Red Hat Bugzilla – Bug 808654
[whql] [netkvm]NDISTest 6.5 job failed on windows 2008
Last modified: 2012-06-20 07:58:43 EDT
Created attachment 574120 [details]
Created attachment 574121 [details]
Created attachment 574123 [details]
For windows 2008 32,the job only passed one time on AD ENV but it always failed in WORKGROUP ENV.
This is a testblocker for netkvm whql test
what's more ,I use freshly installed image .
summary of the issue we hit
1.sometimes NDIS client BSOD ,dump referring to comment #0
2.sometimes NDIS client virtio-nit can not get ip after the job running for a while
3.sometimes job hang at "VLANSendRecv" sub-job ,in the DTM Studio ,the NDIS client shows in no-responsible status ,but ping DTM Studio in the guest works .
(In reply to comment #9)
> summary of the issue we hit
> 1.sometimes NDIS client BSOD ,dump referring to comment #0
> 2.sometimes NDIS client virtio-nit can not get ip after the job running for a
> 3.sometimes job hang at "VLANSendRecv" sub-job ,in the DTM Studio ,the NDIS
> client shows in no-responsible status ,but ping DTM Studio in the guest works .
another thing need to hightlight ,I found "virtio_ioport_write: unexpected address 0x13 value 0x0 " in a loop , and I tried to shutdown the vm ,guest hang at
I check the history of our tesing results ,-23 works fine ,so ,this should be a regression in -24 .
* Please indicate which jobs other than "VLANSendRecv" had issues (Please always when indicating failures in NDISTest 6.5 or NDISTest 6.0 - write which exact sub-test failed).
* ""virtio_ioport_write: unexpected address 0x13 value 0x0 " in a loop" - it means that transfer is stack (we send a packet to QEMU but didn't receive the indication that it was transfer). We specifically added such behavior (writing to virtio ISR register) - in order to get the indication of this situation on the host.
* What is the difference on the host from the previous runs?
* Also - https://bugzilla.redhat.com/show_bug.cgi?id=807967 - if this is a host issues and in any case we have problems after powermanagment events, I see how some of the tests will fail.
(In reply to comment #12)
> * Please indicate which jobs other than "VLANSendRecv" had issues (Please
> always when indicating failures in NDISTest 6.5 or NDISTest 6.0 - write which
> exact sub-test failed).
Today,I tried the bug on two guests windows 2k8 32/64 bits after downgrading qemu to 227.The results were following as below,windows 2008 64 bits hit 7E issue once and windows 2k8-32 hits comments 9.3 issue.Uploaded VLANSendRecv.html log and dump file to \\smamit.eng.lab.tlv.redhat.com\win-team\Public\QE\808654\6.5issue.
Any issues please let me know.
Hello Mike and Min,
Unfortunately I cannot see failure descriptions from CPK (i see only driver verifier logs on job logs) or from log in run000\VlanSendRecv.htm. If you know how to look at the specific failure description - please guid me how to do it.
Regarding crash dumps - there is clear indication from the printout of the MS test driver and our debug printout that transfer is stuck and that's why test driver is calling "int 3".
I will dig dipper into dump files to get more information.
What I need from you:
1. Could you describe what exact gusts are failing? All Win7\Win8? Is there difference between 32bit\64bit?
2. Is there anything new in host network configuration?
3. Are you running with vhost? If yes, is it possible to make a test run without it?
4. Is it possible to test build 24 on older host version?
Created attachment 578893 [details]
An error message of installing win2k8-64 driver
According to Yan, the problem should be fixed in build 27:
Bug report changed to ON_QA status by Errata System.
A QE request has been submitted for advisory RHBA-2011:12358-01
Considering that issue has been fixed on windows 2008 64 and 32 guests via virtio-win-prewhlq-0.1-27 and there was another bug for tracing windows 7 64 bits guest specially.QE think the bug was fixed now.
Thanks for developer's work.
Technical note added. If any revisions are required, please edit the "Technical Notes" field
accordingly. All revisions will be proofread by the Engineering Content Services team.
No Documentation needed.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.