Re: optimization/4382: [3.2/3.3 regression] __builtin_{set,long}jmp with -O3 can crash the compiler

Re: optimization/4382: [3.2/3.3 regression] __builtin_{set,long}jmp with -O3 can crash the compiler

[bangerth]

Old Synopsis: __builtin_{set,long}jmp with -O3 can crash the compiler
New Synopsis: [3.2/3.3 regression] __builtin_{set,long}jmp with -O3 can crash the compiler

State-Changed-From-To: open->analyzed
State-Changed-By: bangerth
State-Changed-When: Fri Dec  6 15:55:04 2002
State-Changed-Why:
    Confirmed. I get with 3.2.2pre and 3.3pre:
    tmp/g> /home/bangerth/bin/gcc-3.2.2-pre/bin/gcc -c -O3 x.c
    x.c: In function `main':
    x.c:21: Internal compiler error in make_edges, at cfgbuild.c:402
    Please submit a full bug report,
    with preprocessed source if appropriate.
    See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
    tmp/g>
    tmp/g>
    tmp/g>
    tmp/g> /home/bangerth/bin/gcc-3.3-pre/bin/gcc -c -O3 x.c
    x.c: In function `main':
    x.c:21: error: Wrong amount of branch edges after unconditional jump 0
    x.c:21: internal compiler error: verify_flow_info failed
    Please submit a full bug report,
    with preprocessed source if appropriate.
    See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
    
    
    This used to work with 2.95, but failed with 3.0 first.
    
    The following slighty simpler code also crashes the compiler,
    but please don't anyone ask me whether this makes any
    sense any more:
    -------------------------------
    void jumpaway(int *ptr) {
      __builtin_longjmp(ptr,1);
    }
    
    int main(void) {
      __builtin_setjmp(0);
      jumpaway(0);
    }
    --------------------------

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=4382

Re: optimization/4382: [3.2/3.3 regression] __builtin_{set,long}jmp with -O3 can crash the compiler

[Janis Johnson]

The following reply was made to PR optimization/4382; it has been noted by GNATS.

From: Janis Johnson <janis187@us.ibm.com>
To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
   nobody@gcc.gnu.org, malekith@pld.org.pl
Cc:  
Subject: Re: optimization/4382: [3.2/3.3 regression] __builtin_{set,long}jmp with 
 -O3 can crash the compiler
Date: Thu, 19 Dec 2002 15:26:15 -0800

 The regression reported in PR optimization/4382 showed up
 starting with this patch:
 
 2000-12-20  Richard Henderson  <rth@redhat.com>
 
         * rtl.h (REG_NON_LOCAL_GOTO): New.
         * rtl.c (reg_note_name): Update.
         * stmt.c (expand_goto): Emit a REG_NON_LOCAL_GOTO note.
         * builtins.c (expand_builtin_longjmp): Likewise.
         * flow.c (make_edges): Check for REG_NON_LOCAL_GOTO and do
         not emit an edge.
 
 I used the original, larger test case provided by the
 submitter.
 
 Output from the compiler with the current mainline is the
 same as that reported by Wolfgang a couple of weeks ago.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=4382