[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] Xen-3.4.0 RC3 Test Report. Xen: #19590 & Xen0: #867


  • To: "'xen-devel@xxxxxxxxxxxxxxxxxxx'" <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Li, Haicheng" <haicheng.li@xxxxxxxxx>
  • Date: Sun, 10 May 2009 13:05:14 +0800
  • Accept-language: en-US
  • Acceptlanguage: en-US
  • Delivery-date: Sat, 09 May 2009 22:05:46 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcnRLObRNSxJ/Q2JQFa4yu3vfyS7Dg==
  • Thread-topic: Xen-3.4.0 RC3 Test Report. Xen: #19590 & Xen0: #867

All,

This is our test report for Xen-3.4.0 RC3. Our RC3 testing covers multiple 
Intel platforms (for both x86 and IPF); and includes full functionality testing 
(VT-x, VT-d, SR-IOV, Power-Management, RAS, NUMA, and Flexmigration), guest 
installation testing (6 guest OSes), guest OS bootup testing (~50 guest OSes), 
as well as stress testing (crashme & helltest). 

For x86 side, 6 new bugs were exposed in our RC3 testing, 10 old bugs from RC1 
have been fixed and verified, there are totally 14 bugs open with RC3. As a 
highlight, VT-d quality got a big improvement; and for PowerManagement, we've 
demoted bug 1451 "T state control failure" to low priority since it happens 
only when SMT is disabled.

For IPF side, basic VT-i testing got pass; however, we also found a new issue 
"SMP VTi domain might hang at start Init, if vcpus are >= pcpus", a patch for 
it from xiantao.zhang@xxxxxxxxx has been sent out to mailing list. 


Status Summary
============================================================================
Feature                         Result
------------------------------------    ---------------------------
VT-x (Ept+Vpid)                 Pass    
VT-x (Shadow)                   Pass
VT-d (Dmar+Intr+Qinval)         Bug (1)
VT-d (Dmar)                     Bug (3)
NUMA                            Pass
PowerMgmt                       Bug (2)
FlexMigration                   Pass
SR-IOV                          Pass
RAS (CPU offline)               Pass
PV driver                       Pass 
Others                          Bug (8)                         

Xen Info:
============================================================================
xen-changeset:   19590:f80cf52a4fb6
dom0-changeset:   867:978499ee4f39

ioemu git: 
commit 438f0453656e54f3f4db248fd280673992d5ad54
Author: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>
Date:   Wed Apr 29 16:28:37 2009 +0100

New Bugs (6):
=====================================================================
1. Fc6/win-XP guest can not reboot right after booting up.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1456
2. PCI configuration space header is corrupted after device pass-through.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1457
(update: the bug has been fixed and verified with staging ioemu commit 
5d767b7b3fac52336f59e5b40d8befa6b1909937.)
3. E1000e NIC failed in FC10 guest with MSI. 
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1458
4. Two onboard 82576 NICs assigned to HVM guest cannot work stable if use INTx 
interrupt.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1459
5. Linux guest panics if it has duplicated BDFs assigned through VT-d.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1460
6. Failed to install FC10.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1461

Fixed Bugs (10):
=====================================================================
1. If VT-d is enabled, x86_64 dom0 panics at booting.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1444
2. [vt-d] Dom0 will hang at booting if we hide device via pciback on some 64 
bit host.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1447
3. Xen/dom0 hangs at booting if we continuously strike the USB keyboard.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1450
4. RHEL5 and Winxp guest can not boot up on some IA32PAE platform.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1445
5. [vt-d] Can not detach the pass-through device from guest. 
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1446
6. [VT-d][MSI] VT-d NIC with MSI enabled does not work in Linux guest with 
kernel 2.6.29.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1443
7. [NUMA] Incorrect NUMA info is reported on PAE host.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1449
8. live migration of smp guest causes guest crash.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1439
9. Failed to start guest with qcow/qcow2 image against c/s 19349
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1430
10. Guest with more than 1G memory can not boot against C/s 19534 on some PAE 
host.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1448

Old Bugs (8):
=====================================================================
1. stubdom based guest hangs when assigning hdc to it.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1373
2. [stubdom]The xm save command hangs while saving <Domain-dm>.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1377
3. [stubdom] cannot restore stubdom based domain.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1378
4. stubdom based guest hangs at starting when using qcow image.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1372
5. [VT-d] failed to reassign some PCI-e NICs.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1379
6. Can not create qcow file with command qemu-img-xen
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1452
7. [ACPI] On some x86_64 platform, `echo mem > /sys/power/state` can not 
trigger Dom0 S3 until using Ctrl^C to terminate the command.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1453
8. T state control failed.
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1451


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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.