This looks like a nice cleanup. One comment:
On Wed, Apr 06, 2011 at 10:21:12PM +0400, Anatoly Sokolov wrote:
> @@ -4958,60 +4958,54 @@
> mark_reload_reg_in_use (unsigned int regno, int opnum, enum reload_type type,
> enum machine_mode mode)
> {
> - unsigned int nregs = hard_regno_nregs[regno][mode];
> - unsigned int i;
> -
> - for (i = regno; i < nregs + regno; i++)
> + switch (type)
> {
> - switch (type)
> - {
> - case RELOAD_OTHER:
> - SET_HARD_REG_BIT (reload_reg_used, i);
> - break;
> + case RELOAD_OTHER:
> + add_to_hard_reg_set (&reload_reg_used, mode, regno);
> + break;
I wonder if it'd be worth breaking out a reg_set_for_reload_type (int
opnum) function and using it here. I think it could be profitably used
in clear_reload_reg_in_use, either in this patch in in a followup.
-Nathan
