This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc compile-time performance
From: Richard Henderson <rth@redhat.com>
Date: Sat, 18 May 2002 21:46:40 -0700
On Sat, May 18, 2002 at 09:20:38PM -0700, David S. Miller wrote:
> + tmpmode = reg_raw_mode[regno];
> + nxtmode = GET_MODE_WIDER_MODE (tmpmode);
> + while (HARD_REGNO_MODE_OK (regno, nxtmode))
> + {
> + tmpmode = nxtmode;
> + nxtmode = GET_MODE_WIDER_MODE (tmpmode);
> + }
You should also stop at nxtmode == VOIDmode.
My original cut had that, but I was sure HARD_REGNO_MODE_OK
would fail on VOIDmode. That is a bad assumption the more
I think about it, thus I'll add the check.