[Bug tree-optimization/94911] New: Failure to optimize comparisons of VLA sizes

[Bug tree-optimization/94911] New: Failure to optimize comparisons of VLA sizes

[gabravier at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

            Bug ID: 94911
           Summary: Failure to optimize comparisons of VLA sizes
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gabravier at gmail dot com
  Target Milestone: ---

inline void assume(_Bool b)
{
    if (!b)
        __builtin_unreachable();
}

_Bool f(int n)
{
    assume(n >= 1);
    typedef int A[n];
    ++n;
    A a;
    int b[n];
    n -= 2;
    typedef int C[n];
    C c;

    return (sizeof(a) < sizeof(b)) && (sizeof(a) > sizeof(c));
}

This C code (this will have different results in C++ with GCC) should always
`return true`. LLVM makes this transformation, but GCC does not.

Also, extra question, is the fact that this always returns `false` when
compiled as C++ normal ? Clang has it return `true` if compiled as C++.

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

--- Comment #1 from Marc Glisse <glisse at gcc dot gnu.org> ---
gcc computes sizeof(a) as 4ul*(size_t)n, and unsigned types don't provide nice
overflow guarantees, so that complicates things.

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Marc Glisse from comment #1)
> gcc computes sizeof(a) as 4ul*(size_t)n, and unsigned types don't provide
> nice overflow guarantees, so that complicates things.

While the C++ front-end does:
(sizetype) ((ssizetype) n + -1) + 1) * 4)

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

--- Comment #3 from Marc Glisse <glisse at gcc dot gnu.org> ---
Since VLA is an extension for compatibility with C, it is strange that it
behaves differently (does one use the value of n at the time of the typedef and
the other at the time of the declaration?). This bug is about the optimization,
maybe a separate report about the C++ behavior would make sense.

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2023-05-13
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
  # RANGE [irange] int [1, +INF] NONZERO 0x7fffffff
  intD.7 n_11(D) = nD.4385;
  sizetype _2;
  _BoolD.4372 _16;

;;   basic block 2, loop depth 0, count 1073741824 (estimated locally), maybe
hot
;;    prev block 0, next block 1, flags: (NEW, REACHABLE, VISITED)
;;    pred:       ENTRY [always]  count:1073741824 (estimated locally)
(FALLTHRU,EXECUTABLE)
  # RANGE [irange] sizetype [1, 2147483647] NONZERO 0x7fffffff
  _1 = (sizetype) n_11(D);
  # RANGE [irange] sizetype [4, 8589934588] NONZERO 0x1fffffffc
  _2 = _1 * 4;
  _16 = _2 > 3;


We have the information to do the folding in optimize but not earlier ...

[Bug tree-optimization/94911] Failure to optimize comparisons of VLA sizes

[gabravier at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911

--- Comment #5 from Gabriel Ravier <gabravier at gmail dot com> ---
Also, as an extra note, w.r.t.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94911#c3, I've just noticed that I
had indeed made a separate bug report at https://gcc.gnu.org/PR94912 (which
ended up being closed as a duplicate of https://gcc.gnu.org/PR68531) - just
wanted to clarify that so nobody ends up filing more duplicates like I almost
just did