Also committed to gcc-6 branch.
PR target/72802
* config/rs6000/rs6000.md (mov_hardfloat): Sort
alternatives. Put loads first, then stores, and reg/reg moves
within same class later. Delete attr length.
testsuite/
* gcc.c-torture/compile/pr72802.c: New.
d
On Mon, Aug 08, 2016 at 04:27:22PM +0930, Alan Modra wrote:
> PR target/72802
> * config/rs6000/rs6000.md (mov_hardflow): Sort
"hardfloat".
> alternatives. Put loads first, then stores, and reg/reg moves
> within same class later. Delete attr length.
> testsuite/
>
On Fri, Aug 05, 2016 at 08:31:45PM -0500, Segher Boessenkool wrote:
> On Sat, Aug 06, 2016 at 10:53:33AM +0930, Alan Modra wrote:
> > On Fri, Aug 05, 2016 at 06:01:47PM -0500, Segher Boessenkool wrote:
> > > I agree reg-reg moves should come after reg-mem moves, but is it such
> > > a good idea to
On Sat, Aug 06, 2016 at 10:53:33AM +0930, Alan Modra wrote:
> On Fri, Aug 05, 2016 at 06:01:47PM -0500, Segher Boessenkool wrote:
> > On Fri, Aug 05, 2016 at 02:20:40PM +0930, Alan Modra wrote:
> > > Here are the reload costs for the various alternatives of
> > > movsf_hardfloat:
> > > "=!r, !r, m
On Fri, Aug 05, 2016 at 06:01:47PM -0500, Segher Boessenkool wrote:
> On Fri, Aug 05, 2016 at 02:20:40PM +0930, Alan Modra wrote:
> > Here are the reload costs for the various alternatives of
> > movsf_hardfloat:
> > "=!r, !r, m, f, ww, ww, !r, f, wb, m, wY, wu, Z,?wn, ?r,*c*l, !r, *h"
> > "
On Fri, Aug 05, 2016 at 02:20:40PM +0930, Alan Modra wrote:
> Here are the reload costs for the various alternatives of
> movsf_hardfloat:
> "=!r, !r, m, f, ww, ww, !r, f, wb, m, wY, wu, Z,?wn, ?r,*c*l, !r, *h"
> "r, m, r, f, ww, j, j, m, wY, f, wb, Z, wu, r, wn, r, *h, 0"
> 617
After fixing the constraint problem, we hit an "insn does not satisfy
its constraints" with -mno-lra on the following insn, a vector load
from mem which has an invalid offset:
(insn 631 630 1122 12 (set (reg:SF 108 31 [orig:260 pretmp_44 ] [260])
(mem:SF (plus:DI (reg:DI 30 30 [orig:338 ivt
