[Bug tree-optimization/82732] malloc+zeroing other than memset not optimized to calloc, so asm output is malloc+memset

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82732

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2025-01-13

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So VRP can do convert it to:
  <bb 2> [local count: 118111600]:
  _1 = (long unsigned int) size_8(D);
  _2 = _1 * 4;
  p_11 = malloc (_2);
  if (size_8(D) != 0)
    goto <bb 3>; [89.00%]
  else
    goto <bb 4>; [11.00%]

  <bb 3> [local count: 105119324]:
  _19 = size_8(D) + 4294967295;
  _15 = (sizetype) _19;
  _14 = (sizetype) size_8(D);
  _7 = _14 * 4;
  __builtin_memset (p_11, 0, _7);


BUT VRP2 is after strlen1 which does the conversion of malloc/memset -> calloc.

I have not looked into why FRE5 could see the redundant _7 being the same as _2
though (note VRP2 also runs after FRE5).

VRP2 can do it because of
Applying pattern match.pd:5271, gimple-match-7.cc:6107
Applying pattern match.pd:237, gimple-match-6.cc:10371
Applying pattern match.pd:4000, gimple-match-8.cc:1587
Applying pattern match.pd:4098, gimple-match-6.cc:16572
gimple_simplified to _15 = (sizetype) size_8(D);
Folded into: _15 = (sizetype) size_8(D);