make NaN behavior consistent with hardfloat

[riscv-isa-sim.git] / riscv / insns / fadd_d.h

require_fp;
softfloat_roundingMode = RM;
FRD = f64_mulAdd(FRS1, 0x3ff0000000000000ULL, FRS2);
set_fp_exceptions;