Hello, Richard. Sorry for bothering you. One more question.
> On 07/19/2018 05:54 AM, Stefan Markovic wrote:
> > + case NM_ADDIUGP_B:
> > + gen_arith_imm(ctx, OPC_ADDIU, rt, 28, u);
> > + break;
>
> Use gen_op_addr_add, since behaves_like('DADDIU[GP.B]').
Did you perhaps mean an implementation similar to this would be appropriate:
case NM_ADDIUGP_B:
if (rt != 0) {
uint32_t offset = extract32(ctx->opcode, 0, 18);
if (offset == 0) {
gen_load_gpr(cpu_gpr[rt], 28);
} else {
TCGv t0;
t0 = tcg_temp_new();
tcg_gen_movi_tl(t0, offset);
gen_op_addr_add(ctx, cpu_gpr[rt], cpu_gpr[28], t0);
tcg_temp_free(t0);
}
}
break;
(this is like NM_ADDIUGP_W implementation)
Aleksandar M.
