BUG:VBox 2.0.6(2008-11-21) does not handle Intel-VT properly

xok · Post by **xok** » 28. Nov 2008, 17:32

Hello,

Just want to report a bug regarding the VirtualBox 2.0.6 (released 2008-11-21)

Intel-VT it is not being properly emulated/used with VirtualBox (is case you are wondering the Intel-VT option is checked).

HOST: Notebook - Asus M50SV with Windows Vista Business.
GUEST: Microsoft Windows Enterprise 2003 R2 VLP - x64

When starting the mounted ISO of Windows it reports:
Attempting to load an x64 operating system, however this CPU is not compatible with x64 mode. Please install a 32-bit X86 operating system.

Setup cannot continue. Press any key to exit.
I was able to install a Virtual Machine with x64 Windows on VMWare v.1.0.5. (Build 80187).

Regards,
RS
_______________________________________________________

ADDENDUM TO THE ABOVE:
Having VMWare v.1.0.5. (Build 80187) running a Win 2003 ent x64 when I start a EMPTY VirtualBox 2.0.6 virtual machine to START INSTALLING Win 2003 ent x64 (with the ISO image) it AUTOMATICALLY REBOOTS THE MACHINE (Windows Vista Business)!

Tested twice and both times it rebooted the machine.

Post by **Sasquatch** » 28. Nov 2008, 22:31

The forum is not meant for bug reports. See the VirtualBox FAQ at the bottom for a link to the real bug tracking system. Register for an account there, as your forum account is not compatible with that system.

Post by **TerryE** » 29. Nov 2008, 16:43

xok, Thanks for this but Sasquatch is right. This forum is largely run by community supporters and the VBox development team rarely read or post to it, so there is little point in posting bug reports here. Use the VBox public Bugtracker to raise tickets for reporting problems in VirtualBox and make suggestions for improvements. This is read by the developers.

However, if you haven't done so, you should read the Forum Posting Guide. This suggest that you should research the forums and VBox site for answers before asking a Q that has been address before. In this case, your bugs aren't; they are known features that have been discussed in other posts.

There are two approaches to Virtualization: H/W assisted and software based. Each has its pros and cons. VBox supports both for 32bit VMs, but only H/W assisted virtualisation for 64bit. AFIAK VMware uses software virtualisation, which works on all 64-bit CPUs. H/W virtualisation only works on VT-x / AMD-V enabled 64-bit CPUs, and many early 64-bit CPU did not have such support. Also many BIOSs by default do not enable this even if your hardware chip supports it.
VBox cannot run inside another (e.g. VMware) VM. Again a stated design limitation.

xok · Post by **xok** » 1. Dec 2008, 21:37

Hello Sasquatch,

I tried to report the bug on "Trac", but I did not had permissions, I thought it was for Admnistrators only and reported it here.

OK, I will subscribe to the bug tracking system and post it there.

Sasquatch wrote:The forum is not meant for bug reports. See the VirtualBox FAQ at the bottom for a link to the real bug tracking system. Register for an account there, as your forum account is not compatible with that system.

Post by **TerryE** » 1. Dec 2008, 21:47

xok, please read my reply. There is no point in raising this as a bug, because it isn't. At best what you are asking for is a functional enhancement.

Post by **sandervl** » 2. Dec 2008, 10:16

You can't run VMWare and VirtualBox at the same time if one if them is using software virtualization and the other VT-x. Blame Intel for this design flaw.

Both programs would need to check for each other's presence and that's not so easy to archieve.

Regarding VT-x not working; post a comment here. Include your VBox.log too please.

xok · Post by **xok** » 13. Dec 2008, 19:57

Thank you for your reply but you did not understood my point.

1 - I am not trying to run VirtualBox inside VMware.

2 - What is happening is when you start VirtualBox when running a instance of VMware, e.g. Windows 2003 R2 VLP 64 Bits (NOT INSIDE VIRTUALBOX) , the computer SHUTDOWNS, REBOOTS, KAPUTZZZ...!

There is no error whatsoever... it is like unplugging the power cord!

If this is by design... OK! But for me it is BUG, ANY kind software should NOT REBOOT a machine...

TerryE wrote:xok, please read my reply. There is no point in raising this as a bug, because it isn't. At best what you are asking for is a functional enhancement.

Post by **sandervl** » 13. Dec 2008, 20:01

Like I said: vent your frustration on Intel. Do not run VMWare and VirtualBox at the same time when using VT-x!

Post by **TerryE** » 13. Dec 2008, 21:08

xok wrote:What is happening is when you start VirtualBox when running a instance of VMware, e.g. Windows 2003 R2 VLP 64 Bits (NOT INSIDE VIRTUALBOX) , the computer SHUTDOWNS, REBOOTS, KAPUTZZZ...

I think that what you are requesting is a functional enhancement for the H/W virtualisation prologue to check that no other VMMs are active before initialising the VT-x based VMM, and if so then raising a "VMM conflict error" before exiting rather than killing the host OS.

So what happens if you are running VBox and then start up VirtualBox? How do you know that VirtualBox caused this panic and not VMware?

Post by **sandervl** » 13. Dec 2008, 23:27

VMWare causes the panic as VirtualBox is in VT-x mode. We detect the reverse already during startup and refuse to continue in this case.

There's no easy way to solve the problem unless you can convince VMWare to participate in the process. We could check if e.g. VMWare or VirtualPC are running, but what about the case when you start them later? Partial solutions are not useful here.

Like I said, Intel designed VT-x in this way so the only solution is for the host OS to offer services to intialize VT-x. Apple did this in Mac OS X. Still everybody using VT-x would have to use this interface.

Post by **vbox4me2** » 14. Dec 2008, 00:12

sandervl wrote:There's no easy way to solve the problem unless you can convince VMWare to participate in the process.

Just buy them out, VB is better anyway

Post by **TerryE** » 14. Dec 2008, 04:14

sandervl wrote:VMWare causes the panic as VirtualBox is in VT-x mode. We detect the reverse already during startup and refuse to continue in this case.

Thanks Sander, maybe xok can try taking this up with the VMware forum and see where he gets