[Bug c++/108967] New: internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[Bug c++/108967] New: internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[malat at debian dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

            Bug ID: 108967
           Summary: internal compiler error: in expand_debug_expr, at
                    cfgexpand.cc:5450
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: malat at debian dot org
  Target Milestone: ---

Created attachment 54555
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54555&action=edit
Preprocessed source

I cannot compile highway on arm64 with gcc-13 anymore.

Compilation fais with:

 % /usr/lib/gcc-snapshot/bin/g++  -freport-bug -DHWY_SHARED_DEFINE
-I"/home/malat/highway-1.0.4~git20230227.dc5c141" -g -O2
-ffile-prefix-map=/home/malat/highway-1.0.4~git20230227.dc5c141=.
-fstack-protector-strong -Wformat -Werror=format-security -DHWY_BROKEN_EMU128=0
-Wdate-time -D_FORTIFY_SOURCE=2 -fPIE -fvisibility=hidden
-fvisibility-inlines-hidden -Wno-builtin-macro-redefined
-D__DATE__=\"redacted\" -D__TIMESTAMP__=\"redacted\" -D__TIME__=\"redacted\"
-fmerge-all-constants -Wall -Wextra -Wconversion -Wsign-conversion -Wvla
-Wnon-virtual-dtor -fmath-errno -fno-exceptions -Werror -DHWY_IS_TEST=1
-DGTEST_HAS_PTHREAD=1 -MD -MT
CMakeFiles/arithmetic_test.dir/hwy/tests/arithmetic_test.cc.o -MF
CMakeFiles/arithmetic_test.dir/hwy/tests/arithmetic_test.cc.o.d -o
CMakeFiles/arithmetic_test.dir/hwy/tests/arithmetic_test.cc.o -c
"/home/malat/highway-1.0.4~git20230227.dc5c141/hwy/tests/arithmetic_test.cc"
 <widen_plus_hi_expr 0xffff69f0f898
    type <vector_type 0xffff333bf348
        type <integer_type 0xffff936c5930 public unsigned HI
            size <integer_cst 0xffffa66e91f8 constant 16>
            unit-size <integer_cst 0xffffa66e9210 constant 2>
            align:16 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0xffff936c5930 precision:16 min <integer_cst 0xffff93c04090 0> max <integer_cst
0xffff93c65bb8 65535>
            pointer_to_this <pointer_type 0xffff59de21f8>>
        unsigned V8HI
        size <integer_cst 0xffffa66e9090 constant 128>
        unit-size <integer_cst 0xffffa66e90a8 constant 16>
        align:128 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0xffff333bf348 nunits:8>
    readonly constant
    arg:0 <vector_cst 0xffff69ee8528
        type <vector_type 0xffff5d64c690 type <integer_type 0xffffa66f0348
unsigned char>
            unsigned V16QI size <integer_cst 0xffffa66e9090 128> unit-size
<integer_cst 0xffffa66e90a8 16>
            align:128 warn_if_not_align:0 symtab:0 alias-set 0 canonical-type
0xffff5d64c690 nunits:16
            pointer_to_this <pointer_type 0xffff31541000>>
        constant npatterns:1 nelts-per-pattern:3
        elt:0:  <integer_cst 0xffff9f9dd6c0 constant 1>
        elt:1:  <integer_cst 0xffff9927fd20 constant 2>
        elt:2:  <integer_cst 0xffff9927fd50 constant 3>>
    arg:1 <vector_cst 0xffff69ee8690 type <vector_type 0xffff5d64c690>
        constant npatterns:1 nelts-per-pattern:3
        elt:0:  <integer_cst 0xffffa66e9108 constant 255>
        elt:1:  <integer_cst 0xffffa66e9168 constant 0> elt:2:  <integer_cst
0xffff9f9dd6c0 1>>>
during RTL pass: expand
In file included from
/home/malat/highway-1.0.4~git20230227.dc5c141/hwy/foreach_target.h:59,
                 from
/home/malat/highway-1.0.4~git20230227.dc5c141/hwy/tests/arithmetic_test.cc:21:
/home/malat/highway-1.0.4~git20230227.dc5c141/hwy/tests/arithmetic_test.cc: In
function 'void hwy::N_EMU128::TestSaturatingArithmeticOverflow::operator()(T,
D) [with T = unsigned char; D = hwy::N_EMU128::Simd<unsigned char, 16, 0>]':
/home/malat/highway-1.0.4~git20230227.dc5c141/hwy/tests/arithmetic_test.cc:126:21:
internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450
  126 |   HWY_NOINLINE void operator()(T /*unused*/, D d) {
      |                     ^~~~~~~~
0x16c4083 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:5450
0x16c5343 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:5311
0x16c34ef expand_debug_expr
        ../../src/gcc/cfgexpand.cc:4490
0x16c5343 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:5311
0x16c3457 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:4463
0x16c5343 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:5311
0x16c35e7 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:4764
0x16c5343 expand_debug_expr
        ../../src/gcc/cfgexpand.cc:5311
0x13b0aa3 expand_debug_locations
        ../../src/gcc/cfgexpand.cc:5648
0x13b0aa3 execute
        ../../src/gcc/cfgexpand.cc:6825
Please submit a full bug report, with preprocessed source.
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-snapshot/README.Bugs> for instructions.
Preprocessed source stored into /tmp/ccEsr3zo.out file, please attach this to
your bugreport.

[Bug c++/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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

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

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Reproduced now, reducing and bisecting.

[Bug c++/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |aarch64
   Last reconfirmed|                            |2023-02-28
             Status|UNCONFIRMED                 |NEW
                 CC|                            |ktkachov at gcc dot gnu.org
     Ever confirmed|0                           |1
   Target Milestone|---                         |13.0
           Keywords|                            |ice-on-valid-code
      Known to fail|                            |13.0

--- Comment #2 from ktkachov at gcc dot gnu.org ---
Confirmed

[Bug c++/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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
           Keywords|                            |ice-checking

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
--- gcc/cfgexpand.cc.jj 2023-02-21 11:44:48.479464567 +0100
+++ gcc/cfgexpand.cc    2023-02-28 15:33:09.806750517 +0100
@@ -5365,6 +5365,10 @@ expand_debug_expr (tree exp)
     case VEC_WIDEN_MULT_ODD_EXPR:
     case VEC_WIDEN_LSHIFT_HI_EXPR:
     case VEC_WIDEN_LSHIFT_LO_EXPR:
+    case VEC_WIDEN_PLUS_HI_EXPR:
+    case VEC_WIDEN_PLUS_LO_EXPR:
+    case VEC_WIDEN_MINUS_HI_EXPR:
+    case VEC_WIDEN_MINUS_LO_EXPR:
     case VEC_PERM_EXPR:
     case VEC_DUPLICATE_EXPR:
     case VEC_SERIES_EXPR:
certainly fixes this, not done yet with reduction though (so far at 300KB now).
Haven't bisected yet, but if it reproduces only on the trunk, it would be
latent
since r11-5160-g9fc9573f9a5e9432e53c when those were introduced.  Though, it is
an checking ICE only, so doesn't reproduce with release branches by default
(and does the right thing then, the ICE is solely about making sure that new
trees which can be handled in expand_debug_expr are handled there).

[Bug c++/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Somewhat reduced for -O2 -g:

template <int N>
struct A {
  using T = unsigned short;
  static constexpr int a = N;
};
template <int N>
struct B { using type = A<N>; };
struct C {
  static constexpr int N = 16 / sizeof (short);
  using type = B<N>::type;
};
template <typename, int, int> using D = C::type;
template <class U> using E = typename U::T;
template <typename, int> struct F {
  unsigned short r[16 / sizeof(short)];
};
template <class U>
F<E<U>, U::a> baz (U);
template <class U>
using G = decltype(baz (U ()));
template <class U, typename T>
G<U> qux (U, T);
template <typename T, int N>
F<T, N>
boo (F<T, N> b)
{
  F<T, N> a;
  for (int i = 0; i < N; ++i)
    a.r[i] = a.r[i] + b.r[i] < T(-1) ? a.r[i] + b.r[i] : T(-1);
  return a;
}
void corge (G<A<8>>);
template <class U, class V>
void
freddy (U, V v, V)
{
  corge (v);
}
template <int U, class V, int W>
struct H {
  static void garply (int, int) { D<int, U, W> d; V() (int(), d); }
};
struct I;
template <int W = 1>
struct J {
  template <typename T>
  void operator () (T) { int m = 1; H<int{}, I, W>::garply (1, m); }
};
struct K {
  void operator () (short t) { J() (t); }
};
struct I {
  template <class T, class U>
  void
  operator () (T, U d)
  {
    auto a = qux (d, T()), b = qux (d, 0);
    F c = boo (b);
    freddy (d, c, a);
  }
};
void
test ()
{
  K k;
  k (short ());
}

[Bug debug/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 54557
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54557&action=edit
gcc13-pr108967.patch

Full untested patch.

[Bug debug/108967] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

--- Comment #6 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:520403f324a6ed8b527f39239709dd0841b992e2

commit r13-6388-g520403f324a6ed8b527f39239709dd0841b992e2
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Mar 1 10:26:46 2023 +0100

    cfgexpand: Handle WIDEN_{PLUS,MINUS}_EXPR and
VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR in expand_debug_expr [PR108967]

    When these tree codes were introduced, expand_debug_expr hasn't been
    updated to handle them.  For the VEC_*_EXPR we currently mostly punt, the
    non-vector ones can be handled easily.  In release compilers this doesn't
    ICE, but with checking we ICE so that we make sure to handle all the needed
    tree codes there.

    2023-03-01  Jakub Jelinek  <jakub@redhat.com>

            PR debug/108967
            * cfgexpand.cc (expand_debug_expr): Handle WIDEN_{PLUS,MINUS}_EXPR
            and VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR.

            * g++.dg/debug/pr108967.C: New test.

[Bug debug/108967] [11/12 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|internal compiler error: in |[11/12 Regression] internal
                   |expand_debug_expr, at       |compiler error: in
                   |cfgexpand.cc:5450           |expand_debug_expr, at
                   |                            |cfgexpand.cc:5450

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for trunk so far.
Bisected with -O2 -g -ftree-vectorize to that r11-5160-g9fc9573f9a5e9432e53c
commit.

[Bug debug/108967] [11/12 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:7b9e837e769474fa5e91958dd2a479837b1176cb

commit r12-9279-g7b9e837e769474fa5e91958dd2a479837b1176cb
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Mar 1 10:26:46 2023 +0100

    cfgexpand: Handle WIDEN_{PLUS,MINUS}_EXPR and
VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR in expand_debug_expr [PR108967]

    When these tree codes were introduced, expand_debug_expr hasn't been
    updated to handle them.  For the VEC_*_EXPR we currently mostly punt, the
    non-vector ones can be handled easily.  In release compilers this doesn't
    ICE, but with checking we ICE so that we make sure to handle all the needed
    tree codes there.

    2023-03-01  Jakub Jelinek  <jakub@redhat.com>

            PR debug/108967
            * cfgexpand.cc (expand_debug_expr): Handle WIDEN_{PLUS,MINUS}_EXPR
            and VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR.

            * g++.dg/debug/pr108967.C: New test.

    (cherry picked from commit 520403f324a6ed8b527f39239709dd0841b992e2)

[Bug debug/108967] [11 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[11/12 Regression] internal |[11 Regression] internal
                   |compiler error: in          |compiler error: in
                   |expand_debug_expr, at       |expand_debug_expr, at
                   |cfgexpand.cc:5450           |cfgexpand.cc:5450

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 12.3 too.

[Bug debug/108967] [11 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.0                        |13.2

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 13.1 is being released, retargeting bugs to GCC 13.2.

[Bug debug/108967] [11 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

--- Comment #11 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:54b240621d372e63d8fcacae591e39b36634ac18

commit r11-10717-g54b240621d372e63d8fcacae591e39b36634ac18
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Mar 1 10:26:46 2023 +0100

    cfgexpand: Handle WIDEN_{PLUS,MINUS}_EXPR and
VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR in expand_debug_expr [PR108967]

    When these tree codes were introduced, expand_debug_expr hasn't been
    updated to handle them.  For the VEC_*_EXPR we currently mostly punt, the
    non-vector ones can be handled easily.  In release compilers this doesn't
    ICE, but with checking we ICE so that we make sure to handle all the needed
    tree codes there.

    2023-03-01  Jakub Jelinek  <jakub@redhat.com>

            PR debug/108967
            * cfgexpand.c (expand_debug_expr): Handle WIDEN_{PLUS,MINUS}_EXPR
            and VEC_WIDEN_{PLUS,MINUS}_{HI,LO}_EXPR.

            * g++.dg/debug/pr108967.C: New test.

    (cherry picked from commit 520403f324a6ed8b527f39239709dd0841b992e2)

[Bug debug/108967] [11 Regression] internal compiler error: in expand_debug_expr, at cfgexpand.cc:5450

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108967

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

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

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 11.4 as well.