[Fedora-xen] Cannot run WinXP on the Xen fedora core 6

[Andrew Cathrow]

I've got WinXP running on the 2647 kernel on FC6. The guest runs well,
but the install was more than a little challenging. I ended up
installing the guest using QEMU (which was tremendously slow) but once
the install had completed the guest runs well.



On Mon, 2006-09-25 at 22:04 +0100, Daniel P. Berrange wrote:

> On Mon, Sep 25, 2006 at 04:51:46PM -0400, Orathai Sukwong wrote:
> > Hi,
> > 
> > I've tried to bring the 32bit WinXP guest up on the Fedore Core 6-xen. My
> > xen release is 2.6.17-1.2647.fc6xen and version #1 SMP Wed Sep 13 13:26:53
> > EDT 2006. xen_caps are xen-3.0-x86_32p hvm-3.0-x86_32 hvm-3.0-x86_32p. My
> > machine is running the Fedora Core 6 i686. The problem is that when I booted
> > the WinXP guest, it crashed. Here are some debugging info:
> > 
> > (XEN) (GUEST: 5) unsupported PCI BIOS function 0x0E
> > (XEN) (GUEST: 5) int13_harddisk: function 15, unmapped device for ELDL=81
> > (XEN) 0f 6f, This opcode isn't handled yet
> > (XEN) handle_mmio: failed to decode instruction
> > (XEN) mmio opcode: va 0xf81393c0, gpa 0xf00783c0, len 3: 0f 6f 06
> > (XEN) domain_crash_sync called from platform.c:869
> > (XEN) Domain 5 (vcpu#0) crashed on cpu#0:
> > (XEN) ----[ Xen-3.0-unstable    Not tainted ]----
> > (XEN) CPU:    0
> > (XEN) EIP:    0008:[<bf902854>]
> > (XEN) EFLAGS: 00010202   CONTEXT: hvm
> > (XEN) eax: f81393c0   ebx: e10875cc   ecx: 0000002a   edx: 00000008
> > (XEN) esi: f81393c0   edi: e10875d2   ebp: f86440d4   esp: f86440c8
> > (XEN) cr0: 80010031   cr3: 001f9ea0
> > (XEN) ds: 0023   es: 0023   fs: 0030   gs: 0000   ss: 0010   cs: 0008
> > (XEN) sh_update_paging_modes: postponing determination of shadow mode
> > (XEN) (file=hvm.c, line=195) Allocated port 3 for hvm.
> > (XEN) vmx_do_launch(): GUEST_CR3<=001f9ea0, HOST_CR3<=2bd03000
> > (XEN) (file=event_channel.c, line=176) EVTCHNOP failure: domain 0, error
> > -22, line 176
> > 
> > domid: 6
> > qemu: the number of cpus is 1
> > shared page at pfn:1ffff, mfn: 21d54
> > buffered io page at pfn:1fffd, mfn: 21d56
> > char device redirected to /dev/pts/2
> > bind interdomain ioctl error 22
> > 
> > Any ideas what went wrong? Please help me. Thank you.
> 
> This is a bug in the HVM emulation layer. In some cases it needs to emulate
> instructions, but it looks like WinXp issues one it was unable to emulate
> yet. I checked upstream  xen-unstable.hg  repository there's still no support
> for this. Please enter a bugzilla ticket with the details from this mail so
> the problem can be tracked.  I'm afraid there's no immediate work around I
> can think of to make this work, so you might want to try Windows 2k instead
> which I believe has been getting more upstream testing than XP
> 
> Regards,
> Dan.


Andrew Cathrow
Red Hat, Inc.

(678) 733 0452 - Mobile
(404) 437 6178 - Home Office

acathrow at redhat.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/fedora-xen/attachments/20060925/241b032b/attachment.htm>