[Bug tree-optimization/103682] New: [12 regression] ICE on atomics: gimple check: expected gimple_assign(error_mark), have gimple_nop() in gimple_assign_rhs_code, at gimple.h:2852

[slyfox at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103682

            Bug ID: 103682
           Summary: [12 regression] ICE on atomics: gimple check: expected
                    gimple_assign(error_mark), have gimple_nop() in
                    gimple_assign_rhs_code, at gimple.h:2852
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: slyfox at gcc dot gnu.org
  Target Milestone: ---

ICE extracted from a real project:

$ cat bug.cpp

  #include <atomic>

  bool bug(std::atomic<unsigned> & ready, unsigned u) {
    return (ready.fetch_and(~u) & u);
  }

$ g++-12.0.0 -Ofast -c bug.cpp
during GIMPLE pass: fab
bug.cpp: In function 'bool bug(std::atomic<unsigned int>&, unsigned int)':
bug.cpp:6:6: internal compiler error: gimple check: expected
gimple_assign(error_mark), have gimple_nop() in gimple_assign_rhs_code, at
gimple.h:2852
    6 | bool bug(std::atomic<unsigned> & ready, unsigned u) {
      |      ^~~
0x20a58f7 internal_error(char const*, ...)
        ???:0
0x7c76dd gimple_check_failed(gimple const*, char const*, int, char const*,
gimple_code, tree_code)
        ???:0

$ g++-12.0.0 -v
Using built-in specs.
COLLECT_GCC=/nix/store/3x3yy5hhikzn8zvhq5fj14a2m9qyxbc8-gcc-12.0.0/bin/g++
COLLECT_LTO_WRAPPER=/nix/store/3x3yy5hhikzn8zvhq5fj14a2m9qyxbc8-gcc-12.0.0/libexec/gcc/x86_64-unknown-linux-gnu/12.0.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with:
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 12.0.0 20211212 (experimental) (GCC)