[Bug c++/110303] New: In debug mode, gcc does not force compile time evaluation for immediate arguments

[Bug c++/110303] New: In debug mode, gcc does not force compile time evaluation for immediate arguments

[denis.yaroshevskij at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

            Bug ID: 110303
           Summary: In debug mode, gcc does not force compile time
                    evaluation for immediate arguments
           Product: gcc
           Version: 13.2.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: denis.yaroshevskij at gmail dot com
  Target Milestone: ---

Minimum example:

The following code

```
#include "immintrin.h"

constexpr int f() { return 0; }
__m128i shuffle(__m128i x) { return _mm_shuffle_epi32(x, f()); }
```

Does not compile in debug

Godbolt: https://godbolt.org/z/x7bfhr4vY

[Bug target/110303] Without -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|In debug mode, gcc does not |Without -O0,
                   |force compile time          |_mm_shuffle_epi32 with a
                   |evaluation for immediate    |constexpr function argument
                   |arguments                   |does not compile
          Component|c++                         |target

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I think this is correct behavior really.

[Bug target/110303] With -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
so _mm_shuffle_epi32 requires a constant but since it is an argument, the
argument, it is not a constant expression requirement.

[Bug target/110303] With -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |DUPLICATE
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Dup of bug 80353. constexpr function does not have to be a full constexpression
in many cases.

*** This bug has been marked as a duplicate of bug 80353 ***

[Bug target/110303] With -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[denis.yaroshevskij at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

--- Comment #4 from Denis Yaroshevskiy <denis.yaroshevskij at gmail dot com> ---
> so _mm_shuffle_epi32 requires a constant but since it is an argument, the argument, it is not a constant expression requirement.

The function is marked constexpr. So it can be a constant if you ask it.

[Bug target/110303] With -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[xry111 at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

Xi Ruoyao <xry111 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |xry111 at gcc dot gnu.org

--- Comment #5 from Xi Ruoyao <xry111 at gcc dot gnu.org> ---
(In reply to Denis Yaroshevskiy from comment #4)
> > so _mm_shuffle_epi32 requires a constant but since it is an argument, the argument, it is not a constant expression requirement.
> 
> The function is marked constexpr. So it can be a constant if you ask it.

The problem is -O0 means "don't ask it".  Constant propagation is an
optimization, but -O0 means don't do optimization.

[Bug target/110303] With -O0, _mm_shuffle_epi32 with a constexpr function argument does not compile

[m.cencora at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110303

m.cencora at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |m.cencora at gmail dot com

--- Comment #6 from m.cencora at gmail dot com ---
(In reply to Denis Yaroshevskiy from comment #4)
> > so _mm_shuffle_epi32 requires a constant but since it is an argument, the argument, it is not a constant expression requirement.
> 
> The function is marked constexpr. So it can be a constant if you ask it.

constexpr at function declaration means that it COULD be evaluated in compile
time, but doesn't force it.

To force it either invoke it in a context that requires a compile-time
evaluation (e.g. static_assert, or initializer of constexpr variable), or mark
it as consteval instead.