[Bug c++/99580] New: False positive -Warray-bounds with -O2

[Bug c++/99580] New: False positive -Warray-bounds with -O2

[sventeam at yandex dot ru]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99580

            Bug ID: 99580
           Summary: False positive -Warray-bounds with -O2
           Product: gcc
           Version: 10.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sventeam at yandex dot ru
  Target Milestone: ---

Created attachment 50383
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50383&action=edit
Source to reproduce the issue

Using default GCC from Ubuntu 20.10.

$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/10/lto-wrapper
OFFLOAD_TARGET_NAMES=nvptx-none:amdgcn-amdhsa:hsa
OFFLOAD_TARGET_DEFAULT=1
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu
10.2.0-13ubuntu1' --with-bugurl=file:///usr/share/doc/gcc-10/README.Bugs
--enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,m2 --prefix=/usr
--with-gcc-major-version-only --program-suffix=-10
--program-prefix=x86_64-linux-gnu- --enable-shared --enable-linker-build-id
--libexecdir=/usr/lib --without-included-gettext --enable-threads=posix
--libdir=/usr/lib --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug
--enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new
--enable-gnu-unique-object --disable-vtable-verify --enable-plugin
--enable-default-pie --with-system-zlib --enable-libphobos-checking=release
--with-target-system-zlib=auto --enable-objc-gc=auto --enable-multiarch
--disable-werror --with-arch-32=i686 --with-abi=m64
--with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic
--enable-offload-targets=nvptx-none=/build/gcc-10-JvwpWM/gcc-10-10.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-10-JvwpWM/gcc-10-10.2.0/debian/tmp-gcn/usr,hsa
--without-cuda-driver --enable-checking=release --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 10.2.0 (Ubuntu 10.2.0-13ubuntu1)

$ g++ -Warray-bounds -Werror -O2 qef_solver.cpp
qef_solver.cpp: In function ‘glm::mat3 jacobi(glm::mat3)’:
qef_solver.cpp:43:19: error: array subscript 3 is above array bounds of
‘glm::mat<3, 3, float>::col_type [3]’ {aka ‘glm::vec<3, float> [3]’}
[-Werror=array-bounds]
   43 |         return col[i];
      |                ~~~^
qef_solver.cpp:52:14: note: while referencing ‘glm::mat<3, 3, float>::col’
   52 |     col_type col[3];
      |              ^~~
qef_solver.cpp:43:19: error: array subscript 3 is above array bounds of
‘glm::mat<3, 3, float>::col_type [3]’ {aka ‘glm::vec<3, float> [3]’}
[-Werror=array-bounds]
   43 |         return col[i];
      |                ~~~^
qef_solver.cpp:52:14: note: while referencing ‘glm::mat<3, 3, float>::col’
   52 |     col_type col[3];
      |              ^~~
cc1plus: all warnings being treated as errors

This warning disappears if we change -O2 to -O1, or change `int j = i + 1` in
line 77 to `int j = i`, or comment out line 82 `max_j = j`.

[Bug tree-optimization/99580] False positive -Warray-bounds with -O2

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99580

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic
     Ever confirmed|0                           |1
          Component|c++                         |tree-optimization
      Known to fail|                            |10.2.0, 11.0, 7.3.0, 8.3.0,
                   |                            |9.2.0
                 CC|                            |msebor at gcc dot gnu.org
   Last reconfirmed|                            |2021-03-14
             Blocks|                            |56456
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Martin Sebor <msebor at gcc dot gnu.org> ---
Confirmed.  The warning first appeared with r238761 AKA "Teach VRP to register
assertions along default switch labels (PR18046)."

A simplified test case is below:

$ cat pr99580.C && -O2 -S -Wall -fdump-tree-vrp1-details pr99580.C
struct vec
{
  int operator[] (int i)
  {
    switch (i) {
    default:
    case 0: return x;
    case 1: return y;
    case 2: return z;
    }
  }

  int x, y, z;
};

vec v[3];

int f ()
{
  int max_el = v[0][1];
  int max_j = 1;

  for (int i = 0; i < 3; i++)
    for (int j = i + 1; j < 3; j++)
      if (v[i][j] > max_el)
        max_j = j;

  return v[max_j][max_j];
}
pr99580.C: In function ‘int f()’:
pr99580.C:28:17: warning: array subscript 3 is above array bounds of ‘vec [3]’
[-Warray-bounds]
   28 |   return v[max_j][max_j];
      |          ~~~~~~~^
pr99580.C:16:5: note: while referencing ‘v’
   16 | vec v[3];
      |     ^

$ grep warning pr99580.C.110t.vrp1 
Array bound warning for MEM <struct vec[3]> [(struct vec *)&v][max_j_27]

And:

Simulating block 14

extract_range_from_stmt visiting:
max_j_27 = ASSERT_EXPR <max_j_21, (unsigned int) max_j_21 + 4294967295 > 1>;
Intersecting
  int ~[1, 2]  EQUIVALENCES: { max_j_21 } (1 elements)
and
  int [1, 2147483646]
to
  int [3, 2147483646]  EQUIVALENCES: { max_j_21 } (1 elements)
Intersecting
  int [3, 2147483646]  EQUIVALENCES: { max_j_21 } (1 elements)
and
  int VARYING
to
  int [3, 2147483646]  EQUIVALENCES: { max_j_21 } (1 elements)
Found new range for max_j_27: int [3, 2147483646]  EQUIVALENCES: { } (0
elements)
Adding destination of edge (14 -> 17) to worklist


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56456
[Bug 56456] [meta-bug] bogus/missing -Warray-bounds

[Bug tree-optimization/99580] False positive -Warray-bounds with nested loops

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99580

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2021-03-14 00:00:00         |2021-10-13
            Summary|False positive              |False positive
                   |-Warray-bounds with -O2     |-Warray-bounds with nested
                   |                            |loops
      Known to fail|                            |11.2.0, 12.0

--- Comment #2 from Martin Sebor <msebor at gcc dot gnu.org> ---
Still present on trunk.