[Bug c++/71372] [6/7 Regression] GCC Optimization "tree-dse" does not respect access to volatile* upon explicit cast.

["rguenth at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71372

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
So pre gimplification we have

(gdb) p debug_generic_expr (0x7ffff69fcbe0)
*FLASH
(gdb) p ((tree)0x7ffff69fcbe0)->base.volatile_flag 
$5 = 0

thus somehow the INDIRECT_REF misses the TREE_THIS_VOLATILE flag.

So it is cp_fold folding

 <indirect_ref 0x7ffff69fc960
    type <integer_type 0x7ffff69ee7e0 short unsigned int volatile unsigned HI
        size <integer_cst 0x7ffff68a8018 constant 16>
        unit size <integer_cst 0x7ffff68a8030 constant 2>
        align 16 symtab 0 alias set -1 canonical type 0x7ffff69ee7e0 precision
16 min <integer_cst 0x7ffff68a8048 0> max <integer_cst 0x7ffff68a8000 65535>
        pointer_to_this <pointer_type 0x7ffff69ee888>>
    side-effects volatile
 ^^^^^^^^^^^^^^^^^^^^^^

via

    unary:

      loc = EXPR_LOCATION (x);
      op0 = cp_fold_maybe_rvalue (TREE_OPERAND (x, 0), rval_ops);

      if (op0 != TREE_OPERAND (x, 0))
        {
          if (op0 == error_mark_node)
            x = error_mark_node;
          else
            x = fold_build1_loc (loc, code, TREE_TYPE (x), op0);
^^^^^^^^^^^

but forgetting to re-apply TREE_THIS_VOLATILE.  This probably seriously breaks
volatile support.