[Bug tree-optimization/60895] [4.8/4.9/4.10 Regression] error: address taken, but ADDRESSABLE bit not set

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60895

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-04-22
                 CC|                            |jakub at gcc dot gnu.org
      Known to work|4.8.2                       |
   Target Milestone|---                         |4.8.3
            Summary|error: address taken, but   |[4.8/4.9/4.10 Regression]
                   |ADDRESSABLE bit not set     |error: address taken, but
                   |                            |ADDRESSABLE bit not set
     Ever confirmed|0                           |1

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r193157, but when the testcase is changed into:
struct C
{
  double elems[3];
};

C
foo ()
{
  C a;
  double *f = a.elems;
  int b;
  for (; b;)
    {
      *f = 0;
      ++f;
    }
  return a;
}

struct J
{
  C c;
  __attribute__((always_inline)) J () : c (foo ()) {}
};

void
bar ()
{
  J ();
}

it starts with -O at r192946.  Both of these changes just change inliner's
decision, so I guess the bug has been latent before that.