Question PC is riddled with different BSODs ?

[johnbl]

people are just doing a hard reset on the laptops. the equivalent for a desktop is to just turn off the power. it will force a full reboot.
it would be better to turn off virtual memory to delete the pagefile then doing a full reboot and turn on virtual memory to create a new pagefile.sys.

i was looking at a memory.dmp file. it includes a lot more debug info and internal logs. it generally shows a lot of errors the people do not notice. deleting the pagefile helps since often some of the problems take a long time to actually cause windows to fail.
note: you can make a registry setting to make windows delete the pagefile on shutdown. this helps for some slow memory leaks, also if you do not do this and you fix the problem the leaks will still be there and will be misleading if someone looks at the memory.dmp file. just fyi

hardware installs that do not complete can have strange system symptoms. Cases where windows plug and play detect that some driver is needed for the hardware, tries to install but can not complete, the installer exits, then the whole process starts over again in a endless loop (every second or so). there are conditions that can hang cpu cores and bugcheck the system.

usb can be very screwy with old drivers and updated usb standards updates.

I think the reason people use the disconnect the charger and hold down the power button is because of windows plug and play.
they do not know they should: start cmd.exe as an admin then run
net.exe stop "plug and play"

before they try to update or modify certain drivers. if they don't do this then they uninstall a driver, windows plug and play detects that it needs the driver and reinstalls it 1 second later and they are back to where they started.

windows control panel device manager has a poor User interface, it does not show the version of the drivers when you install. all the drivers for one hardware device might look identical but they are actually different. when you run the driver installer, the new driver is place on the machine but often is not selected as the new driver. I have seen people update intel network drivers without the new driver being selected as the default driver. I generally have people use the pnputil.exe to delete the old .inf files

 

[johnbl]

looked at a second memory dump dated 10/14/24
looks like a problem with totalcmd64.exe
you should remove the program and retest.

edit: looks like the same dump file (internal debug session time was:
Tue Sep 17 16:53:00.983 2024 )

edit: just assumed it was a new file but yes it was the one I looked at before

 

[Sierrus]

looked at a second memory dump dated 10/14/24
looks like a problem with totalcmd64.exe
you should remove the program and retest.

edit: looks like the same dump file (internal debug session time was:
Tue Sep 17 16:53:00.983 2024 )

edit: just assumed it was a new file but yes it was the one I looked at before

There actually was a new Memory.dmp file Windows made October 11th, here it is if you wanted to give that a look as well.
https://www.mediafire.com/file/qwc0lt950wnocn1/MEMORY.7z/file

In the mean time I changed some BIOS settings related to power savings that I could find (ErP mode, HPET, and disabled onboard LAN) and then I activated highest performance mode in power options and made sure no USB device is allowed to go to sleep so we'll see where it goes from here, I guess? There's a lot of BIOS options that I straight up just don't understand even when I Google them but I think most are related to overclocking. I believe I enabled/disabled everything that might be the cause of issues? If there's any specific option you want me to take a look at lemme know.

Right now the most fullproof plan of triggering a BSOD might be yet another Windows reinstallation with drivers so I'll give that a go as well, and this time I'm avoiding Gigabyte for the most part and I'll try to get Realtek's Ethernet drivers directly. My BIOS actually says there's a 2.5Gb driver on it from 2019, that might also be a problem. 🤷‍♂️

 

[johnbl]

the second memory dump showed that something wrote to the windows registry, the system did a lazy write, later windows tried to flush the data to disk and the data was corrupted.

a lazy write is just sent to the disk but not confirmed as being actually place on the disk. IE it can still be in the disk cache.
windows can not shut down until the cache is flushed to disk. this operation failed and caused the bugcheck.

generally, you would update the drive firmware and hope for a fix. otherwise you would go into disk management and turn off lazy write option for the drive.

Note: this memory dump had a bunch of overclock drivers running and a flash tool running. you want to disable these. since they often do things that are no longer allowed. Windows also reported that your hal.dll has been modified

many vendors have tools that insert files into the operating system from the bios. these cause a lot of problems and I generally don't look at systems that have this running.

I will look at the bugcheck to see if you have any locks or pending installed or high non paged pool memory
--------------------------

looks like the system had 167 commits that were never confirmed from the storage

storage Class Devices

Usage Legend: B = Boot, P = Paging, D = Dump, H = Hiber, R = Removable

FDO # Device ID Usage UP DN FL
-------------------------------------------------------------------------------
ffffcf89cfad10a0 [1,2] 1 KINGSTON SNV2S250G BPD ? ? 3
ffffcf89cfac00a0 [1,2] 0 TOSHIBA HDWD130 ? ? 2
ffffcf89cfae20a0 [1,2] 2 KINGSTON SNV2S500G B ? ? 3


you might look for a firmware update for the kingston device
https://www.kingston.com/en/support/technical/products/snv2s
(just trying to find out why the registry did not flush to disk)

 

[Sierrus]

...I legit have no idea what any of that means, sorry. So now it's the C: drive, I guess? I have no idea what "overclock drivers" and "flash tool" I even have. Flash tool I'm assuming is USB, which is what I'm installing the Windows from at the time of the BSODs. I'm legitimately not overclocking anything apart from XMP for RAM. By "tools that insert files into the OS" I again have no clue what this is. Gigabyte's control center I guess? 🤷‍♂️ That memory dump file was from pretty much the absolute freshest Windows installation with barely any drivers and like 0 programs installed. As for drive firmwares, there aren't any if you mean the Kingston drives I'm installing the Windows on. The Kingston drive manager is reporting 0 issues with either of the NVMEs.

 

[johnbl]

ok, here are some of the software that was running:
amdvbflash.exe
Gv.exe
RGBFusion.exe
EasyTuneEngine
GigabyteDownlo
GCC.exe

drivers:
C:\Windows\system32\AMDRyzenMasterDriver.sys Sun Apr 28 21:52:50 2024
SystemRoot\System32\drivers\AmdTools64.sys Tue May 26 23:45:35 2020
C:\Windows\system32\drivers\gdrv3.sys Mon Jul 29 22:50:48 2024
\SystemRoot\system32\DRIVERS\mtkbtfilterx.sys Wed Mar 23 07:18:48 2022
C:\Windows\system32\drivers\MsIo64.sys Sun Dec 10 22:39:57 2023
C:\Program Files (x86)\GIGABYTE\RGBFusion\GVCIDrv64.sys Fri Dec 6 01:42:17 2019
C:\Windows\system32\drivers\gdrv3.sys Mon Jul 29 22:50:48 2024
C:\Windows\system32\drivers\CtiIo64.sys Mon Jul 8 19:39:19 2024
\SystemRoot\System32\drivers\AmdTools64.sys Tue May 26 23:45:35 2020

 

[johnbl]

ok, here are some of the software that was running:
amdvbflash.exe
Gv.exe
RGBFusion.exe
EasyTuneEngine
GigabyteDownlo
GCC.exe

drivers:
C:\Windows\system32\AMDRyzenMasterDriver.sys Sun Apr 28 21:52:50 2024
SystemRoot\System32\drivers\AmdTools64.sys Tue May 26 23:45:35 2020
C:\Windows\system32\drivers\gdrv3.sys Mon Jul 29 22:50:48 2024
\SystemRoot\system32\DRIVERS\mtkbtfilterx.sys Wed Mar 23 07:18:48 2022
C:\Windows\system32\drivers\MsIo64.sys Sun Dec 10 22:39:57 2023
C:\Program Files (x86)\GIGABYTE\RGBFusion\GVCIDrv64.sys Fri Dec 6 01:42:17 2019
C:\Windows\system32\drivers\gdrv3.sys Mon Jul 29 22:50:48 2024
C:\Windows\system32\drivers\CtiIo64.sys Mon Jul 8 19:39:19 2024
\SystemRoot\System32\drivers\AmdTools64.sys Tue May 26 23:45:35 2020

you would have to look up each item and see if you want it.
for example mtkbtfilterx.sys looks like MediaTek Bluetooth driver

gdvr3.sys is likely the driver for gigabyte control center

i would use the tool if you need it to update bios but then uninstall or disable them after the changes are made.

 

[Sierrus]

So yeah, Kingston definitely doesn't have firmware updates.

As for the running software, I only know RGB Fusion, GCC and the Gigabyte one. I only use them since I was trying to get drivers from Gigabyte and then change the RGB from the orange it was always defaulting to. I can get the drivers manually next time but it usually ends up with the same results.

amdvbflash is apparently a BIOS flash for the GPU? I have no clue why that's on since I never turned that on myself. Same with the AMD-related drivers, I don't install those myself, either Windows does or the MB does and I have no say in that. Unless I'm ok to manually uninstall those myself.

 

[johnbl]

lazy write: the drive lies and tells windows that it has saved the data to disk when it receives the data in the disk RAM buffers. This often works fine because the data gets flushed to disk a second or two later. it is ok as long as the power does not go out during that short timing windows. If there is a bug
then it causes a lot of problems. windows can require that the drive firmware confirm that data was written to the actual disk, it tells the firmware to flush the RAM cache to disk. in your case the object was set to a invalid value (zeror) and the system bugchecked. the problem can be in the drive firmware, the microsoft drivers, or some program messing with the system.

normally, when I see a memory dump with errors indicated in your last one, I would tell people to see if they can get a firmware update for the drive.
if that fails then I assume it is a bug in the current firmware of the device and have them disable the drives lazy writes. see something like this:

Enable or Disable Disk Write Caching in Windows 11/10

Disk Write Caching or Cache is disabled? Learn how to turn on or turn off Disk Write Caching. Helpful if you get Windows Delayed Write Failed Error.

www.thewindowsclub.com

not sure it will apply to the nvme drive.
with updated windows and new specs the problem might be in microsoft driver. like this one:
https://hothardware.com/news/windows-11-24h2-wd-ssd-bsod

 

[Sierrus]

Alrighty, I disabled write caching for every drive and will do that from now on if I reinstall Windows again (probably will tomorrow). It honestly wouldn't surprise me some micro corruption occurred on the drives since it coincides with the BSOD errors I was having related to cache and memory. 🤷‍♂️ If you have any more suggestions what I can modify I'm all ears. Thank you VERY much for looking into this so thoroughly!

 

[johnbl]

Alrighty, I disabled write caching for every drive and will do that from now on if I reinstall Windows again (probably will tomorrow). It honestly wouldn't surprise me some micro corruption occurred on the drives since it coincides with the BSOD errors I was having related to cache and memory. 🤷‍♂️ If you have any more suggestions what I can modify I'm all ears. Thank you VERY much for looking into this so thoroughly!

I would not reinstall, just uninstall software that you are done with.
you can use microsoft autoruns and just disable or remove drivers.
generally, you would uninstall, then make sure your version of windows is good by running the various microsoft utilities.
sfc.exe /scannow
and the
dism.exe command
something like start cmd.exe as an admin and run
dism.exe /online /cleanup-image /restorehealth
(it have been a while you might have to google for the exact command)

after these are ok, and you still get a bugcheck. you would turn on special debugging to force a bugcheck faster. generally using verifier.exe flags

you will also want to do a malware scan since the debugger complained about a modified hal.dll in both bugchecks. (sometimes 3rd party utilities make changes to files)

I would also google "how to force windows to delete the pagefile.sys on system shutdown" and set the registry values. This help for when programs modify windows components in the copy in the pagefile.sys. it will take longer to boot since each start up will load the file from the disk rather than just loading the pagefile.sys. it also helps to dump malware that infect the pagefile.sys

in this last case, windows should refuse to shutdown since it never confirmed that the registry setting got flushed to disk. it would normally just hang on shutdown.
there is a registry setting to tell you what is going on.
google for "how to enable verbose startup and shutdown logging"
and make the registry settings. At least you will have an idea why the system is hanging. (with this enabled window will report to the screen what is waiting for)
my second guess if it is not a firmware bug would be a bug in one of the gigabyte tools. Something that wrote to the registry and did not unlock it so the system could shutdown.

 

[johnbl]

So yeah, Kingston definitely doesn't have firmware updates.

As for the running software, I only know RGB Fusion, GCC and the Gigabyte one. I only use them since I was trying to get drivers from Gigabyte and then change the RGB from the orange it was always defaulting to. I can get the drivers manually next time but it usually ends up with the same results.

amdvbflash is apparently a BIOS flash for the GPU? I have no clue why that's on since I never turned that on myself. Same with the AMD-related drivers, I don't install those myself, either Windows does or the MB does and I have no say in that. Unless I'm ok to manually uninstall those myself.

I have not used this tool before. maybe you have to download the firmware update before it shows up in the list.
(not sure if the tool connects to the update site for you or looks for the update on the local drive.)

from the release notes:
Firmware Revision ELFK0S.6 – November 14, 2022• SNV2S/250G• SNV2S/500G• SNV2S/1000
your boot drive has firmware version elfkos.2 your second drive has the newer firmware from 2022.

see this for how to update: https://vtechinsider.com/how-to-update-kingston-ssd-firmware/
i don't think the tool actually does the download and extraction for you. just lists the available firmware files in some directory

from what i see the same firmware should work on both devices.
if you can not update the firmware, you might consider installing windows on the second device with the newer firmware.
(just if it turns out to be a firmware bug. I could not find any buglist reference to firmware revision elfkos.2

 

[Sierrus]

I have not used this tool before. maybe you have to download the firmware update before it shows up in the list.
(not sure if the tool connects to the update site for you or looks for the update on the local drive.)

from the release notes:
Firmware Revision ELFK0S.6 – November 14, 2022• SNV2S/250G• SNV2S/500G• SNV2S/1000
your boot drive has firmware version elfkos.2 your second drive has the newer firmware from 2022.

see this for how to update: https://vtechinsider.com/how-to-update-kingston-ssd-firmware/
i don't think the tool actually does the download and extraction for you. just lists the available firmware files in some directory

from what i see the same firmware should work on both devices.
if you can not update the firmware, you might consider installing windows on the second device with the newer firmware.
(just if it turns out to be a firmware bug. I could not find any buglist reference to firmware revision elfkos.2

...I legitimately have no idea how you're supposed to download a firmware update other than through the manager. The site isn't providing any links to downloads other than the manager. https://www.kingston.com/en/support/technical/ksm-firmware-update <- this is the only page I see when I navigate to firmware update on the site and it only lists release notes. And as for 1 drive having a different firmware, I tried installing on the other one as well before and it still ran into BSODs.

EDIT: Y'know what, I just sent an email to Kingston customer service to clarify what exactly it is I'm seeing in their manager and on their site because it doesn't really make sense. If there's newer firmware updates for my drives but it's not picking up, then something ain't right. I'mma see what they say, then.

 

[johnbl]

...I legitimately have no idea how you're supposed to download a firmware update other than through the manager. The site isn't providing any links to downloads other than the manager. https://www.kingston.com/en/support/technical/ksm-firmware-update <- this is the only page I see when I navigate to firmware update on the site and it only lists release notes. And as for 1 drive having a different firmware, I tried installing on the other one as well before and it still ran into BSODs.

EDIT: Y'know what, I just sent an email to Kingston customer service to clarify what exactly it is I'm seeing in their manager and on their site because it doesn't really make sense. If there's newer firmware updates for my drives but it's not picking up, then something ain't right. I'mma see what they say, then.

i downloaded the app and looked at the network trace. it did contact servers and it did look for ssd updates.

I suggest you focus on this:
download and run microsoft autoruns from here:
https://learn.microsoft.com/en-us/sysinternals/downloads/autoruns

install it, run it and find the gigabyte driver called
gdvr3.sys
and delete the entry or select the check box so it does not load. then reboot your system.
(it should block your gigabyte control center from making changes)

Note: when you reinstall windows the solid state drive can get backed up on its internal firmware routines. These routines run at system idle time but you can just boot into bios and leave the machine powered on in the bios screen for a hour or two and the firmware can finish its cleanup routines. sometimes this condition can look like firmware bugs in the drive.

I would do both of these, then reboot and see if you get another bugcheck.

it would be nice to delete the pagefile.sys so any corruptions are not saved between reboots.
from google:
# Navigate to the following registry key: HKEY_LOCAL_MACHINESYSTEMCurrentControlSetControlSession ManagerMemory Management # In the right page, double-click on ClearPageFileAtShutdown entry to change its value to 1. # Close the Registry Editor and reboot your machine.

 

[Sierrus]

Alright, sorry for the extended silence, I was waiting for Kingston to come back to me regarding firmware updates and anything else they might say about my SSDs before continuing their usage. Basically they said my specific revisions of the SSDs don't require any firmware updates and that based on the info I provided them there's nothing wrong with the SSDs. HOWEVER, they mentioned

Also, we noticed some high number of Error Info Log Entries. This means that the SSD received a command that we were not able to deal with, does not necessarily mean that the drive has an issue but it just could not handle the command(s) sent.

I have no idea what they meant by that but I don't like the sound of it, lol.

Either way, I reinstalled Windows again, and after tinkering with the BIOS settings from last time, enabling the virtual memory clear at shutdown and the verbose setting, preventing my computer from snoozing USBs and ignoring any Gigabyte programs entirely, I haven't encountered a BSOD during the entire process...........until I realized I ended up installing Windows on my 250G while the boot itself remained on the 500G. 🤦‍♂️Then I had to reinstall AGAIN, and then I had 1 single BSOD after the first Windows update restarted the PC, BUT, I'd like to think this happened only because my USB boot drive stayed in the PC instead of me plugging it out after the installation. I haven't had a BSOD so far, but I'll keep closely monitoring it in the following few weeks to see if anything happens after I strain the PC a bit more extensively with more programs and games. As a sidenote, disabling cache writing on my drives prevented me from restarting the PC so I had to keep that enabled. Even with the verbose setting enabled, the PC just froze on the "Restarting" screen and the circle kept spinning endlessly. I'll just keep that on from now on, but every other suggestion seems to be working fine.

If the combination of changing the BIOS settings, clearing virtual memory and disabling USB snoozing was the fix all along, johnbl you're a GOD and thank you so <Mod Edit> much~ Fingers crossed it remains stable.

 

[johnbl]

Alright, sorry for the extended silence, I was waiting for Kingston to come back to me regarding firmware updates and anything else they might say about my SSDs before continuing their usage. Basically they said my specific revisions of the SSDs don't require any firmware updates and that based on the info I provided them there's nothing wrong with the SSDs. HOWEVER, they mentioned

I have no idea what they meant by that but I don't like the sound of it, lol.

Either way, I reinstalled Windows again, and after tinkering with the BIOS settings from last time, enabling the virtual memory clear at shutdown and the verbose setting, preventing my computer from snoozing USBs and ignoring any Gigabyte programs entirely, I haven't encountered a BSOD during the entire process...........until I realized I ended up installing Windows on my 250G while the boot itself remained on the 500G. 🤦‍♂️Then I had to reinstall AGAIN, and then I had 1 single BSOD after the first Windows update restarted the PC, BUT, I'd like to think this happened only because my USB boot drive stayed in the PC instead of me plugging it out after the installation. I haven't had a BSOD so far, but I'll keep closely monitoring it in the following few weeks to see if anything happens after I strain the PC a bit more extensively with more programs and games. As a sidenote, disabling cache writing on my drives prevented me from restarting the PC so I had to keep that enabled. Even with the verbose setting enabled, the PC just froze on the "Restarting" screen and the circle kept spinning endlessly. I'll just keep that on from now on, but every other suggestion seems to be working fine.

If the combination of changing the BIOS settings, clearing virtual memory and disabling USB snoozing was the fix all along, johnbl you're a GOD and thank you so <Mod Edit> much~ Fingers crossed it remains stable.

disabling the virtual memory and the usb sleep are just workarounds in the hopes that you reboot before the bug is exposed. IE you reboot or shutdown, the state of the machine is cleared from the pagefile and it takes time for the problem to happen again.

I have seen bugs like this many times. one that took 25 sleep wake cycles before you get a bugcheck.
you can also have problems due to specification changes for devices and motherboards and drivers.
these tend to show up in the internal error logs but are hard to figure out who owns the mistake.
Microsoft windows error reporting will report the bugcheck and their systems will debug it and report to the vendors or Microsoft will make a change to prevent the bugcheck. in that case they will send it to your machine in a windows update in the optional update section.

in the "olden" days people would turn off their machines each day and lots of bugs would just disappear. After sleep functions were enabled a lot of bugs in drivers were exposed because they would take longer than 8 hours to crash the system.

 

[ubuysa]

It's vanishingly unlikely to be Windows because it crashed in Safe Mode and that's as pristine a Windows system as you can get. That a Linux distro crashed as well is confirmation that this is not Windows but a hardware problem. In addition, that Linux distro doesn't access your SSD/HDD drives at all so you can probably discount those.

You might try that Linux distro again with all drives, and unnecessary PCIe cards (like networking cards) removed. Also run the Linux distro with one RAM card (in the correct slot, typically A2) and if it crashes then swap the RAM cards and try again. If Linux is crashing from the install media then it's CPU/RAM/motherboard/PSU.

These kinds of hardware problems are difficult to track down, so a careful step-by-step approach is what's needed. Sadly, they often come down to spending money to replace components just to see whether they were at fault.

 

[Sierrus]

It's vanishingly unlikely to be Windows because it crashed in Safe Mode and that's as pristine a Windows system as you can get. That a Linux distro crashed as well is confirmation that this is not Windows but a hardware problem. In addition, that Linux distro doesn't access your SSD/HDD drives at all so you can probably discount those.

You might try that Linux distro again with all drives, and unnecessary PCIe cards (like networking cards) removed. Also run the Linux distro with one RAM card (in the correct slot, typically A2) and if it crashes then swap the RAM cards and try again. If Linux is crashing from the install media then it's CPU/RAM/motherboard/PSU.

These kinds of hardware problems are difficult to track down, so a careful step-by-step approach is what's needed. Sadly, they often come down to spending money to replace components just to see whether they were at fault.

It can't be RAM anymore since a brand new set is experiencing the same issues. I don't even know what would be considered "unnecessary" at this point, the Bluetooth and Wifi? Windows just forces updates on me for them at this point, I can't even turn them off. Unless you mean disable them but they still have drivers for them installed. I do need the onboard LAN since that's my internet connection. And yeah it did crash in Safe Mode but I'm still not sure about Linux. How do crashes occur there? It froze on me and stayed frozen for a while so I had to force restart, is that considered a crash? 🤷‍♂️ I don't know how you access dump files in Linux either.

How does a PSU fail exactly? The errors I've seen people have with PSUs is that most of them can't even get into Windows or past POST since BIOS doesn't respond. And no one can even test it other than "check if it powers components or not". I'm not the guy who can blow off $7k on new PCs from earlier in this thread so all I can do is rely on RMA while I still can. So I can only narrow it down to MB or possibly CPU. The CPU has little reason to be failing since I saw no physical damages on it when I replaced the cooler, the only thing that could worry me are the previous cooler issues I had might've done something to it internally? I literally can't see it being anything else anymore, MB is now my main culprit.