[Bug c++/102319] New: [12 Regression] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[Bug c++/102319] New: [12 Regression] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102319

            Bug ID: 102319
           Summary: [12 Regression] ICE in cpp_atomic_builtins, at
                    c-family/c-cppbuiltin.c:746 since
                    r12-3495-g76b75018b3d053a8
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marxin at gcc dot gnu.org
                CC: jason at gcc dot gnu.org
  Target Milestone: ---

I know that I chose weird param values, but still:

$ gcc
/home/marxin/Programming/gcc/gcc/testsuite/gcc.c-torture/execute/pr93908.c
--param=constructive-interference-size=0 -c
<built-in>: internal compiler error: in cpp_atomic_builtins, at
c-family/c-cppbuiltin.c:746
0x67b440 cpp_atomic_builtins
        /home/marxin/Programming/gcc/gcc/c-family/c-cppbuiltin.c:746
0x67b440 c_cpp_builtins(cpp_reader*)
        /home/marxin/Programming/gcc/gcc/c-family/c-cppbuiltin.c:1495
0x98e0de c_finish_options
        /home/marxin/Programming/gcc/gcc/c-family/c-opts.c:1485
0x9910b1 c_common_parse_file()
        /home/marxin/Programming/gcc/gcc/c-family/c-opts.c:1230
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

and on ppc64le, the following value fails:

$ ppc64le-linux-gnu-gcc
/home/marxin/Programming/gcc/gcc/testsuite/gcc.c-torture/execute/pr93908.c
--param=constructive-interference-size=2147483647
...
warning: NEW ICE #1: in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746
0xdeadbeef cpp_atomic_builtins
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-ppc64le/build/gcc/c-family/c-cppbuiltin.c:746
0xdeadbeef c_cpp_builtins(cpp_reader*)
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-ppc64le/build/gcc/c-family/c-cppbuiltin.c:1495
0xdeadbeef c_finish_options
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-ppc64le/build/gcc/c-family/c-opts.c:1485
0xdeadbeef c_common_parse_file()
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-ppc64le/build/gcc/c-family/c-opts.c:1230

[Bug c++/102319] [12 Regression] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102319

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-09-14
     Ever confirmed|0                           |1
               Host|                            |x86_64-linux-gnu
             Status|UNCONFIRMED                 |NEW
             Target|                            |x86_64-linux,
                   |                            |ppc64le-linux-gnu
   Target Milestone|---                         |12.0

[Bug c++/102319] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102319

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|ice-on-valid-code           |ice-on-invalid-code
            Summary|[12 Regression] ICE in      |ICE in cpp_atomic_builtins,
                   |cpp_atomic_builtins, at     |at
                   |c-family/c-cppbuiltin.c:746 |c-family/c-cppbuiltin.c:746
                   |since                       |since
                   |r12-3495-g76b75018b3d053a8  |r12-3495-g76b75018b3d053a8
   Target Milestone|12.0                        |---

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The param are new so this is not a regression.  Really this is invalid code.

[Bug c++/102319] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102319

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
  /* Macros for C++17 hardware interference size constants.  Either both or
     neither should be set.  */
  gcc_assert (!param_destruct_interfere_size
              == !param_construct_interfere_size);
When it is provided by user, assertion isn't the right thing.
So, either there should be an assertion that if
!global_opts.x_param_destruct_interfere_size
&& !global_opts.x_param_construct_interfere_size
either both are zero or both non-zero (verification that the default setting of
this in the backends set both to non-zero), or there can be a warning or error
otherwise that user explicitly set something that shouldn't be set that way,
or both (assertion for both non-explicit and error if one or both are
explicit).

[Bug c++/102319] ICE in cpp_atomic_builtins, at c-family/c-cppbuiltin.c:746 since r12-3495-g76b75018b3d053a8

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102319

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
@Jason: May I please ping this issue. It's still not addressed after
g:4320a4b717dcccddf230d0b944bfc5a7ae282508.