[PATCH 2/2] fix lock cmpxchg instruction

[Wei Li]

For lock cmpxchg, the situation is more complex. After the instruction
is completed by tcg_gen_atomic_cmpxchg_tl, it needs a branch to judge
if oldv == cmpv or not. The instruction only touches accumulator when
oldv != cmpv.

Signed-off-by: Wei Li <lw945lw945@yahoo.com>
---
 target/i386/tcg/translate.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c
index 05be8d08e6..4fd9c03cb7 100644
--- a/target/i386/tcg/translate.c
+++ b/target/i386/tcg/translate.c
@@ -5360,7 +5360,12 @@ static target_ulong disas_insn(DisasContext *s, CPUState 
*cpu)
                 gen_lea_modrm(env, s, modrm);
                 tcg_gen_atomic_cmpxchg_tl(oldv, s->A0, cmpv, newv,
                                           s->mem_index, ot | MO_LE);
+                label1 = gen_new_label();
+                gen_extu(ot, oldv);
+                gen_extu(ot, cmpv);
+                tcg_gen_brcond_tl(TCG_COND_EQ, oldv, cmpv, label1);
                 gen_op_mov_reg_v(s, ot, R_EAX, oldv);
+                gen_set_label(label1);
             } else {
                 if (mod == 3) {
                     rm = (modrm & 7) | REX_B(s);
-- 
2.30.2