Solved MS14-045 (KB 2993651) not fixing kernel mode BSODs

[NotASpeckOfCere]

Win7 64-bit

Symptoms. Playing video or even Google maps BSODs computer. BlueScreenView tells me that dxgkrnl.sys is causing the problem (Bug Check Code 0x00000116). I've also had the circular re-boot until I turned off the screen-saver Electric Sheep (which plays MPEGs). If I play no videos, the machine is stable. I don't actually see the typical BSOD screen, not surprisingly as this is a Direct X kernel issue: it just goes black. If there is audio (I was watching a TED talk during the last crash), I could hear the speaker continue to talk before the machine reboots or I cold boot it, but the voice a lot slower and lower pitch.

Following this: MS14-045: Description of the security update for kernel-mode drivers: August 12, 2014

• I deleted the cache in safe-mode
• I did the registry hack looking for the fonts that qualified (found none) and deleted the font cache again
• I looked for and could not find the offending 4 updates. I didn't eyeball it, I searched using Control Panels / Programs and Features / Installed Updates / Search Installed Updates without finding any of them.
• I installed the newest KB 2993651 today
• I do not have the Windows Update errors / issues I read that others are having
• Event Viewer tells me that "The previous system shutdown at 12:18:05 PM on 8/30/2014 was unexpected"
• I've attached the 083014-37019-01.dmp from that unexpected shut down from my Windows\minidump folder

Let me know if you need anything else from me. I'm only here because I think I did all my research and have done everything I have found.

Thanks much,
Speck

 

[NoelDP]

I've asked for this to be moved to the BSOD section - but please do the following anyhow, as it may relate. I'm subscribed to the thread, so I'll be able to look at the results.

Please follow the Windows Update Posting Instructions and post the requested data
If the file is too large (8MB compressed), remove the older CBSPersist cab files until the final file is below the limit - you can always post them separately after zipping them. (the forum doesn't allow the upload of bare CAB files, for a number of reasons)

 

[NotASpeckOfCere]

Hi Noel,

This is done:

1. Ran the SURT until it completed
2. SFC scan reports: "Windows Resource Protection did not find any integrity violations."
3. CBS.zip attached (I had to remove all older CAB files to get it under 8MB)

Speck

 

[NoelDP]

There's no significant errors in anything there - although there is one minor (and probably irrelevant) error when installing the patch...

2014-08-30 09:08:16, Info                  CBS    Exec: Staging Package: Package_3_for_KB2993651~31bf3856ad364e35~amd64~~6.1.1.3, Update: 2993651-16_neutral_GDR, PinDeployment: amd64_aee96059509967ef1052dc59cf86dd0b_31bf3856ad364e35_6.1.7601.18577_none_8f34beace38719ad
2014-08-30 09:08:16, Info                  CBS    Exec: This is a PSF Package.  Job has been saved and we are returning to client.
2014-08-30 09:08:16, Info                  CSI    [EMAIL="0000002d@2014/8/30:16:08:16.996"]0000002d@2014/8/30:16:08:16.996[/EMAIL] CSI Transaction @0x18af350 destroyed
2014-08-30 09:08:17, Info                  CBS    Exec: DPX job state saved for one or more packages, aborting the staging and install of execution.
2014-08-30 09:08:17, Info                  CBS    Perf: Stage chain complete.
2014-08-30 09:08:17, Info                  CBS    Failed to stage execution chain. [HRESULT = 0x800f0816 - CBS_E_DPX_JOB_STATE_SAVED]
2014-08-30 09:08:17, Error                 CBS    Failed to process single phase execution. [HRESULT = 0x800f0816 - CBS_E_DPX_JOB_STATE_SAVED]
2014-08-30 09:08:17, Info                  CBS    [B]WER: Failure is not worth reporting[/B] [HRESULT = 0x800f0816 - CBS_E_DPX_JOB_STATE_SAVED]

I'll leave you in the capable hands of the BSOD guys when they can get to you.

 

[NotASpeckOfCere]

Thanks Noel.

Meanwhile, I will try to stress the system by playing videos while I'm doing other things to be sure I still have the problem after taking the steps you had me take.

 

[NotASpeckOfCere]

Update: I played videos for the better part of yesterday to duplicate the BSOD issues but nothing was happening. In the evening, I decided to turn back on the Electric Sheep screen saver, which plays video clips from its cache (while still playing videos with Windows Media Player) and within 30 minutes, I got a reboot (don't know if it BSODed as I was not in the room).

Event Viewer: "The computer has rebooted from a bugcheck. The bugcheck was: 0x00000116 (0xfffffa80157b82e0, 0xfffff8800fa907a0, 0xffffffffc000009a, 0x0000000000000004). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 083114-31028-01."

The MEMORY.DMP is almost 3GB so I can't post that here.

I have attached 083114-31028-01.zip with a dmp from the windows minidump folder.

 

[Arc]

Hi NotASpeckOfCere.

Click on the button below ....



It will download the DM log collector. Right click on the application and run as administrator. It will generate a .zip file on your desktop. Upload the .zip.
http://www.sevenforums.com/tutorials/9733-screenshots-files-upload-post-seven-forums.html
__________________________________________________

*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 116, {fffffa80157b82e0, fffff8800fa907a0, ffffffffc000009a, 4}

Unable to load image \SystemRoot\system32\DRIVERS\nvlddmkm.sys, Win32 error 0n2
*** WARNING: Unable to verify timestamp for nvlddmkm.sys
*** ERROR: Module load completed but symbols could not be loaded for nvlddmkm.sys
Probably caused by : nvlddmkm.sys ( nvlddmkm+9a57a0 )

Followup: MachineOwner
---------

3: kd> !analyze -v
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

VIDEO_TDR_FAILURE (116)
Attempt to reset the display driver and recover from timeout failed.
Arguments:
Arg1: fffffa80157b82e0, Optional pointer to internal TDR recovery context (TDR_RECOVERY_CONTEXT).
Arg2: fffff8800fa907a0, The pointer into responsible device driver module (e.g. owner tag).
Arg3: ffffffffc000009a, Optional error code (NTSTATUS) of the last failed operation.
Arg4: 0000000000000004, Optional internal context dependent data.

Debugging Details:
------------------


FAULTING_IP: 
nvlddmkm+9a57a0
fffff880`0fa907a0 48ff25f94eedff  jmp     qword ptr [nvlddmkm+0x87a6a0 (fffff880`0f9656a0)]

DEFAULT_BUCKET_ID:  GRAPHICS_DRIVER_TDR_FAULT

CUSTOMER_CRASH_COUNT:  1

BUGCHECK_STR:  0x116

PROCESS_NAME:  System

CURRENT_IRQL:  0

ANALYSIS_VERSION: 6.3.9600.16384 (debuggers(dbg).130821-1623) amd64fre

STACK_TEXT:  
fffff880`021f8b48 fffff880`042bc134 : 00000000`00000116 fffffa80`157b82e0 fffff880`0fa907a0 ffffffff`c000009a : nt!KeBugCheckEx
fffff880`021f8b50 fffff880`0428f867 : fffff880`0fa907a0 fffffa80`0ecb5000 00000000`00000000 ffffffff`c000009a : dxgkrnl!TdrBugcheckOnTimeout+0xec
fffff880`021f8b90 fffff880`042bbf43 : fffffa80`ffffd846 ffffffff`fffe7960 fffffa80`157b82e0 00000000`00000000 : dxgkrnl!DXGADAPTER::Reset+0x2a3
fffff880`021f8c40 fffff880`0438c03d : fffffa80`1f16cd50 00000000`00000080 00000000`00000000 fffffa80`0ea7b260 : dxgkrnl!TdrResetFromTimeout+0x23
fffff880`021f8cc0 fffff800`0396773a : 00000000`02b5aae2 fffffa80`0ea23760 fffffa80`0c774040 fffffa80`0ea23760 : dxgmms1!VidSchiWorkerThread+0x101
fffff880`021f8d00 fffff800`036bc8e6 : fffff800`03846e80 fffffa80`0ea23760 fffff800`03854cc0 00000000`00000000 : nt!PspSystemThreadStartup+0x5a
fffff880`021f8d40 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KxStartSystemThread+0x16


STACK_COMMAND:  .bugcheck ; kb

FOLLOWUP_IP: 
nvlddmkm+9a57a0
fffff880`0fa907a0 48ff25f94eedff  jmp     qword ptr [nvlddmkm+0x87a6a0 (fffff880`0f9656a0)]

SYMBOL_NAME:  nvlddmkm+9a57a0

FOLLOWUP_NAME:  MachineOwner

MODULE_NAME: nvlddmkm

IMAGE_NAME:  nvlddmkm.sys

DEBUG_FLR_IMAGE_TIMESTAMP:  53b4446a

FAILURE_BUCKET_ID:  X64_0x116_IMAGE_nvlddmkm.sys

BUCKET_ID:  X64_0x116_IMAGE_nvlddmkm.sys

ANALYSIS_SOURCE:  KM

FAILURE_ID_HASH_STRING:  km:x64_0x116_image_nvlddmkm.sys

FAILURE_ID_HASH:  {1f9e0448-3238-5868-3678-c8e526bb1edc}

Followup: MachineOwner
---------

3: kd> lmvm nvlddmkm
start             end                 module name
fffff880`0f0eb000 fffff880`0fd79000   nvlddmkm T (no symbols)           
    Loaded symbol image file: nvlddmkm.sys
    Image path: \SystemRoot\system32\DRIVERS\nvlddmkm.sys
    Image name: nvlddmkm.sys
    Timestamp:        Wed Jul 02 23:12:02 2014 (53B4446A)
    CheckSum:         00C4C7C9
    ImageSize:        00C8E000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

 

[NotASpeckOfCere]

Thanks Arc,

Please find the attached zip file that is the result of running the utility.

Speck

 

[Arc]

Hi NotASpeckOfCere.

This is a display/graphics related issue. And it started from the time you installed the 340.52 WHQL display driver.

Uninstall this display driver using Display Driver Uninstaller and install any earlier version that worked good for you.

:ar: Drivers | GeForce

View attachment 331553​

When done, stress test the Graphics Card using Furmark.

Video Card - Stress Test with Furmark
​

Let us know the result.

One more place needs attention, as it is a long term issue that never taken care of. The Atheros ethernet driver.

On a special note, it was a great job done to update the Atheros Ethernet driver. Extremely well done on it! You did it yourself?

 

[NotASpeckOfCere]

You know what? I think it's a hot card. Recent results:

1. Reverted driver to previous version from 340.52 WHQL to 337.88 WHQL. Result: still crashed when stressed with video and Electric Sheep.
2. Reverted to an even older driver: 335.23 (March 10th 2014). Same result, several times. I can reproduce it repeatedly within 15 minutes.
3. Ran FurMark fullscreen. The video adapter temp was 68C when it started and quickly (less than a minute) reached 100C. When it hit 102C, the box crashed. I did it again non-fullscreen and it reached 104C before crashing.
4. Right now just typing with no video, it's cruising at 70C.

I've attached a recent ZIP from the DM log collector that includes an incident from this morning.

Is it normal for a stressed GPU to get this hot?

Arc, I think I did update the Atherose driver, but if it was well one, it's probably because I had excellent instructions!

Cheers,
Speck.

PS: I still have my older video card, so if this is due to thermal events, I can swap it out. But if so, I wonder why it took ~6 months to start acting up like this.

 

[Arc]

1. Ran FurMark fullscreen. The video adapter temp was 68C when it started and quickly (less than a minute) reached 100C. When it hit 102C, the box crashed. I did it again non-fullscreen and it reached 104C before crashing.
2. Right now just typing with no video, it's cruising at 70C.

I've attached a recent ZIP from the DM log collector that includes an incident from this morning.

Is it normal for a stressed GPU to get this hot?

Let us see the system temperature in normal condition. Report us the heat of the computer after a couple of hours of your normal usage. Upload a screenshot of the summery tab of Speccy. Alternatively, you can publish a Speccy snapshot too: http://www.sevenforums.com/tutorials/311593-speccy-publish-snapshot-your-system-specs.html .

Apparently you have detected the issue; now it is to determine what exactly is there behind it.

 

[NotASpeckOfCere]

Hi all, this issue is solved.

A fan blade snapped off the video card, getting stuck in there and keeping the fan from turning. A card in the next slot had pins sticking out. At some point over the past few weeks, the pins got a hold of the fan blade. It is now running cool. It was 104C (219F) at peak and 70C (158F) at idle—now it's around 30C at idle and 70-80 when stressed with FurMark.

Resolution: see the attached zip containing a photo-essay of the fix (a Q-tip end super-glued to the video card at a strategic place to keep them apart).

Sorry for taking your time for what turned out to be a hardware failure. It started going down around the same time that MS issued that bad patch, so I assumed that was it. Next time I will check the hardware first.

Cheers,
Speck

 

[NoelDP]

Ye Gods!

No wonder you were having problems.

The important thing is that you have now identified and isolated the problem, and should be on the way to a permanent cure.

Good luck with it!