Xen devel - Oct 2006 - RE: W2K3 Server Installation fails with ACPI=1 on

Steven,
I found where the difference is, it may be my fault:
My changeset number + 1 = your changeset number.
My 11160 is your 11159, and my 11161 is your 11160.
Attached please see those patches,
Thanks,
Winston,

-----Original Message-----
From: Steven Smith [mailto:sos22@hermes.cam.ac.uk] On Behalf Of Steven
Smith
Sent: Wednesday, October 11, 2006 10:33 AM
To: Wang, Winston L
Cc: Stefan Berger; Zheng, Shaohui; xen-devel@lists.xensource.com;
sos22@srcf.ucam.org
Subject: Re: [Xen-devel] W2K3 Server Installation fails with ACPI=1 on
> Sorry to hear more ACPI Windows problem.
> 
> Noticed your issue is W2K3 not W2K:
> 
> W2k3 Server installation issue is caused by HLT instruction; Xin will
> post the fix soon.
Ah, okay.  I''ll be ignoring the Subject: then, will I?
> We also found W2k and XP ACPI mode have boot and installation problem,
> the root cause is new VBD/VNIF PCI device checking in 11160.
xen-unstable.hg$ hg export 11160
# HG changeset patch
# User ssmith@weybridge.uk.xensource.com
# Date 1155815205 -3600
# Node ID 5e3166de7120f474903d58bd74b7e2d12c3d8ea7
# Parent  24827cb8a94596987284c4b934365165468949a2
[NET] Fix ``rx->offset: 10, size: 4294967295'''' bug.

Signed-off-by: Steven Smith <sos22@cam.ac.uk>
...

I don''t think your changeset numbers match up with mine.  What does
hg export 11160 say for your tree?
> 11160 added new HVM PCI device, and 11161 modified 11161''s new PCI
> device ID, If want to verify this need to revert 11161 first, then
> revert 11160.
I''m not sure I understand you here; 11161 modified its own new
PCI device ID?

Steven.
> We have the fix now and will post it soon.
> 
> Regards,
> 
> Winston,
> 
> -----Original Message-----
> From: Steven Smith [mailto:sos22@hermes.cam.ac.uk] On Behalf Of Steven
> Smith
> Sent: Wednesday, October 11, 2006 7:55 AM
> To: Wang, Winston L
> Cc: Stefan Berger; Zheng, Shaohui; xen-devel@lists.xensource.com;
> sos22@srcf.ucam.org
> Subject: Re: [Xen-devel] W2K3 Server Installation fails with ACPI=1 on
> 
> > The problem is cause by the VBD/VNIF PCI device in DM checked in
> > changeset 11160: PCI Device ID: 5853(Xen),class:ff80 (unknown)
> > ACPI Window does not understand this device and can not find device
> > driver associate with it.
> > 
> > To verify this just revert the patches of changset 11161/11160
> > We are working on the fix now, revert patch is not the solution
> > anyway...
> I just tried this, and as far as I can see cset 11161 doesn''t make
any
> difference to the behaviour, while 11160 is fairly obviously
> irrelevant to this kind of thing.
> 
> With or without 11161, I see guests crashing with
> 
> (XEN) sh error: sh_page_fault__shadow_3_guest_2(): disabled-APIC
access:
> not supported
> (XEN) .domain_crash called from multi.c:3197
> (XEN) Domain 3 (vcpu#0) crashed on cpu#1:
> (XEN) ----[ Xen-3.0-unstable  x86_64  debug=y  Not tainted ]----
> (XEN) CPU:    1
> (XEN) RIP:    0008:[<000000008081d1ac>]
> (XEN) RFLAGS: 0000000000010086   CONTEXT: hvm
> (XEN) rax: 00000000ffd01000   rbx: 0000000000000000   rcx:
> 00000000ffdff120
> (XEN) rdx: 00000000fec00000   rsi: 00000000ffd03034   rdi:
> 0000000000000000
> (XEN) rbp: 0000000000060aa4   rsp: 0000000000060a84   r8:
> 0000000000000000
> (XEN) r9:  0000000000000000   r10: 0000000000000000   r11:
> 0000000000000000
> (XEN) r12: 0000000000000000   r13: 0000000000000000   r14:
> 0000000000000000
> (XEN) r15: 0000000000000000   cr0: 0000000080050033   cr4:
> 0000000000000651
> (XEN) cr3: 0000000000fd97a0   cr2: 0000000000000000
> (XEN) ds: 0023   es: 0023   fs: 0030   gs: 0000   ss: 0010   cs: 0008
> qemu-dm[5399]: segfault at 00002aaba96ec000 rip 00002ab993166fe0 rsp
> 00007fffff912188 error 6
> 
> (This is using a 64 bit host, but I see a similar crash with 32 and
> 32p hosts).
> 
> Just to confirm, 11161 in xen-unstable.hg is
> 
> # HG changeset patch
> # User chris@kneesaa.uk.xensource.com
> # Date 1155812483 -3600
> # Node ID 5c1021595e3ce923ff11e338172d4639da2aecb5
> # Parent  5e3166de7120f474903d58bd74b7e2d12c3d8ea7
> [qemu] Add the xen platform device to the qemu PCI bus.
> Useful functionality will come later.
> 
> Signed-off-by: Steven Smith <ssmith@xensource.com>
> Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
> 
> 
> and 11160 is
> 
> 
> # HG changeset patch
> # User ssmith@weybridge.uk.xensource.com
> # Date 1155815205 -3600
> # Node ID 5e3166de7120f474903d58bd74b7e2d12c3d8ea7
> # Parent  24827cb8a94596987284c4b934365165468949a2
> [NET] Fix ``rx->offset: 10, size: 4294967295'''' bug.
> 
> Signed-off-by: Steven Smith <sos22@cam.ac.uk>
> 
> 
> These are the csets you''re talking about, yes?
> 
> Steven.


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

> Steven,
> I found where the difference is, it may be my fault:
> My changeset number + 1 = your changeset number.
> My 11160 is your 11159, and my 11161 is your 11160.
> Attached please see those patches,
> Thanks,
> Winston,
That does mean that your tree has one extra changeset in it that isn''t
in -unstable. I wander if that''s the problem. Might be worth cloning
unstable and then do a ''hg incoming'' in to it.

Ian 

 
> -----Original Message-----
> From: Steven Smith [mailto:sos22@hermes.cam.ac.uk] On Behalf Of Steven
> Smith
> Sent: Wednesday, October 11, 2006 10:33 AM
> To: Wang, Winston L
> Cc: Stefan Berger; Zheng, Shaohui; xen-devel@lists.xensource.com;
> sos22@srcf.ucam.org
> Subject: Re: [Xen-devel] W2K3 Server Installation fails with ACPI=1 on
> 
> > Sorry to hear more ACPI Windows problem.
> >
> > Noticed your issue is W2K3 not W2K:
> >
> > W2k3 Server installation issue is caused by HLT instruction; Xin
will
> > post the fix soon.
> Ah, okay.  I''ll be ignoring the Subject: then, will I?
> 
> > We also found W2k and XP ACPI mode have boot and installation
problem,
> > the root cause is new VBD/VNIF PCI device checking in 11160.
> xen-unstable.hg$ hg export 11160
> # HG changeset patch
> # User ssmith@weybridge.uk.xensource.com
> # Date 1155815205 -3600
> # Node ID 5e3166de7120f474903d58bd74b7e2d12c3d8ea7
> # Parent  24827cb8a94596987284c4b934365165468949a2
> [NET] Fix ``rx->offset: 10, size: 4294967295'''' bug.
> 
> Signed-off-by: Steven Smith <sos22@cam.ac.uk>
> ...
> 
> I don''t think your changeset numbers match up with mine.  What
does
> hg export 11160 say for your tree?
> 
> > 11160 added new HVM PCI device, and 11161 modified 11161''s
new PCI
> > device ID, If want to verify this need to revert 11161 first, then
> > revert 11160.
> I''m not sure I understand you here; 11161 modified its own new
> PCI device ID?
> 
> Steven.
> 
> > We have the fix now and will post it soon.
> >
> > Regards,
> >
> > Winston,
> >
> > -----Original Message-----
> > From: Steven Smith [mailto:sos22@hermes.cam.ac.uk] On Behalf Of
Steven
> > Smith
> > Sent: Wednesday, October 11, 2006 7:55 AM
> > To: Wang, Winston L
> > Cc: Stefan Berger; Zheng, Shaohui; xen-devel@lists.xensource.com;
> > sos22@srcf.ucam.org
> > Subject: Re: [Xen-devel] W2K3 Server Installation fails with ACPI=1
on
> >
> > > The problem is cause by the VBD/VNIF PCI device in DM checked in
> > > changeset 11160: PCI Device ID: 5853(Xen),class:ff80 (unknown)
> > > ACPI Window does not understand this device and can not find
device
> > > driver associate with it.
> > >
> > > To verify this just revert the patches of changset 11161/11160
> > > We are working on the fix now, revert patch is not the solution
> > > anyway...
> > I just tried this, and as far as I can see cset 11161 doesn''t
make
any
> > difference to the behaviour, while 11160 is fairly obviously
> > irrelevant to this kind of thing.
> >
> > With or without 11161, I see guests crashing with
> >
> > (XEN) sh error: sh_page_fault__shadow_3_guest_2(): disabled-APIC
> access:
> > not supported
> > (XEN) .domain_crash called from multi.c:3197
> > (XEN) Domain 3 (vcpu#0) crashed on cpu#1:
> > (XEN) ----[ Xen-3.0-unstable  x86_64  debug=y  Not tainted ]----
> > (XEN) CPU:    1
> > (XEN) RIP:    0008:[<000000008081d1ac>]
> > (XEN) RFLAGS: 0000000000010086   CONTEXT: hvm
> > (XEN) rax: 00000000ffd01000   rbx: 0000000000000000   rcx:
> > 00000000ffdff120
> > (XEN) rdx: 00000000fec00000   rsi: 00000000ffd03034   rdi:
> > 0000000000000000
> > (XEN) rbp: 0000000000060aa4   rsp: 0000000000060a84   r8:
> > 0000000000000000
> > (XEN) r9:  0000000000000000   r10: 0000000000000000   r11:
> > 0000000000000000
> > (XEN) r12: 0000000000000000   r13: 0000000000000000   r14:
> > 0000000000000000
> > (XEN) r15: 0000000000000000   cr0: 0000000080050033   cr4:
> > 0000000000000651
> > (XEN) cr3: 0000000000fd97a0   cr2: 0000000000000000
> > (XEN) ds: 0023   es: 0023   fs: 0030   gs: 0000   ss: 0010   cs:
0008
> > qemu-dm[5399]: segfault at 00002aaba96ec000 rip 00002ab993166fe0 rsp
> > 00007fffff912188 error 6
> >
> > (This is using a 64 bit host, but I see a similar crash with 32 and
> > 32p hosts).
> >
> > Just to confirm, 11161 in xen-unstable.hg is
> >
> > # HG changeset patch
> > # User chris@kneesaa.uk.xensource.com
> > # Date 1155812483 -3600
> > # Node ID 5c1021595e3ce923ff11e338172d4639da2aecb5
> > # Parent  5e3166de7120f474903d58bd74b7e2d12c3d8ea7
> > [qemu] Add the xen platform device to the qemu PCI bus.
> > Useful functionality will come later.
> >
> > Signed-off-by: Steven Smith <ssmith@xensource.com>
> > Signed-off-by: Christian Limpach
<Christian.Limpach@xensource.com>
> >
> >
> > and 11160 is
> >
> >
> > # HG changeset patch
> > # User ssmith@weybridge.uk.xensource.com
> > # Date 1155815205 -3600
> > # Node ID 5e3166de7120f474903d58bd74b7e2d12c3d8ea7
> > # Parent  24827cb8a94596987284c4b934365165468949a2
> > [NET] Fix ``rx->offset: 10, size: 4294967295''''
bug.
> >
> > Signed-off-by: Steven Smith <sos22@cam.ac.uk>
> >
> >
> > These are the csets you''re talking about, yes?
> >
> > Steven.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel