diff options
author | Blue Swirl <blauwirbel@gmail.com> | 2011-07-03 10:42:23 +0000 |
---|---|---|
committer | Blue Swirl <blauwirbel@gmail.com> | 2011-10-23 15:09:13 +0000 |
commit | 2e2f4ade86e335ea7ffa2b5f32e3ad50619fb4e5 (patch) | |
tree | 0b752e454f81d7b63acaef2965fcd835f317ba28 /target-sparc/fop_helper.c | |
parent | 1bccec25e10e9073e21c2f71cd16a0b3ffe06c39 (diff) | |
download | qemu-2e2f4ade86e335ea7ffa2b5f32e3ad50619fb4e5.zip qemu-2e2f4ade86e335ea7ffa2b5f32e3ad50619fb4e5.tar.gz qemu-2e2f4ade86e335ea7ffa2b5f32e3ad50619fb4e5.tar.bz2 |
Sparc: avoid AREG0 for float and VIS ops
Make floating point and VIS ops take a parameter for CPUState instead
of relying on global env.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Diffstat (limited to 'target-sparc/fop_helper.c')
-rw-r--r-- | target-sparc/fop_helper.c | 70 |
1 files changed, 35 insertions, 35 deletions
diff --git a/target-sparc/fop_helper.c b/target-sparc/fop_helper.c index ddd0af9..23502f3 100644 --- a/target-sparc/fop_helper.c +++ b/target-sparc/fop_helper.c @@ -18,7 +18,6 @@ */ #include "cpu.h" -#include "dyngen-exec.h" #include "helper.h" #define DT0 (env->dt0) @@ -26,10 +25,11 @@ #define QT0 (env->qt0) #define QT1 (env->qt1) -#define F_HELPER(name, p) void helper_f##name##p(void) +#define F_HELPER(name, p) void helper_f##name##p(CPUState *env) #define F_BINOP(name) \ - float32 helper_f ## name ## s (float32 src1, float32 src2) \ + float32 helper_f ## name ## s (CPUState * env, float32 src1,\ + float32 src2) \ { \ return float32_ ## name (src1, src2, &env->fp_status); \ } \ @@ -48,14 +48,14 @@ F_BINOP(mul); F_BINOP(div); #undef F_BINOP -void helper_fsmuld(float32 src1, float32 src2) +void helper_fsmuld(CPUState *env, float32 src1, float32 src2) { DT0 = float64_mul(float32_to_float64(src1, &env->fp_status), float32_to_float64(src2, &env->fp_status), &env->fp_status); } -void helper_fdmulq(void) +void helper_fdmulq(CPUState *env) { QT0 = float128_mul(float64_to_float128(DT0, &env->fp_status), float64_to_float128(DT1, &env->fp_status), @@ -80,23 +80,23 @@ F_HELPER(neg, q) #endif /* Integer to float conversion. */ -float32 helper_fitos(int32_t src) +float32 helper_fitos(CPUState *env, int32_t src) { return int32_to_float32(src, &env->fp_status); } -void helper_fitod(int32_t src) +void helper_fitod(CPUState *env, int32_t src) { DT0 = int32_to_float64(src, &env->fp_status); } -void helper_fitoq(int32_t src) +void helper_fitoq(CPUState *env, int32_t src) { QT0 = int32_to_float128(src, &env->fp_status); } #ifdef TARGET_SPARC64 -float32 helper_fxtos(void) +float32 helper_fxtos(CPUState *env) { return int64_to_float32(*((int64_t *)&DT1), &env->fp_status); } @@ -114,64 +114,64 @@ F_HELPER(xto, q) #undef F_HELPER /* floating point conversion */ -float32 helper_fdtos(void) +float32 helper_fdtos(CPUState *env) { return float64_to_float32(DT1, &env->fp_status); } -void helper_fstod(float32 src) +void helper_fstod(CPUState *env, float32 src) { DT0 = float32_to_float64(src, &env->fp_status); } -float32 helper_fqtos(void) +float32 helper_fqtos(CPUState *env) { return float128_to_float32(QT1, &env->fp_status); } -void helper_fstoq(float32 src) +void helper_fstoq(CPUState *env, float32 src) { QT0 = float32_to_float128(src, &env->fp_status); } -void helper_fqtod(void) +void helper_fqtod(CPUState *env) { DT0 = float128_to_float64(QT1, &env->fp_status); } -void helper_fdtoq(void) +void helper_fdtoq(CPUState *env) { QT0 = float64_to_float128(DT1, &env->fp_status); } /* Float to integer conversion. */ -int32_t helper_fstoi(float32 src) +int32_t helper_fstoi(CPUState *env, float32 src) { return float32_to_int32_round_to_zero(src, &env->fp_status); } -int32_t helper_fdtoi(void) +int32_t helper_fdtoi(CPUState *env) { return float64_to_int32_round_to_zero(DT1, &env->fp_status); } -int32_t helper_fqtoi(void) +int32_t helper_fqtoi(CPUState *env) { return float128_to_int32_round_to_zero(QT1, &env->fp_status); } #ifdef TARGET_SPARC64 -void helper_fstox(float32 src) +void helper_fstox(CPUState *env, float32 src) { *((int64_t *)&DT0) = float32_to_int64_round_to_zero(src, &env->fp_status); } -void helper_fdtox(void) +void helper_fdtox(CPUState *env) { *((int64_t *)&DT0) = float64_to_int64_round_to_zero(DT1, &env->fp_status); } -void helper_fqtox(void) +void helper_fqtox(CPUState *env) { *((int64_t *)&DT0) = float128_to_int64_round_to_zero(QT1, &env->fp_status); } @@ -183,34 +183,34 @@ float32 helper_fabss(float32 src) } #ifdef TARGET_SPARC64 -void helper_fabsd(void) +void helper_fabsd(CPUState *env) { DT0 = float64_abs(DT1); } -void helper_fabsq(void) +void helper_fabsq(CPUState *env) { QT0 = float128_abs(QT1); } #endif -float32 helper_fsqrts(float32 src) +float32 helper_fsqrts(CPUState *env, float32 src) { return float32_sqrt(src, &env->fp_status); } -void helper_fsqrtd(void) +void helper_fsqrtd(CPUState *env) { DT0 = float64_sqrt(DT1, &env->fp_status); } -void helper_fsqrtq(void) +void helper_fsqrtq(CPUState *env) { QT0 = float128_sqrt(QT1, &env->fp_status); } #define GEN_FCMP(name, size, reg1, reg2, FS, E) \ - void glue(helper_, name) (void) \ + void glue(helper_, name) (CPUState *env) \ { \ env->fsr &= FSR_FTT_NMASK; \ if (E && (glue(size, _is_any_nan)(reg1) || \ @@ -246,7 +246,7 @@ void helper_fsqrtq(void) } \ } #define GEN_FCMPS(name, size, FS, E) \ - void glue(helper_, name)(float32 src1, float32 src2) \ + void glue(helper_, name)(CPUState *env, float32 src1, float32 src2) \ { \ env->fsr &= FSR_FTT_NMASK; \ if (E && (glue(size, _is_any_nan)(src1) || \ @@ -318,7 +318,7 @@ GEN_FCMP(fcmpeq_fcc3, float128, QT0, QT1, 26, 1); #endif #undef GEN_FCMPS -void helper_check_ieee_exceptions(void) +void helper_check_ieee_exceptions(CPUState *env) { target_ulong status; @@ -352,12 +352,12 @@ void helper_check_ieee_exceptions(void) } } -void helper_clear_float_exceptions(void) +void helper_clear_float_exceptions(CPUState *env) { set_float_exception_flags(0, &env->fp_status); } -static inline void set_fsr(void) +static inline void set_fsr(CPUState *env) { int rnd_mode; @@ -379,16 +379,16 @@ static inline void set_fsr(void) set_float_rounding_mode(rnd_mode, &env->fp_status); } -void helper_ldfsr(uint32_t new_fsr) +void helper_ldfsr(CPUState *env, uint32_t new_fsr) { env->fsr = (new_fsr & FSR_LDFSR_MASK) | (env->fsr & FSR_LDFSR_OLDMASK); - set_fsr(); + set_fsr(env); } #ifdef TARGET_SPARC64 -void helper_ldxfsr(uint64_t new_fsr) +void helper_ldxfsr(CPUState *env, uint64_t new_fsr) { env->fsr = (new_fsr & FSR_LDXFSR_MASK) | (env->fsr & FSR_LDXFSR_OLDMASK); - set_fsr(); + set_fsr(env); } #endif |