target/i386: remove CC_OP_CLR

Just use CC_OP_EFLAGS; it is not that likely that the flags computed by
CC_OP_CLR survive the end of the basic block, in which case there is no
need to spill cc_op_src.

cc_op_src now does need spilling if the XOR is followed by a memory
operation, but this only costs 0.2% extra TCG ops.  They will be recouped
by simplifications in how QEMU evaluates ZF at runtime, which are even
greater with this change.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

diff --git a/target/i386/cpu-dump.c b/target/i386/cpu-dump.c
index dc6723aedeebf4985238622ceb67fd43fd59a7db..a72ed93bd2f34c5a9b0bc9c58974bbd3ac3314bb 100644 (file)
--- a/target/i386/cpu-dump.c
+++ b/target/i386/cpu-dump.c
@@ -91,7 +91,6 @@ static const char * const cc_op_str[] = {
     [CC_OP_BMILGQ] = "BMILGQ",
 
     [CC_OP_POPCNT] = "POPCNT",
-    [CC_OP_CLR] = "CLR",
 };
 
 static void

diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index e2e10f55b2ea49734de8728c04cac0caee1db9e5..50172d7a6f96f98eb4f8b3216fb35e8e96f8150e 100644 (file)
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -1304,7 +1304,6 @@ typedef enum {
     CC_OP_ADCX, /* CC_DST = C, CC_SRC = rest.  */
     CC_OP_ADOX, /* CC_SRC2 = O, CC_SRC = rest.  */
     CC_OP_ADCOX, /* CC_DST = C, CC_SRC2 = O, CC_SRC = rest.  */
-    CC_OP_CLR, /* Z and P set, all other flags clear.  */
 
     CC_OP_MULB, /* modify all flags, C, O = (CC_SRC != 0) */
     CC_OP_MULW,

diff --git a/target/i386/tcg/cc_helper.c b/target/i386/tcg/cc_helper.c
index dbddaa2fcb3d525b36e582678aa37843a8589170..40583c04cf9344934567a2d771f486f58323cf71 100644 (file)
--- a/target/i386/tcg/cc_helper.c
+++ b/target/i386/tcg/cc_helper.c
@@ -104,8 +104,6 @@ target_ulong helper_cc_compute_all(target_ulong dst, target_ulong src1,
 
     case CC_OP_EFLAGS:
         return src1;
-    case CC_OP_CLR:
-        return CC_Z | CC_P;
     case CC_OP_POPCNT:
         return dst ? 0 : CC_Z;
 
@@ -243,7 +241,6 @@ target_ulong helper_cc_compute_c(target_ulong dst, target_ulong src1,
     case CC_OP_LOGICW:
     case CC_OP_LOGICL:
     case CC_OP_LOGICQ:
-    case CC_OP_CLR:
     case CC_OP_POPCNT:
         return 0;
 

diff --git a/target/i386/tcg/emit.c.inc b/target/i386/tcg/emit.c.inc
index fd17a9b1eca35fde29038174ee3169136471ec09..790307dbba8b6ae2159bf3ca9cb3f24af96363b6 100644 (file)
--- a/target/i386/tcg/emit.c.inc
+++ b/target/i386/tcg/emit.c.inc
@@ -1452,19 +1452,12 @@ static void gen_bt_flags(DisasContext *s, X86DecodedInsn *decode, TCGv src, TCGv
      * C is the result of the test, Z is unchanged, and the others
      * are all undefined.
      */
-    switch (s->cc_op) {
-    case CC_OP_DYNAMIC:
-    case CC_OP_CLR:
-    case CC_OP_EFLAGS:
-    case CC_OP_ADCX:
-    case CC_OP_ADOX:
-    case CC_OP_ADCOX:
+    if (s->cc_op == CC_OP_DYNAMIC || CC_OP_HAS_EFLAGS(s->cc_op)) {
         /* Generate EFLAGS and replace the C bit.  */
         cf = tcg_temp_new();
         tcg_gen_setcond_tl(TCG_COND_TSTNE, cf, src, mask);
         prepare_update_cf(decode, s, cf);
-        break;
-    default:
+    } else {
         /*
          * Z was going to be computed from the non-zero status of CC_DST.
          * We can get that same Z value (and the new C value) by leaving
@@ -1475,7 +1468,6 @@ static void gen_bt_flags(DisasContext *s, X86DecodedInsn *decode, TCGv src, TCGv
         decode->cc_dst = cpu_cc_dst;
         decode->cc_op = ((s->cc_op - CC_OP_MULB) & 3) + CC_OP_SARB;
         tcg_gen_shr_tl(decode->cc_src, src, s->T1);
-        break;
     }
 }
 
@@ -4724,7 +4716,8 @@ static void gen_XOR(DisasContext *s, X86DecodedInsn *decode)
         decode->op[2].unit == X86_OP_INT &&
         decode->op[1].n == decode->op[2].n) {
         tcg_gen_movi_tl(s->T0, 0);
-        decode->cc_op = CC_OP_CLR;
+        decode->cc_op = CC_OP_EFLAGS;
+        decode->cc_src = tcg_constant_tl(CC_Z | CC_P);
     } else {
         MemOp ot = decode->op[1].ot;
 

diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c
index dc308f31041ed6fa6e10d75032a2d62f777a45f0..a20fbb019c89e35c3f67828713972f5605c84b2d 100644 (file)
--- a/target/i386/tcg/translate.c
+++ b/target/i386/tcg/translate.c
@@ -309,7 +309,6 @@ static const uint8_t cc_op_live[CC_OP_NB] = {
     [CC_OP_ADCX] = USES_CC_DST | USES_CC_SRC,
     [CC_OP_ADOX] = USES_CC_SRC | USES_CC_SRC2,
     [CC_OP_ADCOX] = USES_CC_DST | USES_CC_SRC | USES_CC_SRC2,
-    [CC_OP_CLR] = 0,
     [CC_OP_POPCNT] = USES_CC_DST,
 };
 
@@ -803,10 +802,6 @@ static void gen_mov_eflags(DisasContext *s, TCGv reg)
         tcg_gen_mov_tl(reg, cpu_cc_src);
         return;
     }
-    if (s->cc_op == CC_OP_CLR) {
-        tcg_gen_movi_tl(reg, CC_Z | CC_P);
-        return;
-    }
 
     dst = cpu_cc_dst;
     src1 = cpu_cc_src;
@@ -897,7 +892,6 @@ static CCPrepare gen_prepare_eflags_c(DisasContext *s, TCGv reg)
                              .reg2 = cpu_cc_src, .use_reg2 = true };
 
     case CC_OP_LOGICB ... CC_OP_LOGICQ:
-    case CC_OP_CLR:
     case CC_OP_POPCNT:
         return (CCPrepare) { .cond = TCG_COND_NEVER };
 
@@ -969,7 +963,6 @@ static CCPrepare gen_prepare_eflags_s(DisasContext *s, TCGv reg)
     case CC_OP_ADCOX:
         return (CCPrepare) { .cond = TCG_COND_TSTNE, .reg = cpu_cc_src,
                              .imm = CC_S };
-    case CC_OP_CLR:
     case CC_OP_POPCNT:
         return (CCPrepare) { .cond = TCG_COND_NEVER };
     default:
@@ -988,7 +981,6 @@ static CCPrepare gen_prepare_eflags_o(DisasContext *s, TCGv reg)
     case CC_OP_ADCOX:
         return (CCPrepare) { .cond = TCG_COND_NE, .reg = cpu_cc_src2,
                              .no_setcond = true };
-    case CC_OP_CLR:
     case CC_OP_POPCNT:
         return (CCPrepare) { .cond = TCG_COND_NEVER };
     case CC_OP_MULB ... CC_OP_MULQ:
@@ -1013,8 +1005,6 @@ static CCPrepare gen_prepare_eflags_z(DisasContext *s, TCGv reg)
     case CC_OP_ADCOX:
         return (CCPrepare) { .cond = TCG_COND_TSTNE, .reg = cpu_cc_src,
                              .imm = CC_Z };
-    case CC_OP_CLR:
-        return (CCPrepare) { .cond = TCG_COND_ALWAYS };
     default:
         {
             MemOp size = (s->cc_op - CC_OP_ADDB) & 3;