[Bug tree-optimization/107837] New: Missed optimization: Using memcpy to load a struct unnecessary uses stack space

[Bug tree-optimization/107837] New: Missed optimization: Using memcpy to load a struct unnecessary uses stack space

[chfast at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107837

            Bug ID: 107837
           Summary: Missed optimization: Using memcpy to load a struct
                    unnecessary uses stack space
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: chfast at gmail dot com
  Target Milestone: ---

I have a simple struct with array uint64_t[4]. When using memcpy() load it from
a storage of bytes and then performing some additional operations, a temporary
object on the stack is created.


struct uint256
{
    unsigned long v[4];
};

void load_bad(uint256* o, const char* src) noexcept
{
    uint256 x;
    __builtin_memcpy(&x, src, sizeof(x));
    uint256 y;
    y.v[0] = __builtin_bswap64(x.v[3]);
    y.v[1] = __builtin_bswap64(x.v[2]);
    y.v[2] = __builtin_bswap64(x.v[1]);
    y.v[3] = __builtin_bswap64(x.v[0]);
    *o = y;
}


load_bad(uint256*, char const*):
        movdqu  xmm0, XMMWORD PTR [rsi]
        movdqu  xmm1, XMMWORD PTR [rsi+16]
        movaps  XMMWORD PTR [rsp-40], xmm0
        mov     rdx, QWORD PTR [rsp-32]
        mov     rax, QWORD PTR [rsp-40]
        movaps  XMMWORD PTR [rsp-24], xmm1
        mov     rsi, QWORD PTR [rsp-16]
        mov     rcx, QWORD PTR [rsp-24]
        bswap   rdx
        bswap   rax
        mov     QWORD PTR [rdi+16], rdx
        bswap   rsi
        bswap   rcx
        mov     QWORD PTR [rdi], rsi
        mov     QWORD PTR [rdi+8], rcx
        mov     QWORD PTR [rdi+24], rax
        ret


The workaround is to use reinterpret_cast.

https://godbolt.org/z/WevYch8nv

[Bug tree-optimization/107837] [10/11/12/13 Regression] Missed optimization: Using memcpy to load a struct unnecessary uses stack space

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107837

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-11-23
            Summary|Missed optimization: Using  |[10/11/12/13 Regression]
                   |memcpy to load a struct     |Missed optimization: Using
                   |unnecessary uses stack      |memcpy to load a struct
                   |space                       |unnecessary uses stack
                   |                            |space
           Keywords|                            |needs-bisection
      Known to work|                            |4.5.3, 4.7.3, 5.1.0, 7.5.0
   Target Milestone|---                         |10.5
      Known to fail|                            |8.1.0
     Ever confirmed|0                           |1

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed, GCC 7 did the right thing ...
Even on the gimple level:

  x$v$0_8 = MEM[(char * {ref-all})src_6(D)];
  x$v$1_9 = MEM[(char * {ref-all})src_6(D) + 8B];
  x$v$2_10 = MEM[(char * {ref-all})src_6(D) + 16B];
  x$v$3_11 = MEM[(char * {ref-all})src_6(D) + 24B];
  _1 = __builtin_bswap64 (x$v$3_11);
  _2 = __builtin_bswap64 (x$v$2_10);
  _3 = __builtin_bswap64 (x$v$1_9);
  _4 = __builtin_bswap64 (x$v$0_8);
  MEM[(struct uint256 *)o_7(D)] = _1;
  MEM[(struct uint256 *)o_7(D) + 8B] = _2;
  MEM[(struct uint256 *)o_7(D) + 16B] = _3;
  MEM[(struct uint256 *)o_7(D) + 24B] = _4;

[Bug tree-optimization/107837] [10/11/12/13 Regression] Missed optimization: Using memcpy to load a struct unnecessary uses stack space since r8-5200

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107837

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |jamborm at gcc dot gnu.org
           Priority|P3                          |P2
           Keywords|needs-bisection             |
            Summary|[10/11/12/13 Regression]    |[10/11/12/13 Regression]
                   |Missed optimization: Using  |Missed optimization: Using
                   |memcpy to load a struct     |memcpy to load a struct
                   |unnecessary uses stack      |unnecessary uses stack
                   |space                       |space since r8-5200

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r8-5200-gd90ffcfb7d105d004cf04911a42935be03256b49

[Bug tree-optimization/107837] [11/12/13/14 Regression] Missed optimization: Using memcpy to load a struct unnecessary uses stack space since r8-5200

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107837

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

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

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10 branch is being closed.