Windows 10: driver_irql_not_less_or_equal (d1) (nvlddmkm.sys)

Hi


When I play demanding games I can get a bsod, the latest game being Rise of the Tomb Raider. This is the first game that bsods within 10 minutes no matter what. And so far this is what I can read out of the .dmp files : "driver_irql_not_less_or_equal (d1)" - (nvlddmkm.sys).

What happens is: without warning or fps drop the screen goes black (disconnects). Then the sound goes on for a bit, before it's cut goes into a loop. Then the PC restarts, or I have to push the restart button.
I have used the Display Driver Uninstaller 15.7.4.0 and reinstalled the Nvidia driver. Daemon tools have also been removed. Have stopped the Nvidia streamer service through msconfig too, just to see if it helped. No luck with any of the "fixes". I have run Furmark test without any issues, and a "sfc/ scannow". I've been monitoring the CPU and GPU temp before the crash, but no spikes in temperature.
*Side note - If I run Intel extreme tuning utility CPU stress test, the CPU reach almost 100 C in about 3 min. But that's under 100% load. But I guess that's unrelated, since the CPU don't get that load or temp during games. Still a indicator that something is wrong. And the Intel Diagnostic tool tells me there is something wrong with CPU Freq (Module - CPUFreq.exe did not create a results file.) , and it does not pass the stress test there as well. But I can't find anything about that in the .dmp file.


I'm not experienced enough to understand the rest. But so far by comparing similar problems others have had, I understand that Rise of the Tomb Raider are sending an invalid call/path to the Nvidia driver and makes it crash. correct me if i'm wrong.

Any tips on what to do?

---

ERROR: Module load completed but symbols could not be loaded for nvlddmkm.sys
Probably caused by : nvlddmkm.sys ( nvlddmkm+4d5acf )

DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
An attempt was made to access a pageable (or completely invalid) address at an
interrupt request level (IRQL) that is too high. This is usually
caused by drivers using improper addresses.
If kernel debugger is available get stack backtrace.
Arguments:
Arg1: 00000000004e7d75, memory referenced
Arg2: 0000000000000002, IRQL
Arg3: 0000000000000000, value 0 = read operation, 1 = write operation
Arg4: fffff800133d5acf, address which referenced memory

:)

 

Windows 10 Driver_irql_not_less_or_equal(tcpip.sys) BSOD

The DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1) error appears to be Driver Verifier initiated so I would discard those results for now.

There were two other errors, SYSTEM_SERVICE_EXCEPTION (3b) and KERNEL_SECURITY_CHECK_FAILURE (139), both appear to be suggesting an issue with the display adapter, such as :

BugCheck 133, {0, 501, 500, 0}

*** 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+ea687 )

I would suggest to uninstall the presently installed NVIDIA display driver and then install the latest version available from NVIDIA.

I would also suggest to update the BIOS to the latest version as the latest BIOS version (1.A) improves PCIE compatibility, etc.

If you need further help I might suggest to start a new topic to get the best help.

 

Phantom USB causing kernel mode bsod

The driver that appeared to cause the latest crash is the NVIDIA graphics driver:

BugCheck D1, {10, a, 1, fffff801d709449d}

*** 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+1a449d )

If this error persists consider using an older version of the driver.

Also, consider re-seating the graphic card.

 

What actually happens is a bit different from what you think and might be a bit above your head.
The Nvidia display drivers priority (IRQL) were too high when they attempted to access invalid memory (memory that is meant for the system only and nothing else), this memory can only be accessed at a minimum IRQL of 1 or 2 (not sure exactly) because IRQL 1 and 2 are for the system (Windows itself), higher is for hardware.
I won't go much more in depth of what is going wrong as that'll take some time to explain.

Please uninstall everything of Nvidia using Display Driver Uninstaller and install drivers from around december/november Nvidia. Be sure the clean install box is checked and only install the Graphics driver and the PhysX driver.
Attachment 64259
Code: 1: kd> lmvm nvlddmkm start end module name fffff800`12f00000 fffff800`13b24000 nvlddmkm T (no symbols) Loaded symbol image file: nvlddmkm.sys Image path: \SystemRoot\system32\DRIVERS\nvlddmkm.sys Image name: nvlddmkm.sys Timestamp: Sat Jan 23 01:25:59 2016 (56A2C897) CheckSum: 00BE5F37 ImageSize: 00C24000 Translations: 0000.04b0 0000.04e4 0409.04b0 0409.04e4[/quote]
Code: ******************************************************************************* * * * Bugcheck Analysis * * * ******************************************************************************* DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1) An attempt was made to access a pageable (or completely invalid) address at an interrupt request level (IRQL) that is too high. This is usually caused by drivers using improper addresses. If kernel debugger is available get stack backtrace. Arguments: Arg1: 00000000004e7d75, memory referenced Arg2: 0000000000000002, IRQL Arg3: 0000000000000000, value 0 = read operation, 1 = write operation Arg4: fffff800133d5acf, address which referenced memory[/quote]

 

Thanks! I'll try that.

 

Here is the new crash report. Should I try an even older driver? This one was from November. This is from the second crash after installing the new driver, the first crash did not create a .dmp file. Could it be the PSU?

 

Have you tried the December version?

 

Tried that now. And it crashed, but only the screen "disconnects". The sound does not go in a loop. And when I ctr.alt.del I can hear some windows sounds. So it's progress I guess. I can try one more time, to see if I get a real crash. This was with the 359.06 version, So I guess I can try the 361.43 next.

*edit. tried with the 361.43 and it's the same result.

*edit- so, maybe I have to go to another forum to get further help?