[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [5189] Convert basic float32 ops to TCG
|
From: |
Blue Swirl |
|
Subject: |
[Qemu-devel] [5189] Convert basic float32 ops to TCG |
|
Date: |
Wed, 10 Sep 2008 19:54:52 +0000 |
Revision: 5189
http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=5189
Author: blueswir1
Date: 2008-09-10 19:54:51 +0000 (Wed, 10 Sep 2008)
Log Message:
-----------
Convert basic float32 ops to TCG
Modified Paths:
--------------
trunk/target-sparc/helper.h
trunk/target-sparc/op_helper.c
trunk/target-sparc/translate.c
Modified: trunk/target-sparc/helper.h
===================================================================
--- trunk/target-sparc/helper.h 2008-09-10 18:18:38 UTC (rev 5188)
+++ trunk/target-sparc/helper.h 2008-09-10 19:54:51 UTC (rev 5189)
@@ -58,12 +58,12 @@
DEF_HELPER(void, helper_ldfsr, (uint32_t new_fsr))
DEF_HELPER(void, helper_check_ieee_exceptions, (void))
DEF_HELPER(void, helper_clear_float_exceptions, (void))
-DEF_HELPER(void, helper_fabss, (void))
-DEF_HELPER(void, helper_fsqrts, (void))
+DEF_HELPER(float32, helper_fabss, (float32 src))
+DEF_HELPER(float32, helper_fsqrts, (float32 src))
DEF_HELPER(void, helper_fsqrtd, (void))
-DEF_HELPER(void, helper_fcmps, (void))
+DEF_HELPER(void, helper_fcmps, (float32 src1, float32 src2))
DEF_HELPER(void, helper_fcmpd, (void))
-DEF_HELPER(void, helper_fcmpes, (void))
+DEF_HELPER(void, helper_fcmpes, (float32 src1, float32 src2))
DEF_HELPER(void, helper_fcmped, (void))
DEF_HELPER(void, helper_fsqrtq, (void))
DEF_HELPER(void, helper_fcmpq, (void))
@@ -71,17 +71,17 @@
#ifdef TARGET_SPARC64
DEF_HELPER(void, helper_ldxfsr, (uint64_t new_fsr))
DEF_HELPER(void, helper_fabsd, (void))
-DEF_HELPER(void, helper_fcmps_fcc1, (void))
+DEF_HELPER(void, helper_fcmps_fcc1, (float32 src1, float32 src2))
+DEF_HELPER(void, helper_fcmps_fcc2, (float32 src1, float32 src2))
+DEF_HELPER(void, helper_fcmps_fcc3, (float32 src1, float32 src2))
DEF_HELPER(void, helper_fcmpd_fcc1, (void))
-DEF_HELPER(void, helper_fcmps_fcc2, (void))
DEF_HELPER(void, helper_fcmpd_fcc2, (void))
-DEF_HELPER(void, helper_fcmps_fcc3, (void))
DEF_HELPER(void, helper_fcmpd_fcc3, (void))
-DEF_HELPER(void, helper_fcmpes_fcc1, (void))
+DEF_HELPER(void, helper_fcmpes_fcc1, (float32 src1, float32 src2))
+DEF_HELPER(void, helper_fcmpes_fcc2, (float32 src1, float32 src2))
+DEF_HELPER(void, helper_fcmpes_fcc3, (float32 src1, float32 src2))
DEF_HELPER(void, helper_fcmped_fcc1, (void))
-DEF_HELPER(void, helper_fcmpes_fcc2, (void))
DEF_HELPER(void, helper_fcmped_fcc2, (void))
-DEF_HELPER(void, helper_fcmpes_fcc3, (void))
DEF_HELPER(void, helper_fcmped_fcc3, (void))
DEF_HELPER(void, helper_fabsq, (void))
DEF_HELPER(void, helper_fcmpq_fcc1, (void))
@@ -97,17 +97,28 @@
F_HELPER_0_0(name ## s); \
F_HELPER_0_0(name ## d); \
F_HELPER_0_0(name ## q)
+#define F_HELPER_DQ_0_0(name) \
+ F_HELPER_0_0(name ## d); \
+ F_HELPER_0_0(name ## q)
-F_HELPER_SDQ_0_0(add);
-F_HELPER_SDQ_0_0(sub);
-F_HELPER_SDQ_0_0(mul);
-F_HELPER_SDQ_0_0(div);
+F_HELPER_DQ_0_0(add);
+F_HELPER_DQ_0_0(sub);
+F_HELPER_DQ_0_0(mul);
+F_HELPER_DQ_0_0(div);
+DEF_HELPER(float32, helper_fadds, (float32 src1, float32 src2))
+DEF_HELPER(float32, helper_fsubs, (float32 src1, float32 src2))
+DEF_HELPER(float32, helper_fmuls, (float32 src1, float32 src2))
+DEF_HELPER(float32, helper_fdivs, (float32 src1, float32 src2))
+
F_HELPER_0_0(smuld);
F_HELPER_0_0(dmulq);
-DEF_HELPER(void, helper_fnegs, (void))
-F_HELPER_SDQ_0_0(ito);
+DEF_HELPER(float32, helper_fnegs, (float32 src))
+F_HELPER_DQ_0_0(ito);
+
+DEF_HELPER(float32, helper_fitos, (int32_t src))
+
#ifdef TARGET_SPARC64
DEF_HELPER(void, helper_fnegd, (void))
DEF_HELPER(void, helper_fnegq, (void))
@@ -119,7 +130,7 @@
F_HELPER_0_0(stoq);
F_HELPER_0_0(qtod);
F_HELPER_0_0(dtoq);
-F_HELPER_0_0(stoi);
+DEF_HELPER(int32_t, helper_fstoi, (float32 src))
F_HELPER_0_0(dtoi);
F_HELPER_0_0(qtoi);
#ifdef TARGET_SPARC64
@@ -176,6 +187,6 @@
VIS_CMPHELPER(cmpne);
#endif
#undef F_HELPER_0_0
-#undef F_HELPER_SDQ_0_0
+#undef F_HELPER_DQ_0_0
#undef VIS_HELPER
#undef VIS_CMPHELPER
Modified: trunk/target-sparc/op_helper.c
===================================================================
--- trunk/target-sparc/op_helper.c 2008-09-10 18:18:38 UTC (rev 5188)
+++ trunk/target-sparc/op_helper.c 2008-09-10 19:54:51 UTC (rev 5189)
@@ -87,9 +87,9 @@
#define F_HELPER(name, p) void helper_f##name##p(void)
#define F_BINOP(name) \
- F_HELPER(name, s) \
+ float32 helper_f ## name ## s (float32 src1, float32 src2) \
{ \
- FT0 = float32_ ## name (FT0, FT1, &env->fp_status); \
+ return float32_ ## name (src1, src2, &env->fp_status); \
} \
F_HELPER(name, d) \
{ \
@@ -120,9 +120,9 @@
&env->fp_status);
}
-F_HELPER(neg, s)
+float32 helper_fnegs(float32 src)
{
- FT0 = float32_chs(FT1);
+ return float32_chs(src);
}
#ifdef TARGET_SPARC64
@@ -138,9 +138,9 @@
#endif
/* Integer to float conversion. */
-F_HELPER(ito, s)
+float32 helper_fitos(int32_t src)
{
- FT0 = int32_to_float32(*((int32_t *)&FT1), &env->fp_status);
+ return int32_to_float32(src, &env->fp_status);
}
F_HELPER(ito, d)
@@ -203,9 +203,9 @@
}
/* Float to integer conversion. */
-void helper_fstoi(void)
+int32_t helper_fstoi(float32 src)
{
- *((int32_t *)&FT0) = float32_to_int32_round_to_zero(FT1, &env->fp_status);
+ return float32_to_int32_round_to_zero(src, &env->fp_status);
}
void helper_fdtoi(void)
@@ -714,9 +714,9 @@
set_float_exception_flags(0, &env->fp_status);
}
-void helper_fabss(void)
+float32 helper_fabss(float32 src)
{
- FT0 = float32_abs(FT1);
+ return float32_abs(src);
}
#ifdef TARGET_SPARC64
@@ -731,9 +731,9 @@
}
#endif
-void helper_fsqrts(void)
+float32 helper_fsqrts(float32 src)
{
- FT0 = float32_sqrt(FT1, &env->fp_status);
+ return float32_sqrt(src, &env->fp_status);
}
void helper_fsqrtd(void)
@@ -776,41 +776,72 @@
} \
env->fsr |= new_fsr; \
}
+#define GEN_FCMPS(name, size, FS, TRAP) \
+ void glue(helper_, name)(float32 src1, float32 src2) \
+ { \
+ target_ulong new_fsr; \
+ \
+ env->fsr &= ~((FSR_FCC1 | FSR_FCC0) << FS); \
+ switch (glue(size, _compare) (src1, src2, &env->fp_status)) { \
+ case float_relation_unordered: \
+ new_fsr = (FSR_FCC1 | FSR_FCC0) << FS; \
+ if ((env->fsr & FSR_NVM) || TRAP) { \
+ env->fsr |= new_fsr; \
+ env->fsr |= FSR_NVC; \
+ env->fsr |= FSR_FTT_IEEE_EXCP; \
+ raise_exception(TT_FP_EXCP); \
+ } else { \
+ env->fsr |= FSR_NVA; \
+ } \
+ break; \
+ case float_relation_less: \
+ new_fsr = FSR_FCC0 << FS; \
+ break; \
+ case float_relation_greater: \
+ new_fsr = FSR_FCC1 << FS; \
+ break; \
+ default: \
+ new_fsr = 0; \
+ break; \
+ } \
+ env->fsr |= new_fsr; \
+ }
-GEN_FCMP(fcmps, float32, FT0, FT1, 0, 0);
+GEN_FCMPS(fcmps, float32, 0, 0);
GEN_FCMP(fcmpd, float64, DT0, DT1, 0, 0);
-GEN_FCMP(fcmpes, float32, FT0, FT1, 0, 1);
+GEN_FCMPS(fcmpes, float32, 0, 1);
GEN_FCMP(fcmped, float64, DT0, DT1, 0, 1);
GEN_FCMP(fcmpq, float128, QT0, QT1, 0, 0);
GEN_FCMP(fcmpeq, float128, QT0, QT1, 0, 1);
#ifdef TARGET_SPARC64
-GEN_FCMP(fcmps_fcc1, float32, FT0, FT1, 22, 0);
+GEN_FCMPS(fcmps_fcc1, float32, 22, 0);
GEN_FCMP(fcmpd_fcc1, float64, DT0, DT1, 22, 0);
GEN_FCMP(fcmpq_fcc1, float128, QT0, QT1, 22, 0);
-GEN_FCMP(fcmps_fcc2, float32, FT0, FT1, 24, 0);
+GEN_FCMPS(fcmps_fcc2, float32, 24, 0);
GEN_FCMP(fcmpd_fcc2, float64, DT0, DT1, 24, 0);
GEN_FCMP(fcmpq_fcc2, float128, QT0, QT1, 24, 0);
-GEN_FCMP(fcmps_fcc3, float32, FT0, FT1, 26, 0);
+GEN_FCMPS(fcmps_fcc3, float32, 26, 0);
GEN_FCMP(fcmpd_fcc3, float64, DT0, DT1, 26, 0);
GEN_FCMP(fcmpq_fcc3, float128, QT0, QT1, 26, 0);
-GEN_FCMP(fcmpes_fcc1, float32, FT0, FT1, 22, 1);
+GEN_FCMPS(fcmpes_fcc1, float32, 22, 1);
GEN_FCMP(fcmped_fcc1, float64, DT0, DT1, 22, 1);
GEN_FCMP(fcmpeq_fcc1, float128, QT0, QT1, 22, 1);
-GEN_FCMP(fcmpes_fcc2, float32, FT0, FT1, 24, 1);
+GEN_FCMPS(fcmpes_fcc2, float32, 24, 1);
GEN_FCMP(fcmped_fcc2, float64, DT0, DT1, 24, 1);
GEN_FCMP(fcmpeq_fcc2, float128, QT0, QT1, 24, 1);
-GEN_FCMP(fcmpes_fcc3, float32, FT0, FT1, 26, 1);
+GEN_FCMPS(fcmpes_fcc3, float32, 26, 1);
GEN_FCMP(fcmped_fcc3, float64, DT0, DT1, 26, 1);
GEN_FCMP(fcmpeq_fcc3, float128, QT0, QT1, 26, 1);
#endif
+#undef GEN_FCMPS
#if !defined(TARGET_SPARC64) && !defined(CONFIG_USER_ONLY) && \
defined(DEBUG_MXCC)
@@ -2220,7 +2251,7 @@
switch(size) {
default:
case 4:
- *((uint32_t *)&FT0) = val;
+ *((uint32_t *)&env->fpr[rd]) = val;
break;
case 8:
*((int64_t *)&DT0) = val;
@@ -2261,7 +2292,7 @@
switch(size) {
default:
case 4:
- val = *((uint32_t *)&FT0);
+ val = *((uint32_t *)&env->fpr[rd]);
break;
case 8:
val = *((int64_t *)&DT0);
Modified: trunk/target-sparc/translate.c
===================================================================
--- trunk/target-sparc/translate.c 2008-09-10 18:18:38 UTC (rev 5188)
+++ trunk/target-sparc/translate.c 2008-09-10 19:54:51 UTC (rev 5189)
@@ -55,6 +55,8 @@
#endif
/* local register indexes (only used inside old micro ops) */
static TCGv cpu_tmp0, cpu_tmp32, cpu_tmp64;
+/* Floating point registers */
+static TCGv cpu_fpr[TARGET_FPREGS];
#include "gen-icount.h"
@@ -102,98 +104,77 @@
/* floating point registers moves */
static void gen_op_load_fpr_FT0(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, ft0));
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, ft0));
}
static void gen_op_load_fpr_FT1(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, ft1));
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, ft1));
}
static void gen_op_store_FT0_fpr(unsigned int dst)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, ft0));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst]));
+ tcg_gen_ld_i32(cpu_fpr[dst], cpu_env, offsetof(CPUSPARCState, ft0));
}
static void gen_op_load_fpr_DT0(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt0) +
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, dt0) +
offsetof(CPU_DoubleU, l.upper));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 1]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt0) +
+ tcg_gen_st_i32(cpu_fpr[src + 1], cpu_env, offsetof(CPUSPARCState, dt0) +
offsetof(CPU_DoubleU, l.lower));
}
static void gen_op_load_fpr_DT1(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt1) +
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, dt1) +
offsetof(CPU_DoubleU, l.upper));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 1]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt1) +
+ tcg_gen_st_i32(cpu_fpr[src + 1], cpu_env, offsetof(CPUSPARCState, dt1) +
offsetof(CPU_DoubleU, l.lower));
}
static void gen_op_store_DT0_fpr(unsigned int dst)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst], cpu_env, offsetof(CPUSPARCState, dt0) +
offsetof(CPU_DoubleU, l.upper));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst]));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, dt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst + 1], cpu_env, offsetof(CPUSPARCState, dt0) +
offsetof(CPU_DoubleU, l.lower));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst + 1]));
}
static void gen_op_load_fpr_QT0(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.upmost));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 1]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_st_i32(cpu_fpr[src + 1], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.upper));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 2]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_st_i32(cpu_fpr[src + 2], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.lower));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 3]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_st_i32(cpu_fpr[src + 3], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.lowest));
}
static void gen_op_load_fpr_QT1(unsigned int src)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt1) +
+ tcg_gen_st_i32(cpu_fpr[src], cpu_env, offsetof(CPUSPARCState, qt1) +
offsetof(CPU_QuadU, l.upmost));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 1]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt1) +
+ tcg_gen_st_i32(cpu_fpr[src + 1], cpu_env, offsetof(CPUSPARCState, qt1) +
offsetof(CPU_QuadU, l.upper));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 2]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt1) +
+ tcg_gen_st_i32(cpu_fpr[src + 2], cpu_env, offsetof(CPUSPARCState, qt1) +
offsetof(CPU_QuadU, l.lower));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[src + 3]));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt1) +
+ tcg_gen_st_i32(cpu_fpr[src + 3], cpu_env, offsetof(CPUSPARCState, qt1) +
offsetof(CPU_QuadU, l.lowest));
}
static void gen_op_store_QT0_fpr(unsigned int dst)
{
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.upmost));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst]));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst + 1], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.upper));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst + 1]));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst + 2], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.lower));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst + 2]));
- tcg_gen_ld_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, qt0) +
+ tcg_gen_ld_i32(cpu_fpr[dst + 3], cpu_env, offsetof(CPUSPARCState, qt0) +
offsetof(CPU_QuadU, l.lowest));
- tcg_gen_st_i32(cpu_tmp32, cpu_env, offsetof(CPUSPARCState, fpr[dst + 3]));
}
/* moves */
@@ -1496,13 +1477,6 @@
}
}
-static GenOpFunc * const gen_fcmps[4] = {
- helper_fcmps,
- helper_fcmps_fcc1,
- helper_fcmps_fcc2,
- helper_fcmps_fcc3,
-};
-
static GenOpFunc * const gen_fcmpd[4] = {
helper_fcmpd,
helper_fcmpd_fcc1,
@@ -1517,13 +1491,6 @@
helper_fcmpq_fcc3,
};
-static GenOpFunc * const gen_fcmpes[4] = {
- helper_fcmpes,
- helper_fcmpes_fcc1,
- helper_fcmpes_fcc2,
- helper_fcmpes_fcc3,
-};
-
static GenOpFunc * const gen_fcmped[4] = {
helper_fcmped,
helper_fcmped_fcc1,
@@ -1538,9 +1505,22 @@
helper_fcmpeq_fcc3,
};
-static inline void gen_op_fcmps(int fccno)
+static inline void gen_op_fcmps(int fccno, TCGv r_rs1, TCGv r_rs2)
{
- tcg_gen_helper_0_0(gen_fcmps[fccno]);
+ switch (fccno) {
+ case 0:
+ tcg_gen_helper_0_2(helper_fcmps, r_rs1, r_rs2);
+ break;
+ case 1:
+ tcg_gen_helper_0_2(helper_fcmps_fcc1, r_rs1, r_rs2);
+ break;
+ case 2:
+ tcg_gen_helper_0_2(helper_fcmps_fcc2, r_rs1, r_rs2);
+ break;
+ case 3:
+ tcg_gen_helper_0_2(helper_fcmps_fcc3, r_rs1, r_rs2);
+ break;
+ }
}
static inline void gen_op_fcmpd(int fccno)
@@ -1553,9 +1533,22 @@
tcg_gen_helper_0_0(gen_fcmpq[fccno]);
}
-static inline void gen_op_fcmpes(int fccno)
+static inline void gen_op_fcmpes(int fccno, TCGv r_rs1, TCGv r_rs2)
{
- tcg_gen_helper_0_0(gen_fcmpes[fccno]);
+ switch (fccno) {
+ case 0:
+ tcg_gen_helper_0_2(helper_fcmpes, r_rs1, r_rs2);
+ break;
+ case 1:
+ tcg_gen_helper_0_2(helper_fcmpes_fcc1, r_rs1, r_rs2);
+ break;
+ case 2:
+ tcg_gen_helper_0_2(helper_fcmpes_fcc2, r_rs1, r_rs2);
+ break;
+ case 3:
+ tcg_gen_helper_0_2(helper_fcmpes_fcc3, r_rs1, r_rs2);
+ break;
+ }
}
static inline void gen_op_fcmped(int fccno)
@@ -1570,9 +1563,9 @@
#else
-static inline void gen_op_fcmps(int fccno)
+static inline void gen_op_fcmps(int fccno, TCGv r_rs1, TCGv r_rs2)
{
- tcg_gen_helper_0_0(helper_fcmps);
+ tcg_gen_helper_0_2(helper_fcmps, r_rs1, r_rs2);
}
static inline void gen_op_fcmpd(int fccno)
@@ -1585,9 +1578,9 @@
tcg_gen_helper_0_0(helper_fcmpq);
}
-static inline void gen_op_fcmpes(int fccno)
+static inline void gen_op_fcmpes(int fccno, TCGv r_rs1, TCGv r_rs2)
{
- tcg_gen_helper_0_0(helper_fcmpes);
+ tcg_gen_helper_0_2(helper_fcmpes, r_rs1, r_rs2);
}
static inline void gen_op_fcmped(int fccno)
@@ -2349,26 +2342,23 @@
xop = GET_FIELD(insn, 18, 26);
switch (xop) {
case 0x1: /* fmovs */
- gen_op_load_fpr_FT0(rs2);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_fpr[rs2]);
break;
case 0x5: /* fnegs */
- gen_op_load_fpr_FT1(rs2);
- tcg_gen_helper_0_0(helper_fnegs);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_helper_1_1(helper_fnegs, cpu_fpr[rd],
+ cpu_fpr[rs2]);
break;
case 0x9: /* fabss */
- gen_op_load_fpr_FT1(rs2);
- tcg_gen_helper_0_0(helper_fabss);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_helper_1_1(helper_fabss, cpu_fpr[rd],
+ cpu_fpr[rs2]);
break;
case 0x29: /* fsqrts */
CHECK_FPU_FEATURE(dc, FSQRT);
- gen_op_load_fpr_FT1(rs2);
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fsqrts);
+ tcg_gen_helper_1_1(helper_fsqrts, cpu_tmp32,
+ cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0x2a: /* fsqrtd */
CHECK_FPU_FEATURE(dc, FSQRT);
@@ -2386,13 +2376,12 @@
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
gen_op_store_QT0_fpr(QFPREG(rd));
break;
- case 0x41:
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
+ case 0x41: /* fadds */
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fadds);
+ tcg_gen_helper_1_2(helper_fadds, cpu_tmp32,
+ cpu_fpr[rs1], cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0x42:
gen_op_load_fpr_DT0(DFPREG(rs1));
@@ -2411,13 +2400,12 @@
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
gen_op_store_QT0_fpr(QFPREG(rd));
break;
- case 0x45:
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
+ case 0x45: /* fsubs */
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fsubs);
+ tcg_gen_helper_1_2(helper_fsubs, cpu_tmp32,
+ cpu_fpr[rs1], cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0x46:
gen_op_load_fpr_DT0(DFPREG(rs1));
@@ -2438,12 +2426,11 @@
break;
case 0x49: /* fmuls */
CHECK_FPU_FEATURE(dc, FMUL);
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fmuls);
+ tcg_gen_helper_1_2(helper_fmuls, cpu_tmp32,
+ cpu_fpr[rs1], cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0x4a: /* fmuld */
CHECK_FPU_FEATURE(dc, FMUL);
@@ -2464,13 +2451,12 @@
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
gen_op_store_QT0_fpr(QFPREG(rd));
break;
- case 0x4d:
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
+ case 0x4d: /* fdivs */
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fdivs);
+ tcg_gen_helper_1_2(helper_fdivs, cpu_tmp32,
+ cpu_fpr[rs1], cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0x4e:
gen_op_load_fpr_DT0(DFPREG(rs1));
@@ -2507,12 +2493,12 @@
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
gen_op_store_QT0_fpr(QFPREG(rd));
break;
- case 0xc4:
- gen_op_load_fpr_FT1(rs2);
+ case 0xc4: /* fitos */
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fitos);
+ tcg_gen_helper_1_1(helper_fitos, cpu_tmp32,
+ cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0xc6:
gen_op_load_fpr_DT1(DFPREG(rs2));
@@ -2565,12 +2551,12 @@
tcg_gen_helper_0_0(helper_fdtoq);
gen_op_store_QT0_fpr(QFPREG(rd));
break;
- case 0xd1:
- gen_op_load_fpr_FT1(rs2);
+ case 0xd1: /* fstoi */
gen_clear_float_exceptions();
- tcg_gen_helper_0_0(helper_fstoi);
+ tcg_gen_helper_1_1(helper_fstoi, cpu_tmp32,
+ cpu_fpr[rs2]);
tcg_gen_helper_0_0(helper_check_ieee_exceptions);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_tmp32);
break;
case 0xd2:
gen_op_load_fpr_DT1(DFPREG(rs2));
@@ -2589,13 +2575,21 @@
break;
#ifdef TARGET_SPARC64
case 0x2: /* V9 fmovd */
- gen_op_load_fpr_DT0(DFPREG(rs2));
- gen_op_store_DT0_fpr(DFPREG(rd));
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd)],
+ cpu_fpr[DFPREG(rs2)]);
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd) + 1],
+ cpu_fpr[DFPREG(rs2) + 1]);
break;
case 0x3: /* V9 fmovq */
CHECK_FPU_FEATURE(dc, FLOAT128);
- gen_op_load_fpr_QT0(QFPREG(rs2));
- gen_op_store_QT0_fpr(QFPREG(rd));
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd)],
+ cpu_fpr[QFPREG(rs2)]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 1],
+ cpu_fpr[QFPREG(rs2) + 1]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 2],
+ cpu_fpr[QFPREG(rs2) + 2]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 3],
+ cpu_fpr[QFPREG(rs2) + 3]);
break;
case 0x6: /* V9 fnegd */
gen_op_load_fpr_DT1(DFPREG(rs2));
@@ -2686,8 +2680,7 @@
cpu_src1 = get_src1(insn, cpu_src1);
tcg_gen_brcondi_tl(gen_tcg_cond_reg[cond], cpu_src1,
0, l1);
- gen_op_load_fpr_FT0(rs2);
- gen_op_store_FT0_fpr(rd);
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_fpr[rs2]);
gen_set_label(l1);
break;
} else if ((xop & 0x11f) == 0x006) { // V9 fmovdr
@@ -2698,8 +2691,8 @@
cpu_src1 = get_src1(insn, cpu_src1);
tcg_gen_brcondi_tl(gen_tcg_cond_reg[cond], cpu_src1,
0, l1);
- gen_op_load_fpr_DT0(DFPREG(rs2));
- gen_op_store_DT0_fpr(DFPREG(rd));
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd)], cpu_fpr[DFPREG(rs2)]);
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd) + 1],
cpu_fpr[DFPREG(rs2) + 1]);
gen_set_label(l1);
break;
} else if ((xop & 0x11f) == 0x007) { // V9 fmovqr
@@ -2711,15 +2704,17 @@
cpu_src1 = get_src1(insn, cpu_src1);
tcg_gen_brcondi_tl(gen_tcg_cond_reg[cond], cpu_src1,
0, l1);
- gen_op_load_fpr_QT0(QFPREG(rs2));
- gen_op_store_QT0_fpr(QFPREG(rd));
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd)], cpu_fpr[QFPREG(rs2)]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 1],
cpu_fpr[QFPREG(rs2) + 1]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 2],
cpu_fpr[QFPREG(rs2) + 2]);
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 3],
cpu_fpr[QFPREG(rs2) + 3]);
gen_set_label(l1);
break;
}
#endif
switch (xop) {
#ifdef TARGET_SPARC64
-#define FMOVCC(size_FDQ, fcc) \
+#define FMOVSCC(fcc) \
{ \
TCGv r_cond; \
int l1; \
@@ -2730,54 +2725,93 @@
gen_fcond(r_cond, fcc, cond); \
tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
0, l1); \
- glue(glue(gen_op_load_fpr_, size_FDQ), T0) \
- (glue(size_FDQ, FPREG(rs2))); \
- glue(glue(gen_op_store_, size_FDQ), T0_fpr) \
- (glue(size_FDQ, FPREG(rd))); \
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_fpr[rs2]); \
gen_set_label(l1); \
tcg_temp_free(r_cond); \
}
+#define FMOVDCC(fcc) \
+ { \
+ TCGv r_cond; \
+ int l1; \
+ \
+ l1 = gen_new_label(); \
+ r_cond = tcg_temp_new(TCG_TYPE_TL); \
+ cond = GET_FIELD_SP(insn, 14, 17); \
+ gen_fcond(r_cond, fcc, cond); \
+ tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
+ 0, l1); \
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd)], \
+ cpu_fpr[DFPREG(rs2)]); \
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd) + 1], \
+ cpu_fpr[DFPREG(rs2) + 1]); \
+ gen_set_label(l1); \
+ tcg_temp_free(r_cond); \
+ }
+#define FMOVQCC(fcc) \
+ { \
+ TCGv r_cond; \
+ int l1; \
+ \
+ l1 = gen_new_label(); \
+ r_cond = tcg_temp_new(TCG_TYPE_TL); \
+ cond = GET_FIELD_SP(insn, 14, 17); \
+ gen_fcond(r_cond, fcc, cond); \
+ tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
+ 0, l1); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd)], \
+ cpu_fpr[QFPREG(rs2)]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 1], \
+ cpu_fpr[QFPREG(rs2) + 1]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 2], \
+ cpu_fpr[QFPREG(rs2) + 2]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 3], \
+ cpu_fpr[QFPREG(rs2) + 3]); \
+ gen_set_label(l1); \
+ tcg_temp_free(r_cond); \
+ }
case 0x001: /* V9 fmovscc %fcc0 */
- FMOVCC(F, 0);
+ FMOVSCC(0);
break;
case 0x002: /* V9 fmovdcc %fcc0 */
- FMOVCC(D, 0);
+ FMOVDCC(0);
break;
case 0x003: /* V9 fmovqcc %fcc0 */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 0);
+ FMOVQCC(0);
break;
case 0x041: /* V9 fmovscc %fcc1 */
- FMOVCC(F, 1);
+ FMOVSCC(1);
break;
case 0x042: /* V9 fmovdcc %fcc1 */
- FMOVCC(D, 1);
+ FMOVDCC(1);
break;
case 0x043: /* V9 fmovqcc %fcc1 */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 1);
+ FMOVQCC(1);
break;
case 0x081: /* V9 fmovscc %fcc2 */
- FMOVCC(F, 2);
+ FMOVSCC(2);
break;
case 0x082: /* V9 fmovdcc %fcc2 */
- FMOVCC(D, 2);
+ FMOVDCC(2);
break;
case 0x083: /* V9 fmovqcc %fcc2 */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 2);
+ FMOVQCC(2);
break;
case 0x0c1: /* V9 fmovscc %fcc3 */
- FMOVCC(F, 3);
+ FMOVSCC(3);
break;
case 0x0c2: /* V9 fmovdcc %fcc3 */
- FMOVCC(D, 3);
+ FMOVDCC(3);
break;
case 0x0c3: /* V9 fmovqcc %fcc3 */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 3);
+ FMOVQCC(3);
break;
-#undef FMOVCC
+#undef FMOVSCC
+#undef FMOVDCC
+#undef FMOVQCC
#define FMOVCC(size_FDQ, icc) \
{ \
TCGv r_cond; \
@@ -2796,32 +2830,87 @@
gen_set_label(l1); \
tcg_temp_free(r_cond); \
}
+#define FMOVSCC(icc) \
+ { \
+ TCGv r_cond; \
+ int l1; \
+ \
+ l1 = gen_new_label(); \
+ r_cond = tcg_temp_new(TCG_TYPE_TL); \
+ cond = GET_FIELD_SP(insn, 14, 17); \
+ gen_cond(r_cond, icc, cond); \
+ tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
+ 0, l1); \
+ tcg_gen_mov_i32(cpu_fpr[rd], cpu_fpr[rs2]); \
+ gen_set_label(l1); \
+ tcg_temp_free(r_cond); \
+ }
+#define FMOVDCC(icc) \
+ { \
+ TCGv r_cond; \
+ int l1; \
+ \
+ l1 = gen_new_label(); \
+ r_cond = tcg_temp_new(TCG_TYPE_TL); \
+ cond = GET_FIELD_SP(insn, 14, 17); \
+ gen_cond(r_cond, icc, cond); \
+ tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
+ 0, l1); \
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd)], \
+ cpu_fpr[DFPREG(rs2)]); \
+ tcg_gen_mov_i32(cpu_fpr[DFPREG(rd) + 1], \
+ cpu_fpr[DFPREG(rs2) + 1]); \
+ gen_set_label(l1); \
+ tcg_temp_free(r_cond); \
+ }
+#define FMOVQCC(icc) \
+ { \
+ TCGv r_cond; \
+ int l1; \
+ \
+ l1 = gen_new_label(); \
+ r_cond = tcg_temp_new(TCG_TYPE_TL); \
+ cond = GET_FIELD_SP(insn, 14, 17); \
+ gen_cond(r_cond, icc, cond); \
+ tcg_gen_brcondi_tl(TCG_COND_EQ, r_cond, \
+ 0, l1); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd)], \
+ cpu_fpr[QFPREG(rs2)]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 1], \
+ cpu_fpr[QFPREG(rs2) + 1]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 2], \
+ cpu_fpr[QFPREG(rs2) + 2]); \
+ tcg_gen_mov_i32(cpu_fpr[QFPREG(rd) + 3], \
+ cpu_fpr[QFPREG(rs2) + 3]); \
+ gen_set_label(l1); \
+ tcg_temp_free(r_cond); \
+ }
case 0x101: /* V9 fmovscc %icc */
- FMOVCC(F, 0);
+ FMOVSCC(0);
break;
case 0x102: /* V9 fmovdcc %icc */
- FMOVCC(D, 0);
+ FMOVDCC(0);
case 0x103: /* V9 fmovqcc %icc */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 0);
+ FMOVQCC(0);
break;
case 0x181: /* V9 fmovscc %xcc */
- FMOVCC(F, 1);
+ FMOVSCC(1);
break;
case 0x182: /* V9 fmovdcc %xcc */
- FMOVCC(D, 1);
+ FMOVDCC(1);
break;
case 0x183: /* V9 fmovqcc %xcc */
CHECK_FPU_FEATURE(dc, FLOAT128);
- FMOVCC(Q, 1);
+ FMOVQCC(1);
break;
-#undef FMOVCC
+#undef FMOVSCC
+#undef FMOVDCC
+#undef FMOVQCC
#endif
case 0x51: /* fcmps, V9 %fcc */
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
- gen_op_fcmps(rd & 3);
+ gen_op_fcmps(rd & 3, cpu_fpr[rs1], cpu_fpr[rs2]);
break;
case 0x52: /* fcmpd, V9 %fcc */
gen_op_load_fpr_DT0(DFPREG(rs1));
@@ -2835,9 +2924,7 @@
gen_op_fcmpq(rd & 3);
break;
case 0x55: /* fcmpes, V9 %fcc */
- gen_op_load_fpr_FT0(rs1);
- gen_op_load_fpr_FT1(rs2);
- gen_op_fcmpes(rd & 3);
+ gen_op_fcmpes(rd & 3, cpu_fpr[rs1], cpu_fpr[rs2]);
break;
case 0x56: /* fcmped, V9 %fcc */
gen_op_load_fpr_DT0(DFPREG(rs1));
@@ -4364,9 +4451,7 @@
switch (xop) {
case 0x20: /* load fpreg */
gen_address_mask(dc, cpu_addr);
- tcg_gen_qemu_ld32u(cpu_tmp32, cpu_addr, dc->mem_idx);
- tcg_gen_st_i32(cpu_tmp32, cpu_env,
- offsetof(CPUState, fpr[rd]));
+ tcg_gen_qemu_ld32u(cpu_fpr[rd], cpu_addr, dc->mem_idx);
break;
case 0x21: /* ldfsr, V9 ldxfsr */
#ifdef TARGET_SPARC64
@@ -4508,9 +4593,7 @@
switch (xop) {
case 0x24: /* store fpreg */
gen_address_mask(dc, cpu_addr);
- tcg_gen_ld_i32(cpu_tmp32, cpu_env,
- offsetof(CPUState, fpr[rd]));
- tcg_gen_qemu_st32(cpu_tmp32, cpu_addr, dc->mem_idx);
+ tcg_gen_qemu_st32(cpu_fpr[rd], cpu_addr, dc->mem_idx);
break;
case 0x25: /* stfsr, V9 stxfsr */
#ifdef TARGET_SPARC64
@@ -4570,7 +4653,6 @@
switch (xop) {
#ifdef TARGET_SPARC64
case 0x34: /* V9 stfa */
- gen_op_load_fpr_FT0(rd);
gen_stf_asi(cpu_addr, insn, 4, rd);
break;
case 0x36: /* V9 stqfa */
@@ -4858,6 +4940,16 @@
"g6",
"g7",
};
+ static const char * const fregnames[64] = {
+ "f0", "f1", "f2", "f3", "f4", "f5", "f6", "f7",
+ "f8", "f9", "f10", "f11", "f12", "f13", "f14", "f15",
+ "f16", "f17", "f18", "f19", "f20", "f21", "f22", "f23",
+ "f24", "f25", "f26", "f27", "f28", "f29", "f30", "f31",
+ "f32", "f33", "f34", "f35", "f36", "f37", "f38", "f39",
+ "f40", "f41", "f42", "f43", "f44", "f45", "f46", "f47",
+ "f48", "f49", "f50", "f51", "f52", "f53", "f54", "f55",
+ "f56", "f57", "f58", "f59", "f60", "f61", "f62", "f63",
+ };
/* init various static tables */
if (!inited) {
@@ -4945,6 +5037,11 @@
cpu_gregs[i] = tcg_global_mem_new(TCG_TYPE_TL, TCG_AREG0,
offsetof(CPUState, gregs[i]),
gregnames[i]);
+ for (i = 0; i < TARGET_FPREGS; i++)
+ cpu_fpr[i] = tcg_global_mem_new(TCG_TYPE_I32, TCG_AREG0,
+ offsetof(CPUState, fpr[i]),
+ fregnames[i]);
+
/* register helpers */
#undef DEF_HELPER
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Qemu-devel] [5189] Convert basic float32 ops to TCG,
Blue Swirl <=