Bugzilla Bug 1586

Environment:
------------
Service Arch (ia32/ia32e/IA64): ia32/ia32e
Guest Arch (ia32/ia32e/IA64): 
Guest OS Type (Linux/Windows):
Change Set: 20990
Hardware: Gulftown-HEDT
Other:
xen-changeset:   20990:46bfb4a318e9

pvops git:
commit 0a61300806b741e5e7a5f96f36e4317fc0329737
Merge: 543318f... 7937752...
Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

ioemu git:
commit f3f98c2c705fbb773e88692e016d828a060793c8
Author: Ian Jackson <ian.jackson@eu.citrix.com>
Date:   Mon Mar 1 16:14:50 2010 +0000


Bug detailed description:
--------------------------
On Gulftown-HEDT, Dom0 S3 resume can not work with latest Xen c/s. It is not a
regression because we did not try Dom0 S3 on this platform before. (NHM-HEDT
can work)

"pci=nomsi" and "iommu=0" can not help on this case.

The serial log shows:
#############
vt-nhm2 login: [  374.661576] PM: Syncing filesystems ... done.
[  374.989073] PM: Preparing system for mem sleep
[  375.002533] Freezing user space processes ... (elapsed 0.00 seconds) done.
[  375.011862] Freezing remaining freezable tasks ... (elapsed 0.00 seconds)
done.
[  375.020015] PM: Entering mem sleep
[  375.023803] Suspending console(s) (use no_console_suspend to debug)
(XEN) Preparing system for ACPI S3 state.Freezing cpus ...
(XEN) Prepare to bring CPU1 down...
(XEN) deleting CPU 1
(XEN) CPU 1 is now offline
(XEN) CPU1 is down
(XEN) Prepare to bring CPU2 down...
(XEN) deleting CPU 2
(XEN) CPU 2 is now offline
(XEN) CPU2 is down
(XEN) Prepare to bring CPU3 down...
(XEN) deleting CPU 3
(XEN) CPU 3 is now offline
(XEN) CPU3 is down
(XEN) Prepare to bring CPU4 down...
(XEN) deleting CPU 4
(XEN) CPU 4 is now offline
(XEN) CPU4 is down
(XEN) Prepare to bring CPU5 down...
(XEN) deleting CPU 5
(XEN) CPU 5 is now offline
(XEN) CPU5 is down
(XEN) Prepare to bring CPU6 down...
(XEN) deleting CPU 6
(XEN) CPU 6 is now offline
(XEN) CPU6 is down
(XEN) Prepare to bring CPU7 down...
(XEN) deleting CPU 7
(XEN) CPU 7 is now offline
(XEN) CPU7 is down
(XEN) Prepare to bring CPU8 down...
(XEN) deleting CPU 8
(XEN) CPU 8 is now offline
(XEN) CPU8 is down
(XEN) Prepare to bring CPU9 down...
(XEN) deleting CPU 9
(XEN) CPU 9 is now offline
(XEN) CPU9 is down
(XEN) Prepare to bring CPU10 down...
(XEN) deleting CPU 10
(XEN) CPU 10 is now offline
(XEN) CPU10 is down
(XEN) Prepare to bring CPU11 down...
(XEN) deleting CPU 11
(XEN) CPU 11 is now offline
(XEN) CPU11 is down
(XEN) deleting CPU 0
(XEN) Entering ACPI S3 state.
#############

Verified with following c/s, the bug is fixed. But there are one new bug found
when intremap enabled. I will open a new bug for it.

xen-changeset:   21047:c5e1a0b720ba

pvops git:
commit 9b258d14a9f01b6be210aeaf55e39c51c48ba58e
Merge: bf3f745... f4be29f...
Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

ioemu git:
commit 435718fb726b80462b4028128301dcf92214cc6a
Author: Ian Jackson <ian.jackson@eu.citrix.com>
Date:   Thu Mar 18 16:48:03 2010 +0000