Multiple BSOD on a New Dell Inspiron Laptop - Cause: Intel RST Driver

Hello, I have lurked on the 7 and 10 forums over the years, and benefited greatly from the tips and information that I have found here, so I thought that it was now time to give something back!

This is my first post: a tale of woe along with my solution, in case it helps someone else.

I bought a brand new Dell laptop (Inspiron 15 3567) at the end of January and I have had the worst out of box experience of any machine that I have ever owned - multiple BSOD on a daily basis for more than a month!

For those in a hurry, the solution was to set SATA Link Power Management to DISABLED, under the "Performance" tab of the Intel Rapid Storage Technology application.
Attachment 271388

The longer version:
I started out with the image of Windows 10 provided by Dell, I ran through set-up then checked for and installed all Windows & driver updates that were offered. Then the BSOD started:
KERNEL_SECURITY_CHECK_FAILURE (139) - lots of these
DRIVER_OVERRAN_STACK_BUFFER (f7)
SYSTEM_SERVICE_EXCEPTION (3b)
DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (7e)

I ran memtest86+ for 12 passes - no errors found.

I booted Linux Mint from a USB drive and ran it live with internet access for 24 hours - faultless, no problems or crashes whilst browsing with Firefox.

From these two tests I was fairly confident that there were no underlying hardware problems, so I downloaded and installed WinDbg and started my education on debugging memory dumps. I googled error codes and tracked down various driver updates, I ran sfc /scannow and DISM multiple times along the way (no problems found); still the BSOD occurred. I removed windows updates in case they were the cause of the problems - no difference.

As a last resort I used the Microsoft MediaCreationTool to downloaded a fresh version of Windows10 v1909 and used this to completely wipe the hard disk (all partitions) and perform a fresh install. I updated all the drivers and set windows updates to "paused for 30 days" - to remove one set of variables from the system.

Then the BSOD started!!

DRIVER_OVERRAN_STACK_BUFFER (f7)
Caused by PROCESS: soffice.bin = open office V2 system tray prog

MULTIPLE_IRP_COMPLETE_REQUESTS (44)
Caused by driver: HIDCLASS.SYS = I2C HID Miniport Driver

DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
Caused by driver: Wdf01000.sys = Kernel Mode Driver Framework Runtime

PFN_LIST_CORRUPT (4e)
Caused by driver: win32k

IRQL_NOT_LESS_OR_EQUAL (0a)
Caused by driver: win32k

SYSTEM_SERVICE_EXCEPTION (3b)
Caused by driver: Firefox.exe

SYSTEM_SERVICE_EXCEPTION (3b)
Caused by Driver: win32kfull.sys

KERNEL_SECURITY_CHECK_FAILURE (139)
Caused by Driver: nwifi = MS NativeWiFi Miniport Driver

DPC_WATCHDOG_VIOLATION (133)

This is the view of system health over the first 10 days:
Attachment 271389

During this time I tried to catch the problem using Driver Verifier but it never caught / tagged anything in the dump files.

Whilst searching for clues to the cause of the crashes I stumbled across a guide to high performance windows settings, at Tweaktown.com, where they offered the following advice: "disable Link Power Management in the Intel Rapid Storage Technology (RST) control panel. Link power management increases I/O latency, resulting in lower performance, which is why we want to disable this feature."

I changed this setting and the BSOD have disappeared; six days on all windows updates have been applied and I have not had a single crash!! Currently Microsoft Windows [Version 10.0.18363.720].
Attachment 271390

It looks to me like the crashes have been caused by the SATA link being put into a low power state, from which it did not wake up in time to service a disk access. This would explain why there was never a consistent driver listed as the cause of the crashes in the dump files.

I have listed the non Microsoft drivers currently in use by my system, in the screenshot below, in case they are of use to someone else.
Attachment 271391

The Intel RST driver (v16.8.3.1003) that is currently in use is available on the Dell website as: "Intel-Rapid-Storage-Technology-Driver-and-Management_SV16A_WIN_16.8.3.1003_A05".

I tried the latest version RST drivers from the Intel website (v17.8.0.1065), but this resulted in WHEA logger and other warnings in the event log, so I reverted back to the Dell version for the time being as I was sick and tired of chasing bugs!

See:
BSOD - Posting Instructions

It tallies with what I witnessed here and also experienced myself and for quite considerable amount of time either.

But it's not an entirely clear-cut case as that I'm afraid, I think there's some delicate interplay with one or two other dependencies here: Intel MEI? Intel's buggy CPU & microcode? Something OS- or its kernel-related? But it's mere speculation on my side...

The funny thing is: I have never (NEVER) experienced such nonsense on my Windows 8.1 x64 laptop since I got it back in mid-2013 - and guess what? I just checked and Link Power Management is enabled - and to my best knowledge I never had to fiddle with it...

krzemien said:

It tallies with what I witnessed here and also experienced myself and for quite considerable amount of time either.

But it's not an entirely clear-cut case as that I'm afraid, I think there's some delicate interplay with one or two other dependencies here: Intel MEI? Intel's buggy CPU & microcode? Something OS- or its kernel-related? But it's mere speculation on my side...

The funny thing is: I have never (NEVER) experienced such nonsense on my Windows 8.1 x64 laptop since I got it back in mid-2013 - and guess what? I just checked and Link Power Management is enabled - and to my best knowledge I never had to fiddle with it...

I agree, there is definitely an interplay between the various software components, but I thought that is what Microsoft was supposed to test for - before they release drivers into their catalogue.

A fresh install with updates from MS should not lead to an unstable system - especially with a bog standard chipset. It smacks of poor quality control, as does releasing updates that result in BSOD and other problems. If the two are combined - a fresh install with dodgy drivers AND a bad update on top, then where does that leave the average person? - pulling their hair out and wailing at Google!

I skipped Win 8, because the user interface was so horrible, so I stayed with Win 7 until the very end. I ran this for years on several laptops and only ever had a BSOD when I had a faulty memory module.

If Win10 updates prove to be too much hassle to maintain over the coming months, then I shall wipe it and resort to Linux Mint!

Some further info on my system configuration:

• The HDD is a Toshiba MQ04ABF100, so a spinner not SSD.
• The performance settings were set to mid-range = "Better Performance" and the laptop is usually running on mains power.

These two factors may make a difference, in terms of when power savings are applied to the SATA link.

- - - Updated - - -

Ztruker said:

A new computer should not have these problems. Contact Dell and start the RMA process to get it replaced.

I agree that a new computer should not have BSOD problems, however, there is plenty of evidence on these and other forums that they do. They well look like hardware faults but, as mine did at first, but I bet that many are actually due to poor software/drivers.

I do not agree that the laptop needs an RMA to Dell for replacement - there is nothing wrong with the hardware, this problem is due entirely to poor software!

During fault finding I tested the laptop using the Dell diagnostics, as well as memtest86+, and no faults were found. It also ran perfectly under a live boot of Linux Mint, and since making the change to the RST driver, it has been running flawlessly under Windows 10 for the last 9 days.

Attachment 271707

That does not look like faulty hardware to me, which is why I have posted my findings in the hope that it will save someone else some grief.

The key trouble is that manufacturer will stick to their guns - and most of the time rightly so - that there's nothing wrong with the hardware. All sorts of UEFI hardware tests executed in non-Windows environment - if they complete without glitch - usually confirm this unless you really can prove otherwise.

But then, with Windows OS added and BSODs suddenly in place, Microsoft will conveniently say that there's obviously nothing wrong with OS either and that you should liaise with manufacturer regarding the possible hardware fault. Which is nonsense and the above post from OP indirectly proves this.

So it's a vicious circle.

And then: if you ditch, say, HP and also Dell, on these arbitrary grounds, then what's left? Mac & OS X?

There is someone in every business with "the buck stops here".

They accept responsibility for any problems related to sales.

They are focused on customer service.

They are focused on the reputation of the company.

The challenge for the consumer is climbing over hurdles to find these people.

I agree with Ztruker.

The product should be returned.

And if there are problems with replacements they should be returned too.

Chaps, I'm not a OP and do not have / report this issue, I merely thrown my two-pence to this thread

[And yes, I had quite similar experiences - I actually registered here then to seek advice - but then guess what? My PC is entirely operational and stable as ever since BSODs magically vanished back in late Oct 2018 despite no hardware/software changes of any sort around that time (at least the ones I would be 100% certain of)].