0

I am trying to run valgrind with waf. The command is ./waf --command-template="valgrind %s" --run program-name.

I receive an internal error for valgrind. What else can I try if even valgrind fails ?

--6446-- VALGRIND INTERNAL ERROR: Valgrind received a signal 11 (SIGSEGV) - exiting
--6446-- si_code=80;  Faulting address: 0x0;  sp: 0x40b3c9df0

valgrind: the 'impossible' happened:
   Killed by fatal signal
==6446==    at 0x38058236: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x38021ADC: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x38021CCD: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x380902A7: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x3809F7D5: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x3809FA5A: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0x380BBEAD: ??? (in /usr/lib/valgrind/memcheck-amd64-linux)
==6446==    by 0xDEADBEEFDEADBEEE: ???
==6446==    by 0xDEADBEEFDEADBEEE: ???
==6446==    by 0xDEADBEEFDEADBEEE: ???
evaG
  • 21
  • 1
  • 6
  • 1
    See the [list of Valgrind bugs](https://bugs.kde.org/buglist.cgi?quicksearch=product%3Avalgrind) and submit a new one if necessary. – Hristo Iliev Jul 08 '12 at 21:47
  • do you have some really optimized libraries/code in your program? I got similar errors when I tried to run valgrind on code that linked to Intel Performance Primitives which uses some fancy memory/cpu instructions tricks under the hood to eek out a bit more speed. – tacaswell Jul 08 '12 at 22:12
  • Much of the code isn't mine so I wouldn't know. – evaG Jul 08 '12 at 22:23

1 Answers1

-1

I have not tried it for a long time but you can use the Boehm GC to detect leaks in C/C++ code.

http://www.hboehm.info/gc/leak.html

Bowie Owens
  • 2,798
  • 23
  • 20