This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: [4.3 Regression] __builtin_strcpy doesn't work with -fstack-protector
On Tue, Jan 20, 2009 at 4:48 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Tue, Jan 20, 2009 at 04:40:40PM -0800, H.J. Lu wrote:
>> I don't think my backport has any optimizations. It only sets
>> memory size correctly.
>
> It contains hunks that are only useful together with the dse.c
> changes. You only want from these the
> expand_movmem_via_rep_mov change, the rest is for the DSE use.
> If you really want, put in an unconditional:
> if (MEM_SIZE (destmem))
> set_mem_size (destmem, NULL_RTX);
> in expand_setmem_via_rep_stos, or just assert it is NULL.
>
One failure is wrong memory size when rep stos is used.
I think we need the change in expand_setmem_via_rep_stos
to set memory size correctly. It isn't just for DSE. It is also
used by scheduler.
--
H.J.