On Wed, Dec 10, 2014 at 9:00 AM, Segher Boessenkool <seg...@kernel.crashing.org> wrote: > On Mon, Dec 08, 2014 at 10:53:21AM -0500, David Edelsohn wrote: >> As we both noticed, there are a few problems with this patch, so I'll >> wait for a revised version. > > Here it is. It took a bit longer because of a latent problem in combine > (ugh!) that caused mysterious failures in guality (double ugh). > > For the carry_in patterns I now have an expander that expands to the > _0 and _m1 cases directly. > > Okay for mainline? > > Cheers, > > > Segher > > > > 2014-12-10 Segher Boessenkool <seg...@kernel.crashing.org> > > gcc/ > PR target/64180 > * config/rs6000/predicates.md (adde_operand): New. > * config/rs6000/rs6000.md (add<mode>3_carry): New. > (*add<mode>3_imm_carry_pos): New. > (*add<mode>3_imm_carry_0): New. > (*add<mode>3_imm_carry_m1): New. > (*add<mode>3_imm_carry_neg): New. > (add<mode>3_carry_in): New. > (*add<mode>3_carry_in): New.
Please name this *add<mode>3_carry_in_internal > (add<mode>3_carry_in_0): New. > (add<mode>3_carry_in_m1): New. > (subf<mode>3_carry): New. > (*subf<mode>3_imm_carry_0): New. > (*subf<mode>3_imm_carry_m1): New. > (subf<mode>3_carry_in): New. > (*subf<mode>3_carry_in): New. Please name this *subf<mode>3_carry_in_internal > (subf<mode>3_carry_in_0): New. > (subf<mode>3_carry_in_m1): New. > (subf<mode>3_carry_in_xx): New. Okay with those changes. Thanks for all of the great improvements! - David
