Windows 7 - getting BSOD with disk.sys ( disk+1 )

[Shneiky]

Got BSODs during Windows loading, opened the minidump and started checking. Error code is:

BugCheck 1000007E, {ffffffffc0000005, 0, fffff8800556b308, fffff8800556ab60}

At the bottom it says that is probably caused by:
disk.sys ( disk+1 )

Is disk+1 the primary or secondary disk?

I have an SSD 840 Evo as primary (less than an year old) and a 5 year old Seagate 1 TB secondary drive divided in 5 partitions. 840 Evo is all green 99% health reported by Samsung Magician. 1.9TB written on it till now.

All drivers are up to date with the exception to the nVidia driver. Been using this driver ever since I installed Windows on the SSD and it has been rock solid.

I do believe it has to do something with my resent up on the overclock. Used to have 4.0 GHz at 1.09, now i am at 4.2 at 1.18. Prime95 is rock solid. RAM is at default 1333 9-9-9. Memtest is rock solid.

Any idea what is causing this?

 

[Darkbreeze]

Have you run Seatools to check the drive health? Magician, which I also have and use, only provides S.M.A.R.T data, it doesn't do any in depth diagnostic checking. I'd run Seatools for Windows (Drive brand or type is irrelevant) and run the short drive self test and long generic, just to rule out any health issues. Run it on all connected drives.

http://www.seagate.com/support/downloads/item/seatools-win-master/

 

[Shneiky]

Doing that right now. Will let it run scans when I am at the office. So in the end:

Is ( disk+1 ) the primary disk or secondary disk?

 

[Darkbreeze]

I have no idea what is being referred to here. Most people probably wouldn't aside from software engineers. I'll assume that it could be the primary disk, but it may be something else entirely like a virtual disk.

 

[johnbl]

disk.sys is the name of a piece of device driver software. It does not refer to a physical disk.

note: some malware attack the disk subsystem driver, you might want to run malwarebytes and
run cmd.exe as an admin then run the system file checker.
sfc.exe /scannow

the bugcheck you got indicated a bad memory address was passed to this driver.

 

[Shneiky]

Sea tools gave a pass. Both drives passed all basic tests.

No malware found. Checked for viruses as well. Made Microsoft Security Essentials perform a deep scan to both drives - nothing. sfc turned out as expected - nothing.

Is it possible this whole issue to be caused by my overclocking settings? I rescently moved from 4.0GHz at 1.09 to 4.2 GHz at 1.18v. The OC is completely stable at full load , but I did not spend that much time to check if it is stable at lower loads. Using LLC level 2 in BIOS (ASrock has LLC level 1 as highest (100%) and LLC 5 as lowest (0%) or reversed Asus settings). Maybe the 0 to load was what caused it? I am not sure that Turbo Boost does not engage before windows loads.

 

[johnbl]

Yes, I tend not to debug systems with overclocking enabled. Just too many problems with the electronics timing

Sea tools gave a pass. Both drives passed all basic tests.

No malware found. Checked for viruses as well. Made Microsoft Security Essentials perform a deep scan to both drives - nothing. sfc turned out as expected - nothing.

Is it possible this whole issue to be caused by my overclocking settings? I rescently moved from 4.0GHz at 1.09 to 4.2 GHz at 1.18v. The OC is completely stable at full load , but I did not spend that much time to check if it is stable at lower loads. Using LLC level 2 in BIOS (ASrock has LLC level 1 as highest (100%) and LLC 5 as lowest (0%) or reversed Asus settings). Maybe the 0 to load was what caused it? I am not sure that Turbo Boost does not engage before windows loads.

 

[Shneiky]

This did appear after a change to the OC. Single instance. Did not happen again yet. Else, everything was running without a hiccup for more than an year.

 

[Darkbreeze]

I'd back off the OC to the previous stable settings for now. If the problem does not recur then further tweaking and stability testing may be necessary for the desired overclock. You could have a lack of sufficient voltage on the memory or cpu.

 

[Shneiky]

Got a new one:

BugCheck 1E, {ffffffffc0000005, 0, 8, 0}
Probably caused by : ntoskrnl.exe ( nt+74ec0 )

Again in Windows preload phase.

----

I have an offset of -0.150 v on the vCore with additional turbo voltage of 0.55. I believe that the voltage is enough for low load and full load (since it did not give any problem) but the voltage is not enough for cross load.

The -0.150 offset with +0.04 additional turbo was completely solid at 4.0 GHz for the past year. Guess I will back down that -0.150 to -0.140/-0.130 and decrease the additional turbo. This way it should stabilize at lower loads.

 

[Darkbreeze]

Well, I think I'm going to back off of this thread as I can see the impending disagreement that's about to start, but I will leave off with this comment, which is of course simply my opinion.

In my opinion, and that of MANY other long time overclocking veterans, if you're going to configure a manual overclock to a specific frequency, it's a poor idea to leave the turbo features enabled. Stabilizing an OC chip requires finding a specific group of settings and voltages that works advantageously with those settings, and using turbo or power saving features is intentionally upsets that balance, causing voltages and multipliers to be in a constant state of change. Use of turbo features obviates the need for a specific overclock configuration since the system is doing the overclocking according to demand on the cpu.

Generally, you want one or the other, not both. Plus, the automatic features of turbo boost or turbo core, automatically negate any specific settings you may have configured for a manual overclock, therefore it renders it pointless. Anyhow, good luck with your issue. I do hope you get it solved.

 

[USAFRet]

Stabilizing an OC chip requires finding a specific group of settings and voltages that works advantageously with those settings

....and then leave it.

 

[Shneiky]

Well darkbreeze, I am sorry to hear that. I never wanted to give the impression I am trying to "argue".

I always overclock through Turbo Boost leaving all power saving features enabled. Always. Overclocking by setting a constant manual voltage and constant frequency is a lazy man's way - sure it is easier and streightforward. This is at least my personal view. Sure, it might let you be more stable above 4.6/4.7, but how many go for that for 24/7 unless their top hobby is OC. Most of my 3D is low load with spikes of highly intensive single/double threaded load, until we get to rendering - and there it is 100% load. I don't want to be running full on while setting up my scene and fiddling with options, but I do want to run full juice while rendering.

Also, both the Turbo and the Speed Step options are dependent on the VID voltage, and once you get that one figured out and play around to know how much your chip needs for each frequency, you can achieve stable overclock with all the power saving and voltage/mhz lowering goodies.

Anyway, just to get down to the issue. The 2nd BSOD that was with unrelated issue to the first one gave me the tip. Changed the offset vCore to -0.140 instead of -0.150 and changed the additional Turbo voltage to +0.055 instead of 0.059. And now I am again rock solid. The -0.150 was good for no turbo or for 4 GHz +0.04 Additional turbo voltage, but it was not enough for 4.2 GHz cross load.

Overclocking with Turbo boost and all the power saving features enabled will stress your system less and spare some electricity. (where I live - I pay 40 cents per KW/h. And that is 40 EUR cents). Also, you can always just go to your Windows power Options and change your maximum CPU at 99% will make your system not engage in Turbo boost. So you change from running OC to running stock whenever you wish from within the OS.

Anyway, at least you pointed me that my drives were not the issue at hand which was useful.

 

[Darkbreeze]

Nah, I didn't mean you. I meant the ones that were about to come out of the woodwork the minute I made that comment.

 

[Shneiky]

Then I wish to excuse myself for the miscommunication.