Red Hat Bugzilla – Bug 489433
Blue screen on W2K8/64 while running WHQL "common scenario stress with IO" test on RHELSCSI
Last modified: 2010-11-09 07:41:14 EST
Description of problem:
Install windows 2003 guest with xen-pv win dirver and do Common Scenario Stress with IO,the windows will have no response.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Install the xenpv-win-32-bit according to the install manual
2.Do the stress testing according to http://msdn.microsoft.com/en-us/library/dd450391.aspx
The windows has no response after suspend/hibernate/wake many times testing RedHat PV NIC Driver
The windows has no response after suspend/hibernate/wake many times testing RHEL SCSI Bus Miniport .
The RedHat PV NIC Driver and RHEL SCSI Bus Miniport should can pass Common Scenario Stress with IO
I'll submit the log later.
Created attachment 335315 [details]
the testing log
This one is failed one to get WHQL unclassified certificate.
I will supply as much information as possible.
Proposed this to blocker. If this bug is resolved, we should be able to obtain WHQL (unclassified).
when doing PC NIC common scenario stress with I/O it can pass.
The scsi miniport bus will test pv nic driver and rhel scsi disk driver,it will fail and go to blue screen.
I will attach the minidump file after go to blue screen from win 2003/32bit.
does this happen with XP(32-bit)?
Created attachment 340483 [details]
the mindump file one
Created attachment 340484 [details]
minidump file two
I have tested with winxp/32 bit, when do the common scenario testing,it will disable the device,then the whole screen hang.the xp system has no response.
I will attach the screen shot.
Created attachment 340663 [details]
the common scenario xp/32 hang cmd dialogue
*** Bug 526400 has been marked as a duplicate of this bug. ***
*** Bug 526404 has been marked as a duplicate of this bug. ***
This is failing for both the RHELSCSI and RHELNET drivers; however, the latter's failure is probably just a consequence of RHELSCSI's.
The screenshot is yet another occurrence of bug 526393. We have to rerun everything after 526393 is fixed, as it look like it is a showstopper for any kind of stress testing.
I still get a PAGEFAULT_IN_NONPAGED_AREA after fixing bug 526393 while testing RHELNET common scenario stess with IO.
After the latest changes, RHELSCSI testing should be done with EnumeratingDevices = 3. In this case, "Common Scenario Stress With IO" fails without hanging the machine.
As this test is basically "Sleep Stress With IO" (which passes) + "Disable Enable With IO", and it fails in the Disable() step, I'm closing this as duplicate.
*** This bug has been marked as a duplicate of bug 526398 ***