[Bug ipa/98000] New: g++-10 internal compiler error: verify_cgraph_node failed

[Bug ipa/98000] New: g++-10 internal compiler error: verify_cgraph_node failed

[ishikawa at yk dot rim.or.jp]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

            Bug ID: 98000
           Summary: g++-10 internal compiler error: verify_cgraph_node
                    failed
           Product: gcc
           Version: og10 (devel/omp/gcc-10)
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ishikawa at yk dot rim.or.jp
                CC: marxin at gcc dot gnu.org
  Target Milestone: ---

Created attachment 49631
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49631&action=edit
Gzipped preprocessed source file that led to the internal compiler error.

While compillng mozilla thunderbird mail client for local build, I get the
following error.

hh:1466:16: internal compiler error: verify_cgraph_node failed
0xdb670e cgraph_node::verify_node()
        ../../src/gcc/cgraph.c:3812
0xda9914 symtab_node::verify()
        ../../src/gcc/symtab.c:1312
0x12caa34 optimize_inline_calls(tree_node*)
        ../../src/gcc/tree-inline.c:5432
0x100c9a3 inline_transform(cgraph_node*)
        ../../src/gcc/ipa-inline-transform.c:736
0x115652d execute_one_ipa_transform_pass
        ../../src/gcc/passes.c:2233
0x115652d execute_all_ipa_transforms(bool)
        ../../src/gcc/passes.c:2272
0xdbc0d3 cgraph_node::expand()
        ../../src/gcc/cgraphunit.c:2293
0xdbd67f expand_all_functions
        ../../src/gcc/cgraphunit.c:2471
0xdbd67f symbol_table::compile()
        ../../src/gcc/cgraphunit.c:2834
0xdbfa6c symbol_table::compile()
        ../../src/gcc/cgraphunit.c:2747
0xdbfa6c symbol_table::finalize_compilation_unit()
        ../../src/gcc/cgraphunit.c:3014
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-10/README.Bugs> for instructions.


I am attaching the preprocessed source file.


The command that led to this bug is as follows (assuming the preprocessed file
is placed in /tmp )

/usr/bin/g++-10 -std=gnu++17 -o Unified_cpp_gfx_harfbuzz_src0.o -c 
-I/NEW-SSD/moz-obj-dir/objdir-tb3/dist/stl_wrappers
-I/NEW-SSD/moz-obj-dir/objdir-tb3/dist/system_wrappers -include
/NEW-SSD/NREF-COMM-CENTRAL/mozilla/config/gcc_hidden.h -U_FORTIFY_SOURCE
-D_FORTIFY_SOURCE=2 -fstack-protector-strong -UDEBUG '-DPACKAGE_VERSION="moz"'
'-DPACKAGE_BUGREPORT="http://bugzilla.mozilla.org/"' -DHAVE_OT=1 -DHAVE_ROUND=1
-DHB_NO_FALLBACK_SHAPE -DHB_NO_MT -DHB_NO_UCD -DHB_NO_UNICODE_FUNCS
-DMOZ_HAS_MOZGLUE -DMOZILLA_INTERNAL_API -DIMPL_LIBXUL
-DSTATIC_EXPORTABLE_JS_API
-I/NEW-SSD/NREF-COMM-CENTRAL/mozilla/gfx/harfbuzz/src
-I/NEW-SSD/moz-obj-dir/objdir-tb3/gfx/harfbuzz/src
-I/NEW-SSD/moz-obj-dir/objdir-tb3/dist/include
-I/NEW-SSD/moz-obj-dir/objdir-tb3/dist/include/nspr
-I/NEW-SSD/moz-obj-dir/objdir-tb3/dist/include/nss -fPIC -DMOZILLA_CLIENT
-include /NEW-SSD/moz-obj-dir/objdir-tb3/mozilla-config.h -Wall -Wempty-body
-Wignored-qualifiers -Woverloaded-virtual -Wpointer-arith -Wsign-compare
-Wtype-limits -Wunreachable-code -Wwrite-strings -Wno-invalid-offsetof
-Wc++2a-compat -Wduplicated-cond -Wimplicit-fallthrough -Wunused-function
-Wunused-variable -Wno-error=maybe-uninitialized
-Wno-error=deprecated-declarations -Wno-error=array-bounds
-Wno-error=coverage-mismatch -Wno-error=free-nonheap-object
-Wno-multistatement-macros -Wno-error=class-memaccess
-Wno-error=deprecated-copy -Wformat -Wformat-overflow=2 -Wno-psabi
-fno-sized-deallocation -fno-aligned-new -fno-builtin-strlen -Wl,--gdb-index
-Dfdatasync=fdatasync -DDEBUG_4GB_CHECK -DUSEHELGRIND=1 -DUSEVALGRIND=1 -DDEBUG
-g -gsplit-dwarf -Werror=sign-compare -Werror=unused-result
-Werror=unused-variable -Werror=format -fuse-ld=gold -fno-exceptions
-fno-strict-aliasing -fno-rtti -ffunction-sections -fdata-sections
-fno-exceptions -fno-math-errno -pthread -pipe -g -g -Og -fvar-tracking
-gdwarf-4 -fvar-tracking-assignments -freorder-blocks -fno-omit-frame-pointer
-funwind-tables -I/usr/include/glib-2.0
-I/usr/lib/x86_64-linux-gnu/glib-2.0/include  -MD -MP -MF
.deps/Unified_cpp_gfx_harfbuzz_src0.o.pp  -fdiagnostics-color 
Unified_cpp_gfx_harfbuzz_src0.cpp

[Bug ipa/98000] g++-10 internal compiler error: verify_cgraph_node failed

[ishikawa at yk dot rim.or.jp]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

--- Comment #1 from ishikawa,chiaki <ishikawa at yk dot rim.or.jp> ---
I noticed a similar Bug 97551.
But I seem to be using different options and I think I may be using a different
construct that triggers the ICE, and thus filed this entry.

I believe more reproducible cases would help.

[Bug ipa/98000] g++-10 internal compiler error: verify_cgraph_node failed

[ishikawa at yk dot rim.or.jp]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

--- Comment #2 from ishikawa,chiaki <ishikawa at yk dot rim.or.jp> ---
I forgot. The g++-10 version is as follows.

ishikawa@ip030:/NREF-COMM-CENTRAL/mozilla$ gcc --version
gcc (Debian 10.2.0-16) 10.2.0
Copyright (C) 2020 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

[Bug ipa/98000] g++-10 internal compiler error: verify_cgraph_node failed

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |marxin at gcc dot gnu.org
   Last reconfirmed|                            |2020-11-26
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
Thank you for the report, it's very likely a different issue.
I'm reducing that right now..

[Bug ipa/98000] g++-10 internal compiler error: verify_cgraph_node failed

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
Reduced test-case:

$ cat pr98000.C
struct {
  template <typename T> T *operator()(T);
} hb_addressof;
template <typename, typename, typename... Ts> static int _hb_cmp_method(Ts...)
{
  return 0;
}
template <typename V, typename K>
inline bool hb_bsearch_impl(K key, V, int compar(void *, void *)) {
  V p;
  int __trans_tmp_2 = *hb_addressof(key) = compar(&__trans_tmp_2, p);
  return false;
}
struct hb_array_t {
  int arrayZ;
};
struct hb_sorted_array_t : hb_array_t {
  void bsearch() {
    hb_bsearch_impl(bsearch_impl_x, &arrayZ, _hb_cmp_method<int, int>);
  }
  int bsearch_impl_x;
};
template <typename Returned> struct hb_lazy_loader_t {
  Returned *operator->();
};
namespace OT {
struct COLR;
}
struct hb_ot_face_t {
  hb_lazy_loader_t<OT::COLR> COLR;
};
struct {
  hb_ot_face_t table;
} hb_ot_color_glyph_get_layers_face;
namespace OT {
struct IntType {
  typedef int wide_type;
  operator wide_type();
};
struct UnsizedArrayOf;
struct OffsetTo {
  template <typename Base> friend UnsizedArrayOf operator+(Base, OffsetTo);
};
template <typename, bool> using LOffsetTo = OffsetTo;
template <typename Type> using LNNOffsetTo = LOffsetTo<Type, false>;
struct UnsizedArrayOf {
  void as_array(int);
};
struct SortedUnsizedArrayOf {
  hb_sorted_array_t __trans_tmp_4;
  void bsearch() { __trans_tmp_4.bsearch(); }
};
struct COLR {
  SortedUnsizedArrayOf __trans_tmp_3;
  int get_glyph_layers() {
    __trans_tmp_3.bsearch();
    int __trans_tmp_3 = numLayers;
    (this + layersZ).as_array(__trans_tmp_3);
  }
  LNNOffsetTo<int> layersZ;
  IntType numLayers;
};
} // namespace OT
void hb_ot_color_glyph_get_layers() {
  hb_ot_color_glyph_get_layers_face.table.COLR->get_glyph_layers();
}

fails with:
g++ pr98000.C -c -fchecking -Og

since r10-7306-g72b3bc895f023bf4

[Bug ipa/98000] [10/11 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |10.3
           Assignee|marxin at gcc dot gnu.org          |unassigned at gcc dot gnu.org
             Status|ASSIGNED                    |NEW
            Summary|g++-10 internal compiler    |[10/11 Regression] ICE
                   |error: verify_cgraph_node   |verify_cgraph_node failed
                   |failed                      |since
                   |                            |r10-7306-g72b3bc895f023bf4
                 CC|                            |hubicka at gcc dot gnu.org

[Bug ipa/98000] [10/11 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[ishikawa at yk dot rim.or.jp]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

--- Comment #5 from ishikawa,chiaki <ishikawa at yk dot rim.or.jp> ---
(In reply to Martin Liška from comment #3)
> Thank you for the report, it's very likely a different issue.
> I'm reducing that right now..

You are very welcome and
thank you for the reduction to simpler case in comment 4.

I used to work on compilers, interpreters, assemblers, etc. many moons ago, but
the current complexity of g++-10, and for that matter C++ and complex software
like mozilla thunderbird,  is beyond a mere user like myself, especially trying
to reduce the memory usage by means of "clone on demand" sounds a very hairy
approach (hairy as in the old Hacker's dictionary meaning).
I AM impressed.

Thank you again.
(I would have used "Happy Hacking" in the old days, but nowadays, hacking only
seems to be taken to carry a very negative sense. Sad.)

[Bug ipa/98000] [10/11 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to ishikawa,chiaki from comment #5)
> (In reply to Martin Liška from comment #3)
> > Thank you for the report, it's very likely a different issue.
> > I'm reducing that right now..
> 
> You are very welcome and
> thank you for the reduction to simpler case in comment 4.
> 
> I used to work on compilers, interpreters, assemblers, etc. many moons ago,
> but the current complexity of g++-10, and for that matter C++ and complex
> software like mozilla thunderbird,  is beyond a mere user like myself,
> especially trying to reduce the memory usage by means of "clone on demand"
> sounds a very hairy approach (hairy as in the old Hacker's dictionary
> meaning).

You're welcome. To be honest, the modern C++ is complicated even for a GCC
developer like me.

For the reduction, I used C-Vise (https://github.com/marxin/cvise), so an
automated tool for reduction. For bisection I use my script which uses
pre-built binaries for all versions of the GCC (from part few years).

> I AM impressed.
> 
> Thank you again.
> (I would have used "Happy Hacking" in the old days, but nowadays, hacking
> only seems to be taken to carry a very negative sense. Sad.)

I'm super-fine with that word. My manager uses it quite often and I like it.

[Bug ipa/98000] [10/11 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-checking,
                   |                            |ice-on-valid-code
           Priority|P3                          |P2

[Bug ipa/98000] [10/11 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.3                        |10.4

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10.3 is being released, retargeting bugs to GCC 10.4.

[Bug ipa/98000] [10/11/12/13 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.4                        |10.5

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 10.4 is being released, retargeting bugs to GCC 10.5.

[Bug ipa/98000] [11/12/13/14 Regression] ICE verify_cgraph_node failed since r10-7306-g72b3bc895f023bf4

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98000

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.5                        |11.5

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10 branch is being closed.