[Bug c++/105186] New: ICE in canonicalize_attr_name, at attribs.h:146

[Bug c++/105186] New: ICE in canonicalize_attr_name, at attribs.h:146

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

            Bug ID: 105186
           Summary: ICE in canonicalize_attr_name, at attribs.h:146
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Affects versions down to at least r5 :


$ cat z1.cc
__attribute__((__int128))


$ g++-12-20220403 -c z1.cc
z1.cc:1:16: internal compiler error: Segmentation fault
    1 | __attribute__((__int128))
      |                ^~~~~~~~
0x10d3b0f crash_signal
        ../../gcc/toplev.cc:322
0x9035f5 tree_check(tree_node*, char const*, int, char const*, tree_code)
        ../../gcc/tree.h:3456
0x9035f5 canonicalize_attr_name
        ../../gcc/attribs.h:146
0x9035f5 cp_parser_gnu_attribute_list
        ../../gcc/cp/parser.cc:28737
0x903831 cp_parser_gnu_attributes_opt
        ../../gcc/cp/parser.cc:28662
0x94439c cp_parser_declaration
        ../../gcc/cp/parser.cc:14844
0x94512d cp_parser_translation_unit
        ../../gcc/cp/parser.cc:5010
0x94512d c_parse_file()
        ../../gcc/cp/parser.cc:48057
0xad39a2 c_common_parse_file()
        ../../gcc/c-family/c-opts.cc:1240

[Bug c++/105186] ICE in canonicalize_attr_name, at attribs.h:146

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mpolacek at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-04-06
     Ever confirmed|0                           |1

--- Comment #1 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Confirmed.  Somewhat better test:

__attribute__((__int128)) int i;

[Bug c++/105186] [9/10/11/12 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2
            Summary|ICE in                      |[9/10/11/12 Regression] ICE
                   |canonicalize_attr_name, at  |in canonicalize_attr_name,
                   |attribs.h:146               |at attribs.h:146
                 CC|                            |jakub at gcc dot gnu.org
   Target Milestone|---                         |9.5

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r5-4106-g78a7c3172fe2e6cd959abb8bfc69f6b0dc747d49

[Bug c++/105186] [9/10/11/12 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 52769
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52769&action=edit
gcc12-pr105186.patch

Untested fix.

[Bug c++/105186] [9/10/11/12 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:083e8e66d2e90992fa83a53bfc3553dfa91abda1

commit r12-8077-g083e8e66d2e90992fa83a53bfc3553dfa91abda1
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Apr 11 10:41:07 2022 +0200

    c-family: Initialize ridpointers for __int128 etc. [PR105186]

    The following testcase ICEs with C++ and is incorrectly rejected with C.
    The reason is that both FEs use ridpointers identifiers for CPP_KEYWORD
    and value or u.value for CPP_NAME e.g. when parsing attributes or OpenMP
    directives etc., like:
             /* Save away the identifier that indicates which attribute
                this is.  */
             identifier = (token->type == CPP_KEYWORD)
               /* For keywords, use the canonical spelling, not the
                  parsed identifier.  */
               ? ridpointers[(int) token->keyword]
               : id_token->u.value;

             identifier = canonicalize_attr_name (identifier);
    I've tried to change those to use ridpointers only if non-NULL and
otherwise
    use the value/u.value even for CPP_KEYWORDS, but that was a large 10 hunks
    patch.

    The following patch instead just initializes ridpointers for the __intNN
    keywords.  It can't be done earlier before we record_builtin_type as there
    are 2 different spellings and if we initialize those ridpointers early, the
    second record_builtin_type fails miserably.

    2022-04-11  Jakub Jelinek  <jakub@redhat.com>

            PR c++/105186
            * c-common.cc (c_common_nodes_and_builtins): After registering
__int%d
            and __int%d__ builtin types, initialize corresponding ridpointers
            entry.

            * c-c++-common/pr105186.c: New test.

[Bug c++/105186] [9/10/11 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11/12 Regression] ICE |[9/10/11 Regression] ICE in
                   |in canonicalize_attr_name,  |canonicalize_attr_name, at
                   |at attribs.h:146            |attribs.h:146

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed on the trunk so far.

[Bug c++/105186] [9/10/11 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:473f2b099bdb3e72591b37f42e566086792f453d

commit r11-9847-g473f2b099bdb3e72591b37f42e566086792f453d
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Apr 11 10:41:07 2022 +0200

    c-family: Initialize ridpointers for __int128 etc. [PR105186]

    The following testcase ICEs with C++ and is incorrectly rejected with C.
    The reason is that both FEs use ridpointers identifiers for CPP_KEYWORD
    and value or u.value for CPP_NAME e.g. when parsing attributes or OpenMP
    directives etc., like:
             /* Save away the identifier that indicates which attribute
                this is.  */
             identifier = (token->type == CPP_KEYWORD)
               /* For keywords, use the canonical spelling, not the
                  parsed identifier.  */
               ? ridpointers[(int) token->keyword]
               : id_token->u.value;

             identifier = canonicalize_attr_name (identifier);
    I've tried to change those to use ridpointers only if non-NULL and
otherwise
    use the value/u.value even for CPP_KEYWORDS, but that was a large 10 hunks
    patch.

    The following patch instead just initializes ridpointers for the __intNN
    keywords.  It can't be done earlier before we record_builtin_type as there
    are 2 different spellings and if we initialize those ridpointers early, the
    second record_builtin_type fails miserably.

    2022-04-11  Jakub Jelinek  <jakub@redhat.com>

            PR c++/105186
            * c-common.c (c_common_nodes_and_builtins): After registering
__int%d
            and __int%d__ builtin types, initialize corresponding ridpointers
            entry.

            * c-c++-common/pr105186.c: New test.

    (cherry picked from commit 083e8e66d2e90992fa83a53bfc3553dfa91abda1)

[Bug c++/105186] [9/10 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11 Regression] ICE in |[9/10 Regression] ICE in
                   |canonicalize_attr_name, at  |canonicalize_attr_name, at
                   |attribs.h:146               |attribs.h:146

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed also for 11.3.

[Bug c++/105186] [9/10 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:73f5db083a73c6125e0203ea50c4d1f0a1c6421a

commit r10-10706-g73f5db083a73c6125e0203ea50c4d1f0a1c6421a
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Apr 11 10:41:07 2022 +0200

    c-family: Initialize ridpointers for __int128 etc. [PR105186]

    The following testcase ICEs with C++ and is incorrectly rejected with C.
    The reason is that both FEs use ridpointers identifiers for CPP_KEYWORD
    and value or u.value for CPP_NAME e.g. when parsing attributes or OpenMP
    directives etc., like:
             /* Save away the identifier that indicates which attribute
                this is.  */
             identifier = (token->type == CPP_KEYWORD)
               /* For keywords, use the canonical spelling, not the
                  parsed identifier.  */
               ? ridpointers[(int) token->keyword]
               : id_token->u.value;

             identifier = canonicalize_attr_name (identifier);
    I've tried to change those to use ridpointers only if non-NULL and
otherwise
    use the value/u.value even for CPP_KEYWORDS, but that was a large 10 hunks
    patch.

    The following patch instead just initializes ridpointers for the __intNN
    keywords.  It can't be done earlier before we record_builtin_type as there
    are 2 different spellings and if we initialize those ridpointers early, the
    second record_builtin_type fails miserably.

    2022-04-11  Jakub Jelinek  <jakub@redhat.com>

            PR c++/105186
            * c-common.c (c_common_nodes_and_builtins): After registering
__int%d
            and __int%d__ builtin types, initialize corresponding ridpointers
            entry.

            * c-c++-common/pr105186.c: New test.

    (cherry picked from commit 083e8e66d2e90992fa83a53bfc3553dfa91abda1)

[Bug c++/105186] [9/10 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:6547662a2afdac90311bcd286f87e98e6bec03d0

commit r9-10147-g6547662a2afdac90311bcd286f87e98e6bec03d0
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Apr 11 10:41:07 2022 +0200

    c-family: Initialize ridpointers for __int128 etc. [PR105186]

    The following testcase ICEs with C++ and is incorrectly rejected with C.
    The reason is that both FEs use ridpointers identifiers for CPP_KEYWORD
    and value or u.value for CPP_NAME e.g. when parsing attributes or OpenMP
    directives etc., like:
             /* Save away the identifier that indicates which attribute
                this is.  */
             identifier = (token->type == CPP_KEYWORD)
               /* For keywords, use the canonical spelling, not the
                  parsed identifier.  */
               ? ridpointers[(int) token->keyword]
               : id_token->u.value;

             identifier = canonicalize_attr_name (identifier);
    I've tried to change those to use ridpointers only if non-NULL and
otherwise
    use the value/u.value even for CPP_KEYWORDS, but that was a large 10 hunks
    patch.

    The following patch instead just initializes ridpointers for the __intNN
    keywords.  It can't be done earlier before we record_builtin_type as there
    are 2 different spellings and if we initialize those ridpointers early, the
    second record_builtin_type fails miserably.

    2022-04-11  Jakub Jelinek  <jakub@redhat.com>

            PR c++/105186
            * c-common.c (c_common_nodes_and_builtins): After registering
__int%d
            and __int%d__ builtin types, initialize corresponding ridpointers
            entry.

            * c-c++-common/pr105186.c: New test.

    (cherry picked from commit 083e8e66d2e90992fa83a53bfc3553dfa91abda1)

[Bug c++/105186] [9/10 Regression] ICE in canonicalize_attr_name, at attribs.h:146

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105186

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.