Cannot run WinXP on the Xen fedora core 6

[Orathai Sukwong]

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.

[Daniel P. Berrange]

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.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  [phone] -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|

[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.

[phone] - Mobile
[phone] - Home Office

acathrow(a)redhat.com