Unexpected shutdowns involving dxgkrnl.sys, ntoskrnl.exe, nvlddmkm.sys

Edited by Milisav111, 27 January 2024 - 05:21 AM.

Hello and welcome to the forum!

The two dumps both point at the graphics driver or graphics card. One is a 0x116 bugcheck, which happens when a graphics hang occurs and the Windows Timeout Detection and Recovery (TDR) feature is unable to recover from the hang. The other is a 0x133, indicating that a collection of DPCs (the back-end of device interrupt processing) ran for too long. We really need the kernel dump to analyse this properly, but there are strong indications that the graphics driver is the cause here too.

The most likely cause is the driver rather than the card. The version of the graphics driver that you have installed is quite old, dating from November 2023...

8: kd> lmvm nvlddmkm
Browse full module list
start             end                 module name
fffff806`43dc0000 fffff806`476d9000   nvlddmkm T (no symbols)           
    Loaded symbol image file: nvlddmkm.sys
    Image path: \SystemRoot\System32\DriverStore\FileRepository\nv_dispig.inf_amd64_1fea8972dc2f0a69\nvlddmkm.sys
    Image name: nvlddmkm.sys
    Browse all global symbols  functions  data
    Timestamp:        Thu Nov  9 23:35:30 2023 (654D50A2)
    CheckSum:         03804E98
    ImageSize:        03919000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4
    Information from resource tables:

The Nvidia drivers website has a more recent driver for your 3090 dated January 2024, I would download and clean install that driver and see how things go.

If it still BSODs on the latest driver, run the SysnativeBSODCollectionApp again and upload a new zip file. I'll walk you through how to tell whether the card is at fault.
 

Edited by ubuysa, 27 January 2024 - 07:07 AM.

Hi, Welcome to BC.

Note : your system is running BIOS Version : 6.02  Released 2023/06/20 and the Current Release is : 8.05  Released : 2024/01/03

There is various CPU fixes, Ram compatibility, UEFI Security fix, Intel Microcode and other extremely important fixes that you have missed.

It is a common procedure to update the Bios these days but of course need to make sure you have no MS Updates coming in or other software updates and to turn off your A/V security.

Following proper protocols established in your Manual which is available here along with your BIOS and Motherboard Drivers you have possibly missed as well :

https://www.asrock.com/MB/Intel/Z790%20Pro%20RS%20WiFi/index.asp#Support

Please note that your : i7-13700KF supports Memory : Up to DDR5 5600  anything faster can and will void the Chip Warranty.   

So your Gskill ram @ 6400 is too fast and can create problems.  If so lower it in Bios to the above speed.

Yes the board can run what ever is on the above link but does not mean that the chip supports this.

As well I understand but am not to sure that Intel can tell if the chip is overclocked or ram is wrong speed which is a form of overclocking.

Please note if did not know there is a WHQL Win10 Driver for your ASUS VG249 Monitor

https://www.asus.com/ca-en/displays-desktops/monitors/tuf-gaming/tuf-gaming-vg249q/helpdesk_download?model2Name=TUF-Gaming-VG249Q

In future please supply the proper name of your :  Zotac GeForce RTX 3090  as there is the Trinity Edition a Trinity Edition OC, AMP Extreme Holo, RTX 3090 Ti, a 24 gig

These cards all have Support Pages on Zotac and the use the WHQL International Driver.  As to what Zotac modifies am not sure. ASUS releases Bios Updates as does MSI releases Firmware Updates

https://www.zotac.com/us/support as well if need to RMA the card info is available.

If you have anything like ASUS Fast Boot enabled in the BIOS suggest turn it off for the same reasons as below :

https://www.windowscentral.com/how-disable-windows-10-fast-startup /11 and WHY !!!

The Nvidia drivers website has a more recent driver for your 3090 dated January 2024, I would download and clean install that driver and see how things go.

Thank you for your replay, I updated drivers as you suggested.

So I'll keep you updated on what's happening.

Thank you for your replay, I was afraid to run BIOS update because I didn't know if system was stable so I wanted to avoid any more damage. 

I had no idea about the RAM speed, I'll decresse it immediately. 

I disabled fast startup, and installed monitor drivers.

GPU is Zotac gaming GeForce RTX 3090 trinity OC

And I'll carry on with BIOS update, and I'll post here if anything comes up.

It happened again today... Here are the files, it involve just ntoskrnl.exe now.

I did all the steps you suggested, and worked on pc normally. I think it's maybe happening when I run more than two apps at the same time.

Today i played Apex and talked on Discord with HWmonitor on and it shut down by itself.

So this covers Repair ,Net and DDU to remove your Video Drivers and then install the newer ones as has been suggested.  As Well Older Games DirectX 2010 for older games.  Note : it will only add drivers and will not copy over or replace newer files.

Retry Install after DDU.

Disabled Media Sources
-----------------------------------

DxDiagx86 shows : HKEY_LOCAL_MACHINE\Software\Classes\MediaFoundation\MediaSources\DoNotUse

No idea if this is a problem or not : it seems to be a various things to cause this error from : Install to Startup to .Net, DirectX and others.  Info gained from :

https://www.google.com/search?q=%22HKEY_LOCAL_MACHINE%5CSoftware%5CClasses%5CMediaFoundation%5CMediaSources%5CDoNotUse%22&client=firefox-b-d&sca_esv=fa2eff043ab95eb5&ei=zhi3ZfCsOsi20PEPqeeRwAo&ved=0ahUKEwjwy6-y0YGEAxVIGzQIHalzBKgQ4dUDCBA&oq=%22HKEY_LOCAL_MACHINE%5CSoftware%5CClasses%5CMediaFoundation%5CMediaSources%5CDoNotUse%22&gs_lp=Egxnd3Mtd2l6LXNlcnAiSyJIS0VZX0xPQ0FMX01BQ0hJTkVcU29mdHdhcmVcQ2xhc3Nlc1xNZWRpYUZvdW5kYXRpb25cTWVkaWFTb3VyY2VzXERvTm90VXNlIkgAUABYAHAAeACQAQCYAQCgAQCqAQC4AQzIAQD4AQHiAwQYACBB&sclient=gws-wiz-serp#ip=1

Sorry on the long URL.  No idea regarding the HKEY Registry Entry as to removal or change.

Could suggest :  None of these will be of harm. Useful as well.  This tool troubleshoots common issues with the setup of Microsoft .NET

https://www.microsoft.com/en-ca/download/details.aspx?id=30135

The DirextX repair is done by removing your Video Card Drivers using DDU (note this URL will be out dated so in future Google DDU and can take from this site or Wagnardsoft)

So follow the directions.

This is also a useful URL for DirectX : NOTE : Do Not Download Anything : https://www.partitionwizard.com/clone-disk/reinstall-directx.html  Note : Is for Info Only.

As well as the link for DirectX 2010 for older games ; https://www.microsoft.com/en-us/download/confirmation.aspx?id=8109

Edited by Pkshadow, 28 January 2024 - 10:42 PM.

Dump File 28th : FAILURE_BUCKET_ID:  0x1a_61941_PAGE_TABLE_RESERVED_BITS_SET_IMAGE_hardware_ram  

Possibly ram not set up properly when you changed the settings in BIOS.    Return to stock ram speed, no XMP.

Can deal further with this at another time.    So crash was ram settings.

Not at all sure need to run this but is good to know :

https://learn.microsoft.com/en-us/answers/questions/645477/bsod-memory-management-with-explorer-exe

Though in this it says faulty ram :

https://www.tenforums.com/bsod-crashes-debugging/176876-windows-10-just-told-me-found-problems-needed-restart.html

Return Ram to Stock Speed, No XMP. 

Did you test the ram as is requested in the Posting Instructions  ??

Anyway.... deal with ram after or if it continues.

@ubuysa will need to look at Stack but as pointed out if only Future Crashes with the same code/s after have returned ram to stock speed and no XMP.

Need to look through Manual I would think as there maybe more than one place have to lock the speed in.

Although the recent BSOD (28th Jan) is a 0x1a MEMORY_MANAGEMENT bugcheck I'm not convinced that this is caused by bad RAM. The argument 1 value in the dump data is 0x61941, indicating that the paging hierarchy is corrupted. Argument 2 contains the virtual address where the problem was found. Displaying the page table entry for that virtual address gives...

PXE at FFFFFDFEFF7FB8B0    PPE at FFFFFDFEFF716010    PDE at FFFFFDFEE2C02628    PTE at FFFFFDC5804C59C8
contains 0A000000048BD863  contains 0A0000013E09A863  contains 0A00000FFCAA4863  contains 0A00FFFBAD23CB7B
pfn 48bd      ---DA--KWEV  pfn 13e09a    ---DA--KWEV  pfn ffcaa4    ---DA--KWEV  pfn fffbad23c CG-DANTKWEV

The PFN of fffbad23c is indicating that the page is 'bad' and that could of course be because the RAM page holding it is bad. As a first step then, and as already suggested by @Pkshadow, disable any XMP profile you may have and run the RAM at stock frequencies. You should also test your RAM with Memtest86 (free). Run Memtest86 twice for a total of 8 iterations of the 13 different tests.

What makes me think that this may not be bad RAM is that this dump also happened during graphics operations and nvlddmkm.sys is again involved. It is possible that a bad driver (or a bad device managed by that driver) could cause this type of BSOD. I can see that you've updated nvlddmkm.sys to the latest version and that tends to point at the graphics card.

What I suggest you do first, is to remove and reseat the graphics card. Be sure that it's firmly seated. Try a different PCIe slot if you can. Also ensure that any extra power connector is fully home at both ends.

If it still BSODs then the only real way to check whether its the card or the driver is to do the following...

1. Download the three most recent versions of the driver (yes, I know you run at least two of these, but please do this).
2. Download DDU.
3. Uninstall the current driver with DDU (the system will reboot)
4. Install the most recent driver version (I know you've tried this but DDU may have made a difference)
5. If it still BSODs then uninstall the current driver with DDU (the system will reboot)
6. Install the one level back version of the driver (you may have already run this version but DDU may have made a difference)
7. If it still BSODs then uninstall the current driver with DDU (the system will reboot)
8. Install the two levels back version of the driver (you may have already run this version but DDU may have made a difference)

If it BSODs on all three of the most recent driver versions, installed with DDU in between, then the problem is most likely with the card. Sadly, given your F series CPU, the only way to confirm is with a swap test. If you get to this point though, before going to the expense and/or trouble of trying another graphics card, post back on here and I'll walk you through enabling Driver Verifier to check whether there is another flaky driver causing these BSODs.

Edited by ubuysa, 29 January 2024 - 04:09 AM.

Crash dmp analysis aside, i I may chime in with an observation;

In both of the OPs Sysnative collections the RAM is reported as the below;

 

G Skill Intl F5-6400J3239G32G 00000000 4800 MHz 32 GB BANK 0 Controller0-ChannelA-DIMM1 8 64

G Skill Intl F5-6400J3239G32G 00000000 4800 MHz 32 GB BANK 0 Controller1-ChannelA-DIMM1 8 64

 

Thank you to both @PhillPower2 and @ubuysa for detailed information.

I agree that you need to check the RAM is populated in the correct slots (the above suggests not as it seems to be operating in single channel mode).  The default SPD speed on this memory is 4800 MT/s @ 1.1v which will be with JEDEC standard timings, which is likely the safest option while trouble shooting.  This should be what the motherboard automatically sets itself to when XMP is disabled and any manual settings overrides are removed.  If or when stability is achieved you can re-enable XMP if required, which runs it at a much higher voltage of 1.4v at 6400 MT/s with more aggressive timings.  This might be a side issue, or it could be a cause of instability, it's hard to say, but it is worth checking anyway.

Agreed re the RAM slots. I'm embarrassed now that I didn't spot that.

Thank you to both @PhillPower2 and @ubuysa for detailed information.

Your welcome Pk