A problem with hal.dll+12a3b - Windows 7 x64

With you all recommending different products it is very confusing.
Click to expand...

Not our goal, sorry. We're just trying to ultimately say that SpeedFan is bad and you should use any slew of tools we're recommending.

Sysnative keeps report that my CSV files are invalid.
Click to expand...

Mostly anything that's not a standard format (like JPG, PNG, etc) will need to be zipped to allow attachment.

I shall provide a link to the memory.dmp when I have had time to zip it and upload it to DropBox.
Click to expand...

Look forward to it, and at this point, I'm pretty certain it's hardware as I think I mentioned before.
 
Hi,

Still having problems and still no solution. I have not upgraded to Windows 10, not sure if I will.

More BSOD's of the SPECIAL_POOL_CORRUPTION type, though they have decreased in frequency, since the pre Windows 10 updates were installed.

There is a new sysmptom, whilst the machine is sitting doing nothing or very little, explorer is running at 30% give or take, and the temps are up aroung high 60's. From the noise from the machine, this is the sort of activity it has been exhibiting whilst a BSOD screen was on display.

Patrick I am not convienced that it is a hardware problem. Again only from the noise level produced by the machine this sort of activity was accopanying BSOD events, since they started occuring. Such stress placed on the processor may have caused the degradation in it's performance.

I don't believe a software issue can be ruled out due to the varied nature of the BSOD types. It suggests to me that a bug is propagating through the drivers, not enough to cause immediate failure, but over time it does cause a failure.

Am I talking nonsense, if so please tell me why.
 
Verifier is enabled in the last crash dump, so we're actually seeing results worth looking into.

Code: 
6: kd> .bugcheck
Bugcheck code 000000C1
Arguments fffff980`c8764fe0 fffff980`c876428b 00000000`005bc024 00000000`00000032

4th arg. is identifying the crash was caused due to the caller freeing an address where nearby bytes located in the same page have a single bit error. So the first thing to look into is who the caller was.

Code: 
6: kd> knL
 # Child-SP          RetAddr           Call Site
00 fffff880`030777e8 fffff800`03323123 nt!KeBugCheckEx
01 fffff880`030777f0 fffff800`0339bfc3 nt!MiCheckSpecialPoolSlop+0x83
02 fffff880`03077830 fffff800`033c7985 nt!MmFreeSpecialPool+0x1d3
03 fffff880`03077970 fffff880`044f36c4 nt!ExAllocatePoolWithTag+0x16bd
04 fffff880`03077a20 fffff880`04590cd8 SRTSP64+0x226c4
05 fffff880`03077a50 fffff880`044ea164 SRTSP64+0xbfcd8
06 fffff880`03077a80 fffff880`0455da00 SRTSP64+0x19164
07 fffff880`03077ab0 fffff880`044ea993 SRTSP64+0x8ca00
08 fffff880`03077af0 fffff880`0455cd96 SRTSP64+0x19993
09 fffff880`03077b20 fffff880`0455c969 SRTSP64+0x8bd96
0a fffff880`03077bd0 fffff800`0352b456 SRTSP64+0x8b969
0b fffff880`03077c00 fffff800`032832c6 nt!PspSystemThreadStartup+0x5a
0c fffff880`03077c40 00000000`00000000 nt!KxStartSystemThread+0x16

SRTSP64 - Symantec's (Norton's) real-time storage protection driver (AutoProtect), calls nt!ExAllocatePoolWithTag+0x16bd, so we know it's our caller.

Get rid of Norton for now and keep me updated.
 
Perhaps I should also mention that my Norton Internet Security subscription expired, and I have replaced it with Norton Security. Probably the same code but a different name.
 
They both mostly use the same kernel drivers AFAIK with the exception of a few due to feature adjustments. But anyway, yeah, remove it and let me know.
 
Guaranteed way to cause a BSOD. Uninstall Norton Security. It happended both times, I uninstalled it in recent days. No MEMORY.DMP created in either case.
 
Another BSOD just now, Norton was uninstalled. the fault was IRQ_ something, I did not make a note of it as I expected the MEMORY.DMP file to be there, it wasn't.

I was using NetBeans to edit some Java code.

I am going to continue with the current setup until there is another BSOD and this time I'll photograph it.
 
Verifier still enabled, settings not changed.

Another BSOD, this time I got a picture,BSOD 20150803_0330.jpg
again no MEMORY.DMP. Though I have the AIDA64 logging. It makes interesting reading. Okay a few details of my usage of the machine I think might be relevant. The logging starts from 19:04 when I switched on the computer after the last crash. I spent 2 hours using NetBeans. At 21:00 approx I went to bed to try and sleep, awoke after a few hours and returned to the computer at approx 00:30, I was on the computer until about 01:30. During this period I noticed strange behaviour of the fans, they were continually speeding up and slowing down. I took a snip'itStange Explorer Behaviour.PNG of the task Manager performance tab.

When I returned to the computer at 03:40, it had crashed and the BSOD was displayed.

If you look at the temps in this zip file; View attachment AIDA64 Logging.zip. During the period from 01:10 until it crashed they were high, 60C range. Which they shouldn't be for an idle computer.

When I set up the logging, I did not know the AIDA64 package so I accepted defaults. The logging did not include the CPU utilisation figures. I have now remedied that. In future they will be included.
 
To be honest at this point I still feel it's hardware, likely either RAM or motherboard given past behavior and current bug check in the screenshot above. The fact that no dumps are generating after a crash is a pretty big red flag to provide truth to that assumption. If it was a software issue, it'd show like before.

The only thing I can really recommended on the software-side of the things is to stop using NetBeans completely and seeing if that stops the crashes. If it does, then we know it is a software issue, just a strange one.
 
Patrick
To be honest at this point I still feel it's hardware, likely either RAM or motherboard given past behavior and current bug check in the screenshot above. The fact that no dumps are generating after a crash is a pretty big red flag to provide truth to that assumption. If it was a software issue, it'd show like before.

The only thing I can really recommended on the software-side of the things is to stop using NetBeans completely and seeing if that stops the crashes. If it does, then we know it is a software issue, just a strange one.​
Click to expand...

Well Patrick to answer your comments, the motherboard has been replaced, IMHO that makes it the source of the problem unlikely. The memory has been tested extensively and on multiple occasions. It has always passed with no errors. As for your comment about NetBeans, I just happened to be using it when it happened. A crash is more likely to happen when the computer is doing nothing.

Which brings me to today's crash. It was a hang today. I returned to the PC at 22:50, after not using it all day, screen blank and dark, not active as it has been in past. No response to mouse or keyboard. BIG RED SWITCH time. When it had rebooted, there was no MEMORY.DMP.

I did get a good record from the AIDA64 logging, this time with CPU utilisation. As you can see the CPU utilisation starts climbing at 0906 approx, and stays at a medium level until the logging stops. Scenarios similar to this seems to have been happening all along, but today was the first time I got a good record of the events.

For the solution, I am looking at the climb in CPU utilisation when the machine is idle. What experience I have with software makes me think that the software is having a war, which neither side wins until the crash.

Does ANYBODY have any useful suggestions as to how to clear this?
 

Attachments

Hi ric80231 ^_^,

Do you still need help with this? In case you still need help, kindly post back. I would be notified via email and will try my best to follow up in 48 hours.

Also, not sure if this has been asked earlier or not, are you getting those crashes in Safe Mode as well?



-Pranav
 
No, The problem has not been solved. I just do not leave the machine on. As I have mentioned many times in the thread the problem most often occurs when the machine is left on. It seems to occur when it tries to hibernate or go to sleep.

IMHO, that means it is Microsoft's problem, you obviously are not testing software for when shutdown occurs adequately, plus the OS seems incapable of controlling errant software. I have provide as much information as I can, but no one seemed interested in carrying it forward and actually debugging the operating system.

I am left thinking that Microsoft is more interested in making money than providing a good service.
 
ric80231 said:
No, The problem has not been solved. I just do not leave the machine on. As I have mentioned many times in the thread the problem most often occurs when the machine is left on. It seems to occur when it tries to hibernate or go to sleep.

IMHO, that means it is Microsoft's problem, you obviously are not testing software for when shutdown occurs adequately, plus the OS seems incapable of controlling errant software. I have provide as much information as I can, but no one seemed interested in carrying it forward and actually debugging the operating system.

I am left thinking that Microsoft is more interested in making money than providing a good service.
Click to expand...

You realize of course that we are not Microsoft nor are we employed by them, everyone here is a volunteer donating our time to help other users.
 
Wrench97 said:
You realize of course that we are not Microsoft nor are we employed by them, everyone here is a volunteer donating our time to help other users.
Click to expand...

Well, Patrick is described as a Microsoft MVP is his tag, as are you. MVP usually stands for Most Valuable Person. How am I wrong in assuming that means you are prized by Microsoft?

Either way it is irrelevant, Microsoft has the problem and they are not addressing it. Windows is an operating sytstem that does not control it's drivers and applications, as a consquence, bugs like this appear and user's such as myself are simply left to endure the inconvience.
 
Actually, Microsoft does test these scenarios, but they test them with their own components. It is up to the driver vendors to test their drivers (as part of the WHQL logo process) to validate that they don't break the process. When Windows bugchecks, it does so because something bad happened, and Windows (or a driver) calls the KeBugCheck routines to shut down the machine to avoid potential hardware failure and/or data loss. The fact that you have bugcheck codes that are all over the spectrum, and driver verifier has found issues multiple times would indicate either you have one of the most unlucky installations of Windows, or you have hardware issues.

An easy way to test if this is hardware or software would be to install STOCK WINDOWS without any drivers other than what installs during setup (offline, so as to not pull anything from Windows Update either), and then stress test the system. If it still fails under heavy load, you have hardware issues. If things work, add drivers and retest - if things fail then, it's drivers. If things still pass, add software and keep retesting until either everything fails, or you are stable. At this point, you're simply blaming precisely because you cannot figure out why things keep failing. I empathize with you here, of course, but you have to understand that Windows is not recompiled every time it is installed, and there are literally hundreds of millions of installations out there that are not behaving the way you are experiencing it. At some point you have to agree that the most logical conclusion should also be the correct one - either everyone with Windows installed would be running into these same issues (and Microsoft would be fixing them post haste), or there's something special about your install. Since the Windows codebase would not be the special part, work forwards - drivers, software, or hardware would be potentially different from other users, and thus the likelihood is that's where the problem lies. It's always possible, however minute the chance may be, that there's a Windows bug you're triggering that throws all these different bugchecks and driver failures. It's just not likely.
 
ric80231 said:
Wrench97 said:
You realize of course that we are not Microsoft nor are we employed by them, everyone here is a volunteer donating our time to help other users.
Click to expand...

Well, Patrick is described as a Microsoft MVP is his tag, as are you. MVP usually stands for Most Valuable Person. How am I wrong in assuming that means you are prized by Microsoft?

Either way it is irrelevant, Microsoft has the problem and they are not addressing it. Windows is an operating sytstem that does not control it's drivers and applications, as a consquence, bugs like this appear and user's such as myself are simply left to endure the inconvience.
Click to expand...

Sysnative is completely independent of Microsoft and any of associated its products and/or services. The MVP award is given by Microsoft to independent volunteers to have shown technical excellence in online forums and through other mediums like blogs; MVPs don't endorse Microsoft or blindly support them either.

https://mvp.microsoft.com/en-us/
 

Has Sysnative Forums helped you? Please consider donating to help us support the site!

Back
Top