RE: [Xen-users] HVM WinXP dom crash

["Petersson, Mats" <Mats.Petersson@xxxxxxx>]

 

---

From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx [mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of jd_jedi@xxxxxxxxxxxxxxxxxxxxx
Sent: 08 October 2006 01:30
To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] HVM WinXP dom crash

Hi
  Trying to get WinXP HVM  to install as domU on AMD. Got latest Xen Kernel and Xen package (NOT unstable or debug).

When start, a window pops up  for few seconds.. then it dies and another one comes up... and it disappears too. (Xen is trying to restart and giving up because the restart happens in 2 seconds..) It seems DomU is not able to read/understand something from Cirrus VGABIOS. 

   Any ideas on getting this to work ? Do I need to change some bios setting ?
   Please see details below.
Thanks
/Jd


OS :
-----
Linux  2.6.17-1.2187_FC5xen0 #1 SMP Mon Sep 11 02:05:21 EDT 2006 i686 athlon i386 GNU/Linux
Xen :
-----
xen-3.0.2-3.FC5

------ config ------
import os, re
arch = os.uname()[4]
if re.search('64', arch):
    arch_libdir = 'lib64'
else:
    arch_libdir = 'lib'
kernel = "/usr/lib/xen/boot/hvmloader"
builder='hvm'
memory = 500
name = "WinXP"
vif = [ 'type=ioemu, bridge=xenbr0' ]
disk = [ 'phy:/dev/VG_PL1/winxp,ioemu:hda,w' ]
device_model = '/usr/' + arch_libdir + '/xen/bin/qemu-dm'
cdrom='/dev/hda'
boot='d'
sdl=1
vnc=0
vncviewer=0
stdvga=0
serial='pty'
ne2000=0
---------------------------

xm dmesg
=======
bunch of lines... similar to below line....

(XEN) (file=traps.c, line=1385) Domain attempted WRMSR c0010041 from 00000001:00000e0e to 00000190:00010e02.
(XEN) (file=traps.c, line=1385) Domain attempted WRMSR c0010041 from 00000001:00000e0e to 00000190:00010e02. 

 

Completely harmless warnings. Explanation: This is the powernow driver (or something else accessing PowerNow! registers) - it's being ignored, so I wouldn't worry about it - there is a patch that adds those MSR's to the msr-white-list in traps.c, but for all intents and purposes, it's only going to slow your processor down (from 11x to 5x multiplier, if I read the tables and binaries right) if it were successfull in writing the register - which may be useful if this is a laptop, but otherwise I wouldn't worry about it. If you don't want to see these things, I'd suggest you remove the PowerNow! driver from Dom0.

 

 
and then...

(XEN) (file=hvm.c, line=196) Allocated port 4 for hvm.
(XEN) (GUEST: 1) HVM Loader
(XEN) (GUEST: 1) Loading ROMBIOS ...
(XEN) (GUEST: 1) Loading Cirrus VGABIOS ...
(XEN) (GUEST: 1) SVM go ...
(XEN) (file=hvm.c, line=385) HVM vcpu 1:0 did a bad hypercall 1480936960.
(XEN) (GUEST: 1) Failed to invoke ROMBIOS 

 

This is wrong and why things aren't working. I'm not sure if it's because your ROMBIOS is build incorrectly or something else...

 

Bad hypercall is definitely not right here - it seems like the BIOS and the Hypervisor are disagreeing on how to "boot" the machine. Since it fails, the system shuts down, which isn't really a crash - it's a failure with a "controlled emergency landing".

 

--

Mats

 
(XEN) Guest shutdown exit
(XEN) domain_crash_sync called from svm.c:2999  
(XEN) Domain 1 (vcpu#0) crashed on cpu#0:
(XEN) ----[ Xen-3.0-unstable    Not tainted ]----
(XEN) CPU:    0
(XEN) EIP:    0000:[<00100039>]
(XEN) EFLAGS: 00000246   CONTEXT: hvm
(XEN) eax: 00000000   ebx: 00000000   ecx: 00129ae8   edx: 000000e9
(XEN) esi: 00000000   edi: 00000000   ebp: 00129ae8   esp: 00129ae8
(XEN) cr0: 0005003b   cr3: 00000000
(XEN) ds: 0000   es: 0000   fs: 0000   gs: 0000   ss: 0000   cs: 0000
==

---

How low will we go? Check out Yahoo! Messenger’s low PC-to-Phone call rates.

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