Question Crash caused by AV_nvlddmkm!unknown_function ?

[bacon_is_tasty]

Hello, since about a week ago my laptop has been randomly crashing. It's very random: sometimes it crashes a couple times in a day, sometimes it goes multiple days without crashing, and I have no idea what triggers it either. I analyzed some of the crash dumps with Windows Debugger and they seem to always be caused by the same reason: AV_nvlddmkm!unknown_function

Here are my computer specs:
Laptop model: OMEN by HP Laptop 16-c0001np
GPU: RTX 3050 Ti
OS: Windows 11 64-bit

Here are 5 recent crash dumps:
- January 15 2024
- January 22 2024
- January 24 2024
- January 24 2024 #2
- January 26 2024

 

[ubuysa]

In all cases the failing driver is nvlddmkm.sys, the Nvidia graphics driver. You can see in the context record from the dump what happened...

CONTEXT:  fffff106f2f35ca0 -- (.cxr 0xfffff106f2f35ca0)
rax=fffff8066d11b201 rbx=000000000018cbd0 rcx=ffffcb0297849000
rdx=ffffcb0297849008 rsi=000000000018bc30 rdi=ffffcb0297849008
rip=fffff8066d290de1 rsp=fffff106f2f366c0 rbp=0000000000000000
 r8=0000000000000000  r9=000000000018cbd0 r10=0000000000000000
r11=fffff106f2f36758 r12=0000000000000000 r13=0000000000000000
r14=ffffa782582f2044 r15=fffff8066d933f60
iopl=0         nv up ei pl nz na pe nc
cs=0010  ss=0000  ds=002b  es=002b  fs=0053  gs=002b             efl=00050202
nvlddmkm+0x670de1:
fffff806`6d290de1 448b3b          mov     r15d,dword ptr [rbx] ds:002b:00000000`0018cbd0=????????
Resetting default scope

The failing instruction (at the bottom) was a MOV instruction using the RBX register as a memory pointer. The ???????? next to the resulting virtual address indicates that this address is invalid (not allocated, paged out, or a bad RAM page). Most usually this is because the driver has messed up the buffer pointers, that seems to be what's happened here.

This could be a faulty driver, or it could be a faulty card. The version of nvlddmkm.sys that you have is very recent...

7: kd> lmvm nvlddmkm
Browse full module list
start             end                 module name
fffff806`6cc20000 fffff806`70540000   nvlddmkm T (no symbols)         
    Loaded symbol image file: nvlddmkm.sys
    Image path: nvlddmkm.sys
    Image name: nvlddmkm.sys
    Browse all global symbols  functions  data
    Timestamp:        Fri Jan 12 20:03:28 2024 (65A17EF0)
    CheckSum:         038126BD
    ImageSize:        03920000
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4
    Information from resource tables:

That may well be the problem. Laptops often use customised graphics drivers, for reasons of power saving, interoperability with graphics switching, and other reasons. The generic drivers from Nvidia and from AMD (for the iGPU) do not contain these customisations and thus may not work properly. I suggest that you uninstall the Nvidia and AMD graphics drivers and use the drivers from the HP website for your model. They may be old but they are guaranteed to be compatible with your laptop, and with each other.

 

[bacon_is_tasty]

I suggest that you uninstall the Nvidia and AMD graphics drivers and use the drivers from the HP website for your model. They may be old but they are guaranteed to be compatible with your laptop, and with each other.

I tried this by uninstalling the drivers w/ Safe Boot and DDU, and then reinstalling the drivers from the HP website, but after a few days, the issue has come back. Here is the new crash dump I just got. The problems seems to be the same.

 

[ubuysa]

Yep, the crash happened in nvlddmkm.sys...

10: kd> .frame /r 8
08 ffffd08f`dc6666f0 00000000`00000000     nvlddmkm+0x670c61
rax=fffff806926db101 rbx=000000000033a0a0 rcx=ffffb688c9799000
rdx=ffffb688c9799008 rsi=0000000000339a60 rdi=ffffb688c9799008
rip=fffff80692850c61 rsp=ffffd08fdc6666f0 rbp=0000000000000000
 r8=0000000000000000  r9=000000000033a0a0 r10=0000000000000000
r11=ffffd08fdc666788 r12=0000000000000000 r13=0000000000000000
r14=ffff80076767d054 r15=fffff80692ef2f40
iopl=0         nv up ei ng nz na pe nc
cs=0010  ss=0000  ds=002b  es=002b  fs=0053  gs=002b             efl=00040282
nvlddmkm+0x670c61:
fffff806`92850c61 448b3b          mov     r15d,dword ptr [rbx] ds:002b:00000000`0033a0a0=????????

Did you update the AMD Radeon graphics driver as well? In laptops the dGPU and iGPU typically cooperate, so you need to ensure that both drivers are updated back to the known working (HP) versions.

This might suggest that the problem may be with the graphics cartd. SInce the HP drivers BSOD then I'd suggest you update the AMD Radeon driver to the latest version via the AMD Driver and Support Tool. Then try the most recent three versions of the Nvidia driver as follows...

• Download the most recent three versions of the driver from the Nvidia website.
• Use DDU to uninstall the current driver
• Install the latest driver.
• If it BSODs use DDU to uninstall it and then install the immediately prior version of the driver.
• If it BSODs use DDU to uninstall it and then install the two level back version of the driver.

If it BSODs on all three of the most recent drivers, using DDU in between, then it's most likely that the graphics card is the problem. One way to test that is to disable the Nvidia card, reboot (so the drivers aren't loaded) and try it just using the AMD Radeon iGPU. Performance will be poor of course but if it doesn't BSOD then it's most likely the Nvidia card.

Note: When manually installing Nvidia drivers always select a 'Custom (Advanced)' install and uncheck everything except the driver. You only need the Nvidia Audio driver if you are sending audio over an HDMI connection. You only need the Nvidia USB-C driver if you have a monitor connected bia USB-C. You only need the PhysX driver if you have games require it. You really don't need GeForce experience at all...