Sign in with
Sign up | Sign in
Your question

BSOD 0x18 what is the cause ?

Last response: in Components
Share
March 9, 2012 8:31:33 AM

On a Windows Vista laptop I have had several blue screens lately. Each time the dump file suggests that the stop code was 0x18 in module ntkrnlpa.exe and the error was "Reference by Pointer".

I understand that this means that reference count of an object is illegal for the current state of the object i.e. an inconsistency in the object's reference count, and that the parameters indicate the type of object and the particular ID of the one at fault. My problem is that these are not consistent. And the drivers have not been updated for years.

How can I map the object to a parameter ? It may be a hardware item that can be replaced if only I knew which.
The last three parameters in my case have been:

(0xFFFFFFFF85A815B8, 0xFFFFFFFF863330E8, 0x2, 0xFFFFFFFFFFFFFFFF)
(0xFFFFFFFF85A815B8, 0xFFFFFFFF8957C710, 0x1, 0x1)
(0xFFFFFFFF85A81640, 0xFFFFFFFF8632F4C0, 0x1, 0x1)

More about : bsod 0x18

March 9, 2012 4:38:37 PM

...wow, this is a rare one.

Quote:
Bug Check 0x18: REFERENCE_BY_POINTER

This indicates that the reference count of an object is illegal for the current state of the object.


Quote:
The reference count of an object is illegal for the current state of the object. Each time a driver uses a pointer to an object, the driver calls a kernel routine to increase the reference count of the object by one. When the driver is done with the pointer, the driver calls another kernel routine to decrease the reference count by one.

Drivers must match calls to the routines that increase (reference) and decrease (dereference) the reference count. This bug check is caused by an inconsistency in the object's reference count. Typically, the inconsistency is caused by a driver that decreases the reference count of an object too many times, making extra calls that dereference the object. This bug check can occur because an object's reference count goes to zero while there are still open handles to the object. It might also occur when the object's reference count drops below zero, whether or not there are open handles to the object.


I could see this being a memory problem, but I would be expecting other BSOD's besides 0x18 [specifically, 0x1A: MEMORY_MANAGEMENT]. And its VERY unlikely the kernel isn't reference counting right...

For the time being, run a memory scan with memtest86/Prime95. Best case would be a RAM problem. Also check and make sure that windows didn't update any drivers on you recently...
March 9, 2012 4:53:52 PM

agree with gamerk316, most likely a bad stick of ram, but it could also be a bad driver, a bad motherboard, a bad CPU, windows corruption due to a failing HDD, etc.

Will the system boot at all? If so then check the error logs and see if there are any red flags there.
Related resources
March 10, 2012 8:15:40 AM

I ran Memtest for 2.25 hours in which time it made 5 passes with no errors.
The PC works fine most of the time - the BSODs are 'only' 3 in the last 4 days.

How can I correlate the parameters in the stop code with hardware IDs ?
May 4, 2012 4:33:22 PM

sorry for the delay in response.

those codes all have to do with memory addresses where the memory had a problem. It does not mean that the memory is bad, but that the issue that caused the crash happened at that location. It could still be any number of things causing the problem.

Install CPUID and run the Intel burn test, or another load test software and see what your temperatures are like. be sure to also run the windows memory test (much shorter, but different type of test), and disc check. Run a registry cleaner (I like CCleaner) as bad registries can cause issues like this (however unlikely).
May 4, 2012 6:51:46 PM

Probably a driver then. Now the problem is figuring out WHICH driver is at fault...
April 26, 2013 8:09:38 PM

I have this exact problem at the moment, and since searching on google led me here, I thought I would add the following...

Nirsoft have a freeware program called bluescreenview and it will highlight the drivers responsible for the crashes.

By he look of it, the culprit seems to be my Radeon HD card driver. Not sure why, but since I installed a new Radeon 6670 to replace a failed 5670 and have been having the problem ever since, although I suspect its while im using the Flash plugin container as it only happens when im browsing and not when I'm gaming

http://www.nirsoft.net/ for the program

I am not affiliated with them, but happily use their product
!