Ubuntu 10.04 Host Freezes After Resuming From Suspend

[vmnewbie]

Hi all,

I was wondering if anyone has encountered a similar problem described below.
This is 100% reproducible for me.

Setup:
Intel Core2 Duo E8400.
Gigabyte EP45T-U3DP with latest BIOS.
Virtualization extension enabled in BIOS.
Ubuntu 10.04 64-bit host OS (2.6.32-21-generic #32-Ubuntu SMP)
Virtualbox 3.1.6 (downloaded from virtualbox.org)
Windows XP Pro 32-bit guest OS

Problem:
The virtual machine can be started with VT-x/AMD-V without any problem when the host is freshly booted. If the host is suspended (S3 suspend to RAM) and subsequently resumed, then Virtualbox will completely freeze the host if I attempt to start any virtual machine. If VT-x/AMD-V is disabled, then the virtual machine will run and won't freeze the host.

(Display 2D/3D acceleration setting does not appear to have any effect on freezing)

Same freezing occurs when VMplayer is run.

There is a ticket 5048 that describes the same problem (http://www.virtualbox.org/ticket/5048), but I don't understand the resolution

Any insight is appreciated.

[yannisf]

Same here! After a lot of research I think that it is a BIOS bug. Gigabyte was not fond to accept that, but in most cases that this issue appears it has to do with a gigabyte mobo. I just hope that gigabyte will issue a BIOS upgrade sometime, or that Virtualbox will work around it.

The point is that the issue exists, and we as end users there is not much we can do. I submitted a ticket http://www.virtualbox.org/ticket/7216.

[mfdias]

I have the exact same issue on my machine (Gigabyte EP45-UD3P motherboard). I even tried updating the BIOS to the latest beta version (FFb - released Oct 2010) but the problem still persists. I tried the workaround described in ticket 5048 (disable VT-x/AMD-V before suspend and re-enable on resume) but that does not work for me. At this point my only options are to either run the VM without VT-x or to hibernate instead of suspend to RAM. FYI my host OS is Ubuntu 10.04 (2.6.32-generic-pae kernel) and I'm running Virtualbox 3.2.12.

[mhanor]

have you tried running rdmsr 0x3a? to check the value of that MSR before entering sleep and after resuming from it?

[psychoboy]

I can confirm the same behavior with the following setup...

Setup:
Intel Core2 Quad Q9400.
Gigabyte GA-E7AUM-DS2H with latest BIOS v.F4
Virtualization extension enabled in BIOS.
Ubuntu 10.10 64-bit host OS (2.6.35-23-server #41-Ubuntu SMP)
Virtualbox 3.2.12 (downloaded from virtualbox.org)
Windows XP Pro 32-bit guest OS

[psychoboy]

have you tried running rdmsr 0x3a? to check the value of that MSR before entering sleep and after resuming from it?

I get a value of 5 in both instances.

[mhanor]

does it hang every time you resume with a running, VT enabled, VM?
if that MSR is always 5 after resuming, then it's not a BIOS bug, unless there's another CPU feature that Virtualbox requires, which the BIOS fails to re-enable it

[psychoboy]

If I start virtualbox and then start a VM, the host(ubuntu 10.04) hangs after coming back from a suspended state. For example... I stop or shut down any VMs, I close virtualbox, and then suspend the host machine. Then wake the host, start virtualbox, start any VM and the host hangs and requires a reset. It does not happen if I disable VT-x, but that makes all my VMs run really slow(almost unusable).

[borisprochazka]

I have a similar problem on my Lenovo T410 with M540 CPU. host SuSE 11.3 and guest WinXP. I useally stop my gusts before suspending the host. After wakening up the host again when I try to star a gust they all goes VERY slow (may be 100 times or more slower then normal). After a reboot of the host OS all guest are running as normal again.

[borisprochazka]

I forgot to mention that I use VBox 4.0.0 r69151. Since this is a different hardware from the above mentioned I don't think this is a BIO or hardware problem

[fixedwheel]

I forgot to mention that I use VBox 4.0.0 r69151.

VBox 4.0.2 is available, did you try that?

Since this is a different hardware from the above mentioned I don't think this is a BIO or hardware problem

original problem above is a buggy BIOS/hardware problem - and whatever yours is, i don't think its closely related (you could have made a thread of your own...)