virtualbox.org

michaln wrote:.....If you can upload it somewhere, please PM me with details.

Can't yet PM you michaln - I need some few days more on this forum.

Link for testapp etc. now ready for download. Please need email address to send info.

Was there ever a fix for this? I'm hitting this error in an XP 32bit emulator, and on the same application it crashes at the same point for a 98SE install, so guessing it's the same, though it just hangs and does not throw an error.
Host is 7-64bit.
I'm on 4.3.10 r93012

Getting this app running was the whole reason I'm working with VM, time to retire the 15 year old laptop, but I need this app for work.

This looks like a 16bit crash. I'd be surprised if you get that in XP, though maybe in the Win16 WOW layer, or whatever it was called.

Tucker wrote:Was there ever a fix for this? I'm hitting this error in an XP 32bit emulator, and on the same application it crashes at the same point for a 98SE install, so guessing it's the same, though it just hangs and does not throw an error.

There is currently no fix because so far no one supplied a testcase which would allow to easily reproduce the problem. If you have something, please use https://www.virtualbox.org/ticket/12646 to provide it.

Should I attach the .vdi for the guest (quite large) or an .iso to install (20mb or so) of the application to the ticket? The application crashes with this error every time I try and run it, so repeatable is certainly the case for me.

Tucker wrote:Should I attach the .vdi for the guest (quite large) or an .iso to install (20mb or so) of the application to the ticket? The application crashes with this error every time I try and run it, so repeatable is certainly the case for me.

You won't be able to attach a file that big, unfortunately. I think an ISO would be preferable, with instructions on how to reproduce unless it's really obvious. Do you have the option to put it up on Dropbox or similar? You can PM me a link and I will grab it.

Since I never got anything...

We're still looking for a testcase which reproducibly shows the GPF. Until then, we won't be spending any time on this issue.

This should be fixed in the next VirtualBox maintenance version.

Thanks Ramshanker. I looked at the Changeset for the fix and it looks like you put in a lot of hard work - much appreciated (not saying I understood it all but I got the general idea). I recently got a chance to get back to looking at this with WinDbg. I'm not an expert at using it but I did see the GP happened in code dealing with save\restore of MPU state so your fix is in line with that. I have three test cases to try - a 16 bit Borland app, a 16 bit VB4 app, and VB.EXE (VB4 16 bit IDE - rare crash but I've seen it a few times). I actively maintain the 16 bit apps (too large\expensive to rewrite at this time), so the fix is welcome and I'll let you know how it went as soon as the release is available. (Sorry, but I wasn't allowed to send these apps.)

If you have not found a fix I have one for you to try. I just posted this on my blog, and this will work for you. The issue is the math coprocessor and 16 bit apps with the newer processors causes the GPF. This works with 32 bit VM's and it has been successfull on vmware fusion and Virtualbox. The link is here http://www.lifeofageekadmin.com/fix-gen ... fault-mod/ but here are the steps.

Download winfloat fromhttp://www.conradshome.com/win31/archive/.
Click on winfloat.exe to download to a directory. Once downloaded double-click to extract the files.
Copy HIDE87.com to c:\windows\system32 directory
Add to lh c:\windows\system32\autoexec.nt file as the first line.
Reboot virtual machine

Try running your 16-bit application. Hope this helps!

I can confirm the HIDE87.com trick solves the problem because it stops using the MPU so effectively it avoids the problem. It also really slows down everything (unbearably for what I work on - Engineering app with lots of calculations). A proper fix is always better, but agreed that this solution will get you going until a fix is in place.

The fix is in version 4.3.16 released yesterday. Really bad timing for such workarounds.

I installed version 4.3.16 and still the same problem I believe has not been resolved yet.

Actually, which problem is that exactly? What host OS, guest OS, and application?

I'm using an application called Siscomex, is a system of import in Brazil pulled one print of the error and tried to attach the file.