The page failed to load.. Dump File Plz Analyze

[chip]

I get the error "The page failed to load" in control panel, but for a strange reason, it only occurs on every 5th attempt at opening a setting.

After this error, explorer.exe will either freeze/crash (and I need to kill it) or it'll just sit there still trying to load the page but allows me to close the window without it crashing.

How I reproduce the error:

1. I open control panel
2. It doesn't really matter which setting I choose to open, but for testing purposes I open "Ease of Access Center"
3. The first 4 times I open Ease of Access Center it loads just fine
4. The 5th time it shows a blank page with that error message centered!?

This seems to happen only on the 5th time and it seems to apply to the majority of the settings in control panel. I am absolutely clueless on the fix. It doesn't do this in safe mode. I suspect it could be a driver or just something in the registry but I'd love to be able to narrow this down and find the culprit without reinstalling clean.

(Dump File Attached)
(Windows 7 Pro)

 

[H2SO4]

Kudos for the desire to get straight to the debugging phase of troubleshooting, but there's a problem- what you're experiencing is not technically a "crash" in the sense that a process terminated in an abnormal way.

Not every error message constitutes a crash.

As a wild guess, is your UxTheme.DLL "patched" perhaps?

 

[chip]

It shouldn't be, as I've never touched that file before. The only theme changes I made was disabling Aero, and changing the colors a bit. I can also replicate this error by creating a second user account (which I believe uses the default Aero theme). I've tried closing every program running in the background and the error still persists, and always on the 5th attempt. I am stuck at figuring this out!

 

[H2SO4]

It shouldn't be, as I've never touched that file before. The only theme changes I made was disabling Aero, and changing the colors a bit. Other than that, I can't think of anything else it could be. I can start uninstalling some files that effect the context menu as a precaution though. Eraser, WinRAR, 7-Zip, and AntiVir, but these are all widely used software and can't imagine that being the cause. =/

As a suggestion, remove "-noDumpOnFirst" from the command line you used to attach the debugger, and then repro the problem again. Actually, you should reboot first before doing anything else.

Without the "-noDumpOnFirst" switch, you might get some DMP files appearing at the time of the problem.

To cut a long story short, a "first-chance exception" is sort of like a half crash (I hate myself for using that analogy). Normally, first-chance exceptions are of very little interest when troubleshooting "real" (second-chance) crashes, hence the use of the "-noDumpOnFirst" switch, but in this instance it may actually be interesting to see those first-chance exception events.

 

[chip]

Tried to use a .Net Cleanup Tool just for the sake of trying it, but it didn't do any good. Also ran "sfc /scannow" and it shows no signs of problems with system files.

Rebooted. Error still exists. Created new dump without "-noDumpOnFirst"

I really appreciate your help, I have my fingers crossed hoping I don't have to reinstall 7.

 

[H2SO4]

Tried to use a .Net Cleanup Tool just for the sake of trying it, but it didn't do any good. Also ran "sfc /scannow" and it shows no signs of problems with system files.

Rebooted. Error still exists. Created new dump without "-noDumpOnFirst"

I really appreciate your help, I have my fingers crossed hoping I don't have to reinstall 7.

Hmm, still no dumpy goodness, just text files. We'd probably have to play with the debugger itself, but doing it over a web forum is like keyhole surgery without being able to see what you're doing, and I'm not confident that I'd be able to justify the time we'd spend on the basis of a guaranteed positive outcome.

Anyway, I can sort of see what's happening on a code level from the info in those log files: a COM object that's being called is somehow dying, although I can't tell which COM object without a dump.

From the two 1st-chance exception stacks in the log:

# ChildEBP RetAddr Args to Child
WARNING: Stack unwind information not available. Following frames may be wrong.
00 03c0ee80 77c98f48 800401fd 00000001 00000000 KERNELBASE!RaiseException+0x54
*** ERROR: Symbol file could not be found. Defaulted to export symbols for C:\Windows\system32\ole32.dll -
01 03c0eea4 778501cc 800401fd 03896290 00000001 RPCRT4!RpcRaiseException+0x4b
02 03c0eeb8 77d056bd 0819b34c 03c0eee4 03c0ef10 ole32!HRGN_UserMarshal+0x92e
03 03c0f2e4 7784ba02 03896290 038964c8 03c0f31c RPCRT4!NdrClientCall2+0xbd
04 03c0f304 7773c8b5 0000000c 0000000c 03c0f330 ole32!WdtpInterfacePointer_UserUnmarshal+0x60d
...

And:

# ChildEBP RetAddr Args to Child
WARNING: Stack unwind information not available. Following frames may be wrong.
00 0009e870 77c98f48 80010012 00000001 00000000 KERNELBASE!RaiseException+0x54
01 0009e894 7784a534 80010012 00000001 0009ecc8 RPCRT4!RpcRaiseException+0x4b
02 0009e8a4 77c9178b 081fb224 80010012 77d05744 ole32!StgGetIFillLockBytesOnFile+0x17204
03 0009ecc8 7784ba02 7775d338 7775bdd8 0009ed00 RPCRT4!NdrFixedArrayBufferSize+0x76a
04 0009ece8 7773c8b5 00000010 00000005 0009ed20 ole32!WdtpInterfacePointer_UserUnmarshal+0x60d

Those two exceptions map to the following:

800401fd = CO_E_OBJNOTCONNECTED

80010012 = RPC_E_SERVER_DIED_DNE

=======================================

Is there a COM or RPC error/warning event in the event logs at the time of the problem? (Run EVENTVWR to get to the event log.)

 

[chip]

Is there a COM or RPC error/warning event in the event logs at the time of the problem? (Run EVENTVWR to get to the event log.)

I don't see any COM, RPC, or any strange errors/warnings in the event viewer.

Steps Taken:
1. Open Event Viewer
2. Replicated Error
3. Looked within the Last hour under the Summary of Administrative Events
4. No errors within time frame of test. I don't see any COM, RPC, or any strange errors that may be related.

 

[H2SO4]

I don't see any COM, RPC, or any strange errors/warnings in the event viewer.

Unfortunately, that makes things more difficult as the limited amount of info doesn't reveal which COM object is going ga-ga.

As a suggestion, test what happens when you use MSCONFIG to specify a "diagnostic" startup and then reboot - can you still repro the symptom in that mode?

 

[chip]

As a suggestion, test what happens when you use MSCONFIG to specify a "diagnostic" startup and then reboot - can you still repro the symptom in that mode?

Yes, I can still repro the error in diagnostic mode. However, that is very odd because I can NOT repro the error in safe mode.

 

[H2SO4]

As a suggestion, test what happens when you use MSCONFIG to specify a "diagnostic" startup and then reboot - can you still repro the symptom in that mode?

Yes, I can still repro the error in diagnostic mode. However, that is very odd because I can NOT repro the error in safe mode.

Another unfortunate outcome.

Diagnostic startup is not quite as sparse as running in safe mode. Hence, the problem in this instance is obviously caused by one of the differences. Suggestions:

1) Test what happens when you disable Aero (pick the win7 basic theme). Is the error reproducible?

2) As a separate test, issue this command on an elevated (run as admin) CMD prompt and reboot:

BCDEDIT /SET VGA ON

After the reboot, you'll be in "basevideo" mode. It would be interesting to know whether the problem is still reproducible because it would mean it's got nothing to do with the video rendering/driver.

 

[chip]

1) Test what happens when you disable Aero (pick the Windows 7 basic theme). Is the error reproducible?

Still reproducible. Aero off/on, Basic, Classic, etc.

2) As a separate test, issue this command on an elevated (run as admin) CMD prompt and reboot:

BCDEDIT /SET VGA ON

Still reproducible, although a very clever suggestion.

 

[H2SO4]

Hmmm. This thing is not making itself easy to troubleshoot

It depends on how far you want to go with it - I'm sure we'll get there eventually, but I can't guarantee that it'll be quicker than simply reinstalling. What worries me a little is the way that it's fine most of the time. Intermittent problems are generally tougher to understand.

One possible "avoidance" strategy is to create another user account. If the problem is not manifested when logged on as a different user, then obviously it's something that's profile-specific.

Otherwise, if you'd like to persist with the debug angle, at least for a while, try attaching the debugger this way please:

cscript adplus.vbs -crash -FullOnFirst -quiet -pn explorer.exe

It may produce multiple dumps hundreds of MB in size, so I'll talk you through some commands in the debugger to distill what I need from all those data.

 

[chip]

Fixed !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Your BCDEDIT suggestion prompted me to disable the graphics card driver from within device manager and from that point I was error-free! (Although I didn't expect it to be any different from the BCDEDIT command, so it was a nice surprise)

Driver is an Intel 82945G. All this time I had thought it was the latest because it was installed by WU which said it was up to date. It was also newer than the one provided by Dell, but apparently there was a newer version and after installing it the error went away like magic! I've been maniacally clicking around my control panel trying to enjoy the moment before I perhaps wake up from a dream.

Thanks for sticking with me on this. I would've lost my mind days ago without your help! (I was nearly ready to reinstall, and it wouldl've led me right back to square 1.)

 

[H2SO4]

Fixed !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Your BCDEDIT suggestion prompted me to disable the graphics card driver from within device manager and from that point I was error-free! (Although I didn't expect it to be any different from the BCDEDIT command, so it was a nice surprise)

Driver is an Intel 82945G. All this time I had thought it was the latest because it was installed by WU which said it was up to date. It was also newer than the one provided by Dell, but apparently there was a newer version and after installing it the error went away like magic! I've been maniacally clicking around my control panel trying to enjoy the moment before I perhaps wake up from a dream.

Thanks for sticking with me on this. I would've lost my mind days ago without your help! (I was nearly ready to reinstall, and it wouldl've led me right back to square 1.)

Glad to hear it

Most of the information in the text files (regarding a 1st-chance exception) seemed to indicate some sort of screen repainting/rendering issue, hence the UxTheme.dll question.

My guess is that it's not the driver itself that's responsible (it also happened in basevideo), but something that it brings with it up in user-mode, although that's pure speculation.