aboutsummaryrefslogtreecommitdiff
path: root/target-ppc/fpu_helper.c
diff options
context:
space:
mode:
authorBlue Swirl <blauwirbel@gmail.com>2012-05-30 04:23:27 +0000
committerAlexander Graf <agraf@suse.de>2012-06-24 01:04:41 +0200
commit8e70394950ed55dd7f0a5f9b5f831cf70089d0df (patch)
tree4ae996a6146bf3df6d1275d3de7558e87534f5b5 /target-ppc/fpu_helper.c
parentbd23cd45ea533580da410eed9c31692abf7584e6 (diff)
ppc: Avoid AREG0 for FPU and SPE helpers
Add an explicit CPUPPCState parameter instead of relying on AREG0. Signed-off-by: Blue Swirl <blauwirbel@gmail.com> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Andreas Färber <afaerber@suse.de> Signed-off-by: Alexander Graf <agraf@suse.de>
Diffstat (limited to 'target-ppc/fpu_helper.c')
-rw-r--r--target-ppc/fpu_helper.c350
1 files changed, 179 insertions, 171 deletions
diff --git a/target-ppc/fpu_helper.c b/target-ppc/fpu_helper.c
index be965d106d..d90636fad0 100644
--- a/target-ppc/fpu_helper.c
+++ b/target-ppc/fpu_helper.c
@@ -17,12 +17,11 @@
* License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/
#include "cpu.h"
-#include "dyngen-exec.h"
#include "helper.h"
/*****************************************************************************/
/* Floating point operations helpers */
-uint64_t helper_float32_to_float64(uint32_t arg)
+uint64_t helper_float32_to_float64(CPUPPCState *env, uint32_t arg)
{
CPU_FloatU f;
CPU_DoubleU d;
@@ -32,7 +31,7 @@ uint64_t helper_float32_to_float64(uint32_t arg)
return d.ll;
}
-uint32_t helper_float64_to_float32(uint64_t arg)
+uint32_t helper_float64_to_float32(CPUPPCState *env, uint64_t arg)
{
CPU_FloatU f;
CPU_DoubleU d;
@@ -51,7 +50,7 @@ static inline int isden(float64 d)
return ((u.ll >> 52) & 0x7FF) == 0;
}
-uint32_t helper_compute_fprf(uint64_t arg, uint32_t set_fprf)
+uint32_t helper_compute_fprf(CPUPPCState *env, uint64_t arg, uint32_t set_fprf)
{
CPU_DoubleU farg;
int isneg;
@@ -107,7 +106,7 @@ uint32_t helper_compute_fprf(uint64_t arg, uint32_t set_fprf)
}
/* Floating-point invalid operations exception */
-static inline uint64_t fload_invalid_op_excp(int op)
+static inline uint64_t fload_invalid_op_excp(CPUPPCState *env, int op)
{
uint64_t ret = 0;
int ve;
@@ -190,7 +189,7 @@ static inline uint64_t fload_invalid_op_excp(int op)
return ret;
}
-static inline void float_zero_divide_excp(void)
+static inline void float_zero_divide_excp(CPUPPCState *env)
{
env->fpscr |= 1 << FPSCR_ZX;
env->fpscr &= ~((1 << FPSCR_FR) | (1 << FPSCR_FI));
@@ -206,7 +205,7 @@ static inline void float_zero_divide_excp(void)
}
}
-static inline void float_overflow_excp(void)
+static inline void float_overflow_excp(CPUPPCState *env)
{
env->fpscr |= 1 << FPSCR_OX;
/* Update the floating-point exception summary */
@@ -224,7 +223,7 @@ static inline void float_overflow_excp(void)
}
}
-static inline void float_underflow_excp(void)
+static inline void float_underflow_excp(CPUPPCState *env)
{
env->fpscr |= 1 << FPSCR_UX;
/* Update the floating-point exception summary */
@@ -239,7 +238,7 @@ static inline void float_underflow_excp(void)
}
}
-static inline void float_inexact_excp(void)
+static inline void float_inexact_excp(CPUPPCState *env)
{
env->fpscr |= 1 << FPSCR_XX;
/* Update the floating-point exception summary */
@@ -253,7 +252,7 @@ static inline void float_inexact_excp(void)
}
}
-static inline void fpscr_set_rounding_mode(void)
+static inline void fpscr_set_rounding_mode(CPUPPCState *env)
{
int rnd_type;
@@ -280,7 +279,7 @@ static inline void fpscr_set_rounding_mode(void)
set_float_rounding_mode(rnd_type, &env->fp_status);
}
-void helper_fpscr_clrbit(uint32_t bit)
+void helper_fpscr_clrbit(CPUPPCState *env, uint32_t bit)
{
int prev;
@@ -290,7 +289,7 @@ void helper_fpscr_clrbit(uint32_t bit)
switch (bit) {
case FPSCR_RN1:
case FPSCR_RN:
- fpscr_set_rounding_mode();
+ fpscr_set_rounding_mode(env);
break;
default:
break;
@@ -298,7 +297,7 @@ void helper_fpscr_clrbit(uint32_t bit)
}
}
-void helper_fpscr_setbit(uint32_t bit)
+void helper_fpscr_setbit(CPUPPCState *env, uint32_t bit)
{
int prev;
@@ -414,7 +413,7 @@ void helper_fpscr_setbit(uint32_t bit)
break;
case FPSCR_RN1:
case FPSCR_RN:
- fpscr_set_rounding_mode();
+ fpscr_set_rounding_mode(env);
break;
default:
break;
@@ -428,7 +427,7 @@ void helper_fpscr_setbit(uint32_t bit)
}
}
-void helper_store_fpscr(uint64_t arg, uint32_t mask)
+void helper_store_fpscr(CPUPPCState *env, uint64_t arg, uint32_t mask)
{
/*
* We use only the 32 LSB of the incoming fpr
@@ -460,10 +459,10 @@ void helper_store_fpscr(uint64_t arg, uint32_t mask)
} else {
env->fpscr &= ~(1 << FPSCR_FEX);
}
- fpscr_set_rounding_mode();
+ fpscr_set_rounding_mode(env);
}
-void helper_float_check_status(void)
+void helper_float_check_status(CPUPPCState *env)
{
if (env->exception_index == POWERPC_EXCP_PROGRAM &&
(env->error_code & POWERPC_EXCP_FP)) {
@@ -475,24 +474,24 @@ void helper_float_check_status(void)
} else {
int status = get_float_exception_flags(&env->fp_status);
if (status & float_flag_divbyzero) {
- float_zero_divide_excp();
+ float_zero_divide_excp(env);
} else if (status & float_flag_overflow) {
- float_overflow_excp();
+ float_overflow_excp(env);
} else if (status & float_flag_underflow) {
- float_underflow_excp();
+ float_underflow_excp(env);
} else if (status & float_flag_inexact) {
- float_inexact_excp();
+ float_inexact_excp(env);
}
}
}
-void helper_reset_fpstatus(void)
+void helper_reset_fpstatus(CPUPPCState *env)
{
set_float_exception_flags(0, &env->fp_status);
}
/* fadd - fadd. */
-uint64_t helper_fadd(uint64_t arg1, uint64_t arg2)
+uint64_t helper_fadd(CPUPPCState *env, uint64_t arg1, uint64_t arg2)
{
CPU_DoubleU farg1, farg2;
@@ -502,12 +501,12 @@ uint64_t helper_fadd(uint64_t arg1, uint64_t arg2)
if (unlikely(float64_is_infinity(farg1.d) && float64_is_infinity(farg2.d) &&
float64_is_neg(farg1.d) != float64_is_neg(farg2.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d))) {
/* sNaN addition */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg1.d = float64_add(farg1.d, farg2.d, &env->fp_status);
}
@@ -516,7 +515,7 @@ uint64_t helper_fadd(uint64_t arg1, uint64_t arg2)
}
/* fsub - fsub. */
-uint64_t helper_fsub(uint64_t arg1, uint64_t arg2)
+uint64_t helper_fsub(CPUPPCState *env, uint64_t arg1, uint64_t arg2)
{
CPU_DoubleU farg1, farg2;
@@ -526,12 +525,12 @@ uint64_t helper_fsub(uint64_t arg1, uint64_t arg2)
if (unlikely(float64_is_infinity(farg1.d) && float64_is_infinity(farg2.d) &&
float64_is_neg(farg1.d) == float64_is_neg(farg2.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d))) {
/* sNaN subtraction */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg1.d = float64_sub(farg1.d, farg2.d, &env->fp_status);
}
@@ -540,7 +539,7 @@ uint64_t helper_fsub(uint64_t arg1, uint64_t arg2)
}
/* fmul - fmul. */
-uint64_t helper_fmul(uint64_t arg1, uint64_t arg2)
+uint64_t helper_fmul(CPUPPCState *env, uint64_t arg1, uint64_t arg2)
{
CPU_DoubleU farg1, farg2;
@@ -550,12 +549,12 @@ uint64_t helper_fmul(uint64_t arg1, uint64_t arg2)
if (unlikely((float64_is_infinity(farg1.d) && float64_is_zero(farg2.d)) ||
(float64_is_zero(farg1.d) && float64_is_infinity(farg2.d)))) {
/* Multiplication of zero by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIMZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIMZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d))) {
/* sNaN multiplication */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg1.d = float64_mul(farg1.d, farg2.d, &env->fp_status);
}
@@ -564,7 +563,7 @@ uint64_t helper_fmul(uint64_t arg1, uint64_t arg2)
}
/* fdiv - fdiv. */
-uint64_t helper_fdiv(uint64_t arg1, uint64_t arg2)
+uint64_t helper_fdiv(CPUPPCState *env, uint64_t arg1, uint64_t arg2)
{
CPU_DoubleU farg1, farg2;
@@ -574,15 +573,15 @@ uint64_t helper_fdiv(uint64_t arg1, uint64_t arg2)
if (unlikely(float64_is_infinity(farg1.d) &&
float64_is_infinity(farg2.d))) {
/* Division of infinity by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIDI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIDI);
} else if (unlikely(float64_is_zero(farg1.d) && float64_is_zero(farg2.d))) {
/* Division of zero by zero */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXZDZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXZDZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d))) {
/* sNaN division */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg1.d = float64_div(farg1.d, farg2.d, &env->fp_status);
}
@@ -591,7 +590,7 @@ uint64_t helper_fdiv(uint64_t arg1, uint64_t arg2)
}
/* fabs */
-uint64_t helper_fabs(uint64_t arg)
+uint64_t helper_fabs(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -601,7 +600,7 @@ uint64_t helper_fabs(uint64_t arg)
}
/* fnabs */
-uint64_t helper_fnabs(uint64_t arg)
+uint64_t helper_fnabs(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -612,7 +611,7 @@ uint64_t helper_fnabs(uint64_t arg)
}
/* fneg */
-uint64_t helper_fneg(uint64_t arg)
+uint64_t helper_fneg(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -622,7 +621,7 @@ uint64_t helper_fneg(uint64_t arg)
}
/* fctiw - fctiw. */
-uint64_t helper_fctiw(uint64_t arg)
+uint64_t helper_fctiw(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -630,12 +629,12 @@ uint64_t helper_fctiw(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXCVI);
} else if (unlikely(float64_is_quiet_nan(farg.d) ||
float64_is_infinity(farg.d))) {
/* qNan / infinity conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXCVI);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXCVI);
} else {
farg.ll = float64_to_int32(farg.d, &env->fp_status);
/* XXX: higher bits are not supposed to be significant.
@@ -647,7 +646,7 @@ uint64_t helper_fctiw(uint64_t arg)
}
/* fctiwz - fctiwz. */
-uint64_t helper_fctiwz(uint64_t arg)
+uint64_t helper_fctiwz(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -655,12 +654,12 @@ uint64_t helper_fctiwz(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXCVI);
} else if (unlikely(float64_is_quiet_nan(farg.d) ||
float64_is_infinity(farg.d))) {
/* qNan / infinity conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXCVI);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXCVI);
} else {
farg.ll = float64_to_int32_round_to_zero(farg.d, &env->fp_status);
/* XXX: higher bits are not supposed to be significant.
@@ -673,7 +672,7 @@ uint64_t helper_fctiwz(uint64_t arg)
#if defined(TARGET_PPC64)
/* fcfid - fcfid. */
-uint64_t helper_fcfid(uint64_t arg)
+uint64_t helper_fcfid(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -682,7 +681,7 @@ uint64_t helper_fcfid(uint64_t arg)
}
/* fctid - fctid. */
-uint64_t helper_fctid(uint64_t arg)
+uint64_t helper_fctid(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -690,12 +689,12 @@ uint64_t helper_fctid(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXCVI);
} else if (unlikely(float64_is_quiet_nan(farg.d) ||
float64_is_infinity(farg.d))) {
/* qNan / infinity conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXCVI);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXCVI);
} else {
farg.ll = float64_to_int64(farg.d, &env->fp_status);
}
@@ -703,7 +702,7 @@ uint64_t helper_fctid(uint64_t arg)
}
/* fctidz - fctidz. */
-uint64_t helper_fctidz(uint64_t arg)
+uint64_t helper_fctidz(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -711,12 +710,12 @@ uint64_t helper_fctidz(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXCVI);
} else if (unlikely(float64_is_quiet_nan(farg.d) ||
float64_is_infinity(farg.d))) {
/* qNan / infinity conversion */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXCVI);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXCVI);
} else {
farg.ll = float64_to_int64_round_to_zero(farg.d, &env->fp_status);
}
@@ -725,7 +724,8 @@ uint64_t helper_fctidz(uint64_t arg)
#endif
-static inline uint64_t do_fri(uint64_t arg, int rounding_mode)
+static inline uint64_t do_fri(CPUPPCState *env, uint64_t arg,
+ int rounding_mode)
{
CPU_DoubleU farg;
@@ -733,43 +733,44 @@ static inline uint64_t do_fri(uint64_t arg, int rounding_mode)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN round */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXCVI);
} else if (unlikely(float64_is_quiet_nan(farg.d) ||
float64_is_infinity(farg.d))) {
/* qNan / infinity round */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXCVI);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXCVI);
} else {
set_float_rounding_mode(rounding_mode, &env->fp_status);
farg.ll = float64_round_to_int(farg.d, &env->fp_status);
/* Restore rounding mode from FPSCR */
- fpscr_set_rounding_mode();
+ fpscr_set_rounding_mode(env);
}
return farg.ll;
}
-uint64_t helper_frin(uint64_t arg)
+uint64_t helper_frin(CPUPPCState *env, uint64_t arg)
{
- return do_fri(arg, float_round_nearest_even);
+ return do_fri(env, arg, float_round_nearest_even);
}
-uint64_t helper_friz(uint64_t arg)
+uint64_t helper_friz(CPUPPCState *env, uint64_t arg)
{
- return do_fri(arg, float_round_to_zero);
+ return do_fri(env, arg, float_round_to_zero);
}
-uint64_t helper_frip(uint64_t arg)
+uint64_t helper_frip(CPUPPCState *env, uint64_t arg)
{
- return do_fri(arg, float_round_up);
+ return do_fri(env, arg, float_round_up);
}
-uint64_t helper_frim(uint64_t arg)
+uint64_t helper_frim(CPUPPCState *env, uint64_t arg)
{
- return do_fri(arg, float_round_down);
+ return do_fri(env, arg, float_round_down);
}
/* fmadd - fmadd. */
-uint64_t helper_fmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
+uint64_t helper_fmadd(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint64_t arg3)
{
CPU_DoubleU farg1, farg2, farg3;
@@ -780,13 +781,13 @@ uint64_t helper_fmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
if (unlikely((float64_is_infinity(farg1.d) && float64_is_zero(farg2.d)) ||
(float64_is_zero(farg1.d) && float64_is_infinity(farg2.d)))) {
/* Multiplication of zero by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIMZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIMZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d) ||
float64_is_signaling_nan(farg3.d))) {
/* sNaN operation */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
/* This is the way the PowerPC specification defines it */
float128 ft0_128, ft1_128;
@@ -798,7 +799,7 @@ uint64_t helper_fmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
float64_is_infinity(farg3.d) &&
float128_is_neg(ft0_128) != float64_is_neg(farg3.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
ft1_128 = float64_to_float128(farg3.d, &env->fp_status);
ft0_128 = float128_add(ft0_128, ft1_128, &env->fp_status);
@@ -810,7 +811,8 @@ uint64_t helper_fmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
}
/* fmsub - fmsub. */
-uint64_t helper_fmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
+uint64_t helper_fmsub(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint64_t arg3)
{
CPU_DoubleU farg1, farg2, farg3;
@@ -822,13 +824,13 @@ uint64_t helper_fmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
(float64_is_zero(farg1.d) &&
float64_is_infinity(farg2.d)))) {
/* Multiplication of zero by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIMZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIMZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d) ||
float64_is_signaling_nan(farg3.d))) {
/* sNaN operation */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
/* This is the way the PowerPC specification defines it */
float128 ft0_128, ft1_128;
@@ -840,7 +842,7 @@ uint64_t helper_fmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
float64_is_infinity(farg3.d) &&
float128_is_neg(ft0_128) == float64_is_neg(farg3.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
ft1_128 = float64_to_float128(farg3.d, &env->fp_status);
ft0_128 = float128_sub(ft0_128, ft1_128, &env->fp_status);
@@ -851,7 +853,8 @@ uint64_t helper_fmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
}
/* fnmadd - fnmadd. */
-uint64_t helper_fnmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
+uint64_t helper_fnmadd(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint64_t arg3)
{
CPU_DoubleU farg1, farg2, farg3;
@@ -862,13 +865,13 @@ uint64_t helper_fnmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
if (unlikely((float64_is_infinity(farg1.d) && float64_is_zero(farg2.d)) ||
(float64_is_zero(farg1.d) && float64_is_infinity(farg2.d)))) {
/* Multiplication of zero by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIMZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIMZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d) ||
float64_is_signaling_nan(farg3.d))) {
/* sNaN operation */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
/* This is the way the PowerPC specification defines it */
float128 ft0_128, ft1_128;
@@ -880,7 +883,7 @@ uint64_t helper_fnmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
float64_is_infinity(farg3.d) &&
float128_is_neg(ft0_128) != float64_is_neg(farg3.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
ft1_128 = float64_to_float128(farg3.d, &env->fp_status);
ft0_128 = float128_add(ft0_128, ft1_128, &env->fp_status);
@@ -894,7 +897,8 @@ uint64_t helper_fnmadd(uint64_t arg1, uint64_t arg2, uint64_t arg3)
}
/* fnmsub - fnmsub. */
-uint64_t helper_fnmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
+uint64_t helper_fnmsub(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint64_t arg3)
{
CPU_DoubleU farg1, farg2, farg3;
@@ -906,13 +910,13 @@ uint64_t helper_fnmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
(float64_is_zero(farg1.d) &&
float64_is_infinity(farg2.d)))) {
/* Multiplication of zero by infinity */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXIMZ);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXIMZ);
} else {
if (unlikely(float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d) ||
float64_is_signaling_nan(farg3.d))) {
/* sNaN operation */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
/* This is the way the PowerPC specification defines it */
float128 ft0_128, ft1_128;
@@ -924,7 +928,7 @@ uint64_t helper_fnmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
float64_is_infinity(farg3.d) &&
float128_is_neg(ft0_128) == float64_is_neg(farg3.d))) {
/* Magnitude subtraction of infinities */
- farg1.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXISI);
+ farg1.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXISI);
} else {
ft1_128 = float64_to_float128(farg3.d, &env->fp_status);
ft0_128 = float128_sub(ft0_128, ft1_128, &env->fp_status);
@@ -938,7 +942,7 @@ uint64_t helper_fnmsub(uint64_t arg1, uint64_t arg2, uint64_t arg3)
}
/* frsp - frsp. */
-uint64_t helper_frsp(uint64_t arg)
+uint64_t helper_frsp(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
float32 f32;
@@ -947,7 +951,7 @@ uint64_t helper_frsp(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN square root */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
f32 = float64_to_float32(farg.d, &env->fp_status);
farg.d = float32_to_float64(f32, &env->fp_status);
@@ -956,7 +960,7 @@ uint64_t helper_frsp(uint64_t arg)
}
/* fsqrt - fsqrt. */
-uint64_t helper_fsqrt(uint64_t arg)
+uint64_t helper_fsqrt(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -964,11 +968,11 @@ uint64_t helper_fsqrt(uint64_t arg)
if (unlikely(float64_is_neg(farg.d) && !float64_is_zero(farg.d))) {
/* Square root of a negative nonzero number */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSQRT);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSQRT);
} else {
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN square root */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg.d = float64_sqrt(farg.d, &env->fp_status);
}
@@ -976,7 +980,7 @@ uint64_t helper_fsqrt(uint64_t arg)
}
/* fre - fre. */
-uint64_t helper_fre(uint64_t arg)
+uint64_t helper_fre(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
@@ -984,14 +988,14 @@ uint64_t helper_fre(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN reciprocal */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg.d = float64_div(float64_one, farg.d, &env->fp_status);
return farg.d;
}
/* fres - fres. */
-uint64_t helper_fres(uint64_t arg)
+uint64_t helper_fres(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
float32 f32;
@@ -1000,7 +1004,7 @@ uint64_t helper_fres(uint64_t arg)
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN reciprocal */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg.d = float64_div(float64_one, farg.d, &env->fp_status);
f32 = float64_to_float32(farg.d, &env->fp_status);
@@ -1010,7 +1014,7 @@ uint64_t helper_fres(uint64_t arg)
}
/* frsqrte - frsqrte. */
-uint64_t helper_frsqrte(uint64_t arg)
+uint64_t helper_frsqrte(CPUPPCState *env, uint64_t arg)
{
CPU_DoubleU farg;
float32 f32;
@@ -1019,11 +1023,11 @@ uint64_t helper_frsqrte(uint64_t arg)
if (unlikely(float64_is_neg(farg.d) && !float64_is_zero(farg.d))) {
/* Reciprocal square root of a negative nonzero number */
- farg.ll = fload_invalid_op_excp(POWERPC_EXCP_FP_VXSQRT);
+ farg.ll = fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSQRT);
} else {
if (unlikely(float64_is_signaling_nan(farg.d))) {
/* sNaN reciprocal square root */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
farg.d = float64_sqrt(farg.d, &env->fp_status);
farg.d = float64_div(float64_one, farg.d, &env->fp_status);
@@ -1034,7 +1038,8 @@ uint64_t helper_frsqrte(uint64_t arg)
}
/* fsel - fsel. */
-uint64_t helper_fsel(uint64_t arg1, uint64_t arg2, uint64_t arg3)
+uint64_t helper_fsel(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint64_t arg3)
{
CPU_DoubleU farg1;
@@ -1048,7 +1053,8 @@ uint64_t helper_fsel(uint64_t arg1, uint64_t arg2, uint64_t arg3)
}
}
-void helper_fcmpu(uint64_t arg1, uint64_t arg2, uint32_t crfD)
+void helper_fcmpu(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint32_t crfD)
{
CPU_DoubleU farg1, farg2;
uint32_t ret = 0;
@@ -1074,11 +1080,12 @@ void helper_fcmpu(uint64_t arg1, uint64_t arg2, uint32_t crfD)
&& (float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d)))) {
/* sNaN comparison */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN);
}
}
-void helper_fcmpo(uint64_t arg1, uint64_t arg2, uint32_t crfD)
+void helper_fcmpo(CPUPPCState *env, uint64_t arg1, uint64_t arg2,
+ uint32_t crfD)
{
CPU_DoubleU farg1, farg2;
uint32_t ret = 0;
@@ -1104,17 +1111,17 @@ void helper_fcmpo(uint64_t arg1, uint64_t arg2, uint32_t crfD)
if (float64_is_signaling_nan(farg1.d) ||
float64_is_signaling_nan(farg2.d)) {
/* sNaN comparison */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXSNAN |
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXSNAN |
POWERPC_EXCP_FP_VXVC);
} else {
/* qNaN comparison */
- fload_invalid_op_excp(POWERPC_EXCP_FP_VXVC);
+ fload_invalid_op_excp(env, POWERPC_EXCP_FP_VXVC);
}
}
}
/* Single-precision floating-point conversions */
-static inline uint32_t efscfsi(uint32_t val)
+static inline uint32_t efscfsi(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1123,7 +1130,7 @@ static inline uint32_t efscfsi(uint32_t val)
return u.l;
}
-static inline uint32_t efscfui(uint32_t val)
+static inline uint32_t efscfui(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1132,7 +1139,7 @@ static inline uint32_t efscfui(uint32_t val)
return u.l;
}
-static inline int32_t efsctsi(uint32_t val)
+static inline int32_t efsctsi(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1145,7 +1152,7 @@ static inline int32_t efsctsi(uint32_t val)
return float32_to_int32(u.f, &env->vec_status);
}
-static inline uint32_t efsctui(uint32_t val)
+static inline uint32_t efsctui(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1158,7 +1165,7 @@ static inline uint32_t efsctui(uint32_t val)
return float32_to_uint32(u.f, &env->vec_status);
}
-static inline uint32_t efsctsiz(uint32_t val)
+static inline uint32_t efsctsiz(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1171,7 +1178,7 @@ static inline uint32_t efsctsiz(uint32_t val)
return float32_to_int32_round_to_zero(u.f, &env->vec_status);
}
-static inline uint32_t efsctuiz(uint32_t val)
+static inline uint32_t efsctuiz(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
@@ -1184,7 +1191,7 @@ static inline uint32_t efsctuiz(uint32_t val)
return float32_to_uint32_round_to_zero(u.f, &env->vec_status);
}
-static inline uint32_t efscfsf(uint32_t val)
+static inline uint32_t efscfsf(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
float32 tmp;
@@ -1196,7 +1203,7 @@ static inline uint32_t efscfsf(uint32_t val)
return u.l;
}
-static inline uint32_t efscfuf(uint32_t val)
+static inline uint32_t efscfuf(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
float32 tmp;
@@ -1208,7 +1215,7 @@ static inline uint32_t efscfuf(uint32_t val)
return u.l;
}
-static inline uint32_t efsctsf(uint32_t val)
+static inline uint32_t efsctsf(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
float32 tmp;
@@ -1224,7 +1231,7 @@ static inline uint32_t efsctsf(uint32_t val)
return float32_to_int32(u.f, &env->vec_status);
}
-static inline uint32_t efsctuf(uint32_t val)
+static inline uint32_t efsctuf(CPUPPCState *env, uint32_t val)
{
CPU_FloatU u;
float32 tmp;
@@ -1240,10 +1247,10 @@ static inline uint32_t efsctuf(uint32_t val)
return float32_to_uint32(u.f, &env->vec_status);
}
-#define HELPER_SPE_SINGLE_CONV(name) \
- uint32_t helper_e##name(uint32_t val) \
- { \
- return e##name(val); \
+#define HELPER_SPE_SINGLE_CONV(name) \
+ uint32_t helper_e##name(CPUPPCState *env, uint32_t val) \
+ { \
+ return e##name(env, val); \
}
/* efscfsi */
HELPER_SPE_SINGLE_CONV(fscfsi);
@@ -1266,11 +1273,11 @@ HELPER_SPE_SINGLE_CONV(fsctsf);
/* efsctuf */
HELPER_SPE_SINGLE_CONV(fsctuf);
-#define HELPER_SPE_VECTOR_CONV(name) \
- uint64_t helper_ev##name(uint64_t val) \
- { \
- return ((uint64_t)e##name(val >> 32) << 32) | \
- (uint64_t)e##name(val); \
+#define HELPER_SPE_VECTOR_CONV(name) \
+ uint64_t helper_ev##name(CPUPPCState *env, uint64_t val) \
+ { \
+ return ((uint64_t)e##name(env, val >> 32) << 32) | \
+ (uint64_t)e##name(env, val); \
}
/* evfscfsi */
HELPER_SPE_VECTOR_CONV(fscfsi);
@@ -1294,7 +1301,7 @@ HELPER_SPE_VECTOR_CONV(fsctsf);
HELPER_SPE_VECTOR_CONV(fsctuf);
/* Single-precision floating-point arithmetic */
-static inline uint32_t efsadd(uint32_t op1, uint32_t op2)
+static inline uint32_t efsadd(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1304,7 +1311,7 @@ static inline uint32_t efsadd(uint32_t op1, uint32_t op2)
return u1.l;
}
-static inline uint32_t efssub(uint32_t op1, uint32_t op2)
+static inline uint32_t efssub(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1314,7 +1321,7 @@ static inline uint32_t efssub(uint32_t op1, uint32_t op2)
return u1.l;
}
-static inline uint32_t efsmul(uint32_t op1, uint32_t op2)
+static inline uint32_t efsmul(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1324,7 +1331,7 @@ static inline uint32_t efsmul(uint32_t op1, uint32_t op2)
return u1.l;
}
-static inline uint32_t efsdiv(uint32_t op1, uint32_t op2)
+static inline uint32_t efsdiv(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1334,10 +1341,10 @@ static inline uint32_t efsdiv(uint32_t op1, uint32_t op2)
return u1.l;
}
-#define HELPER_SPE_SINGLE_ARITH(name) \
- uint32_t helper_e##name(uint32_t op1, uint32_t op2) \
- { \
- return e##name(op1, op2); \
+#define HELPER_SPE_SINGLE_ARITH(name) \
+ uint32_t helper_e##name(CPUPPCState *env, uint32_t op1, uint32_t op2) \
+ { \
+ return e##name(env, op1, op2); \
}
/* efsadd */
HELPER_SPE_SINGLE_ARITH(fsadd);
@@ -1349,10 +1356,10 @@ HELPER_SPE_SINGLE_ARITH(fsmul);
HELPER_SPE_SINGLE_ARITH(fsdiv);
#define HELPER_SPE_VECTOR_ARITH(name) \
- uint64_t helper_ev##name(uint64_t op1, uint64_t op2) \
+ uint64_t helper_ev##name(CPUPPCState *env, uint64_t op1, uint64_t op2) \
{ \
- return ((uint64_t)e##name(op1 >> 32, op2 >> 32) << 32) | \
- (uint64_t)e##name(op1, op2); \
+ return ((uint64_t)e##name(env, op1 >> 32, op2 >> 32) << 32) | \
+ (uint64_t)e##name(env, op1, op2); \
}
/* evfsadd */
HELPER_SPE_VECTOR_ARITH(fsadd);
@@ -1364,7 +1371,7 @@ HELPER_SPE_VECTOR_ARITH(fsmul);
HELPER_SPE_VECTOR_ARITH(fsdiv);
/* Single-precision floating-point comparisons */
-static inline uint32_t efscmplt(uint32_t op1, uint32_t op2)
+static inline uint32_t efscmplt(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1373,7 +1380,7 @@ static inline uint32_t efscmplt(uint32_t op1, uint32_t op2)
return float32_lt(u1.f, u2.f, &env->vec_status) ? 4 : 0;
}
-static inline uint32_t efscmpgt(uint32_t op1, uint32_t op2)
+static inline uint32_t efscmpgt(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1382,7 +1389,7 @@ static inline uint32_t efscmpgt(uint32_t op1, uint32_t op2)
return float32_le(u1.f, u2.f, &env->vec_status) ? 0 : 4;
}
-static inline uint32_t efscmpeq(uint32_t op1, uint32_t op2)
+static inline uint32_t efscmpeq(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
CPU_FloatU u1, u2;
@@ -1391,28 +1398,28 @@ static inline uint32_t efscmpeq(uint32_t op1, uint32_t op2)
return float32_eq(u1.f, u2.f, &env->vec_status) ? 4 : 0;
}
-static inline uint32_t efststlt(uint32_t op1, uint32_t op2)
+static inline uint32_t efststlt(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
/* XXX: TODO: ignore special values (NaN, infinites, ...) */
- return efscmplt(op1, op2);
+ return efscmplt(env, op1, op2);
}
-static inline uint32_t efststgt(uint32_t op1, uint32_t op2)
+static inline uint32_t efststgt(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
/* XXX: TODO: ignore special values (NaN, infinites, ...) */
- return efscmpgt(op1, op2);
+ return efscmpgt(env, op1, op2);
}
-static inline uint32_t efststeq(uint32_t op1, uint32_t op2)
+static inline uint32_t efststeq(CPUPPCState *env, uint32_t op1, uint32_t op2)
{
/* XXX: TODO: ignore special values (NaN, infinites, ...) */
- return efscmpeq(op1, op2);
+ return efscmpeq(env, op1, op2);
}
-#define HELPER_SINGLE_SPE_CMP(name) \
- uint32_t helper_e##name(uint32_t op1, uint32_t op2) \
- { \
- return e##name(op1, op2) << 2; \
+#define HELPER_SINGLE_SPE_CMP(name) \
+ uint32_t helper_e##name(CPUPPCState *env, uint32_t op1, uint32_t op2) \
+ { \
+ return e##name(env, op1, op2) << 2; \
}
/* efststlt */
HELPER_SINGLE_SPE_CMP(fststlt);
@@ -1433,9 +1440,10 @@ static inline uint32_t evcmp_merge(int t0, int t1)
}
#define HELPER_VECTOR_SPE_CMP(name) \
- uint32_t helper_ev##name(uint64_t op1, uint64_t op2) \
+ uint32_t helper_ev##name(CPUPPCState *env, uint64_t op1, uint64_t op2) \
{ \
- return evcmp_merge(e##name(op1 >> 32, op2 >> 32), e##name(op1, op2)); \
+ return evcmp_merge(e##name(env, op1 >> 32, op2 >> 32), \
+ e##name(env, op1, op2)); \
}
/* evfststlt */
HELPER_VECTOR_SPE_CMP(fststlt);
@@ -1451,7 +1459,7 @@ HELPER_VECTOR_SPE_CMP(fscmpgt);
HELPER_VECTOR_SPE_CMP(fscmpeq);
/* Double-precision floating-point conversion */
-uint64_t helper_efdcfsi(uint32_t val)
+uint64_t helper_efdcfsi(CPUPPCState *env, uint32_t val)
{
CPU_DoubleU u;
@@ -1460,7 +1468,7 @@ uint64_t helper_efdcfsi(uint32_t val)
return u.ll;
}
-uint64_t helper_efdcfsid(uint64_t val)
+uint64_t helper_efdcfsid(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1469,7 +1477,7 @@ uint64_t helper_efdcfsid(uint64_t val)
return u.ll;
}
-uint64_t helper_efdcfui(uint32_t val)
+uint64_t helper_efdcfui(CPUPPCState *env, uint32_t val)
{
CPU_DoubleU u;
@@ -1478,7 +1486,7 @@ uint64_t helper_efdcfui(uint32_t val)
return u.ll;
}
-uint64_t helper_efdcfuid(uint64_t val)
+uint64_t helper_efdcfuid(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1487,7 +1495,7 @@ uint64_t helper_efdcfuid(uint64_t val)
return u.ll;
}
-uint32_t helper_efdctsi(uint64_t val)
+uint32_t helper_efdctsi(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1500,7 +1508,7 @@ uint32_t helper_efdctsi(uint64_t val)
return float64_to_int32(u.d, &env->vec_status);
}
-uint32_t helper_efdctui(uint64_t val)
+uint32_t helper_efdctui(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1513,7 +1521,7 @@ uint32_t helper_efdctui(uint64_t val)
return float64_to_uint32(u.d, &env->vec_status);
}
-uint32_t helper_efdctsiz(uint64_t val)
+uint32_t helper_efdctsiz(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1526,7 +1534,7 @@ uint32_t helper_efdctsiz(uint64_t val)
return float64_to_int32_round_to_zero(u.d, &env->vec_status);
}
-uint64_t helper_efdctsidz(uint64_t val)
+uint64_t helper_efdctsidz(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1539,7 +1547,7 @@ uint64_t helper_efdctsidz(uint64_t val)
return float64_to_int64_round_to_zero(u.d, &env->vec_status);
}
-uint32_t helper_efdctuiz(uint64_t val)
+uint32_t helper_efdctuiz(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1552,7 +1560,7 @@ uint32_t helper_efdctuiz(uint64_t val)
return float64_to_uint32_round_to_zero(u.d, &env->vec_status);
}
-uint64_t helper_efdctuidz(uint64_t val)
+uint64_t helper_efdctuidz(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
@@ -1565,7 +1573,7 @@ uint64_t helper_efdctuidz(uint64_t val)
return float64_to_uint64_round_to_zero(u.d, &env->vec_status);
}
-uint64_t helper_efdcfsf(uint32_t val)
+uint64_t helper_efdcfsf(CPUPPCState *env, uint32_t val)
{
CPU_DoubleU u;
float64 tmp;
@@ -1577,7 +1585,7 @@ uint64_t helper_efdcfsf(uint32_t val)
return u.ll;
}
-uint64_t helper_efdcfuf(uint32_t val)
+uint64_t helper_efdcfuf(CPUPPCState *env, uint32_t val)
{
CPU_DoubleU u;
float64 tmp;
@@ -1589,7 +1597,7 @@ uint64_t helper_efdcfuf(uint32_t val)
return u.ll;
}
-uint32_t helper_efdctsf(uint64_t val)
+uint32_t helper_efdctsf(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
float64 tmp;
@@ -1605,7 +1613,7 @@ uint32_t helper_efdctsf(uint64_t val)
return float64_to_int32(u.d, &env->vec_status);
}
-uint32_t helper_efdctuf(uint64_t val)
+uint32_t helper_efdctuf(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u;
float64 tmp;
@@ -1621,7 +1629,7 @@ uint32_t helper_efdctuf(uint64_t val)
return float64_to_uint32(u.d, &env->vec_status);
}
-uint32_t helper_efscfd(uint64_t val)
+uint32_t helper_efscfd(CPUPPCState *env, uint64_t val)
{
CPU_DoubleU u1;
CPU_FloatU u2;
@@ -1632,7 +1640,7 @@ uint32_t helper_efscfd(uint64_t val)
return u2.l;
}
-uint64_t helper_efdcfs(uint32_t val)
+uint64_t helper_efdcfs(CPUPPCState *env, uint32_t val)
{
CPU_DoubleU u2;
CPU_FloatU u1;
@@ -1644,7 +1652,7 @@ uint64_t helper_efdcfs(uint32_t val)
}
/* Double precision fixed-point arithmetic */
-uint64_t helper_efdadd(uint64_t op1, uint64_t op2)
+uint64_t helper_efdadd(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1654,7 +1662,7 @@ uint64_t helper_efdadd(uint64_t op1, uint64_t op2)
return u1.ll;
}
-uint64_t helper_efdsub(uint64_t op1, uint64_t op2)
+uint64_t helper_efdsub(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1664,7 +1672,7 @@ uint64_t helper_efdsub(uint64_t op1, uint64_t op2)
return u1.ll;
}
-uint64_t helper_efdmul(uint64_t op1, uint64_t op2)
+uint64_t helper_efdmul(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1674,7 +1682,7 @@ uint64_t helper_efdmul(uint64_t op1, uint64_t op2)
return u1.ll;
}
-uint64_t helper_efddiv(uint64_t op1, uint64_t op2)
+uint64_t helper_efddiv(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1685,7 +1693,7 @@ uint64_t helper_efddiv(uint64_t op1, uint64_t op2)
}
/* Double precision floating point helpers */
-uint32_t helper_efdtstlt(uint64_t op1, uint64_t op2)
+uint32_t helper_efdtstlt(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1694,7 +1702,7 @@ uint32_t helper_efdtstlt(uint64_t op1, uint64_t op2)
return float64_lt(u1.d, u2.d, &env->vec_status) ? 4 : 0;
}
-uint32_t helper_efdtstgt(uint64_t op1, uint64_t op2)
+uint32_t helper_efdtstgt(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1703,7 +1711,7 @@ uint32_t helper_efdtstgt(uint64_t op1, uint64_t op2)
return float64_le(u1.d, u2.d, &env->vec_status) ? 0 : 4;
}
-uint32_t helper_efdtsteq(uint64_t op1, uint64_t op2)
+uint32_t helper_efdtsteq(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
CPU_DoubleU u1, u2;
@@ -1712,20 +1720,20 @@ uint32_t helper_efdtsteq(uint64_t op1, uint64_t op2)
return float64_eq_quiet(u1.d, u2.d, &env->vec_status) ? 4 : 0;
}
-uint32_t helper_efdcmplt(uint64_t op1, uint64_t op2)
+uint32_t helper_efdcmplt(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
/* XXX: TODO: test special values (NaN, infinites, ...) */
- return helper_efdtstlt(op1, op2);
+ return helper_efdtstlt(env, op1, op2);
}
-uint32_t helper_efdcmpgt(uint64_t op1, uint64_t op2)
+uint32_t helper_efdcmpgt(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
/* XXX: TODO: test special values (NaN, infinites, ...) */
- return helper_efdtstgt(op1, op2);
+ return helper_efdtstgt(env, op1, op2);
}
-uint32_t helper_efdcmpeq(uint64_t op1, uint64_t op2)
+uint32_t helper_efdcmpeq(CPUPPCState *env, uint64_t op1, uint64_t op2)
{
/* XXX: TODO: test special values (NaN, infinites, ...) */
- return helper_efdtsteq(op1, op2);
+ return helper_efdtsteq(env, op1, op2);
}