Memory leak in kernel after having installed Linux

[Son Goku]

Son of a... This makes no sense. Had winXP Pro SP2 running here, all was going well. Then I take everything off my e: drive (second physical HD in the system), moved it to d: (second partition on the first), and changed all associations for programs to point to the first HD.

Installed Linux, and all was stable though I did have the few problems I mentioned, aka getting cedega to compile under 64-bit Linux, etc...

Boot back into Windows, and I wake up to it having crashed. Caused me some lost crunch time in distributed computing projects as well... Then again yesterday it did so, then last night things became rather slow/unresponsive in games. When that was happening in game, it also trashed one of my WUs, one that takes about 800 hours or so to complete, but gives over 16,000 credits... The WU errored out as "file couldn't be read" or computer couldn't access the app to process it or something, right when system perf nearly died...

What's happening now is the paged pool useage in the system kernel is growing while this happens, and there's a system wide 192 MB limit on it in winXP. When things became slow, it had grown to about 125 MB. When I started it last night it was about 20 MB, now it's about 38 MB...

But wtf? Simply installing Linux on a second partition shouldn't be an issue with the Windows kernel should it? And though it includes an NTFS driver, it's set to read only mode (mounts as read only in Linux), so it shouldn't change anything on the partition...

Ran Adaware in case there was something there, my system came up clean. Did a chkdsk d: /f just in case, it rebooted, checked it, and came up clean. Ran an AV check in kaspersky, not finding something. The drivers themselves have not been changed. WTF?

[neo86]

Son of a... This makes no sense. Had winXP Pro SP2 running here, all was going well. Then I take everything off my e: drive (second physical HD in the system), moved it to d: (second partition on the first), and changed all associations for programs to point to the first HD.


What exactly do you mean by that?:confused: You used a tool on the Windows registry to change all references to the previous drive letter to the new one or did you just change shortcut references?

[gnutux]

No it should not, I have Windows and Linux running on the same machine, same hdd in fact, but different partitions. It must be a Windows registry misconfiguration because on Linux, I never heard that it would crash Windows.

gnutux

[Son Goku]

Actually, I didn't have too many links to the e: drive in the first place, as the second HD is 9.1 GB, and I have enough data.... Well I could probably install a couple programs, but eh... Drive's too small.

The only thing that was linked was the swapfile was split between the 2 drivers (which I changed in My Computer/Properties, performance options back to the d: drive before doing anything, moved data, and the programs set to use the data I pointed back to the appropriate folders.

There were no apps to have shortcuts and the like too...

I wouldn't exactly run some unknown util on the registry, as that can be iffy. I'd be more confident going in there myself...though I'm not unfamiliar with the registry structure and how it's organized; though obviously there are values any person would have to look up, where the text isn't human readable (depends on the value type, etc...)

E: didn't really show up by itself much at all in regedt32 when I just checked. Unfortunately some d: did come up on the search, I guess it doesn't sort the drive letter out too much. About the only 2 places were

- Kaspersky lab exlude folder for the link to the Oracle 9i installer. Had that for a class, and anyone who's seen a virus check on those compressed files for an Oracle Enterprise edition install... Haven't used them since DBA-1 class however...

However, in Kaspersky's interface I had already switched that over to d:...(same path would have followed), so hmm...

- The listing of where it had imported the bookmarks from, to get them into IE.

Other then that, no associatians had realy been made, except some disk management stuff listing "there was a drive there" sorta thing...

And no, I've never seen this sort of thing before, and though I haven't installed Linux in awhile (at home, though I have on the machines at uni), I have used it. Baffles me, but that's the only thing I changed when it went from a known working state, to a change in state where the kernel memory being allocated just suddenly turned around and grew for no explicable reason. There's the possibility it was a coincedence, though that was the only known change to have occured between the "before" and the "after". But yes, it doesn't seem to logical this should have happened having partitioned off the drive and installed Linux...