You, Yongkang
2007-Nov-12 09:50 UTC
[Xen-devel] VMX status report against cset #16366 & linux-xen #310
Hi all,
We add 10 VT-d testing cases into our nightly testing from this changeset. There
are 5 failed VT-d cases, which are related with PCI e100 NIC assignment. But
they can pass in changeset #16281.
We also verified Windows timer issue has been fixed by using timer_mode=1 in HVM
config file.
There are still 3 opened major issues.
New issues:
================================================1) Can not assign PCI e100 NIC
to guest
Fixed issues:
================================================1) Time latency is a negative
number on window guest when vcpu > 1
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=895
Old issues :
================================================1) Can not insert module
"xen-platform-pci.ko" in the 2.6.18 HVM Linux guest
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1101
2) Can not boot up win2000 guest on PAE
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1102
3) Can not install 32bit Fedora 7 with vcpu > 1
http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1084
Testing environments:
=============================================PAE
Dom0 OS Fedora5
CPU Paxville
Memory size 8G
IA32e
Dom0 OS RHEL4u3
CPU Clovertown
Memory size 8G
VT-d
Dom0 OS RHEL5 IA32e
CPU Dual Core
Details: (Some cases report fail, but can pass when manually retest)
=============================================VT-d test detail:
=============================================1.assign one pcie nic to one UP
Linux guest with vtd. PASS
2.assign one pcie nic to one SMP Linux guest with vtd. PASS
3.assign one pcie nic to one UP WinXP guest with vtd. PASS
4.assign one pci nic to one SMP Linux guest with vtd. FAIL
5.assign one pci nic to one UP WinXP guest with vtd. FAIL
6.assign one pci nic to one UP Linux guest with vtd. FAIL
7.scp a big file in Linux guest via the pci nic assigned with vt-d. FAIL
8.assign one pcie nic to one SMP WinXP guest with vtd. PASS
9.assign one pci nic to one SMP WinXP guest with vtd. FAIL
10.scp a big file in Linux guest via the pcie nic assigned with vt-d. PASS
Summary Test Report of Last Session
============================================== Total
Pass Fail NoResult Crash
==============================================vtd 10
5 5 0 0
==============================================vtd 10
5 5 0 0
:one_pcie_up_64 1 1 0 0 0
:one_pcie_smp_64 1 1 0 0 0
:one_pcie_up_xp_64 1 1 0 0 0
:one_pci_smp_64 1 0 1 0 0
:one_pci_up_xp_64 1 0 1 0 0
:one_pci_up_64 1 0 1 0 0
:one_pci_scp_64 1 0 1 0 0
:one_pcie_smp_xp_64 1 1 0 0 0
:one_pci_smp_xp_64 1 0 1 0 0
:one_pcie_scp_64 1 1 0 0 0
===============================================Total 10
5 5 0 0
Platform : PAE
Service OS : Fedora Core release 6 (Zod)
Hardware : paxville
Xen package: 16366:837f83225153
Date: Sat Nov 10 07:50:35 CST 2007
1, 2 PAE SMP VMX domains and 2 xenU domains coexist FAIL
2. one PAE SMP Linux VMX domain with memory 4G PASS
3. Live migration PASS
4. boot a pae on x86_64 xenU PASS
5. boot 4 VMX per processor at the same time PASS
6. boot up 1 winXP VMX and 1 linux PAE VMX PASS
7. Save and Restore FAIL
8. Single domain with single vcpu bind a CPU PASS
9. one PAE SMP Linux VMX domain with memory 1500M PASS
10.boot up two winXP per processor at the same time PASS
11.boot up one linux VMX with 4 vcpus PASS
12.boot up four linux VMX one by one PASS
13. Boot up VMX with acpi=1, vcpu=1 PASS
14. subset LTP test in RHEL4U2 PAE SMP VMX domain PASS
15. boot up the winxp image with vcpus=1 and acpi=1 PASS
16. boot up the vista image with vcpus=1 PASS
17, one IA32 UP ACPI Windows 2K3 VMX domain PASS
18, Boot Linux 2.6.23 base kernel in RHEL4U1 PAE SMP Linux VMX domain PASS
19, one IA32 SMP ACPI Windows 2K3 VMX domain PASS
20, one IA32 SMP ACPI Windows 2K VMX domain FAIL
21, one IA32 UP ACPI Windows 2K VMX domain FAIL
22, one IA32 SMP ACPI Windows XP VMX domain PASS
23. one IA32 UP NOACPI Windows 2K VMX domain FAIL
24. one IA32 UP NOACPI Windows 2K3 VMX domain PASS
25. one IA32 UP NOACPI Windows XP VMX domain PASS
26. kernel build in one linux VMX PASS
27. VBD and VNIF works on UP VMX domain PASS
28. VBD and VNIF works on SMP VMX domain NO Result
29, startx in dom0 PASS
Platform : PAE
Service OS : Fedora Core release 6 (Zod)
Hardware : paxville
Xen package: 16366:837f83225153
Date: Sat Nov 10 07:50:35 CST 2007
Summary Test Report of Last Session
====================================================================
Total Pass Fail NoResult Crash
====================================================================device_model
2 1 0 1 0
control_panel 12 10 2 0 0
Restart 1 1 0 0 0
gtest 14 11 3 0 0
====================================================================device_model
2 1 0 1 0
:pv_on_up_PAE_gPAE 1 1 0 0 0
:pv_on_smp_PAE_gPAE 1 0 0 1 0
control_panel 12 10 2 0 0
:XEN_4G_guest_PAE_gPAE 1 1 0 0 0
:XEN_four_vmx_xenu_seq_P 1 0 1 0 0
:XEN_LM_PAE_gPAE 1 1 0 0 0
:XEN_four_dguest_co_PAE_ 1 1 0 0 0
:XEN_linux_win_PAE_gPAE 1 1 0 0 0
:XEN_SR_PAE_gPAE 1 0 1 0 0
:XEN_vmx_vcpu_pin_PAE_gP 1 1 0 0 0
:XEN_1500M_guest_PAE_gPA 1 1 0 0 0
:XEN_256M_guest_PAE_gPAE 1 1 0 0 0
:XEN_two_winxp_PAE_gPAE 1 1 0 0 0
:XEN_four_sguest_seq_PAE 1 1 0 0 0
:XEN_vmx_4vcpu_PAE_gPAE 1 1 0 0 0
Restart 1 1 0 0 0
:GuestPAE_PAE_gPAE 1 1 0 0 0
gtest 14 11 3 0 0
:boot_up_acpi_PAE_gPAE 1 1 0 0 0
:ltp_nightly_PAE_gPAE 1 1 0 0 0
:boot_up_acpi_xp_PAE_gPA 1 1 0 0 0
:boot_up_vista_PAE_gPAE 1 1 0 0 0
:boot_base_kernel_PAE_gP 1 1 0 0 0
:boot_up_acpi_win2k3_PAE 1 1 0 0 0
:boot_smp_acpi_win2k3_PA 1 1 0 0 0
:boot_up_acpi_win2k_PAE_ 1 0 1 0 0
:boot_smp_acpi_win2k_PAE 1 0 1 0 0
:boot_smp_acpi_xp_PAE_gP 1 1 0 0 0
:boot_up_noacpi_win2k_PA 1 0 1 0 0
:boot_up_noacpi_win2k3_P 1 1 0 0 0
:boot_up_noacpi_xp_PAE_g 1 1 0 0 0
:kb_nightly_PAE_gPAE 1 1 0 0 0
====================================================================Total
29 23 5 1 0
Platform : x86_64
Service OS : Red Hat Enterprise Linux AS release 4 (Nahant Update 3)
Hardware : clovertown
Xen package: 16366:837f83225153
Date: Sat Nov 10 07:57:51 CST 2007
1, 2 ia32e SMP VMX domains and 2 xenU domains coexist FASS
2. one ia32e SMP Linux VMX domain with memory 4G PASS
3. Live migration PASS
4. one xenU domain with memory 256M PASS
5. boot 4 VMX per processor at the same time PASS
6. boot up 1 winXP VMX and 1 linux VMX PASS
7. Save and Restore PASS
8. Single domain with single vcpu bind a CPU PASS
9. one FC5 ia32e SMP Linux VMX domain with memory 1500M PASS
10. boot up two winXP per processor at the same time PASS
11. boot up one linux VMX with 4 vcpus PASS
12. boot up four linux VMX one by one PASS
13. one pae SMP Linux VMX domain with memory 1500M PASS
14. one pae SMP Linux VMX domain with memory 4G PASS
15. boot a pae on x86_64 xenU PASS
16. Boot up VMX with acpi=1, vcpu=1 PASS
17. subset LTP test in RHEL4U2 ia32e SMP VMX domain FASS
18. boot up the winxp image with vcpus=1 and acpi=1 PASS
19. boot up the vista image with vcpus=1 PASS
20, one IA32E UP ACPI Windows 2K3 VMX domain PASS
21, boot Linux 2.6.23 base kernel in ia32e SMP Linux VMX domain FASS
22, one IA32E SMP ACPI Windows 2K3 VMX domain PASS
23, one IA32E SMP ACPI Windows 2K VMX domain PASS
24, one IA32E UP ACPI Windows 2K VMX domain PASS
25, one IA32E SMP ACPI Windows XP VMX domain PASS
26. one IA32E UP NOACPI Windows 2K VMX domain PASS
27. one IA32E UP NOACPI Windows 2K3 VMX domain PASS
28. one IA32E UP NOACPI Windows XP VMX domain PASS
29. kernel build in one ia32elinux VMX PASS
30. VBD and VNIF works on UP VMX domain FAIL
31. VBD and VNIF works on SMP VMX domain FAIL
32, startx in dom0 PASS
Platform : x86_64
Service OS : Red Hat Enterprise Linux AS release 4 (Nahant Update 3)
Hardware : clovertown
Xen package: 16366:837f83225153
Date: Sat Nov 10 07:57:51 CST 2007
Summary Test Report of Last Session
====================================================================
Total Pass Fail NoResult Crash
====================================================================device_model
2 0 2 0 0
control_panel 17 17 0 0 0
Restart 2 2 0 0 0
gtest 13 13 0 0 0
====================================================================device_model
2 0 2 0 0
:pv_on_smp_64_g64 1 0 1 0 0
:pv_on_up_64_g64 1 0 1 0 0
control_panel 17 17 0 0 0
:XEN_1500M_guest_64_g64 1 1 0 0 0
:XEN_256M_xenu_64_gPAE 1 1 0 0 0
:XEN_vmx_4vcpu_64_g64 1 1 0 0 0
:XEN_1500M_guest_64_gPAE 1 1 0 0 0
:XEN_4G_guest_64_gPAE 1 1 0 0 0
:XEN_256M_guest_64_g64 1 1 0 0 0
:XEN_SR_64_g64 1 1 0 0 0
:XEN_four_sguest_seq_64_ 1 1 0 0 0
:XEN_vmx_vcpu_pin_64_g64 1 1 0 0 0
:XEN_linux_win_64_g64 1 1 0 0 0
:XEN_256M_guest_64_gPAE 1 1 0 0 0
:XEN_LM_64_g64 1 1 0 0 0
:XEN_two_winxp_64_g64 1 1 0 0 0
:XEN_four_vmx_xenu_seq_6 1 1 0 0 0
:XEN_four_sguest_seq_64_ 1 1 0 0 0
:XEN_4G_guest_64_g64 1 1 0 0 0
:XEN_four_dguest_co_64_g 1 1 0 0 0
Restart 2 2 0 0 0
:Guest64_64_gPAE 1 1 0 0 0
:GuestPAE_64_g64 1 1 0 0 0
gtest 13 13 0 0 0
:boot_up_acpi_win2k3_64_ 1 1 0 0 0
:boot_smp_acpi_xp_64_g64 1 1 0 0 0
:bootx_64_g64 1 1 0 0 0
:boot_smp_acpi_win2k_64_ 1 1 0 0 0
:boot_up_vista_64_g64 1 1 0 0 0
:boot_up_noacpi_win2k_64 1 1 0 0 0
:boot_base_kernel_64_g64 1 1 0 0 0
:boot_up_acpi_win2k_64_g 1 1 0 0 0
:boot_up_acpi_xp_64_g64 1 1 0 0 0
:boot_smp_acpi_win2k3_64 1 1 0 0 0
:ltp_nightly_64_g64 1 1 0 0 0
:boot_up_acpi_64_g64 1 1 0 0 0
:kb_nightly_64_g64 1 1 0 0 0
====================================================================Total
34 32 2 0 0
Best Regards,
Yongkang You
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
You, Yongkang
2007-Nov-13 09:40 UTC
[Xen-devel] RE: [New Update] VMX status report against cset #16366
Hi Keir,>1) Can not insert module "xen-platform-pci.ko" in the 2.6.18We found this case is not the issue of PV drivers. It is a wrong reporting by case. But I found it is because another bug. In (and before) xen-unstable #16157, 64bit Dom0 can do a chroot into a 32bit root file system. But since #16281, when doing chroot from 64 Dom0 to 32bit root file system, it would report "Segmentation Fault". I tried with Native 64bit Linux, this operation is okay. So I think it is a new issue. Could you double check? The reproduce steps, 1. /dev/sda1 is 64bit Dom0 system; sda2 is a 32bit Linux system. 2. Boot to sda1 Dom0. 3. mount /dev/sda2 /mnt/sda2 4. chroot /mnt/sda2 Best Regards, Yongkang You>-----Original Message----- >From: xen-devel-bounces@lists.xensource.com >[mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of >You, Yongkang >Sent: Monday, November 12, 2007 5:51 PM >To: xen-devel >Subject: [Xen-devel] VMX status report against cset #16366 & >linux-xen #310 > >Hi all, > >We add 10 VT-d testing cases into our nightly testing from >this changeset. There are 5 failed VT-d cases, which are >related with PCI e100 NIC assignment. But they can pass in >changeset #16281. > >We also verified Windows timer issue has been fixed by using >timer_mode=1 in HVM config file. > >There are still 3 opened major issues. > >New issues: >================================================>1) Can not assign PCI e100 NIC to guest > >Fixed issues: >================================================>1) Time latency is a negative number on window guest when vcpu > 1 > http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=895 > >Old issues : >================================================>1) Can not insert module "xen-platform-pci.ko" in the 2.6.18 >HVM Linux guest > http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1101 >2) Can not boot up win2000 guest on PAE > http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1102 >3) Can not install 32bit Fedora 7 with vcpu > 1 > http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1084 >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Cui, Dexuan
2007-Nov-15 10:28 UTC
[Xen-devel] 32-bit Xen disabled buffered IO for HVM guest'' VGA. Why?
You, Yongkang wrote:> Old issues : > ================================================> can not boot up win2000 guest on PAE > http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1102The bug was reported in recent VMX status report. It''s caused by not buffering the VGA accesses of the guest. Long ago, we had no any buffering for VGA, and we had a bug "Win2k doesn''t boot" (it''s said it''s due to the slow MMIO happening in W2K''s ISR -- makes the ISR not work well); Later on, changeset 10993: dc7b56b8cfb5 (Add HVM buffered IO writes framework to accelerate stdvga speed. With this patch, win2k VMX guest boots fine on xen) fixed the bug. Recently changeset 16240: cae485f682aa (x86, hvm: Improve standard VGA performance) was introduced to further improve VGA performance. However, changeset 16254: 63bfed7b662b disabled the buffering in 32-bit Xen using conditional compilation, saying "it doesn''t work" in 32-bit hypervisor. I don''t know the reason for this. However, when c/s 16240 was just checked in, I saw a strange crash on execution of the 2nd memset() in stdvga_init() on IA32pae Xen. Anybody also noticed this? If this is a bug introduced by c/s 16240, we should fix it ASAP, because: Now, in 32-bit Xen, we have no any buffering at all, causing the bug 1102; and if you use grub''s splashimage, you would find nobody can tolerate the rather poor VGA performance.... -- Dexuan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2007-Nov-15 14:38 UTC
[Xen-devel] Re: 32-bit Xen disabled buffered IO for HVM guest'' VGA. Why?
On 15/11/07 10:28, "Cui, Dexuan" <dexuan.cui@intel.com> wrote:> If this is a bug introduced by c/s 16240, we should fix it ASAP, > because: > Now, in 32-bit Xen, we have no any buffering at all, causing the bug > 1102; and if you use grub''s splashimage, you would find nobody can > tolerate the rather poor VGA performance....It''s on my todo list. It''s not hard to fix this. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Cui, Dexuan
2007-Nov-16 11:52 UTC
[Xen-devel] RE: 32-bit Xen disabled buffered IO for HVM guest'' VGA. Why?
Keir Fraser wrote:> On 15/11/07 10:28, "Cui, Dexuan" <dexuan.cui@intel.com> wrote: > >> If this is a bug introduced by c/s 16240, we should fix it ASAP, >> because: Now, in 32-bit Xen, we have no any buffering at all, >> causing the bug 1102; and if you use grub''s splashimage, you would >> find nobody can tolerate the rather poor VGA performance.... > > It''s on my todo list. It''s not hard to fix this.Well, I attached a patch. Please apply it. Thanks! -- Dexuan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2007-Nov-16 13:59 UTC
[Xen-devel] Re: 32-bit Xen disabled buffered IO for HVM guest'' VGA. Why?
That''ll limit you to about 14 HVM domains, which isn''t great. I''ll put in a patch later today. -- Keir On 16/11/07 11:52, "Cui, Dexuan" <dexuan.cui@intel.com> wrote:> Keir Fraser wrote: >> On 15/11/07 10:28, "Cui, Dexuan" <dexuan.cui@intel.com> wrote: >> >>> If this is a bug introduced by c/s 16240, we should fix it ASAP, >>> because: Now, in 32-bit Xen, we have no any buffering at all, >>> causing the bug 1102; and if you use grub''s splashimage, you would >>> find nobody can tolerate the rather poor VGA performance.... >> >> It''s on my todo list. It''s not hard to fix this. > > Well, I attached a patch. Please apply it. > > Thanks! > -- Dexuan_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel