[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [4577] transformed TN into temporaries - add local temporar
|
From: |
Fabrice Bellard |
|
Subject: |
[Qemu-devel] [4577] transformed TN into temporaries - add local temporaries usage when needed - optimized fcmovX |
|
Date: |
Sun, 25 May 2008 17:26:41 +0000 |
Revision: 4577
http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=4577
Author: bellard
Date: 2008-05-25 17:26:41 +0000 (Sun, 25 May 2008)
Log Message:
-----------
transformed TN into temporaries - add local temporaries usage when needed -
optimized fcmovX
Modified Paths:
--------------
trunk/target-i386/cpu.h
trunk/target-i386/exec.h
trunk/target-i386/helper_template.h
trunk/target-i386/translate.c
Modified: trunk/target-i386/cpu.h
===================================================================
--- trunk/target-i386/cpu.h 2008-05-25 17:24:00 UTC (rev 4576)
+++ trunk/target-i386/cpu.h 2008-05-25 17:26:41 UTC (rev 4577)
@@ -472,12 +472,6 @@
#define NB_MMU_MODES 2
typedef struct CPUX86State {
-#if TARGET_LONG_BITS > HOST_LONG_BITS
- /* temporaries if we cannot store them in host registers */
- target_ulong t0, t1, t2;
-#endif
- target_ulong t3;
-
/* standard registers */
target_ulong regs[CPU_NB_REGS];
target_ulong eip;
@@ -526,6 +520,7 @@
XMMReg xmm_regs[CPU_NB_REGS];
XMMReg xmm_t0;
MMXReg mmx_t0;
+ target_ulong cc_tmp; /* temporary for rcr/rcl */
/* sysenter registers */
uint32_t sysenter_cs;
Modified: trunk/target-i386/exec.h
===================================================================
--- trunk/target-i386/exec.h 2008-05-25 17:24:00 UTC (rev 4576)
+++ trunk/target-i386/exec.h 2008-05-25 17:26:41 UTC (rev 4577)
@@ -29,60 +29,20 @@
#include "cpu-defs.h"
-/* at least 4 register variables are defined */
register struct CPUX86State *env asm(AREG0);
-#ifndef CPU_NO_GLOBAL_REGS
-
-#if TARGET_LONG_BITS > HOST_LONG_BITS
-
-/* no registers can be used */
-#define T0 (env->t0)
-#define T1 (env->t1)
-#define T2 (env->t2)
-
-#else
-
-/* XXX: use unsigned long instead of target_ulong - better code will
- be generated for 64 bit CPUs */
-register target_ulong T0 asm(AREG1);
-register target_ulong T1 asm(AREG2);
-register target_ulong T2 asm(AREG3);
-
-#endif /* ! (TARGET_LONG_BITS > HOST_LONG_BITS) */
-
-#endif /* ! CPU_NO_GLOBAL_REGS */
-
-#define A0 T2
-
extern FILE *logfile;
extern int loglevel;
-#ifndef reg_EAX
#define EAX (env->regs[R_EAX])
-#endif
-#ifndef reg_ECX
#define ECX (env->regs[R_ECX])
-#endif
-#ifndef reg_EDX
#define EDX (env->regs[R_EDX])
-#endif
-#ifndef reg_EBX
#define EBX (env->regs[R_EBX])
-#endif
-#ifndef reg_ESP
#define ESP (env->regs[R_ESP])
-#endif
-#ifndef reg_EBP
#define EBP (env->regs[R_EBP])
-#endif
-#ifndef reg_ESI
#define ESI (env->regs[R_ESI])
-#endif
-#ifndef reg_EDI
#define EDI (env->regs[R_EDI])
-#endif
-#define EIP (env->eip)
+#define EIP (env->eip)
#define DF (env->df)
#define CC_SRC (env->cc_src)
Modified: trunk/target-i386/helper_template.h
===================================================================
--- trunk/target-i386/helper_template.h 2008-05-25 17:24:00 UTC (rev 4576)
+++ trunk/target-i386/helper_template.h 2008-05-25 17:26:41 UTC (rev 4577)
@@ -287,11 +287,11 @@
if (count > 1)
res |= t0 >> (DATA_BITS + 1 - count);
t0 = res;
- env->t3 = (eflags & ~(CC_C | CC_O)) |
+ env->cc_tmp = (eflags & ~(CC_C | CC_O)) |
(lshift(src ^ t0, 11 - (DATA_BITS - 1)) & CC_O) |
((src >> (DATA_BITS - count)) & CC_C);
} else {
- env->t3 = -1;
+ env->cc_tmp = -1;
}
return t0;
}
@@ -316,11 +316,11 @@
if (count > 1)
res |= t0 << (DATA_BITS + 1 - count);
t0 = res;
- env->t3 = (eflags & ~(CC_C | CC_O)) |
+ env->cc_tmp = (eflags & ~(CC_C | CC_O)) |
(lshift(src ^ t0, 11 - (DATA_BITS - 1)) & CC_O) |
((src >> (count - 1)) & CC_C);
} else {
- env->t3 = -1;
+ env->cc_tmp = -1;
}
return t0;
}
Modified: trunk/target-i386/translate.c
===================================================================
--- trunk/target-i386/translate.c 2008-05-25 17:24:00 UTC (rev 4576)
+++ trunk/target-i386/translate.c 2008-05-25 17:26:41 UTC (rev 4577)
@@ -58,8 +58,9 @@
//#define MACRO_TEST 1
/* global register indexes */
-static TCGv cpu_env, cpu_T[2], cpu_A0, cpu_cc_op, cpu_cc_src, cpu_cc_dst;
-static TCGv cpu_T3;
+static TCGv cpu_env, cpu_A0, cpu_cc_op, cpu_cc_src, cpu_cc_dst, cpu_cc_tmp;
+/* local temps */
+static TCGv cpu_T[2], cpu_T3;
/* local register indexes (only used inside old micro ops) */
static TCGv cpu_tmp0, cpu_tmp1_i64, cpu_tmp2_i32, cpu_tmp3_i32, cpu_tmp4,
cpu_ptr0, cpu_ptr1;
static TCGv cpu_tmp5, cpu_tmp6;
@@ -260,34 +261,34 @@
#define REG_LH_OFFSET 4
#endif
-static inline void gen_op_mov_reg_TN(int ot, int t_index, int reg)
+static inline void gen_op_mov_reg_v(int ot, int reg, TCGv t0)
{
switch(ot) {
case OT_BYTE:
if (reg < 4 X86_64_DEF( || reg >= 8 || x86_64_hregs)) {
- tcg_gen_st8_tl(cpu_T[t_index], cpu_env, offsetof(CPUState,
regs[reg]) + REG_B_OFFSET);
+ tcg_gen_st8_tl(t0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_B_OFFSET);
} else {
- tcg_gen_st8_tl(cpu_T[t_index], cpu_env, offsetof(CPUState,
regs[reg - 4]) + REG_H_OFFSET);
+ tcg_gen_st8_tl(t0, cpu_env, offsetof(CPUState, regs[reg - 4]) +
REG_H_OFFSET);
}
break;
case OT_WORD:
- tcg_gen_st16_tl(cpu_T[t_index], cpu_env, offsetof(CPUState, regs[reg])
+ REG_W_OFFSET);
+ tcg_gen_st16_tl(t0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_W_OFFSET);
break;
#ifdef TARGET_X86_64
case OT_LONG:
- tcg_gen_st32_tl(cpu_T[t_index], cpu_env, offsetof(CPUState, regs[reg])
+ REG_L_OFFSET);
+ tcg_gen_st32_tl(t0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_L_OFFSET);
/* high part of register set to zero */
tcg_gen_movi_tl(cpu_tmp0, 0);
tcg_gen_st32_tl(cpu_tmp0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_LH_OFFSET);
break;
default:
case OT_QUAD:
- tcg_gen_st_tl(cpu_T[t_index], cpu_env, offsetof(CPUState, regs[reg]));
+ tcg_gen_st_tl(t0, cpu_env, offsetof(CPUState, regs[reg]));
break;
#else
default:
case OT_LONG:
- tcg_gen_st32_tl(cpu_T[t_index], cpu_env, offsetof(CPUState, regs[reg])
+ REG_L_OFFSET);
+ tcg_gen_st32_tl(t0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_L_OFFSET);
break;
#endif
}
@@ -295,12 +296,12 @@
static inline void gen_op_mov_reg_T0(int ot, int reg)
{
- gen_op_mov_reg_TN(ot, 0, reg);
+ gen_op_mov_reg_v(ot, reg, cpu_T[0]);
}
static inline void gen_op_mov_reg_T1(int ot, int reg)
{
- gen_op_mov_reg_TN(ot, 1, reg);
+ gen_op_mov_reg_v(ot, reg, cpu_T[1]);
}
static inline void gen_op_mov_reg_A0(int size, int reg)
@@ -329,23 +330,28 @@
}
}
-static inline void gen_op_mov_TN_reg(int ot, int t_index, int reg)
+static inline void gen_op_mov_v_reg(int ot, TCGv t0, int reg)
{
switch(ot) {
case OT_BYTE:
if (reg < 4 X86_64_DEF( || reg >= 8 || x86_64_hregs)) {
goto std_case;
} else {
- tcg_gen_ld8u_tl(cpu_T[t_index], cpu_env, offsetof(CPUState,
regs[reg - 4]) + REG_H_OFFSET);
+ tcg_gen_ld8u_tl(t0, cpu_env, offsetof(CPUState, regs[reg - 4]) +
REG_H_OFFSET);
}
break;
default:
std_case:
- tcg_gen_ld_tl(cpu_T[t_index], cpu_env, offsetof(CPUState, regs[reg]));
+ tcg_gen_ld_tl(t0, cpu_env, offsetof(CPUState, regs[reg]));
break;
}
}
+static inline void gen_op_mov_TN_reg(int ot, int t_index, int reg)
+{
+ gen_op_mov_v_reg(ot, cpu_T[t_index], reg);
+}
+
static inline void gen_op_movl_A0_reg(int reg)
{
tcg_gen_ld32u_tl(cpu_A0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_L_OFFSET);
@@ -511,90 +517,70 @@
}
}
-/* sign does not matter, except for lidt/lgdt call (TODO: fix it) */
-static inline void gen_op_ld_T0_A0(int idx)
+static inline void gen_op_ld_v(int idx, TCGv t0, TCGv a0)
{
int mem_index = (idx >> 2) - 1;
switch(idx & 3) {
case 0:
- tcg_gen_qemu_ld8u(cpu_T[0], cpu_A0, mem_index);
+ tcg_gen_qemu_ld8u(t0, a0, mem_index);
break;
case 1:
- tcg_gen_qemu_ld16u(cpu_T[0], cpu_A0, mem_index);
+ tcg_gen_qemu_ld16u(t0, a0, mem_index);
break;
case 2:
- tcg_gen_qemu_ld32u(cpu_T[0], cpu_A0, mem_index);
+ tcg_gen_qemu_ld32u(t0, a0, mem_index);
break;
default:
case 3:
- tcg_gen_qemu_ld64(cpu_T[0], cpu_A0, mem_index);
+ tcg_gen_qemu_ld64(t0, a0, mem_index);
break;
}
}
+/* XXX: always use ldu or lds */
+static inline void gen_op_ld_T0_A0(int idx)
+{
+ gen_op_ld_v(idx, cpu_T[0], cpu_A0);
+}
+
static inline void gen_op_ldu_T0_A0(int idx)
{
- gen_op_ld_T0_A0(idx);
+ gen_op_ld_v(idx, cpu_T[0], cpu_A0);
}
static inline void gen_op_ld_T1_A0(int idx)
{
+ gen_op_ld_v(idx, cpu_T[1], cpu_A0);
+}
+
+static inline void gen_op_st_v(int idx, TCGv t0, TCGv a0)
+{
int mem_index = (idx >> 2) - 1;
switch(idx & 3) {
case 0:
- tcg_gen_qemu_ld8u(cpu_T[1], cpu_A0, mem_index);
+ tcg_gen_qemu_st8(t0, a0, mem_index);
break;
case 1:
- tcg_gen_qemu_ld16u(cpu_T[1], cpu_A0, mem_index);
+ tcg_gen_qemu_st16(t0, a0, mem_index);
break;
case 2:
- tcg_gen_qemu_ld32u(cpu_T[1], cpu_A0, mem_index);
+ tcg_gen_qemu_st32(t0, a0, mem_index);
break;
default:
case 3:
- tcg_gen_qemu_ld64(cpu_T[1], cpu_A0, mem_index);
+ tcg_gen_qemu_st64(t0, a0, mem_index);
break;
}
}
static inline void gen_op_st_T0_A0(int idx)
{
- int mem_index = (idx >> 2) - 1;
- switch(idx & 3) {
- case 0:
- tcg_gen_qemu_st8(cpu_T[0], cpu_A0, mem_index);
- break;
- case 1:
- tcg_gen_qemu_st16(cpu_T[0], cpu_A0, mem_index);
- break;
- case 2:
- tcg_gen_qemu_st32(cpu_T[0], cpu_A0, mem_index);
- break;
- default:
- case 3:
- tcg_gen_qemu_st64(cpu_T[0], cpu_A0, mem_index);
- break;
- }
+ gen_op_st_v(idx, cpu_T[0], cpu_A0);
}
static inline void gen_op_st_T1_A0(int idx)
{
- int mem_index = (idx >> 2) - 1;
- switch(idx & 3) {
- case 0:
- tcg_gen_qemu_st8(cpu_T[1], cpu_A0, mem_index);
- break;
- case 1:
- tcg_gen_qemu_st16(cpu_T[1], cpu_A0, mem_index);
- break;
- case 2:
- tcg_gen_qemu_st32(cpu_T[1], cpu_A0, mem_index);
- break;
- default:
- case 3:
- tcg_gen_qemu_st64(cpu_T[1], cpu_A0, mem_index);
- break;
- }
+ gen_op_st_v(idx, cpu_T[1], cpu_A0);
}
static inline void gen_jmp_im(target_ulong pc)
@@ -857,9 +843,11 @@
tcg_gen_extu_i32_tl(reg, cpu_tmp2_i32);
}
-static inline void gen_setcc_slow_T0(int op)
+static inline void gen_setcc_slow_T0(DisasContext *s, int jcc_op)
{
- switch(op) {
+ if (s->cc_op != CC_OP_DYNAMIC)
+ gen_op_set_cc_op(s->cc_op);
+ switch(jcc_op) {
case JCC_O:
gen_compute_eflags(cpu_T[0]);
tcg_gen_shri_tl(cpu_T[0], cpu_T[0], 11);
@@ -1151,7 +1139,7 @@
break;
default:
slow_jcc:
- gen_setcc_slow_T0(jcc_op);
+ gen_setcc_slow_T0(s, jcc_op);
tcg_gen_brcondi_tl(inv ? TCG_COND_EQ : TCG_COND_NE,
cpu_T[0], 0, l1);
break;
@@ -1436,7 +1424,8 @@
{
target_ulong mask;
int shift_label;
-
+ TCGv t0, t1;
+
if (ot == OT_QUAD)
mask = 0x3f;
else
@@ -1477,11 +1466,18 @@
if (s->cc_op != CC_OP_DYNAMIC)
gen_op_set_cc_op(s->cc_op);
+ /* XXX: inefficient */
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ t1 = tcg_temp_local_new(TCG_TYPE_TL);
+
+ tcg_gen_mov_tl(t0, cpu_T[0]);
+ tcg_gen_mov_tl(t1, cpu_T3);
+
shift_label = gen_new_label();
tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T[1], 0, shift_label);
- tcg_gen_mov_tl(cpu_cc_src, cpu_T3);
- tcg_gen_mov_tl(cpu_cc_dst, cpu_T[0]);
+ tcg_gen_mov_tl(cpu_cc_src, t1);
+ tcg_gen_mov_tl(cpu_cc_dst, t0);
if (is_right)
tcg_gen_movi_i32(cpu_cc_op, CC_OP_SARB + ot);
else
@@ -1489,6 +1485,9 @@
gen_set_label(shift_label);
s->cc_op = CC_OP_DYNAMIC; /* cannot predict flags after */
+
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
}
static void gen_shift_rm_im(DisasContext *s, int ot, int op1, int op2,
@@ -1556,78 +1555,95 @@
{
target_ulong mask;
int label1, label2, data_bits;
-
+ TCGv t0, t1, t2, a0;
+
+ /* XXX: inefficient, but we must use local temps */
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ t1 = tcg_temp_local_new(TCG_TYPE_TL);
+ t2 = tcg_temp_local_new(TCG_TYPE_TL);
+ a0 = tcg_temp_local_new(TCG_TYPE_TL);
+
if (ot == OT_QUAD)
mask = 0x3f;
else
mask = 0x1f;
/* load */
- if (op1 == OR_TMP0)
- gen_op_ld_T0_A0(ot + s->mem_index);
- else
- gen_op_mov_TN_reg(ot, 0, op1);
+ if (op1 == OR_TMP0) {
+ tcg_gen_mov_tl(a0, cpu_A0);
+ gen_op_ld_v(ot + s->mem_index, t0, a0);
+ } else {
+ gen_op_mov_v_reg(ot, t0, op1);
+ }
- tcg_gen_andi_tl(cpu_T[1], cpu_T[1], mask);
+ tcg_gen_mov_tl(t1, cpu_T[1]);
+ tcg_gen_andi_tl(t1, t1, mask);
+
/* Must test zero case to avoid using undefined behaviour in TCG
shifts. */
label1 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T[1], 0, label1);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t1, 0, label1);
if (ot <= OT_WORD)
- tcg_gen_andi_tl(cpu_tmp0, cpu_T[1], (1 << (3 + ot)) - 1);
+ tcg_gen_andi_tl(cpu_tmp0, t1, (1 << (3 + ot)) - 1);
else
- tcg_gen_mov_tl(cpu_tmp0, cpu_T[1]);
+ tcg_gen_mov_tl(cpu_tmp0, t1);
- gen_extu(ot, cpu_T[0]);
- tcg_gen_mov_tl(cpu_T3, cpu_T[0]);
+ gen_extu(ot, t0);
+ tcg_gen_mov_tl(t2, t0);
data_bits = 8 << ot;
/* XXX: rely on behaviour of shifts when operand 2 overflows (XXX:
fix TCG definition) */
if (is_right) {
- tcg_gen_shr_tl(cpu_tmp4, cpu_T[0], cpu_tmp0);
+ tcg_gen_shr_tl(cpu_tmp4, t0, cpu_tmp0);
tcg_gen_sub_tl(cpu_tmp0, tcg_const_tl(data_bits), cpu_tmp0);
- tcg_gen_shl_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
+ tcg_gen_shl_tl(t0, t0, cpu_tmp0);
} else {
- tcg_gen_shl_tl(cpu_tmp4, cpu_T[0], cpu_tmp0);
+ tcg_gen_shl_tl(cpu_tmp4, t0, cpu_tmp0);
tcg_gen_sub_tl(cpu_tmp0, tcg_const_tl(data_bits), cpu_tmp0);
- tcg_gen_shr_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
+ tcg_gen_shr_tl(t0, t0, cpu_tmp0);
}
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_tmp4);
+ tcg_gen_or_tl(t0, t0, cpu_tmp4);
gen_set_label(label1);
/* store */
- if (op1 == OR_TMP0)
- gen_op_st_T0_A0(ot + s->mem_index);
- else
- gen_op_mov_reg_T0(ot, op1);
+ if (op1 == OR_TMP0) {
+ gen_op_st_v(ot + s->mem_index, t0, a0);
+ } else {
+ gen_op_mov_reg_v(ot, op1, t0);
+ }
/* update eflags */
if (s->cc_op != CC_OP_DYNAMIC)
gen_op_set_cc_op(s->cc_op);
label2 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T[1], 0, label2);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t1, 0, label2);
gen_compute_eflags(cpu_cc_src);
tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, ~(CC_O | CC_C));
- tcg_gen_xor_tl(cpu_tmp0, cpu_T3, cpu_T[0]);
+ tcg_gen_xor_tl(cpu_tmp0, t2, t0);
tcg_gen_lshift(cpu_tmp0, cpu_tmp0, 11 - (data_bits - 1));
tcg_gen_andi_tl(cpu_tmp0, cpu_tmp0, CC_O);
tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, cpu_tmp0);
if (is_right) {
- tcg_gen_shri_tl(cpu_T[0], cpu_T[0], data_bits - 1);
+ tcg_gen_shri_tl(t0, t0, data_bits - 1);
}
- tcg_gen_andi_tl(cpu_T[0], cpu_T[0], CC_C);
- tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, cpu_T[0]);
+ tcg_gen_andi_tl(t0, t0, CC_C);
+ tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, t0);
tcg_gen_discard_tl(cpu_cc_dst);
tcg_gen_movi_i32(cpu_cc_op, CC_OP_EFLAGS);
gen_set_label(label2);
s->cc_op = CC_OP_DYNAMIC; /* cannot predict flags after */
+
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
+ tcg_temp_free(t2);
+ tcg_temp_free(a0);
}
static void *helper_rotc[8] = {
@@ -1666,9 +1682,9 @@
/* update eflags */
label1 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T3, -1, label1);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_cc_tmp, -1, label1);
- tcg_gen_mov_tl(cpu_cc_src, cpu_T3);
+ tcg_gen_mov_tl(cpu_cc_src, cpu_cc_tmp);
tcg_gen_discard_tl(cpu_cc_dst);
tcg_gen_movi_i32(cpu_cc_op, CC_OP_EFLAGS);
@@ -1682,102 +1698,115 @@
{
int label1, label2, data_bits;
target_ulong mask;
+ TCGv t0, t1, t2, a0;
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ t1 = tcg_temp_local_new(TCG_TYPE_TL);
+ t2 = tcg_temp_local_new(TCG_TYPE_TL);
+ a0 = tcg_temp_local_new(TCG_TYPE_TL);
+
if (ot == OT_QUAD)
mask = 0x3f;
else
mask = 0x1f;
/* load */
- if (op1 == OR_TMP0)
- gen_op_ld_T0_A0(ot + s->mem_index);
- else
- gen_op_mov_TN_reg(ot, 0, op1);
+ if (op1 == OR_TMP0) {
+ tcg_gen_mov_tl(a0, cpu_A0);
+ gen_op_ld_v(ot + s->mem_index, t0, a0);
+ } else {
+ gen_op_mov_v_reg(ot, t0, op1);
+ }
tcg_gen_andi_tl(cpu_T3, cpu_T3, mask);
+
+ tcg_gen_mov_tl(t1, cpu_T[1]);
+ tcg_gen_mov_tl(t2, cpu_T3);
+
/* Must test zero case to avoid using undefined behaviour in TCG
shifts. */
label1 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T3, 0, label1);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t2, 0, label1);
- tcg_gen_addi_tl(cpu_tmp5, cpu_T3, -1);
+ tcg_gen_addi_tl(cpu_tmp5, t2, -1);
if (ot == OT_WORD) {
/* Note: we implement the Intel behaviour for shift count > 16 */
if (is_right) {
- tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 0xffff);
- tcg_gen_shli_tl(cpu_tmp0, cpu_T[1], 16);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
- tcg_gen_ext32u_tl(cpu_T[0], cpu_T[0]);
+ tcg_gen_andi_tl(t0, t0, 0xffff);
+ tcg_gen_shli_tl(cpu_tmp0, t1, 16);
+ tcg_gen_or_tl(t0, t0, cpu_tmp0);
+ tcg_gen_ext32u_tl(t0, t0);
- tcg_gen_shr_tl(cpu_tmp4, cpu_T[0], cpu_tmp5);
+ tcg_gen_shr_tl(cpu_tmp4, t0, cpu_tmp5);
/* only needed if count > 16, but a test would complicate */
- tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(32), cpu_T3);
- tcg_gen_shl_tl(cpu_tmp0, cpu_T[0], cpu_tmp5);
+ tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(32), t2);
+ tcg_gen_shl_tl(cpu_tmp0, t0, cpu_tmp5);
- tcg_gen_shr_tl(cpu_T[0], cpu_T[0], cpu_T3);
+ tcg_gen_shr_tl(t0, t0, t2);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
+ tcg_gen_or_tl(t0, t0, cpu_tmp0);
} else {
/* XXX: not optimal */
- tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 0xffff);
- tcg_gen_shli_tl(cpu_T[1], cpu_T[1], 16);
- tcg_gen_or_tl(cpu_T[1], cpu_T[1], cpu_T[0]);
- tcg_gen_ext32u_tl(cpu_T[1], cpu_T[1]);
+ tcg_gen_andi_tl(t0, t0, 0xffff);
+ tcg_gen_shli_tl(t1, t1, 16);
+ tcg_gen_or_tl(t1, t1, t0);
+ tcg_gen_ext32u_tl(t1, t1);
- tcg_gen_shl_tl(cpu_tmp4, cpu_T[0], cpu_tmp5);
+ tcg_gen_shl_tl(cpu_tmp4, t0, cpu_tmp5);
tcg_gen_sub_tl(cpu_tmp0, tcg_const_tl(32), cpu_tmp5);
- tcg_gen_shr_tl(cpu_tmp6, cpu_T[1], cpu_tmp0);
+ tcg_gen_shr_tl(cpu_tmp6, t1, cpu_tmp0);
tcg_gen_or_tl(cpu_tmp4, cpu_tmp4, cpu_tmp6);
- tcg_gen_shl_tl(cpu_T[0], cpu_T[0], cpu_T3);
- tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(32), cpu_T3);
- tcg_gen_shr_tl(cpu_T[1], cpu_T[1], cpu_tmp5);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_T[1]);
+ tcg_gen_shl_tl(t0, t0, t2);
+ tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(32), t2);
+ tcg_gen_shr_tl(t1, t1, cpu_tmp5);
+ tcg_gen_or_tl(t0, t0, t1);
}
} else {
data_bits = 8 << ot;
if (is_right) {
if (ot == OT_LONG)
- tcg_gen_ext32u_tl(cpu_T[0], cpu_T[0]);
+ tcg_gen_ext32u_tl(t0, t0);
- tcg_gen_shr_tl(cpu_tmp4, cpu_T[0], cpu_tmp5);
+ tcg_gen_shr_tl(cpu_tmp4, t0, cpu_tmp5);
- tcg_gen_shr_tl(cpu_T[0], cpu_T[0], cpu_T3);
- tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(data_bits), cpu_T3);
- tcg_gen_shl_tl(cpu_T[1], cpu_T[1], cpu_tmp5);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_T[1]);
+ tcg_gen_shr_tl(t0, t0, t2);
+ tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(data_bits), t2);
+ tcg_gen_shl_tl(t1, t1, cpu_tmp5);
+ tcg_gen_or_tl(t0, t0, t1);
} else {
if (ot == OT_LONG)
- tcg_gen_ext32u_tl(cpu_T[1], cpu_T[1]);
+ tcg_gen_ext32u_tl(t1, t1);
- tcg_gen_shl_tl(cpu_tmp4, cpu_T[0], cpu_tmp5);
+ tcg_gen_shl_tl(cpu_tmp4, t0, cpu_tmp5);
- tcg_gen_shl_tl(cpu_T[0], cpu_T[0], cpu_T3);
- tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(data_bits), cpu_T3);
- tcg_gen_shr_tl(cpu_T[1], cpu_T[1], cpu_tmp5);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_T[1]);
+ tcg_gen_shl_tl(t0, t0, t2);
+ tcg_gen_sub_tl(cpu_tmp5, tcg_const_tl(data_bits), t2);
+ tcg_gen_shr_tl(t1, t1, cpu_tmp5);
+ tcg_gen_or_tl(t0, t0, t1);
}
}
- tcg_gen_mov_tl(cpu_T[1], cpu_tmp4);
+ tcg_gen_mov_tl(t1, cpu_tmp4);
gen_set_label(label1);
/* store */
- if (op1 == OR_TMP0)
- gen_op_st_T0_A0(ot + s->mem_index);
- else
- gen_op_mov_reg_T0(ot, op1);
+ if (op1 == OR_TMP0) {
+ gen_op_st_v(ot + s->mem_index, t0, a0);
+ } else {
+ gen_op_mov_reg_v(ot, op1, t0);
+ }
/* update eflags */
if (s->cc_op != CC_OP_DYNAMIC)
gen_op_set_cc_op(s->cc_op);
label2 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T3, 0, label2);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t2, 0, label2);
- tcg_gen_mov_tl(cpu_cc_src, cpu_T[1]);
- tcg_gen_mov_tl(cpu_cc_dst, cpu_T[0]);
+ tcg_gen_mov_tl(cpu_cc_src, t1);
+ tcg_gen_mov_tl(cpu_cc_dst, t0);
if (is_right) {
tcg_gen_movi_i32(cpu_cc_op, CC_OP_SARB + ot);
} else {
@@ -1785,6 +1814,11 @@
}
gen_set_label(label2);
s->cc_op = CC_OP_DYNAMIC; /* cannot predict flags after */
+
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
+ tcg_temp_free(t2);
+ tcg_temp_free(a0);
}
static void gen_shift(DisasContext *s1, int op, int ot, int d, int s)
@@ -2217,23 +2251,26 @@
static void gen_setcc(DisasContext *s, int b)
{
int inv, jcc_op, l1;
+ TCGv t0;
if (is_fast_jcc_case(s, b)) {
/* nominal case: we use a jump */
- tcg_gen_movi_tl(cpu_T[0], 0);
+ /* XXX: make it faster by adding new instructions in TCG */
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ tcg_gen_movi_tl(t0, 0);
l1 = gen_new_label();
gen_jcc1(s, s->cc_op, b ^ 1, l1);
- tcg_gen_movi_tl(cpu_T[0], 1);
+ tcg_gen_movi_tl(t0, 1);
gen_set_label(l1);
+ tcg_gen_mov_tl(cpu_T[0], t0);
+ tcg_temp_free(t0);
} else {
/* slow case: it is more efficient not to generate a jump,
although it is questionnable whether this optimization is
worth to */
inv = b & 1;
jcc_op = (b >> 1) & 7;
- if (s->cc_op != CC_OP_DYNAMIC)
- gen_op_set_cc_op(s->cc_op);
- gen_setcc_slow_T0(jcc_op);
+ gen_setcc_slow_T0(s, jcc_op);
if (inv) {
tcg_gen_xori_tl(cpu_T[0], cpu_T[0], 1);
}
@@ -4353,6 +4390,7 @@
case 0x1b1: /* cmpxchg Ev, Gv */
{
int label1, label2;
+ TCGv t0, t1, t2, a0;
if ((b & 1) == 0)
ot = OT_BYTE;
@@ -4361,37 +4399,46 @@
modrm = ldub_code(s->pc++);
reg = ((modrm >> 3) & 7) | rex_r;
mod = (modrm >> 6) & 3;
- gen_op_mov_TN_reg(ot, 1, reg);
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ t1 = tcg_temp_local_new(TCG_TYPE_TL);
+ t2 = tcg_temp_local_new(TCG_TYPE_TL);
+ a0 = tcg_temp_local_new(TCG_TYPE_TL);
+ gen_op_mov_v_reg(ot, t1, reg);
if (mod == 3) {
rm = (modrm & 7) | REX_B(s);
- gen_op_mov_TN_reg(ot, 0, rm);
+ gen_op_mov_v_reg(ot, t0, rm);
} else {
gen_lea_modrm(s, modrm, ®_addr, &offset_addr);
- gen_op_ld_T0_A0(ot + s->mem_index);
+ tcg_gen_mov_tl(a0, cpu_A0);
+ gen_op_ld_v(ot + s->mem_index, t0, a0);
rm = 0; /* avoid warning */
}
label1 = gen_new_label();
- tcg_gen_ld_tl(cpu_T3, cpu_env, offsetof(CPUState, regs[R_EAX]));
- tcg_gen_sub_tl(cpu_T3, cpu_T3, cpu_T[0]);
- gen_extu(ot, cpu_T3);
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T3, 0, label1);
+ tcg_gen_ld_tl(t2, cpu_env, offsetof(CPUState, regs[R_EAX]));
+ tcg_gen_sub_tl(t2, t2, t0);
+ gen_extu(ot, t2);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t2, 0, label1);
if (mod == 3) {
label2 = gen_new_label();
- gen_op_mov_reg_T0(ot, R_EAX);
+ gen_op_mov_reg_v(ot, R_EAX, t0);
tcg_gen_br(label2);
gen_set_label(label1);
- gen_op_mov_reg_T1(ot, rm);
+ gen_op_mov_reg_v(ot, rm, t1);
gen_set_label(label2);
} else {
- tcg_gen_mov_tl(cpu_T[1], cpu_T[0]);
- gen_op_mov_reg_T0(ot, R_EAX);
+ tcg_gen_mov_tl(t1, t0);
+ gen_op_mov_reg_v(ot, R_EAX, t0);
gen_set_label(label1);
/* always store */
- gen_op_st_T1_A0(ot + s->mem_index);
+ gen_op_st_v(ot + s->mem_index, t1, a0);
}
- tcg_gen_mov_tl(cpu_cc_src, cpu_T[0]);
- tcg_gen_mov_tl(cpu_cc_dst, cpu_T3);
+ tcg_gen_mov_tl(cpu_cc_src, t0);
+ tcg_gen_mov_tl(cpu_cc_dst, t2);
s->cc_op = CC_OP_SUBB + ot;
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
+ tcg_temp_free(t2);
+ tcg_temp_free(a0);
}
break;
case 0x1c7: /* cmpxchg8b */
@@ -5457,10 +5504,9 @@
(JCC_BE << 1),
(JCC_P << 1),
};
- op1 = fcmov_cc[op & 3] | ((op >> 3) & 1);
- gen_setcc(s, op1);
+ op1 = fcmov_cc[op & 3] | (((op >> 3) & 1) ^ 1);
l1 = gen_new_label();
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T[0], 0, l1);
+ gen_jcc1(s, s->cc_op, op1, l1);
tcg_gen_helper_0_1(helper_fmov_ST0_STN,
tcg_const_i32(opreg));
gen_set_label(l1);
}
@@ -5806,25 +5852,26 @@
case 0x140 ... 0x14f: /* cmov Gv, Ev */
{
int l1;
+ TCGv t0;
+
ot = dflag + OT_WORD;
modrm = ldub_code(s->pc++);
reg = ((modrm >> 3) & 7) | rex_r;
mod = (modrm >> 6) & 3;
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
if (mod != 3) {
gen_lea_modrm(s, modrm, ®_addr, &offset_addr);
- gen_op_ld_T1_A0(ot + s->mem_index);
+ gen_op_ld_v(ot + s->mem_index, t0, cpu_A0);
} else {
rm = (modrm & 7) | REX_B(s);
- gen_op_mov_TN_reg(ot, 1, rm);
+ gen_op_mov_v_reg(ot, t0, rm);
}
- if (s->cc_op != CC_OP_DYNAMIC)
- gen_op_set_cc_op(s->cc_op);
#ifdef TARGET_X86_64
if (ot == OT_LONG) {
/* XXX: specific Intel behaviour ? */
l1 = gen_new_label();
gen_jcc1(s, s->cc_op, b ^ 1, l1);
- tcg_gen_st32_tl(cpu_T[1], cpu_env, offsetof(CPUState,
regs[reg]) + REG_L_OFFSET);
+ tcg_gen_st32_tl(t0, cpu_env, offsetof(CPUState, regs[reg]) +
REG_L_OFFSET);
gen_set_label(l1);
tcg_gen_movi_tl(cpu_tmp0, 0);
tcg_gen_st32_tl(cpu_tmp0, cpu_env, offsetof(CPUState,
regs[reg]) + REG_LH_OFFSET);
@@ -5833,9 +5880,10 @@
{
l1 = gen_new_label();
gen_jcc1(s, s->cc_op, b ^ 1, l1);
- gen_op_mov_reg_T1(ot, reg);
+ gen_op_mov_reg_v(ot, reg, t0);
gen_set_label(l1);
}
+ tcg_temp_free(t0);
}
break;
@@ -6039,6 +6087,8 @@
case 0x1bd: /* bsr */
{
int label1;
+ TCGv t0;
+
ot = dflag + OT_WORD;
modrm = ldub_code(s->pc++);
reg = ((modrm >> 3) & 7) | rex_r;
@@ -6046,17 +6096,20 @@
gen_extu(ot, cpu_T[0]);
label1 = gen_new_label();
tcg_gen_movi_tl(cpu_cc_dst, 0);
- tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_T[0], 0, label1);
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ tcg_gen_mov_tl(t0, cpu_T[0]);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, t0, 0, label1);
if (b & 1) {
- tcg_gen_helper_1_1(helper_bsr, cpu_T[0], cpu_T[0]);
+ tcg_gen_helper_1_1(helper_bsr, cpu_T[0], t0);
} else {
- tcg_gen_helper_1_1(helper_bsf, cpu_T[0], cpu_T[0]);
+ tcg_gen_helper_1_1(helper_bsf, cpu_T[0], t0);
}
gen_op_mov_reg_T0(ot, reg);
tcg_gen_movi_tl(cpu_cc_dst, 1);
gen_set_label(label1);
tcg_gen_discard_tl(cpu_cc_src);
s->cc_op = CC_OP_LOGICB + ot;
+ tcg_temp_free(t0);
}
break;
/************************/
@@ -6725,8 +6778,13 @@
#endif
{
int label1;
+ TCGv t0, t1, t2;
+
if (!s->pe || s->vm86)
goto illegal_op;
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
+ t1 = tcg_temp_local_new(TCG_TYPE_TL);
+ t2 = tcg_temp_local_new(TCG_TYPE_TL);
ot = OT_WORD;
modrm = ldub_code(s->pc++);
reg = (modrm >> 3) & 7;
@@ -6734,55 +6792,61 @@
rm = modrm & 7;
if (mod != 3) {
gen_lea_modrm(s, modrm, ®_addr, &offset_addr);
- gen_op_ld_T0_A0(ot + s->mem_index);
+ gen_op_ld_v(ot + s->mem_index, t0, cpu_A0);
} else {
- gen_op_mov_TN_reg(ot, 0, rm);
+ gen_op_mov_v_reg(ot, t0, rm);
}
- gen_op_mov_TN_reg(ot, 1, reg);
- tcg_gen_andi_tl(cpu_tmp0, cpu_T[0], 3);
- tcg_gen_andi_tl(cpu_T[1], cpu_T[1], 3);
- tcg_gen_movi_tl(cpu_T3, 0);
+ gen_op_mov_v_reg(ot, t1, reg);
+ tcg_gen_andi_tl(cpu_tmp0, t0, 3);
+ tcg_gen_andi_tl(t1, t1, 3);
+ tcg_gen_movi_tl(t2, 0);
label1 = gen_new_label();
- tcg_gen_brcond_tl(TCG_COND_GE, cpu_tmp0, cpu_T[1], label1);
- tcg_gen_andi_tl(cpu_T[0], cpu_T[0], ~3);
- tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_T[1]);
- tcg_gen_movi_tl(cpu_T3, CC_Z);
+ tcg_gen_brcond_tl(TCG_COND_GE, cpu_tmp0, t1, label1);
+ tcg_gen_andi_tl(t0, t0, ~3);
+ tcg_gen_or_tl(t0, t0, t1);
+ tcg_gen_movi_tl(t2, CC_Z);
gen_set_label(label1);
if (mod != 3) {
- gen_op_st_T0_A0(ot + s->mem_index);
+ gen_op_st_v(ot + s->mem_index, t0, cpu_A0);
} else {
- gen_op_mov_reg_T0(ot, rm);
+ gen_op_mov_reg_v(ot, rm, t0);
}
if (s->cc_op != CC_OP_DYNAMIC)
gen_op_set_cc_op(s->cc_op);
gen_compute_eflags(cpu_cc_src);
tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, ~CC_Z);
- tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, cpu_T3);
+ tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, t2);
s->cc_op = CC_OP_EFLAGS;
+ tcg_temp_free(t0);
+ tcg_temp_free(t1);
+ tcg_temp_free(t2);
}
break;
case 0x102: /* lar */
case 0x103: /* lsl */
{
int label1;
+ TCGv t0;
if (!s->pe || s->vm86)
goto illegal_op;
ot = dflag ? OT_LONG : OT_WORD;
modrm = ldub_code(s->pc++);
reg = ((modrm >> 3) & 7) | rex_r;
gen_ldst_modrm(s, modrm, OT_WORD, OR_TMP0, 0);
+ t0 = tcg_temp_local_new(TCG_TYPE_TL);
if (s->cc_op != CC_OP_DYNAMIC)
gen_op_set_cc_op(s->cc_op);
if (b == 0x102)
- tcg_gen_helper_1_1(helper_lar, cpu_T[0], cpu_T[0]);
+ tcg_gen_helper_1_1(helper_lar, t0, cpu_T[0]);
else
- tcg_gen_helper_1_1(helper_lsl, cpu_T[0], cpu_T[0]);
+ tcg_gen_helper_1_1(helper_lsl, t0, cpu_T[0]);
tcg_gen_andi_tl(cpu_tmp0, cpu_cc_src, CC_Z);
label1 = gen_new_label();
tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_tmp0, 0, label1);
- gen_op_mov_reg_T0(ot, reg);
+ gen_op_mov_reg_v(ot, reg, t0);
gen_set_label(label1);
s->cc_op = CC_OP_EFLAGS;
+ tcg_temp_free(t0);
}
break;
case 0x118:
@@ -7029,17 +7093,6 @@
return s->pc;
}
-static void tcg_macro_func(TCGContext *s, int macro_id, const int *dead_args)
-{
- switch(macro_id) {
-#ifdef MACRO_TEST
- case MACRO_TEST:
- tcg_gen_helper_0_1(helper_divl_EAX_T0, cpu_T[0]);
- break;
-#endif
- }
-}
-
void optimize_flags_init(void)
{
#if TCG_TARGET_REG_BITS == 32
@@ -7047,33 +7100,15 @@
#else
assert(sizeof(CCTable) == (1 << 4));
#endif
- tcg_set_macro_func(&tcg_ctx, tcg_macro_func);
-
cpu_env = tcg_global_reg_new(TCG_TYPE_PTR, TCG_AREG0, "env");
-#if TARGET_LONG_BITS > HOST_LONG_BITS
- cpu_T[0] = tcg_global_mem_new(TCG_TYPE_TL,
- TCG_AREG0, offsetof(CPUState, t0), "T0");
- cpu_T[1] = tcg_global_mem_new(TCG_TYPE_TL,
- TCG_AREG0, offsetof(CPUState, t1), "T1");
- cpu_A0 = tcg_global_mem_new(TCG_TYPE_TL,
- TCG_AREG0, offsetof(CPUState, t2), "A0");
-#else
- cpu_T[0] = tcg_global_reg_new(TCG_TYPE_TL, TCG_AREG1, "T0");
- cpu_T[1] = tcg_global_reg_new(TCG_TYPE_TL, TCG_AREG2, "T1");
- cpu_A0 = tcg_global_reg_new(TCG_TYPE_TL, TCG_AREG3, "A0");
-#endif
- cpu_T3 = tcg_global_mem_new(TCG_TYPE_TL,
- TCG_AREG0, offsetof(CPUState, t3), "T3");
-#if defined(__i386__) && (TARGET_LONG_BITS <= HOST_LONG_BITS)
- /* XXX: must be suppressed once there are less fixed registers */
- cpu_tmp1_i64 = tcg_global_reg2_new_hack(TCG_TYPE_I64, TCG_AREG1,
TCG_AREG2, "tmp1");
-#endif
cpu_cc_op = tcg_global_mem_new(TCG_TYPE_I32,
TCG_AREG0, offsetof(CPUState, cc_op),
"cc_op");
cpu_cc_src = tcg_global_mem_new(TCG_TYPE_TL,
TCG_AREG0, offsetof(CPUState, cc_src),
"cc_src");
cpu_cc_dst = tcg_global_mem_new(TCG_TYPE_TL,
TCG_AREG0, offsetof(CPUState, cc_dst),
"cc_dst");
+ cpu_cc_tmp = tcg_global_mem_new(TCG_TYPE_TL,
+ TCG_AREG0, offsetof(CPUState, cc_tmp),
"cc_tmp");
/* register helpers */
@@ -7145,10 +7180,13 @@
printf("ERROR addseg\n");
#endif
+ cpu_T[0] = tcg_temp_new(TCG_TYPE_TL);
+ cpu_T[1] = tcg_temp_new(TCG_TYPE_TL);
+ cpu_A0 = tcg_temp_new(TCG_TYPE_TL);
+ cpu_T3 = tcg_temp_new(TCG_TYPE_TL);
+
cpu_tmp0 = tcg_temp_new(TCG_TYPE_TL);
-#if !(defined(__i386__) && (TARGET_LONG_BITS <= HOST_LONG_BITS))
cpu_tmp1_i64 = tcg_temp_new(TCG_TYPE_I64);
-#endif
cpu_tmp2_i32 = tcg_temp_new(TCG_TYPE_I32);
cpu_tmp3_i32 = tcg_temp_new(TCG_TYPE_I32);
cpu_tmp4 = tcg_temp_new(TCG_TYPE_TL);
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Qemu-devel] [4577] transformed TN into temporaries - add local temporaries usage when needed - optimized fcmovX,
Fabrice Bellard <=