My experienced results about the initial and una tantum partitioning of my UFDs with XP's Disk Management (> "Format-it-&-Forget-it" way) have been very effective for me and then until now I have been able to boot my UFDs on every machine I've tried, but not the same thing occurs with HP Format Tool and/or with a MBR created in a different way.
And yes, the MBR of my UFDs created with the above mentioned way ("Format-it-&-Forget-it") just contains the "magic 55 AA bytes" at the end of the sector. :)

00 00 00 00 00 2C 44 63 6D C3 43 B9 00 00 80 01

01 00 06 FE FF CF 3F 00 00 00 91 3F EF 00 00 00

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

00 00 00 00 00 00 00 00 00 00 00 00 00 00 55 AA
Furthermore performing my some tests about MBR I've noticed that replacing the MBR of my UFD is a practice working related to the partitioning and it is unfortunely NOT working about the pure booting. http://www.boot-land...?...ost&p=62936

2) No need for a filter driver for a single partition.

Yes, I know that: performing "Format-it-&-Forget-it" way then my UFD boots and runs Full XP from the working partition just as a Removable device and without any filter driver (under Windows it is as a removable device, but from BIOS-side my UFD is seen as a Hard Disk just because it is partitioned).

Full XP in Single Booting, running and booting without any boot loader except its own "ntldr" and then without any filter driver from a partitioned UFD (as a Removable device under booted Windows) obtained performing "Format-it-&-Forget-it" way.
Posted Image

So please note that the above result just occurs even if my UFD is partitioned, that's with my UFD that have 2 partitions: where the smaller and last partition is not accessible (with no price to pay) "by design". :)

3) Windows NT by default adopts a geometry of nx64x32
4) Windows 2K/XP/2003/Vista/2008/7 by default adopts a geometry of nx255x63

As always your replies are very instructive and appreciated. :)

5) Latest Windows however use LBA access only, and, due to current size of sticks, it is rare to use anyway CHS filesystems, but some BIOS do use CHS, that is WHY the HP utility may produce an unbootable on some systems device.

I've already mentioned your different post with similar conclusions http://www.boot-land...?...ost&p=14607 about HP Format Tool in my previous reply and may be that it is just so because that way the MBR create on my UFD is more compatible at booting rather than that provided by HP Format Tool.
Maybe a further clarification is needed: I'm talking about the "simple" booting of my UFD, not about the booting of a system device (for "systems device" I assume that you mean a - USB removable - device with installed NT-based operating systems) that is of course secondary to the previous "simple" booting success.
In other words: if I format my UFD with HP Format Tool then even grub4dos does NOT boot from my UFD under certain BIOSes/machines.
"Format-it-&-Forget-it" way (that's the simple double partitioning of my UFD by XP's Disk Management) instead assures the best bootability for me either when it is "simply" related to grub4dos and when it is a Full XP to boot and to run even without any boot loader rather than its own "ntldr".
In my experience the MBR created by XP's Disk Management console on a UFD seen as a Basic Disk under Windows and just partitioned with (at least) 2 partitions is much more compatible rather than the MBR created by HP Format Tool independently if a filter driver is used or not under the just booted Windows from your UFD.
That is a reliably working (and a simple) way, much more compatible for me rather than HP Tool formatting: as matter of fact that way works where HP Format Tool works and it also works where HP Format Tool does NOT work.
About this evidence, and not only, your kind and appreciated opinion is as always more than welcome. :)


i just want to say that im glad that i saw the the final release of grub4dos 0.4.4
but if some users here have tried it i hope that you tell me about its stability

Hi, Nour El2mar!
Generally it is always the latest release of "grub4dos" that would be used, although I've not yet tried it I think it is stable at least "enough". :)
However this thread http://www.boot-land...?...ndpost&p=30 could be hopefully useful. :(

#9 pscEx

pscEx

    Platinum Member

  • Team Reboot
  • 12707 posts
  • Location:Korschenbroich, Germany
  • Interests:What somebody else cannot do.
  •  
    European Union

Posted 10 April 2009 - 09:32 PM

Sorry, I'm really a dummie in this area.

Therefore my question:
Does the current discussion make something in the original post obsolete?
If yes, what?

Peter

#10 wimb

wimb

    Platinum Member

  • Developer
  • 2585 posts
  • Interests:Boot and Install from USB
  •  
    Netherlands

Posted 11 April 2009 - 08:52 AM

@jaclaz
Avoiding HP Format Tool by making MBR + HP patch + (Active) partition table entry manually, and then disconnect - reconnect and format USB-stick with Windows XP disk management, is certainly a possibility.
NTFS Format for Removable Drives requires additionally to Set:
Change USB-stick Drive Properties in "optimize for performance"
I have used it and it works http://www.911cd.net...o...20089&st=24

But would it make any change in bootability of the USB-stick ?
May be, because a well balanced partiton table entry can be made.
It is quite a procedure as compared to using HP Format Tool and perhaps difficult to program.

@online
You observed improved bootability of USB-sticks, when a (small unused) second partition is created on the USB-stick.
This is certainly very interesting and valueable.
I believe this will really make an important change, so that BIOS will more likely see USB-stick as Harddisk, and can explain your success.

Now my question is did you find any influence of the type of bootloader on bootability of USB-sticks.
It could be e.g. that NTLDR bootloader with boot.ini Menu has better bootability then e.g. GRUB4DOS installed in MBR.

Can you say something about this point.

It will be better to move the discussion on bootability of USB-sticks to the thread of online on this subject.
http://www.boot-land...?showtopic=7512

#11 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 11 April 2009 - 11:16 PM

@wimb

First of all: many thanks for your very kind reply. :)

@online
You observed improved bootability of USB-sticks, when a (small unused) second partition is created on the USB-stick.

Yes, I thought about a way in order to increase the bootability of any UFD (do not forget that about the USB-booting every UFD is absolutely the same, because it is in the different BIOSes matched behaviours the real matter) and performing some tests the double partitioning has proved to me to be the best way. :)
So you really can highly increase your UFD bootability when at least 2 partitions are created (by XP's Disk Management) on your UFD.
This is a necessary and sufficient condition: however performing one of the two ways described in the tutorial in my signature you will can do even more (up to 4 primary partitions and up to 3 Multiple Full XP in Dual Booting and/or in Multi Booting).
Furthermore if you create the 1st partition in FAT file-system ("Format-it-&-Boot-it" way) then your UFD bootability would be even further increased (since all reports refer about a related excellence of that condition).
However I've not noticed relevant differences about that, but please note that I have not too many machines around me rather than the possible reports of (many) other users about this experience: then they are more than welcome also and just for further knowledge experienced on a more large-scale.

This is certainly very interesting and valueable.

Many thanks for your appreciation!
Yes, I also think it is so very interesting that I wanted just to share my tutorial (that's my thoughts and my experiences). :)

I believe this will really make an important change, so that BIOS will more likely see USB-stick as Harddisk, and can explain your success.

Yes, it is exactly what I mean and what I describe in my tutorial: many thanks (again) for your appreciation and then for having noticed (the real and) the potential implications of the related practice (that's about the double partitioning of UFDs). :)

Now my question is did you find any influence of the type of bootloader on bootability of USB-sticks.
It could be e.g. that NTLDR bootloader with boot.ini Menu has better bootability then e.g. GRUB4DOS installed in MBR.

Can you say something about this point.

Frankly, in my experience it seems to me almost the same: because the point seems not in the boot loader, in my observations the fundamental point seems in the behaviour of (the most) BIOSes matching any double partitioned UFD just a bit before the following effective booting (I do not know to describe it better, but I'm sure that it is just so): if it so partitioned boots (that's if it becomes really bootable), and it boots (that's it really becomes bootable), then the importance of the kind of boot loader seems to become a minor matter (just as it is for a Hard Disk): at that point "every" boot loader should work the same. :(
Performing "Format-it-&-Forget-it" way your UFD will be seen from the most BIOSes just like a Hard Disk and then you can boot your UFD from (reasonably) the most machines out there indifferently with "grub4dos" or with "ntldr" as boot loader and also in this case I've not noticed great differences in the hardware compatibility (please, read: bootability).
Sure if you want to boot and to run Multiple Full XP from the 2nd, the 3rd and/or the 4th partition of your UFD (just as described in my other tutorial here http://www.boot-land...?...ost&p=62682 ) then a boot loader as "grub4dos" appears as needed: from the 1st partition of your double partitioned UFD, instead, Full XP always boots and runs with-or-without an external boot loader and with-or-without an "integrated" filter driver (please, take a look to the related screenshots in the related threads, both ones in my signature).

It will be better to move the discussion on bootability of USB-sticks to the thread of online on this subject.
http://www.boot-land...?showtopic=7512

I agree. :)

#12 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 12 April 2009 - 03:44 PM

I will try saying this once again only :) :
There is NO DIFFERENCE whatsoever if there is one, two or 23 partitions on the stick.

The HP tool partitions IMPROPERLY the device, creating UNBALANCED CHS and LBA data and not respecting Cylinder boundaries.
This IMPROPER partitioning works nonetheless on most (but not all machines).

By using MBRbatch, or manual calculations and later Windows FORMAT makes the device PROPERLY partitioned/formatted.

Windows 2K/XP/2003 Disk Management (on a device that has the removable bit flipped or using the filter driver) partitions and formats it PROPERLY.

The Hp hack to the standard 2K/XP MBR may improve compatibility on sticks partitioned/formatted by standard windows tools:
http://www.boot-land...?...ic=7468&hl=

@wimb
sure, life its tough, and using MBRbatch is more difficult than using the HP tool.
Using an "intermediate" image might be easier:
http://www.boot-land...?...c=5000&st=1
Simply writing a "proper" app replicating what MBRbatch clumsily does would be the solution, but noone is apparently interested in it.

@online
There is no voodoo :) nor "cabala" linked to "two or more partitions", it is simply the method you are using (as opposed to the HP tool) that produces proper partitioning.

jaclaz

#13 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 12 April 2009 - 11:01 PM

@jaclaz

Ok, normally a reply of kind your above post would be enough for me in order to stop any further discussion, any further participation to the forum(s) and any further ideas/experiences-sharing.
However, at least in this case, I will continue to reply because I respect you.

I will try saying this once again only :)

Take it easy, man: remember rule n. 12
http://www.boot-land...?act=boardrules
12. SMILE! Life is tough, we all know that, when you enter this board, it will be appreciated that you leave your personal problems behind, asking and replyinq questions or however exchanging informations with a "positive" attitude, TAKE IT EASY and enjoy your stay here.
http://www.boot-land...?...ost&p=42769

There is NO DIFFERENCE whatsoever if there is one, two or 23 partitions on the stick.

You are wrong: the difference exists.

And one single partition on your UFD is NOT enough for widest booting.
If you want that your UFD is seen as a Hard Disk from your BIOS (please, read: from the most BIOSes) then you have to partition it with at least 2 (two) partitions (if the smaller and last one is not accessible under Windows, it is not matter).
The double-partitioning of your UFD (that's MY idea) succeeds EXACTLY as the flipping of removable-bit does on your UFD from BIOS-side (that's your UFD is seen as a Hard Disk).

My tests on an ASUS high-level motherboard (almost two years old): you could find a similar motherboard that refuses to boot any UFD that is not as a Hard Disk, then you could perform MY procedure partitioning with only one partition and then take a look if it boots; if it does not boot (and in that case just it does not), then you have to create 2 partitions and then take a look if it boots: and it will boot.
The end.

Just for the record, from your post here http://www.boot-land...?...ost&p=62876

The only way to prove (or disprove) that it's "better" than "standard" MBR will be when you find a PC on which the "standard" one does not work and you try with the "patched version" and it works (or viceversa).


The HP tool partitions IMPROPERLY the device, creating UNBALANCED CHS and LBA data and not respecting Cylinder boundaries.
This IMPROPER partitioning works nonetheless on most (but not all machines).

It is not matter for me, and most of all it is not working for me.
Instead, MY procedure (> double partitioning on your UFD) works.
And it works reasonably from really the most machines out there.

By using MBRbatch, or manual calculations and later Windows FORMAT makes the device PROPERLY partitioned/formatted.

I'm sorry for YOUR batch and/or for other MBR "replacements", but then it does NOT boot for me.
Until now, on certain machines, my UFDs boot only if they are previously and really partitioned/formatted under Windows by XP's Disk Management (after that you can restore any image backup on your UFD and then it will boot).

Windows 2K/XP/2003 Disk Management (on a device that has the removable bit flipped or using the filter driver) partitions and formats it PROPERLY.

This is EXACTLY what I am repeating and describing in my tutorial: but I describe about the (at least) double partitioning (only XP's Disk Management seems to create a widest compatible MBR), and are NOT you (or anyone) to have suggested to me anything about the double partitioning as a way to reach the highest bootability, don't you?

The Hp hack to the standard 2K/XP MBR may improve compatibility on sticks partitioned/formatted by standard windows tools:
http://www.boot-land...?...ic=7468&hl=

You are talking about a SINGLE partition on your UFD.
If you double-partition your UFD then the HP-patch is not longer needed: the reached bootability of your UFD is already the best one possible.

sure, life its tough, and using MBRbatch is more difficult than using the HP tool.
Using an "intermediate" image might be easier:
http://www.boot-land...?...c=5000&st=1
Simply writing a "proper" app replicating what MBRbatch clumsily does would be the solution, but noone is apparently interested in it.

I think that this way (WHEN working) is at least too complicated, instead MY procedure is "quite" a standard procedure, most of all a simple and quick-booting procedure, with excellent, effective and reproducible results.

@online
There is no voodoo nor "cabala"

Your words, not mine.
Excuse me, but how do you allow?

Furthermore, if the above is said by someone who usually tells of "magic" about USB booting it seems to me rather contradictory, don't you? (following bold is mine)

FAQ #10
Question:
I tried everything you suggested, but I still cannot boot from the USB stick, where am I wrong?
Short Answer:
It is quite possible that you did eveything right but did not try hard enough, on many motherboards USB booting seems to be a procedure involving some magic

I hate to say so, but it still seems like a certain amount of black (or white) "magic" is needed when booting from USB.....

http://www.911cd.net...o...st&p=151046

booting from USB involves in some cases a form of "magic", so do not give up if at first try you do not succeed.

http://www.911cd.net...o...st&p=140797

it is simply the method you are using (as opposed to the HP tool) that produces proper partitioning

And that seems little to you?

Furthermore, please note that the double-partitioning of your UFD in order to reach the highest bootability is a MY idea, a MY thought, a MY experience (that I wanted to share).

Anyway, if you are so sure that MY procedure - that's the at least double partitioning - is working why did you never suggest that to anyone before now?
And why anyone has thought, achieved and/or reported any similar way before?


Finally, let me say that it is ALWAYS so: until we know a new knowledge we still MUST stay on the old one.

In my opinion, if the knowledge (ALWAYS limited, until the NEXT step is reached) makes ANYONE less farsighted, I would say to him to (please) read at least a little (but great) piece of Einstein in my signature.
As a classical and relevant example your post here (before to know something more about HP Format Tool)
http://www.boot-land...?...ost&p=14607


Well, now you can also BAN me.

Thank you so very much for your kind attention.


Btw: could you kindly split the related posts existing in this innocent thread to the most related one?
http://www.boot-land...?showtopic=7512
Many thanks in advance.





:)

#14 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 13 April 2009 - 09:01 AM

@online

Let's see if I can convince you how we are saying almost the same thing. :)

You say that in order to have better bootability two conditions must be met for MOST PC's:
  • Xp disk management is used for the partitioning
  • the stick is divided in more than 1 partition

I say that in order to have better bootability one condition must be met for MOST PC's:
  • the stick needs to be properly partitioned (and Windows XP Disk management properly partitions it)

The difference seems to lie with the definition of "MOST".
"My" MOST includes several tens of "couples" sticks/motherboards, with experiments carried by myself or by members of the 911CD or boot-land boards.
"Your" MOST includes your experience, which, no offence whatsoever intended :), may be a little smaller than mine.

Now the questions: :)

  • Have you ever used XP disk management to partition a USB stick with a single partition?
  • Have you ever seen such a stick fail booting?
  • Have you ever seen the same stick, repartitioned with XP disk management into two partitions suddenly start booting on the same machine?

Let's assume that all your answers are YES. :(

More questions that need an answer:

  • Is the fact because there are actually two partitions on the stick (both accessible to windows if a filter driver is installed)?
  • Or is it because there are two entries in the partition table ( the second being a "fake one")?
  • Or is it because the BIOS on that machine can only access a limited in size first partition? (as an example The 1,024 Cylinder (504 MiB / 528 MB) Barrier)

I personally never found a single machine which I wasn't able to boot from with a stick partitioned into a single partition, and I was always able to "fix" sticks that didn't want to boot from users here on on 911CD forums, without the need to add a second partition, but it is well possible that a particular machine needs that, I trust your word for it :).

In which case, I would call it an exception, and not a requirement. <- and this is the point where our opinions diverge

As you very well know, it's almost 15 years :) I go around advising that ANY HD like device should be partitioned in a small FAT16 partition and one Extended with one or more Logical volumes inside it, so I do approve of the method you recommend :) , and I did so even before you even thought about it :(, I simply do not find it a requirement.

To sum up I would rate, from my experience, the levels of compatibility in booting as follows:
  • 100,00% flipping removable bit
  • 90,01% Properly partitioning AND two (or more) partitions (maybe)
  • 90,00% Properly partitioning (which includes XP Disk Management, PeTOUSB - up to 2 or 4 Gb - MBRbatch and manually creating a partition table)
  • 80,00% HP format tool


About the HP MBR hack it is completely unrelated to one or more paritions, as clearly stated it simply skips certain BIOS calls, it is as well NOT a requirement, but rather an exception.

:(

jaclaz

#15 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 13 April 2009 - 09:50 AM

@jaclaz

Please, since a thread with the right related topic is already existing and since I would not wish another thread about the same argument, could you very kindly split this entire thread (just created by you) in my most related and already existing topic?

Make Your Most Bootable USB Flash Drive



here


http://www.boot-land.net/forums/index.php?showtopic=7512




Thanks in advance.



Btw: after your very kind splitting in the most proper and already existing thread I will be pleased to reply to your last post above.
Thanks again in advance.

#16 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 13 April 2009 - 11:42 AM

@jaclaz

Please, since a thread with the right related topic is already existing and since I would not wish another thread about the same argument, could you very kindly split this entire thread (just created by you) in my most related and already existing topic?

Make Your Most Bootable USB Flash Drive



here


http://www.boot-land.net/forums/index.php?showtopic=7512




Thanks in advance.



Btw: after your very kind splitting in the most proper and already existing thread I will be pleased to reply to your last post above.
Thanks again in advance.


No, thanks. :(

Notwithstanding your very kind request :), I won't start administering the board along requests of users, if I find them not appropriate, expecially if they make conditional or vaguely reminding blackmail statements :).

If you want to contribute your ideas or opinions to this thread, you can do it here.

If you don't want to, you are perfectly free not to post.

If you want to post on the other thread, you are free to do so.

If you want to start a new thread, you are free to do so.

As well, should you decide to post anywhere on the board, I will feel free to move or edit your posts whenever and wherever I see it appropriate:
http://www.boot-land...?act=boardrules

8. Has the thread you started or participated in been moved, closed or deleted? Check with the forum moderator via PM. Please do not start new threads asking why other threads were moderated. Forum moderation is not up for public debate, such threads will be deleted and the thread starter warned. The Admin/Mods of this board reserve the right to edit, delete or move posts made on this site.


It is a reknown fact how I am a senseless dictator, and usually behave unfairly and absurdly, applying my supreme will and censorship to anything posted on the board:
http://www.boot-land...?...c=3124&st=0

You can of course contact Nuno and file a complaint about my behaviour. :)
I am sure if there are enough valid and motivated complaints Nuno will decide to fire me :), but until then I will continue administering the board my way. :)

:)

jaclaz

#17 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 13 April 2009 - 12:52 PM

@jaclaz

Really the further moving of already splitted posts does not seem so important (really it does change nothing), and no compliant seems needed. :)

Returning (if also you agree, of course :) ) on the usual (and appreciated) way in order to well communicate, I'm pleased to reply to your above post with all the same respect of all time. :)

@online

Let's see if I can convince you how we are saying almost the same thing.

You say that in order to have better bootability two conditions must be met for MOST PC's:

  • Xp disk management is used for the partitioning
  • the stick is divided in more than 1 partition

I say that in order to have better bootability one condition must be met for MOST PC's:
  • the stick needs to be properly partitioned (and Windows XP Disk management properly partitions it)

Yes, it is just so.

The difference is that I thought and proven (and shared) about a double partitioning - that's creating 2 (two) partitions on your UFD - in order to reach the best bootability of your UFD, and it seems to me that no one else before has ever reported about the double partitioning as the way in order to reach the highest bootability... :(

To sum up I would rate, from my experience, the levels of compatibility in booting as follows:

  • 100,00% flipping removable bit
  • 90,01% Properly partitioning AND two (or more) partitions (maybe)
  • 90,00% Properly partitioning (which includes XP Disk Management, PeTOUSB - up to 2 or 4 Gb - MBRbatch and manually creating a partition table)
  • 80,00% HP format tool

In my opinion, said that in the above manner it is not too clear that I thought and shared about the double partitioning of your UFD as the best way in order to reach the highest bootability from your UFD, and again it seems to me that no one else before has ever reported about the double partitioning as the way in order to reach the highest bootability... :)

Since nobody is able to explain how-&-why ALL or even certain behaviours are mathematical occurring, and since the USB booting still seems a dark area where "magic", "maybe", "trying", etc, are in almost every theory and in every procedure, let me explain how, performing my "Format-it-&-Forget-it" way (that's creating 2 partitions on your UFD) you will be able to reach a better (the highest) bootability rather than a single partitioning and/or the HP Format Tool usage, and/or other utilities/procedures usage.

Now the questions:

1. Have you ever used XP disk management to partition a USB stick with a single partition?
2. Have you ever seen such a stick fail booting?
3. Have you ever seen the same stick, repartitioned with XP disk management into two partitions suddenly start booting on the same machine?

Sometime I doubt that I wrote what I already wrote if you still ask to me about what I already said. :(
However, as already said, my tests on an ASUS high-level motherboard (almost two years old): you could find a similar motherboard that refuses to boot any UFD that is not as a Hard Disk, then you could perform MY procedure partitioning with only one partition and then take a look if it boots; if it does not boot (and in that case just it does not), then you have to create 2 partitions and then take a look if it boots: and it will boot.
And if you state that

The only way to prove (or disprove) that it's "better" than "standard" MBR will be when you find a PC on which the "standard" one does not work and you try with the "patched version" and it works (or viceversa).

so my above experienced issue is a reliable demonstration. :(

Just for the record I repeat that the filter driver is not matter about the bootability and furthermore it seems that different boot loaders are no matter the same (but in these recent posts we are not interested in that and we are not talking about that).
From "the point of view" of the most (all?) BIOSes, if you partition with at least 2 partitions your UFD then it will be seen as a Hard Disk with the best compatibility possible.

There is NO DIFFERENCE whatsoever if there is one, two or 23 partitions on the stick.

jaclaz, let me say one more time and let me prove that instead the difference exists. :)


It seems to me that I finally found where is the matter. :(

I've compared the MBRs of the same UFD where I created in different steps

a. 2 partitions with XP's Disk Management;

b. 1 partition with HP Format Tool;

c. 1 partition with XP's Disk Management.

My findings are in the related screenshots of the 3 different and compared MBRs:

Posted Image

a. When 2 partitions are created on your UFD (with Disk Management console) then the resulting MBR has its Signature with just 2 bytes; this fact achieves the highest bootability from your UFD.
55 AA
b. When HP Format Tool is used then the resulting MBR has its Signature with further 2 (two) bytes added; this fact achieves an average bootability from your UFD;
00 00 55 AA
c. When only one single partition is created on your UFD (with Disk Management console) then the resulting MBR has its Signature with further 12 (twelve) bytes added; this fact achieves a low bootability from your UFD.
00 00 00 00 00 00 00 00 00 00 00 00 55 AA
Furthermore, if you state that the flipping of the removable-bit is 100% working, and I state that my "Format-it-&-Forget-it" way (that's creating 2 partitions on your UFD) succeeds EXACTLY as the flipping of the removable-bit does, then my "Format-it-&-Forget-it" way (that's creating 2 partitions on your UFD) is 100% working.

As matter of fact my "Format-it-&-Forget-it" way works where other ways work, and my "Format-it-&-Forget-it" way works where other ways do NOT work.

And the reason seems :) to be all in those few bytes of difference in the created MBR of your UFD performing my "Format-it-&-Forget-it" way. :(



Thank you (all) for your patience. :(

#18 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 13 April 2009 - 01:28 PM

It seems to me that last line in ALL three MBR's you posted appear :) as:

00 00 00 00 00 00 00 00 00 00 00 00 00 00 55 AA

though you coloured them in a different way. :)

Of course blue zeroes are different from yellow zeroes, but BIOS (generally speaking), are colour blind or daltonic.
(I know in the line above I made them orange, but yellow doesn't display nicely. :))

Now 55AA is the required "Magic bytes" signature to make the sector a valid MBR or bootsector.

The previous 14 bytes are the last 14 bytes of fourth partition entry.

Will you do a personal favour to me? :)

Can you make a MBR with FOUR partition entries (three primaries and one Extended or four primaries)?

Would this somehow disprove your 00's signature theory? :)

jaclaz

#19 wimb

wimb

    Platinum Member

  • Developer
  • 2585 posts
  • Interests:Boot and Install from USB
  •  
    Netherlands

Posted 13 April 2009 - 02:21 PM

then you could perform MY procedure partitioning with only one partition and then take a look if it boots; if it does not boot (and in that case just it does not), then you have to create 2 partitions and then take a look if it boots: and it will boot.
And if you state that
so my above experienced issue is a reliable demonstration. :)

How about removing again the second partition (leaving it as unused space).
Then if it does not boot, you have a better proof,
since in making the second partition, you possibly also induced changes in the first partition :)
Removing the second partition does NOT change the first partition.

#20 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 13 April 2009 - 03:14 PM

since in making the second partition, you possibly also induced changes in the first partition :)
Removing the second partition does NOT change the first partition.


It seems to me like logic is gone on holiday for Easter. :)

I should have done the same, :) , I really made a mistake in taking part to this thread. :(

I will abandon this thread, only thing I can recommend is what I recommended in another instance:
http://www.boot-land...?...c=6554&st=3

But I know it will be completely useless.

I have now to admit total defeat. :)

Off to other attempts to put this:
http://en.wikipedia....ientific_method
into some better use. :)

:)

jaclaz

#21 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 13 April 2009 - 08:20 PM

It seems to me that last line in ALL three MBR's you posted appear :) as:

though you coloured them in a different way. :)

No, I do not.
If you think that about my presumed behaviour, then probably my reply will be useless.

However, the "strange" thing has been that "010 Editor", that's the program I've used in order to compare the 3 MBRs, it is itself that colored in blue the bytes that it found as different.
This fact made me wonder, because taking a simple look to the "55 AA" previous zeros they seem all identical.
But the program pointed (and automatically selected) them as difference bytes (coloring them in blue, tha's selecting them).
I do not know if my description is clear enough (it is a very simple thing, but maybe a bit "difficult" to describe), but (wanting it) you can try by yourself to save the different MBRs with HDHacker in 3 different "dat" files and then comparing them through 010 Editor (of course I think that other HEX editors should work almost the same, sure 010 Editor works so).
When 010 Editor indicates the related differences it will color in blue all the related bytes because it automatically selects them (that's the differences in bytes).
Still it seems to me a strange reason, and then I've imagined in an explanation of that kind.

Posted Image

About the real reason of the effective results of my "Format-it-&-Forget-it" way: if the future will show something of good about that procedure also for so many successful reports, as I think and believe it can show (the reports, not the laughter, are fundamental about the "evidences"), after those hopefully evidences/reports then probably something will be demonstrate also by the real experts.

Finally you (all) are the experts, and about your last words in your last post I think that it does not a nice thing for an expert to ridicule a member that was sharing his good results about UFD booting trying furthermore to add an explanation. :)

Please, do not forget that until some months ago the reason why HP Format Tool appeared to work from a larger amount of machines still was totally unknown and it was discovered (by you, the expert) following the observations of a unaware member. :)

http://www.boot-land...?...ost&p=14607


Btw: probably the USB booting sometime involves something of "magic" (as you like to say)... and about my "Format-it-&-Forget-it" way, although the real reason still seems unknown (it will not surely the FIRST case of this kind :( ), I just know that it works! :)

I hope in so many successful reports (for so many other users, and for me), but also without them, now I have all my UFDs working and booting from EVERY tested machine, and that does NOT occur using HP Format Tool and/or other procedures. :)

#22 maanu

maanu

    Gold Member

  • Advanced user
  • 1134 posts
  •  
    Pakistan

Posted 13 April 2009 - 10:35 PM

well , after reading all three pages , im still confuse :) .. im trying to learn all those thoughts in it ./

online , it is great to have a document like this and i should say research like this. it helps noobs like me who are new to learn this stuff because i believe everyone is STILL learning .

well on topic now ,

i was examining my usb drive (which was made portable with Tom's UBUSB 1.0.44) and has syslinux as base also grub menu too . now when i select my usb drive from 010 editor , it goes very long , and i dont where to look at it.

i have saved my mbr copy from hd hacker , will u plz look at how compatible it can be ? plz change the extension to .dat first

thanks...

Attached Files



#23 online

online

    Silver Member

  • Advanced user
  • 767 posts

Posted 14 April 2009 - 09:42 AM

How about removing again the second partition (leaving it as unused space).
Then if it does not boot, you have a better proof,
since in making the second partition, you possibly also induced changes in the first partition :)
Removing the second partition does NOT change the first partition.

Ok, I will repeat again already said things.

If I create one single partition on any my UFD with "all" known ways (Disk Management, HP Format Tool, PeToUSB, and "all" other ones tried by me in the past, since the gold-time of VistaPE) then on certain machines (not too oldish) my UFDs do NOT boot.

Instead, if I create at least 2 partitions (or more) with Disk Management on ANY my UFD (that's performing my "Format-it-&-Forget-it" way, in my signature), then EVERY my UFD perfectly boots from those certain machines too (expecially an ASUS motherboard not too oldish of high-price that systematically refuse to boot ANYTHING is not as a Hard Disk).

Furthermore if I create 2 partitions on my UFD and then I delete the last one then my UFD does NOT boot.
I've compared the related MBRs and the results are in the screenshot.

Posted Image

You say:

since in making the second partition, you possibly also induced changes in the first partition :)
Removing the second partition does NOT change the first partition.

Frankly I do not see the sense of your statement. :)

It appears quite obvious that the 1st partition does not change if I delete the second just created partition from my UFD, because what that surely changes it is its MBR (with the Partition Table inside).

Here we are talking about MBR and the Partition Table inside it, and the possible changes in the partitions are pointless: if I create 2 partitions on my UFD and then I delete the last one it seems rather obvious that also the MBR and the Partition Table inside it will change, don't you?

If I delete the second partition and left on my UFD only 1 partition, then it does NOT boot (just tried one more time, just for you) and does not boot neither "grldr" (grub4dos) nor "ntldr" (Full XP).

I think (really, I am convinced) that the matter can be in the MBR/Partition Table, but the EXPERT are you not me: do you remember?

I have the impression how if some experts want that my way is not working.
I have the impression how if some experts would have wanted to think by yourselves about the double partitioning as the way in order to increase your UFD bootability (even without still to know the real reason).
Instead, that idea is mine: and I am not an expert...
I hope to be wrong in those my impressions.

However if you take a look to the screenshot above then you probably will convince yourself that your above statement is pointless.
On the other hand you can make the same identical test by yourself and then you can see by yourself what that already appeared as obvious. :)

#24 wimb

wimb

    Platinum Member

  • Developer
  • 2585 posts
  • Interests:Boot and Install from USB
  •  
    Netherlands

Posted 14 April 2009 - 10:12 AM

@online
I am convinced that your method of improving bootability of USB-sticks by making a second partiton works,
but unfortunately I am not able to test it, since all my (about 10 different) USB-sticks NTFS formatted with HP Tool are booting on 3 different computers.
Then I have another computer which is NOT willing to Boot USB-sticks, but changing the stick by creating a second partition did not make any difference.
But again, I believe that there will be enough cases where your method will work, and can make the difference.

If I delete the second partition and left on my UFD only 1 partition, then it does NOT boot (just tried one more time, just for you) and does not boot neither "grldr" (grub4dos) nor "ntldr" (Full XP).

Thanks for doing this test, it even better proofs your success. You are more the expert than me on this subject :)

Have you ever tried to use TinyHexer to compare the MBR's
It can with File > Disk > Open drive ... compare the MBR with a file using option Tools > Compare to file ...

#25 was_jaclaz

was_jaclaz

    Finder

  • Advanced user
  • 7100 posts
  • Location:Gone in the mist
  •  
    Italy

Posted 14 April 2009 - 10:27 AM

@online

I am sorry if you felt ridiculized, but you seem to completely fail to grasp the essence of what I said before, in the effort to prove beyond any reasonable doubt that having two partitions is a requisite to increase bootability of a USB stick.

I will try again:
  • I have already taken as your word on it that you have a particular machine that only boots if the USB stick has more than one partition. :)
  • Such machines, to my knowledge, are very, very rare, up to the point where you are the only one that reported such behaviour. (you cannot define me as an "expert" and right after ignore my experience on the specific field, or, to be more exact, you can :(, but it's not nice :): do make up your mind, you either trust me and try to follow my suggestions or you decide that I am a crazy nut that posts random things in order to undermine your findings)
  • The reason why that particular machine boots if there are two partitions is unknown :), but in my opinion it is to be found in it's BIOS.
  • It's your idea, your finding, your everything, noone is going to deprive you of the paternity of the observation. :)
  • Due to point #2 above, it is my opinion that having two or more partitions is not a "general" requisite for USB bootability, meaning that it will increase the bootability of the stick only in a very small number of cases, at the moment we totaled 1 (one) such report
  • Since it doesn't make ANY harm, expecially because if the stick is partitioned with a "balanced" CHS and LBA you will have anyway some spare sectors at the end of the stick, I find the idea very good, but as said, I don't find it a "requisite", but rather an attempt (or exception) should the "normal", "proper" way be not working.
  • Though I myself have been trying for several years to convey the idea that the "proper" way to partition/format any HD-like media is having an initial smallish FAT16 partition and all the rest as one or more Logical Volumes inside Extended partition, following the good advise by Gilles Vollant, and this is the particular way with which I personally partition/format ALL my HD-like devices, this does not necessarily means that it is needed to do so. (I hope you understand this difference, the whole point is that I try not to make proselytism, only offer alternatives from which everyone is free to choose the one that suits him/her better)

This said, I hope you will recognize this sentences as correct:
  • You observed a certain behaviour of your motherboard, and empirically found that partitioning the stick in two partitions solved the bootability problem.
  • What you (and everybody else) is missing is WHY that particular motherboard behaves like that.

I already posted a few open questions:

More questions that need an answer:

  • Is the fact because there are actually two partitions on the stick (both accessible to windows if a filter driver is installed)?
  • Or is it because there are two entries in the partition table ( the second being a "fake one")?
  • Or is it because the BIOS on that machine can only access a limited in size first partition? (as an example The 1,024 Cylinder (504 MiB / 528 MB) Barrier)

that only you can answer to by experimenting, but that you decided to ignore :), in favour of the, without abusing of my "expert" status, queer theory of different 00's :) .


If I were you, I would repeat the comparison using tinyhexer:
http://www.mirkes.de/files/
http://www.mirkes.de/files/mpth_18.exe

With a net saving of at least €37.76:
http://www.sweetscape.com/010editor/

Commercial Use: $129.95 US
Home/Academic Use: $49.95 US



Different bytes will become green which is a more "neutral" colour, identical bytes will remain "white".

Once you will be convinced yourself that 00's are 00's, you might want to reconsider experimenting in the direction suggested by my open questions, in order to try and find WHAT is causing the behaviour you observed.


jaclaz




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users