[Bug middle-end/57500] New: [4.9 Regression] Segfault in ipa_record_reference

[Bug middle-end/57500] New: [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

            Bug ID: 57500
           Summary: [4.9 Regression] Segfault in ipa_record_reference
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dje at gcc dot gnu.org

Symtab cleanups 2/17 introduces 45 new failures in libstdc++ testsuite on AIX.

#0 
_Z20ipa_record_referenceP15symtab_node_defS0_11ipa_ref_useP18gimple_statement_d
(referring_node=0x70d3ad58, referred_node=0x0, use_type=IPA_REF_ALIAS, 
    stmt=0x0) at /nasfarm/dje/src/src/gcc/ipa-ref.c:338
#1  0x10560b10 in _Z20symtab_resolve_aliasP15symtab_node_defS0_ (
    node=0x70d3ad58, target=0x0) at /nasfarm/dje/src/src/gcc/symtab.c:928
#2  0x10592028 in _Z32cgraph_process_same_body_aliasesv ()
    at /nasfarm/dje/src/src/gcc/cgraphunit.c:660
#3  0x106d2574 in _Z28cp_write_global_declarationsv ()
    at /nasfarm/dje/src/src/gcc/cp/decl2.c:3999
#4  0x10000b14 in _ZL12compile_filev ()
    at /nasfarm/dje/src/src/gcc/toplev.c:558
#5  0x10003734 in _Z11toplev_mainiPPc (argc=9, argv=0x2ff228a8)
    at /nasfarm/dje/src/src/gcc/toplev.c:1872
#6  0x10000410 in main (argc=<error reading variable>, 
    argv=<error reading variable>) at  _start_ :36

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

David Edelsohn <dje at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |powerpc-ibm-aix*
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-06-01
                 CC|                            |hubicka at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from David Edelsohn <dje at gcc dot gnu.org> ---
The failure seems to require -g debugging enabled so this also may be related
to AIX XCOFF stabs debugging.

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

--- Comment #2 from David Edelsohn <dje at gcc dot gnu.org> ---
Created attachment 30240
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=30240&action=edit
preprocessed testsuite/20_util/allocator_traits/members/construct.cc

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

--- Comment #3 from David Edelsohn <dje at gcc dot gnu.org> ---
Unoptimized backtrace:

Program received signal SIGSEGV, Segmentation fault.
0x1056d054 in _ZN7va_heap7reserveIP7ipa_refEEvRP3vecIT_S_8vl_embedEjb (
    v=@0x24: 0x0, reserve=1, exact=false) at /nasfarm/dje/src/src/gcc/vec.h:300
300       v = static_cast <vec<T, va_heap, vl_embed> *> (xrealloc (v, size));

#0  0x1056d054 in _ZN7va_heap7reserveIP7ipa_refEEvRP3vecIT_S_8vl_embedEjb (
    v=@0x24: 0x0, reserve=1, exact=false) at /nasfarm/dje/src/src/gcc/vec.h:300
#1  0x1056ce48 in _ZN3vecIP7ipa_ref7va_heap6vl_ptrE7reserveEjb (this=0x24, 
    nelems=1, exact=false) at /nasfarm/dje/src/src/gcc/vec.h:1468
#2  0x1056f690 in _ZN3vecIP7ipa_ref7va_heap6vl_ptrE9safe_pushERKS1_ (
    this=0x24, obj=@0x2ff225e8: 0x70a42ec8)
    at /nasfarm/dje/src/src/gcc/vec.h:1563
#3  0x1056e294 in
_Z20ipa_record_referenceP15symtab_node_defS0_11ipa_ref_useP18gimple_statement_d
(referring_node=0x70d3ad58, referred_node=0x0, 
    use_type=IPA_REF_ALIAS, stmt=0x0) at /nasfarm/dje/src/src/gcc/ipa-ref.c:53
#4  0x1056c4e4 in _Z20symtab_resolve_aliasP15symtab_node_defS0_ (
    node=0x70d3ad58, target=0x0) at /nasfarm/dje/src/src/gcc/symtab.c:928
#5  0x10b4add4 in _Z32cgraph_process_same_body_aliasesv ()
    at /nasfarm/dje/src/src/gcc/cgraphunit.c:660
#6  0x105594e8 in _Z28cp_write_global_declarationsv ()
    at /nasfarm/dje/src/src/gcc/cp/decl2.c:3999
#7  0x100023ac in _ZL12compile_filev ()
    at /nasfarm/dje/src/src/gcc/toplev.c:558
#8  0x10005658 in _ZL10do_compilev () at /nasfarm/dje/src/src/gcc/toplev.c:1872
#9  0x1000587c in _Z11toplev_mainiPPc (argc=4, argv=0x2ff22914)
    at /nasfarm/dje/src/src/gcc/toplev.c:1948
#10 0x10000410 in main (argc=<error reading variable>, 
    argv=<error reading variable>) at  _start_ :36

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

--- Comment #4 from David Edelsohn <dje at gcc dot gnu.org> ---
The failure occurs with or without -O1/-O2. I am invoking cc1plus with

-g -std=gnu++0x

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.9.0

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |hubicka at gcc dot gnu.org

--- Comment #5 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Mine. Seems alias target got lost somehow.

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

--- Comment #6 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
OK, the problem here is that the symbol is not created yet, but decl exists.
Why it happens with -g only is bit strange. The following should fix it however
Index: cgraphunit.c
===================================================================
--- cgraphunit.c        (revision 199591)
+++ cgraphunit.c        (working copy)
@@ -656,8 +656,11 @@ cgraph_process_same_body_aliases (void)
   symtab_node node;
   FOR_EACH_SYMBOL (node)
     if (node->symbol.cpp_implicit_alias && !node->symbol.analyzed)
-      symtab_resolve_alias (node,
-                           symtab_get_node (node->symbol.alias_target));
+      symtab_resolve_alias
+        (node,
+         TREE_CODE (node->symbol.alias_target) == VAR_DECL
+         ? (symtab_node)varpool_node_for_decl (node->symbol.alias_target)
+         : (symtab_node)cgraph_get_create_node (node->symbol.alias_target));
   cpp_implicit_aliases_done = true;
 }

[Bug middle-end/57500] [4.9 Regression] Segfault in ipa_record_reference

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57500

David Edelsohn <dje at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #7 from David Edelsohn <dje at gcc dot gnu.org> ---
The proposed patch fixes the regression.  I commmitted it with permission.