BSOD - Bad_Pool_Header when using WB 5.01 Beta

What graphics card do you have?

If changing drivers does not help then you may have a faulty video card which is why it continues to happen.

An application cannot be responsible for a BSOD crash.  Buggy drivers / faulty hardware are the most common cause.  Changing to WB 4.6 will result in WB using different apis (4.6 does not support per pixel frames for example)

I've got this installed on an IBM thinkpad T42P with an ATI FireGL T2 128MB.

I've run full system diagnostics, and everything has passed including my video card. I have not had this problem since uninstalling WB 5.01, and prior to that my system was blue screening several times a day. I've rigorously tested it since removing 5.01 and reinstalling 4.6 without any problems. Prior to installing 5.01 I never had this error. I'm sure you know more about your software than I do, but I've had over 12 years of IT experience, so I'm fairly confident that I'm correct. I'd be more than happy to provide you with more technical information if it would be of use.

Again, I've tested everything, and verified that all drivers and software are fully updated. Since 5.01 is still in beta, and I've ruled out other potential problems, I can think of no other cause for this BSOD.

Cheers,
Jason

As I previously said, an application cannot cause a BSOD in a well designed operating system due to apps running in user mode & drivers in kernel mode.  When Windows decides it knows what caused the crash and it blames the drivers, it seems right to assume that the drivers are at fault here.

The drivers are either crashing due to a bug in the drivers, or the hardware doing something unexpected (usually a sign of failure).

WB being in beta makes no difference here.  A beta WB could crash an app, or all apps, but never cause a BAD_POOL_HEADER BSOD unless by chance it exposes an existing bug in the video driver. 

As a I guess I would say the drivers are probably having a problem with alpha blending WB uses.  5.01 betas have improved performance on ATI boards as we worked around an ATI bug, but this will be making the hardware work a little harder.

I would suggest sending the minidump to ATI who should be able to track down the bug in their drivers which is causing your machine to BSOD.  As your card is a FireGL, they might treat it as higher priority than a standard Radeon crash.  Having said that, it is based on a 9600 which is a very old card these days.

Basically there is nothing we can do to resolve this as the problem is clearly not with WB.  WB, like all apps, requires a stable OS and stable drivers.  You are likely to experience the same with WindowFX and depending on the root cause, possibly with Windows Vista too.

Ok, well as I've said.... the problem went away when I removed WB 5.01... I've never had an issue with any other application, and have a very stable system other than the problems that occured with WB 5.01 installed. I posted this information because I thought it would be of help. If it's a driver issue, it's a driver issue that only seems to occur with your software loaded.

I have verified the hardware is fine. I have tested this hardware with many other applications, and run several benchmarks on it without problems. I have WindowFX running without a problem. The ONLY thing that triggers this BSOD is WB 5.01 not any other version... you do the math.

You seem to be missing the most important bit.  Windows correctly told you the crash was in the video driver.  That should be all the evidence you need to prove the video driver has a problem.

WB does not directly speak to the video driver (only via GDI calls), so it is not WB passing bad values to the video driver or anything like that. 

As for why WB may be exposing this, WB makes extensive use of 2D apis which are used less by other apps.  Additionally due to the job WB has, it may have dozens of threads painting at once if the user has suddenly moved windows around.  If the driver bug is timing sensitive then it may only show if an app calls a certain sequence of apis within a certain time period for example.

I believe the 9600 series is known for having problems with AGP fast writes causing crashes.  Turning that off in CCC may help.  The only other workaround for you would be to switch to non per pixel skins which I suspect will help as 4.x seems happy for you.

My original advice still stands.  You need to contact the supplier of the video drivers and report the bug to them.  A minidump should be very useful for them to track down what the problem is.

It is a driver issue, which only appears with our software loaded. It is literally not possible for user-mode code (which is what WindowBlinds uses) to corrupt pool headers. It can just send instructions to drivers, which guarantee to perform them without doing bad things like corrupting pool headers - but, in this case, they failed to do so. Unfortunately just because we happen to be triggering the error does not mean that we can fix the code inside ATI's drivers that doesn't handle that situation correctly.

I think a metaphor would be useful here:

A company builds a bridge. Unfortunately, there is a defect in its manufacture causing a large hole halfway through on the left side. As luck would have it, though, all the traffic is going one way, comes from the USA and drives on the right, so the hole is never encountered. The people who build the bridge make a lot of money by charging tolls and everyone is happy (except perhaps the people who have to pay the tolls).

Then along comes some traffic from the UK that is required by law to drive on the left side of the road - and immediately falls through the hole. Was it really their fault that they drove on the left, or is it the fault of the bridge maker for only testing out the bridge (which should work anywhere) with US traffic?

In the end the only solutions are to stop UK people from driving down the road, or to fix the hole. The latter is the responsibility of the person who built the bridge, because only they know how to fix it.

I'm not trying to argue here, I'm just telling you my findings. Considering I have verified that everything I have control over seems to be fully operational with the latest updates, the only correlation I could find was WB 5.01. I do not know enough about how your product interacts with the drivers/hardware to make any educated statements about what specifically is causing it, I'm simply stating that:

A) I never had this problem prior to installing WB 5.01
I have the most current driver released by ATI for this card. and have performed extensive hardware diagnostics on my system.
C) The video drivers have been out for almost a year and have been well tested
D) Reverting to a previous version of WB resolve the issue.

I have no doubt that you are both making accurate statements, I am just pointing out that WB 5.01 does something that makes my system VERY unstable. When I refer to my system I'm also referring to the most current drivers released for my hardware. If there is a problem with the driver and WB 5.01 compatibility, I would assume you as developers of a said software would want to be informed of it. I know that I as an end user of your software would have loved to have found this thread and a proposed solution long ago instead of spending many hours troubleshooting the problem. There is no reason to get defensive. I'm not trying to point fingers, I'm just trying to inform the community of my experience, which is the purpose of these forums.

In addition, I highly doubt ATI is going to expend the resources to rewrite their drivers due to one application incompatibility, when I can provide no other examples of this problem. If you and your group are so certain that this is an issue with ATI's drivers, then it might be worthwhile to make note of it in a readme or FAQ or knowledgebase so that in the event other people must use the same driver, they at least have an explanation as to why they are having the problem.

Again, as I said in my second post, I would be happy to provide additional technical information if you feel it would be of use to you, I'm just trying to relay my experience and contribute to the community. I've gone through my system with a fine tooth comb trying to figure out another solution, but in the end, the only thing that worked was reverting to an earlier version. Thank you for taking the time to respond.

Cheers,
Jason

Jason: First off, no one thinks that you're doing anything more then reporting a problem and, I might add, doing it better then most. What the others are trying to tell you is you have a problem...the operative word being you. No one doubts WB5 is causing a problem on your machine, but it's only your machine. There are 1000's of ATI users here and in case you haven't noticed, only you, two Stardock guys and myself are in this thread....and we don't have the problem either. You are right about ATI in that they don't have a problem or their boards would be light up. So, for a moment, let's forget about WB and ATI and look for some reason as to why WB5 highlights an otherwise oblivious problem in your system.

Microsoft gives a generic 'bad pool' error as a video adapter error because, frankly, 90% of the time it is. A hot fix was put out in SP2 which I'm presuming you have so for a moment, assume you might be in that 10%. That means it could be chip drivers, sound drivers, corrupted file(s), DEP, Indexing Svc, memory, USB 2.0 ...you get the idea. Whatever it is, it's hardware related and it's Windows throwing up the error. I don't know that I would want WB5 on my lap that bad to go through all that if I was content enough with 4.6, but you paid for it so it might be one of those 'mission' things. If I had to guess I'd say since you can replicate the problem by manipulating files with explorer (Move, copy, delete, etc) that it's corrupted files. Bring up a List view of your drive, click on each folder and the one that pops the error is a winner. Also,if you use any app that uses/creates a virtual drive they could be a problem in themselves. If you're overclocking anything, stop looking.

As to the ATI drivers, Neil was right about the fast write issue and there was another issue with the VPU recovery console, but I'm more interested in what drivers you're using that are a year old and how did you remove them when you tried other drivers?

Update::

I have since installed 5.01.006, and have not had the problem since. This problem only seems to present itself with the beta version.

Just to avoid any longwinded debates again, I'm just posting this for reference purposes. Whether this is driver related or not, it only seems to occur when I have the beta version installed. The problem recurrs shortly after installing WB 5.01 beta.

If you are running with a FireGL T2 on a mobile system, you may experience similar results.