thr3ads.net - Xen devel - [Xen-devel] vmx status report against cset 14912

If this information is useful, please help other people find it:
Share via:

Zhao, Yunfeng

2007-Apr-25 08:51 UTC

[Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

New issues (0)
================================================No new issue.

Fixed issue 
================================================1) Can not detect more than 4G
memory in guest OS
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=971

 Old issues (4):
 ================================================ 1) IA32E/PAE: 32bit Vista RTM
network doesn''t work.
 http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=840
 
 2)time latency is a negative number on window guest when vcpu > 1
 http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=895
 You can reproduce it on SMP windows, it will not happen on UP Windows

 3) Xen0 hang sometimes & a lot of calltraces will be printed on serial
 http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=953

4) HVM VBD doesn''t work well on SMP Linux guest
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=831


Testing environments 
=============================================PAE	
Dom0 OS			Fedora5
CPU		Conroe
Chipset 	 	Intel 965(broadwater) 
Memory size		2G

IA32e	
Dom0 OS			RHEL4u3
CPU		Woodcrest
Chipset 		Intel 5000(Blackford)
Memory size		12G

Details:
=============================================PAE:
+ Build&Boot PAE SMP xen0 without problem
+ Can bootup SMP XenU with  256 M Memory
+ Can bootup unmodified RHEL4U1 IA32 nonPAE SMP 2.6 Linux VMX domain
+ Can bootup unmodified RHEL4U1/RHEL4U2/FC5 PAE SMP 2.6 Linux VMX domain
+ Can bootup RHEL4U1/RHEL4U2/FC5 PAE SMP 2.6 Linux VMX domain & XenU
with different memory size
  (VMX:64MB,128MB,196MB,256MB,512MB,1500MB.XenU:256MB)
+ Can bootup RHEL4U1 PAE SMP Linux VMX domain with base kernel
+ xm-test runs well
+ A selected subset of LTP runs well in RHEL4U2 PAE SMP 2.6 Linux VMX
domain 
+ Can build base kenerl 2.6.16 in RHEL4U1 PAE SMP VMX domain.
+ Can startX in Dom0
+ Can bootup FC5 PAE SMP 2.6 Linux in VMX with acpi enable
+ Network is OK in RHEL4U1/RHEL4U2/FC5 PAE SMP VMX domain
+ Can startX in RHEL4U2 PAE SMP VMX domain
+ Can bootup unmodified IA32 UP nonACPI Windows2K VMX domain 
+ Can bootup unmodified IA32 UP ACPI WindowXP SP2 VMX domain
+ Can bootup unmodified PAE UP ACPI Windows2K3 VMX domain 
+ Can bootup unmodified PAE UP ACPI Windows Beta 2 Vista VMX domain 
+ Can bootup unmodified IA32 SMP Win2K3 VMX domain
+ Can bootup unmodified IA32 SMP WinXP VMX domain
+ Can bootup unmodified IA32 SMP Windows Beta 2 Vista VMX domain
+ Can bootup unmodified PAE SMP Windows RTM Vista VMX domains(network
does not work)   
+ VBD and VNIF work on UP and SMP VMX domain can not work
+ Can not xm save/restore work

IA32E: 
+ Build&Boot IA32E SMP xen0 without problem
+ Can bootup SMP XenU with  256 M Memory
+ Can bootup unmodified FC5 IA32 nonPAE SMP 2.6 Linux VMX domain
+ Can bootup unmodified FC5 PAE SMP 2.6 Linux VMX domain
+ Can bootup unmodified FC5/RHEL4U1/RHEL4U3 IA32E SMP 2.6 Linux in VMX
+ Can bootup FC5/RHEL4U1/RHEL4U3 IA32E SMP VMX domain & xenU with
different memory size.
  (VMX:128MB,256MB,384MB,512MB,1024M,4096MB.XenU:256MB)
+ Can bootup four FC5 IA32E SMP Linux VMX domain in sequence
+ xm-test runs well
+ A selected subset of LTP runs well in RHEL4U1 IA32E SMP 2.6 Linux VMX
domain 
+ Can build base kenerl 2.6.16 in RHEL4U1 IA32E SMP VMX domain.
+ Can boot up FC3 Linux VMX domain with base kernel
+ Can startX in Dom0
+ Can bootup FC5 IA32E 2.6 Linux VMX domain with acpi enable
+ Can startX in RHEL4U3 Ia32E SMP VMX domain
+ Network is OK in FC5/RHEL4U1/RHEL4U3 IA32E SMP VMX domain
+ Can bootup unmodified IA32 UP ACPI WindowXP SP2 VMX domain
+ Can bootup unmodified IA32 UP ACPI Windows 2K3 VMX domain   
+ Can bootup unmodified IA32 SMP Win2K3 VMX domain
+ Can bootup unmodified IA32 SMP WinXP VMX domain
+ Can bootup unmodified PAE SMP ACPI Windows Beta 2 Vista VMX domain   
+ Can bootup unmodified PAE SMP ACPI Windows RTM Vista VMX
domain(network does not work)   
+ Can bootup unmodified IA32E UP WinXP VMX domain
+ Can bootup unmodified IA32E UP Win2K3 VMX domain
+ Can bootup unmodified IA32E SMP Win2K3 VMX domain
+ Can bootup unmodified IA32E SMP WinXP VMX domain
+ Can bootup unmodified IA32E SMP Windows Beta 2 Vista VMX domain
+ Can bootup unmodified IA32E SMP Windows RTM Vista VMX domain
+ Can bootup >4G IA32E SMP Window RTM Vista VMX domain
+ Can bootup >4G IA32E SMP Linux VMX domain
+ VBD and VNIF work on UP and SMP VMX domain can not work
+ Can not xm save/restore work
+ Can 32-bits PAE xenU boot on 64-bits xen0

For details
------------------------------------------------------------
Platform: PAE 
Service OS: FC5, PAE, SMP
Hardware: Conroe
Default guest OS: PAE,SMP, 2.6 kernel
 
1, one xenU domain with memory 256M
PASS
2, one IA32 FC5 nonPAE SMP 2.6 VMX domain with memory 256M	PASS
3, one FC5 PAE SMP Linux VMX domain with memory 512M		PASS
4, one FC5 PAE SMP Linux VMX domain with memory 1500M		PASS
5, one RHEL4u1 Linux VMX domain boot with PAE SMP base kernel(2.6.16)
PASS
6, one FC5 PAE SMP Linux VMX domain with acpi enable		PASS
7, 2 FC5 PAE SMP VMX domains and 2 xenU domains coexist		PASS
9, network in xenu domain
PASS
10, network in RHEL4U1/RHEL4U2/FC5 PAE SMP Linux VMX domain	PASS
11, Startx in RHEL4U2 PAE SMP Linux VMX domain			PASS
12, startx in dom0
PASS
13, xm-test
PASS
14, subset LTP test in RHEL4U2 PAE SMP VMX domain
PASS
15, build Linux 2.6.16 base kernel in RHEL4U1 PAE SMP Linux VMX
domainPASS
16, one IA32 nonACPI Window VMX domain 				PASS
17, IA32 UP nonACPI 2K and XPsp2 Windows VMX domains coexist	 PASS
18, one IA32 UP ACPI Windows 2K VMX domain
PASS
19, one IA32 UP ACPI Windows XP sp2 VMX domain			PASS
20, one IA32 UP ACPI Windows 2K3 VMX domain			PASS
21, one IA32 UP ACPI Windows Beta2 Vista VMX domain
PASS
22, one IA32 SMP Windows 2K3 VMX domain				PASS
23, one IA32 SMP Windows XP VMX domain				PASS
24, one IA32 SMP Windows Beta 2 Vista VMX domain
PASS
25, one IA32 SMP Windows RTM Vista VMX domain			PASS
26. VBD and VNIF works on UP VMX domain			PASS
27. Save and Restore
PASS

Platform : PAE
Service OS : Fedora Core release 5 (Bordeaux)
Hardware : Conroe
Xen package: 14907:ee56c8d8864b
Date: Wed Apr 25 07:42:06 CST 2007

               Summary Test Report of Last Session
====================================================================            
Total   Pass    Fail    NoResult   Crash
====================================================================nightly     
3       3       0         0        0
device_model           2       1       0         1        0
control_panel          19      18      1         0        0
====================================================================nightly     
3       3       0         0        0
 :kernel_build         1       1       0         0        0
 :subset_LTP           1       1       0         0        0
 :dom0x11              1       1       0         0        0
device_model           2       1       0         1        0
 :pv_on_up             1       1       0         0        0
 :pv_on_smp            1       0       0         1        0
control_panel          19      18      1         0        0
 :acpi_nightly         1       1       0         0        0
 :ia32up_acpi_win2k3   1       1       0         0        0
 :ia32up_two_noacpi_   1       1       0         0        0
 :ia32_smp_win2k3      1       1       0         0        0
 :ia32up_acpi_vista    1       1       0         0        0
 :ia32_smp_winxp       1       1       0         0        0
 :ia32up_noacpi_win    1       1       0         0        0
 :ia32up_acpi_winxp    1       1       0         0        0
 :ia32_guest_256M      1       1       0         0        0
 :one_VMX              1       1       0         0        0
 :base_kernel          1       1       0         0        0
 :ia32_vistartm        1       0       1         0        0
 :ia32pae_VMX_1500M    1       1       0         0        0
 :one_XENU             1       1       0         0        0
 :ia32_smp_vista       1       1       0         0        0
 :two_xenu_vmx_seq     1       1       0         0        0
 :guestx11             1       1       0         0        0
 :ia32up_acpi_win2k    1       1       0         0        0
 :saverestore          1       1       0         0        0
====================================================================Total       
24      22      1         1        0
                
Platform: IA32E
service os: RHEL4u3. IA32E. SMP
hardware: Woodcrest
default guest os: IA32E.SMP. 2.6 kernel

1. one xenU domain with memory 256M
PASS
2. one FC5 IA32 nonPAE SMP 2.6 Linux VMX domain with memory 256M
PASS
3. one FC5 PAE SMP Linux VMX domain with 1500M			PASS
4. one FC5 PAE SMP Linux VMX domain with 5120M memory		PASS
5. one FC5 IA32E SMP Linux VMX domain with 512M memory 		PASS
6. one FC5 IA32E SMP Linux VMX domain with 5120M memory		PASS
7. Two 5120M FC5 IA32E Linux VMX domain coexist			PASS
8. one FC3 Linux VMX boot with IA32E SMP base kernel(2.6.16)
PASS	
9. one FC5 IA32E SMP Linux VMX domain with acpi enable		PASS
10. 2 FC5 IA32E SMP Linux VMX domains and 2 xenU VMX domains coexist
PASS
11. Four FC5 IA32E SMP Linux VMX domains coexist
PASS
12. FC3/FC5/RHEL4u1/RHEL4u3 network in IA32E SMP Linux VMX domain
PASS
13. network in XenU domain
PASS
14. Startx in RHEL4U3 IA32E SMP Linux VMX domain
PASS
15. StartX in Dom0
PASS
16. xm-test
PASS
17. subset LTP in RHEL4U1 IA32E Linux VMX domain
PASS
18. Build base kernel in RHEL4U1 IA32E Linux VMX domain		PASS
19. one IA32 UP nonACPI Windows VMX domain
PASS
20. one IA32 UP noACPI 2K3 and XPsp2 Windows VMX domains coexist PASS
21. one IA32 UP ACPI Windows 2K VMX domain
PASS
22. one IA32 UP ACPI Windows XP VMX domain
PASS
23. one IA32 UP ACPI Windows 2K3 VMX domain			PASS
25. one IA32 SMP Windows XP VMX domain				PASS
26. one IA32 SMP Windows 2K3 VMX domain				PASS
27. one IA32 SMP Windows Beta 2 Vista VMX domain
PASS
28.one IA32 SMP Windows RTM Vista VMX domain			PASS
29. one IA32E UP Windows XP VMX domain				PASS
30. one IA32E UP Windows 2K3 VMX domain				PASS
31. one IA32E UP Windows Vista VMX domain
PASS
32. one IA32E SMP Windows XP VMX domain				PASS
33. one IA32E SMP Windows 2K3 VMX domain
PASS
34. one IA32E SMP Windows Beta 2 Vista VMX domain
PASS
35. one IA32E SMP Windows RTM Vista VMX domain 			PASS
36. one IA32E SMP Windows Vista with 5120 M memory
PASS
37. Two 5120M IA32 SMP Win2K3 VMX domain and
	IA32E Linux VMX domain coexist				PASS
38. VBD and VNIF works on UP VMX domain			PASS
39. Save / Restore
PASS
40. boot 32-bits PAE xenU on 64-bits				PASS 


Platform : x86_64
Service OS : Red Hat Enterprise Linux AS release 4 (Nahant Update 3)
Hardware : Woodcrest
Xen package: 14907:ee56c8d8864b
Date: Wed Apr 25 08:13:00 CST 2007

               Summary Test Report of Last Session
====================================================================            
Total   Pass    Fail    NoResult   Crash
====================================================================device_model
2       2       0         0        0
nightly                3       2       1         0        0
control_panel          30      29      1         0        0
====================================================================device_model
2       2       0         0        0
 :pv_on_up             1       1       0         0        0
 :pv_on_smp            1       1       0         0        0
nightly                3       2       1         0        0
 :kernel_build         1       1       0         0        0
 :subset_LTP           1       1       0         0        0
 :dom0x11              1       0       1         0        0
control_panel          30      29      1         0        0
 :ia32e_vistartm       1       1       0         0        0
 :acpi_nightly         1       1       0         0        0
 :ia32up_acpi_win2k3   1       1       0         0        0
 :ia32up_two_noacpi_   1       1       0         0        0
 :fourG_2linux         1       1       0         0        0
 :ia32_smp_win2k3      1       1       0         0        0
 :ia32_smp_winxp       1       1       0         0        0
 :ia32up_noacpi_win    1       1       0         0        0
 :fourG_linuxwindows   1       1       0         0        0
 :four_vmx_seq         1       1       0         0        0
 :ia32up_acpi_winxp    1       1       0         0        0
 :ia32e_smp_winxp      1       1       0         0        0
 :ia32e_up_win2k3      1       1       0         0        0
 :ia32_guest_256M      1       1       0         0        0
 :one_VMX              1       1       0         0        0
 :base_kernel          1       1       0         0        0
 :ia32e_up_winxp       1       1       0         0        0
 :ia32_vistartm        1       0       1         0        0
 :ia32pae_VMX_1500M    1       1       0         0        0
 :fourG_paelinux       1       1       0         0        0
 :ia32e_smp_win2k3     1       1       0         0        0
 :fourG_linux          1       1       0         0        0
 :one_XENU             1       1       0         0        0
 :fourG_windows        1       1       0         0        0
 :ia32_smp_vista       1       1       0         0        0
 :two_xenu_vmx_seq     1       1       0         0        0
 :guestx11             1       1       0         0        0
 :ia32up_acpi_win2k    1       1       0         0        0
 :saverestore          1       1       0         0        0
 :ia32e_smp_vista      1       1       0         0        0
====================================================================Total       
35      33      2         0        0

Yunfeng

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

Keir Fraser

2007-Apr-25 08:56 UTC

head link

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

On 25/4/07 09:51, "Zhao, Yunfeng" <yunfeng.zhao@intel.com>
wrote:
>  3) Xen0 hang sometimes & a lot of calltraces will be printed on serial
>  http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=953
You still see this one? Does it happen often?
> 4) HVM VBD doesn''t work well on SMP Linux guest
> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=831
To be clear: this is an unmodified RHEL4 release kernel, with *latest* PV
drivers from -unstable? Any idea if this issue is specific to the RHEL4
kernel, or can you reproduce with other kernels? Does it very definitely
seem to be specific to multiprocessor guests?

 -- Keir


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

Zhao, Yunfeng

2007-Apr-25 09:27 UTC

head link

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

>>  3) Xen0 hang sometimes & a lot of calltraces will be printed on
serial>>  http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=953
>
>You still see this one? Does it happen often?[Yunfeng] The last time we saw this issue is about 2 weeks
ago.>
>> 4) HVM VBD doesn''t work well on SMP Linux guest
>> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=831
>
>To be clear: this is an unmodified RHEL4 release kernel, with *latest*
PV>drivers from -unstable? Any idea if this issue is specific to the RHEL4
>kernel, or can you reproduce with other kernels? Does it very
definitely>seem to be specific to multiprocessor guests?[Yunfeng] The guest was booted with a 2.16.16 smp kernel which we built
for PV test. It only happens on SMP HVM linux guest.I also saw the same
issue on some SLES SMP HVM guests with SLES release kernel and PV
drivers.

Thanks
YUnfeng

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

Keir Fraser

2007-Apr-25 09:34 UTC

head link

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

On 25/4/07 10:27, "Zhao, Yunfeng" <yunfeng.zhao@intel.com>
wrote:
>> To be clear: this is an unmodified RHEL4 release kernel, with *latest*
> PV
>> drivers from -unstable? Any idea if this issue is specific to the RHEL4
>> kernel, or can you reproduce with other kernels? Does it very
> definitely
>> seem to be specific to multiprocessor guests?
> [Yunfeng] The guest was booted with a 2.16.16 smp kernel which we built
> for PV test. It only happens on SMP HVM linux guest.I also saw the same
> issue on some SLES SMP HVM guests with SLES release kernel and PV
> drivers.
It would be interesting to see what happens if the platform-pci interrupt is
pinned strictly to vcpu0. So, what happens if you boot the guest with
''acpi=0''? Or, what about if you kill any irqbalance programs
that are
running after boot, then ''cat /proc/interrupts'' to find the
xen-platform-pci
interrupt number (x) and then ''echo 1
>/proc/irq/<x>/smp_affinity''?

I couldn''t repro just with SpecJBB but it''s quite possible the
kernel build
is really required to repro. Specjbb itself does very little I/O.

 -- Keir


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

Zhao, Yunfeng

2007-Apr-25 09:44 UTC

head link

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

>> [Yunfeng] The guest was booted with a 2.16.16 smp kernel which we
built>> for PV test. It only happens on SMP HVM linux guest.I also saw the
same>> issue on some SLES SMP HVM guests with SLES release kernel and PV
>> drivers.
>
>It would be interesting to see what happens if the platform-pci
interrupt is>pinned strictly to vcpu0. So, what happens if you boot the guest with
>''acpi=0''? Or, what about if you kill any irqbalance
programs that are
>running after boot, then ''cat /proc/interrupts'' to find
the
xen-platform-pci>interrupt number (x) and then ''echo 1
>/proc/irq/<x>/smp_affinity''?
>
>I couldn''t repro just with SpecJBB but it''s quite possible
the kernel
build>is really required to repro. Specjbb itself does very little I/O.
[Yunfeng] Ok, we will try this, and send out the result.

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

You, Yongkang

2007-Apr-25 11:09 UTC

head link

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues

>> [Yunfeng] The guest was booted with a 2.16.16 smp kernel which we built
>> for PV test. It only happens on SMP HVM linux guest.I also saw the same
>> issue on some SLES SMP HVM guests with SLES release kernel and PV
>> drivers.
>
>It would be interesting to see what happens if the platform-pci interrupt is
>pinned strictly to vcpu0. So, what happens if you boot the guest with
>''acpi=0''? Or, what about if you kill any irqbalance
programs that are
>running after boot, then ''cat /proc/interrupts'' to find
the xen-platform-pci
>interrupt number (x) and then ''echo 1
>/proc/irq/<x>/smp_affinity''?
>
>I couldn''t repro just with SpecJBB but it''s quite possible
the kernel build
>is really required to repro. Specjbb itself does very little I/O.
>
Hi Keir, 

Which kernel and guest did you use? I meet issues in 2.6.16 kernel (I build a
new kernel for the guest). Both UP and SMP has VBD problem in 64bit system. But
IA32 kernel (2.6.16) is okay.

I tried your suggestion way in SMP system. I notice after doing kernel build for
a while, the interrupts number won''t increase any more, and the kernel
build command is hung there.

When I tried UP 64bit guest, the command hang when doing mount VBD disk.

Would it be the issue of guest kernel? I rebuild a 2.6.16 Linux kernel and used
it in RHEL4U3.

Best Regards,
Yongkang (Kangkang) 永康

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

Keir Fraser

2007-Apr-25 11:27 UTC

head link

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues

On 25/4/07 12:09, "You, Yongkang" <yongkang.you@intel.com>
wrote:
> Which kernel and guest did you use? I meet issues in 2.6.16 kernel (I build
a
> new kernel for the guest). Both UP and SMP has VBD problem in 64bit system.
> But IA32 kernel (2.6.16) is okay.
> 
> I tried your suggestion way in SMP system. I notice after doing kernel
build
> for a while, the interrupts number won''t increase any more, and
the kernel
> build command is hung there.
> 
> When I tried UP 64bit guest, the command hang when doing mount VBD disk.
> 
> Would it be the issue of guest kernel? I rebuild a 2.6.16 Linux kernel and
> used it in RHEL4U3.
Well it sounds 64-bit guest specific doesn''t it. I tested with a 32-bit
guest. Also I did not do a kernel build, just specjbb.

 -- Keir


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

You, Yongkang

2007-Apr-25 15:19 UTC

head link

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues

>> Would it be the issue of guest kernel? I rebuild a 2.6.16 Linux kernel
and
>> used it in RHEL4U3.
>
>Well it sounds 64-bit guest specific doesn''t it. I tested with a
32-bit
>guest. Also I did not do a kernel build, just specjbb.
>
Yes, Keir. I didn''t found any issue in 32-bit guest with 2.6.16 kernel.

BTW, I just want to try VBD in Fedora Core 5 (2.6.15). But meeting a lot of PV
driver compile issue.

Best Regards,
Yongkang (Kangkang) 永康

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

Xen devel - Apr 2007 - vmx status report against cset 14912 - 1 fixed issue, 4 old issues

[Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixed issue, 4 old issues

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues

Re: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues

RE: [Xen-devel] vmx status report against cset 14912 - 1 fixedissue, 4 old issues