I recently got a Windows 10 64bit machine, and today I got a system crash shortly after booting, with the “IRQL not less or equal” stopcode. On reboot, I got a “system service exception” crash with dxgkrnl.sys as the culprit twice in a row, before the OS auto-repaired and finally booted. I presume the first BSoD happened during CIS’ auto-update, because it wouldn’t be the first time it happened, and like the other times, after the crash the database version would show as “1” and “never updated”, and I’d have to redownload it. I thought this and other similar BSoDs were caused by a bad HDD, but after replacing it and performing a clean, blank-slate install it seems that this particular one is still happening.
I’ll provide the minidump in the attached files. Password is “kawagoe”.
I think Liosant refers to dxgkrnl.sys which is related to the graphics card driver installed on your system which may cause the intermittent BSOD.
Updating the graphics card driver might solve the BSOD.
Trying to find ways to see if CIS is involved in the BSOD . . .
Did you get your Windows 10 64bit machine with CIS pre-installed?
Have you used your machine for some time, with the latest non-optional drivers for your graphics card installed, without having CIS installed? Did the BSOD occur then too?
Would the BSOD also happen when you uninstall CIS completely? Could you possibly try that out for some time?
It didn’t come with it, but after installing/updating drivers CIS was the first thing I installed.
Still, the crash didn’t happen every time I’d try to update CIS, only at random. I finished this new installation over a week ago and this is the first time it happened since. But like I said, it probably is involved because the AV database gets deleted in the process.
Do I understand it correctly that the BSOD happens during PC boot process and also after PC boot process when you start CIS update manually?
It could be just an unlucky moment in time during PC boot process when the BSOD occurs just when CIS is busy accessing/updating the AV database that corrupts the database.
To be able to analyze this BOSD better I would suggest to try uninstalling CIS and use your machine for some time to check if it runs stable and without getting any BSOD. There are just so many reasons why a BSOD occurs.
No, it happened first during the auto-update (and that’s likely when the database got corrupted) and then twice on boot with dxgkrnl.sys. After which it auto-repaired and booted normally. In fact, later CIS detected the first crash and asked to send a report.
The only other info I can provide is I had trouble loading a Youtube video on Firefox seconds before the first crash happened.
That sounds to me that there might be a close relationship with the video card driver.
How about trying to downgrade your video card driver for one or more versions and leaving CIS installed?
In my experience the latest video card drivers are not always the best choice, older versions may run more stable then newer ones.
In a manner of speaking, it already is. I set the AMD updater to only install the latest stable version instead of the optional updates. Said version does date back to May 15, but.
Well, that’s just it, I’m unable to reproduce it. Unless another database update rolls out maybe.
I rebooted at least once since and it didn’t happen again.
I suppose. It’s just that my previous machine (Win7 32bit) never got a single BSoD in 10+ years of using it, and just ignoring one on the new one seems like an entirely alien concept now.
Well, turns out I got another ntoskrnl.exe crash, this time right before a reboot and with the “kernel security check failure” stopcode… and CIS wasn’t involved at all this time.
I suppose at this point it’s no longer a topic that concerns this forum, but if someone with in-depth knowledge of Windows’ kernel-layer could share their thoughts, I’d still greatly appreciate it.
