[Bug tree-optimization/106513] [10/11/12/13 Regression] bswap pass misses that >>56 for signed types can be replicate the sign bit

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106513

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
      Known to fail|                            |4.5.3, 4.6.4
            Summary|bswap is incorrectly        |[10/11/12/13 Regression]
                   |generated                   |bswap pass misses that >>56
                   |                            |for signed types can be
                   |                            |replicate the sign bit
      Known to work|                            |4.4.7
   Target Milestone|---                         |10.5
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-08-05

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

Better testcase (without the questionable undefined behavior):

typedef long long int int64_t;

__attribute__((noinline)) int64_t
swap64 (int64_t n)
{
  return (((n & (((int64_t) 0xff) )) << 56) |
          ((n & (((int64_t) 0xff) << 8)) << 40) |
          ((n & (((int64_t) 0xff) << 16)) << 24) |
          ((n & (((int64_t) 0xff) << 24)) << 8) |
          ((n & (((int64_t) 0xff) << 32)) >> 8) |
          ((n & (((int64_t) 0xff) << 40)) >> 24) |
          ((n & (((int64_t) 0xff) << 48)) >> 40) |
          ((n & ((int64_t)(0xffull << 56))) >> 56));
}

int main (void)
{
  volatile int64_t n = 0x8000000000000000l;

  if (swap64(n) != 0xffffffffffffff80l)
    __builtin_abort ();

  return 0;
}