Windows XP crash error messages indicate nipalk to have problems

[frische]

Hi

I have hooked two 6009 USB DAQ units up to MacBook runinng Parallels desktop with windows XP inside. LabView 8.5 and DAQmx 8.6 in windows XP. Sometimes while a VI is running windows suddenly crashes. Inspecting the error messages in windows shows that right before the crash nipalk has some problems. Can anybody help with this?

[Posse]

Hi frische

I would like to investigate this, but I will need some more details about the errors that you are seeing. Please post all the information that may be relevant - and maybe some screenshots.

Best regards,

[frische]

Hi Peter

Thanks for your interest. I’m very new to Lab-View, so help from experienced people is really nice. I’m not a computer expert neither, rather an experienced user.

First of all, the system is as I wrote a new MacBook running lates MacOS with Parallels Desktop. This simulates (as you you probably know) a virtual machine, which is a standard Pentium machine with 1 GB RAM. On this machine Windows XP is installed and on this LabView 8.2 Full development system was first installed, then DAQmx 8.6. No other programmes are installed (except what comes with Windows and the MacOS)

With this configuration I experienced crashes of two kinds. Windows crash and LabView shut-down. LabView shut-down was experienced during the execution of a VI. After talking to NI-support in Denmark (thank you Dennis), I upgraded to LabView 8.5 and also updated the 8.6 DAQmx installation, redesigned the VI to a ”producer-consumer” design, and this seemed to have solved the LabView shutdown.

However, I still get the windows crash, which is an imediate shutdown of windows, no warnings or anything, just a blue screen with some text I can’t read before it’s gone and then it rebootes.

The MacOS doesn’t have problems.

I have attached:
- the VI
- the System event-log from windows since the 20th of august
The crash which made me write the mail occurred between 8/28/2007 8:44:29 AM and 8/28/2007 8:44:31 AM. You might want look at that first.
- mini-dumps from windows since the 20th of august

If you need more information or want any specific files, please tell me

[Jeremiah Cox]

Howdy frische,
I could not find the System Event Log attachment.  I notice that the minidumps are attached twice, perhaps it was lost during attachment?  We would like to see the nipalk event log entries that you mention.  Perhaps you could try re-attaching them or just pasting the text into another post?

Regarding the minidumps, they seem to be identical.  They all report a bugcheck in the Windows USB driver (bccode 0xfe BUGCODE_USB_DRIVER, bc_arg1 0x5).  The raw translation from Microsoft is "A hardware failure has occurred due to a bad physical address found in a hardware data structure. This is not due to a driver bug."  This is the first time I have observed this error.  It may be a Parallels issue, but it is challenging to determine more without either a more detailed dump or reproducing the issue in-house.

One option may be to enable "Kernel Memory Dumps" as explained in the Microsoft KB How To Enable Windows XP to Capture a Complete or Kernel Memory Dump.  One of these large "memory.dmp" files may be helpful, though they can take a while to process.

I hope that helps,
Jeremiah

Message Edited by Jeremiah Cox on 09-02-2007 07:41 PM

[frische]

Sorry about the wrong attachment, this one should be the event log. I'll try to see if I can make a large memory dump as you propose.

[frische]

I have made the kernel-dump. It is 60 MB, 15 when compressed. Do you want me to upload it or mail it to you directly?

[Jeremiah Cox]

Howdy frische,
Perhaps you could upload the kernel dump to ftp.ni.com/incoming ?
You can find more information at How Do I Transfer Files With FTP?

Looking at the Event Log, we can probably ignore the NIPALK "Warning"s -- it doesn't seem to like one of the PCI bridges in the Parallels environment, but that probably isn't the source of this issue.  Please open the NIPALK "Error"s, get the full text (these seem to be truncated) and repost them.

Thanks,
Jeremiah

Message Edited by Jeremiah Cox on 09-04-2007 11:31 AM

[frische]

Thank you, Jeremiah

 I don't know how to get the full text of the errors, but I have uploaded the zipped kernel-dump in the incoming folder as FrischeMemorydump.zip

[Jeremiah Cox]

Howdy frische,
I have discovered that the "Error" messages in the Event Log about the 600x are also harmless (they should disappear in a future NI-DAQ release).

Regarding the kernel dump, it continues to point the finger at USB hardware stack.  One of the bugcheck parameters identifies the USB host controller.  I generated Corrective Action Request "4D4CJVAW" to request that someone on the NI-DAQ R&D team look into this issue further.  I recommend that you contact your NI support channel by phone or email (www.ni.com/ask) to help them prioritize and check on the status of this request.

The Parallels virtualization software virtualizes the USB bus.  I think it could be helpful to contact Parallels Technical Support.  If the provide any leads or assistance, please share them with your NI support contacts and this forum.

Best regards,
Jeremiah

[Stuart_G]

frische

To follow up with what Jeremiah and others have stated.  Our group (DAQmx) has already done some testing of USB hardware using parallels and other virtualization software.  We did find issues witht the interaction of some (not all) of our USB devices and Parallels.  I believe the 600x was one of these devices.  What I found is if I boot into Windows on the Macintosh there were no issues but when I go through the Parallels interface there are issues like what you are seeing.  We have contacted Parallels about the issues we are seeing with their software.  Hopefully we, us, them will be able to get a resolution but I don't expect anything in the short term.

Stuart