Windows XP constant 0x50 bugcheck after a recent bugcheck

Hi everyone,

I currently have a Compaq Presario 906US laptop. Originally, I was having an issue where it severely struggled resuming from stand by (throwing a different BSOD nearly every time). While try to diagnose the issue, I tried rebooting after a bug check, and after that point, I am no longer able to boot back into windows.

Just when it seems that the login screen should show up, I get a BSOD mentioning a PAGE_FAULT_IN_NONPAGED_AREA, stop error 0x00000050 (if anyone feels that the parameters may be of any use, I can post them, although I understand that they may not be of much help at the moment).

I've already burned a CD of Hiren's BootCD, and I ran a check disk, which didn't report any errors (oddly). Various memory tests also do not report any issues.

Is it possible that a critical system file has become corrupted? I'd really like to find out what file, instead of jumping straight to a repair install. I'm also willing to try out WinDbg if someone can walk me through it and show me what I'm looking for.

If anyone can help, that would be great.

 

Sorry if I'm bumping, I just have a recent development that I'd like to add.

In Hiren's BCD, I can consistently cause the BSOD to appear when I try to launch EzPcFix, then load the registry hives. The crash will occur about 6 seconds into the process of loading the hives.

Could the registry be corrupted?

 

Welcome to Majorgeeks tgp

If you have any details like the parameters, please go ahead and post them up; with BSOD's we need all the little details we can to be able to piece together what's been happening.

If you can access the minidumps from \Windows\Minidump, please upload some of the most recent. Zip them first.

Right now I'm thinking it might be bad data/blocks on the drive, maybe a bad data cable or weak/faulty PSU.

 

Thank you satrow, I'm glad to be here

Alright, here are the parameters from the most recent BSOD (more on how the bsod occurred later): (0xD284000, 0X0, 0X80802DA8, 0X0)

For some reason, it doesn't seem to want to create minidumps :\ It does report that it successfully dumped the kernel memory (I do have the small dumps selected in the debugging section of the system configuration), but the most recent minidump is from two days ago, before the 0x50 bug checks began occurring. Unless there is some other location wherein the memory is being dumped, I don't know where to look.

I'm very with you on that one at the moment.

Experimenting some more in Hiren's boot CD, I tried opening the offline registry hive(s) with PE Registry Editor. Part way through opening the software hive, the very same BSOD (minus the memory dumping) appeared. And this happened while running on the live cd, mind you.

I think that seems very indicative of a corrupted registry hive, although I've honestly never seen one before. Any recommendations for how I should proceed?

 

Disk/temp file cleaning software like CCleaner might be removing all minidumps, ensure that all such software is deactivated.

Some baseline info on 0x50:

You need to begin with hardware testing, I'd start with memory testing before moving on to hard drive diags: Memtest86+, this needs to be run for 7+ passes, ideally for each stick in each slot, on most laptops that would mean 7+ passes x4. Only a result with 0 fails would put the memory in the clear. (ImgBurn, Caliban's tutorials, Bootable memory diags.)

 

7+ passes? Yikes, that seems a little extreme considering that this issue is happening to a point of being predictable.

Anyhow, I was able to fix the issue by using a program called Registry Restore Wizard, included on Hiren's Boot CD. I had fortunately just turned on System Restore a few days before the crash, and RRW was able to restore my registry from a day before the crash. Turns out my SOFTWARE hive was corrupted! Now I'm making this post right from the laptop

Thank you for your help satrow!

 

Nice fix and well gambled, you obviously have a good feel for what's going in in your laptop.

Re the 7+ passes: bad RAM can often 'pass' several runs.