[Bug c/53105] New: Segmentation fault in is_gimple_min_invariant

[Bug c/53105] New: Segmentation fault in is_gimple_min_invariant

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53105

             Bug #: 53105
           Summary: Segmentation fault in is_gimple_min_invariant
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: dcb314@hotmail.com


Created attachment 27231
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27231
C source code

I just tried to compile the package bristol-0.60.9-2
on gcc-4.8 trunk dated 20120422 on an AMD x86_64 box.

The compiler said

thesermon.c: In function 'thesermon':
thesermon.c:264:1: internal compiler error: Segmentation fault
 thesermon(int samplecount, int sineform)
 ^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Here is a valgrind stack backtrace

==5749== Invalid read of size 2
==5749==    at 0x79E060: is_gimple_min_invariant(tree_node const*)
(gimple.c:2794)
==5749==    by 0xA6AC6E: _Z24expr_invariant_in_loop_pP4loopP9tree_node.part.91
(tree-ssa-loop-ivopts.c:1396)
==5749==    by 0xA6AC7D: _Z24expr_invariant_in_loop_pP4loopP9tree_node.part.91
(tree-ssa-loop-ivopts.c:1396)
==5749==    by 0xEE585B: vect_check_strided_load(gimple_statement_d*,
_loop_vec_info*, tree_node**, tree_node**) (tree-vect-data-refs.c:2730)
==5749==    by 0xEE64B4: vect_analyze_data_refs(_loop_vec_info*, _bb_vec_info*,
int*) (tree-vect-data-refs.c:3144)
==5749==    by 0xB318DA: vect_analyze_loop(loop*) (tree-vect-loop.c:1513)
==5749==    by 0xB459E3: vectorize_loops() (tree-vectorizer.c:203)
==5749==    by 0x8A622C: execute_one_pass(opt_pass*) (passes.c:2176)
==5749==    by 0x8A67E6: execute_pass_list(opt_pass*) (passes.c:2231)
==5749==    by 0x63644D: _ZL24tree_rest_of_compilationP11cgraph_node.isra.28
(cgraphunit.c:1860)
==5749==    by 0x639445: cgraph_expand_function(cgraph_node*)
(cgraphunit.c:1930)
==5749==    by 0x63B707: cgraph_optimize() (cgraphunit.c:1995)
==5749==  Address 0x0 is not stack'd, malloc'd or (recently) free'd

Preprocessed source code attached. Flags -O3 -ffast-math required.

[Bug tree-optimization/53105] Segmentation fault in is_gimple_min_invariant

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53105

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
          Component|c                           |tree-optimization
         Resolution|                            |FIXED

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-04-25 11:07:22 UTC ---
I cannot reproduce this, I believe it was fixed with

2012-04-22  Jan Hubicka  <jh@suse.cz>

        * tree-ssa-loop-ivopts.c (expr_invariant_in_loop_p): Bail out at NULL
        tree refs.