This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: rtx_unchanging_p vs c++ vtable fields
- From: Richard Henderson <rth at redhat dot com>
- To: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 9 Dec 2001 10:57:30 -0800
- Subject: Re: rtx_unchanging_p vs c++ vtable fields
- References: <10112091143.AA12907@vlsi1.ultra.nyu.edu>
On Sun, Dec 09, 2001 at 06:43:13AM -0500, Richard Kenner wrote:
> But let's get back to the first point: why does the /u there causing
> problems: that should always be the safe direction?
Here is a reduced test case. This test doesn't actively fail like
the original because the scheduler doesn't have enough to work with.
But the fact of the aliasing problem can be seen in that we don't
eliminate the dead store:
stq $1,0($16)
stq $2,0($16)
r~
struct base
{
virtual ~base();
base() { }
};
struct derived : public base
{
virtual ~derived();
derived();
const int y;
};
derived::derived() : y(0) { }