Patch: Omit frame pointers on x86

Andrew Haley aph@redhat.com
Thu Jun 2 14:49:00 GMT 2005


Boehm, Hans writes:
 > I'm not sure what impact this has on a build with --enable-gc-debug,
 > which will try to save a partial stack trace in each allocated object.
 > I suspect that in this context it usually uses the backtrace() function,
 > which may just be atrociously slow either way.

I just tried to build i386 with --enable-gc-debug, and it fails to
compile.  I had to do this to build:

Index: configure.ac
===================================================================
RCS file: /cvs/gcc/gcc/boehm-gc/configure.ac,v
retrieving revision 1.12
diff -c -2 -p -r1.12 configure.ac
*** configure.ac	16 Feb 2005 04:10:41 -0000	1.12
--- configure.ac	2 Jun 2005 08:33:47 -0000
*************** AC_ARG_ENABLE(gc-debug,
*** 439,443 ****
  	AC_DEFINE(MAKE_BACK_GRAPH,1)
  	AC_MSG_WARN("Client must not use -fomit-frame-pointer.")
! 	AC_DEFINE(SAVE_CALL_COUNT, 8, [number of call frames saved with objects allocated through the debugging interface])
        ;;
        i[3456]86-*-dgux*)
--- 439,443 ----
  	AC_DEFINE(MAKE_BACK_GRAPH,1)
  	AC_MSG_WARN("Client must not use -fomit-frame-pointer.")
! dnl	AC_DEFINE(SAVE_CALL_COUNT, 8, [number of call frames saved with objects allocated through the debugging interface])
        ;;
        i[3456]86-*-dgux*)

Andrew.



More information about the Java-patches mailing list