Il gio 19 ott 2023, 19:44 Richard Henderson <richard.hender...@linaro.org>
ha scritto:
> On 10/19/23 03:46, Paolo Bonzini wrote:
> > + /*
> > + * Write back flags after last memory access. Some newer ALU
> instructions, as
> > + * well as SSE instructions, write flags in the gen_* function, but
> that can
> > + * cause incorrect tracking of CC_OP for instructions that write to
> both memory
> > + * and flags.
> > + */
> > + if (decode.cc_op != -1) {
> > + if (decode.cc_dst) {
> > + tcg_gen_mov_tl(cpu_cc_dst, decode.cc_dst);
> > + }
> > + if (decode.cc_src) {
> > + tcg_gen_mov_tl(cpu_cc_src, decode.cc_src);
> > + }
> > + if (decode.cc_src2) {
> > + tcg_gen_mov_tl(cpu_cc_src2, decode.cc_src2);
> > + }
> > + if (decode.cc_op == CC_OP_DYNAMIC) {
> > + tcg_gen_mov_i32(cpu_cc_op, decode.cc_op_dynamic);
> > + } else {
> > + assert(!decode.cc_op_dynamic);
> > + }
> > + set_cc_op(s, decode.cc_op);
> > + } else {
> > + assert(!decode.cc_dst);
> > + assert(!decode.cc_src);
> > + assert(!decode.cc_src2);
> > + assert(!decode.cc_op_dynamic);
> > + }
>
> I suggest you use cc_op_live[] to ensure that each output is present if
> USES_CC_* is set,
> and absent otherwise. Obviously that's not possible for CC_OP_DYNAMIC, but
> for everything else...
>
I tried but it didn't work very well. I have shuffled things a bit and I
don't remember why :) so I will give it another go.
Paolo
>
> r~
>
>
