| Message ID | 87mv61cg03.fsf@linaro.org |
|---|---|
| State | Accepted |
| Commit | c1b52c0ef6eeb91b5326eee18bb2652e3f732162 |
| Headers | show |
| Series | Make more use of in_hard_reg_set_p | expand |
On 09/11/2017 11:16 AM, Richard Sandiford wrote: > An upcoming patch will convert hard_regno_nregs into an inline > function, which in turn allows hard_regno_nregs to be used as the > name of a targetm field. This patch rewrites a use that can use > in_hard_reg_set_p instead. > > Tested on aarch64-linux-gnu, x86_64-linux-gnu and powerpc64le-linux-gnu. > Also tested by comparing the testsuite assembly output on at least one > target per CPU directory. OK to install? > > Richard > > > 2017-09-11 Richard Sandiford <richard.sandiford@linaro.org> > > gcc/ > * ira-costs.c (record_operand_costs): Use in_hard_reg_set_p > instead of hard_regno_nregs. OK. jeff
Index: gcc/ira-costs.c =================================================================== --- gcc/ira-costs.c 2017-09-04 11:48:57.531552460 +0100 +++ gcc/ira-costs.c 2017-09-11 17:21:26.382315018 +0100 @@ -1386,7 +1386,7 @@ record_operand_costs (rtx_insn *insn, en cost_classes_t cost_classes_ptr = regno_cost_classes[regno]; enum reg_class *cost_classes = cost_classes_ptr->classes; reg_class_t rclass; - int k, nr; + int k; i = regno == (int) REGNO (src) ? 1 : 0; for (k = cost_classes_ptr->num - 1; k >= 0; k--) @@ -1398,18 +1398,9 @@ record_operand_costs (rtx_insn *insn, en { if (reg_class_size[rclass] == 1) op_costs[i]->cost[k] = -frequency; - else - { - for (nr = 0; - nr < hard_regno_nregs[other_regno][mode]; - nr++) - if (! TEST_HARD_REG_BIT (reg_class_contents[rclass], - other_regno + nr)) - break; - - if (nr == hard_regno_nregs[other_regno][mode]) - op_costs[i]->cost[k] = -frequency; - } + else if (in_hard_reg_set_p (reg_class_contents[rclass], + mode, other_regno)) + op_costs[i]->cost[k] = -frequency; } } }