19 replies to this topic

[Holmes.Sherlock]

I created a bootable USB stick following cdob's method. Everything went fine when tested on VMware Workstation 7, but on real hardware platform, it failed. I used a couple of LENOVO laptops for experiment. One is armed with 4GB DDR3 RAM & Phoneix BIOS where as the other is equipped with 2GB DDR3 RAM & Insyde BIOS. There is 500GB internal HD in both the laptops. On both the machines, the RAM loading completed, setup started as usual till

Setup is staring Windows

but, it was supposed to display the list of available drives detected.

After that, the screen got blank without any error message. The UFD activity also stopped. I tried another 5/6 times with the following variations:

• Uncommenting the line in menu.lst
  

  map --mem /Inst/XP_INST.IMA (fd1)

  which has been told as a workaround for some buggy BIOSes.
  
• Changing the SATA mode to both AHCI & COMPATIBILITY in the BIOS.

Worth mentioning that the XP source I've used is nLited with AHCI drivers needed for the laptop & works perfectly fine when comes to installing XP from CD itself instead of the image.

[ilko]

This looks exactly as this one:
http://www.msfn.org/...all-iso-method/

Does the ~BT & ~LS (used in WinSetupFromUSB 0.2.3 and USB _multiboot) method work on them?
Does WinSetupFromUSB 1.0 beta6 created USB stick fail the same way?

Can you test and report?

[Holmes.Sherlock]

This looks exactly as this one:
http://www.msfn.org/...all-iso-method/

In that case, the computer was restarting. But, in my case, it hangs even before it restarts.

Does the ~BT & ~LS (used in WinSetupFromUSB 0.2.3 and USB _multiboot) method work on them?
Does WinSetupFromUSB 1.0 beta6 created USB stick fail the same way?

I haven't tried that.

Can you test and report?

Sure, I'll try & let u know.

[ilko]

In that case, the computer was restarting. But, in my case, it hangs even before it restarts.

Nope, it hangs the same way with black blank screen, just the first post is a bit misleading.

[Holmes.Sherlock]

Nope, it hangs the same way with black blank screen, just the first post is a bit misleading.

Yes, it is.

[Holmes.Sherlock]

Oh, I forgot to report that in my case also, the USB light stopped flashing while the HDD light did. Looks almost same reported by ilko in the referenced thread at Post #14

In AHCI only the USB light flashes, in IDE only the HDD light.

[Holmes.Sherlock]

@ilko

I've prepared an USB stick with WinSetupFromUSB 1.0 beta6 in the following manner:

1. Selected the USB stick from the drop down menu.
2. Checked the checkbox beside Windows 2000/XP/2003 Setup.
3. Browsed the WinXP SP3 folder extracted on my HDD.
4. Clicked on GO button.

An UFD prepared this way produces BSOD on VMWare & EXACTLY the same output on my physical machine.

Any suggestion or any other method/option to try out & report?

[ilko]

I'd try WinSetupFromUSB 0.2.3, my guess is that it's going to work just fine on the real hardware. May need the modified ntdetect.com if bluescreens with 0x7B:
http://www.msfn.org/...mp;#entry738009
http://www.datafileh...d-dfc48473.html

Try on the real hardware please.

[Holmes.Sherlock]

Try on the real hardware please.

Sorry ilko, you didn't notice the last part of my post.

An UFD prepared this way produces BSOD on VMWare & EXACTLY the same output on my physical machine.

With the BOLD-ed part , I tried to mean that the UFD, prepared with WinsetupfromUSB, also produced EXACTLY SAME results as it was with cdob's method (stated in Post #1), i.e. the screen went blank after "Setup is starting Windows", UFD light turned off & HDD light was blinking showing high HD activity.

[ilko]

I've read it

1.0 beta versions and 0.X.X of WinSetupFromUSB use completely different method. That's why I am asking you to try 0.2.3, which uses default, supported by MS method.

The results on VMWare would be irrelevant, we need to figure out what is going on on some, I guess, newer machines.

[Holmes.Sherlock]

Thank u a lot ilko, seems that 0.2.3 has done the impossible . Though u were not interested in knowing the result on VMWare, I tested on it. It BSOD-ed. But fortunately, on real machine, now it presented the partition selection & F8 EULA agreement screen. As I can't tamper with the XP & 7 installation of my laptop @ this moment, I didn't proceed with the installation further. But as my college project is over & the laptop becomes free for the experiment, I'll give a try for a complete install. At this time, I've certain questions in mind:

• Why didn't u incorporate both the methods used in older & newer version of WinSetupFromUSB in the newest version? Because for such erratic/queer BIOSes like mine, the previous version proved better than the newer. May we expect to have a combined version in near future?
  
• Can u please outline the techniques u are using in 0.2.3 & 1.0 beta versions of the utility?
  
• What modification does modified NTDETECT.COM have?
  
• If u have time & feel it to be worth spending time, can u please investigate why the test ended up with BSOD on VMWare?

[ilko]

I think I've found a workaround, at least with small ISO method and WinSetupFromUSB 1.0 versions.
http://www.msfn.org/...post__p__947272

For now with firadisk method I have no clues, since firadisk driver still needs memory mapped image.

Can you test WinSetupFromUSB 1.0 beta6, editing winsetup.lst and removing --mem option?

...

title First part of Windows XP Professional SP2 setup from partition 0

map --unmap=0:0xff

map --unhook

find --set-root --ignore-cd /WINSETUP/XPpSP2.ISO

map /WINSETUP/XPpSP2.ISO (0xff)

...

[Wonko the Sane]

For now with firadisk method I have no clues, since firadisk driver still needs memory mapped image.

Not really-really.
Experimental support for non-mem mapped image is present.


Wonko

[i]

ok, this maybe odd but i see all dudes (devz) online today in this topic, Sha0, karyonix, erwan.l, wimb, (and obviously Wonko as everyday )

[ilko]

Not really-really.
Experimental support for non-mem mapped image is present.


Wonko

Can you provide more information? Couldn't find anything in version history or the last few pages of the firadisk topic.
Tried firadisk 0.0.1.24 and no --mem mapping with XP SP3 ISO and this results in BSOD 0x7B as expected.

For the record- UBCD4Win iso image based on XP SP3 and firadisk on the same computer worked just fine.

[Holmes.Sherlock]

Can you test WinSetupFromUSB 1.0 beta6, editing winsetup.lst and removing --mem option?

I'll try it today.

[Wonko the Sane]

Can you provide more information? Couldn't find anything in version history or the last few pages of the firadisk topic.
Tried firadisk 0.0.1.24 and no --mem mapping with XP SP3 ISO and this results in BSOD 0x7B as expected.

For the record- UBCD4Win iso image based on XP SP3 and firadisk on the same computer worked just fine.

Start from here:
http://www.boot-land...?...8804&st=246

Same has been added to WinVblock:
http://www.boot-land...?...8168&st=207

But I was talking of .img not .iso.


Wonko

[cdob]

http://www.boot-land...?...st&p=102765

Windows in disk image must know the partition inside disk image. Don't change MBR signature of disk image, or partition start, partition length of Windows partition inside disk image.
...
It probably will not work with PE or Windows setup.

I understand:
some USB drive relating settings are missing at default installation media.
It may be possible to add this settings to a XP installation media.
However this XP installation media would match this USB drive only.
Details are misssing.


There is another approach nowadays.
http://www.vmlite.co...;start=108#4149

You can use VBoot to install an OS to the real harddisk. This way, you can boot from an ISO file, instead of using a real cdrom.

vboot cdrom=(hd0,1)/winxp-sp2.iso floppy=(hd0,1)/vboot.img boot=cdrom

Booting does work at QEMU, no BSOD 0x7B.

[ilko]

...
There is another approach nowadays.
http://www.vmlite.co...;start=108#4149
Booting does work at QEMU, no BSOD 0x7B.

Very nice find, thanks.

[Holmes.Sherlock]

[quote name='ilko' post='114911' date='Nov 28 2010, 07:26 AM']Can you test WinSetupFromUSB 1.0 beta6, editing winsetup.lst and removing --mem option?

[code=auto:0][/quote] U were right. It worked. I guess that it's a problem with Grub4DOS memory mapping for some BIOSes. If it can be resolved, then cdob's method can also be get to work.