1

I keep getting a BSOD on Windows 10 on an almost daily basis at completely random moments. The stop code is usually PFN_LIST_CORRUPT, but looking at the last couple of minidumps I see both "PFN_LIST_CORRUPT", "CACHE_MANAGER" nd "SYSTEM_SERVICE_EXCEPTION" bug check strings. ntoskrnl.exe is always the culprit.

I ran the extended memory diagnostics test and that reported that no errors have been found.

I'm at a loss at what could be going on here. Any other avenues I could try?

Minidumps are available here: https://1drv.ms/f/s!Ao1fjlTRHZdilP9Npt2okkr9QinzGA

Bas
  • 187
  • 1
  • 10
  • Use windbg to analyse your dumps... or upload your dumps and give us a copy. ntoskrnl is often misidentified by simpler tools (e.g. bluescreenview) when it's a driver without its own process. – Bob Mar 28 '18 at 08:12
  • Thanks, I edited the post with a link to the minidump files. – Bas Mar 28 '18 at 08:18
  • 1
    Ok, consistently comes back as memory corruption. Go download a copy of [Memtest86+](http://www.memtest.org/), boot from it, and run it. Each "pass" is a full set of tests. Run as many passes as you are able to (e.g. overnight) - it will keep going forever but I find ~10 passes is reasonably good at detecting memory issues. – Bob Mar 28 '18 at 09:22
  • Thanks, I'll run it overnight. If it finds errors, does that mean replacing hardware, or is Memtest86+ able to mark sections as "bad" so that Windows ignores them? – Bas Mar 28 '18 at 10:03
  • Funnily enough, I've been here before and asked that exact question. If you can, get them replaced (especially if you have a valid warranty). Otherwise, [some answers](https://superuser.com/a/1297660/117590) have come up with ways to mark the bad memory. I should note that the first step might be to reseat them in case it's just a bad connection (but do that after the tests). – Bob Mar 28 '18 at 10:27
  • Just out of curiosity, have you already run thru `chkdsk` and/or `sfc /scannow`? Also per the first error you mention I found https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/bug-check-0x4e--pfn-list-corrupt so just curious if there was a Windows Update or a driver that got updated, etc. when this started to occur? If so, consider rolling it back to the previous version if possible and applicable as something simple to check. I assume you're past some of this already too but wanted to mention just in case so nothing simple is overlooked. – Vomit IT - Chunky Mess Style Mar 28 '18 at 12:52
  • Already ran chckdsk and sfc, yeah. As for updates or drivers.. who knows. This has been going on for a while now, I just now got fed up with it enough to look into the cause. Rolling back updates would mean rolling back months of updates, and that doesn't seem like a good idea either. – Bas Mar 28 '18 at 13:37
  • Would you mind disclosing the make and model of the hardware you are running Windows 10? Also, would you mind telling what version and build of Windows 10 you see when you run `winver`? Lastly, would you tell us what Antivirus and/or anti-malware type software you have installed on this OS? – Vomit IT - Chunky Mess Style Mar 29 '18 at 02:23
  • @PimpJuiceIT: I'm running version 1709, build 16299.19. Make and model of the hardware is a mixm, it's a PC I assembled myself. As for antivirus and malware, only Windows Defender. – Bas Apr 25 '18 at 19:46
  • @Bob Alright, I ran Memtest86+ for 10 hours and 23 minutes... it did thirteen passes and found zero errors. I'm at a loss. – Bas Apr 26 '18 at 08:12
  • Oof. I'm not really sure where to go from here... Actually. Do you run any antivirus software? Try disabling those for a start. While I wouldn't expect them to cause this particular error, they have been known to cause memory corruption. Also, if you can reliably reproduce it, try to reproduce in safe mode to see if you can confirm if some driver is causing it. – Bob Apr 26 '18 at 08:34
  • Hmm. Forgot to ask. Is this something that started recently? If yes, can you think of when it last worked and any changes in between (OS updates, driver updates, hardware changes)? Does it only happen under heavy use or if you just boot and leave the computer running on its own? – Bob Apr 26 '18 at 08:36
  • It happens pretty much all the time, sometime when the pc is idling, sometimes when it is under heavy load. I'm only running Windows Defender, disabling it makes no difference. The problem started a few months ago... I suppose I'll just do a reinstall and see if the problem persists, then start swapping out hardware one by one. – Bas Apr 26 '18 at 10:58
  • I have resolved to swapping out DIMMs, switching cables and ports, and finally switching out harddrives, and I have since determined that my SSD is the culprit. If I run stuff from an old HDD, everything is fine, if I run from the SSD, BSODs happen all the time. It's weird that both SMART and Chkdsk report everything as being fine, though... – Bas May 07 '18 at 07:11
  • The most important info initially required; 1) type, model, and firmware of the SSD? 2) desktop/laptop, bios ver? 3) Is the SSD on the native SATA port or USB? 4) Is the AHCI mode of the connected SATA port enabled(from the BIOS setting). And a copy of the SMART data table? – shanebush May 07 '18 at 07:48
  • It's a 250GB Samsung Evo, I have no idea about the firmware. Bios is AMI V1.3, 27-6-2008. Native SATA, I've tried it with both AHCI and IDE modes, neither work. I'll see if I can get a copy of the SMART table. However, I've since tried a brand new SSD, and it exhibits the exact same problem. So for some reason, my system suddenly decided to be incompatible with SSD drives. – Bas May 10 '18 at 09:26

1 Answers1

0

After parsing the dump file, it refers to FLTMGR.sys. The cause of the fltmgr.sys errors are mainly down to the file being corrupted or unreadable. SYS file errors are typically caused by faulty hardware or corrupt device driver files.

  1. Update Windows 10 and all your drivers.

  2. Temporarily remove your antivirus software

  3. Enter Safe Mode
  4. Check for faulty hardware
daidai
  • 347
  • 1
  • 4
  • I'm curious where you're seeing `fltmgr.sys`. It's a loaded module, but not identified as the faulting module by a cursory windbg analysis. – Bob Mar 28 '18 at 09:53
  • I'm way past that stage... I'm not analyzing minidumps before making sure all my drivers are up to date. – Bas Mar 28 '18 at 10:04