Minidump

I am solving BSOD's that my client has had intermittenly over the last couple of months - the attached being the first minidump that I got on 8 July. After analyzing it with WinDBG, I decided to do some more analysis by using .reload;!analyze -v;r;kv;lmnt. However, I am still none the wiser. With this deeper analysis, I am meant to be able to tell what process, program or device is causing the BSOD. However, I can't find it! Can anyone help?

 

Here is the problem in more detail, and what I've tried. Hope it helps

The specs of the machine are:

3.00GHz Pentium D
2GB RAM
250GB Hard drive
ConRoe 1333 D667 mboard (ASRock)

Since June, I have had intermittent BSOD's. I kept a record of them

8 July - VISTA_DRIVER_FAULT

10 July - CODE_CORRUPTION

12 July - CODE_CORRUPTION

26 August - IRQ_NOT_LESS_OR_EQUAL

26 August - MEMORY_CORRUPTION (2)

26 August - SPECIAL_POOL_DETECTED_MEMORY_CORRUPTION

11 October - SYSTEM_THREAD_NOT_HANDLED

12 October - MEMORY_MANAGEMENT

17 October - PAGE_FAULT_IN_NONPAGED_AREA

27 October - KERNEL_DATA_IMAGE_ERROR

27 October - UNMOUNTABLE_BOOT_DEVICE
(The last two didn't create a minidump for some reason)

Things done so far, to no avail:

Memory modules taken out, one by one - alas, TuffTest, MemTest, Windows Memory Diagnostic say no errors.
New psu - didn't solve errors
Updating drivers - didn't solve anything
Updating Windows - didn't solve anything
Two OS reinstallations due to crashes and Vista Repairs not working, didnt solve anything

What is very odd is this:

I did a TuffTest disk test (Seek Test and Surface Analysis) on his machine and it came up with F004 errors - Sector not found. Then I put the disk in my test machine, same errors. Then I tried it in his machine again with 1 memory module at a time, same errors. However, after testing the disk test again, putting the 2 modules back, the test was error-free!!

Also, in the Event Log, about once an hour, I get "some processor performance power features have been disabled due to a known bad firmware problem. Please contact the manufacturer." I did this last Friday, and still no answer!!

My mate and I (he works with computers as well) both think it's now a motherboard problem, but as it's intermittent, we're struggling!!

 

A mate of mine who has worked with computers for 30 years looked at it as well - this is strange.

I initially thought it was memory problems - so, my mate put his wife's memory into my client's machine, and that worked fine, after several hours of disk testing, motherboard testing, CPU testing, memory testing. Still with her memory in it, he used a piece of software called Glaze, a graphics card tester, thus testing the motherboard and putting it under stress, plus he ran other apps as well, but the computer didn't cough up once.
He ordered some more memory from Crucial, put it in, and it has worked fine since lunchtime today (Friday), (with Glaze running as well).
So, he thinks that the person who built it must have used cheap and nasty, unbranded, memory

 

The computer hasn't coughed up a single BSOD since Friday, however, I did a TuffTest this morning, and it came up with F001 - Invalid Function Request - Bad Command, for both the Seek Test and the Surface Analysis (like it did before) (I did that test about 2 hours ago). Just now I did it again, but this time it was fine; what's going on?

 

It even appears when SP1 isn't installed, and now my client reports another BSOD 'kernel_datapage_input_error' - I've never seen that one before, and Windows didn't create a dump file!

 

Have you run a chkdsk /r on this drive at any point?

It sounds like you may have found one problem, but not resolve the other. It would appear that the filesystem needs repairing, or the drive is bad. This would explain why the dumps were not created either, due to the inability to write to the drive.

 

I've run a chkdsk, with no parameters, and Windows said there were no errors. Also, I ran a Drive Fitness Test, but the disk was reported to be fine

 

New motherboard

I have got my client a new motherboard, as everything else has been eliminated

 

Still got the same errors, although there's a slight modification to the Kernel error message that I got:

Performance power management features on processor0 are disabled due to a known firmware problem (16/1/09 12:03:37)

and at exactly the same time I got the same error message, but this time it was processor1!

Anyway, the only thing that didn't work after the new motherboard was installed, was the internet. I resolved that by uninstalling and installing ZoneAlarm. Then I installed the drivers from the motherboard CD.

The BSOD appeared this morning, IRQL_NOT_LESS_OR_EQUAL. Immediately after, I performed a TuffTest. I did all the Certification Tests, the only thing that failed was the Fixed Disk, with an error code of F100; however, I did a full Surface Analysis, which produced no errors! This test brings up these results more than once!

 

A new motherboard was fitted, and it's done the trick - no more problems

 

On Sunday he brought it back to me saying that it crashed again while updating AVG.
So, I booted up, it got stuck on the screen "Please insert CD or DVD".
So I fixed the boot sector with the Vista installation CD.
I then rebooted; got an error message (not a BSOD) but I couldn't read it as the computer restarted, then froze.
So the Vista CD went back in, but no disks were found!
I did a TuffTest, and a Surface Analysis yielded the error code F000.
The same with the Media Analysis and Repair Test.
So I am going to replace the disk.

He bought the computer from a 'Palicomp' on EBay in March 2008. He first called me in June, how could everything go wrong (almost everything) within a year?

 

My client phoned me yesterday with another problem:

The strange thing is that yesterday it refused to boot up for him, Vista said it couldn't detect a disk, but for me today, it booted fine!
He also got a strange message yesterday, which he didn't tell me about til today, "Security oprions failure"
No clues or errors in the Event Log
Ran a chkdsk - no errors found.
AVG, SpyBot - fine
For some unknown reason, Firefox couldn't connect to the Net, but IE would; I discovered ZA had been uninstalled, but Ron said it was on there before he brought it to me! I installed the latest version, and then put a green tick in the Allow boxes for FF, which solved that - don't know why it disappeared in the first place!
I found out that the error message 'The instruction at 0x768e03fd referenced memory at 0x768e03fd. The required data was not placed in the memory because of an I/O error status of 0x000000e' isn't related to memory!, as 000000 is an unknown address, and that it pionted to network connection problems.
I checked that the network driver was up-to-date, which it was, and that's how far I've got!

 

Hi -

Were you able to get post-SP1 Windows Updates in? I noticed that both the NT Kernel and DirectXG Kernel have Jan 19, 2008, timestamps on them - they are the original SP1 modules. Both have been updated several times since.

Code:

[font=lucida console]
81c0e000 81fb8000   nt       ntkrnlmp.exe Sat Jan 19 05:30:50 2008 (47918B0A)

8bec9000 8bf68000   dxgkrnl  dxgkrnl.sys  Sat Jan 19 05:36:36 2008 (47918C64)
[/font]

The Realtek Ethernet is old - has it been updated? Then there is COMODO - not good for Vista.

Code:

[font=lucida console]
[b]Realtek Ethernet[/b]
8bf87000 8bf96000   Rtlh86   Rtlh86.sys   Tue Sep 26 12:20:27 2006 (45190CFB) 

[b]COMODO[/b]
8ca6b000 8ca83000   cmdguard cmdguard.sys Wed May 21 16:35:39 2008 (4834414B) 
87e16000 87e2a000   inspect  inspect.sys  Wed May 21 16:35:27 2008 (4834413F) 
[/font]

.

I agree w/ your system friend of 30 years. Many strange pieces to a puzzle that don't seem to fit anywhere. Have you validated the Vista install (or re-validated) at the Microsoft Genuine Advantage site? I would suggest doing so as I have seen quite a few users in past few months with validation issues on systems that were already validated either by OEM Vista pre-install or via Genuine Advantage site. I myself had systems here w/ Vista Ultimate full retail version that had been activated and validated, only to learn they were somehow "in-validated". Like the users I mention, the systems showed Vista as "activated" in Control Panel/ System - but a call to MS found that activation status had changed. All fine now and no problems with any since re-validation.
Check site out - www.microsoft.com/genuine

Regards. . .

jcgriff2

.