[Bug c/93537] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[Bug c/93537] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93537

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |needs-bisection

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I can't figure what fixed it ... But it seems to have been fixed in GCC 10.1.0.

[Bug c/93537] [11/12 Regression] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93537

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|gcc 9.2 takes a             |[11/12 Regression] gcc 9.2
                   |Segmentation Violation when |takes a Segmentation
                   |attached file is compiled   |Violation when attached
                   |                            |file is compiled
      Known to fail|                            |11.4.0, 12.1.0, 12.2.0,
                   |                            |9.1.0, 9.5.0
           Keywords|needs-bisection             |
      Known to work|10.1.0                      |8.1.0

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Actually it is still broken in GCC 12.1.0, 12.2.0 and 12.3.0 and 11.4.0. I
don't know why it was working in GCC 12.1.0. Seems to work in GCC 13.1.0 and
13.2.0 though ....

And on the trunk.

[Bug c/93537] [11/12 Regression] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93537

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Created attachment 57909
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57909&action=edit
Reduced testcase

This as reduced as I could get it for now.

[Bug c/93537] [11/12 Regression] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93537

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-invalid-code,
                   |                            |ice-on-valid-code,
                   |                            |needs-bisection
   Target Milestone|---                         |11.5

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Someone should debug this further. because I don't understand how it works in
GCC 10.x series but then fails in the same way again in GCC 11.x/12.x. and
started passing in GCC 13.x.

[Bug c/93537] [11/12 Regression] gcc 9.2 takes a Segmentation Violation when attached file is compiled

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93537

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2020-02-01 00:00:00         |2024-6-12
           Priority|P3                          |P2
      Known to work|                            |13.1.0, 8.4.0

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Re-confirmed on the 12 branch.

(gdb) up
#3  0x0000000000c5650e in pp_c_enumeration_constant (pp=0x3eef950, 
    e=<integer_cst 0x7ffff6a2b450>)
    at /space/rguenther/src/gcc-12-branch/gcc/c-family/c-pretty-print.cc:1068
1068          if (tree_int_cst_equal (DECL_INITIAL (TREE_VALUE (value)), e))
(gdb) l
1063
1064      /* Find the name of this constant.  */
1065      if ((pp->flags & pp_c_flag_gnu_v3) == 0)
1066        for (value = TYPE_VALUES (type); value != NULL_TREE;
1067             value = TREE_CHAIN (value))
1068          if (tree_int_cst_equal (DECL_INITIAL (TREE_VALUE (value)), e))
1069            break;
1070
1071      if (value != NULL_TREE)
1072        pp->id_expression (TREE_PURPOSE (value));
(gdb) p debug_tree (value)
 <tree_list 0x7ffff6b44cd0
    purpose <identifier_node 0x7ffff6b5f500 TS_HANDLE_EXTENSION
        symbol <const_decl 0x7ffff6b71000 TS_HANDLE_EXTENSION type
<enumeral_type 0x7ffff6b70540 TSGenType_E>
            VOID /tmp/t.c:2:28
            align:1 warn_if_not_align:0 initial <integer_cst 0x7ffff6b59468
99>>>
    value <integer_cst 0x7ffff6b59468 type <integer_type 0x7ffff6a275e8 int>
constant 99>>

I'd say we likely get the enumeral type wrong, we assume a CONST_DECL but
have

 <enumeral_type 0x7ffff6b707e0 TSGenTypeE asm_written unsigned SI
    size <integer_cst 0x7ffff6a0bf90 type <integer_type 0x7ffff6a270a8
bitsizetype> constant 32>
    unit-size <integer_cst 0x7ffff6a0bfa8 type <integer_type 0x7ffff6a27000
sizetype> constant 4>
    align:32 warn_if_not_align:0 symtab:-155847024 alias-set -1 canonical-type
0x7ffff6b70540 precision:32 min <integer_cst 0x7ffff6a0bfc0 0> max <integer_cst
0x7ffff6a0bf78 4294967295>
    values <tree_list 0x7ffff6b44cd0
        purpose <identifier_node 0x7ffff6b5f500 TS_HANDLE_EXTENSION symbol
<const_decl 0x7ffff6b71000 TS_HANDLE_EXTENSION>>
        value <integer_cst 0x7ffff6b59468 constant 99>>
    chain <type_decl 0x7ffff6a36688 D.1981>>