Windows 10 gremlins: Reboots from bugcheck 0x000000d1

[Stereodude]

Well, I knew my happiness with my new Server/NAS wouldn't last very long. It seems to have gremlins. I don't know what's wrong with it, but I've got a bad feeling about it...

Sunday was a bugcheck:

The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1 (0xfffff801dfdab4f0, 0x0000000000000002, 0x0000000000000008, 0xfffff801dfdab4f0). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 2d3e6df3-2c02-4b02-92e5-acd5e14f441c.

Monday was a bugcheck:

The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1 (0xfffff800e65eb4f0, 0x0000000000000002, 0x0000000000000008, 0xfffff800e65eb4f0). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: d69de82a-6a68-4b47-bcd1-919de16369ba.

Yesterday was a bugcheck.

The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1 (0xfffff800f658b4f0, 0x0000000000000002, 0x0000000000000008, 0xfffff800f658b4f0). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 5cf67a7c-fa53-4ae9-b5e7-0d28fa275959.

Based on my Googling 0x0d1 is DRIVER_IRQL_NOT_LESS_OR_EQUAL. I'm running ECC RAM and I did memtest it extensively, so it seems unlikely I'm having random RAM corruption so severe to cause the issue not be flagged as an ECC error.

How do I figure out what is causing this? My suspicion is the video driver or Mellanox Connectx-3. The former I've found reported as being disabled due to problems in the device manager before and the latter caused me all sorts of problems when I tried an upgrade install from Windows 7 to 10. I can disable automatic reboot so it will sit at the blue screen and I can see what driver is causing the problem, but that isn't so convenient. Can I get any useful information from the memory dump? The latest one is 1.06gB.

 

[Stereodude]

So I followed the instructions found here and got a 32kB debuglog.txt file, but I'm not entirely sure what to make of it. However these lines lead me to think it's the driver for the ASPEED video "card" in the BMC.

MODULE_NAME: astkmd

IMAGE_NAME: astkmd.sys

PROCESS_NAME: dwm.exe

 

[Stereodude]

I was using the Win 8.1 x64 driver from the v1.01 driver package from ASPEED. I tried installing the driver for Win 8.1 x64 driver package, but Windows says that driver was disabled due to problems with it. They don't have a Windows 10 driver. The driver on ASRock's page for Server 2012r2 won't install. It's a .msi file and the install fails with a less than helpful message. I guess for now I'm going to use the default Microsoft Basic Display Adapter. I guess I could try the Server 2012r2 driver from either the v1.01 package or the v0.99 package. It had a different .sys file than the Win 8.1 x64 driver, so it may actually be different.

 

[Mercutio]

The basic display adapter will display true color in 24 bit. It just doesn't offer audio output support or hardware acceleration features, which is fine considering you're using Aspeed graphics in the first place.

 

[Stereodude]

The basic display adapter will display true color in 24 bit. It just doesn't offer audio output support or hardware acceleration features, which is fine considering you're using Aspeed graphics in the first place.

I mainly interact with the system over remote desktop so it doesn't really matter regardless. Random reboots aren't acceptable though. I have to say RDP Win10 to Win10 box is very nice though. Even over the Internet using the less than stellar upload via my cable modem.