ARM abort() core files

ARM abort() core files

[Hammer, Tim]

About 3 years ago (August, 2008) there was a discussion here about (not) getting a backtrace from abort(3) on ARM:
     http://gcc.gnu.org/ml/gcc/2008-08/msg00060.html 

That thread discussed why core files generated from a call to abort() do not have a stack to review and some possible approaches for modifying the compiler. I can find no resolution to that discussion nor any other discussion of the topic.

Our toolchain vendor is currently only providing GCC 4.4 and I have verified the issue still exists. They informed me today that GCC 4.6 "may be better about it". Can someone confirm that a change has been made and where I can find more information about it?

Thanks!
--
.Tim 
Tim D. Hammer
Software Developer
Global Business & Services Group
Xerox Corporation
M/S 0111-01A
800 Phillips Road
Webster, NY 14580

Phone: 585/427-1684
Fax:      585/231-5596
Mail:     Tim.Hammer@xerox.com

Re: ARM abort() core files

[Ian Lance Taylor]

"Hammer, Tim" <Tim.Hammer@xerox.com> writes:

> About 3 years ago (August, 2008) there was a discussion here about (not) getting a backtrace from abort(3) on ARM:
>      http://gcc.gnu.org/ml/gcc/2008-08/msg00060.html 
>
> That thread discussed why core files generated from a call to abort() do not have a stack to review and some possible approaches for modifying the compiler. I can find no resolution to that discussion nor any other discussion of the topic.
>
> Our toolchain vendor is currently only providing GCC 4.4 and I have verified the issue still exists. They informed me today that GCC 4.6 "may be better about it". Can someone confirm that a change has been made and where I can find more information about it?


This message would have been more appropriate for the mailing list
gcc-help@gcc.gnu.org.  Please take any followups to gcc-help.  Thanks.

This problem was fixed in the gcc 4.5 release with this patch:

http://gcc.gnu.org/ml/gcc-patches/2009-11/msg01234.html

Ian