Windows 10 repost - for reference = Device driver fix @2. NVLDDMKM.SYS blue screen error/system fails (root cause and fix)

  • Thread starter Thread starter Dr.Richard.V.Finney
  • Start date Start date
D

Dr.Richard.V.Finney

This is a re-post of a thread (mostly me talking to me) from Desktops, etc. (due to the wallpaper issue @ 1)


The discussion @ #2 is a resource for MS developers/forum helpers/experts, and folks at NVIDA, ATI, REALTEK (who are supposed to/tasked to, fix this kind of thing), in addition to those of you good enough or brave enough to tackle critical failures.

1 = Qu: Reinstating "travelling" wall-paper/backgrounds ... and; 2 = Fix: the dreaded nvlddmkm.sys (assorted) blue screen dramas.


Hi all,

Keep well - and if protesting (I agree) - keep Covid19 safe ...

1.

Question:

I am still on the ML350G6 - as I've not had time/chance to fix the Z400 yet (well its only been 60 days odd/slack maybe).


Anyway - one of the things I really like/like-most about Windows 10, is the swapping background/wall-paper at boot-up. It is like travelling the world - at its best - without mad traffic, crazy drivers, bed-bugs and bad food ... plus you can often see places you've been, so relive those visits/holidays (good memories), as you wait for the system to get ready for your session/work-day.


This unit (computer) has a Nvida GT220 x 1Gb graphics card, thus has the very buggy 342.01 driver set and BS dramas; which I have been trying to fix (addressed next).


In the process, my computer-chair world travel buzz has come to a halt (much crying), and I only get the beach cave ... which is very nice, but not as much fun as seeing lots of different places.


So my question is ...


How/where do I reset the windows 10 log-on/boot, so I get my ' jollies' (travelling daydreams) back?


{{addendum (not critical): any ideas what 'dang varmit' program/app, might have turned it off?}} <<<IOBit's 'Advanced System Care" system boot speed-up, disabled Bing wall papers>>>



2. The Nvida (assorted) Blue Screen of Death events ... possible solution (when all the others just do not work).


Many folks blame windows/Win10 for this; however my (very detailed) research suggests it is wholly a failing of Nvida circuitry and drivers trying to bridge sloppy programming of the board handler components on a number of Nvida high-end graphics cards - circa 2011~16 (not the Quaddro's).


Symptoms and messages can vary between machines, as well as being 'random', however all result in the "Blue Screen" (or black screen loop); while many involve: 'critcal CPU write'; Video faults, or 'nvlddmkm.sys failed' ... error messages.


As a parallel - say you've got "Chicken-pox" - that is the core illness issue (needs addressing - for a cure). Red spots, High temp, sudden desire to scratch in dirt for worms, building straw nests, laying spotty eggs, etc., are just symptoms.


The Nvida illness is the supporting circuitry around the GPUs - in a number/series of cards - nominally common to the time frame cited above.

{{the system blue screens, shut-down codes/messages, recurring black-screen loops etc. - are just symptoms.}}


These problems have been around for years, and gotten a lot of folks real angry/blaming everyone ... thus lots of IT folks have tried to come up with solutions. I visited lot of tech-head sites and tried a number of the resolutions (fix suggestions); however most did not work or were technically flawed - so could not work - for the GT220 cards.

{{Yes- Included in trials was driver delete and complete fresh reinstall = failed ... Reason I think the 'flush' worked. Why - not sure.}}


One of these 'flawed' solutions showed promise - if you could fix the 'flaw'


That is the solution requiring the user to go into the Drivers folder (Windows system), find nvlddmkm.sys - then rename it to nvlddmkm.old ... followed by finding nvlddmkm.sy_ in the Nvida set-up directory ... copying it to a new/working folder on your drive and expanding it - to be a new copy of nvlddmkm.sys. You then copy that, and paste the new copy (as administrator) in the system folder, where you rename the original to "nvlddmkm.old" ... followed by rebooting.


The 'flaw' in this is that they say to download the latest Nvida driver update. Problem for GT220, 240, 400 and many others - is that the very bad/buggy/maybe-worst version 342.01 driver (Aug 2016 - I think) - is on their site as that 'latest'; so you end-up reinstalling one of the worst drivers - and the problems will recur.


I sat down and wondered: 'if they had fixed it in later drivers for different cards/configurations'; then, 'if you could force the older card to work with a different/not-overly compatible driver set?'


Note: There are truck-loads of reasons not to try this, so do not try it {{experts exempt - does get a bit 'hairy'}}. I was just experimenting (and I do know what I'm doing/risks, etc.)


Right - I ran off and found the 931.35 driver set (~2018 June, I think); extracted it to a working folder, then took out the nvlddmkm.sy_ file (wrong for GT220); then used that as the seed file to replace the renamed nvldmkm.342 ... and rebooted.


Card freaked, after about fifteen minutes of system panic/self-repair (computer - not me, I was watching); it canned the Nvida card and defaulted to the onboard chipset and windows driver with mid-screen sizing.


At that point I thought - "Hmmm, well its back to the drawing board/option II = get a stable earlier version (March 2014) and try the same again."


I renamed nvlddmkm.sys to nvlddmkm.931; then renamed nvlddmkm.342 back to nvlddmkm.sys and rebooted ...


System came back to normal - following some readjustments (automatic/win10 system fixing itself); however my "Drive Booster" (7.4 pro/licensed) kicked up a message to say: "I had 7 drivers to update ..."


Bit strange, as it was all updated before I tried the unsuccessful 342 for 931 switch.


I looked and noticed 75% were Nvida secondary (mostly audios) component drivers; plus, several PCI bridge/similar and a couple of key MB resource cross-handlers - so I thought maybe the failed-stunt (342-931 switch) flushed out the graphics board roms.


The real puzzle is the code errors appear to be in the audio board instructors (that do not seem to play-up), whereas the faults all seem associated with the graphics/video chip instructors ... so you tend to focus on them. Might be where Nvida guys/girls went wrong trying to fix it.


I updated the lot, rebooted and guess-what ... problems gone!


For how long I do not know (because the root cause is bad board support circuitry); however on the HP Prolink bios (dual Xeon, running a gt220), I can force a failure by going into the bios, then just doing an unnecessary 'save' (bios to rom), and F10 exit.


This immediately triggers a nvlddmkm.sys fail/blue screen. At the moment it reboots as it is supposed to (no Nvida fail/blue screen), so I think the process must have flushed bad or redundant code out of the cards roms ... as this occurs before the post stage is finished (before Windows boot).


That same process is what tells you - It is not Windows/Win10 causing the failures. (majority = not)


So - for the tech-heads here (it worked), and FIG-JAM folks at NVIDA - when not bragging about the upcoming developers conference (for folks to develop new crud/crud drivers) - maybe a rom-flushing script could be the go.


Future? Nvida share the same attitudinal failings (executive) as HP. Quaddro cards are great; however I cannot justify buying equipment from suppliers with unreliable and indifferent attitudes (at best) to their customers, and/or suppliers who just walk-away-from or ignore, major product faults.


As an institution - we are dumping HP and switching to Fujitsu or Lenovo. For CAD (dumping Nvida) we will specify ASUS, MET (Ricoh), or Samsung/PRC offerings.


Cheers all,


Rick.


>>added - plus eight hours <<


@1: Whoopie ... It's the "BING wall paper (changer) app" - that someone nice at MS, tagged to this post. Get it ... well worth it.


@2: I read through a couple of hundred post headers to get back here (looking for responses - for #1 = now fixed); and estimate 60>75% are all either the GPU sub board component issues (referred above - ATI and others have similar faults on various boards/chip-sets); or the PCI-USB cross-hander issue (USB2/3 functionality); both of which - essentially - are hardware, not: O/S software, or Win X/10/8/7 in nature (base or cause).


I applaud the efforts of the 'volunteers' in MS Community ... who respond and try to sort-out those 60>75% of complaints/raves - that are issues other than MS/Windows O/S.


~R.



>>>added - plus 8.5 hours <<<


MS developers/Nvida developers/ ATI & Realtek:


I was thinking about the fact the nvlddmkm.sys bugs fixed themselves - post 342,931,342 shock therapy - focusing on the audio instruction sets; then realised this machine and others here (similar problems - not the Quaddros), run their audio through Realtek - NOT NVIDA. That made me think maybe the root dramas are conflicts between the Realtek sound and Nvida/ATI Video (+ sound) components (instruction scripts therein).


Worth throwing in the mix, because truck loads of IT experts at various component manufacturing and computer distributing concerns (HP, et.,al.), plus possibly the Mormon Tabernacle Choir {cheers to Harry Chapin} ... have not got a solution/correct direction in near a decade. Thirty years ago I was an alleged genius (Clintons Mk1, IBM/Intel ~ Pentalogic White Paper, etc.) but I took an Asprin and got better/think-I'm-a-rabbit {{Makka - John - took too many 'blue' asprin, and now he floats around the Caribbean, thinking he's a persecuted chipmonk! [[waves to Peter, Craig, Deb, Suzy {sister Moon}, Will and Cass - we're still laughing ... cheers you guys.]]. Fairly sure I've got to the root of this, so I'll crawl back under a rock and play with BRASS (the radar system). Thread Ends. ~R.

Continue reading...
 
Back
Top