ahci errors with xen 4.1.1, linux 3.0, virtualbox

ahci errors with xen 4.1.1, linux 3.0, virtualbox

[Sven Köhler]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

Hi,

when booting xen 4.1.1, vanilla linux 3.0 (+ vga patch, all 64bit)
inside a virtualbox with 4 CPUs (virtualbox 4.1, my host has 2 cores, 2
threads each), the virtual machine has serious AHCI problems during
boot. Sometimes up to the point, that it doesn't boot at all. However,
usually after a few errors everything works fine.

I tried clocksource=xen, but it didn't improve the behaviour.
When reducing the number of CPUs seems to decreases the likelyhood of
AHCI errors (well, this is my experience so far).

Any ideas, what might be the cause of this?
You find a tgz with dmesg, xm dmesg, etc.

The errors also happened with konrad's 2.6.39 kernel.
The errors never happen with the good old 2.6.18.


Regards,
  Sven

[-- Attachment #2: ahci.tgz --]
[-- Type: application/x-compressed-tar, Size: 13324 bytes --]

[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: ahci errors with xen 4.1.1, linux 3.0, virtualbox

[Sven Köhler]

This might be related:

On another try, xm dmesg complains that the platform times (ACPI PM) has
wrapped unexpectedly 10 or more times. So is it possible that this is a
clocksource issue after all?

Is this a comprehensive list of clocksources that Xen supports?
http://wiki.xensource.com/xenwiki/xenpm#head-253cbbe6cf12fa31e10022610cd7090aa980921f


Regards,
  Sven

Re: ahci errors with xen 4.1.1, linux 3.0, virtualbox

[Sven Köhler]

Hi,

so with clocksource=pit the issue hasn't reoccured so far.

So this boils down to a bug in virtualbox?


Regards,
  Sven