[Bug rtl-optimization/94806] New: Failure to optimize unary minus for 128-bit operand

[gabravier at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94806

            Bug ID: 94806
           Summary: Failure to optimize unary minus for 128-bit operand
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gabravier at gmail dot com
  Target Milestone: ---

__int128 f(__int128 x)
{
    return -x;
}

It would appear like unary minus is badly optimized by GCC.

This is what LLVM outputs for this : 

f(__int128): # @f(__int128)
  mov rax, rdi
  xor edx, edx
  neg rax
  sbb rdx, rsi
  ret

And this is what GCC outputs : 

f(__int128):
  mov rax, rdi
  mov rdx, rsi
  neg rax
  adc rdx, 0
  neg rdx
  ret

GCC's output is obviously worse (unless an `adc`+`neg` pair can somehow be
better than an `sbb`)

(See also https://godbolt.org/z/9AUd79)