Bug 447534 – app-emulation/virtualbox-4.1.22-r1 - VBox/Runtime/common/string/strformatrt.cpp:829:19: internal compiler error: Segmentation fault

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Carsten Milkau 2012-12-16 23:32:49 UTC]

The merge of app-emulation/virtualbox-4.1.22-r1 fails with a segmentation fault on my machine (see build log, looks like gcc). Using gcc (Gentoo 4.6.3 p1.7, pie-0.5.2) 4.6.3. The crash is reproducible for me, and happens at the same location every time.

Reproducible: Always

Steps to Reproduce:
1. emerge '=app-emulation/virtualbox-4.1.22-r1'

Actual Results:  
build error:


Expected Results:  
Succesfull build

[Comment 1 Carsten Milkau 2012-12-16 23:33:48 UTC]

Created attachment 332520 [details]
Output of emerge --info

[Comment 2 Carsten Milkau 2012-12-16 23:36:50 UTC]

Created attachment 332522 [details]
build log

Oops, forgot two lines. Here's the build error:

/var/tmp/portage/app-emulation/virtualbox-4.1.22-r1/work/VirtualBox-4.1.22/src/VBox/Runtime/common/string/strformatrt.cpp: In function ‘rtstrFormatRt’:
/var/tmp/portage/app-emulation/virtualbox-4.1.22-r1/work/VirtualBox-4.1.22/src/VBox/Runtime/common/string/strformatrt.cpp:829:19: internal compiler error: Segmentation fault

[Comment 3 Lars Wendler (Polynomial-C) (RETIRED) 2012-12-23 18:13:33 UTC]

Segmentation faults usually rather indicate a hardware problem. Did you check your system memory (RAM) for errors?
There's sys-apps/memtest86+ for doing such checks.

[Comment 4 Carsten Milkau 2012-12-23 18:30:14 UTC]

Dont beat me but I'm extremely sceptical that memory errors cause such an extremely specific and (locally) perfectly reproducible error. Maybe a CPU bug … but I haven't heard of any for the i7 920. Also, the machine is under heavy load 24/7 and nothing else appears wrong.

Last time I did a memtest was after installing the memory (~1 year ago). The problem doesn't hurt enough to take that pain atm, and considering it doesn't seem a probably cause, unless you definitely want to rule this out, I'd like to just leave it open.

[Comment 5 Ben Kohler 2013-10-03 20:15:53 UTC]

(In reply to Carsten Milkau from comment #4)
> Dont beat me but I'm extremely sceptical that memory errors cause such an
> extremely specific and (locally) perfectly reproducible error. Maybe a CPU
> bug … but I haven't heard of any for the i7 920. Also, the machine is under
> heavy load 24/7 and nothing else appears wrong.
> 
> Last time I did a memtest was after installing the memory (~1 year ago). The
> problem doesn't hurt enough to take that pain atm, and considering it
> doesn't seem a probably cause, unless you definitely want to rule this out,
> I'd like to just leave it open.

Have you had a chance to try this again lately?  One possible cause of a reproducible segfault is a damaged/corrupted file, and by now you have probably upgraded/rebuilt most of the involved packages.  If it's still occurring in current versions, we can try to dig a bit deeper.

[Comment 6 Carsten Milkau 2013-10-05 13:00:52 UTC]

Does not appear in recent versions of the package compiled. Just mentioned it because a gcc crash seemed really unusual to me.

[Comment 7 Ben Kohler 2013-10-05 15:01:18 UTC]

I'm going to go ahead and close this, upstream is pretty active and there's a good chance that IF this was not a local issue, it was fixed in upstream.  Please reopen if you come across this again.

Thanks!