Bug 3077 - VirtualBox 4.3 freezes my system making it dead
: VirtualBox 4.3 freezes my system making it dead
Status: RESOLVED WORKSFORME
Product: Desktop Features
Classification: ROSA Desktop
Component: All
: unspecified
: All Linux
: Normal minor
: ---
Assigned To: Eugene Shatokhin
: ROSA Linux Bugs
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2013-11-05 11:30 MSK by Dmitry
Modified: 2014-09-22 17:08 MSD (History)
2 users (show)

See Also:
RPM Package:
ISO-related:
Bad POT generating:
Upstream: unknown


Attachments
kernel logs (netconsole) (12.53 KB, text/x-log)
2013-11-25 16:02 MSK, Dmitry
Details
Full dmesg log (63.55 KB, text/x-log)
2013-11-25 17:49 MSK, Dmitry
Details
One more log (65.99 KB, text/x-log)
2013-11-27 14:21 MSK, Dmitry
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Dmitry 2013-11-05 11:30:02 MSK
Description of problem:

After upgrade to VirtualBox 4.3 the system freezes during running VM. It is the hard freeze, no any hardware that is working. Fox example, cursor doesn't move, SysRq doesn't work, no problems in logs.
Comment 1 Dmitry 2013-11-05 11:30:37 MSK
Affected 3.8.x and 3.10.x kernels
Comment 2 Alexander Burmashev 2013-11-05 12:22:21 MSK
we do not have vbox 4.3 in official repositories yet and we do not support packages from external sources.
Comment 3 Dmitry 2013-11-25 16:02:41 MSK
Created attachment 2280 [details]
kernel logs (netconsole)

I've collected kernel logs using netconsole. Alex, if you got the same problem, you can attach your logs.
Comment 4 Dmitry 2013-11-25 17:49:27 MSK
Created attachment 2281 [details]
Full dmesg log

Full dmesg log
Comment 5 Denis Silakov 2013-11-26 10:03:08 MSK
Since you use VBox from upstream, I would suggest to send these logs to upstream.
Comment 6 Dmitry 2013-11-27 14:21:31 MSK
Created attachment 2324 [details]
One more log
Comment 7 Eugene Shatokhin 2013-11-28 11:13:57 MSK
Reported upstream:
https://www.virtualbox.org/ticket/12381

The crash occurs in a different place each time and the value of IP register of a faulting instruction is always bad, although different. Looks like a stack corruption because it is the most common way to jump to an invalid memory area.

Unfortunately, none of the logs shows a valid call trace, so it is unclear where the problem happens. 

Kmemcheck proved to be unusable either, it makes the system too slow to work with.

I'll try to find more suitable tools to analyze that problem.
Comment 8 Eugene Shatokhin 2014-09-22 17:08:02 MSD
The problem no longer shows up with VirtualBox 4.3.14 on my machine. Please reopen the bug if it still happens.