6TO4 Adapter error for DP hot-replace test

NTDEV
1

Hi forks,

When I am doing Dynamic Partitioning (DP) hot-replace test for my Xen
para-virtual scsi miniport driver, SimpleIoStress test case print lots
of error said that ‘IcmpSendEcho() had too may consecutive timeouts.
Win32 = 11010 - Error due to lack of resources’ for DeviceID:
ROOT*6TOMP\0004, Name: 6TO4 Adapter. I just know that DP hot-replace
test will simulate a S4 state to try to take all devices in system to D3
state and then resume to D1, at the same time do some IO stress test to
check if all devices work fine. From log file, I find that
SimpleIoStress test for my scsi miniport adapter and generic volume
which adapter reports are pass but lots of errors for 6TO4 adapter.
Driver for 6TO4 adapter are provided by Microsoft and I checked that is
the latest one. Does anybody get the same error like me? Or is there
anything wrong with network environment configuration? I use DHCP to get
IP adress, and it seems that my company network environment has IPv6.

thanks
wayne

2

>

Hi forks,

When I am doing Dynamic Partitioning (DP) hot-replace test for my Xen
para-virtual scsi miniport driver, SimpleIoStress test case print lots
of error said that ‘IcmpSendEcho() had too may consecutive timeouts.
Win32 = 11010 - Error due to lack of resources’ for DeviceID:
ROOT*6TOMP\0004, Name: 6TO4 Adapter. I just know that DP hot-replace
test will simulate a S4 state to try to take all devices in system to
D3
state and then resume to D1, at the same time do some IO stress test
to
check if all devices work fine. From log file, I find that
SimpleIoStress test for my scsi miniport adapter and generic volume
which adapter reports are pass but lots of errors for 6TO4 adapter.
Driver for 6TO4 adapter are provided by Microsoft and I checked that
is
the latest one. Does anybody get the same error like me? Or is there
anything wrong with network environment configuration? I use DHCP to
get
IP adress, and it seems that my company network environment has IPv6.

Wayne,

Are you using SMP or UP? Some versions of Xen have some timer related
issues with some CPU’s where the TSC counters don’t synchronise
properly. Another symptom is things like pings reporting 1ms, -25000ms,
-25000ms, 1ms etc.

James

3

> Are you using SMP or UP? Some versions of Xen have some timer related

issues with some CPU’s where the TSC counters don’t synchronise
properly. Another symptom is things like pings reporting 1ms, -25000ms,
-25000ms, 1ms etc.

I set VCPU = 4, since DP hot-replace test need CPU >= 2. Any patch can
ignore that ?

thanks
wayne

4

If Xen provides a broken view of the TSC, you might try booting the system with /usepmtimer. I have had to do this in the past with other virtualization products that don’t provide a consistent TSC readout (symptoms were negative or erratic ping response times).

  • S

-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of James Harper
Sent: Monday, March 02, 2009 12:36 AM
To: Windows System Software Devs Interest List
Subject: RE: [ntdev] 6TO4 Adapter error for DP hot-replace test

Hi forks,

When I am doing Dynamic Partitioning (DP) hot-replace test for my Xen
para-virtual scsi miniport driver, SimpleIoStress test case print lots
of error said that ‘IcmpSendEcho() had too may consecutive timeouts.
Win32 = 11010 - Error due to lack of resources’ for DeviceID:
ROOT*6TOMP\0004, Name: 6TO4 Adapter. I just know that DP hot-replace
test will simulate a S4 state to try to take all devices in system to
D3
state and then resume to D1, at the same time do some IO stress test
to
check if all devices work fine. From log file, I find that
SimpleIoStress test for my scsi miniport adapter and generic volume
which adapter reports are pass but lots of errors for 6TO4 adapter.
Driver for 6TO4 adapter are provided by Microsoft and I checked that
is
the latest one. Does anybody get the same error like me? Or is there
anything wrong with network environment configuration? I use DHCP to
get
IP adress, and it seems that my company network environment has IPv6.

Wayne,

Are you using SMP or UP? Some versions of Xen have some timer related
issues with some CPU’s where the TSC counters don’t synchronise
properly. Another symptom is things like pings reporting 1ms, -25000ms,
-25000ms, 1ms etc.

James

NTDEV is sponsored by OSR

For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars

To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer

5

Skywing wrote:

If Xen provides a broken view of the TSC, you might try booting the system with /usepmtimer.

Does /userpmtimer affects windows server 2008?

thanks
wayne

6

I got it.

My DTM client and DTM server are running on a network environment with
IPv4 and IPv6. Windows server 2008 will try to use IPv6 with default
setting but server 2003 won’t. It looks like that some SimpleIoStress
test case will try to use IPv6 ping from DTM client to DTM server. After
I install IPv6 support and configure it properly. This case can pass.

thanks
wayne